Citation |

- Permanent Link:
- http://ufdc.ufl.edu/AA00011802/00001
## Material Information- Title:
- Comparison of males and females on math item performance analysis of response patterns
- Creator:
- Feliciano, Sonia, 1938-
- Publication Date:
- 1986
- Language:
- English
- Physical Description:
- x, 117 leaves : ill. ; 28 cm.
## Subjects- Subjects / Keywords:
- Arithmetic ( jstor )
College mathematics ( jstor ) Educational research ( jstor ) Learning ( jstor ) Mathematical variables ( jstor ) Mathematics ( jstor ) Mathematics achievement ( jstor ) Mathematics education ( jstor ) Problem solving ( jstor ) Sex linked differences ( jstor ) Dissertations, Academic -- Foundations of Education -- UF Foundations of Education thesis Ph. D Mathematical ability ( lcsh ) Sex differences ( lcsh ) - Genre:
- bibliography ( marcgt )
non-fiction ( marcgt )
## Notes- Thesis:
- Thesis (Ph. D.)--University of Florida, 1986.
- Bibliography:
- Bibliography: leaves 107-115.
- General Note:
- Typescript.
- General Note:
- Vita.
- Statement of Responsibility:
- by Sonia Feliciano.
## Record Information- Source Institution:
- University of Florida
- Holding Location:
- University of Florida
- Rights Management:
- Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
- Resource Identifier:
- 000887339 ( ALEPH )
15167202 ( OCLC ) AEJ5569 ( NOTIS )
## UFDC Membership |

Downloads |

## This item has the following downloads: |

Full Text |

COMPARISON OF MALES AND FEMALES ON MATH ITEM PERFORMANCE: ANALYSIS OF RESPONSE PATTERNS By SONIA FELICIANO A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1986 ACKNOWLEDGMENTS I would like to acknowledge my gratitude to several people who have influenced my formal education and/or made this study possible. My special thanks to Dr. James Algina, Chairman of the doctoral committee who contributed to the development of my love for research and statistics. He has been insuperable as professor and valued friend; his help and guidance in the preparation and completion of this study were invaluable. I extend my thanks to Dr. Linda Crocker for her advice and help during my doctoral studies at University of Florida. Thanks also go to Dr. Michael Nunnery, member of the doctoral committee. To Dr. Wilson Guertin, who was a friend for me and my family, I extend my special thanks. Thanks are also extended to Dr. Amalia Charneco, past Undersecretary of Education of the Puerto Rico Department of Education, for her continuous support. To my sister Nilda Santaellar who typed the thesis, I give my sincere thanks. Special thanks go to my family and to those friends who provided encouragement throughout this critical period of my life. TABLE OF CONTENTS Page ACKNOWLEDGMENTS.... .......................... ii LIST OF TABLES .. ................ ................... v LIST OF FIGURES...................... ............... vii ABSTRACT....... ...................................... viii CHAPTER I INTRODUCTION.................................. 1 Purpose of the Study ............... ........ 6 Significance of the Study................... 6 Organization of the Study................... 8 II REVIEW OF THE LITERATURE ...................... 9 Sex-related Differences in Incorrect Response Patterns........................ 10 Sex-related Differences in Problem Solving............... ..... ........... .... 14 Cognitive and Affective Variables that Influence Sex Differences in Mathematics Learning and Achievement................ 23 Differences in Formal Mathematics Education....................... 24 Differences in Spatial Ability.......... 26 Differentiated Effect of Affective Variables................... ....... ... 30 Problem Solving Performance and Related Variables.... ..................... ......... 37 Computational Skills and Problem Solving Performance. .......................... 38 Reading and Problem Solving Perfor- mance........................... .... .. 44 Attitudes Toward Problem Solving and Problem Solving Performance........... 50 iii III METHOD..... ... ...... ...... ..... ............... 54 The Sample ................ ................... 54 The Instrument............. ................ 55 Analysis of the Data....................... 57 Analysis of Sex by Option by Year Cross Classifications................ 57 Comparison of Males and Females in Problem Solving Performance........... 66 IV RESULTS........... ..... ....... ...... .. ........ 68 Introduction ............................. 68 Sex-related Differences in the Selection of Incorrect Responses.......... 68 Sex-related Differences in Problem Solving Performance.................... 81 Summary.................................... 93 V DISCUSSION..................................... 101 Summary and Interpretation of the Results ................................... 101 Implications of the Findings and Suggestions for Further Research.......... 103 Sex-related Differences in Incorrect Responses................ 104 Sex-related Differences in Problem Solving Performance................... 105 REFERENCES.......................................... 107 BIOGRAPHICAL SKETCH................................ 116 LIST OF TABLES Table Page 3.1 Hypothetical Probabilities of Option Choice Conditional on Year and Sex.......... 58 3.2 Hypothetical Probabilities of Option Choice Conditional on Year and Sex and Arranged by Sex......... ...... ........... 58 3.3 Hypothetical Joint Probabilities of Yearr Option, and Sex............................. 64 3.4 Hypothetical Probabilities of Option Choice Conditional on Sex and Year.......... 66 4.1 Summary of Significant Tests of Model 3.2..... 71 4.2 Chi-square Values for the Comparison of 75 Models (3.2) and (3.3) with Actual Sample Sizes and with the Corresponding Number of Subjects Needed for Significant Results 75 4.3 Means, Standard Deviations, and t-Test for the Eight Mathematical Variables............ 82 4.4 ANCOVA First Year: Multiplication Covariate 85 4.5 ANCOVA First Year: Division Covariate....... 85 4.6 ANCOVA Second Year: Subtraction Covariate... 86 4.7 Reliability of the Covariates for Each of the Three Years of Test Administra- tion........................................ 91 4.8 ANCOVA First Year: Other Covariates......... 92 4.9 Adjusted Means on Problem Solving, by Covariate and Sex.......................... 94 4.10 ANCOVA Second Year: Other Covariates........ 95 Table Page 4.11 Adjusted Means on Problem Solving, by Covariate and Sex........................ 96 4.12 ANCOVA Third Year........................ 97 4.13 Adjusted Means on Problem Solving, by Covariate and Sex....................... 98 LIST OF FIGURES Figure Page 3.1 Number of Questions by Skill Area in a Basic Skills Test in Mathematics-6.......... 56 4.1 Sex by Multiplication Interaction............. 87 4.2 Sex by Division Interaction ................ 88 4.3 Sex by Subtraction Interaction................ 89 vii Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy COMPARISON OF MALES AND FEMALES ON MATH ITEM PERFORMANCE: ANALYSIS OF RESPONSE PATTERNS By Sonia Feliciano August, 1986 Chairman: James Algina Major Department: Foundations of Education The first objective of this study was to investigate sex differences in the selection of incorrect responses on a mathematics multiple-choice test, and to determine whether these differences were consistent over three consecutive administrations of the test. A second objective was to compare male and female performance in problem solving after controlling for computational skills. The responses of all 6th grade students from the public schools in Puerto Rico who took the Basic Skills Test in Mathematics-6 ("Prueba de Destrezas B&sicas en Matem&ticas-6") during three academic years were used in the analyses relevant to the first objective. viii Log-linear models were used in the analysis of incorrect responses. The results of the analyses showed that for 100 of the 111 items of the test, males and females selected different incorrect options, and this pattern of responses was consistently found during the three years of test administration. However, for the vast majority of the 100 items the male-female differences were relatively small, considering the fact that the number of subjects needed to obtain statistical significance was very large. The responses of approximately 1,000 randomly selected students per academic year were analyzed in the comparison of male and female performance in problem solving. Females outperformed males in problem solving and in six of the seven computational variables. Males showed superiority in equivalence in all the three years, but statistical significance was obtained in only one of the years. Analysis of covariance (ANCOVA) was used in the comparison of male and female performance in problem solving after controlling for computational skills. Seven analyses of covariance tests were conducted, one for each of the covariates. Estimated true scores for observed scores were used in the analyses. The results tend to show that for examinees with similar levels of computational skills, sex-related differences in problem solving performance do not exist. Females retained their superiority in problem solving when equivalence (in all three years) and subtraction (in one year) were the controlling variables. The question of whether male-female differences in problem solving depend on computational skills was answered, partially, in the affirmative. CHAPTER I INTRODUCTION Sex differences in mathematics learning and achievement have been the subject of intensive research. Research done before 1974 has shown that male performance on mathematical achievement tests is superior to female performance by the time they reach upper elementary or junior high school (Fennema, 1976, p. 2). The literature strongly suggests that at the elementary level females outperform males in computation and males excel in mathematical reasoning (Glennon & Callahan, 1968; Jarvis, 1964; Maccoby, 1966). Since 1974, research findings have been less consistent. Fennema (1974), after reviewing 36 studies, found that during secondary school or earlier, sex-related differences in mathematics achievement are not so evident, but that when differences are founds they favor males in high level cognitive tasks (problem solving) and females in low level cognitive tasks (computation). As a result of a further review of the literature, Fennema (1977) concluded that at the elementary level, sex-related differences do not exist at all cognitive levels, from computation to problem solving. Many variables, cognitive affective, and educational, have been investigated since 1974 in relation to sex differences in mathematics learning and achievement. Fennema and Sherman (1977) investigated the effect of differential formal mathematics education. After controlling for the number of years of exposure to the study of mathematics, they found sex differences in only two of the four schools under study. However, in those schools where boys scored higher than girls, differences were also found in their attitudes toward mathematics. Hilton and Berglund (1974) found significant sex differences after controlling for the number of mathematics courses taken, and attributed them to sex differences in interests. "As the boys' interests in science increase related to the girls', their achievement in mathematics increases relative to that of the girls" (p. 234). Wise, Steel, and McDonald (1979) reanalyzed test data collected in a longitudinal study of 400,000 high school students (Project Talent). They found that when the effect of the number of high school mathematics courses was not controlled, no sex differences emerged for 9th graders, but that gains made by boys during the next three years were more than twice that of the girls. These differences between the sexes disappeared when the number of mathematics courses taken was controlled. Results of the 1978 Women and Mathematics National Survey, Survey I, indicated no significant sex differences for 8th grade students on measures of problem solving or algebra. However, females outperformed males in computation and spatial visualization. For the 12th grade students, statistically significant sex differences favoring males were found in problem solving, but not in algebra, computation, or spatial visualization. For males and females who had enrolled in courses beyond general mathematics and who had taken or were enrolled in courses such as pre-calculus, calculus, or geometry, differences in problem solving or spatial visualization did not exist. Sex differences favoring males were found on a total score obtained summing across the computation, problem solving, and algebra subtests (Armstrong, 1979). The mathematics data collected in the second survey by the 1978 National Assessment of Educational Progress showed significant sex differences for both 13- and 17-year-old students. The 13-year-old females outperformed males in the computational subtest and males outscored females by 1 1/2 percentage points in problem solving (statistically significant). No statistically significant differences were found in algebra. No sex differences were found for the 17-year-old group either in the computation subtest or in the algebra subtest. Males surpassed females in problem solving. A reanalysis of the data from the 17-year-old group confirmed male superiority in problem solving after controlling for mathematics preparation. Males who were enrolled or had completed algebra II outperformed the females in computation and problem solving but not in algebra. Males who studied beyond algebra II outscored females on all three subtests: computation, algebra, and problem solving (Armstrong, 1979). Carpenter, Lindquist, Mathews, and Silver (1984) analyzed the results of the Third National Assessment of Educational Progress (NAEP), and compared them with the First and Second Surveys. Between 1978 and 1982, the differences between the average performance of males and females remained stable at each age level. At ages 9 and 13, the overall performance of males and females was not significantly different. At age 17r males scored higher than females by about 3 percentage points. When course background was held constant, achievement differences still existed at age 17. For each category of course background, male achievement exceeded female achievement. Consistent with previous assessments, sex differences in problem solving in favor of males were found for the 17-year-old sample. At ages 9 and 13, no large differences were found between the sexes within any level of course background. Marshall (1981, 1984) investigated sex differences in mathematics performance. She found that males and females excel each other in solving different types of problems. Females were better on items of computation and males were more successful on word-story problem items (problem solving). She also found that females successful performance in the problem solving items was more dependent on their successful performance in the computation items. Males did not need, as much as females, to succeed in the computation items in order to answer correctly the problem solving items. Although the general findings seem to support sex differences in mathematics learning and achievement, the research done does not consistently support superiority for either sex. Most of the research has been concerned with how the sexes differ on subtests or total test scores in mathematics. Moreover, the great majority of the studies deal with correct responses. Sex differences in incorrect responses at the item level have not been fully researched. Only two studies dealing with sex differences in incorrect responses at the item level were found in the research literature (Marshall, 1981, 1983). Marshall investigated whether boys and girls made similar errors in computation and story problems. She analyzed boys' and girls' answers to six mathematics items and found that the sexes made different errors, possibly reflecting different problem solving strategies. Her original findings were supported when she studied the same problem using a large number of items three years later. Purpose of the Study The purpose of this study was twofold. The first was to investigate sex-related differences in the selection of incorrect responses in a multiple-choice achievement test in mathematics. For each test item, the research questions were as follows: 1. Is there a difference in the proportion of males and females choosing each incorrect option? 2. Is the same pattern of differences found in data obtained in three different administrations of the test? The second objective was to investigate sex-related differences in test scores in mathematics problem solving. The following questions were studied: 1. Do males and females differ in problem solving performance? 2. Do sex differences in problem solving persist after controlling for computational skills, and does the differential success of males and females on problem solving items depend on their success on the computation items? Significance of the Study Item response patterns are very useful techniques in the assessment of mathematics learning and achievement. Total test scores can be very misleading in the assessment of student performance and provide no diagnostic information about the nature and seriousness of student errors (Harnisch, 1983). Item response patterns are valuable for the identification of large group differences, including district-to-district, school-to-school, and classroom-to-classroom variations on different subsets of items. The response patterns can provide diagnostic information about the type of understanding the student has on various mathematics topics (e.g.r problem solving). Marshall (1981, 1984) has used the item response pattern technique and her findings indicate sex differences in mathematics performance at the item level. Females outperformed males in computation and males outscored females in problem solving. Also the success of girls in the problem solving items was dependent upon their success in the computation items; for boys, success in the problem solving items did not depend as much on their computational performance. Marshall (1981, 1983) has also reported that males and females differ in the selection of incorrect responses, reflecting differences in reasoning abilities. In Puerto Rico, a high percentage of children promoted to the 7th grade in the public schools does not master the basic skills in mathematics. If 6th grade male and female children can be diagnosed as having different problem solving abilities, as Marshall found with California children, teachers may need to provide tailor-made mathematics instruction for each sex, in order to ensure equal access to formal education and enhance mathematics achievement. Since there are no investigations reported in sex differences in item response patterns in Puerto Rico, research is needed. Organization of the Study A review of the literature on sex differences in mathematics performance is reported in Chapter II. The research methodology is presented in Chapter III. Research questions, sample, instrument, and data analysis are discussed in that chapter. Chapter IV is an exposition of the results of the study. Chapter V contains a summary and interpretation of the results of the study and the implications of the findings together with suggestions for further research. CHAPTER II REVIEW OF THE LITERATURE Sex differences in mathematics learning and achievement have been a subject of concern for educators and psychologists. Many studies found in the literature support the existence of these differences. Boys show superiority in higher level cognitive tasks (problem solving or mathematical reasoning) in the upper elementary years and in the early high school years (Fennemar 1974; Maccoby & Jacklin, 1974). Almost all the research carried out has dealt with analysis of total correct scores in mathematics aptitude and achievement tests or scores in subtests. The literature related to sex differences in incorrect responses, the main subject of the present study, is surprisingly sparse. For the most part, the studies have investigated the differences between the sexes in mathematics learning and achievement and the underlying variables causing the differences. Cognitive and affective variables have been the matter at issue in the establishment of sex differences. Although research in mathematics problem solving, the secondary subject of this investigation, is extensive, most of the studies consider sex differences incidental to the major study findings. The available literature offers very little research directly related to the problem of sex differences in this area. The review of the literature has been divided in four sections. The first section consists of a detailed summary of the available research on sex differences in incorrect responses. The second section deals with sex-related differences in problem solving performance. These sections are directly related to the objectives of the study. The third section is more peripheral, and contains a discussion of the more prevalent issues about the influence of cognitive and affective variables on sex differences in mathematics learning and achievement. The fourth is a summary of the research dealing with variables considered as influential to mathematics problem solving performance. Sex-related Differences in Incorrect Response Patterns Research findings tend to suggest that boys and girls may be approaching problem solving differently (e.g., Fennema and Sherman, 1978; Marshall, 1981, 1983; Meyer, 1978, among others). Marshall (1981) investigated whether 6th grade boys and girls approach mathematical problem solving with different strategies. Her specific interest was whether the sexes made the same errors. She analyzed the responses of 9,000 boys and 9,000 girls to 6 selected items, 2 computation items, and 1 story problem item from each of 2 of the 16 forms of the Survey of Basic Skills test administered during the academic year 1978-79. The Survey is a 30-item achievement test administered every year to all 6th grade children in California through the California Assessment Program. There are 16 forms of the test, to which approximately 9,000 boys and 9,000 girls respond each year. Of the 160 mathematics items contained in the 16 forms of the test, 32 are on measurement and graphing, 28 on number concepts, 28 on whole number arithmetics, 20 on fraction arithmetics, 20 on decimal arithmetics, 20 on geometry, and 12 on probability and statistics. The item analysis performed on the 1978-79 data showed that boys and girls tended to select different incorrect responses. In the first computation item (Form 1 of the test) both sexes reflected similar mistakes in carrying, but in different columns. In the second computation item, both sexes ignored the decimal points and selected the same incorrect response. However, more girls than boys chose this response. In the first computation item (Form 2 of the test) the incorrect choice of both sexes was option cr but the second most frequently selected option was a for boys and b for girls. In the second computation item of this test form, no sex differences were found in response patterns. Approximately 45% of each sex selected option c. The next popular choice for both sexes was option d, selected by approximately 35% of both boys and girls. On the story problem of Form 1, males and females responded alike. Their most popular incorrect response choice was option a for both males and females. The second most popular incorrect choice was option c for both sexes. Response to the story problem in Form 2 showed sex differences in response choice. Including the correct option, 33% of the girls selected option a, 20% chose option c, and 20% option d. For males approximately 25% selected option a and the same percent selected option d. Marshall concluded that although the analysis of incorrect responses does not explain why boys and girls differ in their responses, the analysis shows that boys and girls approach problems in different ways and these varying strategies can be useful in identifying how the sexes differ in reasoning abilities. Two years later, Marshall (1983) analyzed the responses of approximately 300,000 boys and girls to mathematics items contained in the 16 test forms of the Survey of Basic Skills during the years 1977, 1978, and 1979. She used log-linear models (explained in Chapter III) to investigate sex related differences in the selection of incorrect responses, and the consistency of such differences over three years of administration of the test. Based on her findings that sex differences were found in 80% of the items, Marshall classified the students' errors according to Radatz' (1979) five-category error classification. The categories are language (errors in semantics), spatial visualization, mastery, association, and use of irrelevant rules. It was found that girls' errors are more likely to be due to the misuse of spatial information, the use of irrelevant rules, or the choice of an incorrect operation. Girls also make relatively more errors of negative transfer and keyword association. Boys seem more likely than girls to make errors of perseverance and formula interference. Both sexes make language-related errors, but the errors are not the same. Available research is not extensive enough to make definite judgments about the sex-related differences observed in incorrect responses. Clearly more research is needed. Sex-related Differences in Problem Solving It has already been acknowledged that the subject of problem solving has been extensively researched. However, as early as 1969, Kilpatrick criticized the fact that the study of problem solving has not been systematic; some researchers have studied the characteristics of the problem while others have given their attention to the characteristics of the problem solvers. Moreover, differences in the tests used to measure problem solving performance also constitute an obstacle when trying to compare the results of the studies carried out. In order to avoid this pitfall and provide a basis for comparison, the studies reviewed in this section, dealing with sex-related differences in problem solving have been divided in two groups. The first comprises those studies that used the Romberg-Wearne Problem Solving Test. The second contains other relevant studies in which problem solving performance has been measured by means of other instruments. The Romberg-Wearne Problem Solving Test merits special mention because it was the first attempt "to develop a test to overcome the inadequacies of total test scores in explaining the reasons why some students are successful problem solvers and others are not" (Whitaker, 1976, pp. 9, 10). The test is composed of 23 items designed to yield 3 scores: a comprehension score, an application score, and a problem solving score. The comprehension question ascertains whether a child understands the information given in the item stem. The application question assesses the child's mastery of a prerequisite concept or skill of the problem solving question. The problem solving question poses a question whose solution is not immediately available, that is, a situation which does not lend itself to the immediate application of a rule or algorithm. The application and problem solving parts of the test may refer to a common unit of information (the item stem) but the questions are independent in that the response to the application question is not used to respond to the problem solving question. Meyer (1978) Whitaker (1976), Pennema and Sherman (1978), and Sherman (1979) have used the Romberg-Wearne test in their studies. Meyer (1978) investigated whether males and females differ in problem solving performance and examined their prerequisite computational skills and mathematical concepts for the problem solving questions. A sample of 179 students from the 4th grade were administered 19 "reference tests" for intellectual abilities and the Romberg-Wearne test. The analysis showed that males and females were not significantly different in the comprehension, application, and problem solving questions of the test. The sexes differed in only 2 of the 19 reference tests, Spatial Relations and Picture Group Name-Selection. A factor analysis, however, showed differences in the number and composition of the factors. For females, a general mathematics factor was determined by mathematics computation, comprehension, application, and problem solving. For males, the comprehension and application parts determined one factor; problem solving with two other reference tests (Gestalt and Omelet) determined another factor. Meyer concluded that comprehension of the data and mastery of the prerequisite mathematical concepts did not guarantee successful problem solving either for males or for females. Problem solving scores for both sexes were about one third their scores in comprehension and one fourth their scores in application. She also concluded that the sexes may have approached the problem solving questions differently. The methods used by females for solving problem situations may have paralleled their approach to the application parts. Males may have used established rules and algorithms for the application parts, but may have used more of a Gestalt approach to the problem solving situation. Whitaker (1976) investigated the relationship between the mathematical problem performance of 4th grade children and their attitudes toward problem solving, their teachers' attitude toward mathematical problem solving, and related sex and program-type differences. Although his main objective was to construct an attitude scale to measure attitudes toward problem solving, his study is important because his findings support Meyer's regarding the lack of significant sex-related differences in problem solving performance. Performance in the problem solving questions, for both males and females, was much lower than performance in the application questions, and much lower than performance in the comprehension questions. In fact, the mean score for each part of the item, for both males and females, was almost identical to the mean scores obtained by males and females in Meyer's study. Whitaker noted that each application item is more difficult than its preceding comprehension item, and that each problem solving item is more difficult than its preceding application item. No significant sex-related differences were found for any of the three parts of the item (comprehension, application, or problem solving). Fennema and Sherman (1978) investigated sex-related differences in mathematics achievement and cognitive and affective variables related to such differences. They administered the Romberg-Wearne Problem Solving Test to a representative sample of 1320 students (grades 6-8) from Madison, Wisconsin, predominantly middle-class, but including great diversity in SES. The sample consisted of students who had taken a similar number of mathematics courses and were in the top 85% of the class in mathematics achievement. They were tested in 1976. Four high school districts were included. In only one of the high school districts were sex-related differences in application and problem solving founds in favor of males. They concluded that when relevant factors are controlled, sex-related differences in favor of males do not appear often, and when they dor they are not large. Sherman (1980) investigated the causes of the emerging sex-related differences in mathematics performance, in favor of males, during adolescence (grades 8-11). She wanted to know if these differences emerge as a function of sex-related differences in spatial visualization and sociocultural influences that consider math as a male domain. In grade 8, she used the Romberg-Wearne Test and, in grade 11, a mathematical problem solving test derived from the French Kit of Tests. The analysis showed that for girls, problem solving performance remained stable across the years. Mean problem solving performance for boys, however, was higher in grade 11 than in grade 8. No sex-related differences were found in grade 8, but boys outperformed girls in grade 11, where the Stafford test was used. Sherman found that for both sexes problem solving performance in grade 8 was the best predictor of problem solving performance in grade 11. Spatial visualization was a stronger predictor for girls than for boys. Mathematics as a male domain was a good predictor for girls only; the less a girl stereotyped mathematics as a male domain in grade 8, the higher her problem solving score in grade 11. Attitudes toward success in mathematics in grade 8 was a more positive predictor of problem solving performance for boys than for girls; the more positive the attitudes toward success in mathematics in grade 8, the higher their performance in problem solving in grade 11. None of these four studies, all of which used the Romberg-Wearne Mathematics Problem Solving Test, show statistically significant sex-related differences in problem solving performance. In later studies other tests were used to measure this variable (Kaufman, 1984; Marshall, 1981, 1984). Kaufman (1983/1984) investigated if sex differences in problem solving, favoring males, exist in the 5th and 6th grades and if these differences were more pronounced in mathematically gifted students than in students of average mathematical ability. The Iowa Test of Basic Skills and a mathematics problem-solving test were administered to 504 subjects. Males in the average group as well as males in the gifted group outperformed females, but only the gifted group showed statistically significant differences. As a result of her investigations, Marshall (1981) concluded that sex-related differences in mathematics performance may be the result of comparing the sexes on total test scores. If the test contains more computation items than problem solving items, girls will perform better than boys, but if the test contains more problem solving items than computation ones, boys will outperform girls. With this in mind, Marshall investigated sex-related differences in computation and problem solving by analyzing the responses of approximately 18,000 students from grade 6 who had been administered the Survey of Basic Skills Test: Grade 6, during the academic year 1978-79. Two of the 16 test forms of the Survey were used to assess skills such as concepts of whole numbers, fractions, and decimals. These skills were tested both as simple computations and as story problems (problem solving). Two computation items and one story problem item were selected because they were particularly related; both computation items required skills needed in solving the corresponding story problem. It was assumed that correct solution of the computation item correlates with solving the story problem because the story problem requires a similar computation. Marshall found that girls were better in computation and boys were better in problem solving. She also found that boys were much more likely than girls to answer the story problem item correctly after giving incorrect responses to both computation items. Apparently, mastery of the skills required by the computation items is more important for girls than for boys. If girls cannot solve the computation items, they have little chance of solving the related story problem item. For girls, the probability of success in the story problem item after giving successful answers to both computation items is almost 2 1/2 times the probability of success after giving incorrect responses to both computation items. For boys, the probability of success in the story problem item after successful responses to the computation items is about 1 3/4 times the probability of success on the story problem item after incorrect responses to the computation items. Three years later, Marshall (1984) analyzed more in depth these phenomena of sex-related differences. Her interest was twofold. First, she wanted to know if there were differences in the rate of success for boys and girls in solving computation and story problem items. Second, she examined additional factors that interact with sex to influence mathematics performance, such as reading achievement, socio-economic status (SES), primary language, and chronological age. Two questions were raised: Do the probabilibities of successful solving of computation and story problem items increase with reading score? Are these probabilities different for the two sexes? Approximately 270,000 students from the 6th grade were administered the Survey of Basic Skills of the California Assessment Program, during the years 1977, 1978, and 1979. Responses were analyzed using log-linear models. Successful solving of computation items was positively associated with successful solving of story problems. Girls were more successful in computation than boys, and boys were more successful than girls in solving story problems. This finding supports reports from the National Assessment of Educational Progress (NAEP) (Armstrong, 1979). To investigate the effects of reading, SES, language and chronological ager only those test forms containing 2 computation items and 2 story problems were considered for analysis; 32 items from 8 test forms were included in the analysis. The results of these analyses showed that at every level of reading score, 6th grade children were more successful in computation than in story problems. Although the differences were not larger at every reading score boys consistently had higher probabilities of success in story problems than did girls, and girls consistently showed higher probabilities of success in computation than boys. Also, as the reading score increased, the difference between the probability of success in story problems and the probability of success in computation grew larger. This difference grew larger for girls than for boys. Although SES was a major factor in solving computation and story problem items successfully, the effect was similar for each sex. Sex-related differences by primary language or chronological age were not large. This research carried out by Marshall with elementary grade children supports previous research findings that males are better than females in mathematics problem solving (a higher order skill) and females are better than their counterpart males in computation (a lower level skill). Marshall's research also brought out a different aspect of this question: the notion that girls find it more necessary than boys to succeed in the computation items in order to successfully solve the story problem items. Cognitive and Affective Factors That Influence Sex Differences in Mathematics Learning and Achievement The research reviewed in the literature does not provide evidence of any unique variable that could serve as an explanation for the observed sex-related differences in mathematics learning and achievement. However, some issues have been discussed, among which the most prevalent are that sex-related differences in mathematics learning and achievement are a result of differences in formal education; that sex-related differences in mathematics learning and achievement arise from sex differences in spatial visualization; and that sex-related differences result from a differentiated effect of affective variables on the mathematics performance of males and females. Differences in Formal Mathematics Education (Differential Coursework Hypothesis) The basis for the differential coursework hypothesis is the fact that sex-related differences in mathematical learning and achievement show up when comparing groups which are not equal in previous mathematics learning. Atter the 8th grader boys tend to select mathematics courses more otten than girls. Therefore, girls show lower achievement scores in mathematics tests because their mathematics experience is not as strong as the boys' (Fennemar 1975; Fennema & Sherman, 1977; Sherman, 1979). Fennema and Sherman's study (1977) lends additional support to the feasibility of viewing sex differences in mathematics learning and achievement as reflecting something other than a difference in mathematics aptitude. After controlling for previous study of mathematics, they found significant sex differences in mathematics achievement in only two of the four schools under study, making the attribution to sex per se less likely. Controlling for the number of space visualization-related courses, the sex-related differences which originally emerged in spatial visualization scores became non-significant. In the two schools where sex differences in mathematics achievement were found, differences between the sexes were also found in their attitudes toward mathematics. Researchers like Backman (1972), who analyzed data from Project Talent, and Allen and Chambers (1977) have also hypothesized that sex-related differences in mathematics achievement may be related to different curricula followed by males and females. Allen and Chambers attributed male superiority in mathematics problem solving to differences in the number of mathematics courses taken in high school. This issue has been seriously questioned by Astin (1974) Fox (1975ar 1975b), and Benbow and Stanley (1980), among others. Astin and Fox have reported large differences in favor of males among gifted students taking the Scholastic Achievement Test. These differences occur as early as grade 7, when there are no sex differences in the number of courses taken. Benbow and Stanley (1980) compared mathematically precocious boys and girls in the 7th grader with similar mathematics background, and found sizeable sex-related differences favoring boys in mathematical reasoning ability. Five years later, they conducted a follow-up study which showed that boys maintained their superiority in mathematics ability during high school. While Fox attributed sex-related differences in mathematical achievement to differential exposure to mathematical games and activities outside school, Benbow and Stanley suggested that sex-related differences in mathematics performance stem from superior mathematical ability in males, not from differences in mathematics formal education. The differential coursework hypothesis is not totally convincing and, as reported before, it has been challenged by researchers such as Benbow and Stanley (1980). However, Pallas and Alexander (1983) have questioned the generalizaoility of Benbow and Stanley's findings based on the fact that they used highly precocious learners. The differential coursework hypothesis can be accepted only as a partial explanation of differences in mathematics performance found between the sexes. Differences in Spatial Ability The basic premise in this issue is that males and females differ in spatial visualization and this explains differential mathematics learning and achievement. Until recently, sex differences in spatial ability in favor ot males were believed to be a fact and were thought by some to be related to sex differences in mathematical achievement. Research findings in this area have been inconsistent. In 1966, Maccoby stated that "by early school years, boys consistently do better (than girls) on spatial tasks and this difference continues through the high school and college years" (p.26). In 1972, Maccoby and Jacklin said that the differences in spatial ability between the sexes "remain minimal and inconsistent until approximately the ages of 10 or 11, when the superiority of boys becomes consistent in a wide range of populations and tests" (p.41). In 1974, after a comprehensive literature search, Maccoby and Jacklin concluded that sex differences in spatial visualization become more pronounced between upper elementary years and the last year of high school, the years when sex-related differences in mathematics achievement favoring boys emerge. Guay and McDaniel (1977) supported in part Maccoby and Jacklin's 1974 findings. They found that among elementary school children, males had greater high level spatial ability than females, but that males and females were equal in low level spatial ability. This finding is inconsistent with that portion of Maccoby and Jacklin's review that suggests that sex differences become evident only during early adolescence. Cohen and Wilkie (1979) however, stated that in tests measuring distinct spatial tasks, males perform better than females in early adolescence and throughout their life span. Most studies carried out after 1974 have failed to support these sex differences in spatial abilities (Armstrong, 1979; Connor, Serbin, & Schackman, 1977; Fennema & Sherman, 1978; Sherman, 1979). Fennema and Sherman (1978) and Sherman (1979) have explored sex-related differences in mathematical achievement and cognitive and affective variables related to these differences. In a study involving students from grades 6, 7, and 8, from four school districts, Fennema and Sherman found that spatial visualization and problem solving were highly correlated for both sexes (.59 and .60). Even in the school district where sex differences were found in problem solving, no significant sex-related differences were found in spatial visualization. When Sherman (1980) compared groups of males and females in two different grades, 8 and 11, she found no sex-related differences in problem solving or in spatial visualization in grade 8. In grade 11, however, although the sexes differed in their problem solving performance, no sex-related differences were found in spatial vizualization. Even though spatial visualization in grade 8 was the second best predictor of problem solving performance in grade 11, sex differences in grade 11 were not a result of spatial visualization since no differences were found in that skill. In spite of the fact that no sex differences were found in spatial abilities, it is evident that males and females may use them in a different way. Meyer (1978), with an elementary grade sample, and Fennema and Tartree (1983) with an intermediate level sampler found that the influence of spatial visualization on solving mathematics problems is subtler and that males and females use their spatial skills differently in solving word story problems (problems that measure problem solving ability or reasoning). Fennema and Tartree (1983) carried out a three-year longitudinal study which showed that girls and boys with equivalent spatial visualization skills did not solve the same number of items, nor did they use the same processes in solving problems. The results also suggested that a low level of spatial visualization skills was a more debilitating factor for girls than for boys in problem solving performance. Landau (1984) also investigated the relationship between spatial visualization and mathematics achievement. She studied the performance of middle school children in mathematical problems of varying difficulty, and the extent to which a diagramatic representation is likely to facilitate solution. She found that spatial ability was strongly correlated to mathematical problem solving and that the effect of spatial ability was more influential for females. Females made more use of diagrams in the solution of problems, reducing the advantage of males over females in problem solving performance. The issue of sex-related differences in spatial visualization ability as an explanation for sex differences observed in mathematics achievement is less convincing and the findings more contradictory than in the issue of sex differences in formal education. Besides these cognitive issues, other issues, mostly affective in nature have also been studied in trying to explain the origin of these sex differences in mathematics achievement and learning. The studies dealing with these affective variables are reviewed in the next section. Differentiated Effect of Affective Variables Researchers have attempted to explain the effect of sex differences in internal beliefs, interests, and attitudes (affective variables) on mathematics learning and achievement. A brief statement of each explanation precedes the summary of studies conducted that support the explanation. Confidence as lerners of mathematics. Females, more than males, lack confidence in their ability to learn mathematics and this affects their achievement in mathematics and their election of more advanced mathematics courses. Maccoby and Jacklin (1974) reported that self- confidence in terms of grade expectancy and success in particular tasks was found to be consistently lower in women than in men. In 1978, Fennema and Sherman reported that in their study involving students from grades 6 through 12, boys showed a higher level of confidence in mathematics at each grade level. These differences between the sexes occurred in most instances even when no sex-related differences in mathematics achievement were found. The correlation between confidence in mathematics performance and mathematics achievement in this study was higher than for any other affective variable investigated. Sherman reported a similar finding in 1980; in males, the most important factor related to continuation in theoretical mathematics courses was confidence in learning mathematics. This variable weighed more than any of the cognitive variables: mathematics achievement, spatial visualization, general ability, and verbal skill. In the case of females, among the affective variables, confidence in learning mathematics was found to be second in importance to perceived usefulness of mathematics. Probert (1983) supported these findings with college students. A variable that needs discussion within the context of sex differences in confidence as learners of mathematics is causal attribution. Causal attribution models attempt to classify those factors to which one attributes success or failure. The model proposed by Weiner (1974) categorizes four dimensions of attribution ot success and failure: stable and internal, unstable and internal, stable and external, and unstable and external. For example, if one attributes success to an internal, stable attribute, such as ability, then one is confident of being successful in the future and will continue to strive in that area. If one attributes success to an external factor such as a teacher, or to an unstable one, such as effort, then one will not be as confident or success in the future and will cease to strive. Failure attribution patterns work this way: if failure is attributed to unstable causes, such as effort, failure can be avoided in the future and the tendency will be to persist in the task. However, if failure is attributed to a stable cause, such as ability, the belief that one cannot avoid failure will remain. Studies reported by Bar-Tal and Frieze (1977) suggest that males and females tend to exhibit different attributional patterns of success and failure. Males tend to attribute their success to internal causes and their failures to external or unstable ones. Females show a different pattern; they tend to attribute success to external or unstable causes and failures to internal ones. The pattern of attributions, success attributed externally and failure attributed internally, has become hypothesized to show a strong effect on mathematics achievement in females. Kloosterman's (1985) study supported these findings. According to Kloosterman, attributional variables appear to be more important achievement mediators for females than for males, as measured by mathematics word problems. More research is needed in this area. Mathematics as a male domain. Mathematics is an activity more closely related to the male sex domain than to the female sex domain (Eccles et al., 1983). Thus, the mathematical achievement or boys is higher than that of girls. According to John Ernest (1976) in his study Mathematics and Sex, mathematics is a sexist discipline. He attributed sex-related differences in mathematical achievement to the creation by society of sexual stereotypes and attitudes, restrictions, and constraints that promote the idea of the superiority of boys in mathematics. Ernest reported that boys, girls, and teachers, all believe that boys are superior in mathematics, at least by the time students reach adolescence. Bem and Bem (1970) agree and argue that an American woman is trained to "know her place" in society because or the pervasive sex-role concept which results in differential expectations and socialization practices. Plank and Plank (1954) were more specific. They discussed two hypotheses related to this view: the differential cultural reinforcement hypothesis and the masculine identification hypothesis. The differential cultural reinforcement hypothesis states that society in general perceives mathematics as a male domain, giving females less encouragement for excelling in it. The masculine identification hypothesis establishes that achievement and interest in mathematics result from identification with the masculine role. A study related to the differential cultural reinforcement hypothesis is that of Dwyer (1974). Dwyer examined the relationship between sex role standards (the extent to which an individual considers certain activities appropriate to males or females) and achievement in reading and arithmetic. Students from grades 2, 4, 6, 8, 10, and 12 participated in this study. She found that sex role standards contributed significant variance to reading and arithmetic achievement test scores and that the effect was stronger for males than for females. This led to her conclusion that sex-related differences in reading and arithmetic are more a function of the child's perception of these areas as sex-appropriate or sex-inappropriate than of the child's biological sex, individual preference for masculine and feminine sex roles, or liking or disliking reading or mathematics. In a study which agrees with the masculine identification hypothesis, Milton (1957) found that individuals who had received strong masculine orientation performed better in problem solving than individuals who received less masculine orientation. Elton and Rose (1967) found that women with high mathematical aptitude and average verbal aptitude scored higher on the masculinity scale of the Omnibus Personality Inventory (OPI) than those with average scores on both tasks. It is not until adolescence that sex differences in the perception of mathematics as a male domain are found (Fennemar 1976; Stein, 1971; Stein & Smithless, 1969; Verbeker 1983). In a study with 2nd, 6th and 12th graders, Stein and Smithless (1969) found that students' perceptions of spatial, mechanical, and arithmetic skills as masculine became more defined as these students got older. Fennema (1976) considers that the influence each sex exerts upon the other on all aspects of behavior is stronger during adolescence. Since during these years males stereotype mathematics as a male domain, they send this message to females who, in turn, tend to be influenced in their willingness to study or not to study mathematics. Before that stage, girls consider arithmetic feminine, while boys consider it appropriate for both sexes (Bobber 1971). Usefulness of mathematics. Females perceive mathema- tics as less useful to them than males do, and this perception occurs at a very young age. As a results females exert less effort than males to learn or elect to take advanced mathematics courses. Many studies reported before 1976 found that the perception of the usefulness of mathematics for one's future differs for males and females, and is related to course taking plans and behavior (Fox, 1977). If females do not perceive mathematics as useful for their future, they show less interest in the subject than counterpart males. These differences in interest are what Hilton and Berglund (1974) suggest to account for sex-related differences in mathematics achievement. Although the perception of the usefulness of mathematics is still an important predictor of course taking for girls, there is a growing similarity between males and females regarding the usefulness of mathematics (Armstrong & Pricer 1982; Fennema & Sherman, 1977; Moller, 1982/1983). Armstrong and Price investigated the relative influence of selected factors in sex-related differences in mathematics participation. Both males and females selected usefulness of mathematics as the most important factor in deciding whether or not to take more mathematics in high school. Moller's study revealed that both males and females based mathematics course-taking decisions on career usefulness. A Fennema and Sherman (1977) study showed only slight differences between males and females in their feelings about the usefulness of mathematics. In her study of this variable among college students, Probert (1983/1984) did not find any sex-related differences either. These have been the main affective variables researched in attempting to explain the underlying causes of sex-related differences observed in mathematics learning and achievement. In spite ot the great diversity of studies dealing with both cognitive and affective variables, there are no clear-cut findings to render unequivocal support to a particular variable as accounting for these sex-related differences. However, everything seems to point to the fact that affective, rather than cognitive variables play a more significant role in the sex-related differences observed in mathematics performance and learning. In most of the studies dealing with affective variables, findings consistently show that these factors influence mathematics performance in females more than in males. In at least one area confidence as learners of mathematics, Sherman (1980) found that this variable influenced course election more than all the cognitive variables previously discussed. The case for the societal influences on sex roles and expectations to account for the differences in mathematics learning is also supported in one way or another in the studies reported in the literature. Problem Solving Performance and Related Variables Problem solving has been perhaps the most extensively researched area in mathematics education. Published reviews by Kilpatrick (1969) Riedesel (1969), and Suydam and Weaver (1970-1975) attest to this. Much of the research done has focused on identifying the determinants of problem difficulty and the problem features that influence the solution process. At presents no set of variables has been clearly established as a determinant of problem difficulty. Several researchers have investigated the effect or reading and computation on problem solving performance. Others have studied the effect of student attitudes toward problem solving in problem solving learning and achievement. Typically, correlational methods have been used to investigate these questions. Computational Skills and Problem Solving Performance One of the first researchers to study the effect of computation and reading on problem solving performance was Hansen (1944). He investigated the relationship of arithmetical factors, mental factors, and reading factors to achievement in problem solving. Sixth grade students were administered tests in problem solving and categorized as superior achievers (best problem solvers) and inferior archievers (poorest problem solvers). The two groups were compared in selected factors believed to be related to success in arithmetic problem solving: arithmetical, mental and reading factors. After controlling for mental and chronological age, the superior achievers in problem solving surpassed the inferior achievers in mental and arithmetical factors. The superior group did better in only two of the six items under the reading factors: general language ability and the reading of graphs, charts, and tables. The findings suggest that reading factors are not as important as arithmetic and mental factors in problem solving performance. However, these findings should be taken cautiously, as the content of the Gates tests (used to measure reading) is literary and does not include mathematical material. Chase (1960) studied 15 variables in an effort to find out which ones have significant influence on the ability to solve verbal mathematics problems. Only computation, reading to note details, and fundamental knowledge were primarily related to problem solving. Computation accounted for 20.4% of the 32% variance directly associated with problem solving. Chase concluded that a pupil's ability in the mechanics of computation, comprehension of the principles that underline the number systems, and the extent to which important items of information are noticed when reading, are good predictors of the student's ability in solving verbal problems. Balow (1964) investigated the importance of reading ability and computation ability in problem solving performance. He objected to the approaches used by other researchers who in their analyses dichotomized research subjects as "poor" or "good" students, and who ignored the recognize effect of intelligence on reading and on mathematics achievement. Balow administered the Stanford Achievement Test (subtests of reading, arithmetic, and reasoning) and the California Short-Form test of mental ability to a group of 1,400 children from the 6th grade. All levels of achievement were included in the analysis. Analysis of variance and covariance were used and compared. He confirmed the findings of other researchers to the effect that there is a direct relationship between I.Q. and reading ability and between I.Q. and computational skills. The results of the analysis of variance revealed that increases in computation ability were associated with higher achievement in problem solving. A relationship between reading ability and problem solving was also found, but it was not as strong. Significant differences in problem solving performance associated with computational ability were found when intelligence was controlled. Balow concluded that computation is a much more important factor in problem solving than reading ability, and that when I.Q. is taken into consideration, the degree of the relationship between reading and problem solving ability becomes less pronounced. Intelligence tends to confound the relationship between these two variables. Knifong and Holtan (1976, 1977) attempted to investigate the types of difficulties children have in solving word problems. They administered the word problem section of the Metropolitan Achievement Test to 35 children from the 6th grade. Errors were classified in two categories. Category I included clerical and computational errors. Category II included other types of errors, such as average and area errors, use of wrong operation, no response, and erred responses offering no clues. It the student's work indicated the correct procedure and yet the problem was missed because of a computational or clerical errors it was assumed that the problem was read and understood. An analysis of frequencies showed that clerical errors were responsible for 3% of the problems incorrectly solved, computational errors accounted for 49%, and other errors for 48% of the erred problems. Knifong and Holtan concluded that "improved computational skills could have eliminated nearly half or the word problem errors" (p. 111). These computational errors were made in a context where other skills such as reading, interpretation of the problem, and integration of these skills necessary for the solution of word problems, might interact. However, Knifong and Holtan state that their findings neither confirm nor deny that improvement or reading skills will lead to improvement in problem solving. They conclude that "it is difficult to attribute major importance to reading as a source of failure" (p. ll). In a later analysis, looking for evidence of poor reading abilities affecting children's success in word problems, Knifong and Holtan (1977) interviewed the children whose errors fell under the category or "other errors." Students were asked to read each problem aloud and answer these questions: What kind of situation does the problem describe? What does the problem ask you to find? How would you work the problem? Ninety five percent of the students read the problem correctly; 98% explained the kind of situation the problem described in a correct manner; 92% correctly answered what the problem was asking them to find, and 36% correctly answered the question of how to work the problem. The fact that a large percent of the students whose errors were classified as "other errors" (in which reading skills might have been a factor) correctly stated how to work the problem, is strong evidence of their ability to read and interpret the problems correctly. The errors made by this group of students had a distinct origin, unrelated to reading ability. Zalewski (1974) investigated the relative contribution of verbal intelligence, reading comprehension, vocabulary, interpretation of graphs and tables, mathematical concepts, number sentence selection, and computation to successful mathematical word problem solution, and the relationship of the dependent variable to the eight independent variables. She worked with a group of 4th grade children who were administered the subtests of the Iowa Test of Basic Skills (ITBS) and the Weschler Intelligence Scale for Children (WISC). Multiple regression analysis was performed. A correlation of .769 was found between word problem solving and the eight independent variables. Correlations between word problem solving and the independent variables ranged from .363 (verbal intelligence) to .674 (mathematical concepts). Correlations between the independent variables ranged from .369 (verbal intelligence and computation) to .749 (reading comprehension and vocabulary). Mathematical concepts, computation, and number sentence selection were almost as effective as all eight independent variables in predicting achievement in mathematical word problem solving. Mathematical concepts, computations number sentence selection, and reading comprehension accounted for 58% of the variance, whereas all eight predictors accounted for 59% of the variance. The two best predictors were mathematical concepts and computations which accounted for 54% variance. Other variables accounted for about 40% of the variance. The author recommends that the findings of this study be interpreted cautiously because the correlation between the eight independent variables was high, and, according to Zalewski, in a study of this nature where the interest is primarily in the influence of several variables on one dependent variable, a low correlation between the independent variables is required. (p. 2804) In a more recent investigations Exedisis (1983) studied the contribution of reading ability, vocabulary, mathematical concepts, computation, sex, and race on problem-solving performance. The Iowa Test of Basic Skills was administered to a group of 6th, 7th, and 8th grade anglo and black Chicago male and female adolescents. Problem solving was highly correlated to an understanding of basic mathematical concepts, somewhat correlated to race, and weakly correlated to computational and vocabulary skills, sex, and reading ability. Although the findings of these studies show a relationship between computational skills and problem solving achievement, this relationship is not strong enough to be considered the most determinant factor in problem solving achievement, as some of the researchers have been careful to point out. In spite ot the dismissal of reading as a determinant factor in problem solving achievement by some of these same researchers, more recent studies in this area have led others to hold different views. Reading and Problem Solving Performance Martin (1964) studied the contribution of reading comprehension, computation, abstract verbal reasoning, and arithmetic concepts to arithmetic problem solving performance. Fourth and 8th grade students were administered the Iowa tests of Basic Skills and the Lorge-Thorndlike intelligence test (verbal). He found that in the 4th grade the correlations between problem solving and abstract verbal reasoning, reading comprehension, arithmetic concepts, and computation were .61, .64, .66, and .60 respectively, and .56, .68, .69, and .63 in the 8th grade. When computation was held constant, the correlation between problem solving and reading was .52 in grade 4 and .54 in grade 8. When reading was held constant the correlation between problem solving and computation was .43 in grade 4 and .42 in grade 8. Creswell (1982) worked with a sample of anglo and black adolescents from Chicago. Each subject was administered the California Achievement test. Multiple regression was used to analyze the data. The analysis showed that reading is more important than computation in predicting student performance in problem solving. Reading accounted for 49.5% of the variance; computation accounted for 14.6% of the variance. Ballew and Cunningham (1982) worked with 6th grade students in an attempt to find what proportion of students have as their main source of difficulty with word problems each of the following factors: a) computation skills, b) interpretation of the problem, c) reading and, d) integrating these skills in the solution of problems. They also wanted to know if a student can be efficiently diagnosed as having one of the four categories as his/her main difficulty with mathematics word problems. Their study is important because it represents an attempt to demonstrate that multiple factors can interact in the correct solution of a mathematics word problem. They constructed three graded tests from a basal mathematics series for grades 3 through 8. For test 1i the problems were set in pure computational form (the effects or reading, interpretation, as well as the necessity for integration were removed in an effort to measure the computational skills required by the word problems). For test 2, the effects of reading and computation were removed by reading the problems to the students and by giving scores based on whether or not the students set them up properly, in an attempt to measure problem interpretation alone. For test 3, the effect of computation was removed. The test yielded two scores--one by grading the students on whether or not they set up the problems properly and another by grading on the basis of the correct answer. The tests were administered to all 244 students from the 6th grade in two different schools. A diagnostic profile was obtained for each of the 217 students for which complete data were available: a computational score, a problem-interpretation score, a reading score, and a reading-problem solving score. They assumed that if the reading-problem interpretation score was lower (one or more levels lower) than the problem-interpretation score, the difficulty was due to reading ability. If the score of the lowest of the three areas (computation, problem interpretation, and reading-problem interpretation score) was the same as the reading-problem solving score, the student's area of greatest immediate need was either computation, problem interpretation, or reading. If the reading-problem solving score was lower than the lowest of the other three scores, the student's area of greatest immediate need was integration. Analysis of the data revealed that for 19% of the students, problem interpretation was their major difficulty; for 26% of the students, integration (total problem solving) was their greatest immediate need; for another 26%, computation was the major weakness; and for 29%, reading was their greatest immediate need. Seventy five percent of the students demonstrated clear strength in computation, 21% in problem interpretation, and 4% in reading-problem interpretation. An analysis across all students (including those without complete data) showed that 26% of the subjects could not work word problems at a level as high as that at which they could computer interpret problems, and read and interpret problems, when those areas were measured separately. This led them to conclude that knowing the skills or the components of solving word problems is not sufficient for success, since the components must be integrated into a whole process (mastery learning of the components cannot assure mastery of the process). Their analysis also led them to conclude that, in the case of 6th graders, inability to read problems is a major obstacle in solving word problems. Only 12% ot the subjects could read and set up problems correctly at a higher level than they could computer while 60% could compute correctly at a higher level than they could read and set up problems; 44% could set up problems better when they heard them read than when they read the problem themselves. Only 13% could set up problems better when they read them than when they heard them read. Muth (1984) investigated the role of reading and computational skills in the solution of word problems. A group of 200 students from the 6th grade were administered a test of basic skills and a mathematics word problem test. The word problem test consisted of 15 sample items supplied by the National Assessment of Educational Progress. The items were adapted to include some extraneous information and complex syntactic structure. Four versions of the test were constructed by combining two versions of problem information (absence vs. presence of extraneous information) with two versions of syntactic structure (simple vs. complex syntax). Task performance was measured by means of the number of problems answered correctly, number of problems set up correctly, and amount of time spent taking the test. Reading ability and computational ability were both positively correlated with number of correct answers and with number of problems correctly set up, and negatively correlated with test-taking time. Presence of extraneous information was negatively correlated with correct answers and correct set ups and positively correlated with test-taking time. Syntactic complexity was not significantly correlated with any of the performance measures. Results of a multiple regression analysis showed that reading accounted for 46% of the variance in total correct answers and computation accounted for 8%. Reading ability and computational ability uniquely accounted for 14% and 8% of the variance in the number of correct answers, respectively. Extraneous information added significantly to the variance explained in the number of correct answers, but syntactic structure did not. Reading ability accounted for 5% of the variance in test-taking time, but computation did not add significantly to the variance explained by reading. Muth concluded that reading and computation both contribute significantly to success in solving arithmetic word problems, but that reading plays a more significant role than does computation. The studies reviewed in this section show a positive relationship between reading and problem solving performance, but in the case of Ballew and Cunningham (1982) this relationship is not viewed singly but rather as one among the interacting factors that produce successful problem solving. The third variable reviewed is the effect of student attitudes toward problem solving on problem solving performance. Many researchers have tried to demonstrate that this variable is a determinant factor in problem solving achievement. Attitudes Toward Problem Solving and Problem Solving Performance Research studies support the existence of positive and rather stable relationships between student attitudes and achievement in mathematics. Aiken (1970) has suggested that an individual's attitude toward one aspect of the discipline (mathematics), such as problem solving, may be entirely different from his/her attitude toward another phase of the discipline, such as computation. Research, however, has been directed to the use of single, global measures of attitudes toward mathematics rather than to the investigation of attitudes toward a particular phase of the discipline. The studies described below are only part of the few investigations which have examined the relationship between student attitudes and performance in the area of problem solving. Carey (1958) constructed a scale to measure attitudes toward problem solving. Her interest was in general problem solving rather than in mathematical problem solving. Her work constitutes the first attempt to construct a measure of attitudes toward problem solving. The scale was used with a group of college students, and she found, among other things, that problem solving performance is positively related to problem solving attitudes and that, in the case of females, positive modification of attitudes toward problem solving brings a significant gain in problem solving performance. Lindgren, Silvar Faraco, and DaRocha (1964) adapted Carey's scale of attitudes toward problem solving and applied it to a group of 4th grade Brazilian children. Students also answered an arithmetic achievement test, a general intelligence test, and a socioeconomic (SE) scale. A low but significant positive correlation was found between arithmetic achievement and attitudes toward problem solving. A near zero correlation was found between attitudes toward problem solving and intelligence. Since problem solving is one aspect of the discipline of mathematics, this correlation between attitudes and arithmetic achievement can lead to a conclusion or a strong correlation between attitudes toward problem solving and problem solving performance. Whitaker (1976) constructed a student attitude scale to measure some aspects of 4th grade student attitudes toward mathematic problem solving. He included statements reflecting children's beliefs about the nature of various types of mathematical problems, the nature of the problem solving process, the desirability or persevering when solving a problem, and the value of generating several ideas for solving a problem. He correlated student attitudes toward problem solving with their scores in a mathematical test which yielded a comprehension score, an application score, and a problem solving score. He found a significant positive relationship between problem solving performance and student attitude scores on the subscale which measured reactions to such things as problem solving techniques or problem situations, or to the frustration or anxiety experienced when confronted with problem solving situations. In another part of this study, Whitaker investigated the relationship between the attitudes of 4th grade teachers toward problem solving and their students' performance in problem solving. A very weak and nonsignificant negative correlation was found between the teacher's attitudes toward problem solving and student performance. The studies reviewed have confirmed the relationship between problem solving performance and attitudes toward problem solving (Carey, 1958; Lindgren et al., 1964; Whitaker, 1976). However, the results reported in the studies that investigated the relationship between problem solving performance and computation and between reading and problem solving fail to be consistent in their conclusions. Hansen (1944), Chase (1960) Balow (1964), Knifong and Holtan (1976, 1977), and Zalewski (1974) concluded that computation is more strongly related to problem solving than is reading. Martin (1964), Creswell (1982), Ballew and Cunningham (1982), and Muth (1984), concluded that reading ability and mathematical problem solving show a stronger relationship than computation and problem solving. Exedisis's (1983) findings led to the conclusion that the effect or reading and computation in problem solving performance is unimportant. CHAPTER III METHOD The first objective of this study was to investigate sex differences in the selection of incorrect responses in a mathematics multiple-choice achievement test, and to determine whether these differences were consistent over three consecutive administrations of the test. The second objective was to investigate whether males and females differ in problem solving performance, if these differences persist after accounting for computational skills, and if the male-female differences depend on the level of computational skills. This chapter contains descriptions of the sample, the test instrument, and the statistical analysis used in achieving the above mentioned objectives. The Sample To achieve the first objective of the study, all the students who took the Basic Skills Test in Mathematics-6 ("Prueba de Destrezas Basicas en Matem&ticas-6") in each or the three years were included in the study. To achieve the second objective of the study, approximately 1,000 students were selected randomly for each year. For the first year, 1,002 were selected (492 boys and 510 girls); for the second year, 1,013 students were selected (504 boys and 509 girls); and, for the third years 1,013 students were selected (509 boys and 504 girls). The student population in Puerto Rico includes children from the urban and rural zones and comprises children from low and middle socioeconomic levels. Findings can be generalized only to this population. The Instrument The Basic Skills Test in Mathematics-6 is a criterion-referenced test used in the Department ot Education of Puerto Rico as part of the annual assessment program. The test measures academic achievement in operations, mathematical concepts, and story problems. It has a reported split half reliability or .95. The test was designed specifically for Puerto Rico. Its contents and the procedures followed for its development were formulated and reviewed by educators from the mathematics department of the Department of Education of Puerto Rico, in coordination with the Evaluation Center ot the Department of Education and mathematics teachers from the school districts. The emphasis placed on each skill area is depicted in Figure 3.1. Fig. 3.1 Number of Questions by Skill Area in the Basic Skills Test in Mathematics 6 Analysis of the Data Analysis of Sex by Option by Year Cross Classifications Log-linear models were used to analyze the sex by option by year cross classifications for each item. Two topics are addressed in this section, the hypotheses tested using log-linear models and a comparison of the hypotheses tested in this study with those tested by Marshall (1981, 1983). The object of the analysis was to test two hypotheses: 1. The proportion of males and the proportion of females choosing each incorrect option does not vary from year to year. Note that this hypothesis is stated in the null form. 2. Assuming that the first hypothesis is correct, the proportion of males who choose each incorrect option is different from the proportion of females who choose each incorrect option. This hypothesis is stated in the alternate form. In Table 3.1 a hypothetical cross classification of sex, option, and year is presented. Hypothesis 1 is true for this three dimensional contingency table. In Table 3.2 the three dimensional contingency table is rearranged to show the year by option contingency table for each gender. Table 3.1 Hypothetical Probabilities of Option Choice Conditional on Year and Sex Option Year Sex 1 2 3 First M .7 .2 .1 F .5 .3 .2 Second M .7 .2 .1 F .5 .3 .2 Third M .7 .2 .1 F .5 .3 .2 Table 3.2 Hypothetical Probabilities of Option Choice Conditional on Year and Sex and Arranged by Sex Option Sex Year 1 2 3 M First .7 .2 .1 Second .7 .2 .1 Third .7 .2 .1 F First .5 .3 .2 Second .5 .3 .2 Third .5 .3 .2 Inspection of the year by option contingency tables shows that year and option are independent for each gender. Thus, hypothesis 1 is equivalent to the hypothesis that year and option are independent conditional upon gender. Hypothesis 2 is also true for Table 3.1 and Table 3.2. Therefore, when hypothesis 1 is correct, hypothesis 2 is equivalent to the hypothesis that sex and option choice are dependent. In his discussion of the analysis of three dimensional contingency tables, Fienberg (1980) presents the following saturated model for the data: log mijk = Ui+Uj+Uk+Uij+Uik+Ujk+Uijk. (3.1) In this model, mijk is the expected value of the frequency in cell ijk of the three dimensional table. The model states that all three classification factors for a three dimensional contingency table are mutually dependent. In the present research i is the year index, j is the option index, and k is the sex index. Fienberg shows that deleting the terms Uij and Uijk yields a model in which year and option are independent conditional upon sex. This model is log mijk = Ui+Uj+Uk+Uik+Ujk. (3.2) Fienberg also shows that deleting the Ujk term from (3.2) to obtain log mijk= Ui+Uj+Uk+Uik. (3.3) yields a model that specifies that option is independent of sex. Based on Fienberg's presentation, an appropriate analysis for testing the hypotheses is 1. Conduct a likelihood ratio test of the adequacy of model (3.2). If this test is nonsignificant, then the data support the adequacy of the model, and the hypothesis that conditional on gender, options and year are independent. Because model (3.1) is a saturated model, testing the adequacy of model (3.2) is the same as comparing the adequacies of models (3.1) and (3.2). 2. Conduct a likelihood ratio test comparing .the adequacies of models (3.2) and (3.3). If this test is significant, then model (3.2) fits the data better than model (3.3) and the data support the hypothesis that the choice of incorrect option is dependent on sex. To summarized if the first test is nonsignificant and the second test is significant, then the choice of option is dependent on sex, and the pattern of dependency is the same for all three years. A problem arises in interpreting this analysis with the data used in this research. Over the three years, there were responses available from 135,340 students. Even if only 5% of the students answered an item incorrectly, the responses of 7,767 students would be used in analyzing this item. On the other hand, if 90% answered an item incorrectly, the responses of 121,806 students would be used in analyzing the data. As a result of the large sample size the tests described above are likely to be very powerful. In step 1 of the analysis, then, even a very small change from year to year in the proportion of males or females who choose an option is apt to be detected, and the results will indicate that hypothesis 1 is not supported. For step 2, even a very small dependence of option choice on sex is likely to be detected and hypothesis 2 is likely to be supported. In brief, the problem caused by the large sample size is that practically insignificant differences may yield statistical significance. Fortunately, the form of the test statistic used in the likelihood ratio test suggests a reasonable solution to this problem. The test statistic is G2 = 2 ZFo loge (Fo/Fe). (3.4) Here the summation is over all the cells in the contingency table Fo refers to the observed frequency in a cell, and Fe refers to the estimated expected frequency in a cell. Denoting the observed proportion in a cell by Po and the estimated expected proportion in a cell by Per the test statistic can be written G2 = 2 N TPo loge (Po/Pe)- (3.5) where N is the total number of subjects. This form of the test statistic suggests the following strategy. For any significant G2, using Po and Pe calculated from the total data set available for an item, calculate the minimum N required for G2 to be significant. If the minimum N is very larger this suggests that the statistically significant result is not practically significant since it can only be detected in very large samples. Of course, the question remains as to what can be considered a minimum large N. Although there is room for argument, it seems reasonable to claim that if an average of 1000 subjects per year is required to show significance, then the result is not likely to be practically significant. On the basis of this reasoning, it was proposed to ignore all significant results that would be nonsignificant if there were less than 3000 subjects available. In addition, all log-linear model tests were conducted using a .01 level of significance. Since this research is based on Marshall (1981, 1983), it is important to compare the method of analysis used in this study to the one used by Marshall. Marshall also used a two-step analysis. In the first step of her analysis she deleted the Uijk term from (3.1) and tested the adequacy of the model, log mij k = Ui+Uj+Uk+Uij+Uik+Ujk. (3.6) Following this, she deleted the Ujk term to obtain log mijk = Ui+Uj+Uk+Uij+Uik. (3.7) and compared these two models using a likelihood ratio test. If the first test was nonsignificant and the second significant, Marshall claimed that option choice was dependent on sex and the pattern of dependency was the same from year to year. This is the same claim that this study sought to establish. However, the approach used here was to present evidence that model (3.2) fits the data while Marshall tried to show that model (3.6) fits the data. The major difference between the two approaches concerns the operationalization of the concept that the 64 gender-option dependency is the same from year to year. In this study, a three dimensional table was considered to exhibit the same year to year pattern of gender-option dependency if, conditioned on gender, the same proportion of students chose each incorrect response over the three year period. This seems to be a straightforward and natural way to operationalize the concept. To illustrate how Marshall operationalized the concept in question, a hypothetical set of probabilities was constructed conforming to the pattern specified by Marshall. This is displayed in Table 3.3. Table 3.3 Hypothetical Joint Probabilities of Year, Option, and Sex Option Year Sex 1 2 3 ------------------------ First M .120 .045 .015 F .060 .060 .030 I------------ I Second M .144 .022 .019 F .072 .030 .039 I----------------------~ Third M .132 .040 .016 F .066 .054 .039 Note: Probabilities reported are truncated to three places. One important characteristic of this table involves two-by-two subtables of sex cross classified with two of the options for each year. For example, consider the three two-by-two tables obtained by cross classifying sex and option choices one and two for each year. These tables are indicated by dotted lines on Table 3.3. For this table, the ratio of the odds of a male choosing option one to the odds of a female choosing option one is the same (within truncating error) for each year. For example, this odds ratio for the first year is (.120/.045) / (.060/.060) = 2.67. Within the error caused by reporting truncated probabilities the odds ratio for years two and three is the same as that for year one. The equality of these odds ratios is Marshall's operationalization stage of the year- to-year gender-option dependency. To show that the odds ratio can be constant over years, but that the probabilities of option choice conditional on sex and year can change from year to year, for both males and females, the probabilities in Table 3.3 were converted to the probabilities of option choice conditional upon sex and year. These conditional probabilities are reported in Table 3.4. Unlike the probabilities in Table 3.2, those in Table 3.4 change from year to year for both males and females. Table 3.4 Hypothetical Probabilities of Option Choice Conditional on Sex and Year Option Year Sex 1 2 3 First M .666 .250 .083 F .400 .400 .200 Second M .774 .120 .104 F .510 .212 .276 Third M .698 .214 .087 F .431 .352 .215 Note: Probabilities reported are truncated to three places. Which procedure is more appropriate to test the claim that the pattern of male and female option choice remains the same from year to year? It seems more reasonable to test for a pattern like that in Table 3.2 than to test for a pattern like that in Table 3.4, and, consequently, this was the strategy adopted in this study. Comparison of Males and Females in Problem Solving Performance One object of the study was to compare the performance of males and females on problem solving. Two questions were addressed. First, do males and females differ in problem solving performance? Second, do these differences persist when computational skill is controlled for, and do these differences depend on the level of computational skill? Seven analyses of covariance (ANCOVA) tests were conducted, one with each of addition, subtraction, multiplication, division, addition of fractions, decimals in subtraction, and equivalence as covariates. A well known problem that arises in the use of ANCOVA is that an unreliable covariate can cause spurious differences between the sex groups. To solve this problem, Porter (1967/1968) proposed the use of estimated true scores for observed scores. Porter (1967/1968) conducted a simulation that gave empirical support to the adequacy of this strategy. Hunter and Cohen (1974) have provided theoretical support for this strategy. CHAPTER IV RESULTS Introduction The data gathered from 6th grade students from the public schools in Puerto Rico who took the Basic Skills Test in Mathematics-6 during three consecutive years were analyzed in this study. The first objective was to investigate whether boys and girls differ in the selection of incorrect responses, and if the pattern of differences was consistent throughout the three years in which the test was administered. The second objective was to investigate whether males and females differ in problem solving performance, if these differences persist after accounting for computational skills, and if the male-female differences depend on the level of computational skills. The results are discussed in two sections. Study findings in the area of sex differences in the selection of incorrect responses are discussed in the first section. The second section is devoted to an exposition of the findings in the area of sex-related differences in problem solving. Sex-Related Differences in the Selection of Incorrect Responses As indicated in Chapter III, there are two models of interest. The first model indicates that year and option are independent, conditional upon sex, log mijk = Ui+Uj+Uk+Uik+Ujk. (3.2) Substantively, this model implies that the pattern of male and female option choices is consistent over the three years of test administration. The second model indicates that option is independent of sex, log mijk = Ui+Uj+Uk+Uik- (3.3) This model implies that the pattern of option choice is the same for males and females. In order to determine if males and females differed in the selection of incorrect responses and if these differences were stable across the three years of test administration, a two-step test was performed. First, a likelihood ratio test of the adequacy of model (3.2) was conducted for each of the 111 items of the test. Under this test a model fits the data when the X2 value obtained is nonsignificant at a specified alpha level. For 50 (45%) of the items, model (3.2) fitted the data adequately. For these items the pattern of male and female option choices was consistent for the three years. The X2 values for the 61 items for which the likelihood ratio tests were significant at the .01 level are reported in Table 4.1. Also reported in this table are the actual sample sizes and the minimum sample sizes necessary for the likelihood ratio tests to be significant. Of the 61 items, 59 had minimum sample sizes greater than 3,000. Thus, although for the three years both males and females samples had inconsistent option choices on these 61 items, on 59 of the items the inconsistency was relatively minor. Consequently, these 59 items were included in step 2 along with the initial 50 items. In step 2, a likelihood ratio test was performed comparing the adequacy of the models in (3.2) and (3.3). The X2 values associated with models (3.2) and (3.3) for each of the 109 items subjected to step 2 are reported in columns b and c of Table 4.2. Also reported in Table 4.2 is the difference between the two X2 values (see column d). This latter figure is the test statistic for comparing the adequacies of models (3.2) and (3.3). Significant X2 statistics are indicated by asterisks. The X2 statistics were significant for 100 of the items indicating a male-female difference in option choice for these items. In column e of Table 4.2 the actual sample sizes are reported. In column f, the minimum sample sizes necessary for significance are reported. For those 100 items which had significant X2 statistics reported in column d, 94 (94%) had minimum sample sizes greater than 3,000. TABLE 4.1 Summary of Significant Tests of Model 3.2 Item Xe for Actual Sample Min. Sample Number Model 3.2* Size Size for Significance (a) (b) (c) (d) 8 13 23 25 28 29 30 31 34 36 37 42 45 48 49 52 54 56 57 46.15 25.74 29.25 35.90 46.00 61.48 25.19 23.26 58.25 84.36 23.70 23.26 22.41 179.25 27.82 26.01 41.41 21.62 28.51 16,796 15,718 60,377 26,707 84,847 88,396 87,473 32,870 82,257 86,137 37,068 59,140 63,098 97,593 16,766 39,858 74,502 44,107 53,075 3,352 5,624 19,011 6,852 16,988 13,242 31,982 13,015 13,006 9,404 14,405 23,417 25,932 5,014 5,550 14,114 16,570 18,789 17,146 TABLE 4.1 Continued Item Xe for Actual Sample Number Model 3.2* Size (a) (hi ( 50.96 66.43 21.39 55.03 158.36 88.01 56.28 50.51 34.28 48.01 75.29 44.39 14,425.97 24.61 22.75 99.68 79.38 64.56 23.98 --- Min. Sample Size for Significance (d) 7,560 8,387 15,342 8,485 3,028 7,660 6,943 14,352 13,935 12,128 7,647 14,011 54** 19,969 34,077 6,485 7,693 10,957 18,700 41,829 60,494 35,632 50,699 52,068 73,196 42,429 78,708 51,862 63,223 62,511 67,528 84,368 53,358 84,175 70,183 66,304 76,806 48,690 TABLE 4.1 Continued Item XZ for Actual Sample Min. Sample Number Model 3.2* Size Size for Significance (a) (b) (c) (d) 80 75.01 55,277 6,787 81 95.51 57,710 5,565 82 51.99 71,315 12,633 83 20.12 81,021 37,088 84 38.62 78,959 18,830 85 1240.78 41,180 306** 86 102.85 39,903 3,573 87 66.75 48,194 6,650 88 63.01 38,742 5,663 89 43.26 39,648 8,441 90 85.53 53,557 5,767 91 82.53 62,954 7,025 93 78.39 80,723 9,484 94 70.15 87,241 11,454 95 40.59 83,969 19,053 96 29.31 74,287 23,343 98 29.87 69,052 21,291 101 23.77 83,782 32,462 102 61.06 75,708 11,419 104 40.47 72,517 16,503 TABLE 4.1 Continued Item XZ for Actual Sample Min. Sample Number Model 3.2* Size Size for Significance (a) (b) (c) (d) 105 28.66 60,712 19,510 108 26.56 86,261 29,912 111 30.71 86,190 25,849 * p<.Ol ** Minimum sample size less than 3,000 TABLE 4.2 Chi-square Values for the Comparison of Models (3.2) and (3.3) with Actual Sample Sizes and with the Corresponding Number of Subjects Needed for Significant Results Item Xk for X2 for X1 for Actual Num- Model Model Difference Sample ber 3.2 3.3 Size lal 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 (b) 16.11 17.54 20.06 7.54 14.75 11.86 4.60 46.15 10.13 9.54 5.89 12.58 25.74 11.19 16.63 14.51 17.90 (c) 30.94 103.22 47.51 21.87 70.40 23.13 35.73 329.98 109.99 65.29 19.06 198.46 62.37 80.50 17.33 84.29 44.20 (_d) 14.83* 85.68* 27.45* 14.33* 55.65* 11.27* 31.13* 283.83* 99.86* 55.75* 13.17* 185.88* 36.63* 69.31* 0.70 69.78* 26.30* (e) 7,427 11,018 11,497 6,016 7,706 13,046 8,832 16,796 23,075 35,060 24,908 40,423 15,718 30,238 42,889 35,114 43,962 Minimum Sample Size for Signif. (f) 10,061 2,583** 8,733 8,414 2,782** 23,256 5,700 1,189** 4,642 12,634 37,996 4,369 8,621 8,765 1,230,914 10,109 33,582 TABLE 4.2. Continued Item Xe for X for X2 for Actual Minimum Num- Model Model Difference Sample Sample ber 3.2 3.3 Size Size for Signif. (a) (b) (c) (d) (e) (f) 18 11.71 15.68 3.97 52,340 264,864 19 9.37 18.77 9.40* 41,363 88,402 20 12.37 13.46 1.09 55,707 1,026,746 21 11.51 19.03 7.52 49,240 131,547 22 9.63 83.74 74.11* 51,724 14,022 23 29.25 153.74 124.49* 60,377 9,744 24 10.35 42.93 32.58* 68,426 42,194 25 35.90 147.40 111.50* 26,707 4,812 26 18.84 47.30 28.46* 56,717 40,037 27 8.78 222.87 214.09* 78,353 7,353 28 46.00 242.78 196.78* 84,847 8,662 29 61.48 170.22 108.74* 88,396 16,331 30 25.19 53.84 28.65* 87,473 61,338 31 23.26 47.42 24.16* 32,870 27,333 32 15.51 41.69 26.18* 48,777 37,430 33 7.57 48.84 41.27* 39,470 19,214 34 58.25 324.67 266.42* 82,257 6,203 35 6.91 225.66 218.75* 88,069 8,088 TABLE 4.2 Continued Item X2 for a for Xk for Actual Minimum Num- Model Model Difference Sample Sample ber 3.2 3.3 Size Size for Signif. (a) ( (c) (d) (e) (f) 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 84.36 23.70 13.64 13.35 6.67 12.65 23.26 16.29 3.55 22.41 6.56 7.82 179.25 27.82 14.23 16.37 26.01 6.69 41.41 134.56 76.85 84.24 757.60 701.39 603.42 189.11 21.65 115.84 55.90 16.30 76.15 452.64 59.75 240.27 224.54 96.14 27.49 95.41 50.20* 53.15* 70.60* 744.25* 694.72* 590.77* 165.85* 5.36 112.29* 33.49* 9.74* 68.33* 273.39* 31.93* 226.04* 208.17* 70.13* 20.80* 54.00* mE 86,137 37,068 38,870 61,057 56,438 57,145 59,140 33,712 67,192 13,098 56,727 56,555 97,593 16,766 25,410 24,715 39,858 44,694 74,502 34,472 14,011 11,0618 1,648** 1,630** 1,943** 7,154 126,357 12,021 37,851 117,007 16,628 7,172 10,549 2,258** 2,385** 11,418 43,168 27,718 TABLE 4.2 Continued Item Xz for Xa for Xl for Actual Minimum Num- Model Model Difference Sample Sample ber 3.2 3.3 Size Size for Signif. (a) (b) (c) (d) (e) (f) 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 72 73 90.87 70.14 35.78 51.69 73.68 148.39 394.54 300.95 278.34 246.48 94.11 113.88 180.60 225.46 251.66 271.35 208.90 37.80 163.01 140.26* 71.34* 48.52* 7.27 34.20* 22.72* 81.96* 373.15* 245.92* 119.98* 158.47* 37.83* 63.37* 146.32* 177.45* 176.37* 226.96* 191.49* 13.19* 19.53 21.62 28.51 17.49 50.96 66.43 21.39 55.03 158.36 88.01 56.28 50.51 34.28 48.01 75.29 44.39 17.41 24.61 11,770 18,263 146,668 25,873 36,987 14,828 1,918** 4,142 8,719 9,279 22,532 24,953 7,121 7,158 7,121 5,977 8,831 81,271 41,796 44,107 53,075 44,044 41,829 60,494 35,632 50,699 52,068 73,196 42,429 78,708 51,862 63,223 62,511 67,528 84,175 53,358 74 22.75 84,175 12,057 TABLE 4.2 Continued Item X1 for X1 for Xa for Actual Num- Model Model Difference Sample ber 3.2 3.3 Size (aL 75 76 77 78 79 80 81 82 83 84 86 87 88 89 90 91 92 93 94 (b) 8.69 99.68 79.38 64.56 23.98 75.01 95.51 51.99 20.12 38.62 102.85 66.75 63.01 43.26 85.53 82.53 16.36 78.39 70-15 (c) 237.07 167.72 432.36 127.04 72.59 128.13 106.82 108.12 95.18 69.05 433.47 430.40 126.27 289.14 405.04 85.48 93.12 229.02 137.27 (d) 228.38* 68.04* 352.98* 62.48* 48.61* 53.12* 11.31* 56.13* 75.06* 30.43* 330.62* 363.65* 63.26* 245.88* 319.51* 2.95 76.76* 150.63* 67.12* (e) 67,783 70,183 66,304 76,806 48,690 55,277 57,710 71,315 81,021 78,959 39,903 48,194 38,742 39,648 53,557 62,954 88,231 80,723 87,241 Minimum Sample Size for Signif. (f) 5,963 20,723 3,774 24,696 20,123 20,906 102,511 25,525 21,685 52,129 2,425** 2,662** 12,304 3,240 3,368 428,727 23,092 10,766 26,113 TABLE 4.2 Continued Item X1 for Xa for X for Actual Minimum Num- Model Model Difference Sample Sample ber 3.2 3.3 Size Size for Signif. (a) (b) (c) (d) (e) (f) 68.41 212.34 126.96 41.77 9.94 41.48 315.62 437.09 7.97 283.94 48.59 1,682.71 99.20 79.33 18.33 49.11 477.44 27.82* 183.03* 118.36* 11.90* 0.50 30.13* 291.85* 376.03* 1.85 243.47* 19.93* 1,669.34* 86.07* 52.77* 12.60* 31.31* 446.73* 83,969 74,287 50,028 69,052 72,647 60,931 83,782 75,708 50,663 72,517 60,712 71,946 72,482 86,261 76,800 70,980 86,190 60,638 8,154 8,492 116,576 2,918,956 40,627 5,767 4,045 550,173 5,984 61,199 866** 16,918 32,840 112,453 45,544 3,876 * P<.01 ** Minimum sample size less than 3,000 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 40.59 29.31 8.60 29.87 9.44 11.35 23.77 61.06 6.12 40.47 28.66 13.47 13.13 26.56 5.73 17.80 30.71 Sex-related Differences in Problem Solving Performance In this part, results for the two questions related to the second objective of the study are presented. These questions serve as the framework for the presentation. Each question is stated, followed by the results pertaining to that question. Question 1: Do males and females differ in problem solving performance? The responses of 492 males and 510 females who took the Puerto Rico Basic Skills Test-6 during the spring of the first year were analyzed in this study. Also, data from 504 males and 509 females tested in the second year and from 509 males and 504 females tested in the third year were included in the analysis. The mean performance scores and the standard deviations for each of the eight variables are presented in Table 4.3. Results of t-tests are also presented in this table. Females outperformed males in problem solving, a finding consistently present in all the three years of test administration. Over the three-year period the mean differences favored females in all variables except equivalence. The sex-related differences in problem solving were significant (p<.01) for all three years. TABLE 4.3 Means, Standard Deviations, and t-Tests for the Eight Mathematical Subtests Year/ Males Females Subtest X SD X SD t First Problem Solving Addition Subtraction Multi pl i ca ti on Division Fracadd Decsub Equivalence Second Problem Solving Addition Subtraction Multiplication Division Fracadd 3.436 2.32 5.412 1.00 4.550 1.66 4.014 1.80 3.136 1.87 2.475 1.75 2.648 1.83 .77 .77 N = 492 3.632 5.470 4.843 4.148 3.242 2.565 2.28 .98 1.52 1.73 1.77 1.74 3.862 5.592 4.852 4.433 3.580 2.743 3.321 .73 2.45 .86 1.56 1.68 1.91 1.81 1.86 .76 2.82* 3.05* 1.22 3.81* 3.72* 2.32** 5.77* - .83 N = 510 4.015 5.603 4.923 4.550 3.632 2.903 2.36 .86 1.44 1.67 1.86 1.91 2.65* 2.29** .86 3.76* 3.42* 2.94* TABLE 4.3 Continued Year/ Males Females Subtest X SD X SD t Decsub 2.863 1.87 3.440 1.80 5.00* Equivalence .76 .76 .65 .75 -2.32** N = 504 N = 509 Third Problem Solving 3.927 2.49 4.341 2.44 2.67* Addition 5.536 .84 5.704 .64 3.59* Subtraction 4.836 1.52 4.958 1.46 1.30 Multiplication 4.168 1.74 4.541 1.68 3.47* Division 3.343 1.88 3.795 1.83 3.87* Fracadd 2.819 1.89 3.117 1.85 2.53** Decsub 3.021 1.93 3.448 1.85 3.60* Equivalance .830 .82 .800 .78 .60 N = 509 N = 504 Note: The number of items in the problem solving subtest was 9. In each computation subtest, the number of items was 6. An item was included in the computation subtest only if it measured a computation skill required to solve a problem solving item. * p <.01 ** p <.05 Consistent significant differences were also found for addition, multiplication, division, addition of fractions, and subtraction of decimals. For subtraction the difference was not statistically significant. Question 2: Do sex-related differences in problem solving persist when computational skills are controlled for, and is the male-female differences in problem solving dependent on level of computational skills? To address the question of dependence of male-female problem solving differences in computational skills for each year and computation subtest, the possibility of an interaction was investigated. For the first year, statistically significant interactions were found between sex and multiplication, F (1,998) = 8.59, p<.01; and sex and division, F (1,998) = 4.25, p<.01. A significant interaction was found between sex and subtraction in the second year, F (1,1009) = 6.39, p<.05. No significant interactions were found in the third year. Analysis of covariance summary tables are shown as Tables 4.4, 4.5, and 4.6. Also, the three interactions are depicted in Figures 4.1, 4.2, and 4.3. Each figure indicates that at lower levels of computational skills, males outperformed females in problem solving, with the reverse happening at higher levels of computational skills. TABLE 4.4 ANCOVA Summary Table: Multiplication Covariate First Year Source df SS MS F Multiplication (M) 1 1195.90 1195.90 265.66* Sex (S) 1 31.95 31.95 7.07* M x S 1 38.80 38.80 8.59* Error 998 4509.00 4.51 p <.01 TABLE 4.5 ANCOVA Summary Table: Divison Covariate First Year Source df SS MS F Division (D) 1 1195.90 1195.90 264.66* Sex (S) 1 13.94 13.94 3.22 D x S 1 18.39 18.39 4.25* Error 998 4317.00 4.32 * p<.01 TABLE 4.6 ANCOVA Summary Table: Subtraction Covariate Second Year Source df SS MS F Subtraction (S) 1 93.14 593.14 122.80* Sex (S) 1 18.58 18.58 3.85* SxS 1 30.85 30.85 6.39* Error 1009 4873.40 4.83 *p <. 5 87 9 8 7 6 . 0 I~ 5 o ,. 4 3 2 1 1 2 3 4 5 6 Multiplication Fig. 4.1 Sex by Multiplication Interaction. 88 9 8. 7 6 5 4 3 2 1 1 2 3 4 5 6 Division Fig. 4.2 Sex by Division Interaction 89 9 8. 7. 6 : 5 *r* O U) 1 4 1 2 3 4 5 6 Subtraction Fig. 4.3 Sex by Subtraction Interaction The significant interactions found between sex and multiplications and sex and division (first year), and sex and subtraction (second year), answered, in part, the question of whether male-female differences in problem solving performance depend on computational ability. However, the evidence is quite weak. Of 21 possible interactions, only three were significant. No variable exhibited a significant interaction for each of the three years. The analysis of covariance was also used to determine if sex-related differences exist after controlling for computational skills. Analyses were conducted for those variables that did not exhibit significant interactions with sex. As discussed in Chapter III, estimated true scores were used for observed scores to adjust for unreliability of the covariates (the computational subtests). Reliability coefficients calculated for each covariate are shown in Table 4.7. Summaries of the analyses of covariance for the first year are reported in Table 4.8. The results show that females retained their superiority in problem solving performance when equivalence was the controlling variable in the analysis of covariancer the only variable in which males outperformed females (nonsignificant). When the controlling variables were addition, subtraction, addition of fractions, and subtraction with decimals, female |

Full Text |

84
Consistent significant differences were also found for addition/ multiplication/ division/ addition of fractions/ and subtraction of decimals. For subtraction the difference was not statistically significant. Question 2: Do sex-related differences in problem solving persist when computational skills are controlled for/ and is the male-female differences in problem solving dependent on level of computational skills? To address the question of dependence of male-female problem solving differences in computational skills for each year and computation subtest/ the possibility of an interaction was investigated. For the first year/ statistically significant interactions were found between sex and multiplication/ F (1/998) = 8.59/ p second year/ F (1/1009) = 6.39/ p<.05. No significant interactions were found in the third year. Analysis of covariance summary tables are shown as Tables 4.4/ 4.5/ and 4.6. Also/ the three interactions are depicted in Figures 4.1/ 4.2/ and 4.3. Each figure indicates that at lower levels of computational skills/ males outperformed females in problem solving/ with the reverse happening at higher levels of computational skills. 16 of the test. The sexes differed in only 2 of the 19 reference tests Spatial Relations and Picture Group Name-Selection. A factor analysis however showed differences in the number and composition of the factors. For females a general mathematics factor was determined by mathematics computation comprehension application and problem solving. For males the comprehension and application parts determined one factor; problem solving with two other reference tests (Gestalt and Omelet) determined another factor. Meyer concluded that comprehension of the data and mastery of the prerequisite mathematical concepts did not guarantee successful problem solving either for males or for females. Problem solving scores for both sexes were about one third their scores in comprehension and one fourth their scores in application. She also concluded that the sexes may have approached the problem solving questions differently. The methods used by females for solving problem situations may have paralleled their approach to the application parts. Males may have used established rules and algorithms for the application parts but may have used more of a Gestalt approach to the problem solving situation. Whitaker (1976) investigated the relationship between the mathematical problem performance of 4th grade children and their attitudes toward problem solving their teachers' 96 TABLE 4 .11 Adjusted Means on Problem Solving/ by Covariate and Sex Second Year Covariate Adjusted Means Males Females Addition 3 .69 3 .96 Multiplication 3 .80 3.85 Division 3.80 3.85 Fracadd 3.78 3 .87 Decsub 3.90 3 .75 Equivalence 3.46 44.18 62 Here the summation is over all the cells in the contingency table/ F0 refers to the observed frequency in a cell/ and Fe refers to the estimated expected frequency in a cell. Denoting the observed proportion in a cell by PG and the estimated expected proportion in a cell by Pe/ the test statistic can be written G2 = 2 N Â£^P0 1 ge (P0/Pe>- (3.5) where N is the total number of subjects. This form of the test statistic suggests the following strategy. For any significant G2/ using P0 and Pe calculated from the total data set available for an item/ calculate the minimum N required for G2 to be significant. If the minimum N is very large/ this suggests that the statistically significant result is not practically significant since it can only be detected in very large samples. Of course/ the question remains as to what can be considered a minimum large N. Although there is room for argument/ it seems reasonable to claim that if an average of 1000 subjects per year is required to show significance/ then the result is not likely to be practically significant. On the basis of this reasoning/ it was proposed to ignore all significant results that would be nonsignificant if there were less than 3000 subjects available. In addition/ all log-linear 57 Analysis of the Data Analysis of Sex bv Option by Year Cross Classifications Log-linear models were used to analyze the sex by option by year cross classifications for each item. Two topics are addressed in this section/ the hypotheses tested using log-linear models and a comparison of the hypotheses tested in this study with those tested by Marshall (1981/ 1983). The object of the analysis was to test two hypotheses: 1. The proportion of males and the proportion of females choosing each incorrect option does not vary from year to year. Note that this hypothesis is stated in the null form. 2. Assuming that the first hypothesis is correct/ the proportion of males who choose each incorrect option is different from the proportion of females who choose each incorrect option. This hypothesis is stated in the alternate form. In Table 3.1 a hypothetical cross classification of sexz option/ and year is presented. Hypothesis 1 is true for this three dimensional contingency table. In Table 3.2 the three dimensional contingency table is rearranged to show the year by option contingency table for each gender. 76 TABLE 4.2. Continued Item Num ber X2 for Model 3.2 X2 for Model 3.3 X2 for Difference Actual Sample Size Minimum Sample Size for Signif. (a) (b) (c) (d) (e) (f) 18 11.71 15.68 3.97 52/3 40 264/864 19 9.37 18.77 9.40* 41/363 88/402 20 12.37 13.46 1.09 55 /707 1/026 /7 46 21 11.51 19.03 7.52 49/240 131/547 22 9.63 83 .74 74.11* 51/7 24 14/022 23 29.25 153.74 124.49* 60/377 9/744 24 10.35 42.93 32.58* 68/426 42/194 25 35.90 147.40 111.50* 26 /707 4/812 26 18.84 47 .30 28.46* 56/717 40 /037 27 8.78 222.87 214.09* 78/353 7/353 28 46.00 242.78 196.78* 84/847 8/662 29 61.48 170.22 108.74* 88/396 16/331 30 25.19 53 .84 28.65* 87/473 61/338 31 23 .26 47.42 24.16* 32/870 27/333 32 15.51 41.69 26.18* 48/777 37/430 33 7.57 48.84 41.27* 39/470 19/214 34 58.25 324.67 266.42* 82 /257 6/203 35 6.91 225 .66 218.75* 88/069 8/088 60 Fienberg also shows that deleting the Ujterm from (3.2) to obtain log mijk= Ui+Uj+Ufc+Uik. (3.3) yields a model that specifies that option is independent of sex. Based on Fienberg's presentation/ an appropriate analysis for testing the hypotheses is 1. Conduct a likelihood ratio test of the adequacy of model (3.2). If this test is nonsignificant/ then the data support the adequacy of the modelr and the hypothesis that conditional on gender option/ and year are independent. Because model (3.1) is a saturated model/ testing the adequacy of model (3.2) is the same as comparing the adequacies of models (3.1) and (3.2). 2. Conduct a likelihood ratio test comparing the adequacies of models (3.2) and (3.3). If this test is significant/ then model (3.2) fits the data better than model (3.3) and the data support the hypothesis that the choice of incorrect option is dependent on sex. To summarize/ if the first test is nonsignificant and the second test is significant/ then the choice of option is dependent on sex and the pattern of dependency is the same for all three years. 19 solving performance in grade 11. Spatial visualization was a stronger predictor for girls than for boys. Mathematics as a male domain was a good predictor for girls only; the less a girl stereotyped mathematics as a male domain in grade 8# the higher her problem solving score in grade 11. Attitudes toward success in mathematics in grade 8 was a more positive predictor of problem solving performance for boys than for girls; the more positive the attitudes toward success in mathematics in grade 8/ the higher their performance in problem solving in grade 11. None of these four studies# all of which used the Romberg-Wearne Mathematics Problem Solving Test# show statistically significant sex-related differences in problem solving performance. In later studies other tests were used to measure this variable (Kaufman# 1984; Marshall# 1981# 1984). Kaufman (1983/1984) investigated if sex differences in problem solving# favoring males# exist in the 5th and 6th grades and if these differences were more pronounced in mathematically gifted students than in students of average mathematical ability. The Iowa Test of Basic Skills and a mathematics problem-solving test were administered to 504 subjects. Males in the average group as well as males in the gifted group outperformed females# but only the gifted group showed statistically significant differences. 85 TABLE 4.4 ANCOVA Summary Table: Multiplication Covariate First Year Source df SS MS F Multiplication (M) 1 1195.90 1195.90 265.66* Sex (S) 1 31.95 31.95 7.07* M x S 1 38.80 38.80 8.59* Error 998 4509.00 4.51 * p <;.oi TABLE 4.5 ANCOVA Summary Table: Divison Covariate First Year Source df SS MS F Division (D) 1 1195.90 1195.90 264.66* Sex (S) 1 13.94 13.94 3.22 D x S 1 18.39 18.39 4.25* Error 998 4317.00 4.32 * pC.Ol 7 errors (Harnisch# 1983). Item response patterns are valuable for the identification of large group differences# including district-to-district # school-to-school# and classroom-to-classroom variations on different subsets of items. The response patterns can provide diagnostic information about the type of understanding the student has on various mathematics topics (e.g.r problem solving). Marshall (1981# 1984) has used the item response pattern technique and her findings indicate sex differences in mathematics performance at the item level. Females outperformed males in computation and males outscored females in problem solving. Also# the success of girls in the problem solving items was dependent upon their success in the computation items; for boys# success in the problem solving items did not depend as much on their computational performance. Marshall (1981# 1983) has also reported that males and females differ in the selection of incorrect responses# reflecting differences in reasoning abilities. In Puerto Rico# a high percentage of children promoted to the 7th grade in the public schools does not master the basic skills in mathematics. If 6th grade male and female children can be diagnosed as having different problem solving abilities# as Marshall found with California children# teachers may need to provide tailor-made mathematics instruction for each sex# in order to ensure equal access to formal education and enhance mathematics 11 She analyzed the responses of 9*000 boys and 9*000 girls to 6 selected items* 2 computation items* and 1 story problem item from each of 2 of the 16 forms of the Survey of Basic Skills test administered during the academic year 1978-79. The Survey is a 30-item achievement test administered every year to all 6th grade children in California through the California Assessment Program. There are 16 forms of the test* to which approximately 9 *000 boys and 9 *000 girls respond each year. Of the 160 mathematics items contained in the 16 forms of the test* 32 are on measurement and graphing* 28 on number concepts* 28 on whole number arithmetics* 20 on fraction arithmetics* 20 on decimal arithmetics* 20 on geometry* and 12 on probability and statistics. The item analysis performed on the 1978-79 data showed that boys and girls tended to select different incorrect responses. In the first computation item (Form 1 of the test) both sexes reflected similar mistakes in carrying* but in different columns. In the second computation item* both sexes ignored the decimal points and selected the same incorrect response. However* more girls than boys chose this response. In the first computation item (Form 2 of the test) the incorrect choice of both sexes was option Â£* but the second most frequently selected option was a for boys and b for girls. In the second computation item of this 21 story problem item correctly after giving incorrect responses to both computation items. Apparently mastery of the skills required by the computation items is more important for girls than for boys. If girls cannot solve the computation items they have little chance of solving the related story problem item. For girls the probability of success in the story problem item after giving successful answers to both computation items is almost 2 1/2 times the probability of success after giving incorrect responses to both computation items. For boys the probability of success in the story problem item after successful responses to the computation items is about 1 3/4 times the probability of success on the story problem item after incorrect responses to the computation items. Three years later Marshall (1984) analyzed more in depth these phenomena of sex-related differences. Her interest was twofold. First she wanted to know if there were differences in the rate of success for boys and girls in solving computation and story problem items. Second she examined additional factors that interact with sex to influence mathematics performance such as reading achievement socio-economic status (SES) primary language and chronological age. Two questions were raised: Do the probabilibities of successful solving of computation and story problem items increase with reading score? Are these probabilities different for the two sexes? 55 1 /000 students were selected randomly for each year. For the first year / 1 / 00 2 were selected (492 boys and 510 girls); for the second year/ 1/013 students were selected (504 boys and 509 girls); and/ for the third yearz 1/013 students were selected (509 boys and 504 girls). The student population in Puerto Rico includes children from the urban and rural zones and comprises children from low and middle socioeconomic levels. Findings can be generalized only to this population. The Basic Skills Test in Mathematics-6 is a criterion-referenced test used in the Department ot Education of Puerto Rico as part of the annual assessment program. The test measures academic achievement in operations/ mathematical concepts/ and story problems. It has a reported split half reliability ot .95. The test was designed specifically for Puerto Rico. Its contents and the procedures followed for its development were formulated and reviewed by educators from the mathematics department of the Department of Education of Puerto Rico/ in coordination with the Evaluation Center ot the Department of Education and mathematics teachers from the school districts. The emphasis placed on each skill area is depicted in Figure 3.1. 114 Sherman/ J. (197 9) Predicting mathematics performance in high school girls and boys. Journal of Educational Psychology/ 71/ 242-249. Sherman/ J. (1980) Mathematics/ spatial visualization/ and related factors: Changes in girls and boys/ grades 8-11. Journal of Educational Psychology/ 22/ 476-482. Stein/ A. H. (1971). The effects of sex role standards for achievement and sex role preference on three determinants of achievement motiviation. Developmental Psychology/ 4.* 219-231. Stein/ A. H./ & Smithless/ J. (1969). Age and sex differences in children's sex role standards about achievement/ Developmental Psychology/ 1/ 252-259. Suydam/ M. N. (1971). Research on mathematics education (K-12) reported in 1970. Journal for Research in Mathematics Education/ Z, 257-298. Suydam/ M. N. (1972). Research on mathematics education (K-12) reported in 1971. Journal for Research in Mathematics Education/ 2/ 196-232. Suydam/ M. N. (197 3) Research on mathematics education (K-12) reported in 1972. J Mathematics. Education/ / 205-242. Suydam/ M. N. (1974). Research on mathematics education (K-12) recorded in 1973. Journal for Research in Mathematics Education/ 2/ 238-272. Suydam/ M. N./ & Weaver/ J. F. (1975). Research on mathematics education (K-12) recorded in 1974. Journal for Research in Mathematics Education/ Â£/ 253-282. Verbeke/ K. A. (1983). Sex-related differences in mathematically gifted secondary students: An investigation of selected cognitive/ affective/ and educational factors (Doctoral dissertation/ University of Maryland/ 1982) Dissertation Abstracts International/ 42/ 2267A-2268A. Weiner/ B. (1974). Achievement motivation and attribution theory. Morristown/ N J: General Learning Press. 76 TABLE 4.2. Continued Item Num ber X2 for Model 3.2 X2 for Model 3.3 X2 for Difference Actual Sample Size Minimum Sample Size for Signif. (a) tb) (c) . (d). (e) (f) 18 11.71 15.68 3.97 52/340 264,864 19 9.37 18.77 9.40* 41/363 88,402 20 12.37 13.46 1.09 55/707 1/026,746 21 11.51 19.03 7.52 49/240 131/547 22 9.63 83.74 74.11* 51/724 14,022 23 29.25 153.74 124.49* 60/377 9/744 24 10.35 42.93 32.58* 68/426 42/194 25 35.90 147.40 111.50* 26/707 4/812 26 18.84 47.30 28.46* 56/717 40 /037 27 8.78 222.87 214.09* 78,353 7/353 28 46.00 242.78 196.78* 84/847 8,662 29 61.48 170.22 108.74* 88/396 16/331 30 25.19 53.84 28.65* 87/473 61/338 31 23.26 47.42 24.16* 32/870 27,333 32 15.51 41.69 26.18* 48,777 37/430 33 7.57 48.84 41.27* 39,470 19/214 34 58.25 324.67 266.42* 82,257 6/203 35 6.91 225.66 218.75* 88,069 8,088 Log-linear models were used in the analysis of incorrect responses. The results of the analyses showed that for 100 of the 111 items of the test* males and females selected different incorrect optionsr and this pattern of responses was consistently found during the three years of test administration. However for the vast majority of the 100 items the male-female differences were relatively small considering the fact that the number of subjects needed to obtain statistical significance was very large. The responses of approximately 1000 randomly selected students per academic year were analyzed in the comparison of male and female performance in problem solving. Females outperformed males in problem solving and in six of the seven computational variables. Males showed superiority in equivalence in all the three years but statistical significance was obtained in only one of the years. Analysis of covariance (ANCOVA) was used in the comparison of male and female performance in problem solving after controlling for computational skills. Seven analyses of covariance tests were conducted one for each of the covariates. Estimated true scores for observed scores were used in the analyses. The results tend to show that for examinees with similar levels of computational ix 21 story problem item correctly after giving incorrect responses to both computation items. Apparently* mastery of the skills required by the computation items is more important for girls than for boys. If girls cannot solve the computation items* they have little chance of solving the related story problem item. For girls* the probability of success in the story problem item after giving successful answers to both computation items is almost 2 1/2 times the probability of success after giving incorrect responses to both computation items. For boys* the probability of success in the story problem item after successful responses to the computation items is about 1 3/4 times the probability of success on the story problem item after incorrect responses to the computation items. Three years later* Marshall (1984) analyzed more in depth these phenomena of sex-related differences. Her interest was twofold. First* she wanted to know if there were differences in the rate of success for boys and girls in solving computation and story problem items. Second* she examined additional factors that interact with sex to influence mathematics performance* such as reading achievement* socio-economic status (SES) primary language* and chronological age. Two questions were raised: Do the probabilibities of successful solving of computation and story problem items increase with reading score? Are these probabilities different for the two sexes? 37 dealing with both cognitive and affective variables/ there are no clear-cut findings to render unequivocal support to a particular variable as accounting for these sex-related differences. However/ everything seems to point to the fact that affective/ rather than cognitive variables play a more significant role in the sex-related differences observed in mathematics performance and learning. In most of the studies dealing with affective variables/ findings consistently show that these factors influence mathematics performance in females more than in males. In at least one area/ confidence as learners of mathematics/ Sherman (1980) found that this variable influenced course election more than all the cognitive variables previously discussed. The case for the societal influences on sex roles and expectations to account for the differences in mathematics learning is also supported in one way or another in the studies reported in the literature. Problem Solving Performance and Related Variables Problem solving has been perhaps the most extensively researched area in mathematics education. Published reviews by Kilpatrick (1969)/ Riedesel (1969)/ and Suydam and Weaver (1970-1975) attest to this. Much of the research done has focused on identifying the determinants of problem difficulty and the problem features that influence the solution process. 109 Creswell/ J. L. (1982/ February). Sex-related differences in problem solving in rural black; anglo and Chicago adolescents. Paper presented at the annual meeting of the Southwest Educational Research Association. Austin/ TX. (ERIC Document Reproduction Service No. ED 216 895) Dwyer/ C. A. (1974). Influence of children's sex role standards on reading and arithmentic achievement. Journal of Educational Psychology/ Â£Â£/ 811-816. Eccles/ J. ( 1983 ) Sex differences in mathematics participation. In N. Steinkamp & M. Maehr (Eds.)/ Women in Science (pp. 80-100). Greenwich/ CT: JAI Press. Elton/ C. F./ & Rose/ H. A. (1967). Traditional sex attitudes and discrepant ability measures in college women. jQucpai of CQhhseling Psycology* Hr 538-548. Ernest/ J. (1976). Mai.hem3tics and Sex. Santa Barbara: University of California. Exedisis/ R. H. (1983). An investigation of the relation- ship of reading comprehension/ vocabulary/ mathematical concepts/ and computation on problem solving among angloz black/ and Chicago male and female middle school adolescents (Doctoral Dissertation/ University of Houston/ 1982). ni_s.sgrtatieii_Abstracts International/ 41/ 2264A-2265A. Fennema/ E. (1974). Mathematics learning and the sexes: A review. Journal for Research in Mathematics -Educfltioji/ 5./ 126-139. Fennema/ E. (1975). Mathematics/ spatial ability and the sexes. In E. Fennema/ (Ed.)/ Mathematics: What r_esga-rch_savs about_sex differences (pp. 33-44) . Columbus: Ohio State University/ College of Education. (ERIC Document Reproduction Service No. ED 128-195) Fennema/ E. (1976) Influences of selected cognitive afcfcectiahd_educatpn,al_,variables on sex-related _djj.fergji.ces in.mathematics learning and studying. Madison: University of Wisconsin/ Department of Curriculum and Instruction. (ERIC Document Reproduction Service No. ED 142 906) 100 exception was that/ when equivalence was controlled/ sex-related differences in favor of females/ persisted. The question of whether male-female differences in problem solving depend on computational skills was answered partially/ in the affirmative. Interactions/ indicating the dependence/ were significant only for multiplication and division in the first year and for subtraction in the second year. COMPARISON OF MALES AND FEMALES ON MATH ITEM PERFORMANCE: ANALYSIS OF RESPONSE PATTERNS By SONIA FELICIANO A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1986 Ill METHOD 54 The Sample 54 The Instrument 55 Analysis of the Data 57 Analysis of Sex by Option by Year Cross Classifications 57 Comparison of Males and Females in Problem Solving Performance 66 IV RESULTS 68 Introduction 68 Sex-related Differences in the Selection of Incorrect Responses 68 Sex-related Differences in Problem Solving Performance 81 Summary 93 V DISCUSSION 101 Summary and Interpretation of the Results 101 Implications of the Findings and Suggestions for Further Research 103 Sex-related Differences in Incorrect Responses 104 Sex-related Differences in Problem Solving Performance 105 REFERENCES 107 BIOGRAPHICAL SKETCH 116 iv .UL 59 60 61 62 63 64 65 66 67 68 69 70 71 73 74 76 77 78 79 72 TABLE 4.1 Continued X1 for Model 3.2* Actual Sample Size Min. Sample Size for Significance (b) (c) (d) 50.96 41/829 7/560 66.43 60/494 8/387 21.39 35/632 15/342 55.03 50/699 8/485 158.36 52/068 3/028 88.01 73/196 7/660 56.28 42/429 6/943 50.51 78/708 14/352 34.28 51/862 13/935 48.01 63/223 12/128 75.29 62/511 7/647 44.39 67/528 14/011 14/425 .97 84/368 54** 24.61 53/358 19/969 22.75 84/175 34/077 99.68 70/183 6/485 79.38 66/304 7/693 64.56 76/806 10/957 23.98 48/690 18/700 15 The test is composed of 23 items designed to yield 3 scores: a comprehension score an application score and a problem solving score. The comprehension question ascertains whether a child understands the information given in the item stem. The application question assesses the child's mastery of a prerequisite concept or skill of the problem solving question. The problem solving question poses a question whose solution is not immediately available that is a situation which does not lend itself to the immediate application of a rule or algorithm. The application and problem solving parts of the test may refer to a common unit of information (the item stem) but the questions are independent in that the response to the application question is not used to respond to the problem solving question. Meyer (1978) Whitaker (1976) Fennema and Sherman (1978) and Sherman (1979) have used the Romberg-Wearne test in their studies. Meyer (1978) investigated whether males and females differ in problem solving performance and examined their prerequisite computational skills and mathematical concepts for the problem solving questions. A sample of 179 students from the 4th grade were administered 19 "reference tests" for intellectual abilities and the Romberg-Wearne test. The analysis showed that males and females were not significantly different in the comprehension application and problem solving questions 53 Tne studies reviewed have confirmed the relationship between problem solving performance and attitudes toward problem solving (Carey 1958; Lindgren et al. 1964; Whitaker 1976). However the results reportea in the studies that investigated the relationship between problem solving performance and computation and between reading and problem solving fail to be consistent in their conclusions. Hansen (1944) Chase (I960) Balow (1964) Kmfong and Holtan (1976 1977) and Zalewski (1974; concluded that computation is more strongly related to problem solving than is reading. Martin (1964) Creswell (1982) Ballew and Cunningham (1982) and Muth (1984) concluded that reading aoility and mathematical problem solving show a stronger relationsnip than computation and problem solving. Exedisis's (1983) findings led to the conclusion that the effect or reading and computation in problem solving performance is unimportant. 6 Purpose of the Study The purpose of this study was twofold. The first was to investigate sex-related differences in the selection of incorrect responses in a multiple-choice achievement test in mathematics. For each test item* the research questions were as follows: 1. Is there a difference in the proportion of males and females choosing each incorrect option? 2. Is the same pattern of differences found in data obtained in three different administrations of the test? The second objective was to investigate sex-related differences in test scores in mathematics problem solving. The following questions were studied: 1. Do males and females differ in problem solving performance? 2. Do sex differences in problem solving persist after controlling for computational skills* and does the differential success of males and females on problem solving items depend on their success on the computation items? significenge Qf the Study Item response patterns are very useful techniques in the assessment of mathematics learning and achievement. Total test scores can be very misleading in the assessment of student performance and provide no diagnostic information about the nature and seriousness of student 34 achievement and interest in mathematics result from identification with the masculine role. A study related to the differential cultural reinforcement hypothesis is that of Dwyer (1974). Dwyer examined the relationship between sex role standards (the extent to which an individual considers certain activities appropriate to males or females) and achievement in reading and arithmetic. Students from grades 2/ 4* 6/ 8# 10/ and 12 participated in this study. She found that sex role standards contributed significant variance to reading and arithmetic achievement test scores and that the effect was stronger for males than for females. This led to her conclusion that sex-related differences in reading and arithmetic are more a function of the child's perception of these areas as sex-appropriate or sex-inappropriate than of the child's biological sex# individual preference for masculine and feminine sex roles/ or liking or disliking reading or mathematics. In a study which agrees with the masculine identification hypothesis/ Milton (1957) found that individuals who had received strong masculine orientation performed better in problem solving than individuals who received less masculine orientation. Elton and Rose (1967) found that women with high mathematical aptitude and average verbal aptitude scored higher on the masculinity scale of the Omnibus Personality Inventory (OPI) than those with average scores on both tasks. 4 mathematics preparation. Males who were enrolled or had completed algebra II outperformed the females in computation and problem solving but not in algebra. Males who studied beyond algebra II outscored females on all three subtests: computation/ algebra/ and problem solving (Armstrong/ 1979) . Carpenter/ Lindquist/ Mathews/ and Silver (1984) analyzed the results of the Third National Assessment of Educational Progress (NAEP)/ and compared them with the First and Second Surveys. Between 1978 and 1982/ the differences between the average performance of males and females remained stable at each age level. At ages 9 and 13/ the overall performance of males and females was not significantly different. At age 17/ males scored higher than females by about 3 percentage points. When course background was held constant/ achievement differences still existed at age 17. For each category of course background/ male achievement exceeded female achievement. Consistent with previous assessments/ sex differences in problem solving in favor of males were found for the 17-year-old sample. At ages 9 and 13/ no large differences were found between the sexes within any level of course background. Marshall (1981/ 1984) investigated sex differences in mathematics performance. She found that males and females excel each other in solving different types of problems. 80 TABLE 4.2 Continued Item Num ber X2 for Model 3.2 X2 for Model 3.3 X2 for Difference Actual Sample Size Minimum Sample Size for Signif. (a) (b) (c) (d) (e) (f) 95 40 .59 68.41 27 .82* 83 /969 60/638 96 29.31 212.34 183.03* 74/287 8/154 97 8.60 126.96 118.36* 50/028 8,492 98 29.87 41.77 11.90* 69/052 116/576 99 9.44 9.94 0.50 72/647 2/918/956 100 11.35 41.48 30.13* 60/931 40 /627 101 23.77 315.62 291.85* 83 /782 5/767 102 61.06 437.09 376.03* 75/708 4/045 103 6.12 7.97 1.85 50/663 550/173 104 40 .47 283.94 243 .47* 72/517 5/984 105 28.66 48.59 19.93* 60/712 61/199 106 13 .47 1/682 .71 1/669.34* 71/946 866* 107 13.13 99.20 86.07* 72/482 16 ,918 108 26 .56 79.33 52.77* 86/261 32,840 109 5.73 18.33 12.60* 76/800 112/453 110 17.80 49.11 31.31* 70/980 45 ,544 111 30.71 477.44 446 .73* 86/190 3/876 pC.oi Minimum sample size less than 3/000 30 The issue of sex-related differences in spatial visualization ability as an explanation for sex differences observed in mathematics achievement is less convincing and the findings more contradictory than in the issue of sex differences in formal education. Besides these cognitive issues/ other issues/ mostly affective in nature/ have also been studied in trying to explain the origin of these sex differences in mathematics achievement and learning. The studies dealing with these affective variables are reviewed in the next section. Differentiated Effect of Affective Variables Researchers have attempted to explain the effect of sex differences in internal beliefs/ interests/ and attitudes (affective variables) on mathematics learning and achievement. A brief statement of each explanation precedes the summary of studies conducted that support the explanation. than males/ lack confidence in their ability to learn mathematics and this affects their achievement in mathematics and their election of more advanced mathematics courses. Maccoby and Jacklin (1974) reported that self- confidence in terms of grade expectancy and success in particular tasks was found to be consistently lower in women than in men. In 1978/ Fennema and Sherman reported 8 achievement. Since there are no investigations reported in sex differences in item response patterns in Puerto Rico* research is needed. Organization of the Study A review of the literature on sex differences in mathematics performance is reported in Chapter II. The research methodology is presented in Chapter III. Research questions sample instrument and data analysis are discussed in that chapter. Chapter IV is an exposition of the results of the study. Chapter V contains a summary and interpretation of the results of the study and the implications of the findings together with suggestions for further research. LIST OF FIGURES Figure Page 3.1 Number of Questions by Skill Area in a Basic Skills Test in Mathematics-6 56 4.1 Sex by Multiplication Interaction 87 4.2 Sex by Division Interaction 88 4.3 Sex by Subtraction Interaction 89 vii 29 females may use them in a different way. Meyer (1978) , with an elementary grade sample/ and Pennema and Tartree (1983) / with an intermediate level sampler found that the influence of spatial visualization on solving mathematics problems is subtler and that males and females use their spatial skills differently in solving word story problems (problems that measure problem solving ability or reasoning) Fennema and Tartree (1983) carried out a three-year longitudinal study which showed that girls and boys with equivalent spatial visualization skills did not solve the same number of items/ nor did they use the same processes in solving problems. The results also suggested that a low level of spatial visualization skills was a more debilitating factor for girls than for boys in problem solving performance. Landau (1984) also investigated the relationship between spatial visualization and mathematics achievement. She studied the performance of middle school children in mathematical problems of varying difficulty/ and the extent to which a diagramatic representation is likely to facilitate solution. She found that spatial ability was strongly correlated to mathematical problem solving and that the effect of spatial ability was more influential for females. Females made more use of diagrams in the solution of problems/ reducing the advantage of males over females in problem solving performance. lal 8 13 23 25 28 29 30 31 34 36 37 42 45 48 49 52 54 56 57 71 TABLE 4 .1 Summary of Significant Tests of Model 3.2 X1 for odel 3.2* Actual Sample Size Min. Sample Size for Significance (b) (c) (d) 46 .15 16,796 3,352 25.74 15,718 5,624 29.25 60,377 19,011 35.90 26 ,707 6,852 46 .00 84 ,847 16,988 61.48 88,396 13,242 25.19 87 ,473 31,982 23 .26 32,870 13 ,015 58.25 82 ,257 13 ,006 84.36 86 ,137 9,404 23 .70 37,068 14,405 23 .26 59,140 23 ,417 22.41 63,098 25,932 179.25 97 ,593 5,014 27.82 16,766 5,550 26.01 39,858 14,114 41.41 74,502 16,570 21.62 44,107 18,789 28.51 53 ,075 17 ,146 69 are independent/ conditional upon sex, log m i j k = Ui + Uj + Uk+Uik+Uj k* (3.2) Substantively/ this model implies that the pattern of male and female option choices is consistent over the three years of test administration. The second model indicates that option is independent of sex/ log mj.jk = u i+ u j + uk+ ui k (3.3) This model implies that the pattern of option choice is the same for males and females. In order to determine if males and females differed in the selection of incorrect responses and if these differences were stable across the three years of test administration/ a two-step test was performed. First/ a likelihood ratio test of the adequacy of model (3.2) was conducted for each of the 111 items of the test. Under this test a model fits the data when the value obtained is nonsignificant at a specified alpha level. For 50 (45%) of the items/ model (3.2) fitted the data adequately. For these items the pattern of male and female option choices was consistent for the three years. The values for the 61 items for which the likelihood ratio 109 Creswell/ J. L. (1982/ February). Sex-related differences i..n_problem solving in rural black/ anglo and Chicago adolescents. Paper presented at the annual meeting of the Southwest Educational Research Association. Austin/ TX. (ERIC Document Reproduction Service No. ED 216 895) Dwyer/ C. A. ( 197 4) Influence of children's sex role standards on reading and arithmentic achievement. Journal of Educational Psychology/ Â£Â£/ 811-816. Eccles/ J. ( 1 9 83 ) Sex differences in mathematics participation. In M. Steinkamp & M. Maehr (Eds.)/ Women in Science (pp. 80-100). Greenwich/ CT: JAI Press. Elton/ C. F. / & Rose/ H. A. ( 1967). Traditional sex attitudes and discrepant ability measures in college women. Journal of Counseling Psycology/ M* 538-548. Ernest/ J. (1976). Mathematics and Sex. Santa Barbara: University of California. Exedisisz R. H. (1983). An investigation of the relation ship of reading comprehension/ vocabulary/ mathematical concepts/ and computation on problem solving among anglo/ black/ and Chicago male and female middle school adolescents (Doctoral Dissertation/ University of Houston/ 1982). Dissertation,Abstracts International/ 42., 2264A-2265A. Fennema/ E. (1974). Mathematics learning and the sexes: A review. Journal for Research in Mathematics Education/ / 126-139. Fennema/ E. (1975). Mathematics/ spatial ability and the sexes. In E. Fennema/ (Ed.)/ Mathematics: What research savs about sex differences (pp. 33-44). Columbus: Ohio State University/ College of Education. (ERIC Document Reproduction Service No. ED 128-195) Fennema/ E. (1976) Influences of selected cognitive affective/ and educational variables on sex-related djff_exgnc.e.s^,n_mathematics learning and studying. Madison: University of Wisconsin/ Department of Curriculum and Instruction. (ERIC Document Reproduction Service No. ED 142 906) 93 superiority in problem solving lost statistical significance. Adjusted means, associated with the analyses, are reported in Table 4.9. Summaries of the analyses of covariance for the second year are reported in Table 4.10. The analyses show that females retained their superiority in problem solving after controlling for equivalence. When the controlling variables were addition, multiplication, division, addition of fractions, and subtraction with decimals females superiority in problem solving lost statistical significance. Adjusted means are reported in Table 4.11. The analyses of covariance for the third year are summarized in Table 4.12. The results of the analyses show that females maintained their superiority in problem solving when performance on either subtraction or equivalence was controlled. When addition, multiplication, division, addition of fractions, and subtraction of decimals were the controlling variables, female superiority in problem solving lost statistical significance. Adjusted means are reported in Table 4.13. Summary Sex-related differences in mathematics achievement were the subject of this study. The first objective was to investigate whether males and females differ in the selection of incorrect responses, and if the pattern of 32 attributes success to an internal, stable attribute, such as ability, then one is confident of being successful in the future and will continue to strive in that area. If one attributes success to an external factor such as a teacher, or to an unstable one, such as effort, then one will not be as confident of success in the future and will cease to strive. Failure attribution patterns work this way: if failure is attributed to unstable causes, such as effort, failure can be avoided in the future and the tendency will be to persist in the task. However, if failure is attributed to a stable cause, such as ability, the belief that one cannot avoid failure will remain. Studies reported by Bar-Tal and Frieze (1977) suggest that males and females tend to exhibit different attributional patterns of success and failure. Males tend to attribute their success to internal causes and their failures to external or unstable ones. Females show a different pattern; they tend to attribute success to external or unstable causes and failures to internal ones. The pattern of attributions, success attributed externally and failure attributed internally, has become hypothesized to show a strong effect on mathematics achievement in females. Kloosterman's (1985) study supported these findings. According to Kloosterman, attributional variables appear to be more important achievement mediators for females than for males, as measured by mathematics word problems. More research is needed in this area. 50 The studies reviewed in this section show a positive relationship between reading and problem solving performance* but in the case of Ballew and Cunningham (1982)* this relationship is not viewed singly but rather as one among the interacting factors that produce successful problem solving. The third variable reviewed is the effect of student attitudes toward problem solving on problem solving performance. Many researchers have tried to demonstrate that this variable is a determinant factor in problem solving achievement. Attitudes. Toward, Problem Solving and Problem Solving performance Research studies support the existence of positive and rather stable relationships between student attitudes and achievement in mathematics. Aiken (1970) has suggested that an individual's attitude toward one aspect of the discipline (mathematics) such as problem solving* may be entirely different from his/her attitude toward another phase of the discipline* such as computation. Research* however* has been directed to the use of single* global measures of attitudes toward mathematics rather than to the investigation of attitudes toward a particular phase of the discipline. The studies described below are only part of the few investigations which have examined the relationship between 45 He found that in the 4th grade the correlations between problem solving and abstract verbal reasoning reading comprehension/ arithmetic concepts/ and computation were .61/ .64/ .66/ and .60 respectively/ and .56/ .68/ .69/ and .63 in the 8th grade. When computation was held constant/ the correlation between problem solving and reading was .52 in grade 4 and .54 in grade 8. When reading was held constant the correlation between problem solving and computation was .43 in grade 4 and .42 in grade 8. Creswell (1982) worked with a sample of anglo and black adolescents from Chicago. Each subject was administered the California Achievement test. Multiple regression was used to analyze the data. The analysis showed that reading is more important than computation in predicting student performance in problem solving. Reading accounted for 49.5% of the variance; computation accounted for 14.6% of the variance. Ballew and Cunningham (1982) worked with 6th grade students in an attempt to find what proportion of students have as their main source of difficulty with word problems each of the following factors: a) computation skills/ b) interpretation of the problem/ c) reading and/ d) integrating these skills in the solution of problems. They also wanted to know if a student can be efficiently diagnosed as having one of the four categories as his/her main difficulty with mathematics word problems. 49 number of problems set up correctly* and amount of time spent taking the test. Reading ability and computational ability were both positively correlated with number of correct answers and with number of problems correctly set up* and negatively correlated with test-taking time. Presence of extraneous information was negatively correlated with correct answers and correct set ups and positively correlated with test-taking time. Syntactic complexity was not significantly correlated with any of the performance measures. Results of a multiple regression analysis showed that reading accounted for 46% of the variance in total correct answers and computation accounted for 8%. Reading ability and computational ability uniquely accounted for 14% and 8% of the variance in the number of correct answers* respectively. Extraneous information added significantly to the variance explained in the number of correct answers* but syntactic structure did not. Reading ability accounted for 5% of the variance in test-taking time* but computation did not add significantly to the variance explained by reading. Muth concluded that reading and computation both contribute significantly to success in solving arithmetic word problems* but that reading plays a more significant role than does computation. COMPARISON OF MALES AND FEMALES ON MATH ITEM PERFORMANCE: ANALYSIS OF RESPONSE PATTERNS By SONIA FELICIANO A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1986 36 they show less interest in the subject than counterpart males. These differences in interest are what Hilton and Berglund (1974) suggest to account for sex-related differences in mathematics achievement. Although the perception of the usefulness of mathematics is still an important predictor of course taking for girls* there is a growing similarity between males and females regarding the usefulness of mathematics (Armstrong & Price* 1982; Fennema & Sherman* 1977; Moller* 1982/1983). Armstrong and Price investigated the relative influence of selected factors in sex-related differences in mathematics participation. Both males and females selected usefulness of mathematics as the most important factor in deciaing whether or not to take more mathematics in high school. Moller's study revealed that both males and females based mathematics course-taking decisions on career usefulness. A Fennema and Sherman (1977) study showed only slight differences between males and females in their feelings about the usefulness of mathematics. In her study of this variable among college students* Probert (1983/1984) did not find any sex-related differences either. These have been the main affective variables researched in attempting to explain the underlying causes of sex-related differences observed in mathematics learning and achievement. In spite of the great diversity of studies 39 The findings suggest that reading factors are not as important as arithmetic and mental factors in problem solving performance. However these findings should be taken cautiously* as the content of the Gates tests (used to measure reading) is literary and does not include mathematical material. Chase (1960) studied 15 variables in an effort to find out which ones have significant influence on the aDility to solve verbal mathematics problems. Only computation* reading to note details* and fundamental knowledge were primarily related to problem solving. Computation accounted for 20.4% of the 32% variance directly associated with problem solving. Chase concluded that a pupil's ability in the mechanics of computation* comprehension of the principles that underline the number systems* and the extent to which important items of information are noticed when reading* are good predictors of the student's ability in solving verbal problems. Balow (1964) investigated the importance of reading ability and computation ability in problem solving performance. He objected to the approaches used by other researchers who in their analyses dichotomized research subjects as "poor" or "good" students* and who ignored the recognizea effect of intelligence on reading and on mathematics achievement Balow administered the Stanford 33 Mathematics as a male domain. Mathematics is an activity more closely related to the male sex domain than to the female sex domain (Eccles et al.r 1983). Thus* the mathematical achievement ot boys is higher than that of gi rls. According to John Ernest (1976) in his study Mathematics and Sex> mathematics is a sexist discipline. He attributed sex-related differences in mathematical achievement to the creation by society of sexual stereotypes and attitudes* restrictions* and constraints that promote the idea of the superiority of boys in mathematics. Ernest reported that boys* girls* and teachers* all believe that boys are superior in mathematics* at least by the time students reach adolescence. Bern and Bern (1970) agree and argue that an American woman is trained to "know her place" in society because ot the pervasive sex-role concept which results in differential expectations and socialization practices. Plank and Plank (195 4) were more specific. They discussed two hypotheses related to this view: the differential cultural reinforcement hypothesis and the masculine identification hypothesis. The differential cultural reinforcement hypothesis states that society in general perceives mathematics as a male domain* giving females less encouragement for excelling in it. The masculine identification hypothesis establishes that 80 81 82 83 84 85 86 87 88 89 90 91 93 94 95 96 98 101 102 104 73 TABLE 4.1 Continued x* for Model 3.2* Actual Sample Size Min. Sample Size for Significance lb) (c) (d) 75.01 55,277 6,787 95.51 57 *710 5,565 51.99 71/315 12/633 20.12 81/021 37/088 38.62 78/959 18,830 1240.78 41,180 306** 102.85 39/903 3/573 66.75 48,194 6,650 63.01 38/742 5,663 43.26 39,648 8,441 85.53 53/557 5/767 82.53 62,954 7,025 78.39 80/723 9,484 70.15 87,241 11,454 40.59 83,969 19,053 29.31 74,287 23/343 29.87 69/052 21,291 23.77 83,782 32,462 61.06 75/708 11,419 40.47 72,517 16/503 70 tests were significant at the .01 level are reported in Table 4.1. Also reported in this table are the actual sample sizes and the minimum sample sizes necessary for the likelihood ratio tests to be significant. Of the 61 items* 59 had minimum sample sizes greater than 3*000. Thus* although for the three years both males and females samples had inconsistent option choices on these 61 items* on 59 of the items the inconsistency was relatively minor. Consequently* these 59 items were included in step 2 along with the initial 50 items. In step 2* a likelihood ratio test was performed comparing the adequacy of the models in (3.2) and (3.3). The values associated with models (3.2) and (3.3) for each of the 109 items subjected to step 2 are reported in columns b and c of Table 4.2. Also reported in Table 4.2 is the difference between the two values (see column d). This latter figure is the test statistic for comparing the adequacies of models (3.2) and (3.3). Significant X^ statistics are indicated by asterisks. The X^ statistics were significant for 100 of the items indicating a male-female difference in option choice for these items. In column e of Table 4.2 the actual sample sizes are reported. In column f* the minimum sample sizes necessary for significance are reported. For those 100 items which had significant X^ statistics reported in column d* 94 (94%) had minimum sample sizes greater than 3*000. CHAPTER I INTRODUCTION Sex differences in mathematics learning and achievement have been the subject of intensive research. Research done before 1974 has shown that male performance on mathematical achievement tests is superior to female performance by the time they reach upper elementary or junior high school (Fennema# 1976# p. 2). The literature strongly suggests that at the elementary level females outperform males in computation and males excel in mathematical reasoning (Glennon & Callahan# 1968; Jarvis# 1964; Maccoby# 1966). Since 1974# research findings have been less consistent. Fennema (1974)# after reviewing 36 studies# found that during secondary school or earlier# sex-related differences in mathematics achievement are not so evident# but that when differences are found# they favor males in high level cognitive tasks (problem solving) and females in low level cognitive tasks (computation). As a result of a further review of the literature# Fennema (1977) concluded that at the elementary level# sex-related differences do not exist at all cognitive levels# from computation to problem solving. 1 PAGE 1 &203$5,621 2) 0$/(6 $1' )(0$/(6 21 0$7+ ,7(0 3(5)250$1&( $1$/<6,6 2) 5(63216( 3$77(516 %\ 621,$ )(/,&,$12 $ ',66(57$7,21 35(6(17(' 72 7+( *5$'8$7( 6&+22/ 2) 7+( 81,9(56,7< 2) )/25,'$ ,1 3$57,$/ )8/),//0(17 2) 7+( 5(48,5(0(176 )25 7+( '(*5(( 2) '2&725 2) 3+,/2623+< 81,9(56,7< 2) )/25,'$ PAGE 2 $&.12:/('*0(176 ZRXOG OLNH WR DFNQRZOHGJH UD\ JUDWLWXGH WR VHYHUDO SHRSOH ZKR KDYH LQIOXHQFHG P\ IRUPDO HGXFDWLRQ DQGRU PDGH WKLV VWXG\ SRVVLEOH 0\ VSHFLDO WKDQNV WR 'U -DPHV $OJLQDr &KDLUPDQ RI WKH GRFWRUDO FRPPLWWHHr ZKR FRQWULEXWHG WR WKH GHYHORSPHQW RI P\ ORYH IRU UHVHDUFK DQG VWDWLVWLFV +H KDV EHHQ LQVXSHUDEOH DV SURIHVVRU DQG YDOXHG IULHQG KLV KHOS DQG JXLGDQFH LQ WKH SUHSDUDWLRQ DQG FRPSOHWLRQ RI WKLV VWXG\ ZHUH LQYDOXDEOH H[WHQG P\ WKDQNV WR 'U /LQGD &URFNHU IRU KHU DGYLFH DQG KHOS GXULQJ P\ GRFWRUDO VWXGLHV DW 8QLYHUVLW\ RI )ORULGD 7KDQNV DOVR JR WR 'U 0LFKDHO 1XQQHU\r PHPEHU RI WKH GRFWRUDO FRPPLWWHH 7R 'U :LOVRQ *XHUWLQr ZKR ZDV D IULHQG IRU PH DQG P\ IDPLO\r H[WHQG P\ VSHFLDO WKDQNV 7KDQNV DUH DOVR H[WHQGHG WR 'U $PDOLD &KDUQHFRr SDVW 8QGHUVHFUHWDU\ RI (GXFDWLRQ RI WKH 3XHUWR 5LFR 'HSDUWPHQW RI (GXFDWLRQr IRU KHU FRQWLQXRXV VXSSRUW 7R P\ VLVWHU 1LOGD 6DQWDHOODr ZKR W\SHG WKH WKHVLVr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n PDQFH $WWLWXGHV 7RZDUG 3UREOHP 6ROYLQJ DQG 3UREOHP 6ROYLQJ 3HUIRUPDQFH LLL PAGE 4 ,OO 0(7+2' 7KH 6DPSOH 7KH ,QVWUXPHQW $QDO\VLV RI WKH 'DWD $QDO\VLV RI 6H[ E\ 2SWLRQ E\ PAGE 5 /,67 2) 7$%/(6 7DEOH 3DJH +\SRWKHWLFDO 3UREDELOLWLHV RI 2SWLRQ &KRLFH &RQGLWLRQDO RQ PAGE 6 7DEOH 3DJH $GMXVWHG 0HDQV RQ 3UREOHP 6ROYLQJr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r DQG WR GHWHUPLQH ZKHWKHU WKHVH GLIIHUHQFHV ZHUH FRQVLVWHQW RYHU WKUHH FRQVHFXWLYH DGPLQLVWUDWLRQV RI WKH WHVW $ VHFRQG REMHFWLYH ZDV WR FRPSDUH PDOH DQG IHPDOH SHUIRUPDQFH LQ SUREOHP VROYLQJ DIWHU FRQWUROOLQJ IRU FRPSXWDWLRQDO VNLOOV 7KH UHVSRQVHV RI DOO WK JUDGH VWXGHQWV IURP WKH SXEOLF VFKRROV LQ 3XHUWR 5LFR ZKR WRRN WKH %DVLF 6NLOOV 7HVW LQ 0DWKHPDWLFV 3UXHED GH 'HVWUH]DV %Â£VLFDV HQ 0DWHPÂ£WLFDVf GXULQJ WKUHH DFDGHPLF \HDUV ZHUH XVHG LQ WKH DQDO\VHV UHOHYDQW WR WKH ILUVW REMHFWLYH YLLL PAGE 9 /RJOLQHDU PRGHOV ZHUH XVHG LQ WKH DQDO\VLV RI LQFRUUHFW UHVSRQVHV 7KH UHVXOWV RI WKH DQDO\VHV VKRZHG WKDW IRU RI WKH LWHPV RI WKH WHVWr PDOHV DQG IHPDOHV VHOHFWHG GLIIHUHQW LQFRUUHFW RSWLRQVU DQG WKLV SDWWHUQ RI UHVSRQVHV ZDV FRQVLVWHQWO\ IRXQG GXULQJ WKH WKUHH \HDUV RI WHVW DGPLQLVWUDWLRQ +RZHYHU} IRU WKH YDVW PDMRULW\ RI WKH LWHPV WKH PDOHIHPDOH GLIIHUHQFHV ZHUH UHODWLYHO\ VPDOO} FRQVLGHULQJ WKH IDFW WKDW WKH QXPEHU RI VXEMHFWV QHHGHG WR REWDLQ VWDWLVWLFDO VLJQLILFDQFH ZDV YHU\ ODUJH 7KH UHVSRQVHV RI DSSUR[LPDWHO\ } UDQGRPO\ VHOHFWHG VWXGHQWV SHU DFDGHPLF \HDU ZHUH DQDO\]HG LQ WKH FRPSDULVRQ RI PDOH DQG IHPDOH SHUIRUPDQFH LQ SUREOHP VROYLQJ )HPDOHV RXWSHUIRUPHG PDOHV LQ SUREOHP VROYLQJ DQG LQ VL[ RI WKH VHYHQ FRPSXWDWLRQDO YDULDEOHV 0DOHV VKRZHG VXSHULRULW\ LQ HTXLYDOHQFH LQ DOO WKH WKUHH \HDUV} EXW VWDWLVWLFDO VLJQLILFDQFH ZDV REWDLQHG LQ RQO\ RQH RI WKH \HDUV $QDO\VLV RI FRYDULDQFH $1&29$f ZDV XVHG LQ WKH FRPSDULVRQ RI PDOH DQG IHPDOH SHUIRUPDQFH LQ SUREOHP VROYLQJ DIWHU FRQWUROOLQJ IRU FRPSXWDWLRQDO VNLOOV 6HYHQ DQDO\VHV RI FRYDULDQFH WHVWV ZHUH FRQGXFWHG} RQH IRU HDFK RI WKH FRYDULDWHV (VWLPDWHG WUXH VFRUHV IRU REVHUYHG VFRUHV ZHUH XVHG LQ WKH DQDO\VHV 7KH UHVXOWV WHQG WR VKRZ WKDW IRU H[DPLQHHV ZLWK VLPLODU OHYHOV RI FRPSXWDWLRQDO L[ PAGE 10 VNLOOVr VH[UHODWHG GLIIHUHQFHV LQ SUREOHP VROYLQJ SHUIRUPDQFH GR QRW H[LVW )HPDOHV UHWDLQHG WKHLU VXSHULRULW\ LQ SUREOHP VROYLQJ ZKHQ HTXLYDOHQFH LQ DOO WKUHH \HDUVf DQG VXEWUDFWLRQ LQ RQH \HDUf ZHUH WKH FRQWUROOLQJ YDULDEOHV 7KH TXHVWLRQ RI ZKHWKHU PDOHIHPDOH GLIIHUHQFHV LQ SUREOHP VROYLQJ GHSHQG RQ FRPSXWDWLRQDO VNLOOV ZDV DQVZHUHGr SDUWLDOO\r LQ WKH DIILUPDWLYH [ PAGE 11 &+$37(5 ,1752'8&7,21 6H[ GLIIHUHQFHV LQ PDWKHPDWLFV OHDUQLQJ DQG DFKLHYHPHQW KDYH EHHQ WKH VXEMHFW RI LQWHQVLYH UHVHDUFK 5HVHDUFK GRQH EHIRUH KDV VKRZQ WKDW PDOH SHUIRUPDQFH RQ PDWKHPDWLFDO DFKLHYHPHQW WHVWV LV VXSHULRU WR IHPDOH SHUIRUPDQFH E\ WKH WLPH WKH\ UHDFK XSSHU HOHPHQWDU\ RU MXQLRU KLJK VFKRRO )HQQHPD S f 7KH OLWHUDWXUH VWURQJO\ VXJJHVWV WKDW DW WKH HOHPHQWDU\ OHYHO IHPDOHV RXWSHUIRUP PDOHV LQ FRPSXWDWLRQ DQG PDOHV H[FHO LQ PDWKHPDWLFDO UHDVRQLQJ *OHQQRQ t &DOODKDQ -DUYLV 0DFFRE\ f 6LQFH UHVHDUFK ILQGLQJV KDYH EHHQ OHVV FRQVLVWHQW )HQQHPD f DIWHU UHYLHZLQJ VWXGLHV IRXQG WKDW GXULQJ VHFRQGDU\ VFKRRO RU HDUOLHU VH[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFV DFKLHYHPHQW DUH QRW VR HYLGHQW EXW WKDW ZKHQ GLIIHUHQFHV DUH IRXQG WKH\ IDYRU PDOHV LQ KLJK OHYHO FRJQLWLYH WDVNV SUREOHP VROYLQJf DQG IHPDOHV LQ ORZ OHYHO FRJQLWLYH WDVNV FRPSXWDWLRQf $V D UHVXOW RI D IXUWKHU UHYLHZ RI WKH OLWHUDWXUH )HQQHPD f FRQFOXGHG WKDW DW WKH HOHPHQWDU\ OHYHO VH[UHODWHG GLIIHUHQFHV GR QRW H[LVW DW DOO FRJQLWLYH OHYHOV IURP FRPSXWDWLRQ WR SUREOHP VROYLQJ PAGE 12 0DQ\ YDULDEOHV FRJQLWLYH DIIHFWLYH DQG HGXFDWLRQDO KDYH EHHQ LQYHVWLJDWHG VLQFH LQ UHODWLRQ WR VH[ GLIIHUHQFHV LQ PDWKHPDWLFV OHDUQLQJ DQG DFKLHYHPHQW )HQQHPD DQG 6KHUPDQ f LQYHVWLJDWHG WKH HIIHFW RI GLIIHUHQWLDO IRUPDO PDWKHPDWLFV HGXFDWLRQ $IWHU FRQWUROOLQJ IRU WKH QXPEHU RI \HDUV RI H[SRVXUH WR WKH VWXG\ RI PDWKHPDWLFV WKH\ IRXQG VH[ GLIIHUHQFHV LQ RQO\ WZR RI WKH IRXU VFKRROV XQGHU VWXG\ +RZHYHU LQ WKRVH VFKRROV ZKHUH ER\V VFRUHG KLJKHU WKDQ JLUOV GLIIHUHQFHV ZHUH DOVR IRXQG LQ WKHLU DWWLWXGHV WRZDUG PDWKHPDWLFV +LOWRQ DQG %HUJOXQG f IRXQG VLJQLILFDQW VH[ GLIIHUHQFHV DIWHU FRQWUROOLQJ IRU WKH QXPEHU RI PDWKHPDWLFV FRXUVHV WDNHQ DQG DWWULEXWHG WKHP WR VH[ GLIIHUHQFHV LQ LQWHUHVWV $V WKH ER\Vn LQWHUHVWV LQ VFLHQFH LQFUHDVH UHODWHG WR WKH JLUOVn WKHLU DFKLHYHPHQW LQ PDWKHPDWLFV LQFUHDVHV UHODWLYH WR WKDW RI WKH JLUOV S f :LVH 6WHHO DQG 0F'RQDOG f UHDQDO\]HG WHVW GDWD FROOHFWHG LQ D ORQJLWXGLQDO VWXG\ RI KLJK VFKRRO VWXGHQWV 3URMHFW 7DOHQWf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r IHPDOHV RXWSHUIRUPHG PDOHV LQ FRPSXWDWLRQ DQG VSDWLDO YLVXDOL]DWLRQ )RU WKH WK JUDGH VWXGHQWVr VWDWLVWLFDOO\ VLJQLILFDQW VH[ GLIIHUHQFHV IDYRULQJ PDOHV ZHUH IRXQG LQ SUREOHP VROYLQJr EXW QRW LQ DOJHEUDr FRPSXWDWLRQr RU VSDWLDO YLVXDOL]DWLRQ )RU PDOHV DQG IHPDOHV ZKR KDG HQUROOHG LQ FRXUVHV EH\RQG JHQHUDO PDWKHPDWLFV DQG ZKR KDG WDNHQ RU ZHUH HQUROOHG LQ FRXUVHV VXFK DV SUHFDOFXOXVr FDOFXOXVr RU JHRPHWU\r GLIIHUHQFHV LQ SUREOHP VROYLQJ RU VSDWLDO YLVXDOL]DWLRQ GLG QRW H[LVW 6H[ GLIIHUHQFHV IDYRULQJ PDOHV ZHUH IRXQG RQ D WRWDO VFRUH REWDLQHG VXPPLQJ DFURVV WKH FRPSXWDWLRQr SUREOHP VROYLQJr DQG DOJHEUD VXEWHVWV $UPVWURQJr f 7KH PDWKHPDWLFV GDWD FROOHFWHG LQ WKH VHFRQG VXUYH\ E\ WKH 1DWLRQDO $VVHVVPHQW RI (GXFDWLRQDO 3URJUHVV VKRZHG VLJQLILFDQW VH[ GLIIHUHQFHV IRU ERWK DQG \HDUROG VWXGHQWV 7KH \HDUROG IHPDOHV RXWSHUIRUPHG PDOHV LQ WKH FRPSXWDWLRQDO VXEWHVW DQG PDOHV RXWVFRUHG IHPDOHV E\ SHUFHQWDJH SRLQWV LQ SUREOHP VROYLQJ VWDWLVWLFDOO\ VLJQLILFDQWf 1R VWDWLVWLFDOO\ VLJQLILFDQW GLIIHUHQFHV ZHUH IRXQG LQ DOJHEUD 1R VH[ GLIIHUHQFHV ZHUH IRXQG IRU WKH \HDUROG JURXS HLWKHU LQ WKH FRPSXWDWLRQ VXEWHVW RU LQ WKH DOJHEUD VXEWHVW 0DOHV VXUSDVVHG IHPDOHV LQ SUREOHP VROYLQJ $ UHDQDO\VLV RI WKH GDWD IURP WKH \HDUROG JURXS FRQILUPHG PDOH VXSHULRULW\ LQ SUREOHP VROYLQJ DIWHU FRQWUROOLQJ IRU PAGE 14 PDWKHPDWLFV SUHSDUDWLRQ 0DOHV ZKR ZHUH HQUROOHG RU KDG FRPSOHWHG DOJHEUD ,, RXWSHUIRUPHG WKH IHPDOHV LQ FRPSXWDWLRQ DQG SUREOHP VROYLQJ EXW QRW LQ DOJHEUD 0DOHV ZKR VWXGLHG EH\RQG DOJHEUD ,, RXWVFRUHG IHPDOHV RQ DOO WKUHH VXEWHVWV FRPSXWDWLRQ" DOJHEUD" DQG SUREOHP VROYLQJ $UPVWURQJ" f &DUSHQWHU" /LQGTXLVW" 0DWKHZV" DQG 6LOYHU f DQDO\]HG WKH UHVXOWV RI WKH 7KLUG 1DWLRQDO $VVHVVPHQW RI (GXFDWLRQDO 3URJUHVV 1$(3f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f LQYHVWLJDWHG VH[ GLIIHUHQFHV LQ PDWKHPDWLFV SHUIRUPDQFH 6KH IRXQG WKDW PDOHV DQG IHPDOHV H[FHO HDFK RWKHU LQ VROYLQJ GLIIHUHQW W\SHV RI SUREOHPV PAGE 15 )HPDOHV ZHUH EHWWHU RQ LWHPV RI FRPSXWDWLRQ DQG PDOHV ZHUH PRUH VXFFHVVIXO RQ ZRUGVWRU\ SUREOHP LWHPV SUREOHP VROYLQJf 6KH DOVR IRXQG WKDW IHPDOHV VXFFHVVIXO SHUIRUPDQFH LQ WKH SUREOHP VROYLQJ LWHPV ZDV PRUH GHSHQGHQW RQ WKHLU VXFFHVVIXO SHUIRUPDQFH LQ WKH FRPSXWDWLRQ LWHPV 0DOHV GLG QRW QHHG} DV PXFK DV IHPDOHV} WR VXFFHHG LQ WKH FRPSXWDWLRQ LWHPV LQ RUGHU WR DQVZHU FRUUHFWO\ WKH SUREOHP VROYLQJ LWHPV $OWKRXJK WKH JHQHUDO ILQGLQJV VHHP WR VXSSRUW VH[ GLIIHUHQFHV LQ PDWKHPDWLFV OHDUQLQJ DQG DFKLHYHPHQW} WKH UHVHDUFK GRQH GRHV QRW FRQVLVWHQWO\ VXSSRUW VXSHULRULW\ IRU HLWKHU VH[ 0RVW RI WKH UHVHDUFK KDV EHHQ FRQFHUQHG ZLWK KRZ WKH VH[HV GLIIHU RQ VXEWHVWV RU WRWDO WHVW VFRUHV LQ PDWKHPDWLFV 0RUHRYHU} WKH JUHDW PDMRULW\ RI WKH VWXGLHV GHDO ZLWK FRUUHFW UHVSRQVHV 6H[ GLIIHUHQFHV LQ LQFRUUHFW UHVSRQVHV DW WKH LWHP OHYHO KDYH QRW EHHQ IXOO\ UHVHDUFKHG 2QO\ WZR VWXGLHV GHDOLQJ ZLWK VH[ GLIIHUHQFHV LQ LQFRUUHFW UHVSRQVHV DW WKH LWHP OHYHO ZHUH IRXQG LQ WKH UHVHDUFK OLWHUDWXUH 0DUVKDOO} } f 0DUVKDOO LQYHVWLJDWHG ZKHWKHU ER\V DQG JLUOV PDGH VLPLODU HUURUV LQ FRPSXWDWLRQ DQG VWRU\ SUREOHPV 6KH DQDO\]HG ER\Vn DQG JLUOVn DQVZHUV WR VL[ PDWKHPDWLFV LWHPV DQG IRXQG WKDW WKH VH[HV PDGH GLIIHUHQW HUURUV} SRVVLEO\ UHIOHFWLQJ GLIIHUHQW SUREOHP VROYLQJ VWUDWHJLHV +HU RULJLQDO ILQGLQJV ZHUH VXSSRUWHG ZKHQ VKH VWXGLHG WKH VDPH SUREOHP XVLQJ D ODUJH QXPEHU RI LWHPV WKUHH \HDUV ODWHU PAGE 16 3XUSRVH RI WKH 6WXG\ 7KH SXUSRVH RI WKLV VWXG\ ZDV WZRIROG 7KH ILUVW ZDV WR LQYHVWLJDWH VH[UHODWHG GLIIHUHQFHV LQ WKH VHOHFWLRQ RI LQFRUUHFW UHVSRQVHV LQ D PXOWLSOHFKRLFH DFKLHYHPHQW WHVW LQ PDWKHPDWLFV )RU HDFK WHVW LWHPr WKH UHVHDUFK TXHVWLRQV ZHUH DV IROORZV ,V WKHUH D GLIIHUHQFH LQ WKH SURSRUWLRQ RI PDOHV DQG IHPDOHV FKRRVLQJ HDFK LQFRUUHFW RSWLRQ" ,V WKH VDPH SDWWHUQ RI GLIIHUHQFHV IRXQG LQ GDWD REWDLQHG LQ WKUHH GLIIHUHQW DGPLQLVWUDWLRQV RI WKH WHVW" 7KH VHFRQG REMHFWLYH ZDV WR LQYHVWLJDWH VH[UHODWHG GLIIHUHQFHV LQ WHVW VFRUHV LQ PDWKHPDWLFV SUREOHP VROYLQJ 7KH IROORZLQJ TXHVWLRQV ZHUH VWXGLHG 'R PDOHV DQG IHPDOHV GLIIHU LQ SUREOHP VROYLQJ SHUIRUPDQFH" 'R VH[ GLIIHUHQFHV LQ SUREOHP VROYLQJ SHUVLVW DIWHU FRQWUROOLQJ IRU FRPSXWDWLRQDO VNLOOVr DQG GRHV WKH GLIIHUHQWLDO VXFFHVV RI PDOHV DQG IHPDOHV RQ SUREOHP VROYLQJ LWHPV GHSHQG RQ WKHLU VXFFHVV RQ WKH FRPSXWDWLRQ LWHPV" VLJQLILFHQJH 4I WKH 6WXG\ ,WHP UHVSRQVH SDWWHUQV DUH YHU\ XVHIXO WHFKQLTXHV LQ WKH DVVHVVPHQW RI PDWKHPDWLFV OHDUQLQJ DQG DFKLHYHPHQW 7RWDO WHVW VFRUHV FDQ EH YHU\ PLVOHDGLQJ LQ WKH DVVHVVPHQW RI VWXGHQW SHUIRUPDQFH DQG SURYLGH QR GLDJQRVWLF LQIRUPDWLRQ DERXW WKH QDWXUH DQG VHULRXVQHVV RI VWXGHQW PAGE 17 HUURUV +DUQLVFK f ,WHP UHVSRQVH SDWWHUQV DUH YDOXDEOH IRU WKH LGHQWLILFDWLRQ RI ODUJH JURXS GLIIHUHQFHV LQFOXGLQJ GLVWULFWWRGLVWULFW VFKRROWRVFKRRO DQG FODVVURRPWRFODVVURRP YDULDWLRQV RQ GLIIHUHQW VXEVHWV RI LWHPV 7KH UHVSRQVH SDWWHUQV FDQ SURYLGH GLDJQRVWLF LQIRUPDWLRQ DERXW WKH W\SH RI XQGHUVWDQGLQJ WKH VWXGHQW KDV RQ YDULRXV PDWKHPDWLFV WRSLFV HJU SUREOHP VROYLQJf 0DUVKDOO f KDV XVHG WKH LWHP UHVSRQVH SDWWHUQ WHFKQLTXH DQG KHU ILQGLQJV LQGLFDWH VH[ GLIIHUHQFHV LQ PDWKHPDWLFV SHUIRUPDQFH DW WKH LWHP OHYHO )HPDOHV RXWSHUIRUPHG PDOHV LQ FRPSXWDWLRQ DQG PDOHV RXWVFRUHG IHPDOHV LQ SUREOHP VROYLQJ $OVR WKH VXFFHVV RI JLUOV LQ WKH SUREOHP VROYLQJ LWHPV ZDV GHSHQGHQW XSRQ WKHLU VXFFHVV LQ WKH FRPSXWDWLRQ LWHPV IRU ER\V VXFFHVV LQ WKH SUREOHP VROYLQJ LWHPV GLG QRW GHSHQG DV PXFK RQ WKHLU FRPSXWDWLRQDO SHUIRUPDQFH 0DUVKDOO f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} UHVHDUFK LV QHHGHG 2UJDQL]DWLRQ RI WKH 6WXG\ $ UHYLHZ RI WKH OLWHUDWXUH RQ VH[ GLIIHUHQFHV LQ PDWKHPDWLFV SHUIRUPDQFH LV UHSRUWHG LQ &KDSWHU ,, 7KH UHVHDUFK PHWKRGRORJ\ LV SUHVHQWHG LQ &KDSWHU ,,, 5HVHDUFK TXHVWLRQV} VDPSOH} LQVWUXPHQW} DQG GDWD DQDO\VLV DUH GLVFXVVHG LQ WKDW FKDSWHU &KDSWHU ,9 LV DQ H[SRVLWLRQ RI WKH UHVXOWV RI WKH VWXG\ &KDSWHU 9 FRQWDLQV D VXPPDU\ DQG LQWHUSUHWDWLRQ RI WKH UHVXOWV RI WKH VWXG\ DQG WKH LPSOLFDWLRQV RI WKH ILQGLQJV WRJHWKHU ZLWK VXJJHVWLRQV IRU IXUWKHU UHVHDUFK PAGE 19 &+$37(5 ,, 5(9,(: 2) 7+( /,7(5$785( 6H[ GLIIHUHQFHV LQ PDWKHPDWLFV OHDUQLQJ DQG DFKLHYHPHQW KDYH EHHQ D VXEMHFW RI FRQFHUQ IRU HGXFDWRUV DQG SV\FKRORJLVWV 0DQ\ VWXGLHV IRXQG LQ WKH OLWHUDWXUH VXSSRUW WKH H[LVWHQFH RI WKHVH GLIIHUHQFHV %R\V VKRZ VXSHULRULW\ LQ KLJKHU OHYHO FRJQLWLYH WDVNV SUREOHP VROYLQJ RU PDWKHPDWLFDO UHDVRQLQJf LQ WKH XSSHU HOHPHQWDU\ \HDUV DQG LQ WKH HDUO\ KLJK VFKRRO \HDUV )HQQHPD 0DFFRE\ t -DFNOLQ f $OPRVW DOO WKH UHVHDUFK FDUULHG RXW KDV GHDOW ZLWK DQDO\VLV RI WRWDO FRUUHFW VFRUHV LQ PDWKHPDWLFV DSWLWXGH DQG DFKLHYHPHQW WHVWV RU VFRUHV LQ VXEWHVWV 7KH OLWHUDWXUH UHODWHG WR VH[ GLIIHUHQFHV LQ LQFRUUHFW UHVSRQVHV}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f 0DUVKDOO f LQYHVWLJDWHG ZKHWKHU WK JUDGH ER\V DQG JLUOV DSSURDFK PDWKHPDWLFDO SUREOHP VROYLQJ ZLWK GLIIHUHQW VWUDWHJLHV +HU VSHFLILF LQWHUHVW ZDV ZKHWKHU WKH VH[HV PDGH WKH VDPH HUURUV PAGE 21 6KH DQDO\]HG WKH UHVSRQVHV RI r ER\V DQG r JLUOV WR VHOHFWHG LWHPVr FRPSXWDWLRQ LWHPVr DQG VWRU\ SUREOHP LWHP IURP HDFK RI RI WKH IRUPV RI WKH 6XUYH\ RI %DVLF 6NLOOV WHVW DGPLQLVWHUHG GXULQJ WKH DFDGHPLF \HDU 7KH 6XUYH\ LV D LWHP DFKLHYHPHQW WHVW DGPLQLVWHUHG HYHU\ \HDU WR DOO WK JUDGH FKLOGUHQ LQ &DOLIRUQLD WKURXJK WKH &DOLIRUQLD $VVHVVPHQW 3URJUDP 7KHUH DUH IRUPV RI WKH WHVWr WR ZKLFK DSSUR[LPDWHO\ r ER\V DQG r JLUOV UHVSRQG HDFK \HDU 2I WKH PDWKHPDWLFV LWHPV FRQWDLQHG LQ WKH IRUPV RI WKH WHVWr DUH RQ PHDVXUHPHQW DQG JUDSKLQJr RQ QXPEHU FRQFHSWVr RQ ZKROH QXPEHU DULWKPHWLFVr RQ IUDFWLRQ DULWKPHWLFVr RQ GHFLPDO DULWKPHWLFVr RQ JHRPHWU\r DQG RQ SUREDELOLW\ DQG VWDWLVWLFV 7KH LWHP DQDO\VLV SHUIRUPHG RQ WKH GDWD VKRZHG WKDW ER\V DQG JLUOV WHQGHG WR VHOHFW GLIIHUHQW LQFRUUHFW UHVSRQVHV ,Q WKH ILUVW FRPSXWDWLRQ LWHP )RUP RI WKH WHVWf ERWK VH[HV UHIOHFWHG VLPLODU PLVWDNHV LQ FDUU\LQJr EXW LQ GLIIHUHQW FROXPQV ,Q WKH VHFRQG FRPSXWDWLRQ LWHPr ERWK VH[HV LJQRUHG WKH GHFLPDO SRLQWV DQG VHOHFWHG WKH VDPH LQFRUUHFW UHVSRQVH +RZHYHUr PRUH JLUOV WKDQ ER\V FKRVH WKLV UHVSRQVH ,Q WKH ILUVW FRPSXWDWLRQ LWHP )RUP RI WKH WHVWf WKH LQFRUUHFW FKRLFH RI ERWK VH[HV ZDV RSWLRQ er EXW WKH VHFRQG PRVW IUHTXHQWO\ VHOHFWHG RSWLRQ ZDV D IRU ER\V DQG E IRU JLUOV ,Q WKH VHFRQG FRPSXWDWLRQ LWHP RI WKLV PAGE 22 WHVW IRUP QR VH[ GLIIHUHQFHV ZHUH IRXQG LQ UHVSRQVH SDWWHUQV $SSUR[LPDWHO\ b RI HDFK VH[ VHOHFWHG RSWLRQ e 7KH QH[W SRSXODU FKRLFH IRU ERWK VH[HV ZDV RSWLRQ G VHOHFWHG E\ DSSUR[LPDWHO\ b RI ERWK ER\V DQG JLUOV 2Q WKH VWRU\ SUREOHP RI )RUP PDOHV DQG IHPDOHV UHVSRQGHG DOLNH 7KHLU PRVW SRSXODU LQFRUUHFW UHVSRQVH FKRLFH ZDV RSWLRQ D IRU ERWK PDOHV DQG IHPDOHV 7KH VHFRQG PRVW SRSXODU LQFRUUHFW FKRLFH ZDV RSWLRQ MV IRU ERWK VH[HV 5HVSRQVH WR WKH VWRU\ SUREOHP LQ )RUP VKRZHG VH[ GLIIHUHQFHV LQ UHVSRQVH FKRLFH ,QFOXGLQJ WKH FRUUHFW RSWLRQ b RI WKH JLUOV VHOHFWHG RSWLRQ ML b FKRVH RSWLRQ F DQG b RSWLRQ G )RU PDOHV DSSUR[LPDWHO\ b VHOHFWHG RSWLRQ V DQG WKH VDPH SHUFHQW VHOHFWHG RSWLRQ G 0DUVKDOO FRQFOXGHG WKDW DOWKRXJK WKH DQDO\VLV RI LQFRUUHFW UHVSRQVHV GRHV QRW H[SODLQ ZK\ ER\V DQG JLUOV GLIIHU LQ WKHLU UHVSRQVHV WKH DQDO\VLV VKRZV WKDW ER\V DQG JLUOV DSSURDFK SUREOHPV LQ GLIIHUHQW ZD\V DQG WKHVH YDU\LQJ VWUDWHJLHV FDQ EH XVHIXO LQ LQGHQWLI\LQJ KRZ WKH VH[HV GLIIHU LQ UHDVRQLQJ DELOLWLHV 7ZR \HDUV ODWHU 0DUVKDOO f DQDO\]HG WKH UHVSRQVHV RI DSSUR[LPDWHO\ ER\V DQG JLUOV WR PDWKHPDWLFV LWHPV FRQWDLQHG LQ WKH WHVW IRUPV RI WKH 6XUYH\ RI %DVLF 6NLOOV GXULQJ WKH \HDUV DQG 6KH XVHG ORJOLQHDU PRGHOV H[SODLQHG LQ &KDSWHU PAGE 23 ,,,f WR LQYHVWLJDWH VH[ UHODWHG GLIIHUHQFHV LQ WKH VHOHFWLRQ RI LQFRUUHFW UHVSRQVHV DQG WKH FRQVLVWHQF\ RI VXFK GLIIHUHQFHV RYHU WKUHH \HDUV RI DGPLQLVWUDWLRQ RI WKH WHVW %DVHG RQ KHU ILQGLQJV WKDW VH[ GLIIHUHQFHV ZHUH IRXQG LQ b RI WKH LWHPV 0DUVKDOO FODVVLILHG WKH VWXGHQWVn HUURUV DFFRUGLQJ WR 5DGDW]n f ILYHFDWHJRU\ HUURU FODVVLILFDWLRQ 7KH FDWHJRULHV DUH ODQJXDJH HUURUV LQ VHPDQWLFVf VSDWLDO YLVXDOL]DWLRQ PDVWHU\ DVVRFLDWLRQ DQG XVH RI LUUHOHYDQW UXOHV ,W ZDV IRXQG WKDW JLUOVn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e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f PAGE 25 7KH WHVW LV FRPSRVHG RI LWHPV GHVLJQHG WR \LHOG VFRUHV D FRPSUHKHQVLRQ VFRUH} DQ DSSOLFDWLRQ VFRUH} DQG D SUREOHP VROYLQJ VFRUH 7KH FRPSUHKHQVLRQ TXHVWLRQ DVFHUWDLQV ZKHWKHU D FKLOG XQGHUVWDQGV WKH LQIRUPDWLRQ JLYHQ LQ WKH LWHP VWHP 7KH DSSOLFDWLRQ TXHVWLRQ DVVHVVHV WKH FKLOGnV PDVWHU\ RI D SUHUHTXLVLWH FRQFHSW RU VNLOO RI WKH SUREOHP VROYLQJ TXHVWLRQ 7KH SUREOHP VROYLQJ TXHVWLRQ SRVHV D TXHVWLRQ ZKRVH VROXWLRQ LV QRW LPPHGLDWHO\ DYDLODEOH} WKDW LV} D VLWXDWLRQ ZKLFK GRHV QRW OHQG LWVHOI WR WKH LPPHGLDWH DSSOLFDWLRQ RI D UXOH RU DOJRULWKP 7KH DSSOLFDWLRQ DQG SUREOHP VROYLQJ SDUWV RI WKH WHVW PD\ UHIHU WR D FRPPRQ XQLW RI LQIRUPDWLRQ WKH LWHP VWHPf EXW WKH TXHVWLRQV DUH LQGHSHQGHQW LQ WKDW WKH UHVSRQVH WR WKH DSSOLFDWLRQ TXHVWLRQ LV QRW XVHG WR UHVSRQG WR WKH SUREOHP VROYLQJ TXHVWLRQ 0H\HU f} :KLWDNHU f} )HQQHPD DQG 6KHUPDQ f} DQG 6KHUPDQ f KDYH XVHG WKH 5RPEHUJ:HDUQH WHVW LQ WKHLU VWXGLHV 0H\HU f LQYHVWLJDWHG ZKHWKHU PDOHV DQG IHPDOHV GLIIHU LQ SUREOHP VROYLQJ SHUIRUPDQFH DQG H[DPLQHG WKHLU SUHUHTXLVLWH FRPSXWDWLRQDO VNLOOV DQG PDWKHPDWLFDO FRQFHSWV IRU WKH SUREOHP VROYLQJ TXHVWLRQV $ VDPSOH RI VWXGHQWV IURP WKH WK JUDGH ZHUH DGPLQLVWHUHG UHIHUHQFH WHVWV IRU LQWHOOHFWXDO DELOLWLHV DQG WKH 5RPEHUJ:HDUQH WHVW 7KH DQDO\VLV VKRZHG WKDW PDOHV DQG IHPDOHV ZHUH QRW VLJQLILFDQWO\ GLIIHUHQW LQ WKH FRPSUHKHQVLRQ} DSSOLFDWLRQ} DQG SUREOHP VROYLQJ TXHVWLRQV PAGE 26 RI WKH WHVW 7KH VH[HV GLIIHUHG LQ RQO\ RI WKH UHIHUHQFH WHVWV} 6SDWLDO 5HODWLRQV DQG 3LFWXUH *URXS 1DPH6HOHFWLRQ $ IDFWRU DQDO\VLV} KRZHYHU} VKRZHG GLIIHUHQFHV LQ WKH QXPEHU DQG FRPSRVLWLRQ RI WKH IDFWRUV )RU IHPDOHV} D JHQHUDO PDWKHPDWLFV IDFWRU ZDV GHWHUPLQHG E\ PDWKHPDWLFV FRPSXWDWLRQ} FRPSUHKHQVLRQ} DSSOLFDWLRQ} DQG SUREOHP VROYLQJ )RU PDOHV} WKH FRPSUHKHQVLRQ DQG DSSOLFDWLRQ SDUWV GHWHUPLQHG RQH IDFWRU SUREOHP VROYLQJ ZLWK WZR RWKHU UHIHUHQFH WHVWV *HVWDOW DQG 2PHOHWf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} EXW PD\ KDYH XVHG PRUH RI D *HVWDOW DSSURDFK WR WKH SUREOHP VROYLQJ VLWXDWLRQ :KLWDNHU f LQYHVWLJDWHG WKH UHODWLRQVKLS EHWZHHQ WKH PDWKHPDWLFDO SUREOHP SHUIRUPDQFH RI WK JUDGH FKLOGUHQ DQG WKHLU DWWLWXGHV WRZDUG SUREOHP VROYLQJ} WKHLU WHDFKHUVn PAGE 27 DWWLWXGH WRZDUG PDWKHPDWLFDO SUREOHP VROYLQJr DQG UHODWHG VH[ DQG SURJUDPW\SH GLIIHUHQFHV $OWKRXJK KLV PDLQ REMHFWLYH ZDV WR FRQVWUXFW DQ DWWLWXGH VFDOH WR PHDVXUH DWWLWXGHV WRZDUG SUREOHP VROYLQJr KLV VWXG\ LV LPSRUWDQW EHFDXVH KLV ILQGLQJV VXSSRUW 0H\HUnV UHJDUGLQJ WKH ODFN RI VLJQLILFDQW VH[UHODWHG GLIIHUHQFHV LQ SUREOHP VROYLQJ SHUIRUPDQFH 3HUIRUPDQFH LQ WKH SUREOHP VROYLQJ TXHVWLRQVr IRU ERWK PDOHV DQG IHPDOHVr ZDV PXFK ORZHU WKDQ SHUIRUPDQFH LQ WKH DSSOLFDWLRQ TXHVWLRQVr DQG PXFK ORZHU WKDQ SHUIRUPDQFH LQ WKH FRPSUHKHQVLRQ TXHVWLRQV ,Q IDFWr WKH PHDQ VFRUH IRU HDFK SDUW RI WKH LWHPr IRU ERWK PDOHV DQG IHPDOHVr ZDV DOPRVW LGHQWLFDO WR WKH PHDQ VFRUHV REWDLQHG E\ PDOHV DQG IHPDOHV LQ 0H\HUnV VWXG\ :KLWDNHU QRWHG WKDW HDFK DSSOLFDWLRQ LWHP LV PRUH GLIILFXOW WKDQ LWV SUHFHGLQJ FRPSUHKHQVLRQ LWHPr DQG WKDW HDFK SUREOHP VROYLQJ LWHP LV PRUH GLIILFXOW WKDQ LWV SUHFHGLQJ DSSOLFDWLRQ LWHP 1R VLJQLILFDQW VH[UHODWHG GLIIHUHQFHV ZHUH IRXQG IRU DQ\ RI WKH WKUHH SDUWV RI WKH LWHP FRPSUHKHQVLRQr DSSOLFDWLRQr RU SUREOHP VROYLQJf )HQQHPD DQG 6KHUPDQ f LQYHVWLJDWHG VH[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFV DFKLHYHPHQW DQG FRJQLWLYH DQG DIIHFWLYH YDULDEOHV UHODWHG WR VXFK GLIIHUHQFHV 7KH\ DGPLQLVWHUHG WKH 5RPEHUJ:HDUQH 3UREOHP 6ROYLQJ 7HVW WR D UHSUHVHQWDWLYH VDPSOH RI VWXGHQWV JUDGHV f IURP 0DGLVRQr :LVFRQVLQr SUHGRPLQDQWO\ PLGGOHFODVVr EXW LQFOXGLQJ JUHDW GLYHUVLW\ LQ 6(6 7KH VDPSOH FRQVLVWHG RI PAGE 28 VWXGHQWV ZKR KDG WDNHQ D VLPLODU QXPEHU RI PDWKHPDWLFV FRXUVHV DQG ZHUH LQ WKH WRS b RI WKH FODVV LQ PDWKHPDWLFV DFKLHYHPHQW 7KH\ ZHUH WHVWHG LQ )RXU KLJK VFKRRO GLVWULFWV ZHUH LQFOXGHG ,Q RQO\ RQH RI WKH KLJK VFKRRO GLVWULFWV ZHUH VH[UHODWHG GLIIHUHQFHV LQ DSSOLFDWLRQ DQG SUREOHP VROYLQJ IRXQGr LQ IDYRU RI PDOHV 7KH\ FRQFOXGHG WKDW ZKHQ UHOHYDQW IDFWRUV DUH FRQWUROOHGr VH[UHODWHG GLIIHUHQFHV LQ IDYRU RI PDOHV GR QRW DSSHDU RIWHQr DQG ZKHQ WKH\ GRr WKH\ DUH QRW ODUJH 6KHUPDQ f LQYHVWLJDWHG WKH FDXVHV RI WKH HPHUJLQJ VH[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFV SHUIRUPDQFHr LQ IDYRU RI PDOHVr GXULQJ DGROHVFHQFH JUDGHV f 6KH ZDQWHG WR NQRZ LI WKHVH GLIIHUHQFHV HPHUJH DV D IXQFWLRQ RI VH[UHODWHG GLIIHUHQFHV LQ VSDWLDO YLVXDOL]DWLRQ DQG VRFLRFXOWXUDO LQIOXHQFHV WKDW FRQVLGHU PDWK DV D PDOH GRPDLQ ,Q JUDGH r VKH XVHG WKH 5RPEHUJ:HDUQH 7HVW DQGr LQ JUDGH r D PDWKHPDWLFDO SUREOHP VROYLQJ WHVW GHULYHG IURP WKH )UHQFK .LW RI 7HVWV 7KH DQDO\VLV VKRZHG WKDW IRU JLUOVr SUREOHP VROYLQJ SHUIRUPDQFH UHPDLQHG VWDEOH DFURVV WKH \HDUV 0HDQ SUREOHP VROYLQJ SHUIRUPDQFH IRU ER\Vr KRZHYHUr ZDV KLJKHU LQ JUDGH WKDQ LQ JUDGH 1R VH[UHODWHG GLIIHUHQFHV ZHUH IRXQG LQ JUDGH r EXW ER\V RXWSHUIRUPHG JLUOV LQ JUDGH r ZKHUH WKH 6WDIIRUG WHVW ZDV XVHG 6KHUPDQ IRXQG WKDW IRU ERWK VH[HV SUREOHP VROYLQJ SHUIRUPDQFH LQ JUDGH ZDV WKH EHVW SUHGLFWRU RI SUREOHP PAGE 29 VROYLQJ SHUIRUPDQFH LQ JUDGH 6SDWLDO YLVXDOL]DWLRQ ZDV D VWURQJHU SUHGLFWRU IRU JLUOV WKDQ IRU ER\V 0DWKHPDWLFV DV D PDOH GRPDLQ ZDV D JRRG SUHGLFWRU IRU JLUOV RQO\ WKH OHVV D JLUO VWHUHRW\SHG PDWKHPDWLFV DV D PDOH GRPDLQ LQ JUDGH WKH KLJKHU KHU SUREOHP VROYLQJ VFRUH LQ JUDGH $WWLWXGHV WRZDUG VXFFHVV LQ PDWKHPDWLFV LQ JUDGH ZDV D PRUH SRVLWLYH SUHGLFWRU RI SUREOHP VROYLQJ SHUIRUPDQFH IRU ER\V WKDQ IRU JLUOV WKH PRUH SRVLWLYH WKH DWWLWXGHV WRZDUG VXFFHVV LQ PDWKHPDWLFV LQ JUDGH WKH KLJKHU WKHLU SHUIRUPDQFH LQ SUREOHP VROYLQJ LQ JUDGH 1RQH RI WKHVH IRXU VWXGLHV DOO RI ZKLFK XVHG WKH 5RPEHUJ:HDUQH 0DWKHPDWLFV 3UREOHP 6ROYLQJ 7HVW VKRZ VWDWLVWLFDOO\ VLJQLILFDQW VH[UHODWHG GLIIHUHQFHV LQ SUREOHP VROYLQJ SHUIRUPDQFH ,Q ODWHU VWXGLHV RWKHU WHVWV ZHUH XVHG WR PHDVXUH WKLV YDULDEOH .DXIPDQ 0DUVKDOO f .DXIPDQ f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} 0DUVKDOO f FRQFOXGHG WKDW VH[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFV SHUIRUPDQFH PD\ EH WKH UHVXOW RI FRPSDULQJ WKH VH[HV RQ WRWDO WHVW VFRUHV ,I WKH WHVW FRQWDLQV PRUH FRPSXWDWLRQ LWHPV WKDQ SUREOHP VROYLQJ LWHPV} JLUOV ZLOO SHUIRUP EHWWHU WKDQ ER\V} EXW LI WKH WHVW FRQWDLQV PRUH SUREOHP VROYLQJ LWHPV WKDQ FRPSXWDWLRQ RQHV} ER\V ZLOO RXWSHUIRUP JLUOV :LWK WKLV LQ PLQG} 0DUVKDOO LQYHVWLJDWHG VH[UHODWHG GLIIHUHQFHV LQ FRPSXWDWLRQ DQG SUREOHP VROYLQJ E\ DQDO\]LQJ WKH UHVSRQVHV RI DSSUR[LPDWHO\ } VWXGHQWV IURP JUDGH ZKR KDG EHHQ DGPLQLVWHUHG WKH 6XUYH\ RI %DVLF 6NLOOV 7HVW *UDGH } GXULQJ WKH DFDGHPLF \HDU 7ZR RI WKH WHVW IRUPV RI WKH 6XUYH\ ZHUH XVHG WR DVVHVV VNLOOV VXFK DV FRQFHSWV RI ZKROH QXPEHUV} IUDFWLRQV} DQG GHFLPDOV 7KHVH VNLOOV ZHUH WHVWHG ERWK DV VLPSOH FRPSXWDWLRQV DQG DV VWRU\ SUREOHPV SUREOHP VROYLQJf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} PDVWHU\ RI WKH VNLOOV UHTXLUHG E\ WKH FRPSXWDWLRQ LWHPV LV PRUH LPSRUWDQW IRU JLUOV WKDQ IRU ER\V ,I JLUOV FDQQRW VROYH WKH FRPSXWDWLRQ LWHPV} WKH\ KDYH OLWWOH FKDQFH RI VROYLQJ WKH UHODWHG VWRU\ SUREOHP LWHP )RU JLUOV} WKH SUREDELOLW\ RI VXFFHVV LQ WKH VWRU\ SUREOHP LWHP DIWHU JLYLQJ VXFFHVVIXO DQVZHUV WR ERWK FRPSXWDWLRQ LWHPV LV DOPRVW WLPHV WKH SUREDELOLW\ RI VXFFHVV DIWHU JLYLQJ LQFRUUHFW UHVSRQVHV WR ERWK FRPSXWDWLRQ LWHPV )RU ER\V} WKH SUREDELOLW\ RI VXFFHVV LQ WKH VWRU\ SUREOHP LWHP DIWHU VXFFHVVIXO UHVSRQVHV WR WKH FRPSXWDWLRQ LWHPV LV DERXW WLPHV WKH SUREDELOLW\ RI VXFFHVV RQ WKH VWRU\ SUREOHP LWHP DIWHU LQFRUUHFW UHVSRQVHV WR WKH FRPSXWDWLRQ LWHPV 7KUHH \HDUV ODWHU} 0DUVKDOO f DQDO\]HG PRUH LQ GHSWK WKHVH SKHQRPHQD RI VH[UHODWHG GLIIHUHQFHV +HU LQWHUHVW ZDV WZRIROG )LUVW} VKH ZDQWHG WR NQRZ LI WKHUH ZHUH GLIIHUHQFHV LQ WKH UDWH RI VXFFHVV IRU ER\V DQG JLUOV LQ VROYLQJ FRPSXWDWLRQ DQG VWRU\ SUREOHP LWHPV 6HFRQG} VKH H[DPLQHG DGGLWLRQDO IDFWRUV WKDW LQWHUDFW ZLWK VH[ WR LQIOXHQFH PDWKHPDWLFV SHUIRUPDQFH} VXFK DV UHDGLQJ DFKLHYHPHQW} VRFLRHFRQRPLF VWDWXV 6(6f} SULPDU\ ODQJXDJH}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f $UPVWURQJ f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f DQG IHPDOHV DUH EHWWHU WKDQ WKHLU FRXQWHUSDUW PDOHV LQ FRPSXWDWLRQ D ORZHU OHYHO VNLOOf 0DUVKDOOnV UHVHDUFK DOVR EURXJKW RXW D GLIIHUHQW DVSHFW RI WKLV TXHVWLRQ WKH QRWLRQ WKDW JLUOV ILQG LW PRUH QHFHVVDU\ WKDQ ER\V WR VXFFHHG LQ WKH FRPSXWDWLRQ LWHPV LQ RUGHU WR VXFFHVVIXOO\ VROYH WKH VWRU\ SUREOHP LWHPV e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e-&HSFJJ LQ )4IPDOB 0WKHPWLFVB-cGXFDWLRQ 'LIIHUHQWLDO &RXUVHZRUN +\SRWKHVLVf 7KH EDVLV IRU WKH GLIIHUHQWLDO FRXUVHZRUN K\SRWKHVLV LV WKH IDFW WKDW VH[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFDO OHDUQLQJ DQG DFKLHYHPHQW VKRZ XS ZKHQ FRPSDULQJ JURXSV ZKLFK DUH QRW HTXDO LQ SUHYLRXV PDWKHPDWLFV OHDUQLQJ $IWHU WKH WK JUDGH ER\V WHQG WR VHOHFW PDWKHPDWLFV FRXUVHV PRUH RIWHQ WKDQ JLUOV 7KHUHIRUH JLUOV VKRZ ORZHU DFKLHYHPHQW VFRUHV LQ PDWKHPDWLFV WHVWV EHFDXVH WKHLU PDWKHPDWLFV H[SHULHQFH LV QRW DV VWURQJ DV WKH ER\Vn )HQQHPD )HQQHPD t 6KHUPDQ 6KHUPDQ f )HQQHPD DQG 6KHUPDQnV VWXG\ f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} GLIIHUHQFHV EHWZHHQ WKH VH[HV ZHUH DOVR IRXQG LQ WKHLU DWWLWXGHV WRZDUG PDWKHPDWLFV 5HVHDUFKHUV OLNH %DFNPDQ f U ZKR DQDO\]HG GDWD IURP 3URMHFW 7DOHQW} DQG $OOHQ DQG &KDPEHUV f KDYH DOVR K\SRWKHVL]HG WKDW VH[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFV DFKLHYHPHQW PD\ EH UHODWHG WR GLIIHUHQW FXUULFXOD IROORZHG E\ PDOHV DQG IHPDOHV $OOHQ DQG &KDPEHUV DWWULEXWHG PDOH VXSHULRULW\ LQ PDWKHPDWLFV SUREOHP VROYLQJ WR GLIIHUHQFHV LQ WKH QXPEHU RI PDWKHPDWLFV FRXUVHV WDNHQ LQ KLJK VFKRRO 7KLV LVVXH KDV EHHQ VHULRXVO\ TXHVWLRQHG E\ $VWLQ f} )R[ D} Ef} DQG %HQERZ DQG 6WDQOH\ f} DPRQJ RWKHUV $VWLQ DQG )R[ KDYH UHSRUWHG ODUJH GLIIHUHQFHV LQ IDYRU RI PDOHV DPRQJ JLIWHG VWXGHQWV WDNLQJ WKH 6FKRODVWLF $FKLHYHPHQW 7HVW 7KHVH GLIIHUHQFHV RFFXU DV HDUO\ DV JUDGH } ZKHQ WKHUH DUH QR VH[ GLIIHUHQFHV LQ WKH QXPEHU RI FRXUVHV WDNHQ %HQERZ DQG 6WDQOH\ f FRPSDUHG PDWKHPDWLFDOO\ SUHFRFLRXV ER\V DQG JLUOV LQ WKH WK JUDGH} ZLWK VLPLODU PDWKHPDWLFV EDFNJURXQG} DQG IRXQG VL]HDEOH VH[UHODWHG GLIIHUHQFHV IDYRULQJ ER\V LQ PDWKHPDWLFDO UHDVRQLQJ DELOLW\ )LYH \HDUV ODWHU} WKH\ FRQGXFWHG D IROORZXS VWXG\ ZKLFK VKRZHG WKDW ER\V PAGE 36 PDLQWDLQHG WKHLU VXSHULRULW\ LQ PDWKHPDWLFV DELOLW\ GXULQJ KLJK VFKRRO :KLOH )R[ DWWULEXWHG VH[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFDO DFKLHYHPHQW WR GLIIHUHQWLDO H[SRVXUH WR PDWKHPDWLFDO JDPHV DQG DFWLYLWLHV RXWVLGH VFKRRO} %HQERZ DQG 6WDQOH\ VXJJHVWHG WKDW VH[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFV SHUIRUPDQFH VWHP IURP VXSHULRU PDWKHPDWLFDO DELOLW\ LQ PDOHV} QRW IURP GLIIHUHQFHV LQ PDWKHPDWLFV IRUPDO HGXFDWLRQ 7KH GLIIHUHQWLDO FRXUVHZRUN K\SRWKHVLV LV QRW WRWDOO\ FRQYLQFLQJ DQG} DV UHSRUWHX EHURUH} LW KDV EHHQ FKDOOHQJHG E\ UHVHDUFKHUV VXFK DV %HQERZ DQG 6WDQOH\ f +RZHYHU} 3DOODV DQG $OH[DQGHU f KDYH TXHVWLRQHX WKH JHQHUDOL]DQLOLW\ RI %HQERZ DQG 6WDQOH\nV ILQGLQJV EDVHG RQ WKH IDFW WKDW WKH\ XVHG KLJKO\ SUHFRFLRXV OHDUQHUV 7KH GLIIHUHQWLDO FRXUVHZRUN K\SRWKHVLV FDQ EH DFFHSWHG RQO\ DV D SDUWLDO H[SODQDWLRQ RI GLIIHUHQFHV LQ PDWKHPDWLFV SHUIRUPDQFH IRXQG EHWZHHQ WKH VH[HV -8-eeMAJeV-QB6SIOWLDO $ELOLW\ 7KH EDVLF SUHPLVH LQ WKLV LVVXH LV WKDW PDOHV DQG IHPDOHV GLIIHU LQ VSDWLDO YLVXDOL]DWLRQ DQG WKLV H[SODLQV GLIIHUHQWLDO PDWKHPDWLFV OHDUQLQJ DQG DFKLHYHPHQW 8QWLO UHFHQWO\} VH[ GLIIHUHQFHV LQ VSDWLDO DELOLW\ LQ IDYRU RI PDOHV ZHUH EHOLHYHG WR EH D IDFW DQG ZHUH WKRXJKW E\ VRPH WR EH UHODWHG WR VH[ GLIIHUHQFHV LQ PDWKHPDWLFDO DFKLHYHPHQW PAGE 37 5HVHDUFK ILQGLQJV LQ WKLV DUHD KDYH EHHQ LQFRQVLVWHQW ,Q 0DFFRE\ VWDWHG WKDW E\ HDUO\ VFKRRO \HDUV ER\V FRQVLVWHQWO\ GR EHWWHU WKDQ JLUOVf RQ VSDWLDO WDVNV DQG WKLV GLIIHUHQFH FRQWLQXHV WKURXJK WKH KLJK VFKRRO DQG FROOHJH \HDUV Sf ,Q 0DFFRE\ DQG -DFNOLQ VDLG WKDW WKH GLIIHUHQFHV LQ VSDWLDO DELOLW\ EHWZHHQ WKH VH[HV UHPDLQ PLQLPDO DQG LQFRQVLVWHQW XQWLO DSSUR[LPDWHO\ WKH DJHV RI RU ZKHQ WKH VXSHULRULW\ RI ER\V EHFRPHV FRQVLVWHQW LQ D ZLGH UDQJH RI SRSXODWLRQV DQG WHVWV Sf ,Q DIWHU D FRPSUHKHQVLYH OLWHUDWXUH VHDUFK 0DFFRE\ DQG -DFNOLQ FRQFOXGHG WKDW VH[ GLIIHUHQFHV LQ VSDWLDO YLVXDOL]DWLRQ EHFRPH PRUH SURQRXQFHG EHWZHHQ XSSHU HOHPHQWDU\ \HDUV DQG WKH ODVW \HDU RI KLJK VFKRRO WKH \HDUV ZKHQ VH[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFV DFKLHYHPHQW IDYRULQJ ER\V HPHUJH *XD\ DQG 0F'DQLHO f VXSSRUWHG LQ SDUW 0DFFRE\ DQG -DFNOLQnV ILQGLQJV 7KH\ IRXQG WKDW DPRQJ HOHPHQWDU\ VFKRRO FKLOGUHQ PDOHV KDG JUHDWHU KLJK OHYHO VSDWLDO DELOLW\ WKDQ IHPDOHV EXW WKDW PDOHV DQG IHPDOHV ZHUH HTXDO LQ ORZ OHYHO VSDWLDO DELOLW\ 7KLV ILQGLQJ LV LQFRQVLVWHQW ZLWK WKDW SRUWLRQ RI 0DFFRE\ DQG -DFNOLQnV UHYLHZ WKDW VXJJHVWV WKDW VH[ GLIIHUHQFHV EHFRPH HYLGHQW RQO\ GXULQJ HDUO\ DGROHVFHQFH &RKHQ DQG :LONLH f KRZHYHU VWDWHG WKDW LQ WHVWV PHDVXULQJ GLVWLQFW VSDWLDO WDVNV PDOHV SHUIRUP EHWWHU WKDQ IHPDOHV LQ HDUO\ DGROHVFHQFH DQG WKURXJKRXW WKHLU OLIH VSDQ 0RVW VWXGLHV FDUULHG RXW DIWHU PAGE 38 KDYH IDLOHG WR VXSSRUW WKHVH VH[ GLIIHUHQFHV LQ VSDWLDO DELOLWLHV $UPVWURQJr &RQQRUr 6HUELQr t 6FKDFNPDQr )HQQHPD t 6KHUPDQr 6KHUPDQr f )HQQHPD DQG 6KHUPDQ f DQG 6KHUPDQ f KDYH H[SORUHG VH[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFDO DFKLHYHPHQW DQG FRJQLWLYH DQG DIIHFWLYH YDULDEOHV UHODWHG WR WKHVH GLIIHUHQFHV ,Q D VWXG\ LQYROYLQJ VWXGHQWV IURP JUDGHV r r DQG r IURP IRXU VFKRRO GLVWULFWVr )HQQHPD DQG 6KHUPDQ IRXQG WKDW VSDWLDO YLVXDOL]DWLRQ DQG SUREOHP VROYLQJ ZHUH KLJKO\ FRUUHODWHG IRU ERWK VH[HV DQG f (YHQ LQ WKH VFKRRO GLVWULFW ZKHUH VH[ GLIIHUHQFHV ZHUH IRXQG LQ SUREOHP VROYLQJr QR VLJQLILFDQW VH[UHODWHG GLIIHUHQFHV ZHUH IRXQG LQ VSDWLDO YLVXDOL]DWLRQ :KHQ 6KHUPDQ f FRPSDUHG JURXSV RI PDOHV DQG IHPDOHV LQ WZR GLIIHUHQW JUDGHVr DQG r VKH IRXQG QR VH[UHODWHG GLIIHUHQFHV LQ SUREOHP VROYLQJ RU LQ VSDWLDO YLVXDOL]DWLRQ LQ JUDGH ,Q JUDGH r KRZHYHUr DOWKRXJK WKH VH[HV GLIIHUHG LQ WKHLU SUREOHP VROYLQJ SHUIRUPDQFHr QR VH[UHODWHG GLIIHUHQFHV ZHUH IRXQG LQ VSDWLDO YLVXDOL]DWLRQ (YHQ WKRXJK VSDWLDO YLVXDOL]DWLRQ LQ JUDGH ZDV WKH VHFRQG EHVW SUHGLFWRU RI SUREOHP VROYLQJ SHUIRUPDQFH LQ JUDGH r VH[ GLIIHUHQFHV LQ JUDGH ZHUH QRW D UHVXOW RI VSDWLDO YLVXDOL]DWLRQ VLQFH QR GLIIHUHQFHV ZHUH IRXQG LQ WKDW VNLOO ,Q VSLWH RI WKH IDFW WKDW QR VH[ GLIIHUHQFHV ZHUH IRXQG LQ VSDWLDO DELOLWLHVr LW LV HYLGHQW WKDW PDOHV DQG PAGE 39 IHPDOHV PD\ XVH WKHP LQ D GLIIHUHQW ZD\ 0H\HU f ZLWK DQ HOHPHQWDU\ JUDGH VDPSOH DQG 3HQQHPD DQG 7DUWUHH f ZLWK DQ LQWHUPHGLDWH OHYHO VDPSOHU IRXQG WKDW WKH LQIOXHQFH RI VSDWLDO YLVXDOL]DWLRQ RQ VROYLQJ PDWKHPDWLFV SUREOHPV LV VXEWOHU DQG WKDW PDOHV DQG IHPDOHV XVH WKHLU VSDWLDO VNLOOV GLIIHUHQWO\ LQ VROYLQJ ZRUG VWRU\ SUREOHPV SUREOHPV WKDW PHDVXUH SUREOHP VROYLQJ DELOLW\ RU UHDVRQLQJf )HQQHPD DQG 7DUWUHH f FDUULHG RXW D WKUHH\HDU ORQJLWXGLQDO VWXG\ ZKLFK VKRZHG WKDW JLUOV DQG ER\V ZLWK HTXLYDOHQW VSDWLDO YLVXDOL]DWLRQ VNLOOV GLG QRW VROYH WKH VDPH QXPEHU RI LWHPV QRU GLG WKH\ XVH WKH VDPH SURFHVVHV LQ VROYLQJ SUREOHPV 7KH UHVXOWV DOVR VXJJHVWHG WKDW D ORZ OHYHO RI VSDWLDO YLVXDOL]DWLRQ VNLOOV ZDV D PRUH GHELOLWDWLQJ IDFWRU IRU JLUOV WKDQ IRU ER\V LQ SUREOHP VROYLQJ SHUIRUPDQFH /DQGDX f DOVR LQYHVWLJDWHG WKH UHODWLRQVKLS EHWZHHQ VSDWLDO YLVXDOL]DWLRQ DQG PDWKHPDWLFV DFKLHYHPHQW 6KH VWXGLHG WKH SHUIRUPDQFH RI PLGGOH VFKRRO FKLOGUHQ LQ PDWKHPDWLFDO SUREOHPV RI YDU\LQJ GLIILFXOW\ DQG WKH H[WHQW WR ZKLFK D GLDJUDPDWLF UHSUHVHQWDWLRQ LV OLNHO\ WR IDFLOLWDWH VROXWLRQ 6KH IRXQG WKDW VSDWLDO DELOLW\ ZDV VWURQJO\ FRUUHODWHG WR PDWKHPDWLFDO SUREOHP VROYLQJ DQG WKDW WKH HIIHFW RI VSDWLDO DELOLW\ ZDV PRUH LQIOXHQWLDO IRU IHPDOHV )HPDOHV PDGH PRUH XVH RI GLDJUDPV LQ WKH VROXWLRQ RI SUREOHPV UHGXFLQJ WKH DGYDQWDJH RI PDOHV RYHU IHPDOHV LQ SUREOHP VROYLQJ SHUIRUPDQFH PAGE 40 7KH LVVXH RI VH[UHODWHG GLIIHUHQFHV LQ VSDWLDO YLVXDOL]DWLRQ DELOLW\ DV DQ H[SODQDWLRQ IRU VH[ GLIIHUHQFHV REVHUYHG LQ PDWKHPDWLFV DFKLHYHPHQW LV OHVV FRQYLQFLQJ DQG WKH ILQGLQJV PRUH FRQWUDGLFWRU\ WKDQ LQ WKH LVVXH RI VH[ GLIIHUHQFHV LQ IRUPDO HGXFDWLRQ %HVLGHV WKHVH FRJQLWLYH LVVXHV RWKHU LVVXHV PRVWO\ DIIHFWLYH LQ QDWXUH KDYH DOVR EHHQ VWXGLHG LQ WU\LQJ WR H[SODLQ WKH RULJLQ RI WKHVH VH[ GLIIHUHQFHV LQ PDWKHPDWLFV DFKLHYHPHQW DQG OHDUQLQJ 7KH VWXGLHV GHDOLQJ ZLWK WKHVH DIIHFWLYH YDULDEOHV DUH UHYLHZHG LQ WKH QH[W VHFWLRQ 'LIIHUHQWLDWHG (IIHFW RI $IIHFWLYH 9DULDEOHV 5HVHDUFKHUV KDYH DWWHPSWHG WR H[SODLQ WKH HIIHFW RI VH[ GLIIHUHQFHV LQ LQWHUQDO EHOLHIV LQWHUHVWV DQG DWWLWXGHV DIIHFWLYH YDULDEOHVf RQ PDWKHPDWLFV OHDUQLQJ DQG DFKLHYHPHQW $ EULHI VWDWHPHQW RI HDFK H[SODQDWLRQ SUHFHGHV WKH VXPPDU\ RI VWXGLHV FRQGXFWHG WKDW VXSSRUW WKH H[SODQDWLRQ WKDQ PDOHV ODFN FRQILGHQFH LQ WKHLU DELOLW\ WR OHDUQ PDWKHPDWLFV DQG WKLV DIIHFWV WKHLU DFKLHYHPHQW LQ PDWKHPDWLFV DQG WKHLU HOHFWLRQ RI PRUH DGYDQFHG PDWKHPDWLFV FRXUVHV 0DFFRE\ DQG -DFNOLQ f UHSRUWHG WKDW VHOI FRQILGHQFH LQ WHUPV RI JUDGH H[SHFWDQF\ DQG VXFFHVV LQ SDUWLFXODU WDVNV ZDV IRXQG WR EH FRQVLVWHQWO\ ORZHU LQ ZRPHQ WKDQ LQ PHQ ,Q )HQQHPD DQG 6KHUPDQ UHSRUWHG PAGE 41 WKDW LQ WKHLU VWXG\ LQYROYLQJ VWXGHQWV IURP JUDGHV WKURXJK } ER\V VKRZHG D KLJKHU OHYHO RI FRQILGHQFH LQ PDWKHPDWLFV DW HDFK JUDGH OHYHO 7KHVH GLIIHUHQFHV EHWZHHQ WKH VH[HV RFFXUUHG LQ PRVW LQVWDQFHV HYHQ ZKHQ QR VH[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFV DFKLHYHPHQW ZHUH IRXQG 7KH FRUUHODWLRQ EHWZHHQ FRQILGHQFH LQ PDWKHPDWLFV SHUIRUPDQFH DQG PDWKHPDWLFV DFKLHYHPHQW LQ WKLV VWXG\ ZDV KLJKHU WKDQ IRU DQ\ RWKHU DIIHFWLYH YDULDEOH LQYHVWLJDWHG 6KHUPDQ UHSRUWHG D VLPLODU ILQGLQJ LQ LQ PDOHV} WKH PRVW LPSRUWDQW IDFWRU UHODWHG WR FRQWLQXDWLRQ LQ WKHRUHWLFDO PDWKHPDWLFV FRXUVHV ZDV FRQILGHQFH LQ OHDUQLQJ PDWKHPDWLFV 7KLV YDULDEOH ZHLJKHG PRUH WKDQ DQ\ RI WKH FRJQLWLYH YDULDEOHV PDWKHPDWLFV DFKLHYHPHQW} VSDWLDO YLVXDOL]DWLRQ} JHQHUDO DELOLW\} DQG YHUEDO VNLOO ,Q WKH FDVH RI IHPDOHV} DPRQJ WKH DIIHFWLYH YDULDEOHV} FRQILGHQFH LQ OHDUQLQJ PDWKHPDWLFV ZDV IRXQG WR EH VHFRQG LQ LPSRUWDQFH WR SHUFHLYHG XVHIXOQHVV RI PDWKHPDWLFV 3UREHUW f VXSSRUWHG WKHVH ILQGLQJV ZLWK FROOHJH VWXGHQWV $ YDULDEOH WKDW QHHGV GLVFXVVLRQ ZLWKLQ WKH FRQWH[W RI VH[ GLIIHUHQFHV LQ FRQILGHQFH DV OHDUQHUV RI PDWKHPDWLFV LV FDXVDO DWWULEXWLRQ &DXVDO DWWULEXWLRQ PRGHOV DWWHPSW WR FODVVLI\ WKRVH IDFWRUV WR ZKLFK RQH DWWULEXWHV VXFFHVV RU IDLOXUH 7KH PRGHO SURSRVHG E\ :HLQHU f FDWHJRUL]HV IRXU GLPHQVLRQV RI DWWULEXWLRQ RI VXFFHVV DQG IDLOXUH VWDEOH DQG LQWHUQDO} XQVWDEOH DQG LQWHUQDO} VWDEOH DQG H[WHUQDO} DQG XQVWDEOH DQG H[WHUQDO )RU H[DPSOH}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f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n V f VWXG\ VXSSRUWHG WKHVH ILQGLQJV $FFRUGLQJ WR .ORRVWHUPDQ DWWULEXWLRQDO YDULDEOHV DSSHDU WR EH PRUH LPSRUWDQW DFKLHYHPHQW PHGLDWRUV IRU IHPDOHV WKDQ IRU PDOHV DV PHDVXUHG E\ PDWKHPDWLFV ZRUG SUREOHPV 0RUH UHVHDUFK LV QHHGHG LQ WKLV DUHD PAGE 43 0DWKHPDWLFVDV D PDOH GRPDLQ 0DWKHPDWLFV LV DQ DFWLYLW\ PRUH FORVHO\ UHODWHG WR WKH PDOH VH[ GRPDLQ WKDQ WR WKH IHPDOH VH[ GRPDLQ (FFOHV HW DO f 7KXVr WKH PDWKHPDWLFDO DFKLHYHPHQW RW ER\V LV KLJKHU WKDQ WKDW RI JLUOV $FFRUGLQJ WR -RKQ (UQHVW f LQ KLV VWXG\ 0DWKHPDWLFV DQG 6H[ PDWKHPDWLFV LV D VH[LVW GLVFLSOLQH +H DWWULEXWHG VH[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFDO DFKLHYHPHQW WR WKH FUHDWLRQ E\ VRFLHW\ RI VH[XDO VWHUHRW\SHV DQG DWWLWXGHV UHVWULFWLRQV DQG FRQVWUDLQWV WKDW SURPRWH WKH LGHD RI WKH VXSHULRULW\ RI ER\V LQ PDWKHPDWLFV (UQHVW UHSRUWHG WKDW ER\V JLUOV DQG WHDFKHUV DOO EHOLHYH WKDW ER\V DUH VXSHULRU LQ PDWKHPDWLFV DW OHDVW E\ WKH WLPH VWXGHQWV UHDFK DGROHVFHQFH %HUQ DQG %HUQ f DJUHH DQG DUJXH WKDW DQ $PHULFDQ ZRPDQ LV WUDLQHG WR NQRZ KHU SODFH LQ VRFLHW\ EHFDXVH RW WKH SHUYDVLYH VH[UROH FRQFHSW ZKLFK UHVXOWV LQ GLIIHUHQWLDO H[SHFWDWLRQV DQG VRFLDOL]DWLRQ SUDFWLFHV 3ODQN DQG 3ODQN f ZHUH PRUH VSHFLILF 7KH\ GLVFXVVHG WZR K\SRWKHVHV UHODWHG WR WKLV YLHZ WKH GLIIHUHQWLDO FXOWXUDO UHLQIRUFHPHQW K\SRWKHVLV DQG WKH PDVFXOLQH LGHQWLILFDWLRQ K\SRWKHVLV 7KH GLIIHUHQWLDO FXOWXUDO UHLQIRUFHPHQW K\SRWKHVLV VWDWHV WKDW VRFLHW\ LQ JHQHUDO SHUFHLYHV PDWKHPDWLFV DV D PDOH GRPDLQ JLYLQJ IHPDOHV OHVV HQFRXUDJHPHQW IRU H[FHOOLQJ LQ LW 7KH PDVFXOLQH LGHQWLILFDWLRQ K\SRWKHVLV HVWDEOLVKHV WKDW PAGE 44 DFKLHYHPHQW DQG LQWHUHVW LQ PDWKHPDWLFV UHVXOW IURP LGHQWLILFDWLRQ ZLWK WKH PDVFXOLQH UROH $ VWXG\ UHODWHG WR WKH GLIIHUHQWLDO FXOWXUDO UHLQIRUFHPHQW K\SRWKHVLV LV WKDW RI 'Z\HU f 'Z\HU H[DPLQHG WKH UHODWLRQVKLS EHWZHHQ VH[ UROH VWDQGDUGV WKH H[WHQW WR ZKLFK DQ LQGLYLGXDO FRQVLGHUV FHUWDLQ DFWLYLWLHV DSSURSULDWH WR PDOHV RU IHPDOHVf DQG DFKLHYHPHQW LQ UHDGLQJ DQG DULWKPHWLF 6WXGHQWV IURP JUDGHV r DQG SDUWLFLSDWHG LQ WKLV VWXG\ 6KH IRXQG WKDW VH[ UROH VWDQGDUGV FRQWULEXWHG VLJQLILFDQW YDULDQFH WR UHDGLQJ DQG DULWKPHWLF DFKLHYHPHQW WHVW VFRUHV DQG WKDW WKH HIIHFW ZDV VWURQJHU IRU PDOHV WKDQ IRU IHPDOHV 7KLV OHG WR KHU FRQFOXVLRQ WKDW VH[UHODWHG GLIIHUHQFHV LQ UHDGLQJ DQG DULWKPHWLF DUH PRUH D IXQFWLRQ RI WKH FKLOGnV SHUFHSWLRQ RI WKHVH DUHDV DV VH[DSSURSULDWH RU VH[LQDSSURSULDWH WKDQ RI WKH FKLOGnV ELRORJLFDO VH[ LQGLYLGXDO SUHIHUHQFH IRU PDVFXOLQH DQG IHPLQLQH VH[ UROHV RU OLNLQJ RU GLVOLNLQJ UHDGLQJ RU PDWKHPDWLFV ,Q D VWXG\ ZKLFK DJUHHV ZLWK WKH PDVFXOLQH LGHQWLILFDWLRQ K\SRWKHVLV 0LOWRQ f IRXQG WKDW LQGLYLGXDOV ZKR KDG UHFHLYHG VWURQJ PDVFXOLQH RULHQWDWLRQ SHUIRUPHG EHWWHU LQ SUREOHP VROYLQJ WKDQ LQGLYLGXDOV ZKR UHFHLYHG OHVV PDVFXOLQH RULHQWDWLRQ (OWRQ DQG 5RVH f IRXQG WKDW ZRPHQ ZLWK KLJK PDWKHPDWLFDO DSWLWXGH DQG DYHUDJH YHUEDO DSWLWXGH VFRUHG KLJKHU RQ WKH PDVFXOLQLW\ VFDOH RI WKH 2PQLEXV 3HUVRQDOLW\ ,QYHQWRU\ 23,f WKDQ WKRVH ZLWK DYHUDJH VFRUHV RQ ERWK WDVNV PAGE 45 ,W LV QRW XQWLO DGROHVFHQFH WKDW VH[ GLIIHUHQFHV LQ WKH SHUFHSWLRQ RI PDWKHPDWLFV DV D PDOH GRPDLQ DUH IRXQG )HQQHPDr 6WHLQr 6WHLQ t 6PLWKOHVVr 9HUEHNHr f ,Q D VWXG\ ZLWK QGr WK DQG WK JUDGHUVr 6WHLQ DQG 6PLWKOHVV f IRXQG WKDW VWXGHQWVn SHUFHSWLRQV RI VSDWLDOr PHFKDQLFDOr DQG DULWKPHWLF VNLOOV DV PDVFXOLQH EHFDPH PRUH GHILQHG DV WKHVH VWXGHQWV JRW ROGHU )HQQHPD f FRQVLGHUV WKDW WKH LQIOXHQFH HDFK VH[ H[HUWV XSRQ WKH RWKHU RQ DOO DVSHFWV RI EHKDYLRU LV VWURQJHU GXULQJ DGROHVFHQFH 6LQFH GXULQJ WKHVH \HDUV PDOHV VWHUHRW\SH PDWKHPDWLFV DV D PDOH GRPDLQr WKH\ VHQG WKLV PHVVDJH WR IHPDOHV ZKRr LQ WXUQr WHQG WR EH LQIOXHQFHG LQ WKHLU ZLOOLQJQHVV WR VWXG\ RU QRW WR VWXG\ PDWKHPDWLFV %HIRUH WKDW VWDJHr JLUOV FRQVLGHU DULWKPHWLF IHPLQLQHr ZKLOH ER\V FRQVLGHU LW DSSURSULDWH IRU ERWK VH[HV %REEHr f 8VHIXOQHVV RI PDWKHPDWLFV )HPDOHV SHUFHLYH PDWKHPDn WLFV DV OHVV XVHIXO WR WKHP WKDQ PDOHV GRr DQG WKLV SHUFHSWLRQ RFFXUV DW D YHU\ \RXQJ DJH $V D UHVXOWr IHPDOHV H[HUW OHVV HIIRUW WKDQ PDOHV WR OHDUQ RU HOHFW WR WDNH DGYDQFHG PDWKHPDWLFV FRXUVHV 0DQ\ VWXGLHV UHSRUWHG EHIRUH IRXQG WKDW WKH SHUFHSWLRQ RI WKH XVHIXOQHVV RI PDWKHPDWLFV IRU RQHnV IXWXUH GLIIHUV IRU PDOHV DQG IHPDOHVr DQG LV UHODWHG WR FRXUVH WDNLQJ SODQV DQG EHKDYLRU )R[r f ,I IHPDOHV GR QRW SHUFHLYH PDWKHPDWLFV DV XVHIXO IRU WKHLU IXWXUHr PAGE 46 WKH\ VKRZ OHVV LQWHUHVW LQ WKH VXEMHFW WKDQ FRXQWHUSDUW PDOHV 7KHVH GLIIHUHQFHV LQ LQWHUHVW DUH ZKDW +LOWRQ DQG %HUJOXQG f VXJJHVW WR DFFRXQW IRU VH[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFV DFKLHYHPHQW $OWKRXJK WKH SHUFHSWLRQ RI WKH XVHIXOQHVV RI PDWKHPDWLFV LV VWLOO DQ LPSRUWDQW SUHGLFWRU RI FRXUVH WDNLQJ IRU JLUOV} WKHUH LV D JURZLQJ VLPLODULW\ EHWZHHQ PDOHV DQG IHPDOHV UHJDUGLQJ WKH XVHIXOQHVV RI PDWKHPDWLFV $UPVWURQJ t 3ULFH )HQQHPD t 6KHUPDQ 0ROOHU f $UPVWURQJ DQG 3ULFH LQYHVWLJDWHG WKH UHODWLYH LQIOXHQFH RI VHOHFWHG IDFWRUV LQ VH[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFV SDUWLFLSDWLRQ %RWK PDOHV DQG IHPDOHV VHOHFWHG XVHIXOQHVV RI PDWKHPDWLFV DV WKH PRVW LPSRUWDQW IDFWRU LQ GHFLDLQJ ZKHWKHU RU QRW WR WDNH PRUH PDWKHPDWLFV LQ KLJK VFKRRO 0ROOHUnV VWXG\ UHYHDOHG WKDW ERWK PDOHV DQG IHPDOHV EDVHG PDWKHPDWLFV FRXUVHWDNLQJ GHFLVLRQV RQ FDUHHU XVHIXOQHVV $ )HQQHPD DQG 6KHUPDQ f VWXG\ VKRZHG RQO\ VOLJKW GLIIHUHQFHV EHWZHHQ PDOHV DQG IHPDOHV LQ WKHLU IHHOLQJV DERXW WKH XVHIXOQHVV RI PDWKHPDWLFV ,Q KHU VWXG\ RI WKLV YDULDEOH DPRQJ FROOHJH VWXGHQWV 3UREHUW f GLG QRW ILQG DQ\ VH[UHODWHG GLIIHUHQFHV HLWKHU 7KHVH KDYH EHHQ WKH PDLQ DIIHFWLYH YDULDEOHV UHVHDUFKHG LQ DWWHPSWLQJ WR H[SODLQ WKH XQGHUO\LQJ FDXVHV RI VH[UHODWHG GLIIHUHQFHV REVHUYHG LQ PDWKHPDWLFV OHDUQLQJ DQG DFKLHYHPHQW ,Q VSLWH RI WKH JUHDW GLYHUVLW\ RI VWXGLHV PAGE 47 GHDOLQJ ZLWK ERWK FRJQLWLYH DQG DIIHFWLYH YDULDEOHVr WKHUH DUH QR FOHDUFXW ILQGLQJV WR UHQGHU XQHTXLYRFDO VXSSRUW WR D SDUWLFXODU YDULDEOH DV DFFRXQWLQJ IRU WKHVH VH[UHODWHG GLIIHUHQFHV +RZHYHUr HYHU\WKLQJ VHHPV WR SRLQW WR WKH IDFW WKDW DIIHFWLYHr UDWKHU WKDQ FRJQLWLYH YDULDEOHV SOD\ D PRUH VLJQLILFDQW UROH LQ WKH VH[UHODWHG GLIIHUHQFHV REVHUYHG LQ PDWKHPDWLFV SHUIRUPDQFH DQG OHDUQLQJ ,Q PRVW RI WKH VWXGLHV GHDOLQJ ZLWK DIIHFWLYH YDULDEOHVr ILQGLQJV FRQVLVWHQWO\ VKRZ WKDW WKHVH IDFWRUV LQIOXHQFH PDWKHPDWLFV SHUIRUPDQFH LQ IHPDOHV PRUH WKDQ LQ PDOHV ,Q DW OHDVW RQH DUHDr FRQILGHQFH DV OHDUQHUV RI PDWKHPDWLFVr 6KHUPDQ f IRXQG WKDW WKLV YDULDEOH LQIOXHQFHG FRXUVH HOHFWLRQ PRUH WKDQ DOO WKH FRJQLWLYH YDULDEOHV SUHYLRXVO\ GLVFXVVHG 7KH FDVH IRU WKH VRFLHWDO LQIOXHQFHV RQ VH[ UROHV DQG H[SHFWDWLRQV WR DFFRXQW IRU WKH GLIIHUHQFHV LQ PDWKHPDWLFV OHDUQLQJ LV DOVR VXSSRUWHG LQ RQH ZD\ RU DQRWKHU LQ WKH VWXGLHV UHSRUWHG LQ WKH OLWHUDWXUH 3UREOHP 6ROYLQJ 3HUIRUPDQFH BDQGB5HODWHG 9DULDEOHV 3UREOHP VROYLQJ KDV EHHQ SHUKDSV WKH PRVW H[WHQVLYHO\ UHVHDUFKHG DUHD LQ PDWKHPDWLFV HGXFDWLRQ 3XEOLVKHG UHYLHZV E\ .LOSDWULFN fr 5LHGHVHO fr DQG 6X\GDP DQG :HDYHU f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f +H LQYHVWLJDWHG WKH UHODWLRQVKLS RI DULWKPHWLFDO IDFWRUV PHQWDO IDFWRUV DQG UHDGLQJ IDFWRUV WR DFKLHYHPHQW LQ SUREOHP VROYLQJ 6L[WK JUDGH VWXGHQWV ZHUH DGPLQLVWHUHG WHVWV LQ SUREOHP VROYLQJ DQG FDWHJRUL]HG DV VXSHULRU DFKLHYHUV EHVW SUREOHP VROYHUVf DQG LQIHULRU DUFKLHYHUV SRRUHVW SUREOHP VROYHUVf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} DV WKH FRQWHQW RI WKH *DWHV WHVWV XVHG WR PHDVXUH UHDGLQJf LV OLWHUDU\ DQG GRHV QRW LQFOXGH PDWKHPDWLFDO PDWHULDO &KDVH f VWXGLHG YDULDEOHV LQ DQ HIIRUW WR ILQG RXW ZKLFK RQHV KDYH VLJQLILFDQW LQIOXHQFH RQ WKH DQLOLW\ WR VROYH YHUEDO PDWKHPDWLFV SUREOHPV 2QO\ FRPSXWDWLRQ} UHDGLQJ WR QRWH GHWDLOV} DQG IXQGDPHQWDO NQRZOHGJH ZHUH SULPDULO\ UHODWHG WR SUREOHP VROYLQJ &RPSXWDWLRQ DFFRXQWHG IRU b RI WKH b YDULDQFH GLUHFWO\ DVVRFLDWHG ZLWK SUREOHP VROYLQJ &KDVH FRQFOXGHG WKDW D SXSLOnV DELOLW\ LQ WKH PHFKDQLFV RI FRPSXWDWLRQ} FRPSUHKHQVLRQ RI WKH SULQFLSOHV WKDW XQGHUOLQH WKH QXPEHU V\VWHPV} DQG WKH H[WHQW WR ZKLFK LPSRUWDQW LWHPV RI LQIRUPDWLRQ DUH QRWLFHG ZKHQ UHDGLQJ} DUH JRRG SUHGLFWRUV RI WKH VWXGHQWnV DELOLW\ LQ VROYLQJ YHUEDO SUREOHPV %DORZ f LQYHVWLJDWHG WKH LPSRUWDQFH RI UHDGLQJ DELOLW\ DQG FRPSXWDWLRQ DELOLW\ LQ SUREOHP VROYLQJ SHUIRUPDQFH +H REMHFWHG WR WKH DSSURDFKHV XVHG E\ RWKHU UHVHDUFKHUV ZKR LQ WKHLU DQDO\VHV GLFKRWRPL]HG UHVHDUFK VXEMHFWV DV SRRU RU JRRG VWXGHQWV} DQG ZKR LJQRUHG WKH UHFRJQL]HD HIIHFW RI LQWHOOLJHQFH RQ UHDGLQJ DQG RQ PDWKHPDWLFV DFKLHYHPHQW %DORZ DGPLQLVWHUHG WKH 6WDQIRUG PAGE 50 $FKLHYHPHQW 7HVW VXEWHVWV RI UHDGLQJ DULWKPHWLF DQG UHDVRQLQJf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f DWWHPSWHG WR LQYHVWLJDWH WKH W\SHV RI GLIILFXOWLHV FKLOGUHQ KDYH LQ VROYLQJ ZRUG SUREOHPV 7KH\ DGPLQLVWHUHG WKH ZRUG SUREOHP VHFWLRQ RI WKH 0HWURSROLWDQ $FKLHYHPHQW 7HVW WR FKLOGUHQ IURP WKH WK JUDGH (UURUV ZHUH FODVVLILHG LQ WZR PAGE 51 FDWHJRULHV &DWHJRU\ LQFOXGHG FOHULFDO DQG FRPSXWDWLRQDO HUURUV &DWHJRU\ ,, LQFOXGHG RWKHU W\SHV RI HUURUVr VXFK DV DYHUDJH DQG DUHD HUURUV XVH RI ZURQJ RSHUDWLRQr QR UHVSRQVHr DQG HUUHG UHVSRQVHV RIIHULQJ QR FOXHV ,W WKH VWXGHQWnV ZRUN LQGLFDWHG WKH FRUUHFW SURFHGXUH DQG \HW WKH SUREOHP ZDV PLVVHG EHFDXVH RI D FRPSXWDWLRQDO RU FOHULFDO HUURUr LW ZDV DVVXPHG WKDW WKH SUREOHP ZDV UHDG DQG XQGHUVWRRG $Q DQDO\VLV RI IUHTXHQFLHV VKRZHG WKDW FOHULFDO HUURUV ZHUH UHVSRQVLEOH IRU b RI WKH SUREOHPV LQFRUUHFWO\ VROYHGr FRPSXWDWLRQDO HUURUV DFFRXQWHG IRU br DQG RWKHU HUURUV IRU b RI WKH HUUHG SUREOHPV .QLIRQJ DQG +ROWDQ FRQFOXGHG WKDW LPSURYHG FRPSXWDWLRQDO VNLOOV FRXOG KDYH HOLPLQDWHG QHDUO\ KDOI RI WKH ZRUG SUREOHP HUURUV S ,OOf 7KHVH FRPSXWDWLRQDO HUURUV ZHUH PDGH LQ D FRQWH[W ZKHUH RWKHU VNLOOV VXFK DV UHDGLQJr LQWHUSUHWDWLRQ RI WKH SUREOHPr DQG LQWHJUDWLRQ RI WKHVH VNLOOV QHFHVVDU\ IRU WKH VROXWLRQ RI ZRUG SUREOHPVr PLJKW LQWHUDFW +RZHYHUr .QLIRQJ DQG +ROWDQ VWDWH WKDW WKHLU ILQGLQJV QHLWKHU FRQILUP QRU GHQ\ WKDW LPSURYHPHQW RI UHDGLQJ VNLOOV ZLOO OHDG WR LPSURYHPHQW LQ SUREOHP VROYLQJ 7KH\ FRQFOXGH WKDW LW LV GLIILFXOW WR DWWULEXWH PDMRU LPSRUWDQFH WR UHDGLQJ DV D VRXUFH RI IDLOXUH S f ,Q D ODWHU DQDO\VLVr ORRNLQJ IRU HYLGHQFH RI SRRU UHDGLQJ DELOLWLHV DIIHFWLQJ FKLOGUHQnV VXFFHVV LQ ZRUG SUREOHPVr .QLIRQJ DQG +ROWDQ f LQWHUYLHZHG WKH PAGE 52 FKLOGUHQ ZKRVH HUURUV IHOO XQGHU WKH FDWHJRU\ RU RWKHU HUURUV 6WXGHQWV ZHUH DVNHG WR UHDG HDFK SUREOHP DORXG DQG DQVZHU WKHVH TXHVWLRQV :KDW NLQG RI VLWXDWLRQ GRHV WKH SUREOHP GHVFULEH" :KDW GRHV WKH SUREOHP DVN \RX WR ILQG" +RZ ZRXOG \RX ZRUN WKH SUREOHP" 1LQHW\ ILYH SHUFHQW RI WKH VWXGHQWV UHDG WKH SUREOHP FRUUHFWO\ b H[SODLQHG WKH NLQG RI VLWXDWLRQ WKH SUREOHP GHVFULEHG LQ D FRUUHFW PDQQHU b FRUUHFWO\ DQVZHUHG ZKDW WKH SUREOHP ZDV DVNLQJ WKHP WR ILQGr DQG b FRUUHFWO\ DQVZHUHG WKH TXHVWLRQ RI KRZ WR ZRUN WKH SUREOHP 7KH IDFW WKDW D ODUJH SHUFHQW RI WKH VWXGHQWV ZKRVH HUURUV ZHUH FODVVLILHG DV RWKHU HUURUV LQ ZKLFK UHDGLQJ VNLOOV PLJKW KDYH EHHQ D IDFWRUf FRUUHFWO\ VWDWHG KRZ WR ZRUN WKH SUREOHP} LV VWURQJ HYLGHQFH RI WKHLU DELOLW\ WR UHDG DQG LQWHUSUHW WKH SUREOHPV FRUUHFWO\ 7KH HUURUV PDGH E\ WKLV JURXS RI VWXGHQWV KDG D GLVWLQFW RULJLQ} XQUHODWHG WR UHDGLQJ DELOLW\ =DOHZVNL f LQYHVWLJDWHG WKH UHODWLYH FRQWULEXWLRQ RI YHUEDO LQWHOOLJHQFH} UHDGLQJ FRPSUHKHQVLRQ} YRFDEXODU\} LQWHUSUHWDWLRQ RI JUDSKV DQG WDEOHV} PDWKHPDWLFDO FRQFHSWV} QXPEHU VHQWHQFH VHOHFWLRQ} DQG FRPSXWDWLRQ WR VXFFHVVIXO PDWKHPDWLFDO ZRUG SUREOHP VROXWLRQ} DQG WKH UHODWLRQVKLS RI WKH GHSHQGHQW YDULDEOH WR WKH HLJKW LQGHSHQGHQW YDULDEOHV 6KH ZRUNHG ZLWK D JURXS RI WK JUDGH FKLOGUHQ ZKR ZHUH DGPLQLVWHUHG WKH VXEWHVWV RI WKH ,RZD 7HVW RI %DVLF 6NLOOV ,7%6f DQG WKH :HVFKOHU ,QWHOOLJHQFH 6FDOH IRU PAGE 53 &KLOGUHQ :,6&f 0XOWLSOH UHJUHVVLRQ DQDO\VLV ZDV SHUIRUPHG $ FRUUHODWLRQ RI ZDV IRXQG EHWZHHQ ZRUG SUREOHP VROYLQJ DQG WKH HLJKW LQGHSHQGHQW YDULDEOHV &RUUHODWLRQV EHWZHHQ ZRUG SUREOHP VROYLQJ DQG WKH LQGHSHQGHQW YDULDEOHV UDQJHG IURP YHUEDO LQWHOOLJHQFHf WR PDWKHPDWLFDO FRQFHSWVf &RUUHODWLRQV EHWZHHQ WKH LQGHSHQGHQW YDULDEOHV UDQJHG IURP YHUEDO LQWHOOLJHQFH DQG FRPSXWDWLRQf WR UHDGLQJ FRPSUHKHQVLRQ DQG YRFDEXODU\f 0DWKHPDWLFDO FRQFHSWVr FRPSXWDWLRQr DQG QXPEHU VHQWHQFH VHOHFWLRQ ZHUH DOPRVW DV HIIHFWLYH DV DOO HLJKW LQGHSHQGHQW YDULDEOHV LQ SUHGLFWLQJ DFKLHYHPHQW LQ PDWKHPDWLFDO ZRUG SUREOHP VROYLQJ 0DWKHPDWLFDO FRQFHSWVr FRPSXWDWLRQr QXPEHU VHQWHQFH VHOHFWLRQr DQG UHDGLQJ FRPSUHKHQVLRQ DFFRXQWHG IRU b RI WKH YDULDQFHr ZKHUHDV DOO HLJKW SUHGLFWRUV DFFRXQWHG IRU b RI WKH YDULDQFH 7KH WZR EHVW SUHGLFWRUV ZHUH PDWKHPDWLFDO FRQFHSWV DQG FRPSXWDWLRQr ZKLFK DFFRXQWHG IRU b YDULDQFH 2WKHU YDULDEOHV DFFRXQWHG IRU DERXW b RI WKH YDULDQFH 7KH DXWKRU UHFRPPHQGV WKDW WKH ILQGLQJV RI WKLV VWXG\ EH LQWHUSUHWHG FDXWLRXVO\ EHFDXVH WKH FRUUHODWLRQ EHWZHHQ WKH HLJKW LQGHSHQGHQW YDULDEOHV ZDV KLJKr DQGr DFFRUGLQJ WR =DOHZVNLr LQ D VWXG\ RI WKLV QDWXUH ZKHUH WKH LQWHUHVW LV SULPDULO\ LQ WKH LQIOXHQFH RI VHYHUDO YDULDEOHV RQ RQH GHSHQGHQW YDULDEOHr D ORZ FRUUHODWLRQ EHWZHHQ WKH LQGHSHQGHQW YDULDEOHV LV UHTXLUHG S f PAGE 54 ,Q D PRUH UHFHQW LQYHVWLJDWLRQ ([HGLVLV f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f VWXGLHG WKH FRQWULEXWLRQ RI UHDGLQJ FRPSUHKHQVLRQ FRPSXWDWLRQ DEVWUDFW YHUEDO UHDVRQLQJ DQG DULWKPHWLF FRQFHSWV WR DULWKPHWLF SUREOHP VROYLQJ SHUIRUPDQFH )RXUWK DQG WK JUDGH VWXGHQWV ZHUH DGPLQLVWHUHG WKH ,RZD WHVWV RI %DVLF 6NLOOV DQG WKH /RUJH7KRUQGOLNH LQWHOOLJHQFH WHVW YHUEDOf PAGE 55 +H IRXQG WKDW LQ WKH WK JUDGH WKH FRUUHODWLRQV EHWZHHQ SUREOHP VROYLQJ DQG DEVWUDFW YHUEDO UHDVRQLQJ UHDGLQJ FRPSUHKHQVLRQ DULWKPHWLF FRQFHSWV DQG FRPSXWDWLRQ ZHUH DQG UHVSHFWLYHO\ DQG DQG LQ WKH WK JUDGH :KHQ FRPSXWDWLRQ ZDV KHOG FRQVWDQW WKH FRUUHODWLRQ EHWZHHQ SUREOHP VROYLQJ DQG UHDGLQJ ZDV LQ JUDGH DQG LQ JUDGH :KHQ UHDGLQJ ZDV KHOG FRQVWDQW WKH FRUUHODWLRQ EHWZHHQ SUREOHP VROYLQJ DQG FRPSXWDWLRQ ZDV LQ JUDGH DQG LQ JUDGH &UHVZHOO f ZRUNHG ZLWK D VDPSOH RI DQJOR DQG EODFN DGROHVFHQWV IURP &KLFDJR (DFK VXEMHFW ZDV DGPLQLVWHUHG WKH &DOLIRUQLD $FKLHYHPHQW WHVW 0XOWLSOH UHJUHVVLRQ ZDV XVHG WR DQDO\]H WKH GDWD 7KH DQDO\VLV VKRZHG WKDW UHDGLQJ LV PRUH LPSRUWDQW WKDQ FRPSXWDWLRQ LQ SUHGLFWLQJ VWXGHQW SHUIRUPDQFH LQ SUREOHP VROYLQJ 5HDGLQJ DFFRXQWHG IRU b RI WKH YDULDQFH FRPSXWDWLRQ DFFRXQWHG IRU b RI WKH YDULDQFH %DOOHZ DQG &XQQLQJKDP f ZRUNHG ZLWK WK JUDGH VWXGHQWV LQ DQ DWWHPSW WR ILQG ZKDW SURSRUWLRQ RI VWXGHQWV KDYH DV WKHLU PDLQ VRXUFH RI GLIILFXOW\ ZLWK ZRUG SUREOHPV HDFK RI WKH IROORZLQJ IDFWRUV Df FRPSXWDWLRQ VNLOOV Ef LQWHUSUHWDWLRQ RI WKH SUREOHP Ff UHDGLQJ DQG Gf LQWHJUDWLQJ WKHVH VNLOOV LQ WKH VROXWLRQ RI SUREOHPV 7KH\ DOVR ZDQWHG WR NQRZ LI D VWXGHQW FDQ EH HIILFLHQWO\ GLDJQRVHG DV KDYLQJ RQH RI WKH IRXU FDWHJRULHV DV KLVKHU PDLQ GLIILFXOW\ ZLWK PDWKHPDWLFV ZRUG SUREOHPV PAGE 56 7KHLU VWXG\ LV LPSRUWDQW EHFDXVH LW UHSUHVHQWV DQ DWWHPSW WR GHPRQVWUDWH WKDW PXOWLSOH IDFWRUV FDQ LQWHUDFW LQ WKH FRUUHFW VROXWLRQ RI D PDWKHPDWLFV ZRUG SUREOHP 7KH\ FRQVWUXFWHG WKUHH JUDGHG WHVWV IURP D EDVDO PDWKHPDWLFV VHULHV IRU JUDGHV WKURXJK )RU WHVW r WKH SUREOHPV ZHUH VHW LQ SXUH FRPSXWDWLRQDO IRUP WKH HIIHFWV RU UHDGLQJr LQWHUSUHWDWLRQr DV ZHOO DV WKH QHFHVVLW\ IRU LQWHJUDWLRQ ZHUH UHPRYHG LQ DQ HIIRUW WR PHDVXUH WKH FRPSXWDWLRQDO VNLOOV UHTXLUHG E\ WKH ZRUG SUREOHPVf )RU WHVW r WKH HIIHFWV RI UHDGLQJ DQG FRPSXWDWLRQ ZHUH UHPRYHG E\ UHDGLQJ WKH SUREOHPV WR WKH VWXGHQWV DQG E\ JLYLQJ VFRUHV EDVHG RQ ZKHWKHU RU QRW WKH VWXGHQWV VHW WKHP XS SURSHUO\r LQ DQ DWWHPSW WR PHDVXUH SUREOHP LQWHUSUHWDWLRQ DORQH )RU WHVW r WKH HIIHFW RI FRPSXWDWLRQ ZDV UHPRYHG 7KH WHVW \LHOGHG WZR VFRUHVfÂ§RQH E\ JUDGLQJ WKH VWXGHQWV RQ ZKHWKHU RU QRW WKH\ VHW XS WKH SUREOHPV SURSHUO\ DQG DQRWKHU E\ JUDGLQJ RQ WKH EDVLV RI WKH FRUUHFW DQVZHU 7KH WHVWV ZHUH DGPLQLVWHUHG WR DOO VWXGHQWV IURP WKH WK JUDGH LQ WZR GLIIHUHQW VFKRROV $ GLDJQRVWLF SURILOH ZDV REWDLQHG IRU HDFK RI WKH VWXGHQWV IRU ZKLFK FRPSOHWH GDWD ZHUH DYDLODEOH D FRPSXWDWLRQDO VFRUHr D SUREOHPLQWHUSUHWDWLRQ VFRUHr D UHDGLQJ VFRUHr DQG D UHDGLQJSUREOHP VROYLQJ VFRUH 7KH\ DVVXPHG WKDW LI WKH UHDGLQJSUREOHP LQWHUSUHWDWLRQ VFRUH ZDV ORZHU RQH RU PRUH OHYHOV ORZHUf PAGE 57 WKDQ WKH SUREOHPLQWHUSUHWDWLRQ VFRUHU WKH GLIILFXOW\ ZDV GXH WR UHDGLQJ DELOLW\ ,I WKH VFRUH RI WKH ORZHVW RI WKH WKUHH DUHDV FRPSXWDWLRQU SUREOHP LQWHUSUHWDWLRQ DQG UHDGLQJSUREOHP LQWHUSUHWDWLRQ VFRUHf ZDV WKH VDPH DV WKH UHDGLQJSUREOHP VROYLQJ VFRUH WKH VWXGHQWnV DUHD RI JUHDWHVW LPPHGLDWH QHHG ZDV HLWKHU FRPSXWDWLRQ SUREOHP LQWHUSUHWDWLRQ RU UHDGLQJ ,I WKH UHDGLQJSUREOHP VROYLQJ VFRUH ZDV ORZHU WKDQ WKH ORZHVW RI WKH RWKHU WKUHH VFRUHV WKH VWXGHQWnV DUHD RI JUHDWHVW LPPHGLDWH QHHG ZDV LQWHJUDWLRQ $QDO\VLV RI WKH GDWD UHYHDOHG WKDW IRU b RI WKH VWXGHQWV SUREOHP LQWHUSUHWDWLRQ ZDV WKHLU PDMRU GLIILFXOW\ IRU b RI WKH VWXGHQWV LQWHJUDWLRQ WRWDO SUREOHP VROYLQJf ZDV WKHLU JUHDWHVW LPPHGLDWH QHHG IRU DQRWKHU b FRPSXWDWLRQ ZDV WKH PDMRU ZHDNQHVV DQG IRU b UHDGLQJ ZDV WKHLU JUHDWHVW LPPHGLDWH QHHG 6HYHQW\ ILYH SHUFHQW RI WKH VWXGHQWV GHPRQVWUDWHG FOHDU VWUHQJWK LQ FRPSXWDWLRQ b LQ SUREOHP LQWHUSUHWDWLRQ DQG b LQ UHDGLQJSUREOHP LQWHUSUHWDWLRQ $Q DQDO\VLV DFURVV DOO VWXGHQWV LQFOXGLQJ WKRVH ZLWKRXW FRPSOHWH GDWDf VKRZHG WKDW b RI WKH VXEMHFWV FRXOG QRW ZRUN ZRUG SUREOHPV DW D OHYHO DV KLJK DV WKDW DW ZKLFK WKH\ FRXOG FRPSXWH LQWHUSUHW SUREOHPV DQG UHDG DQG LQWHUSUHW SUREOHPV ZKHQ WKRVH DUHDV ZHUH PHDVXUHG VHSDUDWHO\ 7KLV OHG WKHP WR FRQFOXGH WKDW NQRZLQJ WKH VNLOOV RU WKH FRPSRQHQWV RI VROYLQJ ZRUG SUREOHPV LV QRW VXIILFLHQW IRU PAGE 58 VXFFHVV} VLQFH WKH FRPSRQHQWV PXVW EH LQWHJUDWHG LQWR D ZKROH SURFHVV PDVWHU\ OHDUQLQJ RI WKH FRPSRQHQWV FDQQRW DVVXUH PDVWHU\ RI WKH SURFHVVf 7KHLU DQDO\VLV DOVR OHG WKHP WR FRQFOXGH WKDWr LQ WKH FDVH RI WK JUDGHUV} LQDELOLW\ WR UHDG SUREOHPV LV D PDMRU REVWDFOH LQ VROYLQJ ZRUG SUREOHPV 2QO\ b RI WKH VXEMHFWV FRXOG UHDG DQG VHW XS SUREOHPV FRUUHFWO\ DW D KLJKHU OHYHO WKDQ WKH\ FRXOG FRPSXWH} ZKLOH b FRXOG FRPSXWH FRUUHFWO\ DW D KLJKHU OHYHO WKDQ WKH\ FRXOG UHDG DQG VHW XS SUREOHPV b FRXOG VHW XS SUREOHPV EHWWHU ZKHQ WKH\ KHDUG WKHP UHDG WKDQ ZKHQ WKH\ UHDG WKH SUREOHP WKHPVHOYHV 2QO\ b FRXOG VHW XS SUREOHPV EHWWHU ZKHQ WKH\ UHDG WKHP WKDQ ZKHQ WKH\ KHDUG WKHP UHDG 0XWK f LQYHVWLJDWHG WKH UROH RI UHDGLQJ DQG FRPSXWDWLRQDO VNLOOV LQ WKH VROXWLRQ RI ZRUG SUREOHPV $ JURXS RI VWXGHQWV IURP WKH WK JUDGH ZHUH DGPLQLVWHUHG D WHVW RI EDVLF VNLOOV DQG D PDWKHPDWLFV ZRUG SUREOHP WHVW 7KH ZRUG SUREOHP WHVW FRQVLVWHG RI VDPSOH LWHPV VXSSOLHG E\ WKH 1DWLRQDO $VVHVVPHQW RI (GXFDWLRQDO 3URJUHVV 7KH LWHPV ZHUH DGDSWHG WR LQFOXGH VRPH H[WUDQHRXV LQIRUPDWLRQ DQG FRPSOH[ V\QWDFWLF VWUXFWXUH )RXU YHUVLRQV RI WKH WHVW ZHUH FRQVWUXFWHG E\ FRPELQLQJ WZR YHUVLRQV RI SUREOHP LQIRUPDWLRQ DEVHQFH YV SUHVHQFH RI H[WUDQHRXV LQIRUPDWLRQf ZLWK WZR YHUVLRQV RI V\QWDFWLF VWUXFWXUH VLPSOH YV FRPSOH[ V\QWD[f 7DVN SHUIRUPDQFH ZDV PHDVXUHG E\ PHDQV RI WKH QXPEHU RI SUREOHPV DQVZHUHG FRUUHFWO\} PAGE 59 QXPEHU RI SUREOHPV VHW XS FRUUHFWO\r DQG DPRXQW RI WLPH VSHQW WDNLQJ WKH WHVW 5HDGLQJ DELOLW\ DQG FRPSXWDWLRQDO DELOLW\ ZHUH ERWK SRVLWLYHO\ FRUUHODWHG ZLWK QXPEHU RI FRUUHFW DQVZHUV DQG ZLWK QXPEHU RI SUREOHPV FRUUHFWO\ VHW XSr DQG QHJDWLYHO\ FRUUHODWHG ZLWK WHVWWDNLQJ WLPH 3UHVHQFH RI H[WUDQHRXV LQIRUPDWLRQ ZDV QHJDWLYHO\ FRUUHODWHG ZLWK FRUUHFW DQVZHUV DQG FRUUHFW VHW XSV DQG SRVLWLYHO\ FRUUHODWHG ZLWK WHVWWDNLQJ WLPH 6\QWDFWLF FRPSOH[LW\ ZDV QRW VLJQLILFDQWO\ FRUUHODWHG ZLWK DQ\ RI WKH SHUIRUPDQFH PHDVXUHV 5HVXOWV RI D PXOWLSOH UHJUHVVLRQ DQDO\VLV VKRZHG WKDW UHDGLQJ DFFRXQWHG IRU b RI WKH YDULDQFH LQ WRWDO FRUUHFW DQVZHUV DQG FRPSXWDWLRQ DFFRXQWHG IRU b 5HDGLQJ DELOLW\ DQG FRPSXWDWLRQDO DELOLW\ XQLTXHO\ DFFRXQWHG IRU b DQG b RI WKH YDULDQFH LQ WKH QXPEHU RI FRUUHFW DQVZHUVr UHVSHFWLYHO\ ([WUDQHRXV LQIRUPDWLRQ DGGHG VLJQLILFDQWO\ WR WKH YDULDQFH H[SODLQHG LQ WKH QXPEHU RI FRUUHFW DQVZHUVr EXW V\QWDFWLF VWUXFWXUH GLG QRW 5HDGLQJ DELOLW\ DFFRXQWHG IRU b RI WKH YDULDQFH LQ WHVWWDNLQJ WLPHr EXW FRPSXWDWLRQ GLG QRW DGG VLJQLILFDQWO\ WR WKH YDULDQFH H[SODLQHG E\ UHDGLQJ 0XWK FRQFOXGHG WKDW UHDGLQJ DQG FRPSXWDWLRQ ERWK FRQWULEXWH VLJQLILFDQWO\ WR VXFFHVV LQ VROYLQJ DULWKPHWLF ZRUG SUREOHPVr EXW WKDW UHDGLQJ SOD\V D PRUH VLJQLILFDQW UROH WKDQ GRHV FRPSXWDWLRQ PAGE 60 7KH VWXGLHV UHYLHZHG LQ WKLV VHFWLRQ VKRZ D SRVLWLYH UHODWLRQVKLS EHWZHHQ UHDGLQJ DQG SUREOHP VROYLQJ SHUIRUPDQFHr EXW LQ WKH FDVH RI %DOOHZ DQG &XQQLQJKDP fr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f KDV VXJJHVWHG WKDW DQ LQGLYLGXDOnV DWWLWXGH WRZDUG RQH DVSHFW RI WKH GLVFLSOLQH PDWKHPDWLFVf r VXFK DV SUREOHP VROYLQJr PD\ EH HQWLUHO\ GLIIHUHQW IURP KLVKHU DWWLWXGH WRZDUG DQRWKHU SKDVH RI WKH GLVFLSOLQHr VXFK DV FRPSXWDWLRQ 5HVHDUFKr KRZHYHUr KDV EHHQ GLUHFWHG WR WKH XVH RI VLQJOHr JOREDO PHDVXUHV RI DWWLWXGHV WRZDUG PDWKHPDWLFV UDWKHU WKDQ WR WKH LQYHVWLJDWLRQ RI DWWLWXGHV WRZDUG D SDUWLFXODU SKDVH RI WKH GLVFLSOLQH 7KH VWXGLHV GHVFULEHG EHORZ DUH RQO\ SDUW RI WKH IHZ LQYHVWLJDWLRQV ZKLFK KDYH H[DPLQHG WKH UHODWLRQVKLS EHWZHHQ PAGE 61 VWXGHQW DWWLWXGHV DQG SHUIRUPDQFH LQ WKH DUHD RI SUREOHP VROYLQJ &DUH\ f FRQVWUXFWHG D VFDOH WR PHDVXUH DWWLWXGHV WRZDUG SUREOHP VROYLQJ +HU LQWHUHVW ZDV LQ JHQHUDO SUREOHP VROYLQJ UDWKHU WKDQ LQ PDWKHPDWLFDO SUREOHP VROYLQJ +HU ZRUN FRQVWLWXWHV WKH ILUVW DWWHPSW WR FRQVWUXFW D PHDVXUH RI DWWLWXGHV WRZDUG SUREOHP VROYLQJ 7KH VFDOH ZDV XVHG ZLWK D JURXS RI FROOHJH VWXGHQWVr DQG VKH IRXQGr DPRQJ RWKHU WKLQJVr WKDW SUREOHP VROYLQJ SHUIRUPDQFH LV SRVLWLYHO\ UHODWHG WR SUREOHP VROYLQJ DWWLWXGHV DQG WKDWr LQ WKH FDVH RI IHPDOHVr SRVLWLYH PRGLILFDWLRQ RI DWWLWXGHV WRZDUG SUREOHP VROYLQJ EULQJV D VLJQLILFDQW JDLQ LQ SUREOHP VROYLQJ SHUIRUPDQFH /LQGJUHQr 6LOYDr )DUDFRr DQG 'D5RFKD f DGDSWHG &DUH\nV VFDOH RI DWWLWXGHV WRZDUG SUREOHP VROYLQJ DQG DSSOLHG LW WR D JURXS RI WK JUDGH %UD]LOLDQ FKLOGUHQ 6WXGHQWV DOVR DQVZHUHG DQ DULWKPHWLF DFKLHYHPHQW WHVWr D JHQHUDO LQWHOOLJHQFH WHVWr DQG D VRFLRHFRQRPLF 6(f VFDOH $ ORZ EXW VLJQLILFDQW SRVLWLYH FRUUHODWLRQ ZDV IRXQG EHWZHHQ DULWKPHWLF DFKLHYHPHQW DQG DWWLWXGHV WRZDUG SUREOHP VROYLQJ $ QHDU ]HUR FRUUHODWLRQ ZDV IRXQG EHWZHHQ DWWLWXGHV WRZDUG SUREOHP VROYLQJ DQG LQWHOOLJHQFH 6LQFH SUREOHP VROYLQJ LV RQH DVSHFW RI WKH GLVFLSOLQH RI PDWKHPDWLFVr WKLV FRUUHODWLRQ EHWZHHQ DWWLWXGHV DQG DULWKPHWLF DFKLHYHPHQW FDQ OHDG WR D FRQFOXVLRQ RU D VWURQJ PAGE 62 FRUUHODWLRQ EHWZHHQ DWWLWXGHV WRZDUG SUREOHP VROYLQJ DQG SUREOHP VROYLQJ SHUIRUPDQFH :KLWDNHU f FRQVWUXFWHG D VWXGHQW DWWLWXGH VFDOH WR PHDVXUH VRPH DVSHFWV RI WK JUDGH VWXGHQW DWWLWXGHV WRZDUG PDWKHPDWLF SUREOHP VROYLQJ +H LQFOXGHG VWDWHPHQWV UHIOHFWLQJ FKLOGUHQnV EHOLHIV DERXW WKH QDWXUH RI YDULRXV W\SHV RI PDWKHPDWLFDO SUREOHPV} WKH QDWXUH RI WKH SUREOHP VROYLQJ SURFHVV} WKH GHVLUDELOLW\ RU SHUVHYHULQJ ZKHQ VROYLQJ D SUREOHP} DQG WKH YDOXH RI JHQHUDWLQJ VHYHUDO LGHDV IRU VROYLQJ D SUREOHP +H FRUUHODWHG VWXGHQW DWWLWXGHV WRZDUG SUREOHP VROYLQJ ZLWK WKHLU VFRUHV LQ D PDWKHPDWLFDO WHVW ZKLFK \LHOGHG D FRPSUHKHQVLRQ VFRUH} DQ DSSOLFDWLRQ VFRUH} DQG D SUREOHP VROYLQJ VFRUH +H IRXQG D VLJQLILFDQW SRVLWLYH UHODWLRQVKLS EHWZHHQ SUREOHP VROYLQJ SHUIRUPDQFH DQG VWXGHQW DWWLWXGH VFRUHV RQ WKH VXEVFDOH ZKLFK PHDVXUHG UHDFWLRQV WR VXFK WKLQJV DV SUREOHP VROYLQJ WHFKQLTXHV RU SUREOHP VLWXDWLRQV} RU WR WKH IUXVWDWLRQ RU DQ[LHW\ H[SHULHQFHG ZKHQ FRQIURQWHG ZLWK SUREOHP VROYLQJ VLWXDWLRQV ,Q DQRWKHU SDUW RI WKLV VWXG\} :KLWDNHU LQYHVWLJDWHG WKH UHODWLRQVKLS EHWZHHQ WKH DWWLWXGHV RI WK JUDGH WHDFKHUV WRZDUG SUREOHP VROYLQJ DQG WKHLU VWXGHQWVn SHUIRUPDQFH LQ SUREOHP VROYLQJ $ YHU\ ZHDN DQG QRQVLJQLILFDQW QHJDWLYH FRUUHODWLRQ ZDV IRXQG EHWZHHQ WKH WHDFKHUnV DWWLWXGHV WRZDUG SUREOHP VROYLQJ DQG VWXGHQW SHUIRUPDQFH PAGE 63 7KH VWXGLHV UHYLHZHG KDYH FRQILUPHG WKH UHODWLRQVKLS EHWZHHQ SUREOHP VROYLQJ SHUIRUPDQFH DQG DWWLWXGHV WRZDUG SUREOHP VROYLQJ &DUH\} /LQGJUHQ HW DO} :KLWDNHU} f +RZHYHU} WKH UHVXOWV UHSRUWHD LQ WKH VWXGLHV WKDW LQYHVWLJDWHG WKH UHODWLRQVKLS EHWZHHQ SUREOHP VROYLQJ SHUIRUPDQFH DQG FRPSXWDWLRQ DQG EHWZHHQ UHDGLQJ DQG SUREOHP VROYLQJ IDLO WR EH FRQVLVWHQW LQ WKHLU FRQFOXVLRQV +DQVHQ f} &KDVH ,f} %DORZ f} .QLIRQJ DQG +ROWDQ } f} DQG =DOHZVNL FRQFOXGHG WKDW FRPSXWDWLRQ LV PRUH VWURQJO\ UHODWHG WR SUREOHP VROYLQJ WKDQ LV UHDGLQJ 0DUWLQ f} &UHVZHOO f} %DOOHZ DQG &XQQLQJKDP f} DQG 0XWK f} FRQFOXGHG WKDW UHDGLQJ DRLOLW\ DQG PDWKHPDWLFDO SUREOHP VROYLQJ VKRZ D VWURQJHU UHODWLRQVKLS WKDQ FRPSXWDWLRQ DQG SUREOHP VROYLQJ ([HGLVLVnV f ILQGLQJV OHG WR WKH FRQFOXVLRQ WKDW WKH HIIHFW RU UHDGLQJ DQG FRPSXWDWLRQ LQ SUREOHP VROYLQJ SHUIRUPDQFH LV XQLPSRUWDQW PAGE 64 &+$37(5 ,,, 0(7+2' 7KH ILUVW REMHFWLYH RI WKLV VWXG\ ZDV WR LQYHVWLJDWH VH[ GLIIHUHQFHV LQ WKH VHOHFWLRQ RI LQFRUUHFW UHVSRQVHV LQ D PDWKHPDWLFV PXOWLSOHFKRLFH DFKLHYHPHQW WHVWr DQG WR GHWHUPLQH ZKHWKHU WKHVH GLIIHUHQFHV ZHUH FRQVLVWHQW RYHU WKUHH FRQVHFXWLYH DGPLQLVWUDWLRQV RI WKH WHVW 7KH VHFRQG REMHFWLYH ZDV WR LQYHVWLJDWH ZKHWKHU PDOHV DQG IHPDOHV GLIIHU LQ SUREOHP VROYLQJ SHUIRUPDQFHr LI WKHVH GLIIHUHQFHV SHUVLVW DIWHU DFFRXQWLQJ IRU FRPSXWDWLRQDO VNLOOVr DQG LI WKH PDOHIHPDOH GLIIHUHQFHV GHSHQG RQ WKH OHYHO RI FRPSXWDWLRQDO VNLOOV 7KLV FKDSWHU FRQWDLQV GHVFULSWLRQV RI WKH VDPSOHr WKH WHVW LQVWUXPHQWr DQG WKH VWDWLVWLFDO DQDO\VLV XVHG LQ DFKLHYLQJ WKH DERYH PHQWLRQHG REMHFWLYHV 7KH 6DPSOH 7R DFKLHYH WKH ILUVW REMHFWLYH RI WKH VWXG\r DOO WKH VWXGHQWV ZKR WRRN WKH %DVLF 6NLOOV 7HVW LQ 0DWKHPDWLFV 3UXHED GH 'HVWUH]DV %Â£VLFDV HQ 0DWHPÂ£WLFDVf LQ HDFK RU WKH WKUHH \HDUV ZHUH LQFOXGHG LQ WKH VWXG\ 7R DFKLHYH WKH VHFRQG REMHFWLYH RI WKH VWXG\r DSSUR[LPDWHO\ PAGE 65 VWXGHQWV ZHUH VHOHFWHG UDQGRPO\ IRU HDFK \HDU )RU WKH ILUVW \HDU ZHUH VHOHFWHG ER\V DQG JLUOVf IRU WKH VHFRQG \HDU VWXGHQWV ZHUH VHOHFWHG ER\V DQG JLUOVf DQG IRU WKH WKLUG \HDU] VWXGHQWV ZHUH VHOHFWHG ER\V DQG JLUOVf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f SUHVHQWV WKH IROORZLQJ VDWXUDWHG PRGHO IRU WKH GDWD ORJ PLMN 8L 8 M MFL M 8LN8M N8L M N f ,Q WKLV PRGHO PÂMN LV WKH H[SHFWHG YDOXH RI WKH IUHTXHQF\ LQ FHOO LMN RI WKH WKUHH GLPHQVLRQDO WDEOH 7KH PRGHO VWDWHV WKDW DOO WKUHH FODVVLILFDWLRQ IDFWRUV IRU D WKUHH GLPHQVLRQDO FRQWLQJHQF\ WDEOH DUH PXWXDOO\ GHSHQGHQW ,Q WKH SUHVHQW UHVHDUFK L LV WKH \HDU LQGH[ M LV WKH RSWLRQ LQGH[ DQG N LV WKH VH[ LQGH[ )LHQEHUJ VKRZV WKDW GHOHWLQJ WKH WHUPV 8ÂM DQG 8AMN \LHOGV D PRGHO LQ ZKLFK \HDU DQG RSWLRQ DUH LQGHSHQGHQW FRQGLWLRQDO XSRQ VH[ 7KLV PRGHO LV ORJ PLM N 8L8M8N8LN8MN f PAGE 70 )LHQEHUJ DOVR VKRZV WKDW GHOHWLQJ WKH 8MWHUP IURP f WR REWDLQ ORJ PLMN 8L8M8IF8LN f \LHOGV D PRGHO WKDW VSHFLILHV WKDW RSWLRQ LV LQGHSHQGHQW RI VH[ %DVHG RQ )LHQEHUJnV SUHVHQWDWLRQ DQ DSSURSULDWH DQDO\VLV IRU WHVWLQJ WKH K\SRWKHVHV LV &RQGXFW D OLNHOLKRRG UDWLR WHVW RI WKH DGHTXDF\ RI PRGHO f ,I WKLV WHVW LV QRQVLJQLILFDQW WKHQ WKH GDWD VXSSRUW WKH DGHTXDF\ RI WKH PRGHOU DQG WKH K\SRWKHVLV WKDW FRQGLWLRQDO RQ JHQGHU} RSWLRQ DQG \HDU DUH LQGHSHQGHQW %HFDXVH PRGHO f LV D VDWXUDWHG PRGHO WHVWLQJ WKH DGHTXDF\ RI PRGHO f LV WKH VDPH DV FRPSDULQJ WKH DGHTXDFLHV RI PRGHOV f DQG f &RQGXFW D OLNHOLKRRG UDWLR WHVW FRPSDULQJ WKH DGHTXDFLHV RI PRGHOV f DQG f ,I WKLV WHVW LV VLJQLILFDQW WKHQ PRGHO f ILWV WKH GDWD EHWWHU WKDQ PRGHO f DQG WKH GDWD VXSSRUW WKH K\SRWKHVLV WKDW WKH FKRLFH RI LQFRUUHFW RSWLRQ LV GHSHQGHQW RQ VH[ 7R VXPPDUL]H LI WKH ILUVW WHVW LV QRQVLJQLILFDQW DQG WKH VHFRQG WHVW LV VLJQLILFDQW WKHQ WKH FKRLFH RI RSWLRQ LV GHSHQGHQW RQ VH[} DQG WKH SDWWHUQ RI GHSHQGHQF\ LV WKH VDPH IRU DOO WKUHH \HDUV PAGE 71 $ SUREOHP DULVHV LQ LQWHUSUHWLQJ WKLV DQDO\VLV ZLWK WKH GDWD XVHG LQ WKLV UHVHDUFK 2YHU WKH WKUHH \HDUVU WKHUH ZHUH UHVSRQVHV DYDLODEOH IURP r VWXGHQWV (YHQ LI RQO\ b RI WKH VWXGHQWV DQVZHUHG DQ LWHP LQFRUUHFWO\r WKH UHVSRQVHV RI r VWXGHQWV ZRXOG EH XVHG LQ DQDO\]LQJ WKLV LWHP 2Q WKH RWKHU KDQGr LI b DQVZHUHG DQ LWHP LQFRUUHFWO\r WKH UHVSRQVHV RI r VWXGHQWV ZRXOG EH XVHG LQ DQDO\]LQJ WKH GDWD $V D UHVXOW RI WKH ODUJH VDPSOH VL]Hr WKH WHVWV GHVFULEHG DERYH DUH OLNHO\ WR EH YHU\ SRZHUIXO ,Q VWHS RI WKH DQDO\VLVr WKHQr HYHQ D YHU\ VPDOO FKDQJH IURP \HDU WR \HDU LQ WKH SURSRUWLRQ RI PDOHV RU IHPDOHV ZKR FKRRVH DQ RSWLRQ LV DSW WR EH GHWHFWHGr DQG WKH UHVXOWV ZLOO LQGLFDWH WKDW K\SRWKHVLV LV QRW VXSSRUWHG )RU VWHS r HYHQ D YHU\ VPDOO GHSHQGHQFH RI RSWLRQ FKRLFH RQ VH[ LV OLNHO\ WR EH GHWHFWHG DQG K\SRWKHVLV LV OLNHO\ WR EH VXSSRUWHG ,Q EULHIr WKH SUREOHP FDXVHG E\ WKH ODUJH VDPSOH VL]H LV WKDW SUDFWLFDOO\ LQVLJQLILFDQW GLIIHUHQFHV PD\ \LHOG VWDWLVWLFDO VLJQLILFDQFH )RUWXQDWHO\r WKH IRUP RI WKH WHVW VWDWLVWLF XVHG LQ WKH OLNHOLKRRG UDWLR WHVW VXJJHVWV D UHDVRQDEOH VROXWLRQ WR WKLV SUREOHP 7KH WHVW VWDWLVWLF LV f PAGE 72 +HUH WKH VXPPDWLRQ LV RYHU DOO WKH FHOOV LQ WKH FRQWLQJHQF\ WDEOH ) UHIHUV WR WKH REVHUYHG IUHTXHQF\ LQ D FHOO DQG )H UHIHUV WR WKH HVWLPDWHG H[SHFWHG IUHTXHQF\ LQ D FHOO 'HQRWLQJ WKH REVHUYHG SURSRUWLRQ LQ D FHOO E\ 3 DQG WKH HVWLPDWHG H[SHFWHG SURSRUWLRQ LQ D FHOO E\ 3H WKH WHVW VWDWLVWLF FDQ EH ZULWWHQ 1 -3R rH 3RSH` f ZKHUH 1 LV WKH WRWDO QXPEHU RI VXEMHFWV 7KLV IRUP RI WKH WHVW VWDWLVWLF VXJJHVWV WKH IROORZLQJ VWUDWHJ\ )RU DQ\ VLJQLILFDQW XVLQJ 3 DQG 3H FDOFXODWHG IURP WKH WRWDO GDWD VHW DYDLODEOH IRU DQ LWHP FDOFXODWH WKH PLQLPXP 1 UHTXLUHG IRU WR EH VLJQLILFDQW ,I WKH PLQLPXP 1 LV YHU\ ODUJH WKLV VXJJHVWV WKDW WKH VWDWLVWLFDOO\ VLJQLILFDQW UHVXOW LV QRW SUDFWLFDOO\ VLJQLILFDQW VLQFH LW FDQ RQO\ EH GHWHFWHG LQ YHU\ ODUJH VDPSOHV 2I FRXUVH WKH TXHVWLRQ UHPDLQV DV WR ZKDW FDQ EH FRQVLGHUHG D PLQLPXP ODUJH 1 $OWKRXJK WKHUH LV URRP IRU DUJXPHQW LW VHHPV UHDVRQDEOH WR FODLP WKDW LI DQ DYHUDJH RI VXEMHFWV SHU \HDU LV UHTXLUHG WR VKRZ VLJQLILFDQFH WKHQ WKH UHVXOW LV QRW OLNHO\ WR EH SUDFWLFDOO\ VLJQLILFDQW 2Q WKH EDVLV RI WKLV UHDVRQLQJ LW ZDV SURSRVHG WR LJQRUH DOO VLJQLILFDQW UHVXOWV WKDW ZRXOG EH QRQVLJQLILFDQW LI WKHUH ZHUH OHVV WKDQ VXEMHFWV DYDLODEOH ,Q DGGLWLRQ DOO ORJOLQHDU PAGE 73 PRGHO WHVWV ZHUH FRQGXFWHG XVLQJ D OHYHO RI VLJQLILFDQFH 6LQFH WKLV UHVHDUFK LV EDVHG RQ 0DUVKDOO f LW LV LPSRUWDQW WR FRPSDUH WKH PHWKRG RI DQDO\VLV XVHG LQ WKLV VWXG\ WR WKH RQH XVHG E\ 0DUVKDOO 0DUVKDOO DOVR XVHG D WZRVWHS DQDO\VLV ,Q WKH ILUVW VWHS RI KHU DQDO\VLV VKH GHOHWHG WKH 8LMN WHUP IURP f DQG WHVWHG WKH DGHTXDF\ RI WKH PRGHO r J P M M N 8L 8M 8IF8LM 8LN8MN f )ROORZLQJ WKLV VKH GHOHWHG WKH 8MN WHUP WR REWDLQ ORJ PLM 8L8M 8NLM 8LN f DQG FRPSDUHG WKHVH WZR PRGHOV XVLQJ D OLNHOLKRRG UDWLR WHVW ,I WKH ILUVW WHVW ZDV QRQVLJQLILFDQW DQG WKH VHFRQG VLJQLILFDQW 0DUVKDOO FODLPHG WKDW RSWLRQ FKRLFH ZDV GHSHQGHQW RQ VH[ DQG WKH SDWWHUQ RI GHSHQGHQF\ ZDV WKH VDPH IURP \HDU WR \HDU 7KLV LV WKH VDPH FODLP WKDW WKLV VWXG\ VRXJKW WR HVWDEOLVK +RZHYHU WKH DSSURDFK XVHG KHUH ZDV WR SUHVHQW HYLGHQFH WKDW PRGHO f ILWV WKH GDWD ZKLOH 0DUVKDOO WULHG WR VKRZ WKDW PRGHO f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r FRQVLGHU WKH WKUHH WZRE\WZR WDEOHV REWDLQHG E\ FURVV FODVVLI\LQJ VH[ DQG RSWLRQ FKRLFHV RQH DQG WZR IRU HDFK \HDU 7KHVH WDEOHV DUH LQGLFDWHG E\ GRWWHG OLQHV RQ 7DEOH )RU WKLV WDEOHr WKH UDWLR RI WKH RGGV RI D PDOH FKRRVLQJ RSWLRQ RQH WR WKH RGGV RI D IHPDOH FKRRVLQJ RSWLRQ RQH LV WKH VDPH ZLWKLQ WUXQFDWLQJ HUURUf IRU HDFK \HDU )RU H[DPSOHr WKLV RGGV UDWLR IRU WKH ILUVW \HDU LV f f r :LWKLQ WKH HUURU FDXVHG E\ UHSRUWLQJ WUXQFDWHG SUREDELOLWLHVr WKH RGGV UDWLR IRU \HDUV WZR DQG WKUHH LV WKH VDPH DV WKDW IRU \HDU RQH 7KH HTXDOLW\ RI WKHVH RGGV UDWLRV LV 0DUVKDOOnV RSHUDWLRQDOL]DWLRQ VWDJH RI WKH \HDU WR\HDU JHQGHURSWLRQ GHSHQGHQF\ 7R VKRZ WKDW WKH RGGV UDWLR FDQ EH FRQVWDQW RYHU \HDUVr EXW WKDW WKH SUREDELOLWLHV RI RSWLRQ FKRLFH FRQGLWLRQDO RQ VH[ DQG \HDU FDQ FKDQJH IURP \HDU WR \HDUr IRU ERWK PDOHV DQG IHPDOHVr WKH SUREDELOLWLHV LQ 7DEOH ZHUH FRQYHUWHG WR WKH SUREDELOLWLHV RI RSWLRQ FKRLFH FRQGLWLRQDO XSRQ VH[ DQG \HDU 7KHVH FRQGLWLRQDO SUREDELOLWLHV DUH UHSRUWHG LQ 7DEOH 8QOLNH WKH SUREDELOLWLHV LQ 7DEOH r WKRVH LQ 7DEOH FKDQJH IURP \HDU WR \HDU IRU ERWK PDOHV DQG IHPDOHV PAGE 76 7DEOH +\SRWKHWLFDO 3UREDELOLWLHV RI 2SWLRQ &KRLFH &RQGLWLRQDO RQ 6H[ DQG PAGE 77 FRQGXFWHGr RQH ZLWK HDFK RI DGGLWLRQr VXEWUDFWLRQr PXOWLSOLFDWLRQr GLYLVLRQr DGGLWLRQ RI IUDFWLRQVr GHFLPDOV LQ VXEWUDFWLRQr DQG HTXLYDOHQFH DV FRYDULDWHV $ ZHOO NQRZQ SUREOHP WKDW DULVHV LQ WKH XVH RI $1&29$ LV WKDW DQ XQUHOLDEOH FRYDULDWH FDQ FDXVH VSXULRXV GLIIHUHQFHV EHWZHHQ WKH VH[ JURXSV 7R VROYH WKLV SUREOHPr 3RUWHU f SURSRVHG WKH XVH RI HVWLPDWHG WUXH VFRUHV IRU REVHUYHG VFRUHV 3RUWHU f FRQGXFWHG D VLPXODWLRQ WKDW JDYH HPSLULFDO VXSSRUW WR WKH DGHTXDF\ RI WKLV VWUDWHJ\ +XQWHU DQG &RKHQ f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r ORJ P M M N 8Â 8 M 8 N8Â N8 M N f 6XEVWDQWLYHO\ WKLV PRGHO LPSOLHV WKDW WKH SDWWHUQ RI PDOH DQG IHPDOH RSWLRQ FKRLFHV LV FRQVLVWHQW RYHU WKH WKUHH \HDUV RI WHVW DGPLQLVWUDWLRQ 7KH VHFRQG PRGHO LQGLFDWHV WKDW RSWLRQ LV LQGHSHQGHQW RI VH[ ORJ PLMN 8L8M8N8LN f 7KLV PRGHO LPSOLHV WKDW WKH SDWWHUQ RI RSWLRQ FKRLFH LV WKH VDPH IRU PDOHV DQG IHPDOHV ,Q RUGHU WR GHWHUPLQH LI PDOHV DQG IHPDOHV GLIIHUHG LQ WKH VHOHFWLRQ RI LQFRUUHFW UHVSRQVHV DQG LI WKHVH GLIIHUHQFHV ZHUH VWDEOH DFURVV WKH WKUHH \HDUV RI WHVW DGPLQLVWUDWLRQ D WZRVWHS WHVW ZDV SHUIRUPHG )LUVW D OLNHOLKRRG UDWLR WHVW RI WKH DGHTXDF\ RI PRGHO f ZDV FRQGXFWHG IRU HDFK RI WKH LWHPV RI WKH WHVW 8QGHU WKLV WHVW D PRGHO ILWV WKH GDWD ZKHQ WKH YDOXH REWDLQHG LV QRQVLJQLILFDQW DW D VSHFLILHG DOSKD OHYHO )RU bf RI WKH LWHPV PRGHO f ILWWHG WKH GDWD DGHTXDWHO\ )RU WKHVH LWHPV WKH SDWWHUQ RI PDOH DQG IHPDOH RSWLRQ FKRLFHV ZDV FRQVLVWHQW IRU WKH WKUHH \HDUV 7KH ;A YDOXHV IRU WKH LWHPV IRU ZKLFK WKH OLNHOLKRRG UDWLR PAGE 80 WHVWV ZHUH VLJQLILFDQW DW WKH OHYHO DUH UHSRUWHG LQ 7DEOH $OVR UHSRUWHG LQ WKLV WDEOH DUH WKH DFWXDO VDPSOH VL]HV DQG WKH PLQLPXP VDPSOH VL]HV QHFHVVDU\ IRU WKH OLNHOLKRRG UDWLR WHVWV WR EH VLJQLILFDQW 2I WKH LWHPV KDG PLQLPXP VDPSOH VL]HV JUHDWHU WKDQ 7KXV DOWKRXJK IRU WKH WKUHH \HDUV ERWK PDOHV DQG IHPDOHV VDPSOHV KDG LQFRQVLVWHQW RSWLRQ FKRLFHV RQ WKHVH LWHPV RQ RI WKH LWHPV WKH LQFRQVLVWHQF\ ZDV UHODWLYHO\ PLQRU &RQVHTXHQWO\ WKHVH LWHPV ZHUH LQFOXGHG LQ VWHS DORQJ ZLWK WKH LQLWLDO LWHPV ,Q VWHS D OLNHOLKRRG UDWLR WHVW ZDV SHUIRUPHG FRPSDULQJ WKH DGHTXDF\ RI WKH PRGHOV LQ f DQG f 7KH ; YDOXHV DVVRFLDWHG ZLWK PRGHOV f DQG f IRU HDFK RI WKH LWHPV VXEMHFWHG WR VWHS DUH UHSRUWHG LQ FROXPQV E DQG F RI 7DEOH $OVR UHSRUWHG LQ 7DEOH LV WKH GLIIHUHQFH EHWZHHQ WKH WZR ; YDOXHV VHH FROXPQ Gf 7KLV ODWWHU ILJXUH LV WKH WHVW VWDWLVWLF IRU FRPSDULQJ WKH DGHTXDFLHV RI PRGHOV f DQG f 6LJQLILFDQW ; VWDWLVWLFV DUH LQGLFDWHG E\ DVWHULVNV 7KH ; VWDWLVWLFV ZHUH VLJQLILFDQW IRU RI WKH LWHPV LQGLFDWLQJ D PDOHIHPDOH GLIIHUHQFH LQ RSWLRQ FKRLFH IRU WKHVH LWHPV ,Q FROXPQ H RI 7DEOH WKH DFWXDO VDPSOH VL]HV DUH UHSRUWHG ,Q FROXPQ I WKH PLQLPXP VDPSOH VL]HV QHFHVVDU\ IRU VLJQLILFDQFH DUH UHSRUWHG )RU WKRVH LWHPV ZKLFK KDG VLJQLILFDQW ; VWDWLVWLFV UHSRUWHG LQ FROXPQ G bf KDG PLQLPXP VDPSOH VL]HV JUHDWHU WKDQ PAGE 81 7$%/( 6XPPDU\ RI 6LJQLILFDQW 7HVWV RI 0RGHO ,WHP 1XPEHU ;r IRU 0RGHO r $FWXDO 6DPSOH 6L]H 0LQ 6DPSOH 6L]H IRU 6LJQLILFDQFH Df Ef Ff Gf PAGE 82 8/ 7$%/( &RQWLQXHG ; IRU 0RGHO r $FWXDO 6DPSOH 6L]H 0LQ 6DPSOH 6L]H IRU 6LJQLILFDQFH Ef Ff Gf rr PAGE 83 7$%/( &RQWLQXHG [r IRU 0RGHO r $FWXDO 6DPSOH 6L]H 0LQ 6DPSOH 6L]H IRU 6LJQLILFDQFH OEf Ff Gf r rr PAGE 84 7$%/( &RQWLQXHG ,WHP 1XPEHU ;D IRU 0RGHO r $FWXDO 6DPSOH 6L]H 0LQ 6DPSOH 6L]H IRU 6LJQLILFDQFH Df Ef &Ff Gf r S rr 0LQLPXP VDPSOH VL]H OHVV WKDQ PAGE 85 7$%/( &KLVTXDUH 9DOXHV IRU WKH &RPSDULVRQ RI 0RGHOV f DQG f ZLWK $FWXDO 6DPSOH 6L]HV DQG ZLWK WKH &RUUHVSRQGLQJ 1XPEHU RI 6XEMHFWV 1HHGHG IRU 6LJQLILFDQW 5HVXOWV ,WHP 1XPn EHU ;D IRU 0RGHO ;D IRU 0RGHO ;D IRU 'LIIHUHQFH $FWXDO 6DPSOH 6L]H 0LQLPXP 6DPSOH 6L]H IRU 6LJQLI Df OEf Ff Gf Hf If r U r rr r r r rr r r r rr r r r r r r r r PAGE 86 7$%/( &RQWLQXHG ,WHP 1XPn EHU ; IRU 0RGHO ; IRU 0RGHO ; IRU 'LIIHUHQFH $FWXDO 6DPSOH 6L]H 0LQLPXP 6DPSOH 6L]H IRU 6LJQLI Df WEf Ff Gf Hf If r r r r r r r r r r r r r r r PAGE 87 7$%/( &RQWLQXHG ,WHP 1XPn EHU ;r IRU 0RGHO ;D IRU 0RGHO ; IRU 'LIIHUHQFH $FWXDO 6DPSOH 6L]H 0LQLPXP 6DPSOH 6L]H IRU 6LJQLI Df Ef Ff Gf Hf If r r r r rr r rr r rr r r r r r r r r rr r rr r r r PAGE 88 7$%/( &RQWLQXHG ,WHP 1XPn EHU ; IRU 0RGHO ;D IRU 0RGHO ;r IRU 'LIIHUHQFH $FWXDO 6DPSOH 6L]H 0LQLPXP 6DPSOH 6L]H IRU 6LJQLI Df Ef Ff Gf Hf If r r r r r r rr r r r r r r r r r r r r PAGE 89 7$%/( &RQWLQXHG ,WHP 1XPn EHU ;D IRU 0RGHO ;] IRU 0RGHO ;] IRU 'LIIHUHQFH $FWXDO 6DPSOH 6L]H 0LQLPXP 6DPSOH 6L]H IRU 6LJQLI Df OEf Ff Gf Hf If r r r r r r r r r r r rr r rr r r r r r r PAGE 90 7$%/( &RQWLQXHG ,WHP 1XPn EHU ;D IRU 0RGHO ;D IRU 0RGHO ; IRU 'LIIHUHQFH $FWXDO 6DPSOH 6L]H 0LQLPXP 6DPSOH 6L]H IRU 6LJQLI ODf OEf Ff Gf OHOO &If r r r r r r r r r r r r r r r r 3& 0LQLPXP VDPSOH VL]H OHVV WKDQ PAGE 91 6A[AHODMWHG'LLIHUHQFHV LQ 3UREOHP 6ROYLQJ 3HUIRUPDQFH ,Q WKLV SDUWr UHVXOWV IRU WKH WZR TXHVWLRQV UHODWHG WR WKH VHFRQG REMHFWLYH RI WKH VWXG\ DUH SUHVHQWHG 7KHVH TXHVWLRQV VHUYH DV WKH IUDPHZRUN IRU WKH SUHVHQWDWLRQ (DFK TXHVWLRQ LV VWDWHGr IROORZHG E\ WKH UHVXOWV SHUWDLQLQJ WR WKDW TXHVWLRQ 4XHVWLRQ 'R PDOHV DQG IHPDOHV GLIIHU LQ SUREOHP VROYLQJ SHUIRUPDQFH" 7KH UHVSRQVHV RI PDOHV DQG IHPDOHV ZKR WRRN WKH 3XHUWR 5LFR %DVLF 6NLOOV 7HVW GXULQJ WKH VSULQJ RI WKH ILUVW \HDU ZHUH DQDO\]HG LQ WKLV VWXG\ $OVRr GDWD IURP PDOHV DQG IHPDOHV WHVWHG LQ WKH VHFRQG \HDU DQG IURP PDOHV DQG IHPDOHV WHVWHG LQ WKH WKLUG \HDU ZHUH LQFOXGHG LQ WKH DQDO\VLV 7KH PHDQ SHUIRUPDQFH VFRUHV DQG WKH VWDQGDUG GHYLDWLRQV IRU HDFK RI WKH HLJKW YDULDEOHV DUH SUHVHQWHG LQ 7DEOH 5HVXOWV RI WWHVWV DUH DOVR SUHVHQWHG LQ WKLV WDEOH )HPDOHV RXWSHUIRUPHG PDOHV LQ SUREOHP VROYLQJr D ILQGLQJ FRQVLVWHQWO\ SUHVHQW LQ DOO WKH WKUHH \HDUV RI WHVW DGPLQLVWUDWLRQ 2YHU WKH WKUHH\HDU SHULRG WKH PHDQ GLIIHUHQFHV IDYRUHG IHPDOHV LQ DOO YDULDEOHV H[FHSW HTXLYDOHQFH 7KH VH[UHODWHG GLIIHUHQFHV LQ SUREOHP VROYLQJ ZHUH VLJQLILFDQW Sf IRU DOO WKUHH \HDUV PAGE 92 7$%/( 0HDQVr 6WDQGDUG 'HYLDWLRQVr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f S& DQG VH[ DQG GLYLVLRQ ) f S $ VLJQLILFDQW LQWHUDFWLRQ ZDV IRXQG EHWZHHQ VH[ DQG VXEWUDFWLRQ LQ WKH VHFRQG \HDU ) f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fÂ§7 0XOWLSOLFDWLRQ )LJ 6H[ E\ 0XOWLSOLFDWLRQ ,QWHUDFWLRQ PAGE 98 3UREOHP 6ROYLQJ 7 7 7 'LYLVLRQ )LJ 6H[ E\ 'LYLVLRQ ,QWHUDFWLRQ PAGE 99 3UREOHP 6ROYLQJ 6XEWUDFWLRQ )LJ 6H[ E\ 6XEWUDFWLRQ ,QWHUDFWLRQ PAGE 100 7KH VLJQLILFDQW LQWHUDFWLRQV IRXQG EHWZHHQ VH[ DQG PXOWLSOLFDWLRQr DQG VH[ DQG GLYLVLRQ ILUVW \HDUfr DQG VH[ DQG VXEWUDFWLRQ VHFRQG \HDUfr DQVZHUHGr LQ SDUWr WKH TXHVWLRQ RI ZKHWKHU PDOHIHPDOH GLIIHUHQFHV LQ SUREOHP VROYLQJ SHUIRUPDQFH GHSHQG RQ FRPSXWDWLRQDO DELOLW\ +RZHYHUr WKH HYLGHQFH LV TXLWH ZHDN 2I SRVVLEOH LQWHUDFWLRQVr RQO\ WKUHH ZHUH VLJQLILFDQW 1R YDULDEOH H[KLELWHG D VLJQLILFDQW LQWHUDFWLRQ IRU HDFK RI WKH WKUHH \HDUV 7KH DQDO\VLV RI FRYDULDQFH ZDV DOVR XVHG WR GHWHUPLQH LI VH[UHODWHG GLIIHUHQFHV H[LVW DIWHU FRQWUROOLQJ IRU FRPSXWDWLRQDO VNLOOV $QDO\VHV ZHUH FRQGXFWHG IRU WKRVH YDULDEOHV WKDW GLG QRW H[KLELW VLJQLILFDQW LQWHUDFWLRQV ZLWK VH[ $V GLVFXVVHG LQ &KDSWHU ,,,r HVWLPDWHG WUXH VFRUHV ZHUH XVHG IRU REVHUYHG VFRUHV WR DGMXVW IRU XQUHOLDELOLW\ RI WKH FRYDULDWHV WKH FRPSXWDWLRQDO VXEWHVWVf 5HOLDELOLW\ FRHIILFLHQWV FDOFXODWHG IRU HDFK FRYDULDWH DUH VKRZQ LQ 7DEOH 6XPPDULHV RI WKH DQDO\VHV RI FRYDULDQFH IRU WKH ILUVW \HDU DUH UHSRUWHG LQ 7DEOH 7KH UHVXOWV VKRZ WKDW IHPDOHV UHWDLQHG WKHLU VXSHULRULW\ LQ SUREOHP VROYLQJ SHUIRUPDQFH ZKHQ HTXLYDOHQFH ZDV WKH FRQWUROOLQJ YDULDEOH LQ WKH DQDO\VLV RI FRYDULDQFHr WKH RQO\ YDULDEOH LQ ZKLFK PDOHV RXWSHUIRUPHG IHPDOHV QRQVLJQLILFDQWf :KHQ WKH FRQWUROOLQJ YDULDEOHV ZHUH DGGLWLRQr VXEWUDFWLRQr DGGLWLRQ RI IUDFWLRQVr DQG VXEWUDFWLRQ ZLWK GHFLPDOVr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r E\ &RYDULDWH DQG 6H[ )LUVW PAGE 105 7$%/( $1&29$ 6XPPDU\ 7DEOH 2WKHU &RYDULDWHV 6HFRQG PAGE 106 7$%/( $GMXVWHG 0HDQV RQ 3UREOHP 6ROYLQJr E\ &RYDULDWH DQG 6H[ 6HFRQG PAGE 107 7$%/( $1&29$ 6XPPDU\ 7DEOH 7KLUG PAGE 108 7$%/( $GMXVWHG 0HDQV RQ 3UREOHP 6ROYLQJr E\ &RYDULDWH DQG 6H[ 7KLUG PAGE 109 UHVSRQVHV ZDV FRQVLVWHQW WKURXJKRXW WKH \HDUV LQ ZKLFK WKH WHVW ZDV DGPLQLVWHUHG 0DOHV DQG IHPDOHV VHOHFWHG GLIIHUHQW LQFRUUHFW UHVSRQVHV LQ RI WKH LWHPV RI WKH WHVW 7KH SDWWHUQ RI PDOHIHPDOH GLIIHUHQFHV LQ WKH VHOHFWLRQ RI LQFRUUHFW UHVSRQVHV ZDV FRQVLVWHQWO\ IRXQG LQ HDFK RI WKH \HDUV RI WHVW DGPLQLVWUDWLRQ +RZHYHUr IRU WKH YDVW PDMRULW\ RI WKH LWHPVr PDOHIHPDOH GLIIHUHQFHV ZHUH UHODWLYHO\ VPDOO LQ PDJQLWXGHr FRQVLGHULQJ WKH IDFW WKDW WKH QXPEHU RI VXEMHFWV QHHGHG WR REWDLQ VLJQLILFDQFH ZDV YHU\ ODUJH 7KHUHIRUHr WKHVH ILQGLQJV ODFN HGXFDWLRQDO VLJQLILFDQFH $ VHFRQG REMHFWLYH RI WKH VWXG\ ZDV WR LQYHVWLJDWH ZKHWKHU PDOHV DQG IHPDOHV GLIIHU LQ SUREOHP VROYLQJ SHUIRUPDQFHr LI WKHVH GLIIHUHQFHV SHUVLVW DIWHU FRQWUROOLQJ IRU FRPSXWDWLRQDO VNLOOVr DQG LI WKH PDOHIHPDOH GLIIHUHQFHV GHSHQG RQ WKH OHYHO RI FRPSXWDWLRQDO VNLOOV 5HVXOWV RI WKH DQDO\VHV UHSRUWHG LQ WKLV VHFWLRQ JHQHUDOO\ LQGLFDWHG WKDW IHPDOHV RXWSHUIRUPHG PDOHV QRW RQO\ LQ SUREOHP VROYLQJr EXW LQ VL[ RI WKH VHYHQ FRPSXWDWLRQDO YDULDEOHV 0DOHV VXUSDVVHG IHPDOHV LQ HTXLYDOHQFHr EXW VWDWLVWLFDO VLJQLILFDQFH ZDV REWDLQHG IRU RQO\ RQH RI WKH WKUHH \HDUV FRYHUHG LQ WKH VWXG\ 7KH UHVXOWV DOVR WHQGHG WR VKRZ WKDWr IRU H[DPLQHHV ZLWK VLPLODU OHYHOV RI FRPSXWDWLRQDO VNLOOVr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f VXPPDU\ DQG LQWHUSUHWDWLRQ RI WKH UHVXOWV RI WKH VWXG\ DQG Ef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nV ILQGLQJV f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f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f DQG VXEWUDFWLRQ IRU WKH VHFRQG \HDUf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f DQG )HQQHPD DQG 7DUWUHH f ZHUH QRW VXSSRUWHG LQ WKLV LQYHVWLJDWLRQ )HPDOH SHUIRUPDQFH ZDV HTXDO RU KLJKHU WKDQ PDOH SHUIRUPDQFH LQ SUREOHP VROYLQJ 7KH ILQGLQJV RI WKH VWXG\ KDYH VRPHWKLQJ LQ FRPPRQ ZLWK RWKHU LQYHVWLJDWLRQV ZKHUH PDWKHPDWLFV SUREOHP VROYLQJ KDV EHHQ WKH VXEMHFW RI LQWHUHVW WKH IDFW WKDW PDOHV DQG IHPDOHV GR SRRUO\ LQ ZRUGVWRU\ SUREOHP LWHPV SUREOHP VROYLQJf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Â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nV f ILQGLQJV WKDW WKH HIIHFW RI FRPSXWDWLRQDO VNLOOV RQ SUREOHP VROYLQJ SHUIRUPDQFH LV GLIIHUHQW IRU HDFK VH[ +RZHYHU WKH ILQGLQJV RI WKH SUHVHQW LQYHVWLJDWLRQ DUH FRQVLVWHQW ZLWK WKRVH IRXQG E\ 0H\HU f DQG :KLWDNHU f IHPDOH SHUIRUPDQFH LQ SUREOHP VROYLQJ LV QRW VLJQLILFDQWO\ GLIIHUHQW IURP PDOH SHUIRUPDQFH )HPDOHV VKRZHG VXSHULRULW\ LQ SUREOHP VROYLQJ SHUIRUPDQFH LQ DOO WKUHH \HDUV RI WHVW DGPLQLVWUDWLRQ DOWKRXJK WKHLU VXSHULRULW\ ZDV UHWDLQHG RQO\ ZKHQ SHUIRUPDQFH LQ HTXLYDOHQFH ZDV FRQWUROOHG IRU IRU WKH WKUHH \HDUV GDWDf DQG ZKHQ SHUIRUPDQFH LQ VXEWUDFWLRQ ZDV FRQWUROOHG IRU VHFRQG \HDUf VWDWLVWLFDOO\ VLJQLILFDQW RQO\ DIWHU DFFRXQWLQJ IRU HTXLYDOHQFHf 7KH ILQGLQJV RI WKLV VWXG\ DUH DOVR LQ DJUHHPHQW ZLWK WKRVH UHSRUWHG E\ WKH 1DWLRQDO $VVHVVPHQW RI (GXFDWLRQDO 3URJUHVV 1$(3f &DUSHQWHU HW DO] f 5HODWHG WR SUREOHP VROYLQJ WKH\ FRQFOXGH LI LW ZHUH QHFHVVDU\ WR VLQJOH RXW RQH DUHD WKDW GHPDQGV XUJHQW DWWHQWLRQ LW ZRXOG FOHDUO\ EH SUREOHP VROYLQJ $W DOO DJH OHYHOV DQG LQ YLUn WXDOO\ HYHU\ FRQWHQW DUHD SHUIRUPDQFH ZDV H[WUHn PHO\ ORZ RQ H[HUFLVHV UHTXLULQJ SUREOHP VROYLQJ RU DSSOLFDWLRQ RI PDWKHPDWLFDO VNLOOV ,Q JHQHn UDO UHVSRQGHQWV GHPRQVWUDWHG D ODFN RI WKH PRVW EDVLF SUREOHPVROYLQJ VNLOOV 5DWKHU WKDQ DWWHPSWn LQJ WR WKLQN WKURXJK D SUREOHP DQG ILJXUH RXW ZKDW QHHGHG WR EH GRQH WR VROYH WKH SUREOHP PRVW UHVn SRQGHQWV VLPSO\ WULHG WR DSSO\ D VLQJOH DULWKPHWLF PAGE 116 RSHUDWLRQ WR WKH QXPEHUV LQ WKH SUREOHP 7KH UHn VXOWV LQGLFDWH WKDW VWXGHQWV DUH QRW IDPLOLDU ZLWK VXFK EDVLF SUREOHPVROYLQJ VWUDWHJLHV DV GUDZLQJ D SLFWXUH RI D ILJXUH GHVFULEHG LQ D SUREOHP RU FKHFNLQJ WKH UHDVRQDEOHQHVV RI D UHVXOW S f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f $OVR LQ JHQHUDO WKH HIIHFW RI FRPSXWDWLRQDO VNLOOV LV VLPLODU IRU HDFK VH[ %DVHG RQ WKHVH ILQGLQJV LW DSSHDUV WKDW JUHDWHU GLIIHUHQFHV H[LVW ZLWKLQ HDFK VH[ WKDQ EHWZHHQ WKH VH[HV ,Q FRQFOXVLRQ WKH ILQGLQJV RI WKH SUHVHQW VWXG\ KLJKOLJKW WKH QHHG IRU LQYHVWLJDWLRQV RI GLIIHUHQW DSSURDFKHV WR LQVWUXFWLRQDO GHVLJQV DLPHG DW LPSURYLQJ ERWK PDOH DQG IHPDOH SHUIRUPDQFH LQ SUREOHP VROYLQJ PAGE 117 5()(5(1&(6 $LNHQ / 5 f $WWLWXGHV WRZDUG PDWKHPDWLFV 5H PAGE 118 %HUQ 6 / t %HUQ 9 f :H DUH DOO QRQFRQVFLRXV VH[LVWV 3V\FKRORJ\ 7RGD\ %HQERZ & 3 t 6WDQOH\ & f 6H[ GLIIHUHQFHV LQ PDWKHPDWLFDO DELOLW\ )DFW RU DUWLIDFW" 6FLHQFH %HQERZ & 3 t 6WDQOH\ & f &RQVHTXHQFHV LQ KLJK VFKRRO DQG FROOHJH RI VH[ GLIIHUHQFHV LQ PDWKHPDn WLFDO UHDVRQLQJ DELOLW\ $ ORQJLWXGLQDO SHUVSHFWLYH $PHULFDQ (GXFDWLRQDO 5HVHDUFK -RXUQDO %REELH & 1 f 6H[UROH SUHIHUHQFH DQG DFDGHPLF DFKLHYHPHQW 'RFWRUDO GLVVHUWDWLRQ PAGE 119 &UHVZHOO / )HEUXDU\f 6H[UHODWHG GLIIHUHQFHV LQ SUREOHP VROYLQJ LQ UXUDO EODFN DQJOR DQG &KLFDJR DGROHVFHQWV 3DSHU SUHVHQWHG DW WKH DQQXDO PHHWLQJ RI WKH 6RXWKZHVW (GXFDWLRQDO 5HVHDUFK $VVRFLDWLRQ $XVWLQ 7; (5,& 'RFXPHQW 5HSURGXFWLRQ 6HUYLFH 1R (' f 'Z\HU & $ f ,QIOXHQFH RI FKLOGUHQnV VH[ UROH VWDQGDUGV RQ UHDGLQJ DQG DULWKPHQWLF DFKLHYHPHQW -RXUQDO RI (GXFDWLRQDO 3V\FKRORJ\ ee (FFOHV f 6H[ GLIIHUHQFHV LQ PDWKHPDWLFV SDUWLFLSDWLRQ ,Q 1 6WHLQNDPS t 0 0DHKU (GVf :RPHQ LQ 6FLHQFH SS f *UHHQZLFK &7 -$, 3UHVV (OWRQ & ) t 5RVH + $ f 7UDGLWLRQDO VH[ DWWLWXGHV DQG GLVFUHSDQW DELOLW\ PHDVXUHV LQ FROOHJH ZRPHQ M4XFSDL RI &4KKVHOLQJ 3V\FRORJ\r +U (UQHVW f 0DLKHPWLFV DQG 6H[ 6DQWD %DUEDUD 8QLYHUVLW\ RI &DOLIRUQLD ([HGLVLV 5 + f $Q LQYHVWLJDWLRQ RI WKH UHODWLRQ VKLS RI UHDGLQJ FRPSUHKHQVLRQ YRFDEXODU\ PDWKHPDWLFDO FRQFHSWV DQG FRPSXWDWLRQ RQ SUREOHP VROYLQJ DPRQJ DQJOR] EODFN DQG &KLFDJR PDOH DQG IHPDOH PLGGOH VFKRRO DGROHVFHQWV 'RFWRUDO 'LVVHUWDWLRQ 8QLYHUVLW\ RI +RXVWRQ f QLBVVJUWDWLHLLB$EVWUDFWV ,QWHUQDWLRQDO $$ )HQQHPD ( f 0DWKHPDWLFV OHDUQLQJ DQG WKH VH[HV $ UHYLHZ -RXUQDO IRU 5HVHDUFK LQ 0DWKHPDWLFV (GXFIOWLRML )HQQHPD ( f 0DWKHPDWLFV VSDWLDO DELOLW\ DQG WKH VH[HV ,Q ( )HQQHPD (Gf 0DWKHPDWLFV :KDW UBHVJDUFKBVDYV DERXWBVH[ GLIIHUHQFHV SS f &ROXPEXV 2KLR 6WDWH 8QLYHUVLW\ &ROOHJH RI (GXFDWLRQ (5,& 'RFXPHQW 5HSURGXFWLRQ 6HUYLFH 1R (' f )HQQHPD ( f ,QIOXHQFHV RI VHOHFWHG FRJQLWLYH DIFIFHFWLDKGBHGXFDWÂSQDOBYDULDEOHV RQ VH[UHODWHG BGMMIHUJMLFHV LQPDWKHPDWLFV OHDUQLQJ DQG VWXG\LQJ 0DGLVRQ 8QLYHUVLW\ RI :LVFRQVLQ 'HSDUWPHQW RI &XUULFXOXP DQG ,QVWUXFWLRQ (5,& 'RFXPHQW 5HSURGXFWLRQ 6HUYLFH 1R (' f PAGE 120 )HQQHPD} (U t 6KHUPDQ f 6H[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFV DFKLHYHPHQW VSDWLDO YLVXDOL]DWLRQ DQG DIIHFWLYH IDFWRUV $PHULFDQ (GXFDWLRQDO 5HVHDUFK -RXUQDO )HQQHPD ( t 6KHUPDQ $ f 6H[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFV DFKLHYHPHQW DQG UHODWHG IDFWRUV $ IXUWKHU VWXG\ -RXUQDO IRU 5HVHDUFK LQ 0DWKHPDWLFV(GXFDWLRQr K )HQQHPD ( t 7DUWUHH / $ f 5HVHDUFK RQ UHODWLRQVKLS RIBVSDWLVOYLVXDOL]DWL4'BDQGFRQILGHQFH WR PDOHIHPDOH PDWKHPDWLFV DFKLHYHPHQW LQ JUDGHV :DVKLQJWRQ '& 1DWLRQDO 6FLHQFH )RXQGDWLRQ (5,& 'RFXPHQW 5HSURGXFWLRQ 6HUYLFH 1R (' f )LHQEHUJ V H f 7KH$QDO\VLVSI&F46V FLDDVLILHG &DWHJRULFDO 'DWD &DPEULGJH 0$ 7KH 0,7 3UHVV )R[ / + Df 0DWKHPDWLFDOO\ SUHFRFLRXV 0DOH RU IHPDOH ,Q ( )HQQHPD (Gf 0DWKHPDWLFV OHDUQLQJ ZKDWBUe6eDUFKVD\6E4XW VHDBGLIIHUHQFHV SS f &ROXPEXV 2KLR 6WDWH 8QLYHUVLW\ &ROOHJH RI (GXFDWLRQ (5,& 'RFXPHQW 5HSURGXFWLRQ 6HUYLFH 1R (' f )R[ / + Ef 6H[ GLIIHUHQFHV ,PSOLFDWLRQV IRU SURJUDPBSODQQLQJ IRUWKH DFDGHPLFDOO\ JLIWHGW 3DSHU SUHVHQWHG DW WKH /HZLV 0 7HUPDQ 0HPRULDO 6\PSRVLXP RQ ,QWHOOHFWXDO 7DOHQW KHOG DW WKH -RKQV +RSNLQV 8QLYHUVLW\ %DOWLPRUH 0' )R[ / + f 7KH HIIHFWV RI VH[ UROH VRFLDOL]DWLRQ RQ PDWKHPDWLFV SDUWLFLSDWLRQ DQG DFKLHYHPHQW ,Q VKRHPDNHU (Gf ZRPHQBDQG0DWKHPDWLFV 5HVHDUFK 3HUVSHFWLYHV IRU &KDQJH 1,( 3DSHUV LQ (GXFDn WLRQ DQG :RUN 1R f :DVKLQJWRQ & (GXDWLRQ DQG :RUN *URXS 7KH 1DWLRQDO ,QVWLWXWH RI (GXFDWLRQ *OHQQRQ 9 t &DOODKDQ / f (OHPHQWDU\ VFKRRO PDWKHPDWLFV $AJXLGH WF FXUUHQW UHVHDUFK :DVKLQJWRQ & $VVRFLDWLRQ IRU 6XSHUYLVLRQ DQG &XUULFXOXP 'HYHORSPHQW *XD\ 5 % t 0F'DQLHO f 7KH UHODWLRQVKLS EHWZHHQ PDWKHPDWLFV DFKLHYHPHQW DQG VSDFH DELOLW\ DPRQJ HOHPHQWDU\ VFKRRO FKLOGUHQ -RXUQDO IRU 5HVHDUFKLQ0DWKDPDWLFV (GXFDWLRQr er PAGE 121 ,OO +DQVHQr & : f )DFWRUV DVVRFLDWHG ZLWK VXFFHVVIXO DFKLHYHPHQW LQ SUREOHP VROYLQJ LQ VL[WK JUDGH DULWKPHQWLF MRXUQDO Te (GXFDWLRQDO5HVHDURKn +DUQLVFKr / f ,WHP UHVSRQVH SDWWHUQV $SSOLFDWLRQ IRU HGXFDWLRQDO SUDFWLFH -RXUQDO RI (GXFDWLRQDO 0HDVXUHPHQWr r +LOWRQr 7 / r t %HUJOXQGr : f 6H[ GLIIHUHQFHV LQ PDWKHPDWLFV DFKLHYHPHQW $ ORQJLWXGLQDO VWXG\ -RXUQDO RI (GXFDWLRQDO 5HVHDUFKr er +XQWHUr ( r t &RKHQr 6 + f &RUUHFWLQJ IRU XQUHOLDELOLW\ LQ QRQOLQHDU PRGHOV RI DWWLWXGH FKDQJH 3VYFKRPHWULNDr r -DUYLVr 2 7 f %R\JLUO DELOLW\ GLIIHUHQFHV LQ HOHPHQWDU\ VFKRRO DULWKPHQWLF 6FKRRO 6FLHQFH DQG 0WNHPDWLBFBV Â .DXIPDQr $ f 6H[ GLIIHUHQFHV LQ PDWKHPDWLFV UHDVRQLQJ LQ PDWKHPDWLFDOO\ JLIWHG DQG DYHUDJH ILIWK DQG VL[WK JUDGHUV 'RFWRUDO GLVVHUWDWLRQr +RIVWUD 8QLYHUVLW\r f 'LVVHUWDWLRQ $EVWUDFWV ,QWHU -MDIFLRQIO-8 n $$ .LOSDWULFN f 3UREOHP VROYLQJ DQG FUHDWLYH EHKDYLRU LQ PDWKHPDWLFV ,Q : :LOVRQr t / 5 &DUU\ (GVfr 6WXGLHV LQ PDWKHPDWLFV 9RO 1R f 6WDQIRUGr &$ 6FKRRO 0DWKHPDWLFV 6WXG\ *URXS .ORRVWHUPDQr 3 : f $WWULEXWLRQDO WKHRU\r OHDUQHG KHOSOHVVQHVVr DQG DFKLHYHPHQW LQ QLQWK JUDGH PDWKHPDWLFV 'RFWRUDO GLVVHUWDWLRQr 8QLYHUVLW\ RI :LVFRQVLQr f 'LVVHUWDWLRQ $EVWUDFWV ,QWHUQDWLR QDO e $$ .QLIRQJr 'r t +ROWDQr % f $Q DQDO\VLV RI FKLOGUHQnV ZULWWHQ VROXWLRQV WR ZRUG SUREOHPV -RXUQDO IRU 5HVHDUFK LQ 0DWKHPDWLFV (GXFDWLRQr r .QLIRQJr 'r t +ROWDQr % f $ VHDUFK IRU UHDGLQJ GLIILFXOWLHV DPRQJ HUUHG ZRUG SUREOHPV -RXUQDO IRU HVeDUFKB LQ 0DWKHPDWLFV (GXFDWLRQr r PAGE 122 /DQGDXr 0 f 7KH HIIHFWV RI VSDWLDO DELOLW\ DQG SUREOHP SUHVHQWDWLRQ IRUPDW LQ PDWKHPDWLFDO SUREOHP VROYLQJ SHUIRUPDQFH RQ PLGGOH VFKRRO VWXGHQWV 'RFWRUDO GLVVHUWDWLRQr 1RUWKZHVWHUQ 8QLYHUVLW\f 'LVVHUWDWLRQ $EVWUDFWV PWH(3WLRQDOr Â£Âr D $ /LQGJUHQr + & r 6LOYDr ,r )DUDFRr ,r t 'D5RFKDr 1 6 f $WWLWXGHV WRZDUG SUREOHP VROYLQJ DV D IXQFWLRQ RI VXFFHVV LQ DULWKPHWLF LQ %UD]LOLDQ HOHPHQWDU\ VFKRROV -RXUQDO RI (GXFDWLRQDO 5HVHDUFKr tr 0DFFRE\r ( ( f 6H[ GLIIHUHQFHV LQ LQWHOOHFWXDO IXQFWLRQLQJ ,Q ( ( 0DFFRE\ (Gfr 7KH 'HYHORSPHQW RI 6H[ 'LIIHUHQFHV SS f 6WDQIRUGr &$ 7KH 6WDQIRUG 8QLYHUVLW\ 3UHVV 0DFFRE\r ( ( r t -DFNOLQr & 1 f 6H[ GLIIHUHQFHV LQ LQWHOOHFWXDO IXQFWLRQLQJ ,Q $ $QDVWDVLr (Gfr $VVHHVWQHQWLQ D SOXUDOLVWLF VRSLHO[r SS f 3URFHHGLQJV RI WKH ,QYLWDWLRQDO &RQIHUHQFH RQ 7HVWLQJ 3UREOHPVr (GXFDWLRQDO 7HVWLQJ 6HUYLFHr 3ULQFHWRQr 10DFFRE\r ( (r t -DFNOLQr & 1 f 7KH 3V\FKRORJ\ RI 6H[ 'LIIHUHQFHVr 6WDQIRUGr &$ 6WDQIRUG 8QLYHUVLW\ 3UHVV 0DUVKDOOr 6 3 f 6H[ 'LIIHUHQFHV LQ 6L[WK *UDGH &KLOGUHQnV 3UREOHP 6ROYLQJ 3DSHU SUHVHQWHG DW WKH DQQXDO PHHWLQJ RI WKH $PHULFDQ (GXFDWLRQDO 5HVHDUFK $VVRFLDWLRQr /RV $QJHOHVr &$ 0DUVKDOOr 6 3 f 6H[ GLIIHUHQFHV LQ PDWKHPDWLFDO HUURUV $QDO\VLV RI GLVWUDFWRU FKRLFH -RXUQDO IRU 5HVHDUFK LQ 0DWKHPDWLFV (GXFDWLRQr r 0DUVKDOOr 6 r3 f 6H[ GLIIHUHQFHV LQ FKLOGUHQnV PDWKHPDWLFV DFKLHYHPHQW 6ROYLQJ FRPSXWDWLRQV DQG VWRU\ SUREOHPVr -RXUQDO 4I (GXFDWL43AOB3J\FK4O3J\r er 0DUWLQr 0 f 5HDGLQJ FRPSUHKHQVLRQr DEVWUDFW YHUEDO UHDVRQLQJ DQG FRPSXWDWLRQ DV IDFWRUV LQ DULWKPHWLF SUREOHP VROYLQJ 'RFWRUDO GLVVHUWDWLRQr 8QLYHUVLW\ RI ,RZDr f SLVVHUWDWL3QB$EVWUDFW ,QWHUQDWLRQDOr r $$ PAGE 123 0H\HU 5 $ f eJ[U-JODWHGBGLIIJ-JLFHVMSBPDWKHPD WLFDOBSUREOHPB VROYLQJ SHUIRUPDQFH DQG LQWHOOHFWXDO DELOLWLHV 0DGLVRQ 8QLYHUVLW\ RI :LVFRQVLQ 5HVHDUFK DQG 'HYHORSPHQW &HQWHU IRU &RJQLWLYH /HDUQLQJ (5,& 'RFXPHQW 5HSURGXFWLRQ 6HUYLFH 1R (' f 0LOWRQ $ f 7KH HIIHFWV RI VH[UROH LGHQWLILFD FDWLRQ XSRQ SUREOHP VROYLQJ VNLOO -RXUQDO RI $EQRUPDO t 6RFLDO3V\FKRORJ\ 0ROOHU 1 f 7KH LPSDFW RI JHQGHU PDVFXOLQLW\ DQG IHPLQLW\ RQ PDWK DFKLHYHPHQW DQG FRXUVH GHFLVLRQV 'RFWRUDO GLVVHUWLRQ 3XUGXH 8QLYHUVLW\ f +LVV HMW DWLRQ$EVWUDFWHB ,QWHUQDWLRQDO Â£O $ 0XWK f 6ROYLQJ DULWKPHWLF ZRUG SUREOHPV 5ROH RI UHDGLQJ DQG FRPSXWDWLRQDO VNLOOV -RXUQDO RI (GXFDWLRQDO 3V\FKRORJ\ 3DOODV $ 0 t $OH[DQGHU / f 6H[ GLIIHUHQFHV LQ TXDQWLWDWLYH 6$7 SHUIRUPDQFH 1HZ HYLGHQFH RQ WKH GLIIHUHQWLDO FRXUVHZRUN K\SRWKHVLV $PHULFDQ ,GMLADWMASD-5HVHDBUFOL-RXUQDOr 3ODQN ( t 3ODQN 5 f (PRWLRQDO FRPSRQHQWV LQ DULWKPHWLFDO OHDUQLQJ DV VHHQ WKURXJK DXWRELRJUDSKLHV 3V\FKRORJLFDO 6WXGLHV RI WKH &KLOG e 3RUWHU $ & f 7KH HIIHFWV RI XVLQJ IDOOLEOH YDULDEOHV LQ WKH DQDO\VLV RI FRYDULDQFH 'RFWRUDO GLVVHUWDWLRQ 8QLYHUVLW\ RI :LVFRQVLQ f 'LVVHUWDWLILKADEJWMDFW,QWHU3DWLSQDL % 3UREHUW % 6 f 0DWK FRQILGHQFH ZRUNVKRSV $ PXOWLPRGDO JURXS LQWHUYHQWLRQ VWUDWHJ\ LQ PDWKHPDWLFV DQ[LHW\DYRLGDQFH 'RFWRUDO GLVVHUWDWLRQ 8QLYHUVLW\ RI )ORULGD f QLVeHUWDWL4SB$EJWUDFWV ,QWHUQDWLRQDO % 3XHUWR 5LFR 'HSDUWPHQW RI (GXFDWLRQ f %DVLF VNLOOVWHVW LQ PDWKHPDWLFV +DWR 5H\ &HQWHU RI (YDOXDWLRQ 5DGDW] + f (UURU DQDO\VLV LQ PDWKHPDWLFV HGXFDn WLRQ M4XPDL IRUMWHQHDUFK LQ 0DMLHLQDWOBFVBIFMXFD WLRQ OIL 5LHGHVHO $ & f 3UREOHP VROYLQJ 6RPH VXJJHVWLRQV IURP UHVHDUFK $ULWKPHWLF 7HDFKHU PAGE 124 6KHUPDQ f 3UHGLFWLQJ PDWKHPDWLFV SHUIRUPDQFH LQ KLJK VFKRRO JLUOV DQG ER\V -RXUQDO RI (GXFDWLRQDO 3V\FKRORJ\ 6KHUPDQ f 0DWKHPDWLFV VSDWLDO YLVXDOL]DWLRQ DQG UHODWHG IDFWRUV &KDQJHV LQ JLUOV DQG ER\V JUDGHV -RXUQDO RI (GXFDWLRQDO 3V\FKRORJ\ 6WHLQ $ + f 7KH HIIHFWV RI VH[ UROH VWDQGDUGV IRU DFKLHYHPHQW DQG VH[ UROH SUHIHUHQFH RQ WKUHH GHWHUPLQDQWV RI DFKLHYHPHQW PRWLYLDWLRQ 'HYHORSPHQWDO 3V\FKRORJ\ r 6WHLQ $ + t 6PLWKOHVV f $JH DQG VH[ GLIIHUHQFHV LQ FKLOGUHQnV VH[ UROH VWDQGDUGV DERXW DFKLHYHPHQW 'HYHORSPHQWDO 3V\FKRORJ\ 6X\GDP 0 1 f 5HVHDUFK RQ PDWKHPDWLFV HGXFDWLRQ .f UHSRUWHG LQ -RXUQDO IRU 5HVHDUFK LQ 0DWKHPDWLFV (GXFDWLRQ = 6X\GDP 0 1 f 5HVHDUFK RQ PDWKHPDWLFV HGXFDWLRQ .f UHSRUWHG LQ -RXUQDO IRU 5HVHDUFK LQ 0DWKHPDWLFV (GXFDWLRQ 6X\GDP 0 1 f 5HVHDUFK RQ PDWKHPDWLFV HGXFDWLRQ .f UHSRUWHG LQ -"XUQDOBISU 5HVHDUFK LK 0DWKHPDWLFV (GXFDWLRQ Â£ 6X\GDP 0 1 f 5HVHDUFK RQ PDWKHPDWLFV HGXFDWLRQ .f UHFRUGHG LQ -RXUQDO IRU 5HVHDUFK LQ 0DWKHPDWLFV (GXFDWLRQ 6X\GDP 0 1 t :HDYHU ) f 5HVHDUFK RQ PDWKHPDWLFV HGXFDWLRQ .f UHFRUGHG LQ -RXUQDO IRU 5HVHDUFK LQ 0DWKHPDWLFV (GXFDWLRQ e 9HUEHNH $ f 6H[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFDOO\ JLIWHG VHFRQGDU\ VWXGHQWV $Q LQYHVWLJDWLRQ RI VHOHFWHG FRJQLWLYH DIIHFWLYH DQG HGXFDWLRQDO IDFWRUV 'RFWRUDO GLVVHUWDWLRQ 8QLYHUVLW\ RI 0DU\ODQG f 'LVVHUWDWLRQ $EVWUDFWV ,QWHUQDWLRQDO $$ :HLQHU % f $FKLHYHPHQW PRWLYDWLRQ DQG DWWULEXWLRQ WKHRU\ 0RUULVWRZQ 1 *HQHUDO /HDUQLQJ 3UHVV PAGE 125 :KLWDNHU 5 f $ VWXG\ RI WKH UHODWLRQVKLS EHWZHHQ VHOHFWHG QRQFRJQLWLYH IDFWRUV DQG WKH SUREOHP VROYLQJ SHUIRUPDQFH RI IRXUWKJUDGH FKLOGUHQ 7HFK 5HS 1R f 0DGLVRQ 8QLYHUVLW\ RI :LVFRQVLQ 5HVHDUFK DQG 'HYHORSPHQW &HQWHU IRU &RJQLWLYH /HDUQLQJ :LVH / 6WHHO / t 0DF'RQDOG & f 2ULJLQV DQG FDUHHU FRQVHJXHQFHV RI VH[ GLIIHUHQFHV LQ ELJKBBVBFKRRO PDWKHPDWLFV DFKLHYHPHQW *UDQW 1R 1(,* f 3DOR $OWR &$ $PHULFDQ ,QVWLWXWHV IRU 5HVHDUFK =DOHZVNL / f $Q H[SORUDWRU\ VWXG\ WR FRPSDUH WZR SHUIRUPDQFH PHDVXUHV $Q LQWHUYLHZFRGLQJ VFKHPH RI PDWKHPDWLFDO SUREOHP VROYLQJ DQG D ZULWWHQ WHVW 'RFWRUDO 'LVVHUWDWLRQ 8QLYHUVLW\ RU :LVFRQVLQ 0DGLVRQf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} WR -XO\r } VKH ZRUNHG DV JUDGXDWH UHVHDUFK DVVLVWDQW DW WKH )RXQGDWLRQV RI (GXFDWLRQ 'HSDUWPHQW} 8QLYHUVLW\ RI )ORULGD 6KH UHWXUQHG WR 3XHUWR 5LFR LQ $XJXVW} } WR VHUYH DV 6SHFLDO $LGH WR WKH $VVLVWDQW 6HFUHWDU\ IRU WKH 9RFDWLRQDO} 7HFKQLFDO}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Â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Â]]Â£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r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n PDQFH $WWLWXGHV 7RZDUG 3UREOHP 6ROYLQJ DQG 3UREOHP 6ROYLQJ 3HUIRUPDQFH LLL PAGE 133 ,OO0(7+2' 7KH 6DPSOH 7KH ,QVWUXPHQW $QDO\VLV RI WKH 'DWD $QDO\VLV RI 6H[ E\ 2SWLRQ E\ PAGE 134 /,67 2) 7$%/(6 7DEOH 3DJH +\SRWKHWLFDO 3UREDELOLWLHV RI 2SWLRQ &KRLFH &RQGLWLRQDO RQ PAGE 135 7DEOH 3DJH $GMXVWHG 0HDQV RQ 3UREOHP 6ROYLQJr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Â£VLFDV HQ 0DWHPÂ£WLFDVf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f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f DQG VXEWUDFWLRQ LQ RQH \HDUf ZHUH WKH FRQWUROOLQJ YDULDEOHV 7KH TXHVWLRQ RI ZKHWKHU PDOHIHPDOH GLIIHUHQFHV LQ SUREOHP VROYLQJ GHSHQG RQ FRPSXWDWLRQDO VNLOOV ZDV DQVZHUHG SDUWLDOO\ LQ WKH DIILUPDWLYH [ PAGE 140 &+$37(5 ,1752'8&7,21 6H[ GLIIHUHQFHV LQ PDWKHPDWLFV OHDUQLQJ DQG DFKLHYHPHQW KDYH EHHQ WKH VXEMHFW RI LQWHQVLYH UHVHDUFK 5HVHDUFK GRQH EHIRUH KDV VKRZQ WKDW PDOH SHUIRUPDQFH RQ PDWKHPDWLFDO DFKLHYHPHQW WHVWV LV VXSHULRU WR IHPDOH SHUIRUPDQFH E\ WKH WLPH WKH\ UHDFK XSSHU HOHPHQWDU\ RU MXQLRU KLJK VFKRRO )HQQHPD S f 7KH OLWHUDWXUH VWURQJO\ VXJJHVWV WKDW DW WKH HOHPHQWDU\ OHYHO IHPDOHV RXWSHUIRUP PDOHV LQ FRPSXWDWLRQ DQG PDOHV H[FHO LQ PDWKHPDWLFDO UHDVRQLQJ *OHQQRQ t &DOODKDQ -DUYLV 0DFFRE\ f 6LQFH UHVHDUFK ILQGLQJV KDYH EHHQ OHVV FRQVLVWHQW )HQQHPD f DIWHU UHYLHZLQJ VWXGLHV IRXQG WKDW GXULQJ VHFRQGDU\ VFKRRO RU HDUOLHU VH[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFV DFKLHYHPHQW DUH QRW VR HYLGHQW EXW WKDW ZKHQ GLIIHUHQFHV DUH IRXQG WKH\ IDYRU PDOHV LQ KLJK OHYHO FRJQLWLYH WDVNV SUREOHP VROYLQJf DQG IHPDOHV LQ ORZ OHYHO FRJQLWLYH WDVNV FRPSXWDWLRQf $V D UHVXOW RI D IXUWKHU UHYLHZ RI WKH OLWHUDWXUH )HQQHPD f FRQFOXGHG WKDW DW WKH HOHPHQWDU\ OHYHO VH[UHODWHG GLIIHUHQFHV GR QRW H[LVW DW DOO FRJQLWLYH OHYHOV IURP FRPSXWDWLRQ WR SUREOHP VROYLQJ PAGE 141 0DQ\ YDULDEOHVr FRJQLWLYHr DIIHFWLYHr DQG HGXFDWLRQDOr KDYH EHHQ LQYHVWLJDWHG VLQFH LQ UHODWLRQ WR VH[ GLIIHUHQFHV LQ PDWKHPDWLFV OHDUQLQJ DQG DFKLHYHPHQW )HQQHPD DQG 6KHUPDQ f LQYHVWLJDWHG WKH HIIHFW RI GLIIHUHQWLDO IRUPDO PDWKHPDWLFV HGXFDWLRQ $IWHU FRQWUROOLQJ IRU WKH QXPEHU RI \HDUV RI H[SRVXUH WR WKH VWXG\ RI PDWKHPDWLFVr WKH\ IRXQG VH[ GLIIHUHQFHV LQ RQO\ WZR RI WKH IRXU VFKRROV XQGHU VWXG\ +RZHYHUr LQ WKRVH VFKRROV ZKHUH ER\V VFRUHG KLJKHU WKDQ JLUOVr GLIIHUHQFHV ZHUH DOVR IRXQG LQ WKHLU DWWLWXGHV WRZDUG PDWKHPDWLFV +LOWRQ DQG %HUJOXQG f IRXQG VLJQLILFDQW VH[ GLIIHUHQFHV DIWHU FRQWUROOLQJ IRU WKH QXPEHU RI PDWKHPDWLFV FRXUVHV WDNHQr DQG DWWULEXWHG WKHP WR VH[ GLIIHUHQFHV LQ LQWHUHVWV $V WKH ER\Vn LQWHUHVWV LQ VFLHQFH LQFUHDVH UHODWHG WR WKH JLUOVnr WKHLU DFKLHYHPHQW LQ PDWKHPDWLFV LQFUHDVHV UHODWLYH WR WKDW RI WKH JLUOV S f :LVHr 6WHHOr DQG 0F'RQDOG f UHDQDO\]HG WHVW GDWD FROOHFWHG LQ D ORQJLWXGLQDO VWXG\ RI r KLJK VFKRRO VWXGHQWV 3URMHFW 7DOHQWf 7KH\ IRXQG WKDW ZKHQ WKH HIIHFW RI WKH QXPEHU RI KLJK VFKRRO PDWKHPDWLFV FRXUVHV ZDV QRW FRQWUROOHGr QR VH[ GLIIHUHQFHV HPHUJHG IRU WK JUDGHUVr EXW WKDW JDLQV PDGH E\ ER\V GXULQJ WKH QH[W WKUHH \HDUV ZHUH PRUH WKDQ WZLFH WKDW RI WKH JLUOV 7KHVH GLIIHUHQFHV EHWZHHQ WKH VH[HV GLVDSSHDUHG ZKHQ WKH QXPEHU RI PDWKHPDWLFV FRXUVHV WDNHQ ZDV FRQWUROOHG 5HVXOWV RI WKH :RPHQ DQG 0DWKHPDWLFV 1DWLRQDO 6XUYH\r 6XUYH\ ,r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f 7KH PDWKHPDWLFV GDWD FROOHFWHG LQ WKH VHFRQG VXUYH\ E\ WKH 1DWLRQDO $VVHVVPHQW RI (GXFDWLRQDO 3URJUHVV VKRZHG VLJQLILFDQW VH[ GLIIHUHQFHV IRU ERWK DQG \HDUROG VWXGHQWV 7KH \HDUROG IHPDOHV RXWSHUIRUPHG PDOHV LQ WKH FRPSXWDWLRQDO VXEWHVW DQG PDOHV RXWVFRUHG IHPDOHV E\ SHUFHQWDJH SRLQWV LQ SUREOHP VROYLQJ VWDWLVWLFDOO\ VLJQLILFDQWf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f &DUSHQWHU /LQGTXLVW 0DWKHZV DQG 6LOYHU f DQDO\]HG WKH UHVXOWV RI WKH 7KLUG 1DWLRQDO $VVHVVPHQW RI (GXFDWLRQDO 3URJUHVV 1$(3f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f LQYHVWLJDWHG VH[ GLIIHUHQFHV LQ PDWKHPDWLFV SHUIRUPDQFH 6KH IRXQG WKDW PDOHV DQG IHPDOHV H[FHO HDFK RWKHU LQ VROYLQJ GLIIHUHQW W\SHV RI SUREOHPV PAGE 144 )HPDOHV ZHUH EHWWHU RQ LWHPV RI FRPSXWDWLRQ DQG PDOHV ZHUH PRUH VXFFHVVIXO RQ ZRUGVWRU\ SUREOHP LWHPV SUREOHP VROYLQJf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f 0DUVKDOO LQYHVWLJDWHG ZKHWKHU ER\V DQG JLUOV PDGH VLPLODU HUURUV LQ FRPSXWDWLRQ DQG VWRU\ SUREOHPV 6KH DQDO\]HG ER\Vn DQG JLUOVn DQVZHUV WR VL[ PDWKHPDWLFV LWHPV DQG IRXQG WKDW WKH VH[HV PDGH GLIIHUHQW HUURUV SRVVLEO\ UHIOHFWLQJ GLIIHUHQW SUREOHP VROYLQJ VWUDWHJLHV +HU RULJLQDO ILQGLQJV ZHUH VXSSRUWHG ZKHQ VKH VWXGLHG WKH VDPH SUREOHP XVLQJ D ODUJH QXPEHU RI LWHPV WKUHH \HDUV ODWHU PAGE 145 3XUSRVH RI WKH 6WXG\ 7KH SXUSRVH RI WKLV VWXG\ ZDV WZRIROG 7KH ILUVW ZDV WR LQYHVWLJDWH VH[UHODWHG GLIIHUHQFHV LQ WKH VHOHFWLRQ RI LQFRUUHFW UHVSRQVHV LQ D PXOWLSOHFKRLFH DFKLHYHPHQW WHVW LQ PDWKHPDWLFV )RU HDFK WHVW LWHPr WKH UHVHDUFK TXHVWLRQV ZHUH DV IROORZV ,V WKHUH D GLIIHUHQFH LQ WKH SURSRUWLRQ RI PDOHV DQG IHPDOHV FKRRVLQJ HDFK LQFRUUHFW RSWLRQ" ,V WKH VDPH SDWWHUQ RI GLIIHUHQFHV IRXQG LQ GDWD REWDLQHG LQ WKUHH GLIIHUHQW DGPLQLVWUDWLRQV RI WKH WHVW" 7KH VHFRQG REMHFWLYH ZDV WR LQYHVWLJDWH VH[UHODWHG GLIIHUHQFHV LQ WHVW VFRUHV LQ PDWKHPDWLFV SUREOHP VROYLQJ 7KH IROORZLQJ TXHVWLRQV ZHUH VWXGLHG 'R PDOHV DQG IHPDOHV GLIIHU LQ SUREOHP VROYLQJ SHUI RUPDQFH" 'R VH[ GLIIHUHQFHV LQ SUREOHP VROYLQJ SHUVLVW DIWHU FRQWUROOLQJ IRU FRPSXWDWLRQDO VNLOOVI DQG GRHV WKH GLIIHUHQWLDO VXFFHVV RI PDOHV DQG IHPDOHV RQ SUREOHP VROYLQJ LWHPV GHSHQG RQ WKHLU VXFFHVV RQ WKH FRPSXWDWLRQ LWHPV" 6LJQLILFDQFH RI WKH 6WXG\ ,WHP UHVSRQVH SDWWHUQV DUH YHU\ XVHIXO WHFKQLTXHV LQ WKH DVVHVVPHQW RI PDWKHPDWLFV OHDUQLQJ DQG DFKLHYHPHQW 7RWDO WHVW VFRUHV FDQ EH YHU\ PLVOHDGLQJ LQ WKH DVVHVVPHQW RI VWXGHQW SHUIRUPDQFH DQG SURYLGH QR GLDJQRVWLF LQIRUPDWLRQ DERXW WKH QDWXUH DQG VHULRXVQHVV RI VWXGHQW PAGE 146 HUURUV +DUQLVFK" f ,WHP UHVSRQVH SDWWHUQV DUH YDOXDEOH IRU WKH LGHQWLILFDWLRQ RI ODUJH JURXS GLIIHUHQFHV LQFOXGLQJ G L VW ULFWWRGLVWULFW VFKRROWRVFKRRO DQG FODVVURRPWRFODVVURRP YDULDWLRQV RQ GLIIHUHQW VXEVHWV RI LWHPV 7KH UHVSRQVH SDWWHUQV FDQ SURYLGH GLDJQRVWLF LQIRUPDWLRQ DERXW WKH W\SH RI XQGHUVWDQGLQJ WKH VWXGHQW KDV RQ YDULRXV PDWKHPDWLFV WRSLFV HJ SUREOHP VROYLQJf 0DUVKDOO f KDV XVHG WKH LWHP UHVSRQVH SDWWHUQ WHFKQLTXH DQG KHU ILQGLQJV LQGLFDWH VH[ GLIIHUHQFHV LQ PDWKHPDWLFV SHUIRUPDQFH DW WKH LWHP OHYHO )HPDOHV RXWSHUIRUPHG PDOHV LQ FRPSXWDWLRQ DQG PDOHV RXWVFRUHG IHPDOHV LQ SUREOHP VROYLQJ $OVR WKH VXFFHVV RI JLUOV LQ WKH SUREOHP VROYLQJ LWHPV ZDV GHSHQGHQW XSRQ WKHLU VXFFHVV LQ WKH FRPSXWDWLRQ LWHPV IRU ER\V VXFFHVV LQ WKH SUREOHP VROYLQJ LWHPV GLG QRW GHSHQG DV PXFK RQ WKHLU FRPSXWDWLRQDO SHUIRUPDQFH 0DUVKDOO f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r UHVHDUFK LV QHHGHG 2UJDQL]DWLRQ RI WKH Â‘6WXG\ $ UHYLHZ RI WKH OLWHUDWXUH RQ VH[ GLIIHUHQFHV LQ PDWKHPDWLFV SHUIRUPDQFH LV UHSRUWHG LQ &KDSWHU ,, 7KH UHVHDUFK PHWKRGRORJ\ LV SUHVHQWHG LQ &KDSWHU ,,, 5HVHDUFK TXHVWLRQV} VDPSOH} LQVWUXPHQW}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f LQ WKH XSSHU HOHPHQWDU\ \HDUV DQG LQ WKH HDUO\ KLJK VFKRRO \HDUV )HQQHPD 0DFFRE\ t -DFNOLQ f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} DQG FRQWDLQV D GLVFXVVLRQ RI WKH PRUH SUHYDOHQW LVVXHV DERXW WKH LQIOXHQFH RI FRJQLWLYH DQG DIIHFWLYH YDULDEOHV RQ VH[ GLIIHUHQFHV LQ PDWKHPDWLFV OHDUQLQJ DQG DFKLHYHPHQW 7KH IRXUWK LV D VXPPDU\ RI WKH UHVHDUFK GHDOLQJ ZLWK YDULDEOHV FRQVLGHUHG DV LQIOXHQWLDO WR PDWKHPDWLFV SUREOHP VROYLQJ SHUIRUPDQFH 6H[UHODWHG 'LIIHUHQFHV LQ ,QFRUUHFW 5HVSRQVH 3DWWHUQV 5HVHDUFK ILQGLQJV WHQG WR VXJJHVW WKDW ER\V DQG JLUOV PD\ EH DSSURDFKLQJ SUREOHP VROYLQJ GLIIHUHQWO\ HJ} )HQQHPD DQG 6KHUPDQ} 0DUVKDOO} } 0H\HU} } DPRQJ RWKHUVf 0DUVKDOO f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f ERWK VH[HV UHIOHFWHG VLPLODU PLVWDNHV LQ FDUU\LQJ! EXW LQ GLIIHUHQW FROXPQV ,Q WKH VHFRQG FRPSXWDWLRQ LWHP! ERWK VH[HV LJQRUHG WKH GHFLPDO SRLQWV DQG VHOHFWHG WKH VDPH LQFRUUHFW UHVSRQVH +RZHYHU! PRUH JLUOV WKDQ ER\V FKRVH WKLV UHVSRQVH ,Q WKH ILUVW FRPSXWDWLRQ LWHP )RUP RI WKH WHVWf WKH LQFRUUHFW FKRLFH RI ERWK VH[HV ZDV RSWLRQ e! EXW WKH VHFRQG PRVW IUHTXHQWO\ VHOHFWHG RSWLRQ ZDV Â IRU ER\V DQG E IRU JLUOV ,Q WKH VHFRQG FRPSXWDWLRQ LWHP RI WKLV PAGE 151 WHVW IRUP QR VH[ GLIIHUHQFHV ZHUH IRXQG LQ UHVSRQVH SDWWHUQV $SSUR[LPDWHO\ b RI HDFK VH[ VHOHFWHG RSWLRQ F 7KH QH[W SRSXODU FKRLFH IRU ERWK VH[HV ZDV RSWLRQ G VHOHFWHG E\ DSSUR[LPDWHO\ b RI ERWK ER\V DQG JLUOV 2Q WKH VWRU\ SUREOHP RI )RUP ,I PDOHV DQG IHPDOHV UHVSRQGHG DOLNH 7KHLU PRVW SRSXODU LQFRUUHFW UHVSRQVH FKRLFH ZDV RSWLRQ D IRU ERWK PDOHV DQG IHPDOHV 7KH VHFRQG PRVW SRSXODU LQFRUUHFW FKRLFH ZDV RSWLRQ F IRU ERWK VH[HV 5HVSRQVH WR WKH VWRU\ SUREOHP LQ )RUP VKRZHG VH[ GLIIHUHQFHV LQ UHVSRQVH FKRLFH ,QFOXGLQJ WKH FRUUHFW RSWLRQ b RI WKH JLUOV VHOHFWHG RSWLRQ D b FKRVH RSWLRQ & DQG b RSWLRQ G )RU PDOHV DSSUR[LPDWHO\ b VHOHFWHG RSWLRQ D DQG WKH VDPH SHUFHQW VHOHFWHG RSWLRQ G 0DUVKDOO FRQFOXGHG WKDW DOWKRXJK WKH DQDO\VLV RI LQFRUUHFW UHVSRQVHV GRHV QRW H[SODLQ ZK\ ER\V DQG JLUOV GLIIHU LQ WKHLU UHVSRQVHV WKH DQDO\VLV VKRZV WKDW ER\V DQG JLUOV DSSURDFK SUREOHPV LQ GLIIHUHQW ZD\V DQG WKHVH YDU\LQJ VWUDWHJLHV FDQ EH XVHIXO LQ LQGHQWLI\LQJ KRZ WKH VH[HV GLIIHU LQ UHDVRQLQJ DELOLWLHV 7ZR \HDUV ODWHU 0DUVKDOO f DQDO\]HG WKH UHVSRQVHV RI DSSUR[LPDWHO\ ER\V DQG JLUOV WR PDWKHPDWLFV LWHPV FRQWDLQHG LQ WKH WHVW IRUPV RI WKH 6XUYH\ RI %DVLF 6NLOOV GXULQJ WKH \HDUV DQG 6KH XVHG ORJOLQHDU PRGHOV H[SODLQHG LQ &KDSWHU PAGE 152 ,,,f WR LQYHVWLJDWH VH[ UHODWHG GLIIHUHQFHV LQ WKH VHOHFWLRQ RI LQFRUUHFW UHVSRQVHV DQG WKH FRQVLVWHQF\ RI VXFK GLIIHUHQFHV RYHU WKUHH \HDUV RI DGPLQLVWUDWLRQ RI WKH WHVW %DVHG RQ KHU ILQGLQJV WKDW VH[ GLIIHUHQFHV ZHUH IRXQG LQ b RI WKH LWHPV 0DUVKDOO FODVVLILHG WKH VWXGHQWVn HUURUV DFFRUGLQJ WR 5DGDW]n f ILYHFDWHJRU\ HUURU FODVVLILFDWLRQ 7KH FDWHJRULHV DUH ODQJXDJH HUURUV LQ VHPDQWLFVf VSDWLDO YLVXDOL]DWLRQ PDVWHU\ DVVRFLDWLRQ DQG XVH RI LUUHOHYDQW UXOHV ,W ZDV IRXQG WKDW JLUOVn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r DV HDUO\ DV r .LOSDWULFN FULWLFL]HG WKH IDFW WKDW WKH VWXG\ RI SUREOHP VROYLQJ KDV QRW EHHQ V\VWHPDWLF VRPH UHVHDUFKHUV KDYH VWXGLHG WKH FKDUDFWHULVWLFV RI WKH SUREOHP ZKLOH RWKHUV KDYH JLYHQ WKHLU DWWHQWLRQ WR WKH FKDUDFWHULVWLFV RI WKH SUREOHP VROYHUV 0RUHRYHUr GLIIHUHQFHV LQ WKH WHVWV XVHG WR PHDVXUH SUREOHP VROYLQJ SHUIRUPDQFH DOVR FRQVWLWXWH DQ REVWDFOH ZKHQ WU\LQJ WR FRPSDUH WKH UHVXOWV RI WKH VWXGLHV FDUULHG RXW ,Q RUGHU WR DYRLG WKLV SLWIDOO DQG SURYLGH D EDVLV IRU FRPSDULVRQr WKH VWXGLHV UHYLHZHG LQ WKLV VHFWLRQr GHDOLQJ ZLWK VH[UHODWHG GLIIHUHQFHV LQ SUREOHP VROYLQJr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r r SS r f PAGE 154 7KH WHVW LV FRPSRVHG RI LWHPV GHVLJQHG WR \LHOG VFRUHV D FRPSUHKHQVLRQ VFRUHU DQ DSSOLFDWLRQ VFRUHU DQG D SUREOHP VROYLQJ VFRUH 7KH FRPSUHKHQVLRQ TXHVWLRQ DVFHUWDLQV ZKHWKHU D FKLOG XQGHUVWDQGV WKH LQIRUPDWLRQ JLYHQ LQ WKH LWHP VWHP 7KH DSSOLFDWLRQ TXHVWLRQ DVVHVVHV WKH FKLOGnV PDVWHU\ RI D SUHUHTXLVLWH FRQFHSW RU VNLOO RI WKH SUREOHP VROYLQJ TXHVWLRQ 7KH SUREOHP VROYLQJ TXHVWLRQ SRVHV D TXHVWLRQ ZKRVH VROXWLRQ LV QRW LPPHGLDWHO\ DYDLODEOHU WKDW LVU D VLWXDWLRQ ZKLFK GRHV QRW OHQG LWVHOI WR WKH LPPHGLDWH DSSOLFDWLRQ RI D UXOH RU DOJRULWKP 7KH DSSOLFDWLRQ DQG SUREOHP VROYLQJ SDUWV RI WKH WHVW PD\ UHIHU WR D FRPPRQ XQLW RI LQIRUPDWLRQ WKH LWHP VWHPf EXW WKH TXHVWLRQV DUH LQGHSHQGHQW LQ WKDW WKH UHVSRQVH WR WKH DSSOLFDWLRQ TXHVWLRQ LV QRW XVHG WR UHVSRQG WR WKH SUREOHP VROYLQJ TXHVWLRQ 0H\HU f :KLWDNHU f )HQQHPD DQG 6KHUPDQ f DQG 6KHUPDQ f KDYH XVHG WKH 5RPEHUJ:HDUQH WHVW LQ WKHLU VWXGLHV 0H\HU f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} 6SDWLDO 5HODWLRQV DQG 3LFWXUH *URXS 1DPH6HOHFWLRQ $ IDFWRU DQDO\VLV} KRZHYHU} VKRZHG GLIIHUHQFHV LQ WKH QXPEHU DQG FRPSRVLWLRQ RI WKH IDFWRUV )RU IHPDOHV D JHQHUDO PDWKHPDWLFV IDFWRU ZDV GHWHUPLQHG E\ PDWKHPDWLFV FRPSXWDWLRQ FRPSUHKHQVLRQ DSSOLFDWLRQ DQG SUREOHP VROYLQJ )RU PDOHV WKH FRPSUHKHQVLRQ DQG DSSOLFDWLRQ SDUWV GHWHUPLQHG RQH IDFWRU SUREOHP VROYLQJ ZLWK WZR RWKHU UHIHUHQFH WHVWV *HVWDOW DQG 2PHOHWf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f LQYHVWLJDWHG WKH UHODWLRQVKLS EHWZHHQ WKH PDWKHPDWLFDO SUREOHP SHUIRUPDQFH RI WK JUDGH FKLOGUHQ DQG WKHLU DWWLWXGHV WRZDUG SUREOHP VROYLQJ WKHLU WHDFKHUVn PAGE 156 DWWLWXGH WRZDUG PDWKHPDWLFDO SUREOHP VROYLQJ DQG UHODWHG VH[ DQG SURJUDPW\SH GLIIHUHQFHV $OWKRXJK KLV PDLQ REMHFWLYH ZDV WR FRQVWUXFW DQ DWWLWXGH VFDOH WR PHDVXUH DWWLWXGHV WRZDUG SUREOHP VROYLQJ KLV VWXG\ LV LPSRUWDQW EHFDXVH KLV ILQGLQJV VXSSRUW 0H\HUnV UHJDUGLQJ WKH ODFN RI VLJQLILFDQW VH[UHODWHG GLIIHUHQFHV LQ SUREOHP VROYLQJ SHUIRUPDQFH 3HUIRUPDQFH LQ WKH SUREOHP VROYLQJ TXHVWLRQV IRU ERWK PDOHV DQG IHPDOHV ZDV PXFK ORZHU WKDQ SHUIRUPDQFH LQ WKH DSSOLFDWLRQ TXHVWLRQV DQG PXFK ORZHU WKDQ SHUIRUPDQFH LQ WKH FRPSUHKHQVLRQ TXHVWLRQV ,Q IDFW WKH PHDQ VFRUH IRU HDFK SDUW RI WKH LWHP IRU ERWK PDOHV DQG IHPDOHV ZDV DOPRVW LGHQWLFDO WR WKH PHDQ VFRUHV REWDLQHG E\ PDOHV DQG IHPDOHV LQ 0H\HUnV VWXG\ :KLWDNHU QRWHG WKDW HDFK DSSOLFDWLRQ LWHP LV PRUH GLIILFXOW WKDQ LWV SUHFHGLQJ FRPSUHKHQVLRQ LWHP DQG WKDW HDFK SUREOHP VROYLQJ LWHP LV PRUH GLIILFXOW WKDQ LWV SUHFHGLQJ DSSOLFDWLRQ LWHP 1R VLJQLILFDQW VH[UHODWHG GLIIHUHQFHV ZHUH IRXQG IRU DQ\ RI WKH WKUHH SDUWV RI WKH LWHP FRPSUHKHQVLRQ DSSOLFDWLRQ RU SUREOHP VROYLQJf )HQQHPD DQG 6KHUPDQ f LQYHVWLJDWHG VH[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFV DFKLHYHPHQW DQG FRJQLWLYH DQG DIIHFWLYH YDULDEOHV UHODWHG WR VXFK GLIIHUHQFHV 7KH\ DGPLQLVWHUHG WKH 5RPEHUJ:HDUQH 3UREOHP 6ROYLQJ 7HVW WR D UHSUHVHQWDWLYH VDPSOH RI VWXGHQWV JUDGHV f IURP 0DGLVRQ :LVFRQVLQ SUHGRPLQDQWO\ PLGGOHFODVV EXW LQFOXGLQJ JUHDW GLYHUVLW\ LQ 6(6 7KH VDPSOH FRQVLVWHG RI PAGE 157 VWXGHQWV ZKR KDG WDNHQ D VLPLODU QXPEHU RI PDWKHPDWLFV FRXUVHV DQG ZHUH LQ WKH WRS b RI WKH FODVV LQ PDWKHPDWLFV DFKLHYHPHQW 7KH\ ZHUH WHVWHG LQ )RXU KLJK VFKRRO GLVWULFWV ZHUH LQFOXGHG ,Q RQO\ RQH RI WKH KLJK VFKRRO GLVWULFWV ZHUH VH[UHODWHG GLIIHUHQFHV LQ DSSOLFDWLRQ DQG SUREOHP VROYLQJ IRXQG LQ IDYRU RI PDOHV 7KH\ FRQFOXGHG WKDW ZKHQ UHOHYDQW IDFWRUV DUH FRQWUROOHG VH[UHODWHG GLIIHUHQFHV LQ IDYRU RI PDOHV GR QRW DSSHDU RIWHQ DQG ZKHQ WKH\ GR WKH\ DUH QRW ODUJH 6KHUPDQ f LQYHVWLJDWHG WKH FDXVHV RI WKH HPHUJLQJ VH[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFV SHUIRUPDQFH LQ IDYRU RI PDOHV GXULQJ DGROHVFHQFH JUDGHV f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} DOO RI ZKLFK XVHG WKH 5RPEHUJ:HDUQH 0DWKHPDWLFV 3UREOHP 6ROYLQJ 7HVW} VKRZ VWDWLVWLFDOO\ VLJQLILFDQW VH[UHODWHG GLIIHUHQFHV LQ SUREOHP VROYLQJ SHUIRUPDQFH ,Q ODWHU VWXGLHV RWKHU WHVWV ZHUH XVHG WR PHDVXUH WKLV YDULDEOH .DXIPDQ} 0DUVKDOO} } f .DXIPDQ f LQYHVWLJDWHG LI VH[ GLIIHUHQFHV LQ SUREOHP VROYLQJ} IDYRULQJ PDOHV} H[LVW LQ WKH WK DQG WK JUDGHV DQG LI WKHVH GLIIHUHQFHV ZHUH PRUH SURQRXQFHG LQ PDWKHPDWLFDOO\ JLIWHG VWXGHQWV WKDQ LQ VWXGHQWV RI DYHUDJH PDWKHPDWLFDO DELOLW\ 7KH ,RZD 7HVW RI %DVLF 6NLOOV DQG D PDWKHPDWLFV SUREOHPVROYLQJ WHVW ZHUH DGPLQLVWHUHG WR VXEMHFWV 0DOHV LQ WKH DYHUDJH JURXS DV ZHOO DV PDOHV LQ WKH JLIWHG JURXS RXWSHUIRUPHG IHPDOHV} EXW RQO\ WKH JLIWHG JURXS VKRZHG VWDWLVWLFDOO\ VLJQLILFDQW GLIIHUHQFHV PAGE 159 $V D UHVXOW RI KHU LQYHVWLJDWLRQV 0DUVKDOO f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f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r PDVWHU\ RI WKH VNLOOV UHTXLUHG E\ WKH FRPSXWDWLRQ LWHPV LV PRUH LPSRUWDQW IRU JLUOV WKDQ IRU ER\V ,I JLUOV FDQQRW VROYH WKH FRPSXWDWLRQ LWHPVr WKH\ KDYH OLWWOH FKDQFH RI VROYLQJ WKH UHODWHG VWRU\ SUREOHP LWHP )RU JLUOVr WKH SUREDELOLW\ RI VXFFHVV LQ WKH VWRU\ SUREOHP LWHP DIWHU JLYLQJ VXFFHVVIXO DQVZHUV WR ERWK FRPSXWDWLRQ LWHPV LV DOPRVW WLPHV WKH SUREDELOLW\ RI VXFFHVV DIWHU JLYLQJ LQFRUUHFW UHVSRQVHV WR ERWK FRPSXWDWLRQ LWHPV )RU ER\Vr WKH SUREDELOLW\ RI VXFFHVV LQ WKH VWRU\ SUREOHP LWHP DIWHU VXFFHVVIXO UHVSRQVHV WR WKH FRPSXWDWLRQ LWHPV LV DERXW WLPHV WKH SUREDELOLW\ RI VXFFHVV RQ WKH VWRU\ SUREOHP LWHP DIWHU LQFRUUHFW UHVSRQVHV WR WKH FRPSXWDWLRQ LWHPV 7KUHH \HDUV ODWHUr 0DUVKDOO f DQDO\]HG PRUH LQ GHSWK WKHVH SKHQRPHQD RI VH[UHODWHG GLIIHUHQFHV +HU LQWHUHVW ZDV WZRIROG )LUVWr VKH ZDQWHG WR NQRZ LI WKHUH ZHUH GLIIHUHQFHV LQ WKH UDWH RI VXFFHVV IRU ER\V DQG JLUOV LQ VROYLQJ FRPSXWDWLRQ DQG VWRU\ SUREOHP LWHPV 6HFRQGr VKH H[DPLQHG DGGLWLRQDO IDFWRUV WKDW LQWHUDFW ZLWK VH[ WR LQIOXHQFH PDWKHPDWLFV SHUIRUPDQFHr VXFK DV UHDGLQJ DFKLHYHPHQWr VRFLRHFRQRPLF VWDWXV 6(6f r SULPDU\ ODQJXDJHr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f $UPVWURQJ f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f DQG IHPDOHV DUH EHWWHU WKDQ WKHLU FRXQWHUSDUW PDOHV LQ FRPSXWDWLRQ D ORZHU OHYHO VNLOOf 0DUVKDOOn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f 7KH EDVLV IRU WKH GLIIHUHQWLDO FRXUVHZRUN K\SRWKHVLV LV WKH IDFW WKDW VH[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFDO OHDUQLQJ DQG DFKLHYHPHQW VKRZ XS ZKHQ FRPSDULQJ JURXSV ZKLFK DUH QRW HTXDO LQ SUHYLRXV PDWKHPDWLFV OHDUQLQJ $IWHU WKH WK JUDGH ER\V WHQG WR VHOHFW PDWKHPDWLFV FRXUVHV PRUH RIWHQ WKDQ JLUOV 7KHUHIRUH JLUOV VKRZ ORZHU DFKLHYHPHQW VFRUHV LQ PDWKHPDWLFV WHVWV EHFDXVH WKHLU PDWKHPDWLFV H[SHULHQFH LV QRW DV VWURQJ DV WKH ER\Vn )HQQHPD )HQQHPD t 6KHUPDQ 6KHUPDQ f )HQQHPD DQG 6KHUPDQnV VWXG\ f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f ZKR DQDO\]HG GDWD IURP 3URMHFW 7DOHQW DQG $OOHQ DQG &KDPEHUV f KDYH DOVR K\SRWKHVL]HG WKDW VH[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFV DFKLHYHPHQW PD\ EH UHODWHG WR GLIIHUHQW FXUULFXOD IROORZHG E\ PDOHV DQG IHPDOHV $OOHQ DQG &KDPEHUV DWWULEXWHG PDOH VXSHULRULW\ LQ PDWKHPDWLFV SUREOHP VROYLQJ WR GLIIHUHQFHV LQ WKH QXPEHU RI PDWKHPDWLFV FRXUVHV WDNHQ LQ KLJK VFKRRO 7KLV LVVXH KDV EHHQ VHULRXVO\ TXHVWLRQHG E\ $VWLQ f )R[ D Ef DQG %HQERZ DQG 6WDQOH\ f DPRQJ RWKHUV $VWLQ DQG )R[ KDYH UHSRUWHG ODUJH GLIIHUHQFHV LQ IDYRU RI PDOHV DPRQJ JLIWHG VWXGHQWV WDNLQJ WKH 6FKRODVWLF $FKLHYHPHQW 7HVW 7KHVH GLIIHUHQFHV RFFXU DV HDUO\ DV JUDGH ZKHQ WKHUH DUH QR VH[ GLIIHUHQFHV LQ WKH QXPEHU RI FRXUVHV WDNHQ %HQERZ DQG 6WDQOH\ f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} QRW IURP GLIIHUHQFHV LQ PDWKHPDWLFV IRUPDO HGXFDWLRQ 7KH GLIIHUHQWLDO FRXUVHZRUN K\SRWKHVLV LV QRW WRWDOO\ FRQYLQFLQJ DQG DV UHSRUWHX EHURUH LW KDV EHHQ FKDOOHQJHG E\ UHVHDUFKHUV VXFK DV %HQERZ DQG 6WDQOH\ f +RZHYHU 3DOODV DQG $OH[DQGHU f KDYH TXHVWLRQHX WKH JHQHUDOL]DRLOLW\ RI %HQERZ DQG 6WDQOH\n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f RQ VSDWLDO WDVNV DQG WKLV GLIIHUHQFH FRQWLQXHV WKURXJK WKH KLJK VFKRRO DQG FROOHJH \HDUV Sf ,Q 0DFFRE\ DQG -DFNOLQ VDLG WKDW WKH GLIIHUHQFHV LQ VSDWLDO DELOLW\ EHWZHHQ WKH VH[HV UHPDLQ PLQLPDO DQG LQFRQVLVWHQW XQWLO DSSUR[LPDWHO\ WKH DJHV RI RU ZKHQ WKH VXSHULRULW\ RI ER\V EHFRPHV FRQVLVWHQW LQ D ZLGH UDQJH RI SRSXODWLRQV DQG WHVWV S f ,Q DIWHU D FRPSUHKHQVLYH OLWHUDWXUH VHDUFK 0DFFRE\ DQG -DFNOLQ FRQFOXGHG WKDW VH[ GLIIHUHQFHV LQ VSDWLDO YLVXDOL]DWLRQ EHFRPH PRUH SURQRXQFHG EHWZHHQ XSSHU HOHPHQWDU\ \HDUV DQG WKH ODVW \HDU RI KLJK VFKRRO WKH \HDUV ZKHQ VH[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFV DFKLHYHPHQW IDYRULQJ ER\V HPHUJH *XD\ DQG 0F'DQLHO f VXSSRUWHG LQ SDUW 0DFFRE\ DQG -DFNOLQnV ILQGLQJV 7KH\ IRXQG WKDW DPRQJ HOHPHQWDU\ VFKRRO FKLOGUHQ PDOHV KDG JUHDWHU KLJK OHYHO VSDWLDO DELOLW\ WKDQ IHPDOHV EXW WKDW PDOHV DQG IHPDOHV ZHUH HTXDO LQ ORZ OHYHO VSDWLDO DELOLW\ 7KLV ILQGLQJ LV LQFRQVLVWHQW ZLWK WKDW SRUWLRQ RI 0DFFRE\ DQG -DFNOLQnV UHYLHZ WKDW VXJJHVWV WKDW VH[ GLIIHUHQFHV EHFRPH HYLGHQW RQO\ GXULQJ HDUO\ DGROHVFHQFH &RKHQ DQG :LONLH f KRZHYHU VWDWHG WKDW LQ WHVWV PHDVXULQJ GLVWLQFW VSDWLDO WDVNV PDOHV SHUIRUP EHWWHU WKDQ IHPDOHV LQ HDUO\ DGROHVFHQFH DQG WKURXJKRXW WKHLU OLIH VSDQ 0RVW VWXGLHV FDUULHG RXW DIWHU PAGE 167 KDYH IDLOHG WR VXSSRUW WKHVH VH[ GLIIHUHQFHV LQ VSDWLDO DELOLWLHV $UPVWURQJ &RQQRU 6HUELQ t 6FKDFNPDQU )HQQHPD t 6KHUPDQ 6KHUPDQ f )HQQHPD DQG 6KHUPDQ f DQG 6KHUPDQ f KDYH H[SORUHG VH[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFDO DFKLHYHPHQW DQG FRJQLWLYH DQG DIIHFWLYH YDULDEOHV UHODWHG WR WKHVH GLIIHUHQFHV ,Q D VWXG\ LQYROYLQJ VWXGHQWV IURP JUDGHV DQG IURP IRXU VFKRRO GLVWULFWV )HQQHPD DQG 6KHUPDQ IRXQG WKDW VSDWLDO YLVXDOL]DWLRQ DQG SUREOHP VROYLQJ ZHUH KLJKO\ FRUUHODWHG IRU ERWK VH[HV DQG f (YHQ LQ WKH VFKRRO GLVWULFW ZKHUH VH[ GLIIHUHQFHV ZHUH IRXQG LQ SUREOHP VROYLQJ QR VLJQLILFDQW VH[UHODWHG GLIIHUHQFHV ZHUH IRXQG LQ VSDWLDO YLVXDOL]DWLRQ :KHQ 6KHUPDQ f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f } ZLWK DQ HOHPHQWDU\ JUDGH VDPSOH} DQG )HQQHPD DQG 7DUWUHH f } ZLWK DQ LQWHUPHGLDWH OHYHO VDPSOH} IRXQG WKDW WKH LQIOXHQFH RI VSDWLDO YLVXDOL]DWLRQ RQ VROYLQJ PDWKHPDWLFV SUREOHPV LV VXEWOH} DQG WKDW PDOHV DQG IHPDOHV XVH WKHLU VSDWLDO VNLOOV GLIIHUHQWO\ LQ VROYLQJ ZRUG VWRU\ SUREOHPV SUREOHPV WKDW PHDVXUH SUREOHP VROYLQJ DELOLW\ RU UHDVRQLQJf )HQQHPD DQG 7DUWUHH f FDUULHG RXW D WKUHH\HDU ORQJLWXGLQDO VWXG\ ZKLFK VKRZHG WKDW JLUOV DQG ER\V ZLWK HTXLYDOHQW VSDWLDO YLVXDOL]DWLRQ VNLOOV GLG QRW VROYH WKH VDPH QXPEHU RI LWHPV} QRU GLG WKH\ XVH WKH VDPH SURFHVVHV LQ VROYLQJ SUREOHPV 7KH UHVXOWV DOVR VXJJHVWHG WKDW D ORZ OHYHO RI VSDWLDO YLVXDOL]DWLRQ VNLOOV ZDV D PRUH GHELOLWDWLQJ IDFWRU IRU JLUOV WKDQ IRU ER\V LQ SUREOHP VROYLQJ SHUIRUPDQFH /DQGDX f DOVR LQYHVWLJDWHG WKH UHODWLRQVKLS EHWZHHQ VSDWLDO YLVXDOL]DWLRQ DQG PDWKHPDWLFV DFKLHYHPHQW 6KH VWXGLHG WKH SHUIRUPDQFH RI PLGGOH VFKRRO FKLOGUHQ LQ PDWKHPDWLFDO SUREOHPV RI YDU\LQJ GLIILFXOW\} DQG WKH H[WHQW WR ZKLFK D GLDJUDPDWLF UHSUHVHQWDWLRQ LV OLNHO\ WR IDFLOLWDWH VROXWLRQ 6KH IRXQG WKDW VSDWLDO DELOLW\ ZDV VWURQJO\ FRUUHODWHG WR PDWKHPDWLFDO SUREOHP VROYLQJ DQG WKDW WKH HIIHFW RI VSDWLDO DELOLW\ ZDV PRUH LQIOXHQWLDO IRU IHPDOHV )HPDOHV PDGH PRUH XVH RI GLDJUDPV LQ WKH VROXWLRQ RI SUREOHPV}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f RQ PDWKHPDWLFV OHDUQLQJ DQG DFKLHYHPHQW $ EULHI VWDWHPHQW RI HDFK H[SODQDWLRQ SUHFHGHV WKH VXPPDU\ RI VWXGLHV FRQGXFWHG WKDW VXSSRUW WKH H[SODQDWLRQ )HPDOHV PRUH WKDQ PDOHV ODFN FRQILGHQFH LQ WKHLU DELOLW\ WR OHDUQ PDWKHPDWLFV DQG WKLV DIIHFWV WKHLU DFKLHYHPHQW LQ PDWKHPDWLFV DQG WKHLU HOHFWLRQ RI PRUH DGYDQFHG PDWKHPDWLFV FRXUVHV 0DFFRE\ DQG -DFNOLQ f UHSRUWHG WKDW VHOI FRQILGHQFH LQ WHUPV RI JUDGH H[SHFWDQF\ DQG VXFFHVV LQ SDUWLFXODU WDVNV ZDV IRXQG WR EH FRQVLVWHQWO\ ORZHU LQ ZRPHQ WKDQ LQ PHQ ,Q )HQQHPD DQG 6KHUPDQ UHSRUWHG PAGE 170 WKDW LQ WKHLU VWXG\ LQYROYLQJ VWXGHQWV IURP JUDGHV WKURXJK ER\V VKRZHG D KLJKHU OHYHO RI FRQILGHQFH LQ PDWKHPDWLFV DW HDFK JUDGH OHYHO 7KHVH GLIIHUHQFHV EHWZHHQ WKH VH[HV RFFXUUHG LQ PRVW LQVWDQFHV HYHQ ZKHQ QR VH[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFV DFKLHYHPHQW ZHUH IRXQG 7KH FRUUHODWLRQ EHWZHHQ FRQILGHQFH LQ PDWKHPDWLFV SHUIRUPDQFH DQG PDWKHPDWLFV DFKLHYHPHQW LQ WKLV VWXG\ ZDV KLJKHU WKDQ IRU DQ\ RWKHU DIIHFWLYH YDULDEOH LQYHVWLJDWHG 6KHUPDQ UHSRUWHG D VLPLODU ILQGLQJ LQ LQ PDOHV WKH PRVW LPSRUWDQW IDFWRU UHODWHG WR FRQWLQXDWLRQ LQ WKHRUHWLFDO PDWKHPDWLFV FRXUVHV ZDV FRQILGHQFH LQ OHDUQLQJ PDWKHPDWLFV 7KLV YDULDEOH ZHLJKHG PRUH WKDQ DQ\ RI WKH FRJQLWLYH YDULDEOHV PDWKHPDWLFV DFKLHYHPHQW VSDWLDO YLVXDOL]DWLRQ JHQHUDO DELOLW\ DQG YHUEDO VNLOO ,Q WKH FDVH RI IHPDOHV DPRQJ WKH DIIHFWLYH YDULDEOHV FRQILGHQFH LQ OHDUQLQJ PDWKHPDWLFV ZDV IRXQG WR EH VHFRQG LQ LPSRUWDQFH WR SHUFHLYHG XVHIXOQHVV RI PDWKHPDWLFV 3UREHUW f VXSSRUWHG WKHVH ILQGLQJV ZLWK FROOHJH VWXGHQWV $ YDULDEOH WKDW QHHGV GLVFXVVLRQ ZLWKLQ WKH FRQWH[W RI VH[ GLIIHUHQFHV LQ FRQILGHQFH DV OHDUQHUV RI PDWKHPDWLFV LV FDXVDO DWWULEXWLRQ &DXVDO DWWULEXWLRQ PRGHOV DWWHPSW WR FODVVLI\ WKRVH IDFWRUV WR ZKLFK RQH DWWULEXWHV VXFFHVV RU IDLOXUH 7KH PRGHO SURSRVHG E\ :HLQHU f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f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nV f VWXG\ VXSSRUWHG WKHVH ILQGLQJV $FFRUGLQJ WR .ORRVWHUPDQ DWWULEXWLRQDO YDULDEOHV DSSHDU WR EH PRUH LPSRUWDQW DFKLHYHPHQW PHGLDWRUV IRU IHPDOHV WKDQ IRU PDOHV DV PHDVXUHG E\ PDWKHPDWLFV ZRUG SUREOHPV 0RUH UHVHDUFK LV QHHGHG LQ WKLV DUHD PAGE 172 0DWKHPDWLFV DV D PDOH GRPDLQ 0DWKHPDWLFV LV DQ DFWLYLW\ PRUH FORVHO\ UHODWHG WR WKH PDOH VH[ GRPDLQ WKDQ WR WKH IHPDOH VH[ GRPDLQ (FFOHV HW DOU f 7KXVr WKH PDWKHPDWLFDO DFKLHYHPHQW RW ER\V LV KLJKHU WKDQ WKDW RI JL UOV $FFRUGLQJ WR -RKQ (UQHVW f r LQ KLV VWXG\ 0DWKHPDWLFV DQG 6H[! PDWKHPDWLFV LV D VH[LVW GLVFLSOLQH +H DWWULEXWHG VH[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFDO DFKLHYHPHQW WR WKH FUHDWLRQ E\ VRFLHW\ RI VH[XDO VWHUHRW\SHV DQG DWWLWXGHVr UHVWULFWLRQVr DQG FRQVWUDLQWV WKDW SURPRWH WKH LGHD RI WKH VXSHULRULW\ RI ER\V LQ PDWKHPDWLFV (UQHVW UHSRUWHG WKDW ER\Vr JLUOVr DQG WHDFKHUVr DOO EHOLHYH WKDW ER\V DUH VXSHULRU LQ PDWKHPDWLFVr DW OHDVW E\ WKH WLPH VWXGHQWV UHDFK DGROHVFHQFH %HUQ DQG %HUQ f DJUHH DQG DUJXH WKDW DQ $PHULFDQ ZRPDQ LV WUDLQHG WR NQRZ KHU SODFH LQ VRFLHW\ EHFDXVH RW WKH SHUYDVLYH VH[UROH FRQFHSW ZKLFK UHVXOWV LQ GLIIHUHQWLDO H[SHFWDWLRQV DQG VRFLDOL]DWLRQ SUDFWLFHV 3ODQN DQG 3ODQN f ZHUH PRUH VSHFLILF 7KH\ GLVFXVVHG WZR K\SRWKHVHV UHODWHG WR WKLV YLHZ WKH GLIIHUHQWLDO FXOWXUDO UHLQIRUFHPHQW K\SRWKHVLV DQG WKH PDVFXOLQH LGHQWLILFDWLRQ K\SRWKHVLV 7KH GLIIHUHQWLDO FXOWXUDO UHLQIRUFHPHQW K\SRWKHVLV VWDWHV WKDW VRFLHW\ LQ JHQHUDO SHUFHLYHV PDWKHPDWLFV DV D PDOH GRPDLQr JLYLQJ IHPDOHV OHVV HQFRXUDJHPHQW IRU H[FHOOLQJ LQ LW 7KH PDVFXOLQH LGHQWLILFDWLRQ K\SRWKHVLV HVWDEOLVKHV WKDW PAGE 173 DFKLHYHPHQW DQG LQWHUHVW LQ PDWKHPDWLFV UHVXOW IURP LGHQWLILFDWLRQ ZLWK WKH PDVFXOLQH UROH $ VWXG\ UHODWHG WR WKH GLIIHUHQWLDO FXOWXUDO UHLQIRUFHPHQW K\SRWKHVLV LV WKDW RI 'Z\HU f 'Z\HU H[DPLQHG WKH UHODWLRQVKLS EHWZHHQ VH[ UROH VWDQGDUGV WKH H[WHQW WR ZKLFK DQ LQGLYLGXDO FRQVLGHUV FHUWDLQ DFWLYLWLHV DSSURSULDWH WR PDOHV RU IHPDOHVf DQG DFKLHYHPHQW LQ UHDGLQJ DQG DULWKPHWLF 6WXGHQWV IURP JUDGHV ! U DQG SDUWLFLSDWHG LQ WKLV VWXG\ 6KH IRXQG WKDW VH[ UROH VWDQGDUGV FRQWULEXWHG VLJQLILFDQW YDULDQFH WR UHDGLQJ DQG DULWKPHWLF DFKLHYHPHQW WHVW VFRUHV DQG WKDW WKH HIIHFW ZDV VWURQJHU IRU PDOHV WKDQ IRU IHPDOHV 7KLV OHG WR KHU FRQFOXVLRQ WKDW VH[UHODWHG GLIIHUHQFHV LQ UHDGLQJ DQG DULWKPHWLF DUH PRUH D IXQFWLRQ RI WKH FKLOGnV SHUFHSWLRQ RI WKHVH DUHDV DV VH[DSSURSULDWH RU VH[LQDSSURSULDWH WKDQ RI WKH FKLOGnV ELRORJLFDO VH[} LQGLYLGXDO SUHIHUHQFH IRU PDVFXOLQH DQG IHPLQLQH VH[ UROHV RU OLNLQJ RU GLVOLNLQJ UHDGLQJ RU PDWKHPDWLFV ,Q D VWXG\ ZKLFK DJUHHV ZLWK WKH PDVFXOLQH LGHQWLILFDWLRQ K\SRWKHVLVr 0LOWRQ f IRXQG WKDW LQGLYLGXDOV ZKR KDG UHFHLYHG VWURQJ PDVFXOLQH RULHQWDWLRQ SHUIRUPHG EHWWHU LQ SUREOHP VROYLQJ WKDQ LQGLYLGXDOV ZKR UHFHLYHG OHVV PDVFXOLQH RULHQWDWLRQ (OWRQ DQG 5RVH f IRXQG WKDW ZRPHQ ZLWK KLJK PDWKHPDWLFDO DSWLWXGH DQG DYHUDJH YHUEDO DSWLWXGH VFRUHG KLJKHU RQ WKH PDVFXOLQLW\ VFDOH RI WKH 2PQLEXV 3HUVRQDOLW\ ,QYHQWRU\ 23,f WKDQ WKRVH ZLWK DYHUDJH VFRUHV RQ ERWK WDVNV PAGE 174 ,W LV QRW XQWLO DGROHVFHQFH WKDW VH[ GLIIHUHQFHV LQ WKH SHUFHSWLRQ RI PDWKHPDWLFV DV D PDOH GRPDLQ DUH IRXQG )HQQHPDU 6WHLQ 6WHLQ t 6PLWKOHVV 9HUEHNH f ,Q D VWXG\ ZLWK QG WK DQG WK JUDGHUV 6WHLQ DQG 6PLWKOHVV f IRXQG WKDW VWXGHQWVn SHUFHSWLRQV RI VSDWLDO PHFKDQLFDO DQG DULWKPHWLF VNLOOV DV PDVFXOLQH EHFDPH PRUH GHILQHG DV WKHVH VWXGHQWV JRW ROGHU )HQQHPD f FRQVLGHUV WKDW WKH LQIOXHQFH HDFK VH[ H[HUWV XSRQ WKH RWKHU RQ DOO DVSHFWV RI EHKDYLRU LV VWURQJHU GXULQJ DGROHVFHQFH 6LQFH GXULQJ WKHVH \HDUV PDOHV VWHUHRW\SH PDWKHPDWLFV DV D PDOH GRPDLQ WKH\ VHQG WKLV PHVVDJH WR IHPDOHV ZKR LQ WXUQ WHQG WR EH LQIOXHQFHG LQ WKHLU ZLOOLQJQHVV WR VWXG\ RU QRW WR VWXG\ PDWKHPDWLFV %HIRUH WKDW VWDJH JLUOV FRQVLGHU DULWKPHWLF IHPLQLQH ZKLOH ER\V FRQVLGHU LW DSSURSULDWH IRU ERWK VH[HV %REEH f 8VHIXOQHVV RI PDWKHPDWLFV )HPDOHV SHUFHLYH PDWKHPDn WLFV DV OHVV XVHIXO WR WKHP WKDQ PDOHV GR DQG WKLV SHUFHSWLRQ RFFXUV DW D YHU\ \RXQJ DJH $V D UHVXOW IHPDOHV H[HUW OHVV HIIRUW WKDQ PDOHV WR OHDUQ RU HOHFW WR WDNH DGYDQFHG PDWKHPDWLFV FRXUVHV 0DQ\ VWXGLHV UHSRUWHG EHIRUH IRXQG WKDW WKH SHUFHSWLRQ RI WKH XVHIXOQHVV RI PDWKHPDWLFV IRU RQHnV IXWXUH GLIIHUV IRU PDOHV DQG IHPDOHV DQG LV UHODWHG WR FRXUVH WDNLQJ SODQV DQG EHKDYLRU )R[ f ,I IHPDOHV GR QRW SHUFHLYH PDWKHPDWLFV DV XVHIXO IRU WKHLU IXWXUH PAGE 175 WKH\ VKRZ OHVV LQWHUHVW LQ WKH VXEMHFW WKDQ FRXQWHUSDUW PDOHV 7KHVH GLIIHUHQFHV LQ LQWHUHVW DUH ZKDW +LOWRQ DQG %HUJOXQG f VXJJHVW WR DFFRXQW IRU VH[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFV DFKLHYHPHQW $OWKRXJK WKH SHUFHSWLRQ RI WKH XVHIXOQHVV RI PDWKHPDWLFV LV VWLOO DQ LPSRUWDQW SUHGLFWRU RI FRXUVH WDNLQJ IRU JLUOVr WKHUH LV D JURZLQJ VLPLODULW\ EHWZHHQ PDOHV DQG IHPDOHV UHJDUGLQJ WKH XVHIXOQHVV RI PDWKHPDWLFV $UPVWURQJ t 3ULFHr )HQQHPD t 6KHUPDQr 0ROOHUr f $UPVWURQJ DQG 3ULFH LQYHVWLJDWHG WKH UHODWLYH LQIOXHQFH RI VHOHFWHG IDFWRUV LQ VH[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFV SDUWLFLSDWLRQ %RWK PDOHV DQG IHPDOHV VHOHFWHG XVHIXOQHVV RI PDWKHPDWLFV DV WKH PRVW LPSRUWDQW IDFWRU LQ GHFLDLQJ ZKHWKHU RU QRW WR WDNH PRUH PDWKHPDWLFV LQ KLJK VFKRRO 0ROOHUnV VWXG\ UHYHDOHG WKDW ERWK PDOHV DQG IHPDOHV EDVHG PDWKHPDWLFV FRXUVHWDNLQJ GHFLVLRQV RQ FDUHHU XVHIXOQHVV $ )HQQHPD DQG 6KHUPDQ f VWXG\ VKRZHG RQO\ VOLJKW GLIIHUHQFHV EHWZHHQ PDOHV DQG IHPDOHV LQ WKHLU IHHOLQJV DERXW WKH XVHIXOQHVV RI PDWKHPDWLFV ,Q KHU VWXG\ RI WKLV YDULDEOH DPRQJ FROOHJH VWXGHQWVr 3UREHUW f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f IRXQG WKDW WKLV YDULDEOH LQIOXHQFHG FRXUVH HOHFWLRQ PRUH WKDQ DOO WKH FRJQLWLYH YDULDEOHV SUHYLRXVO\ GLVFXVVHG 7KH FDVH IRU WKH VRFLHWDO LQIOXHQFHV RQ VH[ UROHV DQG H[SHFWDWLRQV WR DFFRXQW IRU WKH GLIIHUHQFHV LQ PDWKHPDWLFV OHDUQLQJ LV DOVR VXSSRUWHG LQ RQH ZD\ RU DQRWKHU LQ WKH VWXGLHV UHSRUWHG LQ WKH OLWHUDWXUH 3UREOHP 6ROYLQJ 3HUIRUPDQFH DQG 5HODWHG 9DULDEOHV 3UREOHP VROYLQJ KDV EHHQ SHUKDSV WKH PRVW H[WHQVLYHO\ UHVHDUFKHG DUHD LQ PDWKHPDWLFV HGXFDWLRQ 3XEOLVKHG UHYLHZV E\ .LOSDWULFN f 5LHGHVHO f DQG 6X\GDP DQG :HDYHU f DWWHVW WR WKLV 0XFK RI WKH UHVHDUFK GRQH KDV IRFXVHG RQ LGHQWLI\LQJ WKH GHWHUPLQDQWV RI SUREOHP GLIILFXOW\ DQG WKH SUREOHP IHDWXUHV WKDW LQIOXHQFH WKH VROXWLRQ SURFHVV PAGE 177 $W SUHVHQW} QR VHW RI YDULDEOHV KDV EHHQ FOHDUO\ HVWDEOLVKHG DV D GHWHUPLQDQW RI SUREOHP GLIILFXOW\ 6HYHUDO UHVHDUFKHUV KDYH LQYHVWLJDWHG WKH HIIHFW RI UHDGLQJ DQG FRPSXWDWLRQ RQ SUREOHP VROYLQJ SHUIRUPDQFH 2WKHUV KDYH VWXGLHG WKH HIIHFW RI VWXGHQW DWWLWXGHV WRZDUG SUREOHP VROYLQJ LQ SUREOHP VROYLQJ OHDUQLQJ DQG DFKLHYHPHQW 7\SLFDOO\ FRUUHODWLRQDO PHWKRGV KDYH EHHQ XVHG WR LQYHVWLJDWH WKHVH TXHVWLRQV &RPSXWDWLRQDO 6NLOOV DQG 3UREOHP6ROYLQJ 3HUIRUPDQFH 2QH RI WKH ILUVW UHVHDUFKHUV WR VWXG\ WKH HIIHFW RI FRPSXWDWLRQ DQG UHDGLQJ RQ SUREOHP VROYLQJ SHUIRUPDQFH ZDV +DQVHQ f +H LQYHVWLJDWHG WKH UHODWLRQVKLS RI DULWKPHWLFDO IDFWRUV PHQWDO IDFWRUV DQG UHDGLQJ IDFWRUV WR DFKLHYHPHQW LQ SUREOHP VROYLQJ 6L[WK JUDGH VWXGHQWV ZHUH DGPLQLVWHUHG WHVWV LQ SUREOHP VROYLQJ DQG FDWHJRUL]HG DV VXSHULRU DFKLHYHUV EHVW SUREOHP VROYHUVf DQG LQIHULRU DUFKLHYHUV SRRUHVW SUREOHP VROYHUVf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} WKHVH ILQGLQJV VKRXOG EH WDNHQ FDXWLRXVO\r DV WKH FRQWHQW RI WKH *DWHV WHVWV XVHG WR PHDVXUH UHDGLQJf LV OLWHUDU\ DQG GRHV QRW LQFOXGH PDWKHPDWLFDO PDWHULDO &KDVH f VWXGLHG YDULDEOHV LQ DQ HIIRUW WR ILQG RXW ZKLFK RQHV KDYH VLJQLILFDQW LQIOXHQFH RQ WKH D'LOLW\ WR VROYH YHUEDO PDWKHPDWLFV SUREOHPV 2QO\ FRPSXWDWLRQr UHDGLQJ WR QRWH GHWDLOVr DQG IXQGDPHQWDO NQRZOHGJH ZHUH SULPDULO\ UHODWHG WR SUREOHP VROYLQJ &RPSXWDWLRQ DFFRXQWHG IRU b RI WKH b YDULDQFH GLUHFWO\ DVVRFLDWHG ZLWK SUREOHP VROYLQJ &KDVH FRQFOXGHG WKDW D SXSLOnV DELOLW\ LQ WKH PHFKDQLFV RI FRPSXWDWLRQr FRPSUHKHQVLRQ RI WKH SULQFLSOHV WKDW XQGHUOLQH WKH QXPEHU V\VWHPVr DQG WKH H[WHQW WR ZKLFK LPSRUWDQW LWHPV RI LQIRUPDWLRQ DUH QRWLFHG ZKHQ UHDGLQJr DUH JRRG SUHGLFWRUV RI WKH VWXGHQWnV DELOLW\ LQ VROYLQJ YHUEDO SUREOHPV %DORZ f LQYHVWLJDWHG WKH LPSRUWDQFH RI UHDGLQJ DELOLW\ DQG FRPSXWDWLRQ DELOLW\ LQ SUREOHP VROYLQJ SHUIRUPDQFH +H REMHFWHG WR WKH DSSURDFKHV XVHG E\ RWKHU UHVHDUFKHUV ZKR LQ WKHLU DQDO\VHV GLFKRWRPL]HG UHVHDUFK VXEMHFWV DV SRRU RU JRRG VWXGHQWVr DQG ZKR LJQRUHG WKH UHFRJQL]HD HIIHFW RI LQWHOOLJHQFH RQ UHDGLQJ DQG RQ PDWKHPDWLFV DFKLHYHPHQW %DORZ DGPLQLVWHUHG WKH 6WDQIRUG PAGE 179 $FKLHYHPHQW 7HVW VXEWHVWV RI UHDGLQJ DULWKPHWLF DQG UHDVRQLQJf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f DWWHPSWHG WR LQYHVWLJDWH WKH W\SHV RI GLIILFXOWLHV FKLOGUHQ KDYH LQ VROYLQJ ZRUG SUREOHPV 7KH\ DGPLQLVWHUHG WKH ZRUG SUREOHP VHFWLRQ RI WKH 0HWURSROLWDQ $FKLHYHPHQW 7HVW WR FKLOGUHQ IURP WKH WK JUDGH (UURUV ZHUH FODVVLILHG LQ WZR PAGE 180 FDWHJRULHV &DWHJRU\ LQFOXGHG FOHULFDO DQG FRPSXWDWLRQDO HUURUV &DWHJRU\ ,, LQFOXGHG RWKHU W\SHV RI HUURUV} VXFK DV DYHUDJH DQG DUHD HUURUV XVH RI ZURQJ RSHUDWLRQ QR UHVSRQVH DQG HUUHG UHVSRQVHV RIIHULQJ QR FOXHV ,W WKH VWXGHQWnV ZRUN LQGLFDWHG WKH FRUUHFW SURFHGXUH DQG \HW WKH SUREOHP ZDV PLVVHG EHFDXVH RI D FRPSXWDWLRQDO RU FOHULFDO HUURU LW ZDV DVVXPHG WKDW WKH SUREOHP ZDV UHDG DQG XQGHUVWRRG $Q DQDO\VLV RI IUHTXHQFLHV VKRZHG WKDW FOHULFDO HUURUV ZHUH UHVSRQVLEOH IRU b RI WKH SUREOHPV LQFRUUHFWO\ VROYHG FRPSXWDWLRQDO HUURUV DFFRXQWHG IRU b DQG RWKHU HUURUV IRU b RI WKH HUUHG SUREOHPV .QLIRQJ DQG +ROWDQ FRQFOXGHG WKDW LPSURYHG FRPSXWDWLRQDO VNLOOV FRXOG KDYH HOLPLQDWHG QHDUO\ KDOI RW WKH ZRUG SUREOHP HUURUV S ,OOf 7KHVH FRPSXWDWLRQDO HUURUV ZHUH PDGH LQ D FRQWH[W ZKHUH RWKHU VNLOOV VXFK DV UHDGLQJ LQWHUSUHWDWLRQ RI WKH SUREOHP DQG LQWHJUDWLRQ RI WKHVH VNLOOV QHFHVVDU\ IRU WKH VROXWLRQ RI ZRUG SUREOHPV PLJKW LQWHUDFW +RZHYHU .QLIRQJ DQG +ROWDQ VWDWH WKDW WKHLU ILQGLQJV QHLWKHU FRQILUP QRU GHQ\ WKDW LPSURYHPHQW RW UHDGLQJ VNLOOV ZLOO OHDG WR LPSURYHPHQW LQ SUREOHP VROYLQJ 7KH\ FRQFOXGH WKDW LW LV GLIILFXOW WR DWWULEXWH PDMRU LPSRUWDQFH WR UHDGLQJ DV D VRXUFH RI IDLOXUH S f ,Q D ODWHU DQDO\VLV ORRNLQJ IRU HYLGHQFH RI SRRU UHDGLQJ DELOLWLHV DIIHFWLQJ FKLOGUHQnV VXFFHVV LQ ZRUG SUREOHPV .QLIRQJ DQG +ROWDQ f LQWHUYLHZHG WKH PAGE 181 FKLOGUHQ ZKRVH HUURUV IHOO XQGHU WKH FDWHJRU\ RU RWKHU HUURUV 6WXGHQWV ZHUH DVNHG WR UHDG HDFK SUREOHP DORXG DQG DQVZHU WKHVH TXHVWLRQV :KDW NLQG RI VLWXDWLRQ GRHV WKH SUREOHP GHVFULEH" :KDW GRHV WKH SUREOHP DVN \RX WR ILQG" +RZ ZRXOG \RX ZRUN WKH SUREOHP" 1LQHW\ ILYH SHUFHQW RI WKH VWXGHQWV UHDG WKH SUREOHP FRUUHFWO\ b H[SODLQHG WKH NLQG RI VLWXDWLRQ WKH SUREOHP GHVFULEHG LQ D FRUUHFW PDQQHU b FRUUHFWO\ DQVZHUHG ZKDW WKH SUREOHP ZDV DVNLQJ WKHP WR ILQGr DQG b FRUUHFWO\ DQVZHUHG WKH TXHVWLRQ RI KRZ WR ZRUN WKH SUREOHP 7KH IDFW WKDW D ODUJH SHUFHQW RI WKH VWXGHQWV ZKRVH HUURUV ZHUH FODVVLILHG DV RWKHU HUURUV LQ ZKLFK UHDGLQJ VNLOOV PLJKW KDYH EHHQ D IDFWRUf FRUUHFWO\ VWDWHG KRZ WR ZRUN WKH SUREOHPr LV VWURQJ HYLGHQFH RI WKHLU DELOLW\ WR UHDG DQG LQWHUSUHW WKH SUREOHPV FRUUHFWO\ 7KH HUURUV PDGH E\ WKLV JURXS RI VWXGHQWV KDG D GLVWLQFW RULJLQr XQUHODWHG WR UHDGLQJ DELOLW\ =DOHZVNL f LQYHVWLJDWHG WKH UHODWLYH FRQWULEXWLRQ RI YHUEDO LQWHOOLJHQFHr UHDGLQJ FRPSUHKHQVLRQr YRFDEXODU\r LQWHUSUHWDWLRQ RI JUDSKV DQG WDEOHVr PDWKHPDWLFDO FRQFHSWVr QXPEHU VHQWHQFH VHOHFWLRQr DQG FRPSXWDWLRQ WR VXFFHVVIXO PDWKHPDWLFDO ZRUG SUREOHP VROXWLRQr DQG WKH UHODWLRQVKLS RI WKH GHSHQGHQW YDULDEOH WR WKH HLJKW LQGHSHQGHQW YDULDEOHV 6KH ZRUNHG ZLWK D JURXS RI WK JUDGH FKLOGUHQ ZKR ZHUH DGPLQLVWHUHG WKH VXEWHVWV RI WKH ,RZD 7HVW RI %DVLF 6NLOOV ,7%6f DQG WKH :HVFKOHU ,QWHOOLJHQFH 6FDOH IRU PAGE 182 &KLOGUHQ :,6&f 0XOWLSOH UHJUHVVLRQ DQDO\VLV ZDV SHUIRUPHG $ FRUUHODWLRQ RI ZDV IRXQG EHWZHHQ ZRUG SUREOHP VROYLQJ DQG WKH HLJKW LQGHSHQGHQW YDULDEOHV &RUUHODWLRQV EHWZHHQ ZRUG SUREOHP VROYLQJ DQG WKH LQGHSHQGHQW YDULDEOHV UDQJHG IURP YHUEDO LQWHOOLJHQFHf WR PDWKHPDWLFDO FRQFHSWVf &RUUHODWLRQV EHWZHHQ WKH LQGHSHQGHQW YDULDEOHV UDQJHG IURP YHUEDO LQWHOOLJHQFH DQG FRPSXWDWLRQf WR UHDGLQJ FRPSUHKHQVLRQ DQG YRFDEXODU\f 0DWKHPDWLFDO FRQFHSWV} FRPSXWDWLRQ} DQG QXPEHU VHQWHQFH VHOHFWLRQ ZHUH DOPRVW DV HIIHFWLYH DV DOO HLJKW LQGHSHQGHQW YDULDEOHV LQ SUHGLFWLQJ DFKLHYHPHQW LQ PDWKHPDWLFDO ZRUG SUREOHP VROYLQJ 0DWKHPDWLFDO FRQFHSWV} FRPSXWDWLRQ} QXPEHU VHQWHQFH VHOHFWLRQ} DQG UHDGLQJ FRPSUHKHQVLRQ DFFRXQWHG IRU b RI WKH YDULDQFH} ZKHUHDV DOO HLJKW SUHGLFWRUV DFFRXQWHG IRU b RI WKH YDULDQFH 7KH WZR EHVW SUHGLFWRUV ZHUH PDWKHPDWLFDO FRQFHSWV DQG FRPSXWDWLRQ} ZKLFK DFFRXQWHG IRU b YDULDQFH 2WKHU YDULDEOHV DFFRXQWHG IRU DERXW b RI WKH YDULDQFH 7KH DXWKRU UHFRPPHQGV WKDW WKH ILQGLQJV RI WKLV VWXG\ EH LQWHUSUHWHG FDXWLRXVO\ EHFDXVH WKH FRUUHODWLRQ EHWZHHQ WKH HLJKW LQGHSHQGHQW YDULDEOHV ZDV KLJK} DQG} DFFRUGLQJ WR =DO HZVNL} LQ D VWXG\ RI WKLV QDWXUH ZKHUH WKH LQWHUHVW LV SULPDULO\ LQ WKH LQIOXHQFH RI VHYHUDO YDULDEOHV RQ RQH GHSHQGHQW YDULDEOH} D ORZ FRUUHODWLRQ EHWZHHQ WKH LQGHSHQGHQW YDULDEOHV LV UHTXLUHG S f PAGE 183 ,Q D PRUH UHFHQW LQYHVWLJDWLRQ" ([HGLVLV f VWXGLHG WKH FRQWULEXWLRQ RI UHDGLQJ DELOLW\r YRFDEXODU\r PDWKHPDWLFDO FRQFHSWVr FRPSXWDWLRQr VH[r DQG UDFH RQ SUREOHPVROYLQJ SHUIRUPDQFH 7KH ,RZD 7HVW RI %DVLF 6NLOOV ZDV DGPLQLVWHUHG WR D JURXS RI WKr WKr DQG WK JUDGH DQJOR DQG EODFN &KLFDJR PDOH DQG IHPDOH DGROHVFHQWV 3UREOHP VROYLQJ ZDV KLJKO\ FRUUHODWHG WR DQ XQGHUVWDQGLQJ RI EDVLF PDWKHPDWLFDO FRQFHSWVr VRPHZKDW FRUUHODWHG WR UDFHr DQG ZHDNO\ FRUUHODWHG WR FRPSXWDWLRQDO DQG YRFDEXODU\ VNLOOVr VH[r DQG UHDGLQJ DELOLW\ $OWKRXJK WKH ILQGLQJV RI WKHVH VWXGLHV VKRZ D UHODWLRQVKLS EHWZHHQ FRPSXWDWLRQDO VNLOOV DQG SUREOHP VROYLQJ DFKLHYHPHQWr WKLV UHODWLRQVKLS LV QRW VWURQJ HQRXJK WR EH FRQVLGHUHG WKH PRVW GHWHUPLQDQW IDFWRU LQ SUREOHP VROYLQJ DFKLHYHPHQWr DV VRPH RI WKH UHVHDUFKHUV KDYH EHHQ FDUHIXO WR SRLQW RXW ,Q VSLWH RI WKH GLVPLVVDO RI UHDGLQJ DV D GHWHUPLQDQW IDFWRU LQ SUREOHP VROYLQJ DFKLHYHPHQW E\ VRPH RI WKHVH VDPH UHVHDUFKHUVr PRUH UHFHQW VWXGLHV LQ WKLV DUHD KDYH OHG RWKHUV WR KROG GLIIHUHQW YLHZV +HDGLQBJBDQGB-fU RELHPB 6ROYLQJ 3HUIRUPDQFH 0DUWLQ f VWXGLHG WKH FRQWULEXWLRQ RI UHDGLQJ FRPSUHKHQVLRQr FRPSXWDWLRQr DEVWUDFW YHUEDO UHDVRQLQJr DQG DULWKPHWLF FRQFHSWV WR DULWKPHWLF SUREOHP VROYLQJ SHUIRUPDQFH )RXUWK DQG WK JUDGH VWXGHQWV ZHUH DGPLQLVWHUHG WKH ,RZD WHVWV RI %DVLF 6NLOOV DQG WKH /RUJH7KRUQGOLNH LQWHOOLJHQFH WHVW YHUEDOf PAGE 184 +H IRXQG WKDW LQ WKH WK JUDGH WKH FRUUHODWLRQV EHWZHHQ SUREOHP VROYLQJ DQG DEVWUDFW YHUEDO UHDVRQLQJ} UHDGLQJ FRPSUHKHQVLRQ DULWKPHWLF FRQFHSWV DQG FRPSXWDWLRQ ZHUH DQG UHVSHFWLYHO\ DQG DQG LQ WKH WK JUDGH :KHQ FRPSXWDWLRQ ZDV KHOG FRQVWDQW WKH FRUUHODWLRQ EHWZHHQ SUREOHP VROYLQJ DQG UHDGLQJ ZDV LQ JUDGH DQG LQ JUDGH :KHQ UHDGLQJ ZDV KHOG FRQVWDQW WKH FRUUHODWLRQ EHWZHHQ SUREOHP VROYLQJ DQG FRPSXWDWLRQ ZDV LQ JUDGH DQG LQ JUDGH &UHVZHOO f ZRUNHG ZLWK D VDPSOH RI DQJOR DQG EODFN DGROHVFHQWV IURP &KLFDJR (DFK VXEMHFW ZDV DGPLQLVWHUHG WKH &DOLIRUQLD $FKLHYHPHQW WHVW 0XOWLSOH UHJUHVVLRQ ZDV XVHG WR DQDO\]H WKH GDWD 7KH DQDO\VLV VKRZHG WKDW UHDGLQJ LV PRUH LPSRUWDQW WKDQ FRPSXWDWLRQ LQ SUHGLFWLQJ VWXGHQW SHUIRUPDQFH LQ SUREOHP VROYLQJ 5HDGLQJ DFFRXQWHG IRU b RI WKH YDULDQFH FRPSXWDWLRQ DFFRXQWHG IRU b RI WKH YDULDQFH %DOOHZ DQG &XQQLQJKDP f ZRUNHG ZLWK WK JUDGH VWXGHQWV LQ DQ DWWHPSW WR ILQG ZKDW SURSRUWLRQ RI VWXGHQWV KDYH DV WKHLU PDLQ VRXUFH RI GLIILFXOW\ ZLWK ZRUG SUREOHPV HDFK RI WKH IROORZLQJ IDFWRUV Df FRPSXWDWLRQ VNLOOV Ef LQWHUSUHWDWLRQ RI WKH SUREOHP Ff UHDGLQJ DQG Gf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f )RU WHVW WKH HIIHFWV RI UHDGLQJ DQG FRPSXWDWLRQ ZHUH UHPRYHG E\ UHDGLQJ WKH SUREOHPV WR WKH VWXGHQWV DQG E\ JLYLQJ VFRUHV EDVHG RQ ZKHWKHU RU QRW WKH VWXGHQWV VHW WKHP XS SURSHUO\ LQ DQ DWWHPSW WR PHDVXUH SUREOHP LQWHUSUHWDWLRQ DORQH )RU WHVW WKH HIIHFW RI FRPSXWDWLRQ ZDV UHPRYHG 7KH WHVW \LHOGHG WZR VFRUHVfÂ§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f PAGE 186 WKDQ WKH SUREOHPLQWHUSUHWDWLRQ VFRUH" WKH GLIILFXOW\ ZDV GXH WR UHDGLQJ DELOLW\ ,I WKH VFRUH RI WKH ORZHVW RI WKH WKUHH DUHDV FRPSXWDWLRQ" SUREOHP LQWHUSUHWDWLRQ" DQG UHDGLQJSUREOHP LQWHUSUHWDWLRQ VFRUHf ZDV WKH VDPH DV WKH UHDGLQJSUREOHP VROYLQJ VFRUH" WKH VWXGHQWnV DUHD RI JUHDWHVW LPPHGLDWH QHHG ZDV HLWKHU FRPSXWDWLRQ" SUREOHP LQWHUSUHWDWLRQ" RU UHDGLQJ ,I WKH UHDGLQJSUREOHP VROYLQJ VFRUH ZDV ORZHU WKDQ WKH ORZHVW RI WKH RWKHU WKUHH VFRUHV" WKH VWXGHQWnV DUHD RI JUHDWHVW LPPHGLDWH QHHG ZDV LQWHJUDWLRQ $QDO\VLV RI WKH GDWD UHYHDOHG WKDW IRU b RI WKH VWXGHQWV" SUREOHP LQWHUSUHWDWLRQ ZDV WKHLU PDMRU GLIILFXOW\ IRU b RI WKH VWXGHQWV" LQWHJUDWLRQ WRWDO SUREOHP VROYLQJf ZDV WKHLU JUHDWHVW LPPHGLDWH QHHG IRU DQRWKHU b" FRPSXWDWLRQ ZDV WKH PDMRU ZHDNQHVV DQG IRU b" UHDGLQJ ZDV WKHLU JUHDWHVW LPPHGLDWH QHHG 6HYHQW\ ILYH SHUFHQW RI WKH VWXGHQWV GHPRQVWUDWHG FOHDU VWUHQJWK LQ FRPSXWDWLRQ" b LQ SUREOHP LQWHUSUHWDWLRQ" DQG b LQ UHDGLQJSUREOHP LQWHUSUHWDWLRQ $Q DQDO\VLV DFURVV DOO VWXGHQWV LQFOXGLQJ WKRVH ZLWKRXW FRPSOHWH GDWDf VKRZHG WKDW b RI WKH VXEMHFWV FRXOG QRW ZRUN ZRUG SUREOHPV DW D OHYHO DV KLJK DV WKDW DW ZKLFK WKH\ FRXOG FRPSXWH" LQWHUSUHW SUREOHPV" DQG UHDG DQG LQWHUSUHW SUREOHPV" ZKHQ WKRVH DUHDV ZHUH PHDVXUHG VHSDUDWHO\ 7KLV OHG WKHP WR FRQFOXGH WKDW NQRZLQJ WKH VNLOOV RU WKH FRPSRQHQWV RI VROYLQJ ZRUG SUREOHPV LV QRW VXIILFLHQW IRU PAGE 187 VXFFHVVr VLQFH WKH FRPSRQHQWV PXVW EH LQWHJUDWHG LQWR D ZKROH SURFHVV PDVWHU\ OHDUQLQJ RI WKH FRPSRQHQWV FDQQRW DVVXUH PDVWHU\ RI WKH SURFHVVf 7KHLU DQDO\VLV DOVR OHG WKHP WR FRQFOXGH WKDWr LQ WKH FDVH RI WK JUDGHUVr LQDELOLW\ WR UHDG SUREOHPV LV D PDMRU REVWDFOH LQ VROYLQJ ZRUG SUREOHPV 2QO\ b RI WKH VXEMHFWV FRXOG UHDG DQG VHW XS SUREOHPV FRUUHFWO\ DW D KLJKHU OHYHO WKDQ WKH\ FRXOG FRPSXWHr ZKLOH b FRXOG FRPSXWH FRUUHFWO\ DW D KLJKHU OHYHO WKDQ WKH\ FRXOG UHDG DQG VHW XS SUREOHPV b FRXOG VHW XS SUREOHPV EHWWHU ZKHQ WKH\ KHDUG WKHP UHDG WKDQ ZKHQ WKH\ UHDG WKH SUREOHP WKHPVHOYHV 2QO\ b FRXOG VHW XS SUREOHPV EHWWHU ZKHQ WKH\ UHDG WKHP WKDQ ZKHQ WKH\ KHDUG WKHP UHDG 0XWK f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f ZLWK WZR YHUVLRQV RI V\QWDFWLF VWUXFWXUH VLPSOH YV FRPSOH[ V\QWD[f 7DVN SHUIRUPDQFH ZDV PHDVXUHG E\ PHDQV RI WKH QXPEHU RI SUREOHPV DQVZHUHG FRUUHFWO\r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b RI WKH YDULDQFH LQ WRWDO FRUUHFW DQVZHUV DQG FRPSXWDWLRQ DFFRXQWHG IRU b 5HDGLQJ DELOLW\ DQG FRPSXWDWLRQDO DELOLW\ XQLTXHO\ DFFRXQWHG IRU b DQG b RI WKH YDULDQFH LQ WKH QXPEHU RI FRUUHFW DQVZHUV UHVSHFWLYHO\ ([WUDQHRXV LQIRUPDWLRQ DGGHG VLJQLILFDQWO\ WR WKH YDULDQFH H[SODLQHG LQ WKH QXPEHU RI FRUUHFW DQVZHUV EXW V\QWDFWLF VWUXFWXUH GLG QRW 5HDGLQJ DELOLW\ DFFRXQWHG IRU b RI WKH YDULDQFH LQ WHVWWDNLQJ WLPH EXW FRPSXWDWLRQ GLG QRW DGG VLJQLILFDQWO\ WR WKH YDULDQFH H[SODLQHG E\ UHDGLQJ 0XWK FRQFOXGHG WKDW UHDGLQJ DQG FRPSXWDWLRQ ERWK FRQWULEXWH VLJQLILFDQWO\ WR VXFFHVV LQ VROYLQJ DULWKPHWLF ZRUG SUREOHPV EXW WKDW UHDGLQJ SOD\V D PRUH VLJQLILFDQW UROH WKDQ GRHV FRPSXWDWLRQ PAGE 189 7KH VWXGLHV UHYLHZHG LQ WKLV VHFWLRQ VKRZ D SRVLWLYH UHODWLRQVKLS EHWZHHQ UHDGLQJ DQG SUREOHP VROYLQJ SHUIRUPDQFH EXW LQ WKH FDVH RI %DOOHZ DQG &XQQLQJKDP f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f KDV VXJJHVWHG WKDW DQ LQGLYLGXDOnV DWWLWXGH WRZDUG RQH DVSHFW RI WKH GLVFLSOLQH PDWKHPDWLFVf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f FRQVWUXFWHG D VFDOH WR PHDVXUH DWWLWXGHV WRZDUG SUREOHP VROYLQJ +HU LQWHUHVW ZDV LQ JHQHUDO SUREOHP VROYLQJ UDWKHU WKDQ LQ PDWKHPDWLFDO SUREOHP VROYLQJ +HU ZRUN FRQVWLWXWHV WKH ILUVW DWWHPSW WR FRQVWUXFW D PHDVXUH RI DWWLWXGHV WRZDUG SUREOHP VROYLQJ 7KH VFDOH ZDV XVHG ZLWK D JURXS RI FROOHJH VWXGHQWV DQG VKH IRXQGU DPRQJ RWKHU WKLQJV WKDW SUREOHP VROYLQJ SHUIRUPDQFH LV SRVLWLYHO\ UHODWHG WR SUREOHP VROYLQJ DWWLWXGHV DQG WKDW LQ WKH FDVH RI IHPDOHV SRVLWLYH PRGLILFDWLRQ RI DWWLWXGHV WRZDUG SUREOHP VROYLQJ EULQJV D VLJQLILFDQW JDLQ LQ SUREOHP VROYLQJ SHUIRUPDQFH /LQGJUHQ 6LOYD )DUDFR] DQG 'D5RFKD f DGDSWHG &DUH\nV VFDOH RI DWWLWXGHV WRZDUG SUREOHP VROYLQJ DQG DSSOLHG LW WR D JURXS RI WK JUDGH %UD]LOLDQ FKLOGUHQ 6WXGHQWV DOVR DQVZHUHG DQ DULWKPHWLF DFKLHYHPHQW WHVW D JHQHUDO LQWHOOLJHQFH WHVW DQG D VRFLRHFRQRPLF 6(f VFDOH $ ORZ EXW VLJQLILFDQW SRVLWLYH FRUUHODWLRQ ZDV IRXQG EHWZHHQ DULWKPHWLF DFKLHYHPHQW DQG DWWLWXGHV WRZDUG SUREOHP VROYLQJ $ QHDU ]HUR FRUUHODWLRQ ZDV IRXQG EHWZHHQ DWWLWXGHV WRZDUG SUREOHP VROYLQJ DQG LQWHOOLJHQFH 6LQFH SUREOHP VROYLQJ LV RQH DVSHFW RI WKH GLVFLSOLQH RI PDWKHPDWLFV WKLV FRUUHODWLRQ EHWZHHQ DWWLWXGHV DQG DULWKPHWLF DFKLHYHPHQW FDQ OHDG WR D FRQFOXVLRQ RU D VWURQJ PAGE 191 FRUUHODWLRQ EHWZHHQ DWWLWXGHV WRZDUG SUREOHP VROYLQJ DQG SUREOHP VROYLQJ SHUIRUPDQFH :KLWDNHU f FRQVWUXFWHG D VWXGHQW DWWLWXGH VFDOH WR PHDVXUH VRPH DVSHFWV RI WK JUDGH VWXGHQW DWWLWXGHV WRZDUG PDWKHPDWLF SUREOHP VROYLQJ +H LQFOXGHG VWDWHPHQWV UHIOHFWLQJ FKLOGUHQn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n SHUIRUPDQFH LQ SUREOHP VROYLQJ $ YHU\ ZHDN DQG QRQVLJQLILFDQW QHJDWLYH FRUUHODWLRQ ZDV IRXQG EHWZHHQ WKH WHDFKHUnV DWWLWXGHV WRZDUG SUREOHP VROYLQJ DQG VWXGHQW SHUIRUPDQFH PAGE 192 7QH VWXGLHV UHYLHZHG KDYH FRQILUPHG WKH UHODWLRQVKLS EHWZHHQ SUREOHP VROYLQJ SHUIRUPDQFH DQG DWWLWXGHV WRZDUG SUREOHP VROYLQJ &DUH\} /LQGJUHQ HW DO} :KLWDNHU} f +RZHYHU} WKH UHVXOWV UHSRUWHD LQ WKH VWXGLHV WKDW LQYHVWLJDWHG WKH UHODWLRQVKLS EHWZHHQ SUREOHP VROYLQJ SHUIRUPDQFH DQG FRPSXWDWLRQ DQG EHWZHHQ UHDGLQJ DQG SUREOHP VROYLQJ IDLO WR EH FRQVLVWHQW LQ WKHLU FRQFOXVLRQV +DQVHQ f} &KDVH ,f} %DORZ f} .PIRQJ DQG +ROWDQ } f} DQG =DOHZVNL FRQFOXGHG WKDW FRPSXWDWLRQ LV PRUH VWURQJO\ UHODWHG WR SUREOHP VROYLQJ WKDQ LV UHDGLQJ 0DUWLQ f} &UHVZHOO f} %DOOHZ DQG &XQQLQJKDP f} DQG 0XWK f} FRQFOXGHG WKDW UHDGLQJ DRLOLW\ DQG PDWKHPDWLFDO SUREOHP VROYLQJ VKRZ D VWURQJHU UHODWLRQVQLS WKDQ FRPSXWDWLRQ DQG SUREOHP VROYLQJ ([HGLVLVnV f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r LI WKHVH GLIIHUHQFHV SHUVLVW DIWHU DFFRXQWLQJ IRU FRPSXWDWLRQDO VNLOOVr DQG LI WKH PDOHIHPDOH GLIIHUHQFHV GHSHQG RQ WKH OHYHO RI FRPSXWDWLRQDO VNLOOV 7KLV FKDSWHU FRQWDLQV GHVFULSWLRQV RI WKH VDPSOHr WKH WHVW LQVWUXPHQWr DQG WKH VWDWLVWLFDO DQDO\VLV XVHG LQ DFKLHYLQJ WKH DERYH PHQWLRQHG REMHFWLYHV 7KH 6DPSOH 7R DFKLHYH WKH ILUVW REMHFWLYH RI WKH VWXG\r DOO WKH VWXGHQWV ZKR WRRN WKH %DVLF 6NLOOV 7HVW LQ 0DWKHPDWLFV 3UXHED GH 'HVWUH]DV %Â£VLFDV HQ 0DWHPÂ£WLFDVf LQ HDFK RU WKH WKUHH \HDUV ZHUH LQFOXGHG LQ WKH VWXG\ 7R DFKLHYH WKH VHFRQG REMHFWLYH RI WKH VWXG\r DSSUR[LPDWHO\ PAGE 194 VWXGHQWV ZHUH VHOHFWHG UDQGRPO\ IRU HDFK \HDU )RU WKH ILUVW \HDUU ZHUH VHOHFWHG ER\V DQG JLUOVf IRU WKH VHFRQG \HDU VWXGHQWV ZHUH VHOHFWHG ER\V DQG JLUOVf DQG IRU WKH WKLUG \HDU VWXGHQWV ZHUH VHOHFWHG ER\V DQG JLUOVf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f M 1XPEHUV M ? ? LWHPV f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f SUHVHQWV WKH IROORZLQJ VDWXUDWHG PRGHO IRU WKH GDWD ORJ P L M N L 8M 8N8LM 8LN8MN8 L M N f ,Q WKLV PRGHO PLMN LV WKH H[SHFWHG YDOXH RI WKH IUHTXHQF\ LQ FHOO LMN RI WKH WKUHH GLPHQVLRQDO WDEOH 7KH PRGHO VWDWHV WKDW DOO WKUHH FODVVLILFDWLRQ IDFWRUV IRU D WKUHH GLPHQVLRQDO FRQWLQJHQF\ WDEOH DUH PXWXDOO\ GHSHQGHQW ,Q WKH SUHVHQW UHVHDUFK L LV WKH \HDU LQGH[ M LV WKH RSWLRQ LQGH[ DQG N LV WKH VH[ LQGH[ )LHQEHUJ VKRZV WKDW GHOHWLQJ WKH WHUPV 8LM DQG 8LMN \LHOGV D PRGHO LQ ZKLFK \HDU DQG RSWLRQ DUH LQGHSHQGHQW FRQGLWLRQDO XSRQ VH[ 7KLV PRGHO LV ORJ PLM N 8L 8M 8N8LN8M N f PAGE 199 )LHQEHUJ DOVR VKRZV WKDW GHOHWLQJ WKH 8MN WHUP IURP f WR REWDLQ ORJ PLM N 8L 8M8N 8LN f \LHOGV D PRGHO WKDW VSHFLILHV WKDW RSWLRQ LV LQGHSHQGHQW RI VH[ %DVHG RQ )LHQEHUJ V SUHVHQWDWLRQ} DQ DSSURSULDWH DQDO\VLV IRU WHVWLQJ WKH K\SRWKHVHV LV &RQGXFW D OLNHOLKRRG UDWLR WHVW RI WKH DGHTXDF\ RI PRGHO f ,I WKLV WHVW LV QRQVLJQLILFDQW} WKHQ WKH GDWD VXSSRUW WKH DGHTXDF\ RI WKH PRGHO} DQG WKH K\SRWKHVLV WKDW FRQGLWLRQDO RQ JHQGHU} RSWLRQ} DQG \HDU DUH LQGHSHQGHQW %HFDXVH PRGHO f LV D VDWXUDWHG PRGHO} WHVWLQJ WKH DGHTXDF\ RI PRGHO f LV WKH VDPH DV FRPSDULQJ WKH DGHTXDFLHV RI PRGHOV f DQG f &RQGXFW D OLNHOLKRRG UDWLR WHVW FRPSDULQJ WKH DGHTXDFLHV RI PRGHOV f DQG f ,I WKLV WHVW LV VLJQLILFDQW} WKHQ PRGHO f ILWV WKH GDWD EHWWHU WKDQ PRGHO f DQG WKH GDWD VXSSRUW WKH K\SRWKHVLV WKDW WKH FKRLFH RI LQFRUUHFW RSWLRQ LV GHSHQGHQW RQ VH[ 7R VXPPDUL]H} LI WKH ILUVW WHVW LV QRQVLJQLILFDQW DQG WKH VHFRQG WHVW LV VLJQLILFDQW} WKHQ WKH FKRLFH RI RSWLRQ LV GHSHQGHQW RQ VH[} DQG WKH SDWWHUQ RI GHSHQGHQF\ LV WKH VDPH IRU DOO WKUHH \HDUV PAGE 200 $ SUREOHP DULVHV LQ LQWHUSUHWLQJ WKLV DQDO\VLV ZLWK WKH GDWD XVHG LQ WKLV UHVHDUFK 2YHU WKH WKUHH \HDUV" WKHUH ZHUH UHVSRQVHV DYDLODEOH IURP VWXGHQWV (YHQ LI RQO\ b RI WKH VWXGHQWV DQVZHUHG DQ LWHP LQFRUUHFWO\ WKH UHVSRQVHV RI VWXGHQWV ZRXOG EH XVHG LQ DQDO\]LQJ WKLV LWHP 2Q WKH RWKHU KDQG LI b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rH ))Hf f PAGE 201 +HUH WKH VXPPDWLRQ LV RYHU DOO WKH FHOOV LQ WKH FRQWLQJHQF\ WDEOH ) UHIHUV WR WKH REVHUYHG IUHTXHQF\ LQ D FHOO DQG )H UHIHUV WR WKH HVWLPDWHG H[SHFWHG IUHTXHQF\ LQ D FHOO 'HQRWLQJ WKH REVHUYHG SURSRUWLRQ LQ D FHOO E\ 3* DQG WKH HVWLPDWHG H[SHFWHG SURSRUWLRQ LQ D FHOO E\ 3H WKH WHVW VWDWLVWLF FDQ EH ZULWWHQ 1 eA3 r JH 33H! f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f LW LV LPSRUWDQW WR FRPSDUH WKH PHWKRG RI DQDO\VLV XVHG LQ WKLV VWXG\ WR WKH RQH XVHG E\ 0DUVKDOO 0DUVKDOO DOVR XVHG D WZRVWHS DQDO\VLV ,Q WKH ILUVW VWHS RI KHU DQDO\VLV VKH GHOHWHG WKH 8LMN WHUP IURP f DQG WHVWHG WKH DGHTXDF\ RI WKH PRGHO r J P M M N L 8 M R N 8 L M 8L N M N f f )ROORZLQJ WKLV VKH GHOHWHG WKH 8MN WHUP WR REWDLQ ORJ P Â M N -L 8M 8N8L M 8LN f DQG FRPSDUHG WKHVH WZR PRGHOV XVLQJ D OLNHOLKRRG UDWLR WHVW ,I WKH ILUVW WHVW ZDV QRQVLJQLILFDQW DQG WKH VHFRQG VLJQLILFDQW 0DUVKDOO FODLPHG WKDW RSWLRQ FKRLFH ZDV GHSHQGHQW RQ VH[ DQG WKH SDWWHUQ RI GHSHQGHQF\ ZDV WKH VDPH IURP \HDU WR \HDU 7KLV LV WKH VDPH FODLP WKDW WKLV VWXG\ VRXJKW WR HVWDEOLVK +RZHYHU WKH DSSURDFK XVHG KHUH ZDV WR SUHVHQW HYLGHQFH WKDW PRGHO f ILWV WKH GDWD ZKLOH 0DUVKDOO WULHG WR VKRZ WKDW PRGHO f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r FRQVLGHU WKH WKUHH WZRE\WZR WDEOHV REWDLQHG E\ FURVV FODVVLI\LQJ VH[ DQG RSWLRQ FKRLFHV RQH DQG WZR IRU HDFK \HDU 7KHVH WDEOHV DUH LQGLFDWHG E\ GRWWHG OLQHV RQ 7DEOH )RU WKLV WDEOHr WKH UDWLR RI WKH RGGV RI D PDOH FKRRVLQJ RSWLRQ RQH WR WKH RGGV RI D IHPDOH FKRRVLQJ RSWLRQ RQH LV WKH VDPH ZLWKLQ WUXQFDWLQJ HUURUf IRU HDFK \HDU )RU H[DPSOHr WKLV RGGV UDWLR IRU WKH ILUVW \HDU LV f f :LWKLQ WKH HUURU FDXVHG E\ UHSRUWLQJ WUXQFDWHG SUREDELOLWLHVr WKH RGGV UDWLR IRU \HDUV WZR DQG WKUHH LV WKH VDPH DV WKDW IRU \HDU RQH 7KH HTXDOLW\ RI WKHVH RGGV UDWLRV LV 0DUVKDOOnV RSHUDWLRQDOL]DWLRQ VWDJH RI WKH \HDU WR\HDU JHQGHURSWLRQ GHSHQGHQF\ 7R VKRZ WKDW WKH RGGV UDWLR FDQ EH FRQVWDQW RYHU \HDUVr EXW WKDW WKH SUREDELOLWLHV RI RSWLRQ FKRLFH FRQGLWLRQDO RQ VH[ DQG \HDU FDQ FKDQJH IURP \HDU WR \HDU r IRU ERWK PDOHV DQG IHPDOHVr WKH SUREDELOLWLHV LQ 7DEOH ZHUH FRQYHUWHG WR WKH SUREDELOLWLHV RI RSWLRQ FKRLFH FRQGLWLRQDO XSRQ VH[ DQG \HDU 7KHVH FRQGLWLRQDO SUREDELOLWLHV DUH UHSRUWHG LQ 7DEOH 8QOLNH WKH SUREDELOLWLHV LQ 7DEOH r WKRVH LQ 7DEOH FKDQJH IURP \HDU WR \HDU IRU ERWK PDOHV DQG IHPDOHV PAGE 205 7DEOH +\SRWKHWLFDO 3UREDELOLWLHV RI 2SWLRQ &KRLFH &RQGLWLRQDO RQ 6H[ DQG PAGE 206 FRQGXFWHG RQH ZLWK HDFK RI DGGLWLRQ VXEWUDFWLRQ PXOWLSOLFDWLRQ GLYLVLRQ DGGLWLRQ RI IUDFWLRQV GHFLPDOV LQ VXEWUDFWLRQ DQG HTXLYDOHQFH DV FRYDULDWHV $ ZHOO NQRZQ SUREOHP WKDW DULVHV LQ WKH XVH RI $1&29$ LV WKDW DQ XQUHOLDEOH FRYDULDWH FDQ FDXVH VSXULRXV GLIIHUHQFHV EHWZHHQ WKH VH[ JURXSV 7R VROYH WKLV SUREOHP 3RUWHU f SURSRVHG WKH XVH RI HVWLPDWHG WUXH VFRUHV IRU REVHUYHG VFRUHV 3RUWHU f FRQGXFWHG D VLPXODWLRQ WKDW JDYH HPSLULFDO VXSSRUW WR WKH DGHTXDF\ RI WKLV VWUDWHJ\ +XQWHU DQG &RKHQ f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r f 6XEVWDQWLYHO\ WKLV PRGHO LPSOLHV WKDW WKH SDWWHUQ RI PDOH DQG IHPDOH RSWLRQ FKRLFHV LV FRQVLVWHQW RYHU WKH WKUHH \HDUV RI WHVW DGPLQLVWUDWLRQ 7KH VHFRQG PRGHO LQGLFDWHV WKDW RSWLRQ LV LQGHSHQGHQW RI VH[ ORJ PMMN X L X M XN XL N f f 7KLV PRGHO LPSOLHV WKDW WKH SDWWHUQ RI RSWLRQ FKRLFH LV WKH VDPH IRU PDOHV DQG IHPDOHV ,Q RUGHU WR GHWHUPLQH LI PDOHV DQG IHPDOHV GLIIHUHG LQ WKH VHOHFWLRQ RI LQFRUUHFW UHVSRQVHV DQG LI WKHVH GLIIHUHQFHV ZHUH VWDEOH DFURVV WKH WKUHH \HDUV RI WHVW DGPLQLVWUDWLRQ D WZRVWHS WHVW ZDV SHUIRUPHG )LUVW D OLNHOLKRRG UDWLR WHVW RI WKH DGHTXDF\ RI PRGHO f ZDV FRQGXFWHG IRU HDFK RI WKH LWHPV RI WKH WHVW 8QGHU WKLV WHVW D PRGHO ILWV WKH GDWD ZKHQ WKH YDOXH REWDLQHG LV QRQVLJQLILFDQW DW D VSHFLILHG DOSKD OHYHO )RU bf RI WKH LWHPV PRGHO f ILWWHG WKH GDWD DGHTXDWHO\ )RU WKHVH LWHPV WKH SDWWHUQ RI PDOH DQG IHPDOH RSWLRQ FKRLFHV ZDV FRQVLVWHQW IRU WKH WKUHH \HDUV 7KH YDOXHV IRU WKH LWHPV IRU ZKLFK WKH OLNHOLKRRG UDWLR PAGE 209 WHVWV ZHUH VLJQLILFDQW DW WKH OHYHO DUH UHSRUWHG LQ 7DEOH $OVR UHSRUWHG LQ WKLV WDEOH DUH WKH DFWXDO VDPSOH VL]HV DQG WKH PLQLPXP VDPSOH VL]HV QHFHVVDU\ IRU WKH OLNHOLKRRG UDWLR WHVWV WR EH VLJQLILFDQW 2I WKH LWHPVr KDG PLQLPXP VDPSOH VL]HV JUHDWHU WKDQ r 7KXVr DOWKRXJK IRU WKH WKUHH \HDUV ERWK PDOHV DQG IHPDOHV VDPSOHV KDG LQFRQVLVWHQW RSWLRQ FKRLFHV RQ WKHVH LWHPVr RQ RI WKH LWHPV WKH LQFRQVLVWHQF\ ZDV UHODWLYHO\ PLQRU &RQVHTXHQWO\r WKHVH LWHPV ZHUH LQFOXGHG LQ VWHS DORQJ ZLWK WKH LQLWLDO LWHPV ,Q VWHS r D OLNHOLKRRG UDWLR WHVW ZDV SHUIRUPHG FRPSDULQJ WKH DGHTXDF\ RI WKH PRGHOV LQ f DQG f 7KH YDOXHV DVVRFLDWHG ZLWK PRGHOV f DQG f IRU HDFK RI WKH LWHPV VXEMHFWHG WR VWHS DUH UHSRUWHG LQ FROXPQV E DQG F RI 7DEOH $OVR UHSRUWHG LQ 7DEOH LV WKH GLIIHUHQFH EHWZHHQ WKH WZR YDOXHV VHH FROXPQ Gf 7KLV ODWWHU ILJXUH LV WKH WHVW VWDWLVWLF IRU FRPSDULQJ WKH DGHTXDFLHV RI PRGHOV f DQG f 6LJQLILFDQW ;A VWDWLVWLFV DUH LQGLFDWHG E\ DVWHULVNV 7KH ;A VWDWLVWLFV ZHUH VLJQLILFDQW IRU RI WKH LWHPV LQGLFDWLQJ D PDOHIHPDOH GLIIHUHQFH LQ RSWLRQ FKRLFH IRU WKHVH LWHPV ,Q FROXPQ H RI 7DEOH WKH DFWXDO VDPSOH VL]HV DUH UHSRUWHG ,Q FROXPQ Ir WKH PLQLPXP VDPSOH VL]HV QHFHVVDU\ IRU VLJQLILFDQFH DUH UHSRUWHG )RU WKRVH LWHPV ZKLFK KDG VLJQLILFDQW ;A VWDWLVWLFV UHSRUWHG LQ FROXPQ Gr bf KDG PLQLPXP VDPSOH VL]HV JUHDWHU WKDQ r PAGE 210 ODO 7$%/( 6XPPDU\ RI 6LJQLILFDQW 7HVWV RI 0RGHO ; IRU RGHO r $FWXDO 6DPSOH 6L]H 0LQ 6DPSOH 6L]H IRU 6LJQLILFDQFH Ef Ff Gf PAGE 211 OD/ 7$%/( &RQWLQXHG ; IRU $FWXDO 6DPSOH 0LQ 6DPSOH 0RGHO r 6L]H 6L]H IRU 6LJQLILFDQFH Ef Ff /Âœf rr PAGE 212 Df 7$%/( &RQWLQXHG ;r IRU 0RGHO r $FWXDO 6DPSOH 6L]H 0LQ 6DPSOH 6L]H IRU 6LJQLILFDQFH Ef Ff Gf rr PAGE 213 7$%/( &RQWLQXHG ,WHP 1XPEHU ; IRU 0RGHO r $FWXDO 6DPSOH 6L]H 0LQ 6DPSOH 6L]H IRU 6LJQLILFDQFH ODf Ef Ff Gf r S rr 0LQLPXP VDPSOH VL]H OHVV WKDQ PAGE 214 7$%/( &KLVTXDUH 9DOXHV IRU WKH &RPSDULVRQ RI 0RGHOV f DQG f ZLWK $FWXDO 6DPSOH 6L]HV DQG ZLWK WKH &RUUHVSRQGLQJ 1XPEHU RI 6XEMHFWV 1HHGHG IRU 6LJQLILFDQW 5HVXOWV ,WHP 1XPn EHU ;D IRU 0RGHO ; IRU 0RGHO ;D IRU 'LIIHUHQFH $FWXDO 6DPSOH 6L]H 0LQLPXP 6DPSOH 6L]H IRU 6LJQLI Df Ef Ff Gf Hf If r r r r r rrr r r r r r r r r rrr r r r r r r r r rrr r r r r r r r r r r r r r r r r r r r rr r r r r r r PAGE 215 7$%/( &RQWLQXHG ,WHP 1XPn EHU ; IRU 0RGHO ; IRU 0RGHO ; IRU 'LIIHUHQFH $FWXDO 6DPSOH 6L]H 0LQLPXP 6DPSOH 6L]H IRU 6LJQLI Df Ef Ff Gf Hf If r r r r r r r r r r r r r r r PAGE 216 7$%/( &RQWLQXHG ,WHP 1XPn EHU ;D IRU 0RGHO ;D IRU 0RGHO ;D IRU 'LIIHUHQFH $FWXDO 6DPSOH 6L]H 0LQLPXP 6DPSOH 6L]H IRU 6LJQLI Df Ef Ff Gf Hf If r I r r r r r r r r r rrr r r rrr r r rrr r r r r r r r r r r r r r r r r r r r r r r r r r rrr r r rrr r r r r r r r r r PAGE 217 7$%/( &RQWLQXHG ,WHP 1XPn EHU ;] IRU 0RGHO ; IRU 0RGHO ;D IRU 'LIIHUHQFH $FWXDO 6DPSOH 6L]H 0LQLPXP 6DPSOH 6L]H IRU 6LJQLI Df Ef Ff Gf Hf If r r r r r r rr r r r r r r r r r r r r PAGE 218 7$%/( &RQWLQXHG ,WHP 1XPn EHU ; IRU 0RGHO ;] IRU 0RGHO ;] IRU 'LIIHUHQFH $FWXDO 6DPSOH 6L]H 0LQLPXP 6DPSOH 6L]H IRU 6LJQLI Df Ef Ff Gf Hf If r r r r r r r r r r r rr r rr r r r r r r PAGE 219 7$%/( &RQWLQXHG ,WHP 1XPn EHU ; IRU 0RGHO ; IRU 0RGHO ; IRU 'LIIHUHQFH $FWXDO 6DPSOH 6L]H 0LQLPXP 6DPSOH 6L]H IRU 6LJQLI Df Ef Ff Gf Hf If r r r r r r r r r r r r r r r r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f IRU DOO WKUHH \HDUV PAGE 221 7$%/( 0HDQVr 6WDQGDUG 'HYLDWLRQVr DQG W7HVWV IRU WKH (LJKW 0DWKHPDWLFDO 6XEWHVWV PAGE 222 7$%/( &RQWLQXHG PAGE 223 &RQVLVWHQW VLJQLILFDQW GLIIHUHQFHV ZHUH DOVR IRXQG IRU DGGLWLRQr PXOWLSOLFDWLRQr GLYLVLRQr DGGLWLRQ RI IUDFWLRQVr DQG VXEWUDFWLRQ RI GHFLPDOV )RU VXEWUDFWLRQ WKH GLIIHUHQFH ZDV QRW VWDWLVWLFDOO\ VLJQLILFDQW 4XHVWLRQ 'R VH[UHODWHG GLIIHUHQFHV LQ SUREOHP VROYLQJ SHUVLVW ZKHQ FRPSXWDWLRQDO VNLOOV DUH FRQWUROOHG IRUr DQG LV WKH PDOHIHPDOH GLIIHUHQFHV LQ SUREOHP VROYLQJ GHSHQGHQW RQ OHYHO RI FRPSXWDWLRQDO VNLOOV" 7R DGGUHVV WKH TXHVWLRQ RI GHSHQGHQFH RI PDOHIHPDOH SUREOHP VROYLQJ GLIIHUHQFHV LQ FRPSXWDWLRQDO VNLOOV IRU HDFK \HDU DQG FRPSXWDWLRQ VXEWHVWr WKH SRVVLELOLW\ RI DQ LQWHUDFWLRQ ZDV LQYHVWLJDWHG )RU WKH ILUVW \HDUr VWDWLVWLFDOO\ VLJQLILFDQW LQWHUDFWLRQV ZHUH IRXQG EHWZHHQ VH[ DQG PXOWLSOLFDWLRQr ) rf r S& DQG VH[ DQG GLYLVLRQr ) rf r S& $ VLJQLILFDQW LQWHUDFWLRQ ZDV IRXQG EHWZHHQ VH[ DQG VXEWUDFWLRQ LQ WKH VHFRQG \HDUr ) rf r S@ 1R VLJQLILFDQW LQWHUDFWLRQV ZHUH IRXQG LQ WKH WKLUG \HDU $QDO\VLV RI FRYDULDQFH VXPPDU\ WDEOHV DUH VKRZQ DV 7DEOHV r r DQG $OVRr WKH WKUHH LQWHUDFWLRQV DUH GHSLFWHG LQ )LJXUHV r r DQG (DFK ILJXUH LQGLFDWHV WKDW DW ORZHU OHYHOV RI FRPSXWDWLRQDO VNLOOVr PDOHV RXWSHUIRUPHG IHPDOHV LQ SUREOHP VROYLQJr ZLWK WKH UHYHUVH KDSSHQLQJ DW KLJKHU OHYHOV RI FRPSXWDWLRQDO VNLOOV PAGE 224 7$%/( $1&29$ 6XPPDU\ 7DEOH 0XOWLSOLFDWLRQ &RYDULDWH )LUVW PAGE 225 7$%/( $1&29$ 6XPPDU\ 7DEOH 6XEWUDFWLRQ &RYDULDWH 6HFRQG PAGE 226 3UREOHP 6ROYLQJ f 7 f, 0XOWLSOLFDWLRQ )LJ 6H[ E\ 0XOWLSOLFDWLRQ ,QWHUDFWLRQ PAGE 227 3UREOHP 6ROYLQJ n U U W 'LYLVLRQ )LJ 6H[ E\ 'LYLVLRQ ,QWHUDFWLRQ PAGE 228 3UREOHP 6ROYLQJ 6XEWUDFWLRQ )LJ 6H[ E\ 6XEWUDFWLRQ ,QWHUDFWLRQ PAGE 229 7KH VLJQLILFDQW LQWHUDFWLRQV IRXQG EHWZHHQ VH[ DQG PXOWLSOLFDWLRQA DQG VH[ DQG GLYLVLRQ ILUVW \HDUf DQG VH[ DQG VXEWUDFWLRQ VHFRQG \HDUf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f 5HOLDELOLW\ FRHIILFLHQWV FDOFXODWHG IRU HDFK FRYDULDWH DUH VKRZQ LQ 7DEOH 6XPPDULHV RI WKH DQDO\VHV RI FRYDULDQFH IRU WKH ILUVW \HDU DUH UHSRUWHG LQ 7DEOH 7KH UHVXOWV VKRZ WKDW IHPDOHV UHWDLQHG WKHLU VXSHULRULW\ LQ SUREOHP VROYLQJ SHUIRUPDQFH ZKHQ HTXLYDOHQFH ZDV WKH FRQWUROOLQJ YDULDEOH LQ WKH DQDO\VLV RI FRYDULDQFH WKH RQO\ YDULDEOH LQ ZKLFK PDOHV RXWSHUIRUPHG IHPDOHV QRQVLJQLILFDQWf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f VXPPDU\ DQG LQWHUSUHWDWLRQ RI WKH UHVXOWV RI WKH VWXG\ DQG Ef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nV ILQGLQJV f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f IRU WKH VWXG\ RI VH[UHODWHG GLIIHUHQFHVr LW FRQWLQXHV WR EH D SURPLVLQJ WHFKQLTXH LQ WKH GLVFRYHU\ RI KRZ FKLOGUHQ DSSURDFK PDWKHPDWLFV SUREOHP VROYLQJr PRUH VR WKDQ DQDO\VHV WKDW ORRN RQO\ DW WKH LWHPV DQVZHUHG FRUUHFWO\ RU DW WRWDO WHVW VFRUHV 7KH VHFRQG REMHFWLYH RI WKH VWXG\ ZDV WR LQYHVWLJDWH VH[UHODWHG GLIIHUHQFHV LQ SUREOHP VROYLQJ SHUIRUPDQFHr WR ILQG RXW LI VLJQLILFDQW GLIIHUHQFHV SHUVLVWHG DIWHU FRQWUROOLQJ IRU FRPSXWDWLRQDO VNLOOVr DQG LI WKH GLIIHUHQFHV GHSHQGHG RQ WKH OHYHO RI FRPSXWDWLRQDO VNLOOV 7KH UHVXOWV VKRZHG WKDW IHPDOHV RXWSHUIRUPHG PDOHV LQ SUREOHP VROYLQJ DQG LQ VL[ RI WKH VHYHQ FRPSXWDWLRQDO VNLOOV 0DOHV VKRZHG VXSHULRULW\ LQ HTXLYDOHQFH LQ HDFK RI WKH WKUHH \HDUVr EXW VWDWLVWLFDO VLJQLILFDQFH ZDV REWDLQHG RQO\ IRU WKH VHFRQG \HDU 7KH UHVXOWV DOVR VKRZHG WKDW IHPDOHV UHWDLQHG WKHLU VXSHULRULW\ LQ SUREOHP VROYLQJr RQO\ ZKHQ HTXLYDOHQFH IRU WKH WKUHH \HDUV GDWDf DQG VXEWUDFWLRQ IRU WKH VHFRQG \HDUf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f DQG )HQQHPD DQG 7DUWUHH f ZHUH QRW VXSSRUWHG LQ WKLV LQYHVWLJDWLRQ )HPDOH SHUIRUPDQFH ZDV HTXDO RU KLJKHU WKDQ PDOH SHUIRUPDQFH LQ SUREOHP VROYLQJ 7KH ILQGLQJV RI WKH VWXG\ KDYH VRPHWKLQJ LQ FRPPRQ ZLWK RWKHU LQYHVWLJDWLRQV ZKHUH PDWKHPDWLFV SUREOHP VROYLQJ KDV EHHQ WKH VXEMHFW RI LQWHUHVW WKH IDFW WKDW PDOHV DQG IHPDOHV GR SRRUO\ LQ ZRUGVWRU\ SUREOHP LWHPV SUREOHP VROYLQJf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ÂH[UBUHODWHG'LIIHUHQFHV LQ 3UREOHP 6ROYLQJ 3HUIRUPDQFH 7KH UHVXOWV RI WKLV VWXG\ GR QRW JLYH VWURQJ VXSSRUW WR 0DUVKDOOnV f ILQGLQJV WKDW WKH HIIHFW RI FRPSXWDWLRQDO VNLOOV RQ SUREOHP VROYLQJ SHUIRUPDQFH LV GLIIHUHQW IRU HDFK VH[ +RZHYHU WKH ILQGLQJV RI WKH SUHVHQW LQYHVWLJDWLRQ DUH FRQVLVWHQW ZLWK WKRVH IRXQG E\ 0H\HU f DQG :KLWDNHU f IHPDOH SHUIRUPDQFH LQ SUREOHP VROYLQJ LV QRW VLJQLILFDQWO\ GLIIHUHQW IURP PDOH SHUIRUPDQFH )HPDOHV VKRZHG VXSHULRULW\ LQ SUREOHP VROYLQJ SHUIRUPDQFH LQ DOO WKUHH \HDUV RI WHVW DGPLQLVWUDWLRQ DOWKRXJK WKHLU VXSHULRULW\ ZDV UHWDLQHG RQO\ ZKHQ SHUIRUPDQFH LQ HTXLYDOHQFH ZDV FRQWUROOHG IRU IRU WKH WKUHH \HDUV GDWDf DQG ZKHQ SHUIRUPDQFH LQ VXEWUDFWLRQ ZDV FRQWUROOHG IRU VHFRQG \HDUf VWDWLVWLFDOO\ VLJQLILFDQW RQO\ DIWHU DFFRXQWLQJ IRU HTXLYDOHQFHf 7KH ILQGLQJV RI WKLV VWXG\ DUH DOVR LQ DJUHHPHQW ZLWK WKRVH UHSRUWHG E\ WKH 1DWLRQDO $VVHVVPHQW RI (GXFDWLRQDO 3URJUHVV 1$(3f &DUSHQWHU HW DO] f 5HODWHG WR SUREOHP VROYLQJ WKH\ FRQFOXGH LI LW ZHUH QHFHVVDU\ WR VLQJOH RXW RQH DUHD WKDW GHPDQGV XUJHQW DWWHQWLRQ LW ZRXOG FOHDUO\ EH SUREOHP VROYLQJ $W DOO DJH OHYHOV DQG LQ YLUn WXDOO\ HYHU\ FRQWHQW DUHD SHUIRUPDQFH ZDV H[WUHn PHO\ ORZ RQ H[HUFLVHV UHTXLULQJ SUREOHP VROYLQJ RU DSSOLFDWLRQ RI PDWKHPDWLFDO VNLOOV ,Q JHQHn UDO UHVSRQGHQWV GHPRQVWUDWHG D ODFN RI WKH PRVW EDVLF SUREOHPVROYLQJ VNLOOV 5DWKHU WKDQ DWWHPSWn LQJ WR WKLQN WKURXJK D SUREOHP DQG ILJXUH RXW ZKDW QHHGHG WR EH GRQH WR VROYH WKH SUREOHP PRVW UHVn SRQGHQWV VLPSO\ WULHG WR DSSO\ D VLQJOH DULWKPHWLF PAGE 245 RSHUDWLRQ WR WKH QXPEHUV LQ WKH SUREOHP 7KH UHn VXOWV LQGLFDWH WKDW VWXGHQWV DUH QRW IDPLOLDU ZLWK VXFK EDVLF SUREOHPVROYLQJ VWUDWHJLHV DV GUDZLQJ D SLFWXUH RI D ILJXUH GHVFULEHG LQ D SUREOHP RU FKHFNLQJ WKH UHDVRQDEOHQHVV RI D UHVXOW S f 7KH IDFW WKDW LQ JHQHUDO} DIWHU FRQWUROOLQJ IRU FRPSXWDWLRQDO VNLOOV} WKHUH ZHUH QR VWDWLVWLFDOO\ VLJQLILFDQW GLIIHUHQFHV EHWZHHQ WKH SUREOHP VROYLQJ SHUIRUPDQFH RI WKH PDOHV DQG IHPDOHV ZKR SDUWLFLSDWHG LQ WKLV VWXG\ ZRXOG VHHP WR LPSO\ WKDW WHDFKHUV QHHG QRW ZRUU\ DERXW GHVLJQLQJ GLIIHUHQW WHDFKLQJ VWUDWHJLHV IRU WKH VH[HV +RZHYHUA EHFDXVH WKH SHUIRUPDQFH LQ SUREOHPVROYLQJ LV ORZ IRU ERWK VH[HV} WHDFKHUV PD\ EH HQFRXUDJHG WR HPSKDVL]H SUREOHP VROYLQJ LQ WKHLU GDLO\ WHDFKLQJ 7KLV VWXG\ ZDV GHVLJQHG WR LQYHVWLJDWH VH[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFDO SHUIRUPDQFH 7KH ILQGLQJV FOHDUO\ GHPRQVWUDWHG WKDW WK JUDGH IHPDOHV DQG PDOHV IURP WKH SXEOLF VFKRROV LQ 3XHUWR 5LFR DUH HTXDOO\ JRRG RU HTXDOO\ EDG LQ WKH VROXWLRQ RI ZRUGVWRU\ LWHPV SUREOHP VROYLQJf $OVR LQ JHQHUDO WKH HIIHFW RI FRPSXWDWLRQDO VNLOOV LV VLPLODU IRU HDFK VH[ %DVHG RQ WKHVH ILQGLQJV LW DSSHDUV WKDW JUHDWHU GLIIHUHQFHV H[LVW ZLWKLQ HDFK VH[ WKDQ EHWZHHQ WKH VH[HV ,Q FRQFOXVLRQ WKH ILQGLQJV RI WKH SUHVHQW VWXG\ KLJKOLJKW WKH QHHG IRU LQYHVWLJDWLRQV RI GLIIHUHQW DSSURDFKHV WR LQVWUXFWLRQDO GHVLJQV DLPHG DW LPSURYLQJ ERWK PDOH DQG IHPDOH SHUIRUPDQFH LQ SUREOHP VROYLQJ PAGE 246 5()(5(1&(6 $LNHQ / 5 f $WWLWXGHV WRZDUG PDWKHPDWLFV 5HYLHZ RI (GXFDWLRQDO 5HVDUFK ,4 $OOHQ 5 + t &KDPEHUV / f $ FRPSDULVRQ RI WKH PDWKHPDWLFV DFKLHYHPHQW RI PDOHV DQG IHPDOHV %XOOHWLQ 1R f 0DGLVRQ :LVFRQVLQ 6WDWH 'HSDUWPHQW RI 3XEOLF ,QVWUXFWLRQ 'LYLVLRQ IRU 0DQDJHPHQW DQG 3ODQQLQJ 6HUYLFH (5,& 'RFXPHQW 5HSURGXFWLRQ 6HUYLFHV 1R (' f $UPVWURQJ 0 f $FKLHYHPHQW DQG SDUWLFLSDWLRQ RI :SPHBQBLQBPDWKHPDWLeVM Â/QB4B\HM\LH: 5HSRUW RI D WZR \HDU VWXG\f 'HQYHU &2 (GXFDWLRQDO &RPPLVVLRQ RI WKH 6WDWHV 1DWLRQDO $VVHVVPHQW RI (GXFDWLRQDO 3URJUHVV (5,& 'RFXPHQW 5HSURGXFWLRQ 6HUYLFH 1R (' f $UPVWURQJ 0 t 3ULFH 5 $ f &RUUHODWHV DQG SUHGLFWRUV RI ZRPHQnV PDWKHPDWLFV SDUWLFLSDWLRQ -RXUQDO IRU 5HVHDUFK LQ 0DWKHPDWLFV (GXFDWLRQ $VWLQ + 6 f 2YHUYLHZ RI WKH ILQGLQJV ,Q + $VWLQ + 6XQLHZLFN t 6 'ZHFN (GVf :RPHQ $ %LEOLRJUDSK\ RQ WKHLU (GXFDWLRQ DQG &DUHHUV SS f 1HZ PAGE 247 %HUQr 6 / r t %HUQ 9 f :H DUH DOO QRQFRQVFLRXV VH[LVWV 3V\FKRORJ\ 7RGD\V r %HQERZr & 3r t 6WDQOH\r & f 6H[ GLIIHUHQFHV LQ PDWKHPDWLFDO DELOLW\ )DFW RU DUWLIDFW" 6FLHQFHr r %HQERZr & 3r t 6WDQOH\r & f &RQVHTXHQFHV LQ KLJK VFKRRO DQG FROOHJH RI VH[ GLIIHUHQFHV LQ PDWKHPDn WLFDO UHDVRQLQJ DELOLW\ $ ORQJLWXGLQDO SHUVSHFWLYH $PHULFDQ (GXFDWLRQDO 5HVHDUFK -RXUQDOr r %REELHr & 1 f 6H[UROH SUHIHUHQFH DQG DFDGHPLF DFKLHYHPHQW 'RFWRUDO GLVVHUWDWLRQr PAGE 248 &UHVZHOO / )HEUXDU\f 6H[UHODWHG GLIIHUHQFHV LQBSUREOHP VROYLQJ LQ UXUDO EODFN DQJOR DQG &KLFDJR DGROHVFHQWV 3DSHU SUHVHQWHG DW WKH DQQXDO PHHWLQJ RI WKH 6RXWKZHVW (GXFDWLRQDO 5HVHDUFK $VVRFLDWLRQ $XVWLQ 7; (5,& 'RFXPHQW 5HSURGXFWLRQ 6HUYLFH 1R (' f 'Z\HU & $ f ,QIOXHQFH RI FKLOGUHQnV VH[ UROH VWDQGDUGV RQ UHDGLQJ DQG DULWKPHQWLF DFKLHYHPHQW -RXUQDO RI (GXFDWLRQDO 3V\FKRORJ\ ee (FFOHV f 6H[ GLIIHUHQFHV LQ PDWKHPDWLFV SDUWLFLSDWLRQ ,Q 0 6WHLQNDPS t 0 0DHKU (GVf :RPHQ LQ 6FLHQFH SS f *UHHQZLFK &7 -$, 3UHVV (OWRQ & ) t 5RVH + $ f 7UDGLWLRQDO VH[ DWWLWXGHV DQG GLVFUHSDQW DELOLW\ PHDVXUHV LQ FROOHJH ZRPHQ -RXUQDO RI &RXQVHOLQJ 3V\FRORJ\ 0r (UQHVW f 0DWKHPDWLFV DQG 6H[ 6DQWD %DUEDUD 8QLYHUVLW\ RI &DOLIRUQLD ([HGLVLV] 5 + f $Q LQYHVWLJDWLRQ RI WKH UHODWLRQn VKLS RI UHDGLQJ FRPSUHKHQVLRQ YRFDEXODU\ PDWKHPDWLFDO FRQFHSWV DQG FRPSXWDWLRQ RQ SUREOHP VROYLQJ DPRQJ DQJOR EODFN DQG &KLFDJR PDOH DQG IHPDOH PLGGOH VFKRRO DGROHVFHQWV 'RFWRUDO 'LVVHUWDWLRQ 8QLYHUVLW\ RI +RXVWRQ f 'LVVHUWDWLRQ$EVWUDFWV ,QWHUQDWLRQDO $$ )HQQHPD ( f 0DWKHPDWLFV OHDUQLQJ DQG WKH VH[HV $ UHYLHZ -RXUQDO IRU 5HVHDUFK LQ 0DWKHPDWLFV (GXFDWLRQ Â )HQQHPD ( f 0DWKHPDWLFV VSDWLDO DELOLW\ DQG WKH VH[HV ,Q ( )HQQHPD (Gf 0DWKHPDWLFV :KDW UHVHDUFK VDYV DERXW VH[ GLIIHUHQFHV SS f &ROXPEXV 2KLR 6WDWH 8QLYHUVLW\ &ROOHJH RI (GXFDWLRQ (5,& 'RFXPHQW 5HSURGXFWLRQ 6HUYLFH 1R (' f )HQQHPD ( f ,QIOXHQFHV RI VHOHFWHG FRJQLWLYH DIIHFWLYH DQG HGXFDWLRQDO YDULDEOHV RQ VH[UHODWHG GMIIBH[JQFHVAQBPDWKHPDWLFV OHDUQLQJ DQG VWXG\LQJ 0DGLVRQ 8QLYHUVLW\ RI :LVFRQVLQ 'HSDUWPHQW RI &XUULFXOXP DQG ,QVWUXFWLRQ (5,& 'RFXPHQW 5HSURGXFWLRQ 6HUYLFH 1R (' f PAGE 249 )HQQHPDU ( r t 6KHUPDQU f 6H[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFV DFKLHYHPHQWV VSDWLDO YLVXDOL]DWLRQV DQG DIIHFWLYH IDFWRUV $PHULFDQ (GXFDWLRQDO5HVHDUFK -RXUQDOV n )HQQHPDV (V t 6KHUPDQV $ f 6H[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFV DFKLHYHPHQW DQG UHODWHG IDFWRUV $ IXUWKHU VWXG\ -RXUQDO IRU 5HVHDUFK LQ 0DWKHPDWLFV (GXFDWLRQ V )HQQHPDV (V t 7DUWUHHV / $ f 5HVHDUFK RQ UHODWLRQVKLS RI VSDWLDO YLVXDOL]DWLRQ DQG FRQILGHQFH WR PDOHIHPDOH PDWKHPDWLFV DFKLHYHPHQW LQ JUDGHV :DVKLQJWRQV '& 1DWLRQDO 6FLHQFH )RXQGDWLRQ (5,& 'RFXPHQW 5HSURGXFWLRQ 6HUYLFH 1R (' f )LHQEHUJV 6 ( f 7KH $QDO\VLV RI &URVV&ODVVLILHG &DWHJRULFDO 'DWD &DPEULGJHV 0$ 7KH 0,7 3UHVV )R[V / + Df 0DWKHPDWLFDOO\ SUHFRFLRXV 0DOH RU IHPDOH ,Q ( )HQQHPD (Gf 0DWKHPDWLFV OHDUQLQJ :KDW UHVHDUFK VD\VDERXW VH[ GLIIHUHQFHV SS f &ROXPEXV 2KLR 6WDWH 8QLYHUVLW\V &ROOHJH RI (GXFDWLRQ (5,& 'RFXPHQW 5HSURGXFWLRQ 6HUYLFH 1R (' f )R[V / + Ef 6H[ GLIIHUHQFHV ,PSOLFDWLRQV IRU SURJUDP SODQQLQJ IRU WKH DFDGHPLFDOO\ JLIWHG 3DSHU SUHVHQWHG DW WKH /HZLV 0 7HUPDQ 0HPRULDO 6\PSRVLXP RQ ,QWHOOHFWXDO 7DOHQW KHOG DW WKH -RKQV +RSNLQV 8QLYHUVLW\r %DOWLPRUHr 0' )R[r / + f 7KH HIIHFWV RI VH[ UROH VRFLDOL]DWLRQ RQ PDWKHPDWLFV SDUWLFLSDWLRQ DQG DFKLHYHPHQW ,Q 6KRHPDNHU (Gfr :RPHQBDQGB0DWKHPDWLFV5HVHDUFK 3HUVSHFWLYHV IRU &KDQJHr 1,( 3DSHUV LQ (GXFDn WLRQ DQG :RUN 1R f :DVKLQJWRQr & (GXDWLRQ DQG :RUN *URXSr 7KH 1DWLRQDO ,QVWLWXWH RI (GXFDWLRQ *OHQQRQr 9 -r t &DOODKDQr / f (OHPHQWDU\ VFKRRO PDWKHPDWLFV $BJXLGHBWSBFXU UHQW UHVHDUFKr :DVKLQJWRQr & $VVRFLDWLRQ IRU 6XSHUYLVLRQ DQG &XUULFXOXP 'HYHORSPHQW *XD\r 5 %r t 0F'DQLHOr f 7KH UHODWLRQVKLS EHWZHHQ PDWKHPDWLFV DFKLHYHPHQW DQG VSDFH DELOLW\ DPRQJ HOHPHQWDU\ VFKRRO FKLOGUHQ -RXUQDO IRU 5HVHDUFK LQ 0DWKHPDWLFV (GXFDWLRQr r PAGE 250 ,OO +DQVHQ & : f )DFWRUV DVVRFLDWHG ZLWK VXFFHVVIXO DFKLHYHPHQW LQ SUREOHP VROYLQJ LQ VL[WK JUDGH DULWKPHQWLF -RXUQDO RI (GXFDWLRQDO 5HVHDUFK +DUQLVFK / f ,WHP UHVSRQVH SDWWHUQV $SSOLFDWLRQ IRU HGXFDWLRQDO SUDFWLFH -RXUQDO RI (GXFDWLRQDO 0HDVXUHPHQW 6/n +LOWRQ 7 / t %HUJOXQG : f 6H[ GLIIHUHQFHV LQ PDWKHPDWLFV DFKLHYHPHQW $ ORQJLWXGLQDO VWXG\ -RXUQDO RI (GXFDWLRQDO 5HVHDUFK e +XQWHU ( t &RKHQ 6 + f &RUUHFWLQJ IRU XQUHOLDELOLW\ LQ QRQOLQHDU PRGHOV RI DWWLWXGH FKDQJH 3V\FKRPHWULND -DUYLV 7 f %R\JLUO DELOLW\ GLIIHUHQFHV LQ HOHPHQWDU\ VFKRRO DULWKPHQWLF 6FKRRO 6FLHQFH DQG 0DWKHPDWLFV .DXIPDQ $ f 6H[ GLIIHUHQFHV LQ PDWKHPDWLFV UHDVRQLQJ LQ PDWKHPDWLFDOO\ JLIWHG DQG DYHUDJH ILIWK DQG VL[WK JUDGHUV 'RFWRUDO GLVVHUWDWLRQ +RIVWUD 8QLYHUVLW\ f 'LVVHUWDWLRQ $EVWUDFWV ,QWHUn QDWLRQDO $$ .LOSDWULFN f 3UREOHP VROYLQJ DQG FUHDWLYH EHKDYLRU LQ PDWKHPDWLFV ,Q : :LOVRQ t / 5 &DUU\ (GVf 6WXGLHV LQ PDWKHPDWLFV 9RO 1R f 6WDQIRUG &$ 6FKRRO 0DWKHPDWLFV 6WXG\ *URXS .ORRVWHUPDQ 3 : f $WWULEXWLRQDO WKHRU\ OHDUQHG KHOSOHVVQHVV DQG DFKLHYHPHQW LQ QLQWK JUDGH PDWKHPDWLFV 'RFWRUDO GLVVHUWDWLRQ 8QLYHUVLW\ RI :LVFRQVLQ f 'LVVHUWDWLRQ $EVWUDFWV ,QWHUQDWLRn QDO e $$ .QLIRQJ t +ROWDQ % f $Q DQDO\VLV RI FKLOGUHQnV ZULWWHQ VROXWLRQV WR ZRUG SUREOHPV -RXUQDO IRU 5HVHDUFK LQ 0DWKHPDWLFVB(GXFDWLRQ .QLIRQJ t +ROWDQ % f $ VHDUFK IRU UHDGLQJ GLIILFXOWLHV DPRQJ HUUHG ZRUG SUREOHPV -RXUQDO IRU 5HVHDUFK LQ 0DWKHPDWLFV (GXFDWLRQ e PAGE 251 /DQGD8I 0 f 7KH HIIHFWV RI VSDWLDO DELOLW\ DQG SUREOHP SUHVHQWDWLRQ IRUPDW LQ PDWKHPDWLFDO SUREOHP VROYLQJ SHUIRUPDQFH RQ PLGGOH VFKRRO VWXGHQWV 'RFWRUDO GLVVHUWDWLRQr 1RUWKZHVWHUQ 8QLYHUVLW\f 'LVVHUWDWLRQ $EVWUDFWV ,QWHUQDWLRQDO" Â $ $ /LQGJUHQ + & 6LOYD )DUDFR} t 'D5RFKD 1 6 f $WWLWXGHV WRZDUG SUREOHP VROYLQJ DV D IXQFWLRQ RI VXFFHVV LQ DULWKPHWLF LQ %UD]LOLDQ HOHPHQWDU\ VFKRROV -RXUQDO RI (GXFDWLRQDO 5HVHDUFK Â 0DFFRE\ ( ( f 6H[ GLIIHUHQFHV LQ LQWHOOHFWXDO IXQFWLRQLQJ ,Q ( ( 0DFFRE\ (Gf 7KH 'HYHORSPHQW RI 6H[ 'LIIHUHQFHV SS f 6WDQIRUG &$ 7KH 6WDQIRUG 8QLYHUVLW\ 3UHVV 0DFFRE\ ( ( t -DFNOLQ & 1 f 6H[ GLIIHUHQFHV LQ LQWHOOHFWXDO IXQFWLRQLQJ ,Q $ $QDVWDVL (Gf $VVHVVPHQW LQ D SOXUDOLVWLF VRFLHW\ SS f 3URFHHGLQJV RI WKH ,QYLWDWLRQDO &RQIHUHQFH RQ 7HVWLQJ 3UREOHPV (GXFDWLRQDO 7HVWLQJ 6HUYLFH 3ULQFHWRQ 10DFFRE\ ( ( t -DFNOLQ & 1 f 7KH 3V\FKRORJ\ RI 6H[ 'LIIHUHQFHV 6WDQIRUG &$ 6WDQIRUG 8QLYHUVLW\ 3UHVV 0DUVKDOO 6 3 f 6H[ 'LIIHUHQFHV LQ 6L[WK *UDGH &KLOGUHQnV 3UREOHP 6ROYLQJ 3DSHU SUHVHQWHG DW WKH DQQXDO PHHWLQJ RI WKH $PHULFDQ (GXFDWLRQDO 5HVHDUFK $VVRFLDWLRQ /RV $QJHOHV &$ 0DUVKDOO 6 3 f 6H[ GLIIHUHQFHV LQ PDWKHPDWLFDO HUURUV $QDO\VLV RI GLVWUDFWRU FKRLFH -RXUQDO IRU 5HVHDUFK LQ 0DWKHPDWLFV (GXFDWLRQ 0DUVKDOO 6 nS f 6H[ GLIIHUHQFHV LQ FKLOGUHQnV PDWKHPDWLFV DFKLHYHPHQW 6ROYLQJ FRPSXWDWLRQV DQG VWRU\ SUREOHPV -RXUQDO RI (GXFDWLRQDO 3V\FKRORJ\ O$I 0DUWLQ 0 f 5HDGLQJ FRPSUHKHQVLRQ DEVWUDFW YHUEDO UHDVRQLQJ DQG FRPSXWDWLRQ DV IDFWRUV LQ DULWKPHWLF SUREOHP VROYLQJ 'RFWRUDO GLVVHUWDWLRQ 8QLYHUVLW\ RI ,RZD f 'LVVHUWDWLRQ $EVWUDFWV ,QWHUQDWLRQDO $$ PAGE 252 0H\HU" 5 $ f 6H[UHODWHG GLIIHUHQFHV LQ PDWKHPDn WLFDO SUREOHP VROYLQJ SHUIRUPDQFH DQG LQWHOOHFWXDO DELOLWLHV 0DGLVRQ 8QLYHUVLW\ RI :LVFRQVLQ" 5HVHDUFK DQG 'HYHORSPHQW &HQWHU IRU &RJQLWLYH /HDUQLQJ (5,& 'RFXPHQW 5HSURGXFWLRQ 6HUYLFH 1R (' f 0LOWRQ" $ f 7KH HIIHFWV RI VH[UROH LGHQWLILFD FDWLRQ XSRQ SUREOHP VROYLQJ VNLOO -RXUQDO RI $EQRUPDO t 6RFLDO3V\FKRORJ\" Â" 0ROOHU" 1 f 7KH LPSDFW RI JHQGHU" PDVFXOLQLW\" DQG IHPLQLW\ RQ PDWK DFKLHYHPHQW DQG FRXUVH GHFLVLRQV 'RFWRUDO GLVVHUWLRQ" 3XUGXH 8QLYHUVLW\" f QLVBVHUWDWLRQB$EVWUDFWV ,QWHUQDWLRQDO" $ 0XWK" f 6ROYLQJ DULWKPHWLF ZRUG SUREOHPV 5ROH RI UHDGLQJ DQG FRPSXWDWLRQDO VNLOOV -RXUQDO RI (GXFDWLRQDO 3V\FKRORJ\" 3DOODV" $ 0" t $OH[DQGHU" / f 6H[ GLIIHUHQFHV LQ TXDQWLWDWLYH 6$7 SHUIRUPDQFH 1HZ HYLGHQFH RQ WKH GLIIHUHQWLDO FRXUVHZRUN K\SRWKHVLV $PHULFDQ (GXFDWLRQDO 5HVHDUFK -RXUQDO" 3ODQN" (" t 3ODQN" 5 f (PRWLRQDO FRPSRQHQWV LQ DULWKPHWLFDO OHDUQLQJ DV VHHQ WKURXJK DXWRELRJUDSKLHV 3V\FKRORJLFDO 6WXGLHV RI WKH &KLOG" 3RUWHU" $ & f 7KH HIIHFWV RI XVLQJ IDOOLEOH YDULDEOHV LQ WKH DQDO\VLV RI FRYDULDQFH 'RFWRUDO GLVVHUWDWLRQ" 8QLYHUVLW\ RI :LVFRQVLQ" f 'LVVHUWDWLRQ $EVWUDFW ,QWHUQDWLRQDO" e" % 3UREHUW" % 6 f 0DWK FRQILGHQFH ZRUNVKRSV $ PXOWLPRGDO JURXS LQWHUYHQWLRQ VWUDWHJ\ LQ PDWKHPDWLFV DQ[LHW\DYRLGDQFH 'RFWRUDO GLVVHUWDWLRQ" 8QLYHUVLW\ RI )ORULGD" f 'LVVHUWDWLRQ $EVWUDFWV ,QWHUQDWLRQDO" % 3XHUWR 5LFR 'HSDUWPHQW RI (GXFDWLRQ f %DVLF VNLOOV WHVW LQ PDWKHPDWLFV +DWR 5H\ &HQWHU RI (YDOXDWLRQ 5DGDW]" + f (UURU DQDO\VLV LQ PDWKHPDWLFV HGXFD WLRQ -RXUQDO IRU 5HVHDUFK LQ 0DWKHPDWLFV (GXFDn WLRQ " fÂ§ 5LHGHVHO" $ & f 3UREOHP VROYLQJ 6RPH VXJJHVWLRQV IURP UHVHDUFK $ULWKPHWLF 7HDFKHU" PAGE 253 6KHUPDQ f 3UHGLFWLQJ PDWKHPDWLFV SHUIRUPDQFH LQ KLJK VFKRRO JLUOV DQG ER\V -RXUQDO RI (GXFDWLRQDO 3V\FKRORJ\ 6KHUPDQ f 0DWKHPDWLFV VSDWLDO YLVXDOL]DWLRQ DQG UHODWHG IDFWRUV &KDQJHV LQ JLUOV DQG ER\V JUDGHV -RXUQDO RI (GXFDWLRQDO 3V\FKRORJ\ 6WHLQ $ + f 7KH HIIHFWV RI VH[ UROH VWDQGDUGV IRU DFKLHYHPHQW DQG VH[ UROH SUHIHUHQFH RQ WKUHH GHWHUPLQDQWV RI DFKLHYHPHQW PRWLYLDWLRQ 'HYHORSPHQWDO 3V\FKRORJ\ 6WHLQ $ + t 6PLWKOHVV f $JH DQG VH[ GLIIHUHQFHV LQ FKLOGUHQnV VH[ UROH VWDQGDUGV DERXW DFKLHYHPHQW 'HYHORSPHQWDO 3V\FKRORJ\ 6X\GDP 0 1 f 5HVHDUFK RQ PDWKHPDWLFV HGXFDWLRQ .f UHSRUWHG LQ -RXUQDO IRU 5HVHDUFK LQ 0DWKHPDWLFV (GXFDWLRQ =! 6X\GDP 0 1 f 5HVHDUFK RQ PDWKHPDWLFV HGXFDWLRQ .f UHSRUWHG LQ -RXUQDO IRU 5HVHDUFK LQ 0DWKHPDWLFV (GXFDWLRQ 6X\GDP 0 1 f 5HVHDUFK RQ PDWKHPDWLFV HGXFDWLRQ .f UHSRUWHG LQ -RXUQDO IRU 5HVHDUFK LQ 0DWKHPDWLFV (GXFDWLRQ 6X\GDP 0 1 f 5HVHDUFK RQ PDWKHPDWLFV HGXFDWLRQ .f UHFRUGHG LQ -RXUQDO IRU 5HVHDUFK LQ 0DWKHPDWLFV (GXFDWLRQ Â 6X\GDP 0 1 t :HDYHU ) f 5HVHDUFK RQ PDWKHPDWLFV HGXFDWLRQ .f UHFRUGHG LQ -RXUQDO IRU 5HVHDUFK LQ 0DWKHPDWLFV (GXFDWLRQ e 9HUEHNH $ f 6H[UHODWHG GLIIHUHQFHV LQ PDWKHPDWLFDOO\ JLIWHG VHFRQGDU\ VWXGHQWV $Q LQYHVWLJDWLRQ RI VHOHFWHG FRJQLWLYH DIIHFWLYH DQG HGXFDWLRQDO IDFWRUV 'RFWRUDO GLVVHUWDWLRQ 8QLYHUVLW\ RI 0DU\ODQG f 'LVVHUWDWLRQ $EVWUDFWV ,QWHUQDWLRQDO $$ :HLQHU % f $FKLHYHPHQW PRWLYDWLRQ DQG DWWULEXWLRQ WKHRU\ 0RUULVWRZQ 1*HQHUDO /HDUQLQJ 3UHVV PAGE 254 :KLWDNHU 5 f $ VWXG\ RI WKH UHODWLRQVKLS EHWZHHQ VHOHFWHG QRQFRJQLWLYH IDFWRUV DQG WKH SUREOHP VROYLQJ SHUIRUPDQFH RI IRXUWKJUDGH FKLOGUHQ 7HFK 5HS 1R f 0DGLVRQ 8QLYHUVLW\ RI :LVFRQVLQ 5HVHDUFK DQG 'HYHORSPHQW &HQWHU IRU &RJQLWLYH /HDUQLQJ :LVHU / .U 6WHHO} /U t 0DF'RQDOG & f 2ULJLQV DQG FDUHHU FRQVHTXHQFHV RI VH[ GLIIHUHQFHV LQ KLJK VFKRRO PDWKHPDWLFV DFKLHYHPHQW *UDQW 1R 1(,* f 3DOR $OWR &$ $PHULFDQ ,QVWLWXWHV IRU 5HVHDUFK =DOHZVNL / f $Q H[SORUDWRU\ VWXG\ WR FRPSDUH WZR SHUIRUPDQFH PHDVXUHV $Q LQWHUYLHZFRGLQJ VFKHPH RI PDWKHPDWLFDO SUREOHP VROYLQJ DQG D ZULWWHQ WHVW 'RFWRUDO 'LVVHUWDWLRQ 8QLYHUVLW\ RU :LVFRQVLQ 0DGLVRQf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r WR -XO\r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onsistent significant differences were also found for addition* multiplication* division* addition of fractions* and subtraction of decimals. For subtraction the difference was not statistically significant. Question 2: Do sex-related differences in problem solving persist when computational skills are controlled for* and is the male-female differences in problem solving dependent on level of computational skills? To address the question of dependence of male-female problem solving differences in computational skills for each year and computation subtest* the possibility of an interaction was investigated. For the first year* statistically significant interactions were found between sex and multiplication* F (1*998) = 8.59* p second year* F (1 *1009) = 6 .39 p<].05. No significant interactions were found in the third year. Analysis of covariance summary tables are shown as Tables 4.4* 4.5* and 4.6. Also* the three interactions are depicted in Figures 4.1* 4.2* and 4.3. Each figure indicates that at lower levels of computational skills* males outperformed females in problem solving* with the reverse happening at higher levels of computational skills. 98 TABLE 4.13 Adjusted Means on Problem Solving* by Covariate and Sex Third Year Covariate Adjusted Males Means Females Addition 4.09 4.18 Subtraction 3.98 4.29 Multiplication 4.10 4.17 Division 4.17 4.16 Addfrac 4.07 4.20 Subdec 4.15 4.12 Equivalence 3.79 4.45 18 students who had taken a similar number of mathematics courses and were in the top 85% of the class in mathematics achievement. They were tested in 1976. Four high school districts were included. In only one of the high school districts were sex-related differences in application and problem solving found# in favor of males. They concluded that when relevant factors are controlled/ sex-related differences in favor of males do not appear often/ and when they do/ they are not large. Sherman (1980) investigated the causes of the emerging sex-related differences in mathematics performance/ in favor of males/ during adolescence (grades 8-11). She wanted to know if these differences emerge as a function of sex-related differences in spatial visualization and sociocultural influences that consider math as a male domain. In grade 8/ she used the Romberg-Wearne Test and/ in grade 11/ a mathematical problem solving test derived from the French Kit of Tests. The analysis showed that for girls/ problem solving performance remained stable across the years. Mean problem solving performance for boys/ however/ was higher in grade 11 than in grade 8. No sex-related differences were found in grade 8/ but boys outperformed girls in grade 11/ where the Stafford test was used. Sherman found that for both sexes problem solving performance in grade 8 was the best predictor of problem 40 Achievement Test (subtests of reading/ arithmetic/ and reasoning) and the California Short-Form test of mental ability to a group of 1/400 children from the 6th grade. All levels of achievement were included in the analysis. Analysis of variance and covariance were used and compared. He confirmed the findings of other researchers to the effect that there is a direct relationship between I.Q. and reading ability/ and between I.Q. and computational skills. The results of the analysis of variance revealed that increases in computation ability were associated with higher achievement in problem solving. A relationship between reading ability and problem solving was also found/ but it was not as strong. Significant differences in problem solving performance associated with computational ability were found when intelligence was controlled. Balow concluded that computation is a much more important factor in problem solving than reading ability/ and that when I.Q. is taken into consideration/ the degree of the relationship between reading and problem solving ability becomes less pronounced. Intelligence tends to confound the relationship between these two variables. Knifong and Holtan (1976/ 1977) attempted to investigate the types of difficulties children have in solving word problems. They administered the word problem section of the Metropolitan Achievement Test to 35 children from the 6th grade. Errors were classified in two 74 TABLE 4.1 Continued Item Number Xa for Model 3.2* Actual Sample Size Min. Sample Size for Significance (a) ... (b) Cc) (d) 105 28.66 60/712 19/510 108 26.56 86/261 29/912 111 30.71 86/190 25/849 * p<.01 ** Minimum sample size less than 3/000 laL 59 60 61 62 63 64 65 66 67 68 69 70 71 73 74 76 77 78 79 72 TABLE 4.1 Continued X2 for Actual Sample Min. Sample Model 3.2* Size Size for Significance (b) (c) L) 50.96 41,829 7,560 66.43 60,494 8,387 21.39 35,632 15,342 55.03 50,699 8,485 158.36 52,068 3 ,028 88.01 73,196 7,660 56 .28 42,429 6 ,943 50.51 78,708 14,352 34.28 51,862 13,935 48.01 63,223 12,128 75.29 62,511 7 ,647 44.39 67 ,528 14,011 14,425.97 84 ,368 54** 24.61 53 ,358 19,969 22.75 84 ,175 34,077 99.68 70,183 6,485 79.38 66,304 7,693 64.56 76,806 10,957 23 .98 48,690 18,700 61 A problem arises in interpreting this analysis with the data used in this research. Over the three yearsr there were responses available from 135*340 students. Even if only 5% of the students answered an item incorrectly* the responses of 7*767 students would be used in analyzing this item. On the other hand* if 90% answered an item incorrectly* the responses of 121*806 students would be used in analyzing the data. As a result of the large sample size* the tests described above are likely to be very powerful. In step 1 of the analysis* then* even a very small change from year to year in the proportion of males or females who choose an option is apt to be detected* and the results will indicate that hypothesis 1 is not supported. For step 2* even a very small dependence of option choice on sex is likely to be detected and hypothesis 2 is likely to be supported. In brief* the problem caused by the large sample size is that practically insignificant differences may yield statistical significance. Fortunately* the form of the test statistic used in the likelihood ratio test suggests a reasonable solution to this problem. The test statistic is (3.4) 15 The test is composed of 23 items designed to yield 3 scores: a comprehension scorer an application scorer and a problem solving score. The comprehension question ascertains whether a child understands the information given in the item stem. The application question assesses the child's mastery of a prerequisite concept or skill of the problem solving question. The problem solving question poses a question whose solution is not immediately availabler that isr a situation which does not lend itself to the immediate application of a rule or algorithm. The application and problem solving parts of the test may refer to a common unit of information (the item stem) but the questions are independent in that the response to the application question is not used to respond to the problem solving question. Meyer (1978) / Whitaker (1976) / Fennema and Sherman (1978)/ and Sherman (1979) have used the Romberg-Wearne test in their studies. Meyer (1978) investigated whether males and females differ in problem solving performance and examined their prerequisite computational skills and mathematical concepts for the problem solving questions. A sample of 179 students from the 4th grade were administered 19 "reference tests" for intellectual abilities and the Romberg-Wearne test. The analysis showed that males and females were not significantly different in the comprehension/ application/ and problem solving questions 55 1/000 students were selected randomly for each year. For the first yearr 1/002 were selected (492 boys and 510 girls); for the second year/ 1/013 students were selected (504 boys and 509 girls); and/ for the third year/ 1/013 students were selected (509 boys and 504 girls). The student population in Puerto Rico includes children from the urban and rural zones and comprises children from low and middle socioeconomic levels. Findings can be generalized only to this population. The Instrument The Basic Skills Test in Mathematics-6 is a criter ion-referenced test used in the Department of Education of Puerto Rico as part of the annual assessment program. The test measures academic achievement in operations/ mathematical concepts/ and story problems. It has a reported split half reliability or .95. The test was designed specifically for Puerto Rico. Its contents and the procedures followed for its development were formulated and reviewed by educators from the mathematics department of the Department of Education of Puerto Rico/ in coordination with the Evaluation Center of the Department of Education and mathematics teachers from the school districts. The emphasis placed on each skill area is depicted in Figure 3.1. Problem Solving 88 9 8. 7- 5- 4- 3- 2. 1 r r t 1 2 3 Division Fig. 4.2 Sex by Division Interaction 4 5 6 5 Females were better on items of computation and males were more successful on word-story problem items (problem solving). She also found that females successful performance in the problem solving items was more dependent on their successful performance in the computation items. Males did not need as much as females to succeed in the computation items in order to answer correctly the problem solving items. Although the general findings seem to support sex differences in mathematics learning and achievement the research done does not consistently support superiority for either sex. Most of the research has been concerned with how the sexes differ on subtests or total test scores in mathematics. Moreover the great majority of the studies deal with correct responses. Sex differences in incorrect responses at the item level have not been fully researched. Only two studies dealing with sex differences in incorrect responses at the item level were found in the research literature (Marshall 1981 1983). Marshall investigated whether boys and girls made similar errors in computation and story problems. She analyzed boys' and girls' answers to six mathematics items and found that the sexes made different errors possibly reflecting different problem solving strategies. Her original findings were supported when she studied the same problem using a large number of items three years later. 70 tests were significant at the .01 level are reported in Table 4.1. Also reported in this table are the actual sample sizes and the minimum sample sizes necessary for the likelihood ratio tests to be significant. Of the 61 items# 59 had minimum sample sizes greater than 3#000. Thus# although for the three years both males and females samples had inconsistent option choices on these 61 items# on 59 of the items the inconsistency was relatively minor. Consequently# these 59 items were included in step 2 along with the initial 50 items. In step 2# a likelihood ratio test was performed comparing the adequacy of the models in (3.2) and (3.3). The X2 values associated with models (3.2) and (3.3) for each of the 109 items subjected to step 2 are reported in columns b and c of Table 4.2. Also reported in Table 4.2 is the difference between the two X2 values (see column d) This latter figure is the test statistic for comparing the adequacies of models (3.2) and (3.3). Significant X2 statistics are indicated by asterisks. The X2 statistics were significant for 100 of the items indicating a male-female difference in option choice for these items. In column e of Table 4.2 the actual sample sizes are reported. In column f# the minimum sample sizes necessary for significance are reported. For those 100 items which had significant X2 statistics reported in column d# 94 (94%) had minimum sample sizes greater than 3 #000 100 exception was that/ when equivalence was controlled/ sex-related differences in favor of females/ persisted. The question of whether male-female differences in problem solving depend on computational skills was answered partially/ in the affirmative. Interactions/ indicating the dependence/ were significant only for multiplication and division in the first year and for subtraction in the second year. 2 Many variables* cognitive* affective* and educational* have been investigated since 1974 in relation to sex differences in mathematics learning and achievement. Fennema and Sherman (1977) investigated the effect of differential formal mathematics education. After controlling for the number of years of exposure to the study of mathematics* they found sex differences in only two of the four schools under study. However* in those schools where boys scored higher than girls* differences were also found in their attitudes toward mathematics. Hilton and Berglund (1974) found significant sex differences after controlling for the number of mathematics courses taken* and attributed them to sex differences in interests. "As the boys' interests in science increase related to the girls'* their achievement in mathematics increases relative to that of the girls" (p. 234) . Wise* Steel* and McDonald (1979) reanalyzed test data collected in a longitudinal study of 400*000 high school students (Project Talent). They found that when the effect of the number of high school mathematics courses was not controlled* no sex differences emerged for 9th graders* but that gains made by boys during the next three years were more than twice that of the girls. These differences between the sexes disappeared when the number of mathematics courses taken was controlled. Results of the 1978 Women and Mathematics National Survey* Survey I* 13 III) to investigate sex related differences in the selection of incorrect responses/ and the consistency of such differences over three years of administration of the test. Based on her findings that sex differences were found in 80% of the items/ Marshall classified the students' errors according to Radatz' (1979) five-category error classification. The categories are language (errors in semantics)/ spatial visualization/ mastery/ association/ and use of irrelevant rules. It was found that girls' errors are more likely to be due to the misuse of spatial information/ the use of irrelevant rules/ or the choice of an incorrect operation. Girls also make relatively more errors of negative transfer and keyword association. Boys seem more likely than girls to make errors of perseverance and formula interference. Both sexes make language-related errors/ but the errors are not the same. Available research is not extensive enough to make definite judgments about the sex-related differences observed in incorrect responses. Clearly more research is needed. CHAPTER IV RESULTS Introduction The data gathered from 6th grade students from the public schools in Puerto Rico who took the Basic Skills Test in Mathematics-6 during three consecutive years were analyzed in this study. The first objective was to investigate whether boys and girls differ in the selection of incorrect responses, and if the pattern of differences was consistent throughout the three years in which the test was administered. The second objective was to investigate whether males and females differ in problem solving performance, if these differences persist after accounting for computational skills, and if the male-female differences depend on the level of computational skills. The results are discussed in two sections. Study findings in the area of sex differences in the selection of incorrect responses are discussed in the first section. The second section is devoted to an exposition of the findings in the area of sex-related differences in problem solving. Sex-Related Differences in the Selection of Incorrect Responses As indicated in Chapter III, there are two models of interest. The first model indicates that year and option 68 86 TABLE 4.6 ANCOVA Summary Table: Subtraction Covariate Second Year Source df SS MS F Subtraction (S) 1 93.14 593.14 122.80* Sex (S) 1 18.58 18.58 3.85* SxS 1 30.85 30.85 6.39* Error 1009 4873.40 4.83 *p<.05 25 emerged in spatial visualization scores became non-significant. In the two schools where sex differences in mathematics achievement were found/ differences between the sexes were also found in their attitudes toward mathematics. Researchers like Backman (1972) / who analyzed data from Project Talent/ and Allen and Chambers (1977) have also hypothesized that sex-related differences in mathematics achievement may be related to different curricula followed by males and females. Allen and Chambers attributed male superiority in mathematics problem solving to differences in the number of mathematics courses taken in high school. This issue has been seriously questioned by Astin (1974)/ Fox (1975a/ 1975b)/ and Benbow and Stanley (1980)/ among others. Astin and Fox have reported large differences in favor of males among gifted students taking the Scholastic Achievement Test. These differences occur as early as grade 7/ when there are no sex differences in the number of courses taken. Benbow and Stanley (1980) compared mathematically precocious boys and girls in the 7th grade/ with similar mathematics background/ and found sizeable sex-related differences favoring boys in mathematical reasoning ability. Five years later/ they conducted a follow-up study which showed that boys 95 TABLE 4.10 ANCOVA Summary Table: Other Covariates Second Year Source df ss MS F Addition 1 206.97 206.97 39.57* Sex 1 19.12 19.12 3.66 Error 1010 5277 .90 5.23 Multiplication 1 1026.68 1026.68 232.80* Sex 1 .80 .80 .18 Error 1010 4458.30 4.41 Division 1 1117 .46 1117.46 258.67* Sex 1 .64 .64 .15 Error 1010 4367.50 4.32 Fracadd 1 1341.13 1341.13 327.10* Sex 1 2.03 2.03 .50 Error 1010 4143.80 4.10 Decsub 1 1453 .91 1453.91 364.39* Sex 1 4.88 4.88 1.22 Error 1010 4031.10 3 .99 Equivalence 1 797.11 797.11 171.79* Sex 1 127.33 127.33 27.44* Error 1010 4687.90 4.64 * P<-05 25 emerged in spatial visualization scores became non-significant. In the two schools where sex differences in mathematics achievement were found differences between the sexes were also found in their attitudes toward mathematics. Researchers like Backman (1972) r who analyzed data from Project Talent and Allen and Chambers (1977) have also hypothesized that sex-related differences in mathematics achievement may be related to different curricula followed by males and females. Allen and Chambers attributed male superiority in mathematics problem solving to differences in the number of mathematics courses taken in high school. This issue has been seriously questioned by Astin (1974) Fox (1975a 1975b) and Benbow and Stanley (1980) among others. Astin and Fox have reported large differences in favor of males among gifted students taking the Scholastic Achievement Test. These differences occur as early as grade 7 when there are no sex differences in the number of courses taken. Benbow and Stanley (1980) compared mathematically precocious boys and girls in the 7th grade with similar mathematics background and found sizeable sex-related differences favoring boys in mathematical reasoning ability. Five years later they conducted a follow-up study which showed that boys skills* sex-related differences in problem solving performance do not exist. Females retained their superiority in problem solving when equivalence (in all three years) and subtraction (in one year) were the controlling variables. The question of whether male-female differences in problem solving depend on computational skills was answered* partially* in the affirmative. x 22 Approximately 270/000 students from the 6th grade were administered the Survey of Basic Skills of the California Assessment Program/ during the years 1977/ 1978/ and 1979. Responses were analyzed using log-linear models. Successful solving of computation items was positively associated with successful solving of story problems. Girls were more successful in computation than boys/ and boys were more successful than girls in solving story problems. This finding supports reports from the National Assessment of Educational Progress (NAEP) (Armstrong/ 1979). To investigate the effects of reading/ SES/ language/ and chronological age/ only those test forms containing 2 computation items and 2 story problems were considered for analysis; 32 items from 8 test forms were included in the analysis. The results of these analyses showed that at every level of reading score/ 6th grade children were more successful in computation than in story problems. Although the differences were not large/ at every reading score boys consistently had higher probabilities of success in story problems than did girls/ and girls consistently showed higher probabilities of success in computation than boys. Also/ as the reading score increased/ the difference between the probability of success in story problems and the probability of success in ACKNOWLEDGMENTS I would like to acknowledge ray gratitude to several people who have influenced my formal education and/or made this study possible. My special thanks to Dr. James Algina* Chairman of the doctoral committee* who contributed to the development of my love for research and statistics. He has been insuperable as professor and valued friend; his help and guidance in the preparation and completion of this study were invaluable. I extend my thanks to Dr. Linda Crocker for her advice and help during my doctoral studies at University of Florida. Thanks also go to Dr. Michael Nunnery* member of the doctoral committee. To Dr. Wilson Guertin* who was a friend for me and my family* I extend my special thanks. Thanks are also extended to Dr. Amalia Charneco* past Undersecretary of Education of the Puerto Rico Department of Education* for her continuous support. To my sister Nilda Santaella* who typed the thesis* I give my sincere thanks. Special thanks go to my family and to those friends who provided encouragement throughout this critical period of my life. ii 42 children whose errors fell under the category or "other errors." Students were asked to read each problem aloud and answer these questions: What kind of situation does the problem describe? What does the problem ask you to find? How would you work the problem? Ninety five percent of the students read the problem correctly; 98% explained the kind of situation the problem described in a correct manner; 92% correctly answered what the problem was asking them to find* and 36% correctly answered the question of how to work the problem. The fact that a large percent of the students whose errors were classified as "other errors" (in which reading skills might have been a factor) correctly stated how to work the problem is strong evidence of their ability to read and interpret the problems correctly. The errors made by this group of students had a distinct origin unrelated to reading ability. Zalewski (1974) investigated the relative contribution of verbal intelligence reading comprehension vocabulary interpretation of graphs and tables mathematical concepts number sentence selection and computation to successful mathematical word problem solution and the relationship of the dependent variable to the eight independent variables. She worked with a group of 4th grade children who were administered the subtests of the Iowa Test of Basic Skills (ITBS) and the Weschler Intelligence Scale for 64 gender-option dependency is the same from year to year. In this study/ a three dimensional table was considered to exhibit the same year to year pattern of gender-option dependency if / conditioned on gender/ the same proportion of students chose each incorrrect response over the three year period. This seems to be a straightforward and natural way to operationalize the concept. To illustrate how Marshall operationalized the concept in question/ a hypothetical set of probabilities was constructed conforming to the pattern specified by Marshall. This is displayed in Table 3.3. Table 3 .3 Hypothetical Joint Probabilities of Year/ Option/ and Sex Year Sex 1 Option 2 3 First M .120 .045 Â¡ .015 F .060 .060 Â¡ .030 Second M .144 .022 | .019 F .072 .030 Â¡ .039 Third M .132 .040 ! .016 F .066 .054 Â¡ J .039 Note: Probabilities reported are truncated to three places. 67 conducted/ one with each of addition/ subtraction/ multiplication/ division/ addition of fractions/ decimals in subtraction/ and equivalence as covariates. A well known problem that arises in the use of ANCOVA is that an unreliable covariate can cause spurious differences between the sex groups. To solve this problem/ Porter (1967/1968) proposed the use of estimated true scores for observed scores. Porter (1967/1968) conducted a simulation that gave empirical support to the adequacy of this strategy. Hunter and Cohen (1974) have provided theoretical support for this strategy. 104 to contribute to the body of knowledge regarding such differences in mathematical learning and achievement. This section of Chapter V comprises implications of the findings and recommendations evolving from the present investigation. Sex-related Differences in Incorrect Responses The main purpose of education is to impart formal education equally for the sexes. Analysis of the data on incorrect responses showed that for 100 of the 111 items of the test/ males and females selected different incorrect responses/ and this pattern of responses was consistently found during the three years of test administration. However/ with a sample of less than 3/000 subjects/ statistical significant results were obtained for only 6 items. Therefore/ it seems that/ at least for 6th grade children/ educators need not concern themselves with varying teaching techniques for the sexes. This study/ however/ has value apart from the investigation of sex-related differences. It was found that/ year after year/ 6th graders from the public schools in Puerto Rico tend to make the same errors in the solution of mathematical problems. It is important/ therefore/ that teachers provide the learning strategies necessary to enable students to obtain a better understanding of the concepts of numbers. Table Page 4.11 Adjusted Means on Problem Solving* by Covariate and Sex 96 4.12 ANCOVA Third Year 97 4.13 Adjusted Means on Problem Solving* by Covariate and Sex 98 vi 90 The significant interactions found between sex and multiplication* and sex and division (first year)* and sex and subtraction (second year)* answered* in part* the question of whether male-female differences in problem solving performance depend on computational ability. However* the evidence is quite weak. Of 21 possible interactions* only three were significant. No variable exhibited a significant interaction for each of the three years. The analysis of covariance was also used to determine if sex-related differences exist after controlling for computational skills. Analyses were conducted for those variables that did not exhibit significant interactions with sex. As discussed in Chapter III* estimated true scores were used for observed scores to adjust for unreliability of the covariates (the computational subtests). Reliability coefficients calculated for each covariate are shown in Table 4.7. Summaries of the analyses of covariance for the first year are reported in Table 4.8. The results show that females retained their superiority in problem solving performance when equivalence was the controlling variable in the analysis of covariance* the only variable in which males outperformed females (nonsignificant). When the controlling variables were addition* subtraction* addition of fractions* and subtraction with decimals* female 59 Inspection of the year by option contingency tables shows that year and option are independent for each gender. Thus/ hypothesis 1 is equivalent to the hypothesis that year and option are independent conditional upon gender. Hypothesis 2 is also true for Table 3.1 and Table 3.2. Therefore/ when hypothesis 1 is correct/ hypothesis 2 is equivalent to the hypothesis that sex and option choice are dependent. In his discussion of the analysis of three dimensional contingency tables/ Fienberg (1980) presents the following saturated model for the data: log m i j k = i + Uj + Uk+Uij + Uik+Ujk+U i j k. (3.1) In this model/ mijk is the expected value of the frequency in cell ijk of the three dimensional table. The model states that all three classification factors for a three dimensional contingency table are mutually dependent. In the present research i is the year index/ j is the option index/ and k is the sex index. Fienberg shows that deleting the terms Uij and Uijk yields a model in which year and option are independent conditional upon sex. This model is log mij k Ui + Uj + Uk+Uik+Uj k. (3.2) 75 TABLE 4.2 Chi-square Values for the Comparison of Models (3.2) and (3.3) with Actual Sample Sizes and with the Corresponding Number of Subjects Needed for Significant Results Item Num ber Xa- for Model 3.2 X1 for Model 3.3 Xa- for Difference Actual Sample Size Minimum Sample Size for Signif. (a) (b) (c) (d) (e) (f) 1 16.11 30.94 14.83* 7*427 10*061 2 17.54 103.22 85 .68* 11*018 2*583** 3 20.06 47.51 27.45* 11*497 8*733 4 7.54 21.87 14.33* 6*016 8*414 5 14.75 70.40 55.65* 7*706 2*782** 6 11.86 23 .13 11.27* 13*046 23 *256 7 4.60 35.73 31.13* 8*832 5*700 8 46.15 329.98 283.83* 16*796 1*189** 9 10.13 109.99 99.86* 23 *075 4*642 10 9.54 65.29 55.75* 35*060 12*634 11 5.89 19.06 13.17* 24*908 37 *996 12 12.58 198.46 185.88* 40 *423 4*369 13 25.74 62.37 36.63* 15*718 8*621 14 11.19 80 .50 69.31* 30*238 8*765 15 16.63 17.33 0.70 42*889 1*230*914 16 14.51 84.29 69.78* 35*114 10*109 17 17.90 44.20 26 .30* 43 *962 33 *582 41 categories. Category I included clerical and computational errors. Category II included other types of errors such as average and area errors/ use of wrong operation/ no response/ and erred responses offering no clues. It the student's work indicated the correct procedure and yet the problem was missed because of a computational or clerical error/ it was assumed that the problem was read and understood. An analysis of frequencies showed that clerical errors were responsible for 3% of the problems incorrectly solved/ computational errors accounted for 49%/ and other errors for 48% of the erred problems. Knifong and Holtan concluded that "improved computational skills could have eliminated nearly half ot the word problem errors" (p. Ill). These computational errors were made in a context where other skills such as reading/ interpretation of the problem/ and integration of these skills necessary for the solution of word problems/ might interact. However/ Knifong and Holtan state that their findings neither confirm nor deny that improvement ot reading skills will lead to improvement in problem solving. They conclude that "it is difficult to attribute major importance to reading as a source of failure" (p. 111). In a later analysis/ looking for evidence of poor reading abilities affecting children's success in word problems/ Knifong and Holtan (1977) interviewed the 69 are independent/ conditional upon sex* log m j j k = U + U j + U k+U k+U j k (3.2) Substantively/ this model implies that the pattern of male and female option choices is consistent over the three years of test administration. The second model indicates that option is independent of sex/ log mijk = Ui+Uj+Uk+Uik. (3.3) This model implies that the pattern of option choice is the same for males and females. In order to determine if males and females differed in the selection of incorrect responses and if these differences were stable across the three years of test administration/ a two-step test was performed. First/ a likelihood ratio test of the adequacy of model (3.2) was conducted for each of the 111 items of the test. Under this test a model fits the data when the value obtained is nonsignificant at a specified alpha level. For 50 (45%) of the items/ model (3.2) fitted the data adequately. For these items the pattern of male and female option choices was consistent for the three years. The X^ values for the 61 items for which the likelihood ratio 97 TABLE 4.12 ANCOVA Summary Table Third Year Source df ss MS F Addition 1 205.96 205.96 34.85* Sex 1 1.88 1.88 .32 Error 1010 5967.60 5.91 Subtraction 1 886.42 886.42 169.48* Sex 1 24.39 24.39 4.66* Error 1010 5287.2 5.23 Multiplication 1 1402.42 1402.42 297.12* Sex 1 1.05 1.05 .22 Error 1010 4771.20 4.72 Division 1 2028.49 2028.49 494.75* Sex 1 1.11 1.11 .27 Error 1010 4145.10 4.10 Fracadd 1 1733.72 1733.72 393.97* Sex 1 4.64 4.64 1.05 Error 1010 4439.89 4.40 Decsub 1 2116.84 2116.84 526.57* Sex 1 .15 .15 .04 Error 1010 4056.76 4.02 Equivalence 1 1186.92 1186.92 240.26* Sex 1 59.85 59.85 12.12* Error 1010 4986.69 4.94 * p<. 05 66 Table 3.4 Hypothetical Probabilities of Option Choice Conditional on Sex and Year Year Sex 1 Option 2 3 First M .666 .250 .083 F .400 .400 .200 Second M .774 .120 .104 F .510 .212 .276 Third M .698 .214 .087 F .431 .352 .215 Note: Probabilities reported are truncated to three places. Which procedure is more appropriate to test the claim that the pattern of male and female option choice remains the same from year to year? It seems more reasonable to test for a pattern like that in Table 3.2 than to test for a pattern like that in Table 3.4 / and/ consequently/ this was the strategy adopted in this study. Comparison of Males and Females in Problem Solving Performance One object of the study was to compare the performance of males and females on problem solving. Two questions were addressed. Firstr do males and females differ in problem solving performance? Second/ do these differences persist when computational skill is controlled for/ and do these differences depend on the level of computational skill? Seven analyses of covariance (ANCOVA) tests were 85 TABLE 4.4 ANCOVA Summary Table: Multiplication Covariate First Year Source df SS MS F Multiplication (M) 1 1195.90 1195.90 265.66* Sex (S) 1 31.95 31.95 7.07* M x S 1 38.80 38.80 8.59* Error 998 4509.00 4.51 * p <;.oi TABLE 4.5 ANCOVA Summary Table: Divison Covariate First Year Source df SS MS F Division (D) 1 1195.90 1195.90 264.66* Sex (S) 1 13.94 13.94 3.22 D x S 1 18.39 18.39 4.25* Error 998 4317.00 4.32 * p<.01 64 gender-option dependency is the same from year to year. In this study/ a three dimensional table was considered to exhibit the same year to year pattern of gender-option dependency if/ conditioned on gender/ the same proportion of students chose each incorrrect response over the three year period. This seems to be a straightforward and natural way to operationalize the concept. To illustrate how Marshall operationalized the concept in question/ a hypothetical set of probabilities was constructed conforming to the pattern specified by Marshall. This is displayed in Table 3.3. Table 3.3 Hypothetical Joint Probabilities of Year/ Option/ and Sex Year Sex 1 Option 2 3 First M .120 .045 | .015 F .060 .060 Â¡ .030 Second M .144 .022** .019 F .072 .030 ; .039 Third M .132 .040*1 .016 F .066 .054 Â¡ .039 Note: Probabilities reported are truncated to three places. 102 response was related to the use of a large number of subjects in the analysis. One must question whether analysis of incorrect responses has any educational value. Although the results of this study do not support the use of this analysis (error analysis) for the study of sex-related differences* it continues to be a promising technique in the discovery of how children approach mathematics problem solving* more so than analyses that look only at the items answered correctly or at total test scores. The second objective of the study was to investigate sex-related differences in problem solving performance* to find out if significant differences persisted after controlling for computational skills* and if the differences depended on the level of computational skills. The results showed that females outperformed males in problem solving and in six of the seven computational skills. Males showed superiority in equivalence in each of the three years* but statistical significance was obtained only for the second year. The results also showed that females retained their superiority in problem solving* only when equivalence (for the three years data) and subtraction (for the second year) were the controlling variables. Analysis of covariance was also used to answer the question of whether male-female differences in problem solving depend on the level of computational skills. The 12 test form/ no sex differences were found in response patterns. Approximately 45% of each sex selected option c. The next popular choice for both sexes was option d/ selected by approximately 35% of both boys and girls. On the story problem of Form If males and females responded alike. Their most popular incorrect response choice was option a for both males and females. The second most popular incorrect choice was option c for both sexes. Response to the story problem in Form 2 showed sex differences in response choice. Including the correct option/ 33% of the girls selected option a./ 20% chose option C/ and 20% option d. For males/ approximately 25% selected option a and the same percent selected option d. Marshall concluded that although the analysis of incorrect responses does not explain why boys and girls differ in their responses/ the analysis shows that boys and girls approach problems in different ways and these varying strategies can be useful in indentifying how the sexes differ in reasoning abilities. Two years later/ Marshall (1983) analyzed the responses of approximately 300/000 boys and girls to mathematics items contained in the 16 test forms of the Survey of Basic Skills during the years 1977/ 1978/ and 1979. She used log-linear models (explained in Chapter 42 children whose errors fell under the category or "other errors." Students were asked to read each problem aloud and answer these questions: What kind of situation does the problem describe? What does the problem ask you to find? How would you work the problem? Ninety five percent of the students read the problem correctly; 98% explained the kind of situation the problem described in a correct manner; 92% correctly answered what the problem was asking them to find* and 36% correctly answered the question of how to work the problem. The fact that a large percent of the students whose errors were classified as "other errors" (in which reading skills might have been a factor) correctly stated how to work the problem* is strong evidence of their ability to read and interpret the problems correctly. The errors made by this group of students had a distinct origin* unrelated to reading ability. Zalewski (1974) investigated the relative contribution of verbal intelligence* reading comprehension* vocabulary* interpretation of graphs and tables* mathematical concepts* number sentence selection* and computation to successful mathematical word problem solution* and the relationship of the dependent variable to the eight independent variables. She worked with a group of 4th grade children who were administered the subtests of the Iowa Test of Basic Skills (ITBS) and the Weschler Intelligence Scale for 34 achievement and interest in mathematics result from identification with the masculine role. A study related to the differential cultural reinforcement hypothesis is that of Dwyer (1974). Dwyer examined the relationship between sex role standards (the extent to which an individual considers certain activities appropriate to males or females) and achievement in reading and arithmetic. Students from grades 2/ 4 > 6 > 81 10 r and 12 participated in this study. She found that sex role standards contributed significant variance to reading and arithmetic achievement test scores and that the effect was stronger for males than for females. This led to her conclusion that sex-related differences in reading and arithmetic are more a function of the child's perception of these areas as sex-appropriate or sex-inappropriate than of the child's biological sex individual preference for masculine and feminine sex roles/ or liking or disliking reading or mathematics. In a study which agrees with the masculine identification hypothesis* Milton (1957) found that individuals who had received strong masculine orientation performed better in problem solving than individuals who received less masculine orientation. Elton and Rose (1967) found that women with high mathematical aptitude and average verbal aptitude scored higher on the masculinity scale of the Omnibus Personality Inventory (OPI) than those with average scores on both tasks. 67 conducted* one with each of addition* subtraction* multiplication* division* addition of fractions* decimals in subtraction* and equivalence as covariates. A well known problem that arises in the use of ANCOVA is that an unreliable covariate can cause spurious differences between the sex groups. To solve this problem* Porter (1967/1968) proposed the use of estimated true scores for observed scores. Porter (1967/1968) conducted a simulation that gave empirical support to the adequacy of this strategy. Hunter and Cohen (1974) have provided theoretical support for this strategy. 103 question was answered/ partially/ in the affirmative. Interactions/ indicating the dependence/ were significant for multiplication and division in the first year and for subtration in the second year. In general/ the sex-related differences in mathematics problem-solving/ in favor of males performance/ reported by Marshall (1981/ 1984)/ and Fennema and Tartree (1983) were not supported in this investigation. Female performance was equal or higher than male performance in problem solving. The findings of the study have something in common with other investigations where mathematics problem solving has been the subject of interest/ the fact that males and females do poorly in word-story problem items (problem solving) The mean score in problem solving for males and females in the 9-item problem solving subtest was 3.436 and 3.862 respectively for the first year/ 3.632 and 4.015 for the second year/ and 3.927 and 4.341 for the third year. Implications of the Findings and Suggestions for Further Research The main purpose of this study was to investigate sex-related differences in the selection of incorrect responses. Another objective was to compare male and female performance after accounting for computational ability. This type of research is conducted in an attempt 46 Their study is important because it represents an attempt to demonstrate that multiple factors can interact in the correct solution of a mathematics word problem. They constructed three graded tests from a basal mathematics series for grades 3 through 8. For test lr the problems were set in pure computational form (the effects or reading/ interpretation/ as well as the necessity for integration were removed in an effort to measure the computational skills required by the word problems) . For test 2/ the effects of reading and computation were removed by reading the problems to the students and by giving scores based on whether or not the students set them up properly/ in an attempt to measure problem interpretation alone. For test 3/ the effect of computation was removed. The test yielded two scoresone by grading the students on whether or not they set up the problems properly and another by grading on the basis of the correct answer. The tests were administered to all 244 students from the 6th grade in two different schools. A diagnostic profile was obtained for each of the 217 students for which complete data were available: a computational score/ a probiem-interpretation score/ a reading score/ and a reading-problem solving score. They assumed that if the reading-prob1em interpretation score was lower (one or more levels lower) 106 operation to the numbers in the problem. The re sults indicate that students are not familiar with such basic problem-solving strategies as drawing a picture of a figure described in a problem or checking the reasonableness of a result, (p. 338) The fact that, in general after controlling for computational skills there were no statistically significant differences between the problem solving performance of the males and females who participated in this study would seem to imply that teachers need not worry about designing different teaching strategies for the sexes. However^ because the performance in problem-solving is low for both sexes teachers may be encouraged to emphasize problem solving in their daily teaching. This study was designed to investigate sex-related differences in mathematical performance. The findings clearly demonstrated that 6th grade females and males from the public schools in Puerto Rico are equally good or equally bad in the solution of word-story items (problem solving). Also/ in general/ the effect of computational skills is similar for each sex. Based on these findings/ it appears that greater differences exist within each sex than between the sexes. In conclusion/ the findings of the present study highlight the need for investigations of different approaches to instructional designs aimed at improving both male and female performance in problem solving. 32 attributes success to an internal, stable attribute, such as ability, then one is confident of being successful in the future and will continue to strive in that area. If one attributes success to an external factor such as a teacher, or to an unstable one, such as effort, then one will not be as confident of success in the future and will cease to strive. Failure attribution patterns work this way: if failure is attributed to unstable causes, such as effort, failure can be avoided in the future and the tendency will be to persist in the task. However, if failure is attributed to a stable cause, such as ability, the belief that one cannot avoid failure will remain. Studies reported by Bar-Tal and Frieze (1977) suggest that males and females tend to exhibit different attributional patterns of success and failure. Males tend to attribute their success to internal causes and their failures to external or unstable ones. Females show a different pattern; they tend to attribute success to external or unstable causes and failures to internal ones. The pattern of attributions, success attributed externally and failure attributed internally, has become hypothesized to show a strong effect on mathematics achievement in females. K1oosterman' s (1985) study supported these findings. According to Kloosterman, attributional variables appear to be more important achievement mediators for females than for males, as measured by mathematics word problems. More research is needed in this area. 108 Bern/ S. L. / & Bern D. V. (1970). We are all nonconscious sexists. Psychology Today/ 22-24, 261 115-116. Benbow/ C. P./ & Stanley/ J. C. (1980). Sex differences in mathematical ability: Fact or artifact? Science/ 210/ 1262-1264. Benbow/ C. P./ & Stanley/ J. C. (1982). Consequences in high school and college of sex differences in mathema tical reasoning ability: A longitudinal perspective. American .Educational Research Journal/ 12./ 598 622. Bobbie/ C. N. (1971) Sex-role preference and academic achievement (Doctoral dissertation/ Yeshiva Univer sity) Dissertation Abstg^ 1818B-1819B. Carey/ G. L. (1958) Sex differences in problem solving performance as a function of attitude differences. Journal Of Abnormal. $tn.d Social psychology/ 22/ 256-260. Carpenter/ T. P./ Corbitt/ M. K./ Kepner/ H. S./ Lindquist/ M. M./ & Reys/ R. E. (1980). Results of the second NAEP mathematics assessment: Secondary school. Mathematics Teacher/ 22/ 329-338. Carpenter/ T. P. / Lindquist/ M. M./ Mathews/ W./ & Silver/ E. A. (1984). Results of the third NAED mathematics assessment: Secondary school. Mathematics Teacher/ 22/ 652-659. Chase/ C. I. (1960). The position of certain variables in the prediction of problem solving in arithmetic. Journal of Educational Research/ 14./ 9-15. Cohen/ D./ & Wilkie/ F. (1979). Sex-related differences in cognition among the elderly. In M. A. Witting and A. C. Petersen (Eds.)/ Sex-related differences in C.ognjtjve_J.unct,iojiin.qj Developmental Issues/ New York: Academic Press. Connor/ J. M. / Serbin/ L. A./ & Schakmanz M. (1977). Sex differences in children's response to training on a visual-spatial test. Developmental Psychology/ 12/ 392-394. CHAPTER IV RESULTS Introduction The data gathered from 6th grade students from the public schools in Puerto Rico who took the Basic Skills Test in Mathematics-6 during three consecutive years were analyzed in this study. The first objective was to investigate whether boys and girls differ in the selection of incorrect responses/ and if the pattern of differences was consistent throughout the three years in which the test was administered. The second objective was to investigate whether males and females differ in problem solving performance/ if these differences persist after accounting for computational skills/ and if the male-female differences depend on the level of computational skills. The results are discussed in two sections. Study findings in the area of sex differences in the selection of incorrect responses are discussed in the first section. The second section is devoted to an exposition of the findings in the area of sex-related differences in problem solving. Sex-Related Differences in the Selection of ,Incorr_egt.. Response As indicated in Chapter III/ there are two models of interest. The first model indicates that year and option 68 94 TABLE 4.9 Adjusted Means on Problem Solving* by Covariate and Sex First Year Covariate Adjusted Means Males Females Addition 3.53 3.78 Subtraction 3.55 3.76 Fracadd 3.56 3.74 Decsub 3.73 3.58 Equivalence 3.38 3.91 REFERENCES Aiken/ L. R. (1970). Attitudes toward mathematics. Review of Educational Resarch/ IQ./ 551-596 . Allen/ R. H./ & Chambers/ D. L. (1977). A comparison of the mathematics achievement of males and females (Bulletin No. 9194). Madison: Wisconsin State Department of Public Instruction/ Division for Management and Planning Service. (ERIC Document Reproduction Services No. ED 159 073) Armstrong/ J. M. (1979) Achievement and participation of W.pme_n_in_.math.ematiÂ£sj Ln_Q_y.ejyi-e.W-. (Report of a two year study). Denver/ CO: Educational Commission of the States/ National Assessment of Educational Progress. (ERIC Document Reproduction Service No. ED 184 878) Armstrong/ J. M./ & Price/ R. A. (1982). Correlates and predictors of women's mathematics participation. Journal for Research in Mathematics Education/ 12/ 99-109. Astin/ H. S. ( 197 4) Overview of the findings. In H. Astin/ H. Suniewick/ & S. Dweck (Eds.) / Women: A Bibliography on their Education and Careers. (pp. 1-10). New York: Behavorial Publications. Backman/ M. E. (1972). Patterns of mental abilities: Ethnic/ socioeconomic and sex differences. American Educational Research Journal/ 1/ 1-11. Ballew/ H./ & Cunningham/ W. (1982). Diagnosing strengths and weaknesses of sixth-grade students in solving word problems. Journal for Research in Mathematics Education/ 13/ 202-210. Balow/ I. H. (1964). Reading and computation ability as determinants of problem solving. The Arithmetic Teacher/ 11/ 18-22. Bar-Tal / D. / & Frieze/ J. H. ( 1977 ) Achievement motivation for males and females as a determinant of attribution for success and failure. Sex roles. 1/ 301-313. 107 83 TABLE 4.3 Continued Year/ Subtest Males X SD Females X SD t Decsub 2.863 1.87 3.440 1.80 5.00* Equivalence .76 .76 .65 .75 -2.32** N = 504 N = 509 Third Problem Solving 3.927 2.49 4.341 2.44 2.67* Addition 5 .536 .84 5.704 .64 3.59* Subtraction 4.836 1.52 4.958 1.46 1.30 Multiplication 4.168 1.74 4.541 1.68 3 .47* Division 3.343 1.88 3.795 1.83 3.87* Fracadd 2.819 1.89 3 .117 1.85 2.53** Decsub 3.021 1.93 3.448 1.85 3.60* Equivalanee .830 .82 .800 .78 - .60 N = 509 N = 504 Note: The number of items in the problem solving subtest was 9 . In each computation subtestr the number of items was 6. An item was included in the computation subtest only if it measured a computation skill required to solve a problem solving item. * ** p <.01 P <.05 38 At present# no set of variables has been clearly established as a determinant of problem difficulty. Several researchers have investigated the effect of reading and computation on problem solving performance. Others have studied the effect of student attitudes toward problem solving in problem solving learning and achievement. Typically# correlational methods have been used to investigate these questions. Computational_S_k_U.ls, and_.Pr,QbJ.etn Solving Performance One of the first researchers to study the effect of computation and reading on problem solving performance was Hansen (1944). He investigated the relationship of arithmetical factors# mental factors# and reading factors to achievement in problem solving. Sixth grade students were administered tests in problem solving and categorized as superior achievers (best problem solvers) and inferior archievers (poorest problem solvers) The two groups were compared in selected factors believed to be related to success in arithmetic problem solving: arithmetical# mental and reading factors. After controlling for mental and chronological age# the superior achievers in problem solving surpassed the inferior achievers in mental and arithmetical factors. The superior group did better in only two of the six items under the reading factors: general language ability and the reading of graphs# charts# and tables. 58 Table 3.1 Hypothetical Probabilities of Option Choice Conditional on Year and Sex Year Sex 1 Option 2 3 First M .7 .2 .1 F .5 .3 .2 Second M .7 .2 .1 F .5 .3 .2 Third M .7 .2 .1 F .5 .3 .2 Hypothetical Conditional on Table 3.2 Probabilities of Year and Sex and Option Choice Arranged by Sex Option Sex Year 1 2 3 M First .7 .2 .1 Second .7 .2 .1 Third .7 .2 .1 F First .5 .3 .2 Second .5 .3 .2 Third .5 .3 .2 5 Females were better on items of computation and males were more successful on word-story problem items (problem solving). She also found that females successful performance in the problem solving items was more dependent on their successful performance in the computation items. Males did not need/ as much as females/ to succeed in the computation items in order to answer correctly the problem solving items. Although the general findings seem to support sex differences in mathematics learning and achievement/ the research done does not consistently support superiority for either sex. Most of the research has been concerned with how the sexes differ on subtests or total test scores in mathematics. Moreover/ the great majority of the studies deal with correct responses. Sex differences in incorrect responses at the item level have not been fully researched. Only two studies dealing with sex differences in incorrect responses at the item level were found in the research literature (Marshall/ 1981/ 1983). Marshall investigated whether boys and girls made similar errors in computation and story problems. She analyzed boys' and girls' answers to six mathematics items and found that the sexes made different errors/ possibly reflecting different problem solving strategies. Her original findings were supported when she studied the same problem using a large number of items three years later. Problem Solving 87 9 . 8 7 6 . 5 - 4 - 3 - 2 - 1 - T 1 1 1 I 2 3 4 5 6 Multiplication Fig. 4.1 Sex by Multiplication Interaction 78 TABLE 4.2 Continued Item Num ber X1 for Model 3.2 Xa- for Model 3.3 X* for Difference Actual Sample Size Minimum Sample Size for Signif. (a) . (b) (c) (d) (e) (f) 55 19.53 90.87 71.34* 41/796 11/770 56 21.62 70.14 48.52* 44/107 18/263 57 28.51 35.78 7.27 53/075 146/668 58 17.49 51.69 34.20* 44/044 25/873 59 50.96 73.68 22.72* 41/829 36/987 60 66.43 148.39 81.96* 60/494 14/828 61 21.39 394.54 373.15* 35/632 1/918** 62 55.03 300.95 245.92* 50/699 4/142 63 158.36 278.34 119.98* 52/068 8/719 64 88.01 246.48 158.47* 73/196 9/279 65 56.28 94.11 37.83* 42/429 22/532 66 50.51 113.88 63.37* 78/708 24/953 67 34.28 180.60 146.32* 51/862 7/121 68 48.01 225.46 177.45* 63/223 7/158 69 75.29 251.66 176.37* 62/511 7/121 70 44.39 271.35 226.96* 67/528 5/977 72 17.41 208.90 191.49* 84/175 8/831 73 24.61 37.80 13.19* 53/358 81/271 74 22.75 163.01 140.26* 84/175 12,057 66 Table 3.4 Hypothetical Probabilities of Option Choice Conditional on Sex and Year Year Sex 1 Option 2 3 First M .666 .250 .083 F .400 .400 .200 Second M .774 .120 .104 F .510 .212 .276 Third M .698 .214 .087 F .431 .352 .215 Note: Probabilities reported are truncated to three places. Which procedure is more appropriate to test the claim that the pattern of male and female option choice remains the same from year to year? it seems more reasonable to test for a pattern like that in Table 3.2 than to test for a pattern like that in Table 3.4* and* consequently* this was the strategy adopted in this study. C-Omparisoa of_Males_and Females in Problem Solving Pejfbrmance One object of the study was to compare the performance of males and females on problem solving. Two questions were addressed. First* do males and females differ in problem solving performance? Second* do these differences persist when computational skill is controlled for* and do these differences depend on the level of computational skill? Seven analyses of covariance (ANCOVA) tests were BIOGRAPHICAL SKETCH Sonia Feliciano was born on March 17/ 1938/ in Mayaguez/ Puerto Rico. She graduated from Eugenio Ma. de Hostos High School/ in her hometown. She received a degree in business administration/ with a major in finance/ from the University of Puerto Rico in June/ 1961. After graduation she worked as junior accountant for Pittsburgh Plate Glass/ International. During the academic year of 1969-70 she started her career in education/ obtaining the degree of Master in Education from the University of Puerto Rico in June/ 1975. From January/ 1971 to June/ 1985/ she served in the public school system of Puerto Rico in different positions. She worked as educational researcher until July/ 1977/ when was appointed Director of the Research Center of the Commonwealth Department of Education. During the academic year of 1978-79/ she initiated doctoral studies in the Faculty of Philosophy and Educational Sciences of the Universidad Complutense de Madrid/ in Madrid/ Spain. She completed the coursework requirements and then transferred to University of Florida during the fall of 1979. 116 27 Research findings in this area have been inconsistent. In 1966 / Maccoby stated that "by early school years/ boys consistently do better (than girls) on spatial tasks and this difference continues through the high school and college years" (p.26). In 1972/ Maccoby and Jacklin said that the differences in spatial ability between the sexes "remain minimal and inconsistent until approximately the ages of 10 or 11/ when the superiority of boys becomes consistent in a wide range of populations and tests" (p- 41) In 1974 / after a comprehensive literature search/ Maccoby and Jacklin concluded that sex differences in spatial visualization become more pronounced between upper elementary years and the last year of high school/ the years when sex-related differences in mathematics achievement favoring boys emerge. Guay and McDaniel (1977) supported in part Maccoby and Jacklin's 1974 findings. They found that among elementary school children/ males had greater high level spatial ability than females/ but that males and females were equal in low level spatial ability. This finding is inconsistent with that portion of Maccoby and Jacklin's review that suggests that sex differences become evident only during early adolescence. Cohen and Wilkie (1979) however/ stated that in tests measuring distinct spatial tasks/ males perform better than females in early adolescence and throughout their life span. Most studies carried out after 36 they show less interest in the subject than counterpart males. These differences in interest are what Hilton and Berglund (1974) suggest to account for sex-related differences in mathematics achievement. Although the perception of the usefulness of mathematics is still an important predictor of course taking for girls there is a growing similarity between males and females regarding the usefulness of mathematics (Armstrong & Price/ 1982; Fennema & Sherman/ 1977; Moller/ 1982/1983) Armstrong and Price investigated the relative influence of selected factors in sex-related differences in mathematics participation. Both males and females selected usefulness of mathematics as the most important factor in deciaing whether or not to take more mathematics in high school. Moller's study revealed that both males and females based mathematics course-taking decisions on career usefulness. A Fennema and Sherman (1977) study showed only slight differences between males and females in their feelings about the usefulness of mathematics. In her study of this variable among college students/ Probert (1983/1984) did not find any sex-related differences either. These have been the main affective variables researched in attempting to explain the underlying causes of sex-related differences observed in mathematics learning and achievement. In spite of the great diversity of studies TABLE OF CONTENTS Page ACKNOWLEDGMENTS LIST OF TABLES V LIST OF FIGURES V ABSTRACT viii CHAPTER I INTRODUCTION 1 Purpose of the Study 6 Significance of the Study 6 Organization of the Study 8 II REVIEW OF THE LITERATURE 9 Sex-related Differences in Incorrect Response Patterns 10 Sex-related Differences in Problem Solving 14 Cognitive and Affective Variables that Influence Sex Differences in Mathematics Learning and Achievement 23 Differences in Formal Mathematics Education 24 Differences in Spatial Ability 26 Differentiated Effect of Affective Variables 30 Problem Solving Performance and Related Variables 37 Computational Skills and Problem Solving Performance 38 Reading and Problem Solving Perfor mance 44 Attitudes Toward Problem Solving and Problem Solving Performance 50 iii CHAPTER I INTRODUCTION Sex differences in mathematics learning and achievement have been the subject of intensive research. Research done before 1974 has shown that male performance on mathematical achievement tests is superior to female performance by the time they reach upper elementary or junior high school (Fennema/ 1976/ p. 2). The literature strongly suggests that at the elementary level females outperform males in computation and males excel in mathematical reasoning (Glennon & Callahan/ 1968; Jarvis/ 1964; Maccoby/ 1966). Since 1974/ research findings have been less consistent. Fennema (1974)/ after reviewing 36 studies/ found that during secondary school or earlier/ sex-related differences in mathematics achievement are not so evident/ but that when differences are found/ they favor males in high level cognitive tasks (problem solving) and females in low level cognitive tasks (computation) As a result of a further review of the literature/ Fennema (1977) concluded that at the elementary level/ sex-related differences do not exist at all cognitive levels/ from computation to problem solving. 1 S^x^-elajted-Diiferences in Problem Solving Performance 81 In this part* results for the two questions related to the second objective of the study are presented. These questions serve as the framework for the presentation. Each question is stated* followed by the results pertaining to that question. Question 1: Do males and females differ in problem solving performance? The responses of 492 males and 510 females who took the Puerto Rico Basic Skills Test-6 during the spring of the first year were analyzed in this study. Also* data from 504 males and 509 females tested in the second year and from 509 males and 504 females tested in the third year were included in the analysis. The mean performance scores and the standard deviations for each of the eight variables are presented in Table 4.3. Results of t-tests are also presented in this table. Females outperformed males in problem solving* a finding consistently present in all the three years of test administration. Over the three-year period the mean differences favored females in all variables except equivalence. The sex-related differences in problem solving were significant (p<.01) for all three years. 86 TABLE 4.6 ANCOVA Summary Table: Subtraction Covariate Second Year Source df SS MS F Subtraction (S) 1 93 .14 593 .14 122.80* Sex (S) 1 18.58 18.58 3 .85* SxS 1 30.85 30.85 6.39* Error 1009 4873.40 4.83 *P < 05 14 jLexrxe-latÂ£_d_Differ enees in Problem Solving It has already been acknowledged that the subject of problem solving has been extensively researched. However/ as early as 1969/ Kilpatrick criticized the fact that the study of problem solving has not been systematic? some researchers have studied the characteristics of the problem while others have given their attention to the characteristics of the problem solvers. Moreover/ differences in the tests used to measure problem solving performance also constitute an obstacle when trying to compare the results of the studies carried out. In order to avoid this pitfall and provide a basis for comparison/ the studies reviewed in this section/ dealing with sex-related differences in problem solving/ have been divided in two groups. The first comprises those studies that used the Romberg-Wearne Problem Solving Test. The second contains other relevant studies in which problem solving performance has been measured by means of other instruments. The Romberg-Wearne Problem Solving Test merits special mention because it was the first attempt "to develop a test to overcome the inadequacies of total test scores in explaining the reasons why some students are successful problem solvers and others are not" (Whitaker/ 1976/ pp. 9/ 10) 63 model tests were conducted using a .01 level of significance. Since this research is based on Marshall (1981/ 1983)/ it is important to compare the method of analysis used in this study to the one used by Marshall. Marshall also used a two-step analysis. In the first step of her analysis she deleted the Uijk term from (3.1) and tested the adequacy of the model/ 1 g m j. j k = i + U j + o k+ U i j + Ui k+ j k (3.6) Following this/ she deleted the Ujk term to obtain log m j k = (Ji + Uj + Uk+Ui j + Uik. (3.7) and compared these two models using a likelihood ratio test. If the first test was nonsignificant and the second significant/ Marshall claimed that option choice was dependent on sex and the pattern of dependency was the same from year to year. This is the same claim that this study sought to establish. However/ the approach used here was to present evidence that model (3.2) fits the data while Marshall tried to show that model (3.6) fits the data. The major difference between the two approaches concerns the operationalization of the concept that the 16 of the test. The sexes differed in only 2 of the 19 reference tests Spatial Relations and Picture Group Name-Selection. A factor analysis however showed differences in the number and composition of the factors. For females/ a general mathematics factor was determined by mathematics computation/ comprehension/ application/ and problem solving. For males/ the comprehension and application parts determined one factor; problem solving with two other reference tests (Gestalt and Omelet) determined another factor. Meyer concluded that comprehension of the data and mastery of the prerequisite mathematical concepts did not guarantee successful problem solving either for males or for females. Problem solving scores for both sexes were about one third their scores in comprehension and one fourth their scores in application. She also concluded that the sexes may have approached the problem solving questions differently. The methods used by females for solving problem situations may have paralleled their approach to the application parts. Males may have used established rules and algorithms for the application parts/ but may have used more of a Gestalt approach to the problem solving situation. Whitaker (1976) investigated the relationship between the mathematical problem performance of 4th grade children and their attitudes toward problem solving/ their teachers' 45 He found that in the 4th grade the correlations between problem solving and abstract verbal reasoning/ reading comprehension/ arithmetic concepts/ and computation were .61/ .64/ .66/ and .60 respectively/ and .56/ .68/ .69/ and .63 in the 8th grade. When computation was held constant/ the correlation between problem solving and reading was .52 in grade 4 and .54 in grade 8. When reading was held constant the correlation between problem solving and computation was .43 in grade 4 and .42 in grade 8. Creswell (1982) worked with a sample of anglo and black adolescents from Chicago. Each subject was administered the California Achievement test. Multiple regression was used to analyze the data. The analysis showed that reading is more important than computation in predicting student performance in problem solving. Reading accounted for 49.5% of the variance; computation accounted for 14.6% of the variance. Ballew and Cunningham (1982) worked with 6th grade students in an attempt to find what proportion of students have as their main source of difficulty with word problems each of the following factors: a) computation skills/ b) interpretation of the problem/ c) reading and/ d) integrating these skills in the solution of problems. They also wanted to know if a student can be efficiently diagnosed as having one of the four categories as his/her main difficulty with mathematics word problems. 112 Landau* M. (1984). The effects of spatial ability and problem presentation format in mathematical problem solving performance on middle school students (Doctoral dissertation* Northwestern University). Dissertation. Abstracts., mteEP3ti.onal* * 442a- 443 A. Lindgren* H. C. Silva* I.* Faraco* I.* & DaRocha* N. S. (1964). Attitudes toward problem solving as a function of success in arithmetic in Brazilian elementary schools. Journal of Educational Research* 5&* 44-45. Maccoby* E. E. (1966) Sex differences in intellectual functioning. In E. E. Maccoby (Ed.)* The Development of Sex Differences (pp. 25-55) Stanford* CA: The Stanford University Press. Maccoby* E. E. & Jacklin* C. N. (1972). Sex differences in intellectual functioning. In A. Anastasi* (Ed.)* Asseestnent.in a pluralistic, sopi-elx* (pp. 37-55). Proceedings of the 1972 Invitational Conference on Testing Problems* Educational Testing Service* Princeton* NJ. Maccoby* E. E.* & Jacklin* C. N. (1974). The Psychology of Sex Differences* Stanford* CA: Stanford University Press. Marshall* S. P. (1981) Sex Differences in Sixth Grade Children's Problem Solving. Paper presented at the annual meeting of the American Educational Research Association* Los Angeles* CA. Marshall* S. P. (1983). Sex differences in mathematical errors: Analysis of distractor choice. Journal for Research in Mathematics Education* 14.* 325-336. Marshall* S. *P. (1984). Sex differences in children's mathematics achievement: Solving computations and story problems* Journal. Qf Ed.ucati.Q.P.^l_Pgy.ch.Q.l.P.gy* 2Â£* 194-204. Martin* M. D. (1964). Reading comprehension* abstract verbal reasoning and computation as factors in arithmetic problem solving (Doctoral dissertation* University of Iowa* 1963) pissertatiPn_Ab.strac.t.5 International* 24 4547A-4548A. Problem Solving 87 9 8 - 6 5 - 4 - 3 - 2 - 1 - T 1 I 1 1 1 1 2 3 4 5 6 Multiplication Fig. 4.1 Sex by Multiplication Interaction. 24 learning and achievement arise from sex differences in spatial visualization; and that sex-related differences result from a differentiated effect of affective variables on the mathematics performance of males and females. D.,iffÂ£JC-epcgg.. in. F,Qfmal_ M3th.em.3tics_JÂ¡ducation (Differential Coursework Hypothesis) The basis for the differential coursework hypothesis is the fact that sex-related differences in mathematical learning and achievement show up when comparing groups which are not equal in previous mathematics learning. After the 8th grade/ boys tend to select mathematics courses more often than girls. Therefore/ girls show lower achievement scores in mathematics tests because their mathematics experience is not as strong as the boys' (Fennema/ 1975; Fennema & Sherman/ 1977; Sherman/ 1979). Fennema and Sherman's study (1977) lends additional support to the feasibility of viewing sex differences in mathematics learning and achievement as reflecting something other than a difference in mathematics aptitude. After controlling for previous study of mathematics/ they found significant sex differences in mathematics achievement in only two of the four schools under study/ making the attribution to sex per se less likely. Controlling for the number of space visualization-related courses/ the sex-related differences which originally 79 TABLE 4.2 Continued Item Num ber X1 for Model 3.2 Xz for Model 3 .3 Xz for Difference Actual Sample Size Minimum Sample Size for Signif. (a) (b) (c) (d) (e) (f) 75 8.69 237.07 228.38* 67 /783 5/963 76 99.68 167 .72 68.04* 70/183 20/723 77 79.38 432.36 352.98* 66/304 3/774 78 64.56 127 .04 62.48* 76/806 24/696 79 23.98 72.59 48.61* 48/690 20/123 80 75.01 128.13 53 .12* 55/277 20 /906 81 95.51 106.82 11.31* 57 /710 102/511 82 51.99 108.12 56.13* 71/315 25/525 83 20.12 95.18 75.06* 81 /021 21/685 84 38.62 69.05 30.43* 78/959 52/129 86 102.85 433.47 330.62* 39/903 2/425** 87 66.75 430.40 363.65* 48/194 2/662** 88 63.01 126 .27 63.26* 38/742 12/304 89 43 .26 289.14 245.88* 39/648 3 /240 90 85.53 405.04 319.51* 53 /557 3/368 91 82 .53 85.48 2.95 62/954 428/727 92 16.36 93 .12 76.76* 88/231 23 /092 93 78.39 229.02 150.63* 80 ,122 10/766 94 70.15 137 .27 67.12* 87 /241 26 /113 80 TABLE 4.2 Continued Item Num ber Xa for Model 3.2 Xa for Model 3.3 X2 for Difference Actual Sample Size Minimum Sample Size for Signif. la) lb) (c) (d) lell . Cf) 95 40.59 68.41 27.82* 83/969 60/638 96 29.31 212.34 183.03* 74/287 8/154 97 8.60 126.96 118.36* 50/028 8/492 98 29.87 41.77 11.90* 69/052 116/576 99 9.44 9.94 0.50 72/647 2/918/956 100 11.35 41.48 30.13* 60/931 40/627 101 23.77 315.62 291.85* 83/782 5/767 102 61.06 437.09 376.03* 75/708 4/045 103 6.12 7.97 1.85 50/663 550/173 104 40.47 283.94 243.47* 72/517 5/984 105 28.66 48.59 19.93* 60/712 61/199 106 13.47 1/682.71 1/669.34* 71/946 866*1 107 13.13 99.20 86.07* 72/482 16/918 108 26.56 79.33 52.77* 86/261 32/840 109 5.73 18.33 12.60* 76/800 112/453 110 17.80 49.11 31.31* 70/980 45/544 111 30.71 477.44 446.73* 86/190 3/876 PC.01 Minimum sample size less than 3/000 112 LandaUf M. (1984). The effects of spatial ability and problem presentation format in mathematical problem solving performance on middle school students (Doctoral dissertation* Northwestern University). Dissertation Abstracts International? 4/ 442A- 443 A. Lindgren/ H. C./ Silva/ I./ Faraco I./ & DaRocha/ N. S. (1964). Attitudes toward problem solving as a function of success in arithmetic in Brazilian elementary schools. Journal of Educational Research/ 2., 44-45 . Maccoby/ E. E. (1966). Sex differences in intellectual functioning. In E. E. Maccoby (Ed.)/ The Development of Sex Differences (pp. 25-55). Stanford/ CA: The Stanford University Press. Maccoby/ E. E. / & Jacklin/ C. N. (1972). Sex differences in intellectual functioning. In A. Anastasi/ (Ed.)/ Assessment in a pluralistic society, (pp. 37-55) . Proceedings of the 1972 Invitational Conference on Testing Problems/ Educational Testing Service/ Princeton/ NJ. Maccoby/ E. E./ & Jacklin/ C. N. (1974) The Psychology of Sex Differences/ Stanford/ CA: Stanford University Press. Marshall/ S. P. (1981) Sex Differences in Sixth Grade Children's Problem Solving. Paper presented at the annual meeting of the American Educational Research Association/ Los Angeles/ CA. Marshall/ S. P. (1983). Sex differences in mathematical errors: Analysis of distractor choice. Journal for Research in Mathematics Education/ 14./ 325-336 . Marshall/ S. 'p. (1984). Sex differences in children's mathematics achievement: Solving computations and story problems/ Journal of Educational Psychology/ lAf 194-204. Martin/ M. D. (1964). Reading comprehension/ abstract verbal reasoning and computation as factors in arithmetic problem solving (Doctoral dissertation/ University of Iowa/ 1963) Dissertation Abstracts International/ 24/ 4547A-4548A. 47 than the problem-interpretation score? the difficulty was due to reading ability. If the score of the lowest of the three areas (computation? problem interpretation? and reading-problem interpretation score) was the same as the reading-problem solving score? the student's area of greatest immediate need was either computation? problem interpretation? or reading. If the reading-problem solving score was lower than the lowest of the other three scores? the student's area of greatest immediate need was integration. Analysis of the data revealed that for 19% of the students? problem interpretation was their major difficulty; for 26% of the students? integration (total problem solving) was their greatest immediate need; for another 26%? computation was the major weakness; and for 29%? reading was their greatest immediate need. Seventy five percent of the students demonstrated clear strength in computation? 21% in problem interpretation? and 4% in reading-problem interpretation. An analysis across all students (including those without complete data) showed that 26% of the subjects could not work word problems at a level as high as that at which they could compute? interpret problems? and read and interpret problems? when those areas were measured separately. This led them to conclude that knowing the skills or the components of solving word problems is not sufficient for 6 Purpose of the Study The purpose of this study was twofold. The first was to investigate sex-related differences in the selection of incorrect responses in a multiple-choice achievement test in mathematics. For each test item* the research questions were as follows: 1. Is there a difference in the proportion of males and females choosing each incorrect option? 2. Is the same pattern of differences found in data obtained in three different administrations of the test? The second objective was to investigate sex-related differences in test scores in mathematics problem solving. The following questions were studied: 1. Do males and females differ in problem solving perf ormance? 2. Do sex differences in problem solving persist after controlling for computational skillsf and does the differential success of males and females on problem solving items depend on their success on the computation items? Significance of the Study Item response patterns are very useful techniques in the assessment of mathematics learning and achievement. Total test scores can be very misleading in the assessment of student performance and provide no diagnostic information about the nature and seriousness of student 65 One important characteristic of this table involves two-by-two subtables of sex cross classified with two of the options for each year. For example* consider the three two-by-two tables obtained by cross classifying sex and option choices one and two for each year. These tables are indicated by dotted lines on Table 3.3. For this table* the ratio of the odds of a male choosing option one to the odds of a female choosing option one is the same (within truncating error) for each year. For example* this odds ratio for the first year is (.120/.045) / (.060/.060) * 2.67. Within the error caused by reporting truncated probabilities* the odds ratio for years two and three is the same as that for year one. The equality of these odds ratios is Marshall's operationalization stage of the year- to-year gender-option dependency. To show that the odds ratio can be constant over years* but that the probabilities of option choice conditional on sex and year can change from year to year* for both males and females* the probabilities in Table 3.3 were converted to the probabilities of option choice conditional upon sex and year. These conditional probabilities are reported in Table 3.4. Unlike the probabilities in Table 3.2* those in Table 3.4 change from year to year for both males and females. 77 TABLE 4.2 Continued Item Num ber Xa for Model 3.2 Xa for Model 3 .3 Xa for Difference Actual Sample Size Minimum Sample Size for Signif. (a) (b) (c) (d) (e) (f) 36 84.36 134.56 50 .20* 86 f137 34*472 37 23.70 76.85 53 .15* 37 *06 8 14*011 38 13.64 84 .24 70.60* 38*870 11*0618 39 13.35 757.60 744.25* 61*057 1*648** 40 6.67 701.39 694.72* 56 *438 1*630** 41 12.65 603.42 590.77* 57 *145 1*943** 42 23 .26 189.11 165.85* 59*140 7*154 43 16.29 21.65 5.36 33*712 126 *357 44 3.55 115.84 112.29* 67*192 12*021 45 22.41 55.90 33.49* 13*098 37*851 46 6.56 16.30 9.74* 56 *727 117 *007 47 7.82 76.15 68.33* 56 *555 16*628 48 179.25 452.64 273.39* 97 *593 7*172 49 27 .82 59.75 31.93* 16*766 10*549 50 14.23 240.27 226.04* 25*410 2*258** 51 16.37 224.54 208.17* 24 *715 2*385** 52 26.01 96 .14 70.13* 39*858 11*418 53 6.69 27.49 20.80* 44*694 43 *168 54 41.41 95.41 54.00* 74*502 27*718 92 TABLE 4.8 ANCOVA Summary Table: Other Covariates First Year Source df SS MS F Addition 1 271.57 271.57 49.74* Sex 1 15.54 15.54 2.85 Error 999 5453.90 5.46 Subtraction 1 771.78 771.78 155.60* Sex 1 10.73 10.73 2.16 Error 999 4953.70 4.96 Fracadd 1 1412.11 1412.11 326.87* Sex 1 7.54 7.54 1.75 Error 999 4313.30 4.32 Decsub 1 1281.10 1281.10 287.89* Sex 1 4.94 4.94 1.11 Error 999 4444.30 4.45 Equivalence 1 811.44 811.44 164.93* Sex 1 69.69 69.69 14.16* Error 999 4914.00 4.92 * p .05 UNIVERSITY OF FLORIDA 3 1262 08554 6660 77 TABLE 4.2 Continued Item Num ber X* for Model 3.2 Xa for Model 3.3 X2- for Difference Actual Sample Size Minimum Sample Size for Signif. (a) (b) (c) . (d) . (e) (f) 36 84.36 134.56 50.20* 86,137 34,472 37 23.70 76.85 53.15* 37,068 14,011 38 13.64 84.24 70.60* 38,870 11,0618 39 13.35 757.60 744.25* 61,057 1,648** 40 6.67 701.39 694.72* 56,438 1,630** 41 12.65 603.42 590.77* 57,145 1,943** 42 23 .26 189.11 165.85* 59,140 7,154 43 16.29 21.65 5.36 33,712 126,357 44 3.55 115.84 112.29* 67,192 12,021 45 22.41 55.90 33.49* 13,098 37,851 46 6.56 16.30 9.74* 56,727 117,007 47 7.82 76.15 68.33* 56,555 16,628 48 179.25 452.64 273.39* 97,593 7,172 49 27.82 59.75 31.93* 16,766 10,549 50 14.23 240.27 226.04* 25,410 2,258** 51 16.37 224.54 208.17* 24,715 2,385** 52 26.01 96.14 70.13* 39,858 11,418 53 6.69 27.49 20.80* 44,694 43,168 54 41.41 95.41 54.00* 74,502 27,718 CHAPTER II REVIEW OF THE LITERATURE Sex differences in mathematics learning and achievement have been a subject of concern for educators and psychologists. Many studies found in the literature support the existence of these differences. Boys show superiority in higher level cognitive tasks (problem solving or mathematical reasoning) in the upper elementary years and in the early high school years (Fennema/ 1974; Maccoby & Jacklin/ 1974) . Almost all the research carried out has dealt with analysis of total correct scores in mathematics aptitude and achievement tests or scores in subtests. The literature related to sex differences in incorrect responses/ the main subject of the present study/ is surprisingly sparse. For the most part/ the studies have investigated the differences between the sexes in mathematics learning and achievement and the underlying variables causing the differences. Cognitive and affective variables have been the matter at issue in the establishment of sex differences. Although research in mathematics problem solving/ the secondary subject of this investigation/ is extensive/ most 9 48 success* since the components must be integrated into a whole process (mastery learning of the components cannot assure mastery of the process). Their analysis also led them to conclude that* in the case of 6th graders* inability to read problems is a major obstacle in solving word problems. Only 12% of the subjects could read and set up problems correctly at a higher level than they could compute* while 60% could compute correctly at a higher level than they could read and set up problems; 44% could set up problems better when they heard them read than when they read the problem themselves. Only 13% could set up problems better when they read them than when they heard them read. Muth ( 1984 ) investigated the role of reading and computational skills in the solution of word problems. A group of 200 students from the 6th grade were administered a test of basic skills and a mathematics word problem test. The word problem test consisted of 15 sample items supplied by the National Assessment of Educational Progress. The items were adapted to include some extraneous information and complex syntactic structure. Four versions of the test were constructed by combining two versions of problem information (absence vs. presence of extraneous information) with two versions of syntactic structure (simple vs. complex syntax) Task performance was measured by means of the number of problems answered correctly* 104 to contribute to the body of knowledge regarding such differences in mathematical learning and achievement. This section of Chapter V comprises implications of the findings and recommendations evolving from the present investigation. ex-related,,Differences jp incorrect Responses The main purpose of education is to impart formal education equally for the sexes. Analysis of the data on incorrect responses showed that for 100 of the 111 items of the test/ males and females selected different incorrect responses/ and this pattern of responses was consistently found during the three years of test administration. However/ with a sample of less than 3/000 subjects/ statistical significant results were obtained for only 6 items. Therefore/ it seems that/ at least for 6th grade children/ educators need not concern themselves with varying teaching techniques for the sexes. This study/ however/ has value apart from the investigation of sex-related differences. It was found that/ year after year/ 6th graders from the public schools in Puerto Rico tend to make the same errors in the solution of mathematical problems. It is important/ therefore/ that teachers provide the learning strategies necessary to enable students to obtain a better understanding of the concepts of numbers. 23 computation grew larger. This difference grew larger for girls than for boys. Although SES was a major factor in solving computation and story problem items successfully/ the effect was similar for each sex. Sex-related differences by primary language or chronological age were not large. This research carried out by Marshall with elementary grade children supports previous research findings that males are better than females in mathematics problem solving (a higher order skill) and females are better than their counterpart males in computation (a lower level skill). Marshall's research also brought out a different aspect of this question: the notion that girls find it more necessary than boys to succeed in the computation items in order to successfully solve the story problem items. Cognitive and Affective Factors That Influence_Sex Differences in Mathematics Learning and Achievement The research reviewed in the literature does not provide evidence of any unique variable that could serve as an explanation for the observed sex-related differences in mathematics learning and achievement. However/ some issues have been discussed/ among which the most prevalent are that sex-related differences in mathematics learning and achievement are a result of differences in formal education; that sex-related differences in mathematics 30 The issue of sex-related differences in spatial visualization ability as an explanation for sex differences observed in mathematics achievement is less convincing and the findings more contradictory than in the issue of sex differences in formal education. Besides these cognitive issues/ other issues/ mostly affective in nature/ have also been studied in trying to explain the origin of these sex differences in mathematics achievement and learning. The studies dealing with these affective variables are reviewed in the next section. Differentiated Effect of Affective Variables Researchers have attempted to explain the effect of sex differences in internal beliefs/ interests/ and attitudes (affective variables) on mathematics learning and achievement. A brief statement of each explanation precedes the summary of studies conducted that support the explanation. ;. Females/ more than males/ lack confidence in their ability to learn mathematics and this affects their achievement in mathematics and their election of more advanced mathematics courses. Maccoby and Jacklin (1974) reported that self- confidence in terms of grade expectancy and success in particular tasks was found to be consistently lower in women than in men. In 1978/ Fennema and Sherman reported 43 Children (WISC). Multiple regression analysis was performed. A correlation of .769 was found between word problem solving and the eight independent variables. Correlations between word problem solving and the independent variables ranged from .363 (verbal intelligence) to .674 (mathematical concepts). Correlations between the independent variables ranged from .369 (verbal intelligence and computation) to .749 (reading comprehension and vocabulary). Mathematical concepts* computation* and number sentence selection were almost as effective as all eight independent variables in predicting achievement in mathematical word problem solving. Mathematical concepts* computation* number sentence selection* and reading comprehension accounted for 58% of the variance* whereas all eight predictors accounted for 59% of the variance. The two best predictors were mathematical concepts and computation* which accounted for 54% variance. Other variables accounted for about 40% of the variance. The author recommends that the findings of this study be interpreted cautiously because the correlation between the eight independent variables was high* and* according to Zalewski* in a study of this nature where the interest is primarily in the influence of several variables on one dependent variable* a low correlation between the independent variables is required. (p. 2804) 99 responses was consistent throughout the years in which the test was administered. Males and females selected different incorrect responses in 100 of the 111 items of the test. The pattern of male-female differences in the selection of incorrect responses was consistently found in each of the years of test administration. However/ for the vast majority of the 100 items/ male-female differences were relatively small in magnitude/ considering the fact that the number of subjects needed to obtain significance was very large. Therefore/ these findings lack educational significance. A second objective of the study was to investigate whether males and females differ in problem solving performance/ if these differences persist after controlling for computational skills/ and if the male-female differences depend on the level of computational skills. Results of the analyses reported in this section generally indicated that females outperformed males not only in problem solving/ but in six of the seven computational variables. Males surpassed females in equivalence/ but statistical significance was obtained for only one of the three years covered in the study. The results also tended to show that/ for examinees with similar levels of computational skills/ sex differences in problem solving did not exist. The only 103 question was answered/ partially/ in the affirmative. Interactions/ indicating the dependence/ were significant for multiplication and division in the first year and for subtration in the second year. In general/ the sex-related differences in mathematics problem-solving/ in favor of males performance/ reported by Marshall (1981/ 1984)/ and Fennema and Tartree (1983) were not supported in this investigation. Female performance was equal or higher than male performance in problem solving. The findings of the study have something in common with other investigations where mathematics problem solving has been the subject of interest/ the fact that males and females do poorly in word-story problem items (problem solving) The mean score in problem solving for males and females in the 9-item problem solving subtest was 3.436 and 3.862 respectively for the first year/ 3.632 and 4.015 for the second year/ and 3.927 and 4.341 for the third year. Implications of the Findings and Suggestions for Further Research The main purpose of this study was to investigate sex-related differences in the selection of incorrect responses. Another objective was to compare male and female performance after accounting for computational ability. This type of research is conducted in an attempt 43 Children (WISC). Multiple regression analysis was performed. A correlation of .769 was found between word problem solving and the eight independent variables. Correlations between word problem solving and the independent variables ranged from .363 (verbal intelligence) to .674 (mathematical concepts). Correlations between the independent variables ranged from .369 (verbal intelligence and computation) to .749 (reading comprehension and vocabulary). Mathematical concepts computation and number sentence selection were almost as effective as all eight independent variables in predicting achievement in mathematical word problem solving. Mathematical concepts computation number sentence selection and reading comprehension accounted for 58% of the variance whereas all eight predictors accounted for 59% of the variance. The two best predictors were mathematical concepts and computation which accounted for 54% variance. Other variables accounted for about 40% of the variance. The author recommends that the findings of this study be interpreted cautiously because the correlation between the eight independent variables was high and according to Zal ewski in a study of this nature where the interest is primarily in the influence of several variables on one dependent variable a low correlation between the independent variables is required. (p. 2804) 65 One important characteristic of this table involves two-by-two subtables of sex cross classified with two of the options for each year. For example* consider the three two-by-two tables obtained by cross classifying sex and option choices one and two for each year. These tables are indicated by dotted lines on Table 3.3. For this table* the ratio of the odds of a male choosing option one to the odds of a female choosing option one is the same (within truncating error) for each year. For example* this odds ratio for the first year is (.120/.045) / (.060/.060) = 2.67. Within the error caused by reporting truncated probabilities* the odds ratio for years two and three is the same as that for year one. The equality of these odds ratios is Marshall's operationalization stage of the year- to-year gender-option dependency. To show that the odds ratio can be constant over years* but that the probabilities of option choice conditional on sex and year can change from year to year * for both males and females* the probabilities in Table 3.3 were converted to the probabilities of option choice conditional upon sex and year. These conditional probabilities are reported in Table 3.4. Unlike the probabilities in Table 3.2* those in Table 3.4 change from year to year for both males and females. MATH CONCEPTS Fractions j 2 items) j / Numbers j \ \ (18 items ) \ / Fig. 3.1 Number of Questions by Skill Area in the Basic Skills Test in Mathematics 6 CHAPTER III METHOD The first objective of this study was to investigate sex differences in the selection of incorrect responses in a mathematics multiple-choice achievement test* and to determine whether these differences were consistent over three consecutive administrations of the test. The second objective was to investigate whether males and females differ in problem solving performance* if these differences persist after accounting for computational skills* and if the male-female differences depend on the level of computational skills. This chapter contains descriptions of the sample* the test instrument* and the statistical analysis used in achieving the above mentioned objectives. The Sample To achieve the first objective of the study* all the students who took the Basic Skills Test in Mathematics-6 ("Prueba de Destrezas Bsicas en Matemticas-6") in each or the three years were included in the study. To achieve the second objective of the study* approximately 54 REFERENCES Aiken/ L. R. (1970). Attitudes toward mathematics. Re.Y,iÂ£M_Qf. Education2_Reg.ar.Â£h/ 10, 551-596. Allen, R. H., & Chambers, D. L. (1977). A comparison of the mathematics achievement of males and females (Bulletin No. 9194). Madison: Wisconsin State Department of Public Instruction, Division for Management and Planning Service. (ERIC Document Reproduction Services No. ED 159 073) Armstrong, j. m. (1979). ftgh.ie-yejnent_and..participatiQn pf women in mathematics: An overview. (Report of a two year study). Denver, CO: Educational Commission of the States, National Assessment of Educational Progress. (ERIC Document Reproduction Service No. ED 184 878) Armstrong, J. M., & Price, R. A. (1982). Correlates and predictors of women's mathematics participation, journal f. 99-109. Astin, H. S. (1974). Overview of the findings. In H. Astin, H. Suniewick, & S. Dweck (Eds.), Women: A Bibliography on their Education and Careers. (pp. 1-10). New York: Behavorial Publications. Backman, M. E. (1972). Patterns of mental abilities: Ethnic, socioeconomic and sex differences. American Educational Research Journal, Â£, 1-11. Ballew, H., & Cunningham, W. (1982). Diagnosing strengths and weaknesses of sixth-grade students in solving word problems. Journal for Research in Mathematics Education, 11, 202-210. Balow, I. H. (1964). Reading and computation ability as determinants of problem solving. The Arithmetic Teacher, 11, 18-22. Bar-Tal, D., & Frieze, J. H. (1977). Achievement motivation for males and females as a determinant of attribution for success and failure. Sex roles, 1, 301-313. 107 Table Page 4.11 Adjusted Means on Problem Solving* by Covariate and Sex 96 4.12 ANCOVA Third Year 97 4.13 Adjusted Means on Problem Solving* by Covariate and Sex 98 vi 35 It is not until adolescence that sex differences in the perception of mathematics as a male domain are found (Fennema* 1976; Stein* 1971? Stein & Smithless* 1969; Verbeke* 1983). In a study with 2nd* 6th and 12th graders* Stein and Smithless (1969) found that students' perceptions of spatial* mechanical* and arithmetic skills as masculine became more defined as these students got older. Fennema (1976) considers that the influence each sex exerts upon the other on all aspects of behavior is stronger during adolescence. Since during these years males stereotype mathematics as a male domain* they send this message to females who* in turn* tend to be influenced in their willingness to study or not to study mathematics. Before that stage* girls consider arithmetic feminine* while boys consider it appropriate for both sexes (Bobbe* 1971) . Usefulness of mathematics. Females perceive mathema tics as less useful to them than males do* and this perception occurs at a very young age. As a result* females exert less effort than males to learn or elect to take advanced mathematics courses. Many studies reported before 1976 found that the perception of the usefulness of mathematics for one's future differs for males and females* and is related to course taking plans and behavior (Fox* 1977). If females do not perceive mathematics as useful for their future* 92 TABLE 4.8 ANCOVA Summary Table: Other Covariates First Year Source df SS MS F Addition 1 271.57 271.57 49.74* Sex 1 15.54 15.54 2.85 Error 999 5453.90 5.46 Subtraction 1 771.78 771.78 155.60* Sex 1 10.73 10.73 2.16 Error 999 4953.70 4.96 Fracadd 1 1412.11 1412.11 326.87* Sex 1 7.54 7.54 1.75 Error 999 4313.30 4.32 Decsub 1 1281.10 1281.10 287.89* Sex 1 4.94 4.94 1.11 Error 999 4444.30 4.45 Equivalence 1 811.44 811.44 164.93* Sex 1 69.69 69.69 14.16* Error 999 4914.00 4.92 * p .05 114 Sherman/ J. (1979) Predicting mathematics performance in high school girls and boys. Journal of Educational Psychology/ 71/ 242-249. Sherman/ J. (1980). Mathematics/ spatial visualization/ and related factors: Changes in girls and boys/ grades 8-11. Journal of Educational Psychology/ 22./ 476-482. Stein/ A. H. (1971). The effects of sex role standards for achievement and sex role preference on three determinants of achievement motiviation. Developmental Psychology/ 4./ 219-231. Stein/ A. H./ & Smithless/ J. (1969). Age and sex differences in children's sex role standards about achievement/ Developmental Psychology/ 1/ 252-259. Suydam/ M. N. (1971) Research on mathematics education (K-12) reported in 1970. Journal for Research in Mathematics Education/ Z> 257-298. Suydam/ M. N. (1972). Research on mathematics education (K-12) reported in 1971. Journal for Research in Mathematics Education/ 196-232. Suydam/ M. N. (1973) Research on mathematics education (K-12) reported in 1972. Journal for Research in Mathematics Education/ 1/ 205-242. Suydam/ M. N. (1974). Research on mathematics education (K-12) recorded in 1973. Journal for Research in Mathematics Education/ / 238-272. Suydam/ M. N. / & Weaver/ J. F. ( 1975) Research on mathematics education (K-12) recorded in 1974. Journal for Research in Mathematics Education/ Â£/ 253-282. Verbeke/ K. A. (1983). Sex-related differences in mathematically gifted secondary students: An investigation of selected cognitive/ affective/ and educational factors (Doctoral dissertation/ University of Maryland/ 1982) Dissertation Abstracts International/ 43/ 2267A-2268A. Weiner/ B. (1974). Achievement motivation and attribution theory. Morristown/ NJ: General Learning Press. Ill Hansen/ C. W. (1944). Factors associated with successful achievement in problem solving in sixth grade arithmentic. Journal of Educational Research# 38 # 111-119. Harnisch# D. L. (1983). Item response patterns: Application for educational practice. Journal of Educational Measurement# 2SL' 191-205. Hilton# T. L. # & Berglund# G. W. (1974). Sex differences in mathematics achievement: A longitudinal study. Journal of. Educational Research# Â£2# 231-237. Hunter# J. E. # & Cohen# S. H. ( 1974). Correcting for unreliability in nonlinear models of attitude change. Psychometrika# 39 # 445-468. Jarvis# 0. T. (1964). Boy-girl ability differences in elementary school arithmentic. School Science and Mathematics# 64 # 657-659. Kaufman# A. (1984). Sex differences in mathematics reasoning in mathematically gifted and average fifth and sixth graders (Doctoral dissertation# Hofstra University# 1983) Dissertation Abstracts Inter national# 45 # 1094A-1095A. Kilpatrick I. (1969). Problem solving and creative behavior in mathematics. In J. W. Wilson# & L. R. Carry (Eds.)# Studies in mathematics (Vol. No. 19). Stanford# CA: School Mathematics Study Group. Kloosterman# P. W. (1985). Attributional theory# learned helplessness# and achievement in ninth grade mathematics (Doctoral dissertation# University of Wisconsin# 1984) Dissertation Abstracts Internatio nal# 4Â£/ 919A-920A. Knifong# J. D. # & Holtan# B. ( 1976) An analysis of children's written solutions to word problems. Journal for Research in Mathematics_Education# 2# 106-112. Knifong# J. D.# & Holtan# B. (1977). A search for reading difficulties among erred word problems. Journal for Research in Mathematics Education# Â£# 227-230. 24 learning and achievement arise from sex differences in spatial visualization; and that sex-related differences result from a differentiated effect of affective variables on the mathematics performance of males and females. Differences in Formal Mathematics Education (Differential Coursework Hypothesis) The basis for the differential coursework hypothesis is the fact that sex-related differences in mathematical learning and achievement show up when comparing groups which are not equal in previous mathematics learning. After the 8th grade/ boys tend to select mathematics courses more often than girls. Therefore/ girls show lower achievement scores in mathematics tests because their mathematics experience is not as strong as the boys' (Fennema/ 1975; Fennema & Sherman/ 1977; Sherman/ 1979). Fennema and Sherman's study (1977) lends additional support to the feasibility of viewing sex differences in mathematics learning and achievement as reflecting something other than a difference in mathematics aptitude. After controlling for previous study of mathematics/ they found significant sex differences in mathematics achievement in only two of the four schools under study/ making the attribution to sex per se less likely. Controlling for the number of space visualization-related courses/ the sex-related differences which originally Problem Solving 88 T T T 1 1 2 3 Division Fig. 4.2 Sex by Division Interaction 4 5 6 95 TABLE 4.10 ANCOVA Summary Table: Other Covariates Second Year Source df ss MS F Addition 1 206.97 206.97 39.57* Sex 1 19.12 19.12 3.66 Error 1010 5277.90 5.23 Multiplication 1 1026.68 1026.68 232.80* Sex 1 .80 .80 .18 Error 1010 4458.30 4.41 Division 1 1117 .46 1117.46 258.67* Sex 1 .64 .64 .15 Error 1010 4367.50 4.32 Fracadd 1 1341.13 1341.13 327.10* Sex 1 2.03 2.03 .50 Error 1010 4143.80 4.10 Decsub 1 1453 .91 1453.91 364.39* Sex 1 4.88 4.88 1.22 Error 1010 4031.10 3.99 Equivalence 1 797.11 797.11 171.79* Sex 1 127.33 127.33 27.44* Error 1010 4687.90 4.64 * P<.05 33 Mathematics.as. a male domain. Mathematics is an activity more closely related to the male sex domain than to the female sex domain (Eccles et al.# 1983). Thus* the mathematical achievement ot boys is higher than that of girls. According to John Ernest (1976) # in his study Mathematics and Sex/ mathematics is a sexist discipline. He attributed sex-related differences in mathematical achievement to the creation by society of sexual stereotypes and attitudes# restrictions# and constraints that promote the idea of the superiority of boys in mathematics. Ernest reported that boys# girls# and teachers# all believe that boys are superior in mathematics# at least by the time students reach adolescence. Bern and Bern (1970) agree and argue that an American woman is trained to "know her place" in society because ot the pervasive sex-role concept which results in differential expectations and socialization practices. Plank and Plank (1954) were more specific. They discussed two hypotheses related to this view: the differential cultural reinforcement hypothesis and the masculine identification hypothesis. The differential cultural reinforcement hypothesis states that society in general perceives mathematics as a male domain# giving females less encouragement for excelling in it. The masculine identification hypothesis establishes that CHAPTER V DISCUSSION This cnapter comprises two sections: a) summary and interpretation of the results of the study/ and b) implications of the findings/ with suggestions for further research. Summary and Interpretation of the Results The main objective of this study was to investigate whether there were sex-related differences in incorrect responses/ and if the pattern of response was consistent over three consecutive administrations of the mathematics test. Males and females selected different incorrect responses in 100 of the 111 items of the test and this pattern of response was consistenly found in each of the three years of test administration. The results tend to support/ in part/ Marshall's findings (1981/ 1983) of observed sex-related differences in incorrect responses/ but in no way do they point to a conclusion that 6th grade boys and girls from the public schools in Puerto Rico exhibit different problem solving strategies. The significant differentiated pattern of 101 44 In a more recent investigation# Exedisis (1983) studied the contribution of reading ability# vocabulary# mathematical concepts# computation# sex# and race on problem-solving performance. The Iowa Test of Basic Skills was administered to a group of 6th# 7th# and 8th grade anglo and black Chicago male and female adolescents. Problem solving was highly correlated to an understanding of basic mathematical concepts# somewhat correlated to race# and weakly correlated to computational and vocabulary skills# sex# and reading ability. Although the findings of these studies show a relationship between computational skills and problem solving achievement# this relationship is not strong enough to be considered the most determinant factor in problem solving achievement# as some of the researchers have been careful to point out. In spite of the dismissal of reading as a determinant factor in problem solving achievement by some of these same researchers# more recent studies in this area have led others to hold different views. Reading end problem. SQlvinfl-JLgj.f.Qirniangg Martin (1964) studied the contribution of reading comprehension# computation# abstract verbal reasoning# and arithmetic concepts to arithmetic problem solving performance. Fourth and 8th grade students were administered the Iowa tests of Basic Skills and the Lorge-Thorndlike intelligence test (verbal). I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate/ in scope and quality/ as a dissertation for the degree of Doctor of Philosophy. Jaimes Algina/Chai rman Processor of Foundations of Education I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate/ in scope and quality/ as a dissertation for the degree of Doctor of Philosophy. 2U J- Linda Crocker Professor of Foundations of Education I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate/ in scope and quality/ as a dissertation for the degree of Doctor of Philosophy. Leadership This dissertation was submitted to the Graduate Faculty of the College of Education and to the Graduate School/ and was accepted as partial fulfillment for the degree of Doctor of Philosophy. August/ 1986 Cnairman/ Founp lations. of Education A/uJ Cl zzUfh Dean/ College of EducSrtion Dean/ Graduate School (a) 80 81 82 83 84 85 86 87 88 89 90 91 93 94 95 96 98 101 102 104 73 TABLE 4.1 Continued X* for Model 3.2* Actual Sample Size Min. Sample Size for Significance (b) (c) (d) 75.01 55/277 6,787 95 .51 57 ,710 5,565 51.99 71/315 12,633 20.12 81/021 37 ,088 38.62 78/959 18,830 1240 .78 41,180 306** 102.85 39/903 3 ,573 66.75 48,194 6,650 63.01 38,742 5/663 43 .26 39,648 8,441 85 .53 53 ,557 5,767 82.53 62,954 7/025 78.39 80 ,723 9,484 70.15 87,241 11,454 40.59 83 ,969 19,053 29.31 74,287 23 ,343 29.87 69,052 21,291 23 .77 83 ,782 32/462 61.06 75/708 11/419 40.47 72 ,517 16 ,503 CHAPTER III METHOD The first objective of this study was to investigate sex differences in the selection of incorrect responses in a mathematics multiple-choice achievement test, and to determine whether these differences were consistent over three consecutive administrations of the test. The second objective was to investigate whether males and females differ in problem solving performance* if these differences persist after accounting for computational skills* and if the male-female differences depend on the level of computational skills. This chapter contains descriptions of the sample* the test instrument* and the statistical analysis used in achieving the above mentioned objectives. The Sample To achieve the first objective of the study* all the students who took the Basic Skills Test in Mathematics-6 ("Prueba de Destrezas Bsicas en Matemticas-6") in each or the three years were included in the study. To achieve the second objective of the study* approximately 54 7 errors (Harnisch? 1983). Item response patterns are valuable for the identification of large group differences/ including d i st rict-to-district/ school-to-school/ and classroom-to-classroom variations on different subsets of items. The response patterns can provide diagnostic information about the type of understanding the student has on various mathematics topics (e.g. / problem solving). Marshall (1981/ 1984) has used the item response pattern technique and her findings indicate sex differences in mathematics performance at the item level. Females outperformed males in computation and males outscored females in problem solving. Also/ the success of girls in the problem solving items was dependent upon their success in the computation items; for boys/ success in the problem solving items did not depend as much on their computational performance. Marshall (1981/ 1983) has also reported that males and females differ in the selection of incorrect responses/ reflecting differences in reasoning abilities. In Puerto Rico/ a high percentage of children promoted to the 7th grade in the public schools does not master the basic skills in mathematics. If 6th grade male and female children can be diagnosed as having different problem solving abilities/ as Marshall found with California children/ teachers may need to provide tailor-made mathematics instruction for each sex/ in order to ensure equal access to formal education and enhance mathematics 49 number of problems set up correctly/ and amount of time spent taking the test. Reading ability and computational ability were both positively correlated with number of correct answers and with number of problems correctly set up/ and negatively correlated with test-taking time. Presence of extraneous information was negatively correlated with correct answers and correct set ups and positively correlated with test-taking time. Syntactic complexity was not significantly correlated with any of the performance measures. Results of a multiple regression analysis showed that reading accounted for 46% of the variance in total correct answers and computation accounted for 8%. Reading ability and computational ability uniquely accounted for 14% and 8% of the variance in the number of correct answers/ respectively. Extraneous information added significantly to the variance explained in the number of correct answers/ but syntactic structure did not. Reading ability accounted for 5% of the variance in test-taking time/ but computation did not add significantly to the variance explained by reading. Muth concluded that reading and computation both contribute significantly to success in solving arithmetic word problems/ but that reading plays a more significant role than does computation. 39 The findings suggest that reading factors are not as important as arithmetic and mental factors in problem solving performance. However/ these findings should be taken cautiously as the content of the Gates tests (used to measure reading) is literary and does not include mathematical material. Chase (1960) studied 15 variables in an effort to find out which ones have significant influence on the anility to solve verbal mathematics problems. Only computation reading to note details and fundamental knowledge were primarily related to problem solving. Computation accounted for 20.4% of the 32% variance directly associated with problem solving. Chase concluded that a pupil's ability in the mechanics of computation comprehension of the principles that underline the number systems and the extent to which important items of information are noticed when reading are good predictors of the student's ability in solving verbal problems. Balow (1964) investigated the importance of reading ability and computation ability in problem solving performance. He objected to the approaches used by other researchers who in their analyses dichotomized research subjects as "poor" or "good" students and who ignored the recognizea effect of intelligence on reading and on mathematics achievement. Balow administered the Stanford 31 that in their study involving students from grades 6 through 12, boys showed a higher level of confidence in mathematics at each grade level. These differences between the sexes occurred in most instances even when no sex-related differences in mathematics achievement were found. The correlation between confidence in mathematics performance and mathematics achievement in this study was higher than for any other affective variable investigated. Sherman reported a similar finding in 1980; in males, the most important factor related to continuation in theoretical mathematics courses was confidence in learning mathematics. This variable weighed more than any of the cognitive variables: mathematics achievement, spatial visualization, general ability, and verbal skill. In the case of females, among the affective variables, confidence in learning mathematics was found to be second in importance to perceived usefulness of mathematics. Probert (1983) supported these findings with college students. A variable that needs discussion within the context of sex differences in confidence as learners of mathematics is causal attribution. Causal attribution models attempt to classify those factors to which one attributes success or failure. The model proposed by Weiner (1974) categorizes four dimensions of attribution of success and failure: stable and internal, unstable and internal, stable and external, and unstable and external. For example, if one 93 superiority in problem solving lost statistical significance. Adjusted meansr associated with the analyses/ are reported in Table 4.9. Summaries of the analyses of covariance for the second year are reported in Table 4.10. The analyses show that females retained their superiority in problem solving after controlling for equivalence. When the controlling variables were additionr multiplication/ division/ addition of fractions/ and subtraction with decimals females superiority in problem solving lost statistical significance. Adjusted means are reported in Table 4.11. The analyses of covariance for the third year are summarized in Table 4.12. The results of the analyses show that females maintained their superiority in problem solving when performance on either subtraction or equivalence was controlled. When addition/ multiplication/ division/ addition of fractions/ and subtraction of decimals were the controlling variables/ female superiority in problem solving lost statistical significance. Adjusted means are reported in Table 4.13. Summary Sex-related differences in mathematics achievement were the subject of this study. The first objective was to investigate whether males and females differ in the selection of incorrect responses/ and if the pattern of 98 TABLE 4.13 Adjusted Means on Problem Solving/ by Covariate and Sex Third Year Covariate Adjusted Means Males Females Addition 4.09 4.18 Subtraction 3.98 4.29 Multiplication 4.10 4.17 Division 4.17 4.16 Addf rac 4.07 4.20 Subdec 4 .15 4.12 Equivalence 3 .79 4.45 Problem Solving 89 Subtraction Fig. 4.3 Sex by Subtraction Interaction 51 student attitudes and performance in the area of problem solving. Carey (1958) constructed a scale to measure attitudes toward problem solving. Her interest was in general problem solving rather than in mathematical problem solving. Her work constitutes the first attempt to construct a measure of attitudes toward problem solving. The scale was used with a group of college students/ and she foundr among other things/ that problem solving performance is positively related to problem solving attitudes and that/ in the case of females/ positive modification of attitudes toward problem solving brings a significant gain in problem solving performance. Lindgren/ Silva/ Faracoz and DaRocha (1964) adapted Carey's scale of attitudes toward problem solving and applied it to a group of 4th grade Brazilian children. Students also answered an arithmetic achievement test/ a general intelligence test/ and a socioeconomic (SE) scale. A low but significant positive correlation was found between arithmetic achievement and attitudes toward problem solving. A near zero correlation was found between attitudes toward problem solving and intelligence. Since problem solving is one aspect of the discipline of mathematics/ this correlation between attitudes and arithmetic achievement can lead to a conclusion or a strong 2 Many variables# cognitive/ affective# and educational# have been investigated since 1974 in relation to sex differences in mathematics learning and achievement. Fennema and Sherman (1977) investigated the effect of differential formal mathematics education. After controlling for the number of years of exposure to the study of mathematics# they found sex differences in only two of the four schools under study. However# in those schools where boys scored higher than girls# differences were also found in their attitudes toward mathematics. Hilton and Berglund (1974) found significant sex differences after controlling for the number of mathematics courses taken# and attributed them to sex differences in interests. "As the boys' interests in science increase related to the girls'# their achievement in mathematics increases relative to that of the girls" (p. 234). Wise# Steel# and McDonald (1979) reanalyzed test data collected in a longitudinal study of 400#000 high school students (Project Talent). They found that when the effect of the number of high school mathematics courses was not controlled# no sex differences emerged for 9th graders# but that gains made by boys during the next three years were more than twice that of the girls. These differences between the sexes disappeared when the number of mathematics courses taken was controlled. Results of the 1978 Women and Mathematics National Survey# Survey I# 28 1974 have failed to support these sex differences in spatial abilities (Armstrong* 1979; Connor* Serbin* & Schackman* 1977; Fennema & Sherman* 1978; Sherman* 1979). Fennema and Sherman (1978) and Sherman (1979) have explored sex-related differences in mathematical achievement and cognitive and affective variables related to these differences. In a study involving students from grades 6* 7* and 8* from four school districts* Fennema and Sherman found that spatial visualization and problem solving were highly correlated for both sexes (.59 and .60). Even in the school district where sex differences were found in problem solving* no significant sex-related differences were found in spatial visualization. When Sherman (1980) compared groups of males and females in two different grades* 8 and 11* she found no sex-related differences in problem solving or in spatial visualization in grade 8. In grade 11* however* although the sexes differed in their problem solving performance* no sex-related differences were found in spatial visualization. Even though spatial visualization in grade 8 was the second best predictor of problem solving performance in grade 11* sex differences in grade 11 were not a result of spatial visualization since no differences were found in that skill. In spite of the fact that no sex differences were found in spatial abilities* it is evident that males and 83 TABLE 4.3 - Continued Year/ Subtest Males X SD Females X SD t Decsub 2.863 1.87 3.440 1.80 5.00* Equivalence .76 .76 .65 .75 -2.32** N = 504 N = 509 Third Problem Solving 3.927 2.49 4.341 2.44 2.67* Addition 5.536 .84 5.704 .64 3.59* Subtraction 4.836 1.52 4.958 1.46 1.30 Multiplication 4.168 1.74 4.541 1.68 3.47* Division 3.343 1.88 3.795 1.83 3.87* Fracadd 2.819 1.89 3.117 1.85 2.53** Decsub 3.021 1.93 3.448 1.85 3.60* Equivalance .830 .82 .800 .78 - .60 N = 509 N = 504 Note: The number of items in the problem solving subtest was 9. In each computation subtest the number of items was 6. An item was included in the computation subtest only if it measured a computation skill required to solve a problem solving item. * p <.01 ** p <.05 12 test form/ no sex differences were found in response patterns. Approximately 45% of each sex selected option Â£. The next popular choice for both sexes was option d/ selected by approximately 35% of both boys and girls. On the story problem of Form 1/ males and females responded alike. Their most popular incorrect response choice was option a. for both males and females. The second most popular incorrect choice was option js for both sexes. Response to the story problem in Form 2 showed sex differences in response choice. Including the correct option/ 33% of the girls selected option ji/ 20% chose option c / and 20% option d. For males/ approximately 25% selected option s and the same percent selected option d. Marshall concluded that although the analysis of incorrect responses does not explain why boys and girls differ in their responses/ the analysis shows that boys and girls approach problems in different ways and these varying strategies can be useful in indentifying how the sexes differ in reasoning abilities. Two years later/ Marshall (1983) analyzed the responses of approximately 300/000 boys and girls to mathematics items contained in the 16 test forms of the Survey of Basic Skills during the years 1977/ 1978/ and 1979. She used log-linear models (explained in Chapter 106 operation to the numbers in the problem. The re sults indicate that students are not familiar with such basic problem-solving strategies as drawing a picture of a figure described in a problem or checking the reasonableness of a result, (p. 338) The fact that/ in general/ after controlling for computational skills/ there were no statistically significant differences between the problem solving performance of the males and females who participated in this study would seem to imply that teachers need not worry about designing different teaching strategies for the sexes. However/ because the performance in problem-solving is low for both sexes/ teachers may be encouraged to emphasize problem solving in their daily teaching. This study was designed to investigate sex-related differences in mathematical performance. The findings clearly demonstrated that 6th grade females and males from the public schools in Puerto Rico are equally good or equally bad in the solution of word-story items (problem solving). Also/ in general/ the effect of computational skills is similar for each sex. Based on these findings/ it appears that greater differences exist within each sex than between the sexes. In conclusion/ the findings of the present study highlight the need for investigations of different approaches to instructional designs aimed at improving both male and female performance in problem solving. 82 TABLE 4.3 Means* Standard Deviations* and t-Tests for the Eight Mathematical Subtests Year/ Subtest Males X SD Females X SD t First Problem Solving 3.436 2.32 3.862 2.45 2.82* Addition 5.412 1.00 5.592 .86 3.05* Subtraction 4.550 1.66 4.852 1.56 1.22 Multi plicati on 4.014 1.80 4.433 1.68 3.81* Division 3.136 1.87 3.580 1.91 3.72* Fracadd 2.475 1.75 2.743 1.81 2.32** Decsub 2.648 1.83 3.321 1.86 5.77* Equivalence .77 .77 .73 .76 - .83 N = 492 N = 510 Second Problem Solving 3.632 2.28 4.015 2.36 2.65* Addition 5.470 .98 5.603 .86 2.29** Subtraction 4.843 1.52 4.923 1.44 .86 Multiplication 4.148 1.73 4.550 1.67 3.76* Division 3.242 1.77 3.632 1.86 3.42* Fracadd 2.565 1.74 2.903 1.91 2.94* 20 As a result of her investigations/ Marshall (1981) concluded that sex-related differences in mathematics performance may be the result of comparing the sexes on total test scores. If the test contains more computation items than problem solving items/ girls will perform better than boys/ but if the test contains more problem solving items than computation ones/ boys will outperform girls. With this in mind/ Marshall investigated sex-related differences in computation and problem solving by analyzing the responses of approximately 18/000 students from grade 6 who had been administered the Survey of Basic Skills Test: Grade 6/ during the academic year 1978-79. Two of the 16 test forms of the Survey were used to assess skills such as concepts of whole numbers/ fractions/ and decimals. These skills were tested both as simple computations and as story problems (problem solving). Two computation items and one story problem item were selected because they were particularly related; both computation items required skills needed in solving the corresponding story problem. It was assumed that correct solution of the computation item correlates with solving the story problem because the story problem requires a similar computation. Marshall found that girls were better in computation and boys were better in problem solving. She also found that boys were much more likely than girls to answer the 51 student attitudes and performance in the area of problem solving. Carey (1958) constructed a scale to measure attitudes toward problem solving. Her interest was in general problem solving rather than in mathematical problem solving. Her work constitutes the first attempt to construct a measure of attitudes toward problem solving. The scale was used with a group of college students* and she found* among other things* that problem solving performance is positively related to problem solving attitudes and that* in the case of females* positive modification of attitudes toward problem solving brings a significant gain in problem solving performance. Lindgren* Silva* Faraco* and DaRocha (1964) adapted Carey's scale of attitudes toward problem solving and applied it to a group of 4th grade Brazilian children. Students also answered an arithmetic achievement test* a general intelligence test* and a socioeconomic (SE) scale. A low but significant positive correlation was found between arithmetic achievement and attitudes toward problem solving. A near zero correlation was found between attitudes toward problem solving and intelligence. Since problem solving is one aspect of the discipline of mathematics* this correlation between attitudes and arithmetic achievement can lead to a conclusion or a strong 47 than the problem-interpretation scorer the difficulty was due to reading ability. If the score of the lowest of the three areas (computationr problem interpretation/ and reading-problem interpretation score) was the same as the reading-problem solving score/ the student's area of greatest immediate need was either computation/ problem interpretation/ or reading. If the reading-problem solving score was lower than the lowest of the other three scores/ the student's area of greatest immediate need was integration. Analysis of the data revealed that for 19% of the students/ problem interpretation was their major difficulty; for 26% of the students/ integration (total problem solving) was their greatest immediate need; for another 26%/ computation was the major weakness; and for 29%/ reading was their greatest immediate need. Seventy five percent of the students demonstrated clear strength in computation/ 21% in problem interpretation/ and 4% in reading-problem interpretation. An analysis across all students (including those without complete data) showed that 26% of the subjects could not work word problems at a level as high as that at which they could compute/ interpret problems/ and read and interpret problems/ when those areas were measured separately. This led them to conclude that knowing the skills or the components of solving word problems is not sufficient for 62 Here the summation is over all the cells in the contingency table/ F0 refers to the observed frequency in a cell/ and Fe refers to the estimated expected frequency in a cell. Denoting the observed proportion in a cell by P0 and the estimated expected proportion in a cell by Pe/ the test statistic can be written G2 = 2 N J2Po 1 9e (Po/pe}- (3.5) where N is the total number of subjects. This form of the test statistic suggests the following strategy. For any significant G2/ using P0 and Pe calculated from the total data set available for an item/ calculate the minimum N required for G2 to be significant. If the minimum N is very large/ this suggests that the statistically significant result is not practically significant since it can only be detected in very large samples. Of course/ the question remains as to what can be considered a minimum large N. Although there is room for argument/ it seems reasonable to claim that if an average of 1000 subjects per year is required to show significance/ then the result is not likely to be practically significant. On the basis of this reasoning/ it was proposed to ignore all significant results that would be nonsignificant if there were less than 3000 subjects available. In addition/ all log-linear BIOGRAPHICAL SKETCH Sonia Feliciano was born on March 17/ 1938/ in Mayaguez/ Puerto Rico. She graduated from Eugenio Ma. de Hostos High School/ in her hometown. She received a degree in business administration/ with a major in finance/ from the University of Puerto Rico in June/ 1961. After graduation she worked as junior accountant for Pittsburgh Plate Glass/ International. During the academic year of 1969-70 she started her career in education/ obtaining the degree of Master in Education from the University of Puerto Rico in June/ 1975. From January/ 1971 to June/ 1985/ she served in the public school system of Puerto Rico in different positions. She worked as educational researcher until July/ 1977/ when was appointed Director of the Research Center of the Commonwealth Department of Education. During the academic year of 1978-79/ she initiated doctoral studies in the Faculty of Philosophy and Educational Sciences of the Universidad Complutense de Madrid/ in Madrid/ Spain. She completed the coursework requirements and then transferred to University of Florida during the fall of 1979. 116 52 correlation between attitudes toward problem solving and problem solving performance. Whitaker (1976) constructed a student attitude scale to measure some aspects of 4th grade student attitudes toward mathematic problem solving. He included statements reflecting children's beliefs about the nature of various types of mathematical problems the nature of the problem solving process the desirability or persevering when solving a problem and the value of generating several ideas for solving a problem. He correlated student attitudes toward problem solving with their scores in a mathematical test which yielded a comprehension score an application score and a problem solving score. He found a significant positive relationship between problem solving performance and student attitude scores on the subscale which measured reactions to such things as problem solving techniques or problem situations or to the frustation or anxiety experienced when confronted with problem solving situations. In another part of this study Whitaker investigated the relationship between the attitudes of 4th grade teachers toward problem solving and their students' performance in problem solving. A very weak and nonsignificant negative correlation was found between the teacher's attitudes toward problem solving and student performance. 20 As a result of her investigations Marshall (1981) concluded that sex-related differences in mathematics performance may be the result of comparing the sexes on total test scores. If the test contains more computation items than problem solving items girls will perform better than boys but if the test contains more problem solving items than computation ones boys will outperform girls. With this in mind Marshall investigated sex-related differences in computation and problem solving by analyzing the responses of approximately 18000 students from grade 6 who had been administered the Survey of Basic Skills Test: Grade 6 during the academic year 1978-79. Two of the 16 test forms of the Survey were used to assess skills such as concepts of whole numbers fractions and decimals. These skills were tested both as simple computations and as story problems (problem solving). Two computation items and one story problem item were selected because they were particularly related; both computation items required skills needed in solving the corresponding story problem. It was assumed that correct solution of the computation item correlates with solving the story problem because the story problem requires a similar computation. Marshall found that girls were better in computation and boys were better in problem solving. She also found that boys were much more likely than girls to answer the 91 TABLE 4.7 Reliability Three of the Years of Covariates for Each of Test Administration the Year Covariate First Second Third Addition .579 .610 .454 Subtraction .756 .724 .737 Multiplication .729 .729 .734 Division .716 .680 .714 Fracadd .709 .728 .748 Decsub .720 .716 .742 Equivalence .421 .394 .488 LIST OF TABLES Table Page 3.1 Hypothetical Probabilities of Option Choice Conditional on Year and Sex 58 3.2 Hypothetical Probabilities of Option Choice Conditional on Year and Sex and Arranged by Sex 58 3.3 Hypothetical Joint Probabilities of Year/ Option/ and Sex 64 3.4 Hypothetical Probabilities of Option Choice Conditional on Sex and Year 66 4.1 Summary of Significant Tests of Model 3.2 71 4.2 Chi-square Values for the Comparison of 75 Models (3.2) and (3.3) with Actual Sample Sizes and with the Corresponding Number of Subjects Needed for Significant Results 75 4.3 Means/ Standard Deviations/ and t-Test for the Eight Mathematical Variables 82 4.4 ANCOVA First Year: Multiplication Covariate 85 4.5 ANCOVA First Year: Division Covariate 85 4.6 ANCOVA Second Year: Subtraction Covariate... 86 4.7 Reliability of the Covariates for Each of the Three Years of Test Administra tion 91 4.8 ANCOVA First Year: Other Covariates 92 4.9 Adjusted Means on Problem Solving/ by Covariate and Sex 94 4.10 ANCOVA Second Year: Other Covariates 95 v LIST OF TABLES Table Page 3.1 Hypothetical Probabilities of Option Choice Conditional on Year and Sex 58 3.2 Hypothetical Probabilities of Option Choice Conditional on Year and Sex and Arranged by Sex 58 3.3 Hypothetical Joint Probabilities of Yearr Option/ and Sex 64 3.4 Hypothetical Probabilities of Option Choice Conditional on Sex and Year 66 4.1 Summary of Significant Tests of Model 3.2 71 4.2 Chi-square Values for the Comparison of 75 Models (3.2) and (3.3) with Actual Sample Sizes and with the Corresponding Number of Subjects Needed for Significant Results 75 4.3 Means/ Standard Deviations/ and t-Test for the Eight Mathematical Variables 82 4.4 ANCOVA First Year: Multiplication Covariate 85 4.5 ANCOVA First Year: Division Covariate 85 4.6 ANCOVA Second Year: Subtraction Covariate... 86 4.7 Reliability of the Covariates for Each of the Three Years of Test Administra tion 91 4.8 ANCOVA First Year: Other Covariates 92 4.9 Adjusted Means on Problem Solving/ by Covariate and Sex 94 4.10 ANCOVA Second Year: Other Covariates 95 v 22 Approximately 270/000 students from the 6th grade were administered the Survey of Basic Skills of the California Assessment Program/ during the years 1977/ 1978/ and 1979. Responses were analyzed using log-linear models. Successful solving of computation items was positively associated with successful solving of story problems. Girls were more successful in computation than boys/ and boys were more successful than girls in solving story problems. This finding supports reports from the National Assessment of Educational Progress (NAEP) (Armstrong/ 1979) . To investigate the effects of reading/ SES/ language/ and chronological age/ only those test forms containing 2 computation items and 2 story problems were considered for analysis; 32 items from 8 test forms were included in the analysis. The results of these analyses showed that at every level of reading score/ 6th grade children were more successful in computation than in story problems. Although the differences were not large/ at every reading score boys consistently had higher probabilities of success in story problems than did girls/ and girls consistently showed higher probabilities of success in computation than boys. Also/ as the reading score increased/ the difference between the probability of success in story problems and the probability of success in 94 TABLE 4.9 Adjusted Means on Problem Solvingf by Covariate and Sex First Year Covariate Adjusted Means Males Females Addition 3 .53 3 .78 Subtraction 3 .55 3.76 Fracadd 3.56 3.74 Decsub 3.73 3.58 Equivalence 3 .38 3.91 17 attitude toward mathematical problem solving/ and related sex and program-type differences. Although his main objective was to construct an attitude scale to measure attitudes toward problem solving/ his study is important because his findings support Meyer's regarding the lack of significant sex-related differences in problem solving performance. Performance in the problem solving questions/ for both males and females/ was much lower than performance in the application questions/ and much lower than performance in the comprehension questions. In fact/ the mean score for each part of the item/ for both males and females/ was almost identical to the mean scores obtained by males and females in Meyer's study. Whitaker noted that each application item is more difficult than its preceding comprehension item/ and that each problem solving item is more difficult than its preceding application item. No significant sex-related differences were found for any of the three parts of the item (comprehension/ application/ or problem solving). Fennema and Sherman (1978) investigated sex-related differences in mathematics achievement and cognitive and affective variables related to such differences. They administered the Romberg-Wearne Problem Solving Test to a representative sample of 1320 students (grades 6-8) from Madison/ Wisconsin/ predominantly middle-class/ but including great diversity in SES. The sample consisted of 19 solving performance in grade 11. Spatial visualization was a stronger predictor for girls than for boys. Mathematics as a male domain was a good predictor for girls only; the less a girl stereotyped mathematics as a male domain in grade 8/ the higher her problem solving score in grade 11. Attitudes toward success in mathematics in grade 8 was a more positive predictor of problem solving performance for boys than for girls; the more positive the attitudes toward success in mathematics in grade 8/ the higher their performance in problem solving in grade 11. None of these four studies all of which used the Romberg-Wearne Mathematics Problem Solving Test show statistically significant sex-related differences in problem solving performance. In later studies other tests were used to measure this variable (Kaufman 1984; Marshall 1981 1984) . Kaufman (1983/1984) investigated if sex differences in problem solving favoring males exist in the 5th and 6th grades and if these differences were more pronounced in mathematically gifted students than in students of average mathematical ability. The Iowa Test of Basic Skills and a mathematics problem-solving test were administered to 504 subjects. Males in the average group as well as males in the gifted group outperformed females but only the gifted group showed statistically significant differences. 7 4 TABLE 4.1 Continued Item Number X2 for Model 3.2* Actual Sample Size Min. Sample Size for Significance la) (b) (c) (d) 105 28.66 60/712 19/510 108 26 .56 86/261 29/912 111 30.71 86/190 25/849 * p<.01 ** Minimum sample size less than 3/000 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy COMPARISON OF MALES AND FEMALES ON MATH ITEM PERFORMANCE: ANALYSIS OF RESPONSE PATTERNS By Sonia Feliciano August/ 1986 Chairman: James Algina Major Department: Foundations of Education The first objective of this study was to investigate sex differences in the selection of incorrect responses on a mathematics multiple-choice test/ and to determine whether these differences were consistent over three consecutive administrations of the test. A second objective was to compare male and female performance in problem solving after controlling for computational skills. The responses of all 6th grade students from the public schools in Puerto Rico who took the Basic Skills Test in Mathematics-6 ("Prueba de Destrezas Bsicas en Matemticas-6") during three academic years were used in the analyses relevant to the first objective. VXll 14 Sex-related Differences in Problem Solving It has already been acknowledged that the subject of problem solving has been extensively researched. However* as early as 1969* Kilpatrick criticized the fact that the study of problem solving has not been systematic; some researchers have studied the characteristics of the problem while others have given their attention to the characteristics of the problem solvers. Moreover* differences in the tests used to measure problem solving performance also constitute an obstacle when trying to compare the results of the studies carried out. In order to avoid this pitfall and provide a basis for comparison* the studies reviewed in this section* dealing with sex-related differences in problem solving* have been divided in two groups. The first comprises those studies that used the Romberg-Wearne Problem Solving Test. The second contains other relevant studies in which problem solving performance has been measured by means of other instruments. The Romberg-Wearne Problem Solving Test merits special mention because it was the first attempt "to develop a test to overcome the inadequacies of total test scores in explaining the reasons why some students are successful problem solvers and others are not" (Whitaker* 1976* pp. 9* 10) . 57 Analysis of the Data Analysis of Sex bv Option by Year Cross Classifications Log-linear models were used to analyze the sex by option by year cross classifications for each item. Two topics are addressed in this section/ the hypotheses tested using log-linear models and a comparison of the hypotheses tested in this study with those tested by Marshall (1981/ 1983) . The object of the analysis was to test two hypotheses: 1. The proportion of males and the proportion of females choosing each incorrect option does not vary from year to year. Note that this hypothesis is stated in the null form. 2. Assuming that the first hypothesis is correct/ the proportion of males who choose each incorrect option is different from the proportion of females who choose each incorrect option. This hypothesis is stated in the alternate form. In Table 3.1 a hypothetical cross classification of sex/ option/ and year is presented. Hypothesis 1 is true for this three dimensional contingency table. In Table 3.2 the three dimensional contingency table is rearranged to show the year by option contingency table for each gender. 82 TABLE 4.3 Means* Standard Deviations* and t-Tests for the Eight Mathematical Subtests Year/ Subtest Males X SD Females X SD t First Problem Solving 3.436 2.32 3.862 2.45 2.82* Addition 5.412 1.00 5.592 .86 3.05* Subtraction 4.550 1.66 4.852 1.56 1.22 Multi plication 4.014 1.80 4.433 1.68 3.81* Division 3.136 1.87 3.580 1.91 3.72* Fracadd 2.475 1.75 2.743 1.81 2.32** Decsub 2.648 1.83 3.321 1.86 5.77* Equivalence .77 .77 .73 .76 - .83 N = 492 N = 510 Second Problem Solving 3.632 2.28 4.015 2.36 2.65* Addition 5.470 .98 5.603 .86 2.29** Subtraction 4.843 1.52 4.923 1.44 .86 Multi plication 4 .148 1.73 4.550 1.67 3.76* Division 3.242 1.77 3.632 1.86 3.42* Fracadd 2.565 1.74 2.903 1.91 2.94* CHAPTER V DISCUSSION This cnapter comprises two sections: a) summary and interpretation of the results of the study; and b) implications of the findings/ with suggestions for further research. Summary and Interpretation of the Results The main objective of this study was to investigate whether there were sex-related differences in incorrect responses/ and if the pattern of response was consistent over three consecutive administrations of the mathematics test. Males and females selected different incorrect responses in 100 of the 111 items of the test and this pattern of response was consistenly found in each of the three years of test administration. The results tend to support/ in part/ Marshall's findings (1981/ 1983) of observed sex-related differences in incorrect responses/ but in no way do they point to a conclusion that 6th grade boys and girls from the public schools in Puerto Rico exhibit different problem solving strategies. The significant differentiated pattern of 101 31 that in their study involving students from grades 6 through 12 boys showed a higher level of confidence in mathematics at each grade level. These differences between the sexes occurred in most instances even when no sex-related differences in mathematics achievement were found. The correlation between confidence in mathematics performance and mathematics achievement in this study was higher than for any other affective variable investigated. Sherman reported a similar finding in 1980; in males the most important factor related to continuation in theoretical mathematics courses was confidence in learning mathematics. This variable weighed more than any of the cognitive variables: mathematics achievement spatial visualization general ability and verbal skill. In the case of females among the affective variables confidence in learning mathematics was found to be second in importance to perceived usefulness of mathematics. Probert (1983) supported these findings with college students. A variable that needs discussion within the context of sex differences in confidence as learners of mathematics is causal attribution. Causal attribution models attempt to classify those factors to which one attributes success or failure. The model proposed by Weiner (1974) categorizes four dimensions of attribution of success and failure: stable and internal unstable and internal stable and external and unstable and external. For example if one 97 TABLE 4.12 ANCOVA Summary Table Third Year Source df ss MS F Addition 1 205.96 205 .96 34.85* Sex 1 1.88 1.88 .32 Error 1010 5967.60 5.91 Subtraction 1 886.42 886.42 169.48* Sex 1 24.39 24.39 4.66* Error 1010 5287.2 5.23 Multi pi icati on 1 1402.42 1402.42 297.12* Sex 1 1.05 1.05 .22 Error 1010 4771.20 4.72 Division 1 2028.49 2028.49 494.75* Sex 1 1.11 1.11 .27 Error 1010 4145 .10 4.10 Fracadd 1 1733.72 1733.72 393.97* Sex 1 4.64 4.64 1.05 Error 1010 4439.89 4 .40 Decsub 1 2116.84 2116.84 526 .57* Sex 1 .15 .15 .04 Error 1010 4056 .76 4.02 Equivalence 1 1186.92 1186.92 240.26* Sex 1 59.85 59.85 12.12* Error 1010 4986.69 4.94 * p<. 05 35 It is not until adolescence that sex differences in the perception of mathematics as a male domain are found (Fennemar 1976; Stein/ 1971; Stein & Smithless/ 1969; Verbeke/ 1983). In a study with 2nd/ 6th and 12th graders/ Stein and Smithless (1969) found that students' perceptions of spatial/ mechanical/ and arithmetic skills as masculine became more defined as these students got older. Fennema (1976) considers that the influence each sex exerts upon the other on all aspects of behavior is stronger during adolescence. Since during these years males stereotype mathematics as a male domain/ they send this message to females who/ in turn/ tend to be influenced in their willingness to study or not to study mathematics. Before that stage/ girls consider arithmetic feminine/ while boys consider it appropriate for both sexes (Bobbe/ 1971). Usefulness of mathematics. Females perceive mathema tics as less useful to them than males do/ and this perception occurs at a very young age. As a result/ females exert less effort than males to learn or elect to take advanced mathematics courses. Many studies reported before 1976 found that the perception of the usefulness of mathematics for one's future differs for males and females/ and is related to course taking plans and behavior (Fox/ 1977). If females do not perceive mathematics as useful for their future/ 52 correlation between attitudes toward problem solving and problem solving performance. Whitaker (1976) constructed a student attitude scale to measure some aspects of 4th grade student attitudes toward mathematic problem solving. He included statements reflecting children's beliefs about the nature of various types of mathematical problems/ the nature of the problem solving process/ the desirability or persevering when solving a problem/ and the value of generating several ideas for solving a problem. He correlated student attitudes toward problem solving with their scores in a mathematical test which yielded a comprehension score/ an application score/ and a problem solving score. He found a significant positive relationship between problem solving performance and student attitude scores on the subscale which measured reactions to such things as problem solving techniques or problem situations/ or to the frustation or anxiety experienced when confronted with problem solving situations. In another part of this study/ Whitaker investigated the relationship between the attitudes of 4th grade teachers toward problem solving and their students' performance in problem solving. A very weak and nonsignificant negative correlation was found between the teacher's attitudes toward problem solving and student performance. I certify that opinion it conforms presentation and is as a dissertation for I have read this study and that in my to acceptable standards of scholarly fully adequatef in scope and qualityf the degree of Doctor of Philosophy. is Algina/ Cl iessor of Foi irman dations of Education I certify that opinion it conforms presentation and is as a dissertation for I have read this study and that in my to acceptable standards of scholarly fully adequatef in scope and qualityf the degree of Doctor of Philosophy. Linda Crocker Professor of Foundations of Education I certify that opinion it conforms presentation and is as a dissertation for I have read this study and that in my to acceptable standards of scholarly fully adequatef in scope and qualityf the degree of Doctor of Philosophy. .chael Professor r njary of Educational Leadership This dissertation was submitted to the Graduate Faculty of the College of Education and to the Graduate Schoolf and was accepted as partial fulfillment for the degree of Doctor Augustf Deanf Graduate School 10 of the studies consider sex differences incidental to the major study findings. The available literature offers very little research directly related to the problem of sex differences in this area. The review of the literature has been divided in four sections. The first section consists of a detailed summary of the available research on sex diferences in incorrect responses. The second section deals with sex-related differences in problem solving performance. These sections are directly related to the objectives of the study. The third section is more peripheral/ and contains a discussion of the more prevalent issues about the influence of cognitive and affective variables on sex differences in mathematics learning and achievement. The fourth is a summary of the research dealing with variables considered as influential to mathematics problem solving performance. Sex-related Differences in incorrect^Response Patterns Research findings tend to suggest that boys and girls may be approaching problem solving differently (e.g./ Fennema and Sherman# 1978; Marshall# 1981# 1983; Meyer# 1978# among others). Marshall (1981) investigated whether 6th grade boys and girls approach mathematical problem solving with different strategies. Her specific interest was whether the sexes made the same errors. 60 Fienberg also shows that deleting the Ujk term from (3.2) to obtain log mij k = Ui + Uj+Uk + Uik. (3.3) yields a model that specifies that option is independent of sex. Based on Fienberg1 s presentation an appropriate analysis for testing the hypotheses is 1. Conduct a likelihood ratio test of the adequacy of model (3.2). If this test is nonsignificant then the data support the adequacy of the model and the hypothesis that conditional on gender option and year are independent. Because model (3.1) is a saturated model testing the adequacy of model (3.2) is the same as comparing the adequacies of models (3.1) and (3.2). 2. Conduct a likelihood ratio test comparing the adequacies of models (3.2) and (3.3). If this test is significant then model (3.2) fits the data better than model (3.3) and the data support the hypothesis that the choice of incorrect option is dependent on sex. To summarize if the first test is nonsignificant and the second test is significant then the choice of option is dependent on sex and the pattern of dependency is the same for all three years. 117 From May 1982 to July* 1983 she worked as graduate research assistant at the Foundations of Education Department University of Florida. She returned to Puerto Rico in August 1983 to serve as Special Aide to the Assistant Secretary for the Vocational Technical and High Skills Educational Programs of the Commonwealth Department of Education. 110 Fennema E.r & Sherman/ J. (1977). Sex-related differences in mathematics achievement/ spatial visualization/ and affective factors. American Educational Research Journal/ 14/ 51-71. Fennema/ E. / & Sherman/ J. A. (1978). Sex-related differences in mathematics achievement and related factors. A further study. Journal for Research in Mathematics...Education* h 189-204. Fennema/ E./ & Tartree/ L. A. (1983). Research on relationship of._spatisl...visualizatiQD,_and.confidence to male/female mathematics achievement in grades 6-8. Washington/ DC: National Science Foundation. (ERIC Document Reproduction Service No. ED 232 853) Fienberg/ s. e. (1980). Th.e...Analys.is..pf.CcQS.s=.ciaasified Categorical Data. Cambridge/ MA: The MIT Press. Fox/ L. H. (1975a). Mathematically precocious: Male or female. In E. Fennema (Ed.) Mathematics learning: what_rÂ£SÂ£arch--say-S-3bQut sea;_differences (pp. 1-12). Columbus: Ohio State University/ College of Education. (ERIC Document Reproduction Service No. ED 128 195) Fox/ L. H. (1975b). Sex differences: Implications for program_planning for..the academically giftedt Paper presented at the Lewis M. Terman Memorial Symposium on Intellectual Talent held at the Johns Hopkins University/ Baltimore/ MD. Fox/ L. H. (1977) The effects of sex role socialization on mathematics participation and achievement. In J. shoemaker (Ed.)/ women_and...Mathema.tics: Research Perspectives for Change/ (N.I.E. Papers in Educa tion and Work: No. 8). Washington/ D. C.: Eduation and Work Group/ The National Institute of Education. Glennon/ V. J./ & Callahan/ L. G. (1968) Elementary school mathematics; A^guide tc current research/ Washington/ D. C.: Association for Supervision and Curriculum Development. Guay/ R. B. / & McDaniel/ D. (1977). The relationship between mathematics achievement and space ability among elementary school children. Journal for Research.-in.-Mathamatics Education* Â£* 211-215. 18 students who had taken a similar number of mathematics courses and were in the top 85% of the class in mathematics achievement. They were tested in 1976. Four high school districts were included. In only one of the high school districts were sex-related differences in application and problem solving found* in favor of males. They concluded that when relevant factors are controlled* sex-related differences in favor of males do not appear often* and when they do* they are not large. Sherman (1980) investigated the causes of the emerging sex-related differences in mathematics performance* in favor of males* during adolescence (grades 8-11). She wanted to know if these differences emerge as a function of sex-related differences in spatial visualization and sociocultural influences that consider math as a male domain. In grade 8* she used the Romberg-Wearne Test and* in grade 11* a mathematical problem solving test derived from the French Kit of Tests. The analysis showed that for girls* problem solving performance remained stable across the years. Mean problem solving performance for boys* however* was higher in grade 11 than in grade 8. No sex-related differences were found in grade 8* but boys outperformed girls in grade 11* where the Stafford test was used. Sherman found that for both sexes problem solving performance in grade 8 was the best predictor of problem TABLE OF CONTENTS Page ACKNOWLEDGMENTS ii LIST OF TABLES V LIST OF FIGURES vii ABSTRACT viii CHAPTER I INTRODUCTION 1 Purpose of the Study 6 Significance of the Study 6 Organization of the Study 8 II REVIEW OF THE LITERATURE 9 Sex-related Differences in Incorrect Response Patterns 10 Sex-related Differences in Problem Solving 14 Cognitive and Affective Variables that Influence Sex Differences in Mathematics Learning and Achievement 23 Differences in Formal Mathematics Education 24 Differences in Spatial Ability 26 Differentiated Effect of Affective Variables 30 Problem Solving Performance and Related Variables 37 Computational Skills and Problem Solving Performance 38 Reading and Problem Solving Perfor mance 44 Attitudes Toward Problem Solving and Problem Solving Performance 50 iii MATH CONCEPTS Fig. 3.1 Number of Questions by Skill Area in the Basic Skills Test in Mathematics 6 xml version 1.0 encoding UTF-8 REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd INGEST IEID E7GIIQLHZ_4RJU7A INGEST_TIME 2012-09-24T12:49:43Z PACKAGE AA00011802_00001 AGREEMENT_INFO ACCOUNT UF PROJECT UFDC FILES 28 1974 have failed to support these sex differences in spatial abilities (Armstrong/ 1979; Connor/ Serbin, & Schackmanr 1977; Fennema & Sherman/ 1978; Sherman/ 1979). Fennema and Sherman (1978) and Sherman (1979) have explored sex-related differences in mathematical achievement and cognitive and affective variables related to these differences. In a study involving students from grades 6/ 7/ and 8/ from four school districts/ Fennema and Sherman found that spatial visualization and problem solving were highly correlated for both sexes (.59 and .60). Even in the school district where sex differences were found in problem solving/ no significant sex-related differences were found in spatial visualization. When Sherman (1980) compared groups of males and females in two different grades/ 8 and 11/ she found no sex-related differences in problem solving or in spatial visualization in grade 8. In grade 11/ however/ although the sexes differed in their problem solving performance/ no sex-related differences were found in spatial vizualization. Even though spatial visualization in grade 8 was the second best predictor of problem solving performance in grade 11/ sex differences in grade 11 were not a result of spatial visualization since no differences were found in that skill. In spite of the fact that no sex differences were found in spatial abilities/ it is evident that males and 3 indicated no significant sex differences for 8th grade students on measures of problem solving or algebra. However* females outperformed males in computation and spatial visualization. For the 12th grade students* statistically significant sex differences favoring males were found in problem solving* but not in algebra* computation* or spatial visualization. For males and females who had enrolled in courses beyond general mathematics and who had taken or were enrolled in courses such as pre-calculus* calculus* or geometry* differences in problem solving or spatial visualization did not exist. Sex differences favoring males were found on a total score obtained summing across the computation* problem solving* and algebra subtests (Armstrong* 1979). The mathematics data collected in the second survey by the 1978 National Assessment of Educational Progress showed significant sex differences for both 13- and 17-year-old students. The 13-year-old females outperformed males in the computational subtest and males outscored females by 1 1/2 percentage points in problem solving (statistically significant). No statistically significant differences were found in algebra. No sex differences were found for the 17-year-old group either in the computation subtest or in the algebra subtest. Males surpassed females in problem solving. A reanalysis of the data from the 17-year-old group confirmed male superiority in problem solving after controlling for 113 Meyer# R. A. (1978). Â£g.xrJg.la.ted_diffgJ.gi3c.es.j,p_math.ema:- ti,cal_problem_ solving performance and intellectual abilities. Madison: University of Wisconsin# Research and Development Center for Cognitive Learning. (ERIC Document Reproduction Service No. ED 166 023) . Milton# G. A. (1957) The effects of sex-role identifica- cation upon problem solving skill. Journal of Abnormal & Social.Psychology# 55.# 208-212. Moller# N. J. (1983). The impact of gender# masculinity# and feminity on math achievement and course decisions (Doctoral dissertion# Purdue University# 1982). Hiss ej.t a.tion,Abstrac.te_ International# l# 2584A. Muth# K. D. (1984). Solving arithmetic word problems: Role of reading and computational skills. Journal of Educational. Psychology# 15# 205-210. Pallas# A. M.# & Alexander# K. L. (1983). Sex differences in quantitative SAT performance: New evidence on the differential coursework hypothesis. American Idji^atj^p.aJ.,Res.ea_rc.li.Journal* 21# 165-182. Plank# E.# & Plank# R. (1954). Emotional components in arithmetical learning as seen through autobiographies. Psychological Studies of the Child# Â£# 274-296. Porter# A. C. (1968). The effects of using fallible variables in the analysis of covariance (Doctoral dissertation# University of Wisconsin# 1967). Di.s.s.er.tatifih-^abgtjact.,.Inte.r.P,a-t.ip.nai# 25# 3517B. Probert# B. S. (1984). Math confidence workshops: A multimodal group intervention strategy in mathematics anxiety/avoidance. (Doctoral dissertation# University of Florida# 1983). nisÂ£.er.tatiQp_Abgtracts International# 2231B. Puerto Rico Department of Education. (1980). Basic skills,test in mathematics-6. Hato Rey: Center of Evaluation. Radatz# H. (1979) Error analysis in mathematics educa tion. jQumai fo.r.jtene.arch .in Majieinatl_c,s_fcjuc.a- tion# lfi.# 163-172. Riedesel# A. C. (1969) Problem solving: Some suggestions from research. Arithmetic Teacher# 15# 54-58. 13 III) to investigate sex related differences in the selection of incorrect responses/ and the consistency of such differences over three years of administration of the test. Based on her findings that sex differences were found in 80% of the items/ Marshall classified the students' errors according to Radatz' (1979) five-category error classification. The categories are language (errors in semantics)/ spatial visualization/ mastery/ association/ and use of irrelevant rules. It was found that girls' errors are more likely to be due to the misuse of spatial information/ the use of irrelevant rules/ or the choice of an incorrect operation. Girls also make relatively more errors of negative transfer and keyword association. Boys seem more likely than girls to make errors of perseverance and formula interference. Both sexes make language-related errors/ but the errors are not the same. Available research is not extensive enough to make definite judgments about the sex-related differences observed in incorrect responses. Clearly more research is needed. 105 exr_related.,Differences in Problem Solving Performance The results of this study do not give strong support to Marshall's (1981/ 1984) findings that the effect of computational skills on problem solving performance is different for each sex. However/ the findings of the present investigation are consistent with those found by Meyer (1978) and Whitaker (1976): female performance in problem solving is not significantly different from male performance. Females showed superiority in problem solving performance in all three years of test administration although their superiority was retained only when performance in equivalence was controlled for (for the three years data) and when performance in subtraction was controlled for (second year) (statistically significant only after accounting for equivalence) . The findings of this study are also in agreement with those reported by the National Assessment of Educational Progress (NAEP) (Carpenter et al.z 1980). Related to problem solving/ they conclude/ if it were necessary to single out one area that demands urgent attention/ it would clearly be problem solving. At all age levels/ and in vir tually every content area/ performance was extre mely low on exercises requiring problem solving or application of mathematical skills. In gene ral/ respondents demonstrated a lack of the most basic problem-solving skills. Rather than attempt ing to think through a problem and figure out what needed to be done to solve the problem/ most res pondents simply tried to apply a single arithmetic 17 attitude toward mathematical problem solving* and related sex and program-type differences. Although his main objective was to construct an attitude scale to measure attitudes toward problem solving* his study is important because his findings support Meyer's regarding the lack of significant sex-related differences in problem solving performance. Performance in the problem solving questions* for both males and females* was much lower than performance in the application questions* and much lower than performance in the comprehension questions. In fact* the mean score for each part of the item* for both males and females* was almost identical to the mean scores obtained by males and females in Meyer's study. Whitaker noted that each application item is more difficult than its preceding comprehension item* and that each problem solving item is more difficult than its preceding application item. No significant sex-related differences were found for any of the three parts of the item (comprehension* application* or problem solving). Fennema and Sherman (1978) investigated sex-related differences in mathematics achievement and cognitive and affective variables related to such differences. They administered the Romberg-Wearne Problem Solving Test to a representative sample of 1320 students (grades 6-8) from Madison* Wisconsin* predominantly middle-class* but including great diversity in SES. The sample consisted of 26 maintained their superiority in mathematics ability during high school. While Fox attributed sex-related differences in mathematical achievement to differential exposure to mathematical games and activities outside school Benbow and Stanley suggested that sex-related differences in mathematics performance stem from superior mathematical ability in males not from differences in mathematics formal education. The differential coursework hypothesis is not totally convincing and as reporteu berore it has been challenged by researchers such as Benbow and Stanley (1980). However Pallas and Alexander (1983) have questioneu the generalizanility of Benbow and Stanley's findings based on the fact that they used highly precocious learners. The differential coursework hypothesis can be accepted only as a partial explanation of differences in mathematics performance found between the sexes. JUJ-Â£Â£j^gÂ£.s-Jn_Sp.fltial Ability The basic premise in this issue is that males and females differ in spatial visualization and this explains differential mathematics learning and achievement. Until recently sex differences in spatial ability in favor of males were believed to be a fact and were thought by some to be related to sex differences in mathematical achievement CHAPTER II REVIEW OF THE LITERATURE Sex differences in mathematics learning and achievement have been a subject of concern for educators and psychologists. Many studies found in the literature support the existence of these differences. Boys show superiority in higher level cognitive tasks (problem solving or mathematical reasoning) in the upper elementary years and in the early high school years (Fennema/ 1974; Maccoby & Jacklin/ 1974) . Almost all the research carried out has dealt with analysis of total correct scores in mathematics aptitude and achievement tests or scores in subtests. The literature related to sex differences in incorrect responses the main subject of the present study/ is surprisingly sparse. For the most part/ the studies have investigated the differences between the sexes in mathematics learning and achievement and the underlying variables causing the differences. Cognitive and affective variables have been the matter at issue in the establishment of sex differences. Although research in mathematics problem solving/ the secondary subject of this investigation/ is extensive/ most 9 3 indicated no significant sex differences for 8th grade students on measures of problem solving or algebra. However/ females outperformed males in computation and spatial visualization. For the 12th grade students/ statistically significant sex differences favoring males were found in problem solving/ but not in algebra/ computation/ or spatial visualization. For males and females who had enrolled in courses beyond general mathematics and who had taken or were enrolled in courses such as pre-calculus/ calculus/ or geometry/ differences in problem solving or spatial visualization did not exist. Sex differences favoring males were found on a total score obtained summing across the computation/ problem solving/ and algebra subtests (Armstrong/ 1979). The mathematics data collected in the second survey by the 1978 National Assessment of Educational Progress showed significant sex differences for both 13- and 17-year-old students. The 13-year-old females outperformed males in the computational subtest and males outscored females by 1 1/2 percentage points in problem solving (statistically significant) No statistically significant differences were found in algebra. No sex differences were found for the 17-year-old group either in the computation subtest or in the algebra subtest. Males surpassed females in problem solving. A reanalysis of the data from the 17-year-old group confirmed male superiority in problem solving after controlling for 79 TABLE 4.2 Continued Item Num ber Xa for Model 3.2 Xz for Model 3.3 Xz for Difference Actual Sample Size Minimum Sample Size for Signif. (a) lb) . (c) . (d)... (e) (f) 75 8.69 237.07 228.38* 67/783 5/963 76 99.68 167.72 68.04* 70/183 20/723 77 79.38 432.36 352.98* 66/304 3/774 78 64.56 127.04 62.48* 76/806 24/696 79 23.98 72.59 48.61* 48/690 20/123 80 75.01 128.13 53.12* 55/277 20/906 81 95.51 106.82 11.31* 57/710 102/511 82 51.99 108.12 56.13* 71/315 25/525 83 20.12 95.18 75.06* 81/021 21/685 84 38.62 69.05 30.43* 78/959 52/129 86 102.85 433.47 330.62* 39/903 2/425** 87 66.75 430.40 363.65* 48/194 2/662** 88 63.01 126.27 63.26* 38/742 12/304 89 43.26 289.14 245.88* 39/648 3/240 90 85.53 405.04 319.51* 53/557 3/368 91 82.53 85.48 2.95 62/954 428/727 92 16.36 93.12 76.76* 88/231 23/092 93 78.39 229.02 150.63* 80/723 10/766 94 70-15 137.27 67.12* 87/241 26/113 41 categories. Category I included clerical and computational errors. Category II included other types of errors* such as average and area errors/ use of wrong operation* no response* and erred responses offering no clues. It the student's work indicated the correct procedure and yet the problem was missed because of a computational or clerical error* it was assumed that the problem was read and understood. An analysis of frequencies showed that clerical errors were responsible for 3% of the problems incorrectly solved* computational errors accounted for 49%* and other errors for 48% of the erred problems. Knifong and Holtan concluded that "improved computational skills could have eliminated nearly half of the word problem errors" (p. Ill). These computational errors were made in a context where other skills such as reading* interpretation of the problem* and integration of these skills necessary for the solution of word problems* might interact. However* Knifong and Holtan state that their findings neither confirm nor deny that improvement of reading skills will lead to improvement in problem solving. They conclude that "it is difficult to attribute major importance to reading as a source of failure" (p. 111). In a later analysis* looking for evidence of poor reading abilities affecting children's success in word problems* Knifong and Holtan (1977) interviewed the 59 Inspection of the year by option contingency tables shows that year and option are independent for each gender. Thus/ hypothesis 1 is equivalent to the hypothesis that year and option are independent conditional upon gender. Hypothesis 2 is also true for Table 3.1 and Table 3.2. Therefore/ when hypothesis 1 is correct/ hypothesis 2 is equivalent to the hypothesis that sex and option choice are dependent. In his discussion of the analysis of three dimensional contingency tables/ Fienberg (1980) presents the following saturated model for the data: log mijk = Ui + U j + jc+i j + Uik+Uj k+Ui j k- (3.1) In this model/ mjk is the expected value of the frequency in cell ijk of the three dimensional table. The model states that all three classification factors for a three dimensional contingency table are mutually dependent. In the present research i is the year index/ j is the option index/ and k is the sex index. Fienberg shows that deleting the terms Uj and U^jk yields a model in which year and option are independent conditional upon sex. This model is log mij k Ui+Uj+Uk+Uik+Ujk (3.2) 29 females may use them in a different way. Meyer (1978) with an elementary grade sample and Fennema and Tartree (1983) with an intermediate level sample found that the influence of spatial visualization on solving mathematics problems is subtle and that males and females use their spatial skills differently in solving word story problems (problems that measure problem solving ability or reasoning). Fennema and Tartree (1983) carried out a three-year longitudinal study which showed that girls and boys with equivalent spatial visualization skills did not solve the same number of items nor did they use the same processes in solving problems. The results also suggested that a low level of spatial visualization skills was a more debilitating factor for girls than for boys in problem solving performance. Landau (1984) also investigated the relationship between spatial visualization and mathematics achievement. She studied the performance of middle school children in mathematical problems of varying difficulty and the extent to which a diagramatic representation is likely to facilitate solution. She found that spatial ability was strongly correlated to mathematical problem solving and that the effect of spatial ability was more influential for females. Females made more use of diagrams in the solution of problems reducing the advantage of males over females in problem solving performance. Problem Solving 89 Subtraction Fig. 4.3 Sex by Subtraction Interaction 44 In a more recent investigation? Exedisis (1983) studied the contribution of reading ability* vocabulary* mathematical concepts* computation* sex* and race on problem-solving performance. The Iowa Test of Basic Skills was administered to a group of 6th* 7th* and 8th grade anglo and black Chicago male and female adolescents. Problem solving was highly correlated to an understanding of basic mathematical concepts* somewhat correlated to race* and weakly correlated to computational and vocabulary skills* sex* and reading ability. Although the findings of these studies show a relationship between computational skills and problem solving achievement* this relationship is not strong enough to be considered the most determinant factor in problem solving achievement* as some of the researchers have been careful to point out. In spite of the dismissal of reading as a determinant factor in problem solving achievement by some of these same researchers* more recent studies in this area have led others to hold different views. Headin_g_and_J)r obiem_ Solving Performance Martin (1964) studied the contribution of reading comprehension* computation* abstract verbal reasoning* and arithmetic concepts to arithmetic problem solving performance. Fourth and 8th grade students were administered the Iowa tests of Basic Skills and the Lorge-Thorndlike intelligence test (verbal). 38 At present no set of variables has been clearly established as a determinant of problem difficulty. Several researchers have investigated the effect of reading and computation on problem solving performance. Others have studied the effect of student attitudes toward problem solving in problem solving learning and achievement. Typically/ correlational methods have been used to investigate these questions. Computational Skills and Problem.Solving Performance One of the first researchers to study the effect of computation and reading on problem solving performance was Hansen (1944). He investigated the relationship of arithmetical factors/ mental factors/ and reading factors to achievement in problem solving. Sixth grade students were administered tests in problem solving and categorized as superior achievers (best problem solvers) and inferior archievers (poorest problem solvers). The two groups were compared in selected factors believed to be related to success in arithmetic problem solving: arithmetical/ mental and reading factors. After controlling for mental and chronological age/ the superior achievers in problem solving surpassed the inferior achievers in mental and arithmetical factors. The superior group did better in only two of the six items under the reading factors: general language ability and the reading of graphs/ charts/ and tables. 26 maintained their superiority in mathematics ability during high school. While Fox attributed sex-related differences in mathematical achievement to differential exposure to mathematical games and activities outside school/ Benbow and Stanley suggested that sex-related differences in mathematics performance stem from superior mathematical ability in males not from differences in mathematics formal education. The differential coursework hypothesis is not totally convincing and/ as reporteu berore/ it has been challenged by researchers such as Benbow and Stanley (1980). However/ Pallas and Alexander (1983) have questioneu the generalizaoility of Benbow and Stanley's findings based on the fact that they used highly precocious learners. The differential coursework hypothesis can be accepted only as a partial explanation of differences in mathematics performance found between the sexes. Differenees in Spatial Ability The basic premise in this issue is that males and females differ in spatial visualization and this explains differential mathematics learning and achievement. Until recently/ sex differences in spatial ability in favor of males were believed to be a fact and were thought by some to be related to sex differences in mathematical achievement. 108 Bern* S. L. & Bern D. V. (1970). We are all nonconscious sexists. Psychology Todays 22-24* 26/ 115-116. Benbow* C. P.* & Stanley* J. C. (1980). Sex differences in mathematical ability: Fact or artifact? Science* 210 1262-1264. Benbow* C. P.* & Stanley* J. C. (1982). Consequences in high school and college of sex differences in mathema tical reasoning ability: A longitudinal perspective. American Educational Research Journal* 12* 598 622. Bobbie* C. N. (1971). Sex-role preference and academic achievement (Doctoral dissertation* Yeshiva Univer sity) Â£iss.ejtation Abstracts International* 22* 1818B-1819B. Carey* G. L. (1958). Sex differences in problem solving performance as a function of attitude differences. Journal of Abnormal and Social Psychology* 2Â£* 256-260. Carpenter* T. P.* Corbitt* M. K.* Kepner* H. S.* Lindquist* M. M. & Reys* R. E. (1980). Results of the second NAEP mathematics assessment: Secondary school. Mathematics Teacher* 73 329-338. Carpenter* T. P. Lindquist* M. M. Mathews* W. & Silver* E. A. (1984). Results of the third NAED mathematics assessment: Secondary school. Mathematics Teacher* 16.* 652-659. Chase* C. I. (1960). The position of certain variables in the prediction of problem solving in arithmetic. Journal of Educational Research* 24/ 9-15. Cohen* D.* & Wilkie* F. (1979). Sex-related differences in cognition among the elderly. In M. A. Witting and A. C. Petersen (Eds.)* Sex-related differences in cognitive functioning: Developmental Issues* New York: Academic Press. Connor* J. M. Serbin* L. A.* & Schakman* M. (1977). Sex differences in children's response to training on a visual-spatial test. Developmental Psychology* 12* 392-394. 37 dealing with both cognitive and affective variables* there are no clear-cut findings to render unequivocal support to a particular variable as accounting for these sex-related differences. However* everything seems to point to the fact that affective* rather than cognitive variables play a more significant role in the sex-related differences observed in mathematics performance and learning. In most of the studies dealing with affective variables* findings consistently show that these factors influence mathematics performance in females more than in males. In at least one area* confidence as learners of mathematics* Sherman (1980) found that this variable influenced course election more than all the cognitive variables previously discussed. The case for the societal influences on sex roles and expectations to account for the differences in mathematics learning is also supported in one way or another in the studies reported in the literature. Problem Solving Performance _and_Rela.ted Variables Problem solving has been perhaps the most extensively researched area in mathematics education. Published reviews by Kilpatrick (1969)* Riedesel (1969)* and Suydam and Weaver (1970-1975) attest to this. Much of the research done has focused on identifying the determinants of problem difficulty and the problem features that influence the solution process. LIST OF FIGURES Figure Page 3.1 Number of Questions by Skill Area in a Basic Skills Test in Mathematics-6 56 4.1 Sex by Multiplication Interaction 87 4.2 Sex by Division Interaction 88 4.3 Sex by Subtraction Interaction 89 vii 8 achievement. Since there are no investigations reported in sex differences in item response patterns in Puerto Rico research is needed. Organization of the Study A review of the literature on sex differences in mathematics performance is reported in Chapter II. The research methodology is presented in Chapter III. Research questions sample instrument and data analysis are discussed in that chapter. Chapter IV is an exposition of the results of the study. Chapter V contains a summary and interpretation of the results of the study and the implications of the findings together with suggestions for further research. 23 computation grew larger. This difference grew larger for girls than for boys. Although SES was a major factor in solving computation and story problem items successfully/ the effect was similar for each sex. Sex-related differences by primary language or chronological age were not large. This research carried out by Marshall with elementary grade children supports previous research findings that males are better than females in mathematics problem solving (a higher order skill) and females are better than their counterpart males in computation (a lower level skill). Marshall's research also brought out a different aspect of this question: the notion that girls find it more necessary than boys to succeed in the computation items in order to successfully solve the story problem items. Â£ognitive_and_Affective Factors That Influence Sex Differences in Mathematics Learning and Achievement The research reviewed in the literature does not provide evidence of any unique variable that could serve as an explanation for the observed sex-related differences in mathematics learning and achievement. However/ some issues have been discussed/ among which the most prevalent are that sex-related differences in mathematics learning and achievement are a result of differences in formal education; that sex-related differences in mathematics 115 Whitaker, D. R. (1976). A study of the relationship between selected noncognitive factors and the problem solving performance of fourth-grade children. (Tech. Rep. No. 396). Madison: University of Wisconsin/ Research and Development Center for Cognitive Learning. Wiser L. K.r Steel L.r & MacDonald/ C. (1979). Origins and career consequences of sex differences in high school mathematics achievement (Grant No. NEI-G 7 8-001 ) Palo Alto/ CA : American Institutes for Research. Zalewski/ D. L. (1974). An exploratory study to compare two performance measures: An interview-coding scheme of mathematical problem solving and a written test (Doctoral Dissertation/ University or Wisconsin/ Madison). Dissertation Abstracts International/ 15./ 3797A. 91 TABLE 4.7 Reliability of the Covariates for Each of the Three Years of Test Administration Covariate First Year Second Third Addition .579 .610 .454 Subtraction .756 .724 .737 Multiplication .729 .729 .734 Division .716 .680 .714 Fracadd .709 .728 .748 Decsub .720 .716 .742 Equivalence .421 .394 .488 skills/ sex-related differences in problem solving performance do not exist. Females retained their superiority in problem solving when equivalence (in all three years) and subtraction (in one year) were the controlling variables. The question of whether male-female differences in problem solving depend on computational skills was answered/ partially/ in the affirmative. x 110 Fennemar E. & Shermanr J. (1977). Sex-related differences in mathematics achievements spatial visualizations and affective factors. American Educational.Research Journals 14' 51-71. Fennemas E.s & Shermans J. A. (1978). Sex-related differences in mathematics achievement and related factors. A further study. Journal for Research in Mathematics .Education/ 9s 189-204. Fennemas E.s & Tartrees L. A. (1983). Research on relationship of spatial visualization and confidence to male/female mathematics achievement in grades 6-8. Washingtons DC: National Science Foundation. (ERIC Document Reproduction Service No. ED 232 853) Fienbergs S. E. (1980). The Analysis of .Cross-Classified Categorical Data. Cambridges MA: The MIT Press. Foxs L. H. (1975a). Mathematically precocious: Male or female. In E. Fennema (Ed.) Mathematics learning: What research says.about sex differences (pp. 1-12). Columbus: Ohio State Universitys College of Education. (ERIC Document Reproduction Service No. ED 128 195) Foxs L. H. (1975b). Sex differences: Implications for program planning for the academically gifted. Paper presented at the Lewis M. Terman Memorial Symposium on Intellectual Talent held at the Johns Hopkins University* Baltimore* MD. Fox* L. H. (1977). The effects of sex role socialization on mathematics participation and achievement. In J. Shoemaker (Ed.)* Women_and_Mathe.mati,csResearch Perspectives for Change* (N.I.E. Papers in Educa tion and Work: No. 8). Washington* D. C. : Eduation and Work Group* The National Institute of Education. Glennon* V. J.* & Callahan* L. G. (1968) Elementary school mathematics: A_guide._tp_..cur rent research* Washington* D. C.: Association for Supervision and Curriculum Development. Guay* R. B.* & McDaniel* D. (1977). The relationship between mathematics achievement and space ability among elementary school children. Journal for Research in Mathematics Education* 8* 211-215. 61 A problem arises in interpreting this analysis with the data used in this research. Over the three years? there were responses available from 135/340 students. Even if only 5% of the students answered an item incorrectly/ the responses of 7/767 students would be used in analyzing this item. On the other hand/ if 90% answered an item incorrectly/ the responses of 121/806 students would be used in analyzing the data. As a result of the large sample size/ the tests described above are likely to be very powerful. In step 1 of the analysis/ then/ even a very small change from year to year in the proportion of males or females who choose an option is apt to be detected/ and the results will indicate that hypothesis 1 is not supported. For step 2/ even a very small dependence of option choice on sex is likely to be detected and hypothesis 2 is likely to be supported. In brief/ the problem caused by the large sample size is that practically insignificant differences may yield statistical significance. Fortunately/ the form of the test statistic used in the likelihood ratio test suggests a reasonable solution to this problem. The test statistic is - 2 ^ F0 l9e (F0/Fe). (3.4) Ill Hansen* C. W. (1944) Factors associated with successful achievement in problem solving in sixth grade arithmentic. journal qÂ£ Educational,-Researoh' 22.1 111-119. Harnisch* D. L. (1983). Item response patterns: Application for educational practice. Journal of Educational Measurement* 22* 191-205. Hilton* T. L. & Berglund* G. W. (1974). Sex differences in mathematics achievement: A longitudinal study. Journal of Educational Research* Â£2* 231-237. Hunter* J. E. & Cohen* S. H. (1974). Correcting for unreliability in nonlinear models of attitude change. Psvchometrika* 22* 445-468. Jarvis* O. T. (1964). Boy-girl ability differences in elementary school arithmentic. School Science and Mtkemati_c_s/ 4, 657-659. Kaufman* A. (1984). Sex differences in mathematics reasoning in mathematically gifted and average fifth and sixth graders (Doctoral dissertation* Hofstra University* 1983) Dissertation Abstracts Inter- JjafcionflJU 15-' 1094A-1095A. Kilpatrick I. (1969). Problem solving and creative behavior in mathematics. In J. W. Wilson* & L. R. Carry (Eds.)* Studies in mathematics (Vol. No. 19). Stanford* CA: School Mathematics Study Group. Kloosterman* P. W. (1985). Attributional theory* learned helplessness* and achievement in ninth grade mathematics (Doctoral dissertation* University of Wisconsin* 1984) Dissertation Abstracts Internatio- nal, 4Â£/ 919A-920A. Knifong* J. D.* & Holtan* B. (1976). An analysis of children's written solutions to word problems. Journal .for Research in Mathematics Education* 1* 106-112. Knifong* J. D.* & Holtan* B. (1977). A search for reading difficulties among erred word problems. Journal for 2esÂ£.arch_, in Mathematics .Education* 2* 227-230. 50 The studies reviewed in this section show a positive relationship between reading and problem solving performance/ but in the case of Ballew and Cunningham (1982)/ this relationship is not viewed singly but rather as one among the interacting factors that produce successful problem solving. The third variable reviewed is the effect of student attitudes toward problem solving on problem solving performance. Many researchers have tried to demonstrate that this variable is a determinant factor in problem solving achievement. Attitudes Toward Problem Solving and Problem Solving Performance Research studies support the existence of positive and rather stable relationships between student attitudes and achievement in mathematics. Aiken (1970) has suggested that an individual's attitude toward one aspect of the discipline (mathematics) / such as problem solving/ may be entirely different from his/her attitude toward another phase of the discipline/ such as computation. Research/ however/ has been directed to the use of single/ global measures of attitudes toward mathematics rather than to the investigation of attitudes toward a particular phase of the discipline. The studies described below are only part of the few investigations which have examined the relationship between 96 TABLE 4.11 Adjusted Means on Problem Solving* by Covariate and Sex Second Year Adjusted Means Covariate Males Females Addition 3.69 3.96 Multiplication 3.80 3.85 Division 3.80 3.85 Fracadd 3.78 3.87 Decsub 3.90 3.75 Equivalence 3.46 44.18 Log-linear models were used in the analysis of incorrect responses. The results of the analyses showed that for 100 of the 111 items of the testr males and females selected different incorrect options/ and this pattern of responses was consistently found during the three years of test administration. However/ for the vast majority of the 100 items the male-female differences were relatively small/ considering the fact that the number of subjects needed to obtain statistical significance was very large. The responses of approximately 1/000 randomly selected students per academic year were analyzed in the comparison of male and female performance in problem solving. Females outperformed males in problem solving and in six of the seven computational variables. Males showed superiority in equivalence in all the three years/ but statistical significance was obtained in only one of the years. Analysis of covariance (ANCOVA) was used in the comparison of male and female performance in problem solving after controlling for computational skills. Seven analyses of covariance tests were conducted/ one for each of the covariates. Estimated true scores for observed scores were used in the analyses. The results tend to show that for examinees with similar levels of computational IX 11 She analyzed the responses of 9>000 boys and 9>000 girls to 6 selected items> 2 computation items> and 1 story problem item from each of 2 of the 16 forms of the Survey of Basic Skills test administered during the academic year 1978-79. The Survey is a 30-item achievement test administered every year to all 6th grade children in California through the California Assessment Program. There are 16 forms of the test> to which approximately 9 >000 boys and 9 >000 girls respond each year. Of the 160 mathematics items contained in the 16 forms of the test> 32 are on measurement and graphing> 28 on number concepts> 28 on whole number arithmetics> 20 on fraction arithmetics> 20 on decimal arithmeti cs > 20 on geometry > and 12 on probability and statistics. The item analysis performed on the 1978-79 data showed that boys and girls tended to select different incorrect responses. In the first computation item (Form 1 of the test) both sexes reflected similar mistakes in carrying> but in different columns. In the second computation item> both sexes ignored the decimal points and selected the same incorrect response. However> more girls than boys chose this response. In the first computation item (Form 2 of the test) the incorrect choice of both sexes was option Â£> but the second most frequently selected option was for boys and b for girls. In the second computation item of this 63 model tests were conducted using a .01 level of significance. Since this research is based on Marshall (1981/ 1983)/ it is important to compare the method of analysis used in this study to the one used by Marshall. Marshall also used a two-step analysis. In the first step of her analysis she deleted the Uijk term from (3.1) and tested the adequacy of the model/ 1 g m j. j k = Ui + Uj + Ufc+Uij + Uik+Ujk. (3.6) Following this/ she deleted the Ujk term to obtain log mij = Ui+Uj + Uk+ij + Uik. (3.7) and compared these two models using a likelihood ratio test. If the first test was nonsignificant and the second significant/ Marshall claimed that option choice was dependent on sex and the pattern of dependency was the same from year to year. This is the same claim that this study sought to establish. However/ the approach used here was to present evidence that model (3.2) fits the data while Marshall tried to show that model (3.6) fits the data. The major difference between the two approaches concerns the operationalization of the concept that the 115 Whitaker/ D. R. (1976). A study of the relationship between selected noncognitive factors and the problem solving performance of fourth-grade children. (Tech. Rep. No. 396). Madison: University of Wisconsin/ Research and Development Center for Cognitive Learning. Wise/ L. K./ Steel/ L./ & MacDonald/ C. (1979). Origins and career conseguences of sex differences in bigh__s_chool mathematics achievement (Grant No. NEI-G 78-001). Palo Alto/ CA : American Institutes for Research. Zalewski/ D. L. (1974). An exploratory study to compare two performance measures: An interview-coding scheme of mathematical problem solving and a written test (Doctoral Dissertation/ University or Wisconsin/ Madison). Dissertation Abstracts International/ 15./ 3797A. 27 Research findings in this area have been inconsistent. In 1966 / Maccoby stated that "by early school years/ boys consistently do better (than girls) on spatial tasks and this difference continues through the high school and college years" (p.26). In 1972/ Maccoby and Jacklin said that the differences in spatial ability between the sexes "remain minimal and inconsistent until approximately the ages of 10 or 11/ when the superiority of boys becomes consistent in a wide range of populations and tests" (p.41). In 1974/ after a comprehensive literature search/ Maccoby and Jacklin concluded that sex differences in spatial visualization become more pronounced between upper elementary years and the last year of high school/ the years when sex-related differences in mathematics achievement favoring boys emerge. Guay and McDaniel (1977) supported in part Maccoby and Jacklin's 1974 findings. They found that among elementary school children/ males had greater high level spatial ability than females/ but that males and females were equal in low level spatial ability. This finding is inconsistent with that portion of Maccoby and Jacklin's review that suggests that sex differences become evident only during early adolescence. Cohen and Wilkie (1979) however/ stated that in tests measuring distinct spatial tasks/ males perform better than females in early adolescence and throughout their life span. Most studies carried out after 78 TABLE 4.2 Continued Item Num ber Xz for Model 3.2 X1 for Model 3.3 Xa for Difference Actual Sample Size Minimum Sample Size for Signif. (a) (b) (c) (d) (e) (f) 55 19.53 90 .87 71.34* 41,7 96 11,770 56 21.62 70.14 48.52* 44,107 18,263 57 28.51 35.78 7.27 53 ,075 146,668 58 17.49 51.69 34.20* 44,044 25,873 59 50.96 73.68 22.72* 41,829 36 ,987 60 66.43 148.39 81.96* 60,494 14,828 61 21.39 394.54 373.15* 35,632 1,918** 62 55.03 300.95 245.92* 50,699 4,142 63 158.36 278.34 119.98* 52 ,068 8,719 64 88.01 246.48 158.47* 73,196 9,279 65 56 .28 94.11 37.83* 42,429 22,532 66 50.51 113.88 63.37* 78,708 24,953 67 34.28 180 .60 146.32* 51,862 7,121 68 48.01 225.46 177.45* 63,223 7,158 69 75.29 251.66 176.37* 62,511 7,121 70 44.39 271.35 226 .96* 67,528 5,977 72 17.41 208.90 191.49* 84 ,175 8,831 73 24.61 37.80 13.19* 53 ,358 81,271 74 22.75 163.01 140 .26* 84,175 12,057 75 TABLE 4.2 Chi-square Values for the Comparison of Models (3.2) and (3.3) with Actual Sample Sizes and with the Corresponding Number of Subjects Needed for Significant Results Item Num ber Xa- for Model 3.2 Xa- for Model 3.3 Xa for Difference Actual Sample Size Minimum Sample Size for Signif. (a) lb) (c) (d) (e) (f) 1 16.11 30.94 14.83* 7 r 427 10,061 2 17.54 103.22 85.68* 11,018 2,583** 3 20.06 47.51 27.45* 11,497 8,733 4 7.54 21.87 14.33* 6,016 8,414 5 14.75 70.40 55.65* 7,706 2,782** 6 11.86 23.13 11.27* 13,046 23,256 7 4.60 35.73 31.13* 8,832 5,700 8 46.15 329.98 283.83* 16,796 1,189** 9 10.13 109.99 99.86* 23,075 4,642 10 9.54 65.29 55.75* 35,060 12,634 11 5.89 19.06 13.17* 24,908 37,996 12 12.58 198.46 185.88* 40 ,423 4,369 13 25.74 62.37 36.63* 15,718 8,621 14 11.19 80.50 69.31* 30,238 8,765 15 16.63 17.33 0.70 42,889 1,230,914 16 14.51 84.29 69.78* 35,114 10,109 17 17.90 44.20 26.30* 43,962 33,582 113 Meyer? R. A. (1978). Sex-related differences in mathema tical problem solving performance and intellectual abilities. Madison: University of Wisconsin? Research and Development Center for Cognitive Learning. (ERIC Document Reproduction Service No. ED 166 023). Milton? G. A. (1957) The effects of sex-role identifica- cation upon problem solving skill. Journal of Abnormal & Social,Psychology? 5? 208-212. Moller? N. J. (1983). The impact of gender? masculinity? and feminity on math achievement and course decisions (Doctoral dissertion? Purdue University? 1982). nis_sertation_Abstracts International? 41/ 2584A. Muth? K. D. (1984). Solving arithmetic word problems: Role of reading and computational skills. Journal of Educational Psychology? 76 ? 205-210. Pallas? A. M.? & Alexander? K. L. (1983). Sex differences in quantitative SAT performance: New evidence on the differential coursework hypothesis. American Educational Research Journal? 21/ 165-182. Plank? E.? & Plank? R. (1954). Emotional components in arithmetical learning as seen through autobiographies. Psychological Studies of the Child? 2.? 274-296. Porter? A. C. (1968). The effects of using fallible variables in the analysis of covariance (Doctoral dissertation? University of Wisconsin? 1967). Dissertation Abstract International? 2Â£? 3517B. Probert? B. S. (1984). Math confidence workshops: A multimodal group intervention strategy in mathematics anxiety/avoidance. (Doctoral dissertation? University of Florida? 1983). Dissertation Abstracts International? 44 ? 2231B. Puerto Rico Department of Education. (1980). Basic skills test in mathematics-6. Hato Rey: Center of Evaluation. Radatz? H. (1979) Error analysis in mathematics educa- tion. Journal for Research in Mathematics Educa tion ? 10 ? 163172. Riedesel? A. C. (1969). Problem solving: Some suggestions from research. Arithmetic Teacher? 16 ? 54-58. 48 success since the components must be integrated into a whole process (mastery learning of the components cannot assure mastery of the process). Their analysis also led them to conclude that* in the case of 6th graders inability to read problems is a major obstacle in solving word problems. Only 12% of the subjects could read and set up problems correctly at a higher level than they could compute while 60% could compute correctly at a higher level than they could read and set up problems; 44% could set up problems better when they heard them read than when they read the problem themselves. Only 13% could set up problems better when they read them than when they heard them read. Muth (1984) investigated the role of reading and computational skills in the solution of word problems. A group of 200 students from the 6th grade were administered a test of basic skills and a mathematics word problem test. The word problem test consisted of 15 sample items supplied by the National Assessment of Educational Progress. The items were adapted to include some extraneous information and complex syntactic structure. Four versions of the test were constructed by combining two versions of problem information (absence vs. presence of extraneous information) with two versions of syntactic structure (simple vs. complex syntax). Task performance was measured by means of the number of problems answered correctly 102 response was related to the use of a large number of subjects in the analysis. One must question whether analysis of incorrect responses has any educational value. Although the results of this study do not support the use of this analysis (error analysis) for the study of sex-related differences# it continues to be a promising technique in the discovery of how children approach mathematics problem solving# more so than analyses that look only at the items answered correctly or at total test scores. The second objective of the study was to investigate sex-related differences in problem solving performance# to find out if significant differences persisted after controlling for computational skills# and if the differences depended on the level of computational skills. The results showed that females outperformed males in problem solving and in six of the seven computational skills. Males showed superiority in equivalence in each of the three years# but statistical significance was obtained only for the second year. The results also showed that females retained their superiority in problem solving# only when equivalence (for the three years data) and subtraction (for the second year) were the controlling variables. Analysis of covariance was also used to answer the question of whether male-female differences in problem solving depend on the level of computational skills. The 117 From May* 1982 to July* 1983 / she worked as graduate research assistant at the Foundations of Education Department/ University of Florida. She returned to Puerto Rico in August/ 1983 / to serve as Special Aide to the Assistant Secretary for the Vocational/ Technical/ and High Skills Educational Programs of the Commonwealth Department of Education. 46 Their study is important because it represents an attempt to demonstrate that multiple factors can interact in the correct solution of a mathematics word problem. They constructed three graded tests from a basal mathematics series for grades 3 through 8. For test 1* the problems were set in pure computational form (the effects or reading* interpretation* as well as the necessity for integration were removed in an effort to measure the computational skills required by the word problems). For test 2* the effects of reading and computation were removed by reading the problems to the students and by giving scores based on whether or not the students set them up properly* in an attempt to measure problem interpretation alone. For test 3* the effect of computation was removed. The test yielded two scoresone by grading the students on whether or not they set up the problems properly and another by grading on the basis of the correct answer. The tests were administered to all 244 students from the 6th grade in two different schools. A diagnostic profile was obtained for each of the 217 students for which complete data were available: a computational score* a problem-interpretation score* a reading score* and a reading-problem solving score. They assumed that if the reading-problem interpretation score was lower (one or more levels lower) 71 TABLE 4.1 Summary of Significant Tests of Model 3.2 Item Number X* for Model 3.2* Actual Sample Size Min. Sample Size for Significance (a) (b) (c) (d) 8 46.15 16/796 3/352 13 25.74 15/718 5/624 23 29.25 60/377 19/011 25 35.90 26/707 6/852 28 46.00 84/847 16/988 29 61.48 88/396 13/242 30 25.19 87/473 31/982 31 23.26 32/870 13/015 34 58.25 82/257 13/006 36 84.36 86/137 9/404 37 23.70 37/068 14/405 42 23.26 59/140 23/417 45 22.41 63/098 25/932 48 179.25 97/593 5,014 49 27.82 16/766 5/550 52 26.01 39/858 14/114 54 41.41 74/502 16/570 56 21.62 44/107 18/789 57 28.51 53/075 17/146 58 Table 3.1 Hypothetical Probabilities of Option Choice Conditional on Year and Sex Year Sex 1 Option 2 3 First H .7 .2 .1 F .5 .3 .2 Second H .7 .2 .1 F .5 .3 .2 Third M .7 .2 .1 F .5 .3 .2 Table 3.2 Hypothetical Conditional on Probabilities of Year and Sex and Option Choice Arranged by Sex Sex Year 1 Option 2 3 M First .7 .2 .1 Second .7 .2 .1 Third .7 .2 .1 F First .5 .3 .2 Second .5 .3 .2 Third .5 .3 .2 90 The significant interactions found between sex and multiplication^ and sex and division (first year)/ and sex and subtraction (second year)/ answered/ in part/ the question of whether male-female differences in problem solving performance depend on computational ability. However/ the evidence is quite weak. Of 21 possible interactions/ only three were significant. No variable exhibited a significant interaction for each of the three years. The analysis of covariance was also used to determine if sex-related differences exist after controlling for computational skills. Analyses were conducted for those variables that did not exhibit significant interactions with sex. As discussed in Chapter III/ estimated true scores were used for observed scores to adjust for unreliability of the covariates (the computational subtests). Reliability coefficients calculated for each covariate are shown in Table 4.7. Summaries of the analyses of covariance for the first year are reported in Table 4.8. The results show that females retained their superiority in problem solving performance when equivalence was the controlling variable in the analysis of covariance/ the only variable in which males outperformed females (nonsignificant). When the controlling variables were addition/ subtraction/ addition of fractions/ and subtraction with decimals/ female ACKNOWLEDGMENTS I would like to acknowledge my gratitude to several people who have influenced my formal education and/or made this study possible. My special thanks to Dr. James Alginar Chairman of the doctoral committee/ who contributed to the development of my love for research and statistics. He has been insuperable as professor and valued friend; his help and guidance in the preparation and completion of this study were invaluable. I extend my thanks to Dr. Linda Crocker for her advice and help during my doctoral studies at University of Florida. Thanks also go to Dr. Michael Nunnery/ member of the doctoral committee. To Dr. Wilson Guertin* who was a friend for me and my family/ I extend my special thanks. Thanks are also extended to Dr. Amalia Charneco/ past Undersecretary of Education of the Puerto Rico Department of Education/ for her continuous support. To my sister Nilda Santaella/ who typed the thesis/ I give my sincere thanks. Special thanks go to my family and to those friends who provided encouragement throughout this critical period of my life. ii 10 of the studies consider sex differences incidental to the major study findings. The available literature offers very little research directly related to the problem of sex differences in this area. The review of the literature has been divided in four sections. The first section consists of a detailed summary of the available research on sex diferences in incorrect responses. The second section deals with sex-related differences in problem solving performance. These sections are directly related to the objectives of the study. The third section is more peripheral and contains a discussion of the more prevalent issues about the influence of cognitive and affective variables on sex differences in mathematics learning and achievement. The fourth is a summary of the research dealing with variables considered as influential to mathematics problem solving performance. Sex-related Differences in Incorrect Response Patterns Research findings tend to suggest that boys and girls may be approaching problem solving differently (e.g. Fennema and Sherman 1978; Marshall 1981 1983; Meyer 1978 among others). Marshall (1981) investigated whether 6th grade boys and girls approach mathematical problem solving with different strategies. Her specific interest was whether the sexes made the same errors. UNIVERSITY OF FLORIDA 3 1262 08554 6660 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy COMPARISON OF MALES AND FEMALES ON MATH ITEM PERFORMANCE: ANALYSIS OF RESPONSE PATTERNS By Sonia Feliciano Augustf 1986 Chairman: James Algina Major Department: Foundations of Education The first objective of this study was to investigate sex differences in the selection of incorrect responses on a mathematics multiple-choice test* and to determine whether these differences were consistent over three consecutive administrations of the test. A second objective was to compare male and female performance in problem solving after controlling for computational skills. The responses of all 6th grade students from the public schools in Puerto Rico who took the Basic Skills Test in Mathematics-6 ("Prueba de Destrezas Bsicas en Matemticas-6") during three academic years were used in the analyses relevant to the first objective. viii 53 The studies reviewed have confirmed the relationship between problem solving performance and attitudes toward problem solving (Carey 1958? Lindgren et al. 1964? Whitaker 1976). However the results reportea in the studies that investigated the relationship between problem solving performance and computation and between reading and problem solving fail to be consistent in their conclusions. Hansen (1944) Chase (I960) Balow (1964) Knifong and Holtan (1976 1977) and Zalewski (1974? concluded that computation is more strongly related to problem solving than is reading. Martin (1964) Creswell (1982) Ballew and Cunningham (1982) and Muth (1984) concluded that reading aoility and mathematical problem solving show a stronger relationship than computation and problem solving. Exedisis's (1983) findings led to the conclusion that the effect or reading and computation in problem solving performance is unimportant. IllMETHOD 54 The Sample 54 The Instrument 55 Analysis of the Data 57 Analysis of Sex by Option by Year Cross Classifications 57 Comparison of Males and Females in Problem Solving Performance 66 IV RESULTS 6 8 Introduction 68 Sex-related Differences in the Selection of Incorrect Responses 68 Sex-related Differences in Problem Solving Performance 81 Summary 93 V DISCUSSION 101 Summary and Interpretation of the Results 101 Implications of the Findings and Suggestions for Further Research 103 Sex-related Differences in Incorrect Responses 104 Sex-related Differences in Problem Solving Performance 105 REFERENCES 107 BIOGRAPHICAL SKETCH 116 IV Sex-related Differences in Problem Solving Performance 81 In this part/ results for the two questions related to the second objective of the study are presented. These questions serve as the framework for the presentation. Each question is stated/ followed by the results pertaining to that question. Question 1: Do males and females differ in problem solving performance? The responses of 492 males and 510 females who took the Puerto Rico Basic Skills Test-6 during the spring of the first year were analyzed in this study. Also/ data from 504 males and 509 females tested in the second year and from 509 males and 504 females tested in the third year were included in the analysis. The mean performance scores and the standard deviations for each of the eight variables are presented in Table 4.3. Results of t-tests are also presented in this table. Females outperformed males in problem solving/ a finding consistently present in all the three years of test administration. Over the three-year period the mean differences favored females in all variables except equivalence. The sex-related differences in problem solving were significant (p 4 mathematics preparation. Males who were enrolled or had completed algebra II outperformed the females in computation and problem solving but not in algebra. Males who studied beyond algebra II outscored females on all three subtests: computation? algebra? and problem solving (Armstrong? 1979). Carpenter? Lindquist? Mathews? and Silver (1984) analyzed the results of the Third National Assessment of Educational Progress (NAEP)? and compared them with the First and Second Surveys. Between 1978 and 1982? the differences between the average performance of males and females remained stable at each age level. At ages 9 and 13? the overall performance of males and females was not significantly different. At age 17? males scored higher than females by about 3 percentage points. When course background was held constant? achievement differences still existed at age 17. For each category of course background? male achievement exceeded female achievement. Consistent with previous assessments? sex differences in problem solving in favor of males were found for the 17-year-old sample. At ages 9 and 13? no large differences were found between the sexes within any level of course background. Marshall (1981? 1984) investigated sex differences in mathematics performance. She found that males and females excel each other in solving different types of problems. 105 Sexrrelated Dif fgrgnce_s_in. Problem Solving Performance The results of this study do not give strong support to Marshall's (1981/ 1984) findings that the effect of computational skills on problem solving performance is different for each sex. However/ the findings of the present investigation are consistent with those found by Meyer (1978) and Whitaker (1976): female performance in problem solving is not significantly different from male performance. Females showed superiority in problem solving performance in all three years of test administration although their superiority was retained only when performance in equivalence was controlled for (for the three years data) and when performance in subtraction was controlled for (second year) (statistically significant only after accounting for equivalence). The findings of this study are also in agreement with those reported by the National Assessment of Educational Progress (NAEP) (Carpenter et al.z 1980). Related to problem solving/ they conclude/ if it were necessary to single out one area that demands urgent attention/ it would clearly be problem solving. At all age levels/ and in vir tually every content area/ performance was extre mely low on exercises requiring problem solving or application of mathematical skills. In gene ral/ respondents demonstrated a lack of the most basic problem-solving skills. Rather than attempt ing to think through a problem and figure out what needed to be done to solve the problem/ most res pondents simply tried to apply a single arithmetic 40 Achievement Test (subtests of reading, arithmetic, and reasoning) and the California Short-Form test of mental ability to a group of 1,400 children from the 6th grade. All levels of achievement were included in the analysis. Analysis of variance and covariance were used and compared. He confirmed the findings of other researchers to the effect that there is a direct relationship between I.Q. and reading ability, and between I.Q. and computational skills. The results of the analysis of variance revealed that increases in computation ability were associated with higher achievement in problem solving. A relationship between reading ability and problem solving was also found, but it was not as strong. Significant differences in problem solving performance associated with computational ability were found when intelligence was controlled. Balow concluded that computation is a much more important factor in problem solving than reading ability, and that when I.Q. is taken into consideration, the degree of the relationship between reading and problem solving ability becomes less pronounced. Intelligence tends to confound the relationship between these two variables. Knifong and Holtan (1976, 1977) attempted to investigate the types of difficulties children have in solving word problems. They administered the word problem section of the Metropolitan Achievement Test to 35 children from the 6th grade. Errors were classified in two 99 responses was consistent throughout the years in which the test was administered. Males and females selected different incorrect responses in 100 of the 111 items of the test. The pattern of male-female differences in the selection of incorrect responses was consistently found in each of the years of test administration. However* for the vast majority of the 100 items* male-female differences were relatively small in magnitude* considering the fact that the number of subjects needed to obtain significance was very large. Therefore* these findings lack educational significance. A second objective of the study was to investigate whether males and females differ in problem solving performance* if these differences persist after controlling for computational skills* and if the male-female differences depend on the level of computational skills. Results of the analyses reported in this section generally indicated that females outperformed males not only in problem solving* but in six of the seven computational variables. Males surpassed females in equivalence* but statistical significance was obtained for only one of the three years covered in the study. The results also tended to show that* for examinees with similar levels of computational skills* sex differences in problem solving did not exist. The only |