<%BANNER%>

Predictors of Academic Achievement, Social Adjustment, and Intention to Persist

Permanent Link: http://ufdc.ufl.edu/UFE0024988/00001

Material Information

Title: Predictors of Academic Achievement, Social Adjustment, and Intention to Persist A Bioecological Analysis of College Retention
Physical Description: 1 online resource (118 p.)
Language: english
Creator: Cordell-Mcnulty, Kristi
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: college, motivation, retention
Educational Psychology -- Dissertations, Academic -- UF
Genre: Educational Psychology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Persistence of students to graduation from college is a major problem. Most studies of college retention, however, focus on relatively few variables, and questions remain regarding important predictors of retention. In this dissertation Bronfenbrenner and Morris s (1998) bioecological theory was used to create and test a model of predictors of retention from five social contexts (the self-system, family, peers, school, and culture) that have been shown to predict GPA, social adjustment, and intention to persist. The self-system includes personality, identity, perceived academic self-efficacy, achievement goals, and theory of intelligence. The family context includes parental involvement, feelings of relatedness to parents, and parents support of their children s autonomy. The peer context includes relatedness to off-campus friends, and the school context includes relatedness to instructor and, on-campus friends, sense of belonging to the institution, and participation in extracurricular activities. The cultural context includes gender and ethnicity. Participants were 299 students from a southeastern university who completed a survey online. Structural equation modeling was used to estimate the relationships in the model. Results indicated that with the exception of off-campus friends in the peer context, significant relationships to at least one of the outcome variables were found for every ecological context. In the parent context, parents support of their children s autonomy predicted social adjustment. In the college context, sense of belonging and relatedness to on-campus friends had direct and indirect relationships to social adjustment, and participation in extracurricular activities had a direct relationship to social adjustment. In the self-system, entity beliefs had a direct and an indirect relationship to GPA. Conscientiousness had an indirect relationship to intention to persist, and high school achievement predicted social adjustment. In the social context, gender predicted social adjustment with males more likely to be socially adjusted than females. Ethnicity predicted GPA, with White and Asian students reporting higher GPAs. Of the process variables, perceived academic self-efficacy had direct and indirect relationships to all three outcome variables. Performance-avoid goals had negative direct and indirect relationships to social adjustment and GPA. These findings offer ideas from multiple contexts that could be useful in increasing college retention, especially if adopted simultaneously.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Kristi Cordell-Mcnulty.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Ashton, Patricia T.
Local: Co-adviser: Miller, Scott A.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024988:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024988/00001

Material Information

Title: Predictors of Academic Achievement, Social Adjustment, and Intention to Persist A Bioecological Analysis of College Retention
Physical Description: 1 online resource (118 p.)
Language: english
Creator: Cordell-Mcnulty, Kristi
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: college, motivation, retention
Educational Psychology -- Dissertations, Academic -- UF
Genre: Educational Psychology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Persistence of students to graduation from college is a major problem. Most studies of college retention, however, focus on relatively few variables, and questions remain regarding important predictors of retention. In this dissertation Bronfenbrenner and Morris s (1998) bioecological theory was used to create and test a model of predictors of retention from five social contexts (the self-system, family, peers, school, and culture) that have been shown to predict GPA, social adjustment, and intention to persist. The self-system includes personality, identity, perceived academic self-efficacy, achievement goals, and theory of intelligence. The family context includes parental involvement, feelings of relatedness to parents, and parents support of their children s autonomy. The peer context includes relatedness to off-campus friends, and the school context includes relatedness to instructor and, on-campus friends, sense of belonging to the institution, and participation in extracurricular activities. The cultural context includes gender and ethnicity. Participants were 299 students from a southeastern university who completed a survey online. Structural equation modeling was used to estimate the relationships in the model. Results indicated that with the exception of off-campus friends in the peer context, significant relationships to at least one of the outcome variables were found for every ecological context. In the parent context, parents support of their children s autonomy predicted social adjustment. In the college context, sense of belonging and relatedness to on-campus friends had direct and indirect relationships to social adjustment, and participation in extracurricular activities had a direct relationship to social adjustment. In the self-system, entity beliefs had a direct and an indirect relationship to GPA. Conscientiousness had an indirect relationship to intention to persist, and high school achievement predicted social adjustment. In the social context, gender predicted social adjustment with males more likely to be socially adjusted than females. Ethnicity predicted GPA, with White and Asian students reporting higher GPAs. Of the process variables, perceived academic self-efficacy had direct and indirect relationships to all three outcome variables. Performance-avoid goals had negative direct and indirect relationships to social adjustment and GPA. These findings offer ideas from multiple contexts that could be useful in increasing college retention, especially if adopted simultaneously.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Kristi Cordell-Mcnulty.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Ashton, Patricia T.
Local: Co-adviser: Miller, Scott A.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024988:00001


This item has the following downloads:


Full Text

PAGE 1

1 PREDICT ORS OF ACADEMIC ACHIEVEMENT, SOCIAL ADJUSTMENT, AND INTENTION TO PERSIST : A BIOECOLOGICAL ANALYSIS OF COLLEGE RETENTION By KRISTI L. CORDELL MCNULTY 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 2009

PAGE 2

2 2009 Kristi L. CordellMcNulty

PAGE 3

3 To Calvin Hobbes, and Evan

PAGE 4

4 ACKNOWLEDGMENTS B ecause my research emphasizes social support and this dissertation tested a bioecological model, I think it would be appropriate to organize my acknowledgments around my college, peer, and family microsystems. First I would like to start with my college microsystem. None of what I accomplished over the last four years would have been possible without the help of my advisor, Dr. Patricia Ashton. She has shown me new ways to think and pushed me to grow beyond my intellectual limits. I also want to thank Dr. James Algina for all his help with statistics over the past four years and especially for his help with this dissertation. In addition I want to thank Dr. Scott Miller and Dr. Julia Graber for all their encouragement over the last four years. I also wish to thank Elaine Green for her guidance in getting through all the hoop jumping that is part of grad school and her friendship over the last four years. Also in my college microsystem, I want to thank all the friends I have made in the program, especially Tes ia Jenni, and Michelle. Without them, there are times I think I might not have made it through. In my peer microsystem, I want to thank my three best friends from home, Stacey, Heather, and Dawn. They have been some of my biggest supporters. They have always been there when I needed someone to talk to. In addition I want to thank my church family who has also cheered me on, especially Beth and Randi who have always listened to me whine and expressed their empathy as graduate students themselves. I also need to thank the countless number of preschoolers I have come in contact with every Sunday morning. They have kept me grounded, reminded me of the important things in life, and many times I have been humbled by the wisdom of a preschooler. In my family micr osystem I have to thank my Mom and Dave, Jackie and Marvin, Pappy Web and Grammy Cordell, and Grammy Fay and Pappy Bub. These individuals had the mos t impact on me while growing up and without any of them, I would not have become the hard

PAGE 5

5 working person I am today. I also owe most of my success to the support of my husband, Evan, who put his dreams on hold to follow me across the country, not once, but twice. Finally I wish to thank Calvin and Hobbes for their relentless support even though I forced them to adapt to many new situations over the last few years. They truly loved me unconditionally.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ...............................................................................................................4 LIST OF TABLES ...........................................................................................................................9 LIST OF FIGURES .......................................................................................................................10 ABSTRACT ...................................................................................................................................11 CHAPTER 1 INTRODUCTION ..................................................................................................................13 Statement of the Problem ........................................................................................................13 Theoretical Background of the Study .....................................................................................13 Bronfenbre nners Bioecological Model .................................................................................18 Psychological Processes .........................................................................................................20 Identity .............................................................................................................................20 Achievement Goal Theory ..............................................................................................30 Perceived Academic Self Efficacy ..................................................................................31 The Bioecological Systems .....................................................................................................34 The Family Microsystem .................................................................................................34 Parental involvement ................................................................................................34 Relatedness to parents ..............................................................................................37 Parental autonomy support .......................................................................................39 The Peer Microsystem .....................................................................................................42 The College Microsystem ................................................................................................45 Relatedness to on -campus friends ............................................................................45 Relatedness to instructors .........................................................................................46 Belonging on campus ...............................................................................................48 Extracurricular activities ..........................................................................................51 The Self System ..............................................................................................................52 Personality ................................................................................................................53 Theory of intelligence ..............................................................................................55 Previous academic achievement ...............................................................................58 The Macrosystem ............................................................................................................59 Gender ......................................................................................................................60 Ethnicity ...................................................................................................................61 Socioeconomic status ...............................................................................................61 The Chronosystem ...........................................................................................................62 Purpose of the Study ...............................................................................................................63 R esearch Hypotheses ..............................................................................................................64 Theoretical Significance .........................................................................................................65 Practical Significance .............................................................................................................65

PAGE 7

7 2 METHOD ...............................................................................................................................69 Participants .............................................................................................................................69 Instruments .............................................................................................................................69 Predictor Variables ..........................................................................................................69 Identity. ....................................................................................................................69 Goal orientation. .......................................................................................................71 Perceived academ ic self efficacy. ............................................................................72 Parental involvement. ...............................................................................................73 Parental autonomy support. ......................................................................................74 Relatedness. ..............................................................................................................74 Sense of belonging on campus. ................................................................................75 Theory of intelligence. .............................................................................................75 Personality. ...............................................................................................................76 Socioeconomic status. ..............................................................................................77 Demographic information. .......................................................................................78 Outcome Variables ..........................................................................................................78 Social adjustment. ....................................................................................................78 Academic achievement. ...........................................................................................81 Intention to persist. ...................................................................................................81 Analyses ..................................................................................................................................81 3 RESULTS ...............................................................................................................................83 Descriptive Statistics ..............................................................................................................83 Analysis of the Proposed Model .............................................................................................83 Research Hypotheses ..............................................................................................................85 4 DISCUSSION .........................................................................................................................96 The Bioecological Systems .....................................................................................................96 The Family Microsytem ..................................................................................................96 The Peer Microsystem .....................................................................................................97 The College Microystem .................................................................................................97 The Self System ..............................................................................................................99 The Macrosytem ............................................................................................................100 Psychological Process Variables ..........................................................................................101 Identity ...........................................................................................................................101 Achievement Goals .......................................................................................................101 Perceived Academic Self Efficacy ................................................................................102 Outcome Variables ...............................................................................................................102 Limitations of the Study .......................................................................................................103 Implications of the Findings for Future Research ................................................................103 Conclusions...........................................................................................................................104 LIST OF REFERENCES .............................................................................................................106

PAGE 8

8 BIOGRAPHICAL SKETCH .......................................................................................................118

PAGE 9

9 LIST OF TABLES Table page 1-1 Proposed relationships in the model ..................................................................................68 3-1 Demographic data ..............................................................................................................89 3-2 Means and standard deviations of variables ......................................................................90 3-3 Correlation matrix ..............................................................................................................91 3-4 Total, direct, and indirect effects in the revised model ......................................................93 3-5 Effects proposed in original model and effects in revised model ......................................95

PAGE 10

10 LIST OF FIGURES Figure page 1-1 Proposed Bioecological Model of College Student Retention ..........................................67

PAGE 11

11 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 PREDICT ORS OF ACADEMIC ACHIEVEMENT, SOCIAL ADJUSTMENT, AND INTENTION TO PERSIST: A BIOECOLOGICAL ANALYSIS OF COLLEGE RETENTION By Kristi L. CordellMcNulty December 2009 C hair: Patricia Ashton Cochair: Scott Miller Major: Educational Psychology Persistence of students to graduation from college is a major problem. Most studies of college retention however, focus on relatively few variables and questions remain regarding important predictors of retention. In this dissertation Bronfenbrenner and Morriss (1998) bioecological theory was used to creat e and test a model of predictor s o f retention from five social contexts (the self sys tem family, peers, school, and culture) that have been shown to predict GPA social adjustment, and intention to persist. The self system includes personality, identity, perceived academic self efficacy, achievement goal s, and theory of intelligence. The family context includes parental involvement, feelings of relatedness to parents, and parent s support of their childrens autonomy. The peer context i ncludes relatedness to offcampus friend s, and the school context includes relatedness to instructor and, on-campus friends sense of belonging to the institution and participation in extracurricular activities. The cultural context includes gender and ethnicity. Participants were 29 9 students from a southeastern university who complete d a survey online. S tr uctural equation modeling was used to estimate the relationships in the model. Results indicated that with the exception of offcamp us friends in the peer context significant

PAGE 12

12 relationships to at least one of the outcome variables were found for every ecol ogical context. In the parent context parents support of their childrens autonomy predicted social adjustment. In the college context sense of belonging and relatedness to oncampus friends had direct and indirect r elationships to social adjustmen t an d participation in extracurricular activities had a direct relationship to social adjustment. In the self system entity beliefs had a direct and an indirect relationship to GPA. Conscientiousness had an indirect relationship to intention to persist and high school achievement predicted social adjustment. In the social context, gender predicted social adjustment with males more likely to be socially adjusted than females. Ethnicity predicted GPA, with White and Asian students reporting higher GPAs. Of the process variables, perceived academic self efficacy had direct and indirect relationships to all three outcome variables Performance -avoid goals had negative direct and indirect relationships to social adjustment and GPA. These findings offer ideas from m ultiple contexts that could be useful in increasing college retention, especially if adopted simultaneously.

PAGE 13

13 CHAPTER 1 INTRODUCTION Statement of the Problem The graduation rate of college students remains a serious national concern. A comparison of stu dents who entered a 4-year college in 1989 with those entering in 1995 showed that the rate of completion of the bachelors degree within 5 years remained steady at about 53% ( National Center for Education Statistics, 2005b). More recently, only 56.4% of the students entering college in Fall 2000 had graduated by Summer 2006 (National Center for Higher Education Management Systems 2007). These data show little improvement in the bachelors degree graduation rate over the last two decades. In addition, 36.9% of all undergraduates were enrolled in 2-year institutions in Fall 2005 (National Center for Education Statistics, 2007). Of the students who entered a 2-year institutions in Fall 2002, only 32.5% graduated by Spring 2006 (National Center for Education Statistics, 2007). Failure of so many students to complete a college degree in a timely fashion represents a significant economic loss to the nation and a serious personal loss to students. The purpose of this study wa s to test a model of college success an d retention based on a bioecological perspective. This new bioecological systems model of college retention focuses on the different ecological systems that predict college success and the psychological processes that mediate the relationship between the different systems and college success. In this chapter I describe the theoretical models and empirical research that provided the basis for the variables that were included in the model. Theoretical Background of the Study The problem of college retention has prompted many research studies in the last 30 years. Numerous studies have been completed, and several theories have been proposed Many variables have been linked with college retention. Tintos (1975, 1993) interactionalist theory is

PAGE 14

14 one of the most w idely researched theories of college retention. Tintos model postulates that college student attrition needs to be examined as a longitudinal process of interactions between the student and the academic and social realms of college. The first part of the model indicates that students enter college with various personal characteristics that influence their decisions to leave school without completing a degree. These personal characteristics include family background, individual attributes, and precollege s chooling experiences that shape students commitment to the educational institution and educational goals. In Tintos model, goal commitment refers to how determined students are to finish college and institutional commitment refers to their determination to complet e their degrees at the college where they are currently enrolled. According to Tinto, t hese initial commitments affect the level of students integration into the academic and social realms of the college. Academic integration consists of t he str uctural dimension, which refers to meeting the explicit standards of the university (i.e., GPA), and the normative dimension, which refers to students identification with the norms of the academic system. Social integration refers to the fit between the s tudent and the social system of the university, including interactions with peers and faculty. Tinto proposed that academic and social integration along with initial institutional and goal commitment impact subsequent institutional and goal commitment. The greater the amount of subsequent institutional and goal commitment the more likely the student will persist in college (Tinto, 1975, 1993). Recently Braxton Sullivan, and Johnson (1997) presented revisions to Tintos (1975, 1993) theory. They argued that research has only partially supported Tintos theory, and research on traditional residential colleges and commuter colleges has provided little empirical support for Tintos theory. Braxton et al. examined further the empirical support for Tintos model. They divided the model into 13 testable propositions. For example, the first proposition is student

PAGE 15

15 entry characteristics affect the level of initial commitment to the institution (p. 112). Then Braxton et al. examined peer reviewed studies of Tintos model to find empirical support for the 13 propositions. Of these 13 propositions they concluded that, for residential universities, strong empirical support was evident for five of the propositions, but for commuter colleges strong empirical support was evident for only two of the propositions. For residential colleges, the five empirically supported propositions were initial goal commitment affects the level of social integration, the greater the level of social integration, the greater the level of su bsequent institutional commitment, initial institutional commitment affects the subsequent level of institutional commitment, initial goal commitment affects the subsequent level of goal commitment, and subsequent level of institutional commitment af fects the level of persistence (p. 135). For commuter colleges, the two empirically supported propositions were student entry characteristics affect the level of initial commitment to the institution, and initial institutional commitment affects the su bsequent level of institutional commitment (Braxton et al., 1997, p. 135; Braxton, Hirschy, & McClendon, 2004). On the basis of these findings, Braxton and colleagues created two revised models of Tintos theory, one for residential colleges and one for commuter colleges. In the residential college model, social integration is emphasized because of the importance of social interaction and community as students are living on campus. At commuter colleges, academic integration is emphasized because there is l ess opportunity for social interaction at commuter colleges where most students are not on campus, other than for class (Braxton & Hirschy, 2005). Another prominent model in the college retention literature is Bean and Eatons (2000) psychological model of college retention. Bean and Eaton contended that most college retention models, including Tintos (1975, 1993), have emphasized sociological theories to explain why

PAGE 16

16 students leave college. Bean and Eaton, in contrast, argued that leaving college is a beh avior and that behaviors are psychologically motivated; therefore, psychological theories can inform retention models. The ir model consists of eight parts : entry characteristics, environmental interactions, psychological processes, psychological outcomes, intermediate outcomes, attitudes, intention, and behavior. Entry characteristics consist of student variables such as past performance, personality, motivations, skills, and abilities. Entry characteristics influence the institutional environment, which includes environmental interactions, psychological processes, psychological outcomes, and intermediate outcomes. Environmental interactions include bureaucratic, academic, social, and external to the institution interactions. The environmental interactions a re expected to predict psychological processes, which in turn predict the psychological outcomes. For example, according to the model when students make selfefficacy judgments and assessments (psychological process), they will exhibit either positive or negative perceived self efficacy (psychological outcome) The psychological outcomes predict the immediate outcomes of academic and social integration and academic performance. These immediate outcomes then predict attitudes such as institutional fit and institutional commitment. Attitudes predict intention, which in the retention model, is the intention to persist. Subsequently intention predicts behavior, that is, persistence. Neuville et al. (2007) compared Tintos (1975, 1993) model and a psychological process model. Specifically they compared three structural equation models: a version of Tintos model, an expectancy -value model, and a combined model of Tintos theory and expectancy-value. All three models had the same outcome variables, academic performance and intention to persist, and all three models had the same input variables : mothers education, students high school GPA during the last year of high school, and certainty of study choice; that is the students

PAGE 17

17 estimate of their certainty of their major and if they would continue in the same major if they did not succeed academically. In addition, the Tinto model included academic and social integration variables and institutional commitment and academic engagement; the expectancy -value model included expectancy perceptions at time 1and time 2, value perceptions, and academic engagement, and the integrative model included the variables from both models. Participants were 2,637 firstyear students at a university in Belgium. The researchers collected data during the first weeks of the semester in September ( Time 1) and again in November (T ime 2). Results indicated that the expectancy value model was a significantly better fit than the other two models. Within the expectancyvalue model, certainty of st udy choice ( = 37, p < .001) was the largest predictor of persistence followed by expectancy perceptions at T ime 2 ( = 20, p < .001) and value perceptions ( = 19, p < .001). High school grade ( = 21, p < .001) and expectancy perceptions at T ime 2 ( = 21, p < .001) were the strongest predictors of performance followed by academic engagement ( = 10, p < .001), intention to persist ( = 09, p < .001), and mothers education ( = 08, p < .001). This study highlights the importance of motivational constructs in predicting college success and retention and is one of the first studies to compare structural models of motivational theory and Tintos theory. However, this study examined relatively few variables and because of this limitation, their results may be spurious. I examined a more comprehensive set of predictor variables of college retention In summary, the purpose of this dissertation wa s to identify variables that predict college success and retention. Many studies have been conducted on this topic from a variety of perspectives, each one focusing on relatively few variables. Several researchers have proposed models to explain the relationships among these variables. Tinto's (1975, 1993) model has been used widely for this purpose. More recently, noticing the lack of psychological processes in

PAGE 18

18 Tinto's model, Bean and Eaton (2000) developed a model of college retention that focuses on psychological processes and how these processes impact retention. In an effort to provide further clarificatio n of the variables related to college success and retention, the purpose of this study is to use Bronfenbrenner's (1979, Bronfenbrenner & Morris, 1998) bioecological systems model as a basis for a structural model that incorporates the models of Tinto and Bean and Eaton. This new bioecological systems model of college retention focus es on the different ecological systems that predict college students performance and persistence and the psychological processes that mediate the relationship between the different systems and college success. Bronfenbrenners Bioecological Model Typically, only a few variables such as personality, previous academic achievement, and socioeconomic status are included in studies of college retention and success, and consequently, it is unclear whether findings are spurious because one or more critical variables have been left out of the model. In this study I examined a more comprehensive set of variables that may be predictive of college retention or success using Bronfenbrenners (1979; Bronfenbrenner & Morris, 1998) bioecological model as a framework for identifying the most relevant variables to include in the study. Bronfenbrenner and Morriss (1998) bioecological model focuses on the four components of person, process, context, and time (PPCT) and the relationships among them. The person component includes the unique characteristics of the individual. The context component includes the environment in which development is occurring. The process component comprises the interact ion occur ring between person and context components, and the time component includes the time encompassed in the study. In this study I examined variables for each component of Bronfenbrenner and Morriss model.

PAGE 19

19 In addition to the four components of person, process, context, and time, Bronfenbrenner (1979; Bronfenbrenner & Morris, 1998) described spheres of environmental influences in his theory: the microsystem, mesosystem, exosystem, macrosystem, and chronosystem. In this study I have integrated the pers on, process, context, and time variables into the microsystem, macrosystem, and chronosystems. The microsystem includes activities, roles, and relationships that tend to occur on an everyday, faceto face basis. The macrosystem includes overall patterns of the society and culture. The chronosystem represents the relationship of time to the other systems. In their later work, Bronfenbrenner and Morris also emphasized examining peoples characteristics and how they influence the environment, which they referred to as the self system The focus of Bronfenbrenner and Morris s (1998) bioecological theory on the multiple contexts that influence development may offer a useful framework for identifying multiple sources of influence on the retention of college student. In this study I examine d the family microsystem, through the process variables of parental involvement, parental autonomy support, and parental relatedness; the peer microsystem through the process variable of off -campus friend relatedness; the college microsystem was represented in the process variables of instructor relatedness, on -campus friend relatedness, and sense of belonging on campus and the person variable of extracurricular activities; the self system was assessed through the person variables of personality, and previous academic achiev ement, and the process variable of theory of intelligence ; the macrosystem was assessed through the context variables of gender, ethnicity, and socioeconomic status, and the chronosystem, was assessed through t he time variable of age. In addition, I examined the psychological process variables of identity, academic self efficacy, and goal orientation as potential mediators between the systems and the outcome variables.

PAGE 20

20 In m y bioecological model of college reten tion (see Figure 1-1) the microsystems of family, peers, college, and the selfsystem comprise the academic and social realms that Tinto (1975, 1993) emphasized in his model. Following the microsystems in my model are the psychological processes. These processes include identity, perceived academic self efficacy, and goal orientations. These variables were chosen because they were hypothesized as psychological processes in Bean and Eatons (2000) model and because the literature supports the relationship be tween these variables and college success. Next in the model are the two immediate outcomes of the psychological process variables: academic achievement and social adjustment followed by intention to persist. In the sections that follow, I focus on each of the psychological process variables because they offer explanatory mechanisms that account for the relationships between the bioecological systems and the outcome variables. Then I discuss each of Bronfenbrenner and Morriss bioecological systems. Psychol ogical Processes Identity According to Arnett (2006), most identity exploration occurs between the ages of 18 and 25 during the period he referred to as emerging adulthood, not during adolescence as popularly believed. Arnett argued that in our society most identity exploration does not take place in adolescence because children are still living under the supervision of their parents. During emerging adulthood, however, most children leave home for the first time, often going to college, and this relative i ndependence typically allows for easier identity exploration. Successful exploration is associated with positive life outcomes (Berzonsky & Kuk, 2000; Luyckx, Goossens, Soenens, Beyers, & Vansteenkiste, 2005). For example, identity is positively related to adjustment and negatively related to depressive symptoms and substance abuse in college students (Luyckx, Goossens, Soenens, & Beyers, 2006, Luyckx, Soenens, Goossens, &

PAGE 21

21 Vansteenkiste, 2007). Students who experience success in college and ultimately graduate are likely to have a well established identity (Robinson, 2003). Erikson (1968) conceived of identity as the crisis that characterizes the fifth stage of his psychosocial stages of development: the identity versus role confusion stage. He defined identity as the awareness that there is a self -sameness and continuity to the egos synthesizing methods, the style of ones individuality, and that this style coincides with the sameness and continuity of ones meaning for significant others in the imm ediate community (p. 50) In the identity versus role confusion stage, Erikson proposed that adolescents must seek out their own ideological and occupational goals and through this exploration can either successfully achieve an identity or experience a st ate of role confusion where they are unsure of their goals. Elaborating on Eriksons (1968) stage theory, Marcia (1993a) conceived of identity formation along two dimensions, exploration and commitment. Exploration is the cognitive and behavioral departure from a fixed occupational or ideological position to investigate other possibilities, and commitment is the dedication to the new position or direction (Marcia, 1993b). From his conception of identity formation along two dimensions, Marcia (1993a) de scri bed four identity statuses: identity achievement, foreclosure, identity diffusion, and moratorium. Identity achievement occurs when people make their own decisions about occupational and ideological goals and are committed to pursuing their goals. People i n the identity achievement status are considered high on the exploration dimension because they have examined different alternatives in coming to their decisions and because they have made a decision they are also considered high on commitment. The foreclosure status is achieved when people have committed to occupational and ideological goals without exploration. These goals are set by others, usually the parents, instead of the individual. People in the foreclosure status are considered low on exploration because they never engaged in their own search but high on the commitment dimension because

PAGE 22

22 they have made a decision about their identity. Identity diffusion occurs when people have no real ideological or occupational direction. People in identity diffusion are low on both exploration and commitment because they have not engaged in an ideological or occupational search and they have not made a commitment to any goals. Finally, people in the moratorium status are experiencing an identity crisis. They are hi gh on the exploration dimension because they are currently searching for their own goals, but they are low on commitment because they have not made a decision yet (Marcia, 1980, 1993b). Recently Luyckx Goossens, et al. (2006) used confirmatory factor anal ysis to test a fourdimension model of identity formation. Participants were 565 freshmen from a large university in Belgium. The researchers used the Utrecht Groningen Identity Development Scale (U GIDS; Meeus & Dekovic, 1995) to assess Identification with Commitment and Exploration in Depth in the ideological and interpersonal domains, and the Ego Identity Process Questionnaire (EIPQ; Balistreri, Busch-Rossnagel, & Geisinger, 1995) to assess Commitment Making and Exploration in Breadth in the ideological and interpersonal domains. The four dimensions are Commitment Making, Identification with Commitment, Exploration in Depth, and Exploration in Breadth. Commitment Making refers to whether people have made an actual identity commitment; in contrast, Identif ication with Commitment refers to the extent people feel certain and confident about their commitment. Exploration in Breadth refers to the extent which people have explored different identity alternatives, and Exploration in Depth refers to the extent that people have explored their current commitment in detail. Luyckx Goossens, et al. found that the fourdimension model of identity was a better fit to the data than the two and three-dimension models. They assessed academic and social adjustment using a shortened version of the Student Adaptation to College Questionnaire (SACQ; Baker & Siryk, 1984), and they used the shortened

PAGE 23

23 version of the Center for Epidemiologic Studies Depression Scale (CESD; Radloff, 1977) to measure depressive symptoms. The results indicated that Commitment Making ( r = .19, p < .001) and I dentification with Commitment ( r = .38, p < .001) were positively related to social adjustment. Commitment Making ( r = .33, .p < .001) and Identification with Commitment ( r = .42, p < .001) were also related to academic adjustment. Commitment Making ( r = -.25, p < .001) and Identification with Commitment ( r = -.34, p < .001) were negatively related to depressive symptoms. Commitment Making was negatively related to substance use ( r = -.12, p < .001). Exploration in Depth was positively related to academic adjustment ( r = .31, p < .001) and negatively related to substance use ( r = -.12, p < .01). Exploration in Breadth was positively related to depressive symptoms ( r = .20, p < .001) and substance use ( r = .16, p < .001). Exploration in Breadth could signal an identity crisis, which could explain the relationship with depressive symptoms and substance use. In sum, Luyckx, Goossens, et al indicated that the identity dimensions of Commitment Making and Commitment with Identification have positive relationships with social and academic adjustment and negative relationships with depression and substance abuse. Of particular importance, their study expanded upon Marcias (1993a) original conception of ident ity by separating identity into four dimensions, instead of the two dimensions Marcia originally proposed. Having four dimensions instead of two allows for a more thorough examination of identity development. However, it is unclear whet her the results of t he Luyckx Goossens, et al. study conducted with freshmen at a university in Belgium will generalize to a predominantly female, upper-division, middleclass sample of American university students. In another study to refine the four-dimension conception of identity, Luyckx et al. (2005) conducted a cluster analysis of the responses of 565 freshmen in educational psychology classes

PAGE 24

24 in a university in Belgium to two identity measures : The U -GIDS (Meeus & Dekovic, 1995) was used to assess Identification with Commitment and Exploration in Depth in the domains of education and friendship, and the EIPQ (Balistreri et al., 1995) was used to assess Commitment Making and Exploration in Breadth in the ideological and interpersonal domains. Academic and social adjustme nt were assessed using a shortened version of the SACQ (Baker & Siryk, 1984). The shortened version of the CESD (Radloff, 1977) was used to measure depressive symptoms. The Rosenberg Self Esteem Scale (RSES; Rosenberg, 1965) was used to measure self esteem and substance use was measured by asking participants if they had used drugs or drank too much in the last 6 months. Using the four dimensions of identity, Exploration in Breadth, Exploration in Depth, Commitment Making, and Identification with Commitment, Luyckx et al. concluded that the data fit best into five clusters instead of the four Marcia (1993a) conceived. However, four of these clusters were consistent with Marcias definition of the four identity statuses and had similar relationships with criterion variables. Participants in the Achievement Cluster were high in Exploration in Breadth and Commitment Making, which is consistent with Marcias definition of Achievement. They were also high in Exploration in Depth and Identification with Commitment. Participants in the Foreclosure Cluster were low in Exploration in Breadth and high in Commitment Making which is consistent with Marcias definition of Foreclosure They were also moderately high in both Identification with Commitment and Exploration in Depth. Participants in Moratorium Cluster were high in Exploration in Breadth and low in Commitment Making, which is consistent with Marcias definition of Moratorium. They were also low in Identification with Commitment and moderately high in Exploration in Depth. Luyckx et al. found that two types of clusters for Diffusion, which they identified as the Diffused Diffusion Cluster and the Carefree Diffusion Cluster. Participants in the Diffused Diffusion Cluster were

PAGE 25

25 low in Commitment Making and moderate in Exploration in Breadth. They were also low in Identification with Commitment and Exploration in Depth. Students in the Carefree Diffusion Cluster were moderate in Commitment Making and low in Exploration in Breadth. They were also low in Exploration in D epth and low to moderate in Identification with Commitment. L uyckx et al. found that for social adjustment, students in the Achievement ( M = 3. 83), Foreclosure ( M = 3.70), and Carefree Diffusion c lusters ( M = 3.62) scored high, whereas students in the Mora torium ( M = 3.59) and Diffused Diffusion clusters ( M = 3.31) scored low on social adjustment. For academic adjustment, students in the Achievement ( M = 3.57) and Foreclosure clusters ( M = 3.45) scored high, whereas students in the Diffused Diffusion cluster ( M = 2.75) scored low. For depressive symptoms students in the Moratorium ( M = 1.95) and Diffused Diffusion c lusters ( M = 2.13 ) scored high For selfesteem students in the Moratorium ( M = 2.89) and Diffused Diffusion c lusters ( M = 2.71) scored lowest. Students in the Moratorium cluster also reported the most substance use ( M = 2.01) whereas students in the Foreclosure Cluster had the lowest amount of substance use ( M = 1.69). All these means were significantly different from each other. These findings s uggest that experiencing the Moratorium or Diffused Diffusion identity status may have a negative impact on college students. Like the Luyckx, Goosens, et al. study (2006), i t is unclear whether these results from a sam ple of f irst year predominantly femal e (85%) middle class Cuscasian students at a university in Belgium are generalizable to a predominantly female, middleclass sample of predominantly upper-division American university students. In a study of identity and academic outcomes, Berzonsky and Kuk (2000) examined how identity status relates to the transition to college. Participants were 363 freshmen from a medium-sized university in New York. The researchers measured identity status with the

PAGE 26

26 Objective Measure of Ego Identity Status (OM EIS; Adams Shea, & Fitch, 1979) and used the Student Developmental Task and Lifestyle Inventory (SDTLI; Winston & Miller, 1987) to assess the extent to which students were prepared to adapt to college by measuring students academic autonomy, mature interpersonal relationships, and educational purpose. Berzonsky and Kuk found that students who were in the Achievement and Moratorium statuses were better prepared to act autonomously in college without needing excessive reassurance from others. This finding was replicated by Boyd, Hunt, Kandell, and Lucas (2003) with 2,818 firstyear students at a large east coast university who completed the Identity Styles Inventory (ISI; Berzonsky, 1992) and the University New Student Census (UNSC). The UNSC asked students about thei r academic expectations and retention. Boyd et al. determined that first year college students with an informative identity style (Achievement or Moratorium) were more likely to be prepared for college. In particular, these students perceived themselves as prepared for college and were open to learning new skills that might help them succeed in college. Thus, the positive re lationship between Moratorium and college adjustment in the Berzonsky and Kuk study and the Boyd et al. study conflict with the Luyckx et al. (2005) finding that students in Moratorium experienced low academic and social adjustment, raising the question of whether Moratorium is a positive predictor of persistence and success in college. To explore this question further Luyckx et al. (2008) created a new identity measure in an attempt to clarify the difference between positive and negative exploration in the Moratorium stage. Specifically they proposed another dimension of exploration called Ruminative Exploration as a potential explanation of the mixed findings in the literature. They described Ruminative Exploration as occurring when people become stuck in the exploration process mulling over different alternatives and hav ing trouble making commitments To study this new dimension Luyckx et al. created a new measure, the Dimensions

PAGE 27

27 of Identity Development Scale (DIDS) to assess five identity dimensions, Commitment Making, Identification with Commitment, Exploration in Breadth, Exploration in Depth, and Ruminative Exploration. The researchers administered the DIDS to a sample of 263 freshmen at a university in Belgium. The sample was 72.6% female with a mean age of 19.4. Confirmatory factor analyses revealed that the fivefactor model of identity was a significantly better fit than a four fact or model of identity ( df = 265, SatorraBentler scaled chi -square = 658.93, RMSEA = .07, CFI = .94) A cluster analysis was conducted to capture the interactions of the identity statuses. Six clusters emerged, including Achievement, Diffused Diffusion, Carefree Diffusion, Ruminative Moratorium, Foreclosure, and Undifferentiated. The Achievement cluster scored high on Commitment Making and Identification with Commitment, moderately high on Exploration in Breadth and Exploration in Depth, and low on Ruminative Exploration. The Diffused Diffusion cluster scored low on Commitment Making and Identification with Commitment, intermediate on Exploration in Breadth and Exploration in Depth, and high on Ruminative Exploration. The Carefree Diffusion cluster scored low on Commitment Making, Identification with Commitment, Exploration in Breadth, and Exploration in Depth, and intermediate on Ruminative Exploration. The Ruminative Exploration cluster scored low on Commitment Making and Identification with Commitment, and high on Exploration in Breadth, Exploration in Depth, and Ruminative Exploration. The Foreclosure cluster scored high on Commitment Making and Identification with Commitment, and low on Exploration in Breadth, Exploration in Depth, and Ruminative Exploration. The Undifferentiated cluster scored intermediate on all dimensions ex cept Exploration in Breadth, which was moderately high. Selfesteem, depressive symptoms, anxiety symptoms, self-reflection, and self-rumination were also assessed in this study. Self -esteem was assessed using the RSES (Rosenberg, 1965). Depressive

PAGE 28

28 symptoms were assessed using 12 items of the CES-D (Radloff, 1977). Anxiety symptoms were assessed using the revised Symptom Checklist (SCL-90-R; Arrindell & Ellema, 1986). Selfreflection a nd self-rumination were assessed using the Rumination-Reflection Questionnaire (RRQ; Trapnell & Campbell, 1999). For selfesteem, the Achievement ( M = 3.30) and Foreclosure clusters ( M = 3.34) scored the highest and the Diffused Diffusion ( M = 2.67) and Ru minative Moratorium clusters ( M = 2.52) scored the lowest. For depression symptoms, the Achievement ( M = 1.78), Foreclosure ( M = 1.57), and Carefree Diffusion clusters ( M = 1.81) scored the lowest and the Diffused Diffusion ( M = 2.35) and Ruminative Moratorium clusters ( M = 2.38) scored the highest. For anxiety symptoms, the Achievement ( M = 2.02), Foreclosure ( M = 1.82), and Carefree Diffusion clusters ( M = 1.69) scored the lowest and the Diffused Diffusion ( M =2.52) and Ruminative Moratorium clusters ( M = 2.50) scored the highest. For self reflection, the Achievement ( M = 3.81) and Ruminative Moratorium clusters ( M = 3.69) scored the highest and the Carefree Diffusion cluster ( M = 3.26) scored the lowest. For self-rumination, the Ruminative Moratorium ( M = 4.02) and Diffused Diffusion clusters ( M = 3.76) scored the highest and the Achievement ( M = 3.37) Carefree Diffusion ( M = 3.30), and Foreclosure clusters ( M = 3.01) scored the lowest. All these means were significantly different from each other. From th ese data, Luyckx et al. concluded that commitment is what separates successful from unsuccessful identity development because the Achievement and Foreclosure clusters both scored high on both dimensions of identity and they were the most adaptive clusters. Therefore in my study I only examined the commitment dimensions and their relationship s to academic achievement and social adjustment. Robinson (2003) examined whether identity mediated the relationship between academic integration and both shortand long-term persistence in college in a sample of 212 undergraduate

PAGE 29

29 students. Robinson defined academic integration as the students perception of their academic performance and intellectual development, and social integration as the quality of students relat ionships with both their peers and faculty and used the Institutional Integration Scale (Pascarella & Terenzini, 1980) to measure academic and social integration. Robinson measured identity with the identity subscale of the Ego Development Scale (Ochse & Plug, 1986), and shortterm persistence by the Persistence at the Institution (PAI) and longterm persistence by the Persistence in Higher Education (PHE) subscales of the Undergraduate Persistence Intention Measure (UPI; Robinson, 1996). The results indica ted that identity did mediate the relationship between academic integration and short term persistence. Identity was the only variable that predicted longterm persistence in college ( = .31, p = .05). Academic integration did not predict longterm persistence in college, suggesting that identity may become more important for academic success as students progress through college. African American students were overrepresented in this sample (35%) because some of the students were recruited from a Black studies course, and these results may not generalize to all college populations. This study indicates that identity does predict college success and persistence, though its exact role remains unclear because Robinson did not measure identity with the more discriminating dimensions of Commitment Making and Identification with Commitment. I examined if identity, as measured by Commitment Making and Identification with Commitment, predicts ac ademic achievement, social adjustment, and intention to persist in college students. Studies have linked identity, in particular Commitment Making and Identification with Commitment, to academic adjustment (Luyckx et al., 2005; Luyckx, Goossens, et al., 2006), and to social adjustment (Luyckx et al., 2005; Luyckx, Goossens, et al., 2006). One study has also linked identity with persistence (Robinson, 2003). B ecause previous

PAGE 30

30 research has established a link between identity and adjustment, I hypothesized that identity is positively related to social and academic achievement in my model. On the basis of Robinsons study, I also hypothesized that identity predicts persistence in college students. Achievement Goal T heory Achievement goal theory is comprised of t wo main types of goals: mastery goals, also referred to as learning goals and performance goals. Mastery goals focus on learning the task for self -improvement, meeting a challenge, or gaining new understanding and insight Performance goals focus on demonstrating ability or competence, especially relative to others, such as receiving the highest grade or being the best in the class. Performance goals are often divided into two types, approach and avoidance. Performance approach goals focus on demonstrating competence and outperforming others. Performance avoid goals focus on trying to avoid looking incompetent (Schunk, Pintrich, & Meece, 2008). Dweck (1999) has reported positive effects of mastery goals on learning in experimental studies, and Schunk et a l. described studies that have shown relationships between students mastery goals and their selfreports of use of self regulatory learning strategies in education from elementary school through college. At the college level, Hsieh, Sullivan, and Guerra (2007) examined the relationship of goal orientation and perceived self -efficacy to GPA. Participants were 112 undergraduates from a large university in the southwest that serves predominantly Hispanic students. Fortysix percent of the sample was Hispanic, and 41% was Caucasian. Goal orientation was measured using the Achievement Goal Orientation Inventory (Elliot & Church, 1997). Performance-avoid and mastery goals were the strongest predictors of GPA. Performanceavoid goals were negatively related to GPA ( r = -.35, p < .01) and mastery goals were positively related to GPA ( r = .40, p < .01) ; however, performance-approach goals were not related to GPA. This study suggests th at mastery goals may be important to academic success in college. Students with mast ery goals

PAGE 31

31 were more likely to succeed academically ; however, further research is needed. As the study was correlational, the direction of the relationship is unclear, and Hsieh et al. did not control for prior ability To explore further the relationship between mastery goals on the basis of the results of Hsieh et al., i n my model I hypothesized that mastery goals predict academic achievement and performance avoid goals are negatively related to academic achievement. Although I found no study that has examined the relationship of achievement goals to college persistence or social adjustment Bean and Eaton (2000) hypothesized in their model that the psychological processes affect both academic and social integration. On this basis, I hypothesized that mastery goals predict social adjustment and performance-avoid goals predict social adjustment. Perceived A cademic S elf -E fficacy Perceived self -efficacy is defined as the extent to which one feels confident to perform a task (Bandura, 1997; Gaskill & Woolfolk Hoy, 2002). Perceived academic self efficacy is the belief in ones ability to successfully perform academic tasks. Many studies have shown that students with strong academic self -efficacy beliefs display higher levels of motivation and skills and earn hi gher grades than students with weaker levels of perceived academic self efficacy (Schunk et al., 2008). Experimental research has shown that raising perceptions of selfefficacy increases students perseverance and achievement (Bandura, 1997). In their metaanalysis of the relationship of self-efficacy beliefs to academic outcomes, Multon, Brown, and Lent (1991) found that students perceived selfefficacy was related to their academic performance at all educational levels, but the relationship was stronges t for college students ( r = .35). In a study examining predictors of college performance, Elias and MacDonald (2007) examined the effects of prior achievement and perceived academic self efficacy on college achievement in 202 undergraduates, using the Aca demic Self Efficacy Scale (ASES; Elias &

PAGE 32

32 Loomis, 2000). Results indicated that both academic self-efficacy and high school GPA predicted college GPA, but their regression analysis showed that perceived academic self efficacy accounted for a greater proportion of variance in college GPA than did high school GPA. In a meta-analysis of 109 studies examining psychosocial and study skill factors and their relationship to college outcomes, Robbins, Lauver, Le, Davis, Langley, and Carlstrom (2004) found that perce ived academic self efficacy predicted retention ( true score correlation = .36) and GPA ( true score correlation = .50) In another study of 364 college students, Bembenutty (2007) found that perceived academic selfefficacy was significantly correlated with course grades for Caucasian males ( r = .62), Caucasian females ( r = .62), minority males ( r = .51) and minority females ( r = .75). Perceived academic self efficacy was measured using the academic self efficacy subscale of the Motivated Strategies for Lear ning Questionnaire (MSLQ; Pintrich, Smith, Garcia, & McKeachie, 1991). In sum, t hese three studies all indicate that perceived academic self efficacy predicts GPA. Perceived self -efficacy has also been linked to social outcomes in college. DeWitz and Wals h (2002) examined the relationship between college satisfaction and perceived selfefficacy in 312 undergraduates from a large Midwestern university using the College SelfEfficacy Inventory (CSEI; Solberg, OBrien, Villareal, Kennel, & Davis, 1993). The C SEI measures perceived self efficacy in various areas of college life including courses and roommates. College satisfaction was measured using the College Student Satisfaction Questionnaire, Form D (CSSQ; Betz, Betz, & Menne, 1989). One of the subscales of the CSSQ is social life, which measures the perception of ones social network and whether there are opportunities to interact socially. Results indicated that perceived college self efficacy was related to overall college satisfaction ( = 48). An ANOVA comparing highand low selfefficacy groups indicated a significant

PAGE 33

33 difference in overall college satisfaction between the low and highself efficacy groups ( F = 43.91, p < .0001). A MANOVA was conducted comparing the high and low selfefficacy groups on the five subscales. For the social life subscale, there was a significant difference between the high and low groups, indicating that students with higher perceived selfefficacy for college had higher social life satisfaction ( F = 30.41, p < .0001). A possible interpretation of this finding is that students with high perceived academic self efficacy are likely to be more successful in academics and therefore less stressed, so they were more likely to enjoy their social life than students with low perceived self efficacy. However, it is also possible that students with higher life satisfaction may have higher perceived self efficacy due to their positive feelings about their social life. In relation to other types of motivation, perceived self -efficacy has been linked to goal orientations. At the college level, H sie h et al. (2007) examined the relationship of goal orientation and self-efficacy on GPA. Participants were 112 undergraduates from a large university in the southwest that predominantly serves Hispanic students. Forty-six percent of the sample was Hispanic and 41% was Caucasian. Goal orientation was measured using the Achievement Goal Orientation Inventory (Elliot & Church, 1997). Perceived academic selfefficacy was measured using items from the Patterns of Adaptive Learning Survey (PALS; Midgley, Maehr, & Urdan, 1993). The results showed that students with higher perceived self efficacy reported significantly more mastery goals than students with lower perceived self efficacy ( F = 13.16, p < .001). There was no significant difference in the number of performance -approach and performance-avoid goals for students for students with high and low perceived academic self efficacy.

PAGE 34

34 In sum, perceived academic self efficacy has been linked with academic achievement (Bembenutty, 2007; Elias & MacDonald, 2007; Multon et al., 1991; Robbins et al., 2004), retention (Robbins et al., 2004), and social adjustment (DeWitz & Walsh, 2002), and mastery goals (Hsieh et al., 2007). On the basis of those findings, I hypothesized in my model that perceived academic self efficacy predicts mastery goals, academic achievement, social ad justment, and intention to persist. The Bioecological Systems The Family Microsystem Within the family microsystem, the process variables of parental involvement, parental autonomy support, and parental relatedness w ere examined. The research evidence indicates that each of these variables has been linked to one or more of the outcome variables of academic achievement, soci al adjustment, and intention to persist. In the following sections, e ach variable is reviewed separately, and any research that connects the family microsystem variables to any of the psychological processes variables is also discussed. Parental involveme nt Grolnick (2003) defined parental involvement as the provision of resources by the parent for the child (p. 16). Many studies have shown the relationship of parental involvement to students school success in elementary and secondary schools (see Hoov er -Dempsey & Sandler, 1995, for a review of this literature). It is likely that parents can and do play a role in whether their children persist to graduate from college. Recent research on parental involvement during the college years suggests that it may reduce the rate of attrition among college students. Researchers have reported relationships between parental involvement in college and their students educational aspirations (McCarron & Inkelas, 2006), college choice (Bers, 2005), self-regulation (Hofe r, Kennedy, & Hurd, 2006), academic and emotional adjustment ( Duchesne, Ratelle, Larose, &

PAGE 35

35 Guay, 2007), achievement, autonomy, competence, and relatedness (Ratelle, Larose, Guay, & Sencal 2005), and perceived selfefficacy (Cordell -McNulty & Ashton, 2006). In a study of parental involvement at the college level, Ratelle, Larose, Guay, and Sencal (2005) investigated a mediational model in which college students perceptions of parental involvement predicted persistence in a science curriculum by fosterin g the self -processes of autonomy, competence, and relatedness. Participants were recruited during their last year of high school in Quebec, Canada. The data were analyzed for the 262 participants who had completed all measures. Parental involvement was measured by assessing students perceptions of their parents involvement in their vocational decisionmaking process using a scale adapted from Barnes and Olson (1992). The researchers measured science achievement by the students self reports of their high school science grades and persistence in a science program by checking to see if students were still enrolled in a science program during their third year of college. The results showed that students perceptions of their parents involvement were positive ly associated with science achievement ( r = .14, p < .05), but perceived parental involvement was not a predictor of persistence in the science program. Students perceptions of parental involvement were positively associated with their autonomy ( r = .34, p < .001), competence ( r = .24, p < .001), and relatedness ( r = .28, p < .001). Analysis of the authors mediation model showed that students perceptions of parental involvement predicted their feelings of relatedness ( = 31, p < .05) and autonomy in the science program ( = 22, p < .05). On the basis of these findings, the authors proposed that perceived parental involvement predicts student selfprocesses but not persistence, which was an outcome of the self-processes. This study shows that perceived parental involvement was related to science achievement; however a major weakness of this study is that the authors analyzed students grades in high school, not in college. I

PAGE 36

36 investigate d whether parental involvement predicts college achievement. Another weakness of this study is that a large percentage (71%) of the sample of college students lived at home with their parents. It is possible that parental involvement had a significant relationship to students science achievement because many of these parti cipants lived with their parents. It is unclear whether the findings generalize to college samples where most of the students do not live with their parents. I examined a sample of university undergraduates with only a small percentage who live at home with their parents. In regard to perceived academic self -efficacy, in a study of 346 students at a state university in the Northeast, Cordell-McNulty and Ashton (2006) found a positive relationship between parental involvement in college and perceived academi c self efficacy using the self efficacy subscale of the Motivated Strategies for Learning Questionnaire (MSLQ; Pintrich et al., 1991). Perceptions of parental involvement was measured by questions from the student survey of the High School and Family Partnership (Epstein, Connors, & Salinas, 1993) that were adapted to measure aspects of college students perceptions of parental involvement. Students who knew their parents had expectations for them ( t = 3.41, p = .001) and received praise from their parents ( t = 3.39, p = .001) were more likely to have higher perceived academic self efficacy. Byars -Winston and Fouad (2008) examined the relationship between students perceptions of their parents involvement and students perceived selfefficacy in math and sc ience using the Math/Science Self -Efficacy Scale (Smith & Fouad, 1999) and the Parental Involvement Scale (Ferry Fouad, & Smith, 2000), which measured students perceptions of parental encouragement and expectations for math, science, and career choices. Participants were 227 undergraduates from two Midwestern universities. Results indicated that students

PAGE 37

37 perceptions of their parents involvement directly predicted perceived mathematics and science self efficacy ( = 13). On the basis of these studies showing that parental involvement is related to perceived academic self efficacy in c ollege students ( Byars -Winston & Fouad, 2008; CordellMcNulty & Ashton, 2006) and academic achievement (Ratelle et al., 2005) I hypothesized in my model that parental involvement predicts perceived academic self efficacy and academic achievement. R elatedness to p arent s Relatedness refers to the human need to feel connected to others and involves loving and caring for others and being loved and cared for by others and is one of the three basic psychological needs according to self -determination theory (Deci & Ryan, 2000). Relatedness is sometimes called belongingness and more recently in the counseling literature, mattering to others. Baumeister and Leary (1995) described the belongingness hypothesis as human beings having a pervasive drive to form and maintain at least a minimum quality of lasting, positive, and significant interpersonal relationships. Research has shown that a lack of relatedness, belongingness, or mattering to others predicts negative psychological outcomes. Baumeister and Leary (1995) reviewed a wide range of research showing that being rejected, ignored, or excluded, that is experiencing a lack of relatedness, is associated with strong negative feeling s such as depression and anxiety. The absence of mattering, a counseling term for relatedness (Dixon Rayle, 2006), has also been linked to higher levels of depression and academic stress ( Dixon Rayle & Chung, 2007). Cutrona, Cole, Colangelo, Assouline, and Russell (1994) examined whether parental social support predicted GPA for undergraduates. They conducted three studies. In the first study participants were 418 undergraduates at a Midwestern university. Parental social support was measured using the Social Provisions Scale Parent Version (SPS -P; Cutrona, 1989; Cutrona &

PAGE 38

38 Russell, 1987). Causal modeling showed that parental social support predicted college GPA ( = .19, p < .01). In the second study participants were 378 undergraduates from a Midwestern university. For this study the entire Social Provisions Scale (SPSSS) was administered to assess social support from parents, friends, and romantic partners. Once again, parental social support predicted GPA ( = 14, p < .05), but social support from friends and romantic partners did not. These studies are unique in that they measure perceptions of parental support during college and GPA during college. On the basis of these findings, I examined whether relatedness to parent s predicts academic achievement in college students. In relation to identity formation, Luyckx, Goossens, et al. (2006) analyzed students perceptions of supportive parenting and identity with 565 freshmen at a University in Belgium. The authors assessed supportive parenting by assessing e motional separation from parents using a shortened version of the Psychological Separation Inventory (PSI; Hoffman, 1984). The sample consisted of Belgian college students who were predominantly Caucasian and over 85% female. They reported that among college students emotional separation from their parents was negatively related to Commitment Making ( = -.17, p < .01) and Exploration in Depth ( = -.19, p < .001), indicating that college students who experienced relatedness to parent s were more like to make an identity commitment and seek out information about that commitment. As mentioned earlier, the major concern about this study is whether the results generalize to U.S. students. The results from Luyckx, Goossens, et al. indicate a relationship between parental relatedness and identity formation, in particular that parental relatedness is positively related to Commitment Making. In relation to goal theory, Moller, Elliot, and Friedman (2008) examined the relationship between achievement goals and parental closeness. Two studies were conducted, one examining

PAGE 39

39 parental closeness after an exam and prior to receiving feedback, and one after receiving feedback. Par ticipants in the first study were 182 undergraduates. The researchers measured achievement goals using the Achievement Goal Questionnaire (Pekrun, Elliot, & Maier, 2006) and perceived closeness to parents with a measure created by McGregor and Elliot (2005). This measure consisted of two questions, Right now, how close do you feel to your mother/father? Results indicated that mastery approach goals were positive predictors of perceived closeness to parents after taking the exam ( = 19, p < .01) Partici pants in the second study were 227 undergraduates. The only difference in Study 2 was that perceptions of closeness to parents was measured again after students received a grade for the exam. Results indicated that students with mastery approach goals reported feeling closer to their parents after receiving feedback ( = 10, p < .05). For students with performance-avoid goals, feelings of closeness to their parents depended on performance outcomes. However, the direction of causality of this relationship cannot be determined from this study. I hypothesized that parental relatedness predicts students tendency to form mastery goals because if students feel close to their parents they are less worried about impressing them and are more likely to focus on learning for understanding. In summary, r esearch has linked relatedness to parents with Commitment Making (Luyckx, Goossens, et al., 2006), mastery goals (Moller et al., 2008), and academic achievement (Cutrona et al., 1994). On the basis of these studies, I hypothesized in my model that relatedness to parent s predict s Co mmitment Making, mastery goals and academic achievement Parental autonomy support Need for autonomy, defined as the desire to have agreement between ones actions and experiences and ones sense of self, is one of the three basic psychological needs po stulated in self -determination theory (SDT) (Deci & Ryan, 2000). Research has shown that children whose parents provide autonomy support that is, who support their childrens need for autonomy, tend

PAGE 40

40 to be more self regulated and have higher academic achievement than children whose parents provide less support of their autonomy (Grolnick, 2003). Ratelle et al. (2005) examined the relationship between parental autonomy support and their college students persistence in a science program. The researchers meas ured parental autonomy support by assessing students perceptions of their parents autonomy support in relation to their decision to pursue science studies in college. Items were adapted from Paulson, Marchant, and Rothlisberg (1994) and Robinson et al. (1995). A sample item is my parents allowed me to have my own point of view regarding my choice of program. Ratelle et al. found that parental autonomy support was positively associated with students science achievement ( r = .20, p < .01) and that parent al autonomy support predicted persistence in the college science program 2 years later ( = 20, p < .05) A weakness of this study is that the researchers only assessed science achievement in high school, not in college. I will examine parental autonomy support during college to see if it predicts college achievement. In a more recent exami nation of the conceptualization of parental autonomy support, Soenens et al. (2007) completed two studies and found that parental autonomy support, conceptualized as promotion of volitional functioning, predicted psychological adjustment in college student s. Participants in the first study were 390 firstyear students at a Belgian university. Parental autonomy support was measured using items from the Autonomy Support scale of the Perceptions of Parents Scale (POPS; Grolnick Ryan, & Deci, 1991) and one ite m from Silk, Morris, Kanaya, and Steinberg (2003). Psychosocial functioning was assessed by examining three constructs : depression, selfesteem, and social well being. Depression was measured using the CES -D (Radloff, 1977). Selfesteem was measured using the RSES (Rosenberg, 1965), and social wellbeing was measured using the Social Well Being scale from

PAGE 41

41 the Louvain WellBeing Scale (Marcoen, Van Cotthem, Billiet, & Beyers, 2002). The results of this study indicated that parental autonomy support predicted psychosocial functioning ( = 05, p < .05). In the ir second study the researchers studied 495 undergraduates from a Belgian university using the same measures except that they measured social well being using the Social Adjustment subscale of the SACQ (Baker & Siryk, 1984). This study confirmed the results from the first study. Parental autonomy support predicted psychosocial functioning ( = 34, p < .001). However, most Belgian college students live at home; therefore, these students have frequent conta ct with their parents. It is unclear whether these results will generalize to college students in the U.S. who do not live at home. In contrast, I examined whether the relationship of parental autonomy support to college achievement, social adjustment, and intention to persist is evident in a sample of students who typically do not live at home. In relation to identity, Luyckx Goossens, et al. (2006) analyzed identity and supportive parenting with 565 freshmen at a university in Belgium by measuring paren tal responsiveness, autonomy support, and psychological control. Parental responsiveness items were from the Child Report on Parent Behavior for Older Children and Adolescents (CRPBI -30; Schludermann & Schludermann, 1988). Psychological control items were from the Parental Psychological Control ScaleYouth SelfReport (PCS YSR; Barber, 1996). Autonomy support items were from the Perceptions of Parents Scales (POPS; Grolnick et al., 1991). The researchers found a positive relationship between a supportive parent -child relationship and Commitment Making ( = 16, p < .01) and Identification with Commitment ( = 21, p < .001) In another study Luyckx, Soenens, Vansteenkiste, et al. (2007) examined the relationship of parental psychological control to identity formation in college students. Parental psychological control is generally considered to be the opposite of parental autonomy support and an inhibitor of autonomy in adolescents. The

PAGE 42

42 researchers conducted a longitudinal study with 565 Belgian undergraduat es using the UtrechtGroningen Identity Development Scale (UGIDS; Meeus & Dekovic, 1995 ) to measure Identification with Commitment and Exploration in Depth and the Ego Identity Process Questionnaire (EIPQ; Balistreri et al., 1995) to measure Commitment Ma king and Exploration in Breadth. Parental psychological control was measured using the Psychological Control Scale Youth Self-Report (Barber, 1996). Students were assessed on these measures five times, with 6 months between each data collection. Luyckx, Soenens, Vansteenkiste, et al. found that parental psychological control was associated with less Commitment Making ( = -.10, p < .05) and less Identification with Commitment ( = -.12, p < .05), suggesting that parental autonomy support should be associated with more Commitment Making and Identification with Commitment. However, the results of Luyckx Goossens, et al. and Luyckx, Soenens, Vansteenkiste, et al. may not be generalize to other populations. The participants in these studies were all Caucasian Belgian university students. I examined whether these results generalize to a sample of U.S. college students. In s umm ar y, research has linked parental autonomy support with Commitment Making and Identifi cation with Commitment (Luyckx Goossens, et al., 2006; Luyckx, Soenens, Vansteenkiste, et al., 2007), academic achievement (Ratelle et al., 2005), social adjustment (Soenens et al., 2007), and persistence (Ratelle et al., 2005). In my model, I hypothesized that parental autonomy support is positively related to Commitment Making, Identification with Commitment, academic achievement, social adjustment, and intention to persist The Peer Microsystem Peers play an important role in the lives of adults during emerging adulthood (Arnett, 2000, 2006). Because of the potential role of peers in this period, I examined the peer microsystem. The peer microsystem consists of the process variable of off -campus friend

PAGE 43

43 relatedness. The research evidence indicates that peer relatedness has been linked to identity development and social adjustment. Peer relationships tend to become more important as students age through adolescence and emerging adulthood, and it is likely that peer relationships play an important role in the lives of most college students. In her study of college outcomes, Fischer (2007) used data from the National Longitudinal Survey of Freshmen. Equal numbers of White, Black, Hispanic, and Asian students were sampled from each college resulting in a total of 3,924 participants. Outcome variables for this study were college GPA and whether students left their initial college by the end of their junior year. Fischer found that, for minority students, greater involvement in formal social activities as measured by extracurricular activities and volunteering was related to higher grades ( = 08, p < .05 for Asian students, = 17, p < .001 for Hispanic students, and = 11, p < .01 for Black students), but for Black and White students, having more ties off campus, measured by the number of visits home and how many friends the student fr equently contacted off campus, was negatively related to their grades ( = -.15, p < .05 for Black students and = .14, p < .10 for White students). For all ethnic groups, a strong relationship between relatedness to peers and college retention was found. However, the direction of the relationship differed for on-campus and off-campus friends. A s scores for informal ties on campus increased, measured by the number of close on-campus friends, hours spent partying per week, and hours spent with friends per week, the likelihood that students would leave college decreased ( = -.89, p < .05 for White students, = -1.65, p < .001 for Asian students, = -1.56, p < .01 for Hispanic students, and = -1.20, p < .01 for Black students) Off campus ties however, were related to a greater probability of leaving college for Black ( = 1.26, p < .05) and White students ( = 1.43, p < .05) Th ese findings suggest that for college students the relationship of peer relatedness to

PAGE 44

44 academic achievement and persistence may differ for on -campus and off-campus friends, so I examined related ness to both off-campus and oncampus friend s. In a study of the importance of friends, Buote et al. (2007) examined the relationship between friendship and adjustment during the first year of college in 702 first-year students from six different Canadian universitie s. The SACQ (Baker & Siryk, 1984) was used to measure adjustment, and the McGill Friendship Questionnaire Friends Functions (MFQFF, short form; Mendelson & Aboud, 1999) was used to measure friendship. The MFQFF measures six functions of friendship in reference to a particular friend: stimulating companionship, help, intimacy, reliable alliance, self -validation, and emotional security. Buote et al. found that overall friendship quality predicted all forms of university adjustment, academic ( t = 4.32, p < .001), social ( t = 15.53, p < .001) and institutional attachment ( t = 10.95, p < .001) In a study examining social support and adjustment of 115 firstyear undergraduates at a midsized Canadian university, Friedlander Reid, Shupak, and Cribbie (2007) examined the relationship between social support from friends and adjustment to college. Data were collected during their first and second semesters. Social support from friends was assessed using the Multidimensional Scale of Perceived Social Suppor t (MSPSS; Zimet, Dahlem, Zimet, & Farley, 1988). Adjustment was measured usin g the SACQ (Barker & Siryk, 1989 ). From the fall to spring semester, increas ed social support from friends predicted improved overall adjustment ( = 19, p < .05) social adjustment ( = 20, p < .05), and personalemotional adjustment ( = 19, p < .05). The results of these two studies ( Buote et al., 2007; Friedlander et al., 2007) illustrate the positive relationship between friends and adjustment, but neither study examined if relatedness to peer s is related to academic achievement or intention to persist.

PAGE 45

45 In summary, research has linked relatedness to offcampus friend s with social adjustment (Buote et al., 2007; Friedlander et al., 2007). On the basis of these studies, I hypothesized that relatedness to offcampus friend s predict s social adjustment in my model. The College Microsystem The college environment is likely to play an important role in the success of college students. The college microsystem is a part of social integration in Tintos (1975) model. The college microsystem includes the process variables of relatedness to instructor, relatedness to oncampus friend s, and belonging on campus, and the person variable of extracurricula r activities. The research evidence indicates that each of these variables has been linked to one or more of the outcome variables of academic achievement, social adjustment, and intention to persist. Each variable is reviewed separately and research that connects the college microsystem variables to any of the psychological processes variables is also discussed. Relatedness to o n-c ampus friend s As mentioned earlier in the section on peer relatedness, college students tend to interact with on-campus and off-campus peers. In particular, Fischer (2007) in her study of data from the National Longitudinal Survey of Freshmen obtained results suggesting that onand offcampus friendships with peers may differ in their relationships to students grades and retention. For Black and White students, having more ties off-campus was negative ly related to their grades and retention, and, for all ethnic groups, having more informal ties on campus was related to retention. Because of the differen t relationships between relatedness to onand offcampus friend s and the outcome variables of academic achievement, social adjustment, and intention to persist in college I examined relationships with relatedness to oncampus friend separately from relatedness to offcampus fri end s to see if differences exist in my analysis Because informal on campus ties have been related to persistence (Fischer, 2007), in my model I hypothesized that

PAGE 46

46 relatedness to oncampus friend s predict s intention to persist I also predict ed that oncamp us relatedness to friend s is related to social adjustment because of the research mentioned earlier in the section on offcampus friends linking friends to social adjustment (Buote et al., 2007; Friedlander et al., 2007). I also hypothesized that relatedness to oncampus friends predict s social ad justment and intention to persist. Relatedness to instructors Teacher student relationships have always been an important research topic in elementary and secondary education. Instructor-student relationships have not been as thoroughly researched at the college level. Some researchers have examined students relatedness to their instructor and found positive relationships. Larose, Tarabulsy, and Cyrenne (2005) examined relatedness as a moderating factor in mento ring relationships and student adjustment for 40 firstyear college students. Participants were 40 students from a private college who were identified as at -risk on the basis of low grades in high school. The mean age of the participants was 18.6 years, and 72% were female. All participants received 10 hours of mentoring from an instructor. Before beginning the mentoring program participants social adjustment was assessed with the SACQ (Baker & Siryk, 1984) to measure adjustment. After completing the 10 -hour mentoring program, participants completed an adapted version of the Inventory of P arent and P eer A ttachment ( IPPA; Armsden & Greenberg, 1987) a measure of their perceived relatedness to their mentor Five months after completing the mentoring program, the SACQ was re administered to participants. The results showed that students with high relatedness with their mentors had better social adjustment and institutional attachment compared to students with low relatedness and students who did not go through the mentoring program. Those students who had low relatedness were less well adjusted to college and received lower grades. The researchers found that students with high relatedness reported better social adjustment and institutional attachment. Students

PAGE 47

47 w ith low relatedness reported worse academic and emotional adjustment. Larose et al. highlighted the potential positive effects of students relatedness to their instructor; however, they assessed relatedness to mentors, not relatedness to instructors. All of the mentors were instructors, but this study examined the mentoring relationship in particular. I examine d relatedness to instructor instead of relatedness to mentors. In her study comparing college outcomes by ethnic group, Fischer (2007) assessed f or mal academic ties with professors by asking students how often they asked questions in class and after class, asked questions when information was unclear, visited professors in their office to ask questions, and visited professors in their office to talk. Results indicated that more connections with professors were related to higher GPA for all ethnic groups ( = 07, p < .001 for White students, = 06, p < .01 for Asian students, = 09, p < .001 for Hispanic students, and = 04, p < .10 for Black students) These relationships though modest suggest that positive relationship s with instructors contribute to students GPA. One study has linked relatedness to instructor to perceived academic self -efficacy. In a study of 172 community college students, Cordell-McNulty and Ashton (2008) assessed students sense of relatedness to their instructors with four items from a scale developed by Furrer and Skinner (2003) and p erceived academic self -efficacy using the MSLQ (Pintrich et al., 1991). C onfirmatory fact or analysis showed that relatedness to instructor loaded on two factors identified as Instructor Acceptance and Ignored by Instructor Perceived academic self -efficacy also loaded on two factors identified as Perceived Self Efficacy Outcome and Perceived Self Efficacy Mastery Perceived Self Efficacy Ma stery consist ed of five items measuring students belief that they can master the course content. Perceived Self -E fficacy Outcome consist ed of three items measuring students belief in their ability to achiev e a successful outcome in the course. Results showed that Instructor Acceptance

PAGE 48

48 predicted both Perceived Self Efficacy Outcome ( F = 5.00, p < .05) and Perceived Self Efficacy Mastery ( F = 5.16, p < .05) Though the researchers measured perceived self effic acy and not achievement, academic self efficacy has been linked with academic outcomes (Bembenutty, 2007; Multon et al., 1991). R esearch has linked relatedness to instructor with academic achievement (Fischer, 2007; Larose et al., 2005), social adjustment (Larose et al., 2005), and perceived academic self efficacy (Cordell -McNulty & Ashton, 2008). In my model, I hypothesized positive relationship s between relatedness to instructor and academic achievement, social adjustment, and perceived academic self eff icacy. Belonging on campus Sense of belonging refers to the feeling that one is valued as a part of the environment (Hagerty, Lynch -Saucer, Patusky, Bouwsema, & Collier, 1992). Belonging on campus is feeling a part of the campus environment. Many college retention studies include sense of belonging, but it is often combined with other factors including institutional fit and commitment (Hausmann, Schofield, & Woods, 2007). Hurtado and Carter (1997) and Hausmann et al. argued that a sense of belonging on campus needs to be examined in research as a separate construct from active involvement in the community. Hurtado and Carter and Hausmann et al. also suggested that belonging to campus might be an important contributor to minority adolescents persistence and achievements in college in particular. In a study of the contribution of campus climate to adjustment, Mounts (2004) examined sense of belonging on campus and college adjustment in a sample of 319 college freshmen. Sense of belonging was measured using the sense of belonging subscale of the Perceived Cohesion Scale (PCS; Bollen & Hoyle, 1990). Anxiety and depression were measured using the Beck Anxiety Inventory (BAI; Beck, Epstein, Brown, & Steer, 1988) and the Beck Depression

PAGE 49

49 Inventory (BDI; Beck, Ward, Mendelson, Mock, & Erbaugh, 1961). For White adolescents a higher sense of belonging on campus was related to lower levels of anxiety ( = -.14, p < .05), and for African American adolescents a higher sense of belonging on campus was related to lower levels of depression ( = -.41, p < .001), indicating that sense of belonging is related to adjustment, but the relationships may differ for different ethnic groups. Because sense of belonging was negatively related to anxiety and depression, it is likely that sense of belonging is positively related to social adjustment in college students. In a study of 324 students at a small liberal arts college Ostrove and Long (2007) also found a link between sense of belonging and social adjustment. Sense of belonging was measured using 1 item from the SACQ (Baker & Siryk, 1999), I feel that I fit in well as part of the college environment, and another item the authors created, Overall, to what extent do you feel you belong at [college name]. Cronbachs alpha was .8 5 for this 2 item measure. Social adjustment was measured using the social adjustment subscale of the SACQ, minus the one item used to assess sense of belonging. Results indicated that sense of belonging was significantly related to social adjustment ( = .58, p < .001). These studies emphasize the relationship of sense of belonging to social adjustment. Hausmann et al. (2007) examined sense of belonging and intention to persist in 365 firstyear college students from a large mid Atlantic university. Sense of belonging was measured using the sense of belonging subscale of the PCS (Bollen & Hoyle, 1990). Persistence was measured by asking students whether they intended to complete their degree at their institution. Results showed that sense of belonging predi cted intention to persist at the beginning of the year (parameter estimate = .30, p < .001); however, sense of belonging declined significantly over the year. The authors argued that as students progress through the school year the practicality of getting a college degree may become more important than sense of belonging in determining

PAGE 50

50 intention to persist. Even though Hausmann et al. found that sense of belonging decreased as students progressed through their first year of college, I hypothesized that sense of belonging predict s college student persistence. In relation to perceived self -efficacy, Freeman, Anderman, and Jensen (2007) examined sense of belonging in 238 college freshmen at a university in the southeast using the perceived self efficacy subscale of the MSLQ (Pintrich et al., 1991). To measure sense of class belonging, an adapted version of the Psychological Sense of School Membership (PSSM; Goodenow, 1993) was used. Results indicated that students sense of class belonging was strongly associa ted with perceived academic self efficacy adjustment ( = 58, p < .001). In my model I hypothesized that sense of belonging predict s perceived academic self efficacy. Recently, Rodgers and Summers (2008) proposed a revision to Bean and Eatons (2000) mo del for examining the retention of African American students at predominantly White institutions. In particular, they proposed that attitudes precede rather than follow the psychological processes so sense of belonging directly affects all the psychological process variables : self -efficacy, goal orientations, attributions, and intrinsic motivation. Because of this theoretical relationship proposed in the Rodgers and Summers model, I hypothesized that sense of belonging is positively related to Commitment Making, Identification with Commitment, mastery goals, and pe rceived academic self efficacy and is negatively related to performanceavoid goals. Research has also linked sense of belonging with adjustment (Mounts, 2004; Ostrove & Long, 2007), intention to persist (Hausmann et al., 2007), and perceived academic self -efficacy (Freeman et al., 2007). Consistent with these studies, in my model I hypothesized that sense of belonging predict s perceived academic self efficacy, social adjustment, and intention to persist.

PAGE 51

51 Extracurricular activities Involvement in extracurricular activities may also be an important predictor of college attrition. Numerous researchers have reported that student involvement in extracurricular activities is positively related to educa tional aspirations, bachelors degree attainment, and graduate school attendance (Moore, Lovell, McGann, & Wyrick, 1998). In a study of 172 community college students, Cordell-McNulty and Ashton (2008) found that participation in extracurricular activities predicted GPA ( F = 4.64, p < .03). In a longitudinal study of 695 participants from middle school through age 20, Mahoney, Cairns, and Farmer (2003) found that consistent participation in extracurricular activities was positively associated with education al status at age 20, suggesting that the more students are academically and socially involved, the more they are likely to graduate. Participation in extracurricular activities may be especially important during students first year of college when relatio nships and affiliations are forming (Tinto, 2005). In her study of college outcomes, Fischer (2007) used data from the National Longitudinal Survey of Freshmen. Equal numbers of White, Black, Hispanic, and Asian students were sampled from each college resu lting in a total of 3,924 participants. Outcome variables for this study were college GPA and whether students left their initial college by the end of their junior year. Fischer found that for minority students, greater involvement in formal social activities as measured by extracurricular activities and volunteering, was weakly related to higher grades ( = 08, p < .05 for Asian students, = 17, p < .001 for Hispanic students, and = 11, p < .01 for Black students). However, g reater involvement in ex tracurricular activities was stron gly negatively related to leaving college for minority students ( = -1.89, p < .05 for Asian students, = -1.80, p < .001 for Hispanic students, and = -1.87, p < .05 for Black students). These studies show a weak posit ive relationship between participation in extracurricular activities and GPA but a strong positive relationship with persistence.

PAGE 52

52 Bohnert, Aikins, and Edidin (2007) examined the relationship between extracurricular activities and social adaptation du ring the transition to college in 85 adolescents who completed measures during the summer before their first year in college ( Time 1) and towards the end of their first year in college ( T ime 2). Extracurricular activities were measured by the number of activities students participated in (breadth) and the number of hours each week students spent on extracurricular activities (intensity). Friendship quality was assessed using the Friendship Quality Questionnaire (FQQ; Parker & Asher, 1993). Loneliness and social dissatisfaction were assessed using the Loneliness and Social Dissatisfaction Questionnaire (Asher & Wheeler, 1985). The researchers found that the number of hours a student participated in extracurricular activities each week (intensity) was significantly correlated with friendship quality ( r = .27, p < .05) and loneliness ( r = -.27, p < .05). The number of activities a student participated in each week (breadth) was negatively correlated with loneliness ( r = -.23, p < .05). In summary, previous research demonstrates relationships between participation in extracurricular activities and academic achievement (Cordell -McNulty & Ashton, 2008; Fischer, 2007), social adaptation (Bohnert et al., 2007), and intention to persist (Fischer, 2007; Moore et al., 1998). On the basis of these studies, in my model I hypothesize d that involvement in extracurricular activities predict academic achievement, social adjustment, and intention to persist. The Self System Research suggests that several student characteristics are related to success during college. Because of the importance of the person in college success, I examined the self system. The self system in this study consists of the person variables of personality and academic ability, and the process variable of theory of intelligence. The research evidence indicates that each of these variables has been linked to one or more of the outcome variables of academic

PAGE 53

53 achievement, social adjustment, and intention to persist. Each variable is reviewed separately, and research that connects these self system variables to any of the psychological processes variables is also discussed. Personality Personality, in particular the five -factor model of personality, has often been linked with academic outcomes (Nguyen, Allen, & Fraccastoro, 2005; Wolfe & Johnson, 1995). The five factor model is one of the most popular conceptions of personality traits. The model is composed of Extraversion, Agreeableness, Conscientiousness, Neuroticism, and Openness. Extraversion refers to being assertive, active, enthusiastic, and energetic. Agreeableness refers to being cooperative, generous, appreciative, and trustful. Conscientiousness refers to being responsible, orderly, organized, and reliable. Neuroticism refers to being anxious, tense, and unstable and Openness refers to being imaginative, artistic curious, and insightful (John & Srivastava, 1999; McCrae & John, 1992). Of these five personality factors, conscientiousness has been consistently shown to predict academic achievement. Nguyen et al. investigated the relationship between the five -factor model and college GPA. Participants were 360 undergraduate and graduate students at a Southern university. Personality was measured using the Big 5 Personality Inventory developed by Goldberg (1999). They found that conscientiousness was a stronger predictor of overall GPA than the other four personality traits ( t = 2.916, p < .01) In a similar study, Wolfe and Johnson examined personality and college outcome in 201 undergraduate students at a state university in New York. Personality was measured using the Big Five Inventory (John, Donahue, & Kentle, 1991). Wolfe and Johnson found that conscientiousness was the second strongest predictor of college GPA ( = .31) after average grade earned in high school ( = .43). Because of this research linking conscientiousness as a strong predictor of academic achievement, I examined the relationship of conscientiousness to GPA in my study.

PAGE 54

54 Numerous studies linking personality and identity have been conducted. For example, Luyckx et al. (2005) examined personality and identity with 565 freshmen in educational psychology classes in a university in Belgium, using the Utrecht Groningen Identity Development Scale (U GID S; Meeus & Dekovic, 1995) to assess Identification with Commitment and Exploration in Depth and the Ego Identity Process Questionnaire (EIPQ; Balistreri et al., 1995 ) to assess Commitment Making and Exploration in Breadth. Conscientiousness and openness were assessed using the Dutch version of the NEOFFI ( Ho ekstra et al., 1996). In relationship to the five identity categories Luyckx et al. found that college students in the Achievement category were high on conscientiousness ( M = 3.58) and openness ( M = 3. 60). Students in the Foreclosure category were high on conscientiousness ( M = 3.53), but low on openness ( M = 3.35). Students in Moratorium were low on conscientiousness ( M = 3.37), but high on openness ( M = 3.60). Students in both the Diffused Diffusion and Carefree Diffusion categories scored low on both conscientiousness ( M = 3.13, 3.27) and openness ( M = 3.48, 3.37). All these means were significantly different from other means. These results are consistent with what is expected in each identity status. In a study examining the relationship between identity status and personality, Clancy and Dollinger (1993) examined all five personality traits in 198 undergraduates at a large Midwestern university. Identity was measured using the Extended Objective Meas ure of Ego Identity Status (EOMEIS; Adams, Bennion, & Huh, 1989), and personality was measured using the NEO Personality Inventory (NEO PI; Costa & McCrae, 1992). They found a significant positive relationship between conscientiousness and Identity Achievement ( r = .30, p < .001), and significant negative relationships between conscientiousness and Moratorium ( r = -.22, p < .001) and Diffusion ( r = .38, p < .001). These results are consistent with the findings of Luyckx et al.

PAGE 55

55 In another study of personality and identity, Luyckx, Soenens, and Goossens (2006) examined the relationship between personality and identity in 351 college women at a university in Belgium. The UGIDS ( Meeus & Dekovic, 1995) was used to assess Identification with Commitment, and th e EIPQ ( Balistreri et al., 1995 ) was used to assess Commitment Making and Exploration in Breadth. Personality was assessed using the Dutch version of the NEOFFI (Ho ekstra et al., 1996). The researchers found that Commitment Making and Identification with Commitment were negatively related to neuroticism and positively related to extraversion, agreeableness, and conscientiousness. Using structural equation modeling they reported that conscientiousness at T ime 1 positively predicted Commitment Making at T ime 3 ( = .09, p < .05) and positively predicted Identification with Commitment at T ime 5 ( = .12, p < .05) The major weakness of this study is that all participants were female, so these results cannot be generalized to male college populations. These studies all indicate a strong relationship between identity and personality and in particular a relationship between conscientiousness and Identity Achievement. In previous research conscientiousness has been associated with academic achievement (Nguyen et al., 2005; Wolfe & Johnson, 1995), identity achievement (Clancy & Dollinger, 1993; Luyckx et al., 2005), and Commitment Making and Identification with Commitment (Luyckx et al., 2006). In my model I hypothesized that conscientiousness is positively associated w ith academic achievement, Commitment Making, and Identification with Commitment. Theory of i ntelligence Implicit theories of intelligence refer to the way people understand their own intelligence. Dweck (1999) proposed two types of implicit theories of i ntelligence : entity and incremental. Entity theorists believe that their intelligence is fixed and cannot be changed. In contrast, incremental theorists believe that intelligence is malleable and can be increased. Entity theorists

PAGE 56

56 often have performance go als and make attributions of success and failure to ability ; in contrast, incremental theorists often have mastery goals and make attributions of success and failure to effort or lack of effort (Dweck, 1999). Dweck examined college students implicit theor ies of intelligence and whether they wanted to be challenged or receive a good grade in a course. Sixty eight percent of the incremental theorists wanted to be challenged versus only 35% of entity theorists, indicating that students with an incremental the ory of intelligence are more likely to endorse mastery goals and students with an entity theor y of intelligence are more likely to endorse performance goals. Aronson, Fried, and Good (2002) conducted an experimental study where they attempted to manipula te African American and Caucasian students theories of intelligence. Participants were 79 undergraduates at Stanford University, who were randomly assigned to one of six conditions. For each ethnicity there was a malleable pen pal group, a control pen pal group, and a non pen pal group. Both pen pal groups were told that they were writing letters to a struggling middle -school student. The malleable group viewed a presentation about how new research has shown that intelligence is not fixed. The control pen pal group viewed a presentation about how everyone has intellectual strengths and weaknesses. Both groups then wrote letters to the pen pals and gave presentations about what they wrote. Several days later the participants completed a measure about their theories of intelligence. Nine weeks later the participants again completed a measure about their theories of intelligence, and their semester grades were obtained. Results indicated that both African American and Caucasian students in the malleable pen pal groups had higher GPAs than students in the control groups. For African Americans the differences between the malleable pen pal group and the control pen pal group ( t = 2.19, p <.05) and the non pen pal group ( t = 2.24, p < .05) in GPA were significant. F or Caucasians the differences between the

PAGE 57

57 malleable pen pal group and the control pen pal group ( t = 1.76, p < .09) and the non pen pal group ( t = 1.82, p < .08) in GPA were marginally significant. However, the researchers did not measure persistence, so w e do not know if incremental beliefs would eventually impact persistence. However, Zuckerman, Gagne, and Nafshi (2001) examined the relationship between students implicit beliefs about their major and students persistence in their choice of a major. Pa rticipants were 186 undergraduates at the University of Rochester. Entity beliefs were measured with a scale the authors created. A sample item is ones interests are what they are; they cannot be changed to fit a particular major. There was a significan t entity by perceived competence interaction ( F = 6.83, p < .01), which indicated that students with entity beliefs who believed they were not doing well in their major were more likely to choose a new major than were incremental theorists who were also no t doing well in the major. However, the researchers measured persistence in the major, not persistence in college. Robins and Pals (2002) examined the implicit selftheories of college s tudents in the academic domain with d ata from the Longitudinal Study of Self and Personality Development. Participants were 363 undergraduate students at the University of California at Berk e ley. Data w ere collected throughout the 4 years of college. Implicit self theories were measured with five items adapted from Erdley and Dweck (1993). Performance and learning goal orientations were assessed with items the authors created. Cronbachs alphas for the performance and learning goal orientation scales were .84 and .78 respectively. Affective response to academic achievement w as measured using the Positive and Negative Affect Scale (PANAS; Watson, Clark, & Tellegen, 1998). Results indicated that entity theorists were more likely to endorse performance goals and experience negative affect about their academic performance. In con trast, incremental

PAGE 58

58 theorists were more likely to endorse learning goals and experience positive affect about their academic performance. Path analysis was also used to test the overall implicit self theories model. The final model the authors tested was significant and had a good fit to the data. In the model, entity orientation had a direct effect ( r = .30, p < .05) on performance goals and a direct negative effect on learning goals ( r = -.26, p < .05). In summary, e ntity beliefs about intelligence have been found to be negatively related to GPA (Aronson et al., 2002), persistence (Zuckerman et al., 2001), and mastery goals and positively related to performance goals (Robins & Pals, 2002). Consistent with those relationships I hypothesized in my model a n egative relationship between entity beliefs and mastery goals, GPA, and intention to persist. I also hypothesized a positive relationship between entity beliefs and performance-avoid goals. Previous academic achievement Previous academic achievement (SAT scores and high school GPA) is often used as a predictor of college retention (Robbins et al., 2004). In their meta-analysis of 109 studies of college outcomes, Robbins et al. found that high school GPA, SAT, and ACT scores predicted retention and GPA in a study of college students. Robbins, Allen, Casillas, Peterson, and Le (2006) conducted a largescale study of college outcomes in 14,464 students from 48 2and 4year colleges in the U.S. The resear chers found that the ACT composite score and high school GPA predicted retention and college GPA in 2and 4year institutions (for retention, odds ratio = 1.24, p < .03 for ACT scores at 2 year institutions, odds ratio = 1.28, p < .001 for ACT scores at 4 year institutions, odds ratio = 1.24, p < .001 for high school GPA at 2-year institutions, and odds ratio = 1.25, p < .001 for high school GPA at 4 -year institutions, and for college GPA, = .18, p < .001 for ACT scores at 2 year institutions, = .30, p < .001 for ACT score at 4year institutions, = .24, p < .001 for high school GPA at 2-year institutions, and = .28, p < .001 for

PAGE 59

59 high school GPA at 4year institutions) at 2 and 4ye ar institutions. In a study examining predictors of college performance, Elias and MacDonald (2007) examined the relationship of prior achievement on college achievement in 202 undergraduates. Results indicated that high school GPA predicted college GPA ( = .38, p < .001). In a comparison of college outcomes for different ethnic groups, Fischer (2007) used data from the National Longitudinal Survey of Freshmen. Equal numbers of White, Black, Hispanic, and Asian students were sampled from each college resul ting in a total of 3,924 participants. Results showed that, for all students, high school GPA significantly predicted college GPA. In another study, Mattson (2007) also found that high school GPA was significantly related to academic success in 591 student s who were identified as being at risk at a highly selective, private research university in the Southwest Students were identified as high risk if they had a low high school GPA or low standardized test scores. Results indicated that high school GPA was significantly correlated with firstsemester college GPA ( r = .18, p <. 01) and firstyear college GPA ( r = .20, p < .01). In summary, the studies show that high school GPA predicts college GPA. In sum previous academic achievement has been linked with academic achievement in college (El ias & MacDonald, 2007; Fischer, 2007; Mattson, 2007; Robbins et al., 2004; Robbins et al., 2006) and college retention (Robbins et al., 2004; Robbins et al., 2006). On the basis of these studies I hypothesized that previous academic achievement in high school predicts college GPA in my model. The Macrosystem In 2002-2003, women earned over 57% of all bachelors degrees, and Caucasian students earned 70% of all bachelors degrees (National Center for Education Statistics, 2005a ). Many researchers have noted that gender, ethnicity, and SES are often predictors of college success

PAGE 60

60 (Astin, 1997; Fischer, 2007). In t his study I examined the context variables of gender, ethnicity, and socioeconomic status within the macrosystem. Gender Gender, when included as a predictor variable, is significantly related to college success. Astin (1997) pointed out that gender is one of the factors that colleges can use to predict the retention rate of their students. Robbins et al. (2006) in a study of 14,464 students from 48 2and 4-year colleges in the U.S. found that gender predicted college GPA ( = -.12, p < .001 for 4year institutions and = -.07, p < .07 for 2year institutions) with females having higher GPAs than males. In a study of 172 community college students, CordellMcNulty and Ashton (2008) also found that gender predicted GPA ( F = 3.83, p < .05) with females having higher self reported GPAs than male students. Examining predictors of success in students at-risk at a highly selective, private research university in the Southwest, Mattson (2007) found that gender was significantly related to academic success in 591 students with a low high school GPA or low standardized test scores. Results indicated that gender was significan tly correlated with firstsemester college GPA ( r = .16, p < .01) and firstyear college GPA ( r = .19, p < .01), indicating that females had higher college GPA s than males. Using data from the National Longitudinal Survey of Freshmen, Fischer (2007) sampled equal numbers of White, Black, Hispanic, and Asian students from each college resulting in a total of 3,924 participants. Outcome variables for this study were college GPA and whether students left their initial college by the end of their junior year. For all students, being male was negatively related to cumulative college GPA. This relationship was significant for White ( = -.07, p < .01), Hispanic ( = -.11, p < .05), and Black males ( = -.10, p < .001) All these studies show that being female pre dicts higher GPA. One recent study linked college success with males. Gupta, Harris, Carrier, and Cohen (2006) examined predictors of success of 451 undergraduates in an upperlevel mathematics course at a

PAGE 61

61 public university in the Northeast. Results indica ted that gender was a significant predictor of final course grade with male students receiving higher grades. This result suggests that in some courses or perhaps majors, male students may outperform female students, but female students are more likely to have higher cumulative college GPAs than male students. Several studies have indicated that females have higher cumulative GPAs in college than males (Cordell -McNulty & Ashton, 2008; Fischer, 2007; Mattson, 2007; Robbins et al., 2006). On the basis of the se studies I control l ed for gender in my model. Ethnicity In her examination of college outcomes by ethnicity in 3,924 students from four ethnic groups, Fischer (2007) argued that when examining college outcomes and adjustment, ethnicity cannot be ignored because minority students experience college differently. Astin (1997) argued that ethnicity is one of the factors that can be used to help colleges predict the expected retention rate of their students. In a study of college outcomes, Robbins et al. (2006) examined 14,464 students from 48 2and 4-year colleges in the U.S. and found that ethnicity predicted college GPA at both 2( p < .001) and 4year institutions ( p < .001). In a study of social support and perceived academic self -efficacy in 172 community college students, CordellMcNulty and Ashton (2008) found that ethnicity was a significant predictor of GPA ( F = 8.40, p < .0001), with White students and students who identified themselves as other, having higher GPAs than Black and Hispanic students. On the basis of these studies linking e thnicity to academic achievement in college (Cordell -McNulty & Ashton, 2008; Robbins et al., 2006), I control led for ethnicity in my model. Socioeconomic status SES is often included as a predictor of college outc omes. In their meta -analysis of 109 studies examining psychosocial and studyskill factors and their relationship to college

PAGE 62

62 outcomes, Robbins et al. (2004) found that SES positively predicted retention and college GPA. In their study of college outcomes, Robbins et al. (2006) examined 14,464 students from 48 2and 4-year colleges in the U.S. and found that SES predicted retention (odds ratio = 1.12, p < .001) and college GPA ( = -.80, p < .001) at 4 year institutions with students with higher SES having higher GPAs and persistence. Of note, SES was not a significant predictor of GPA or persistence at 2 -year colleges. A possible reason for this finding is that there was less va riation in SES at 2 year colleges than in the 4 -year sample. Fischer (2007) used data from the National Longitudinal Survey of Freshmen to examine differences in college outcomes by ethnicity. Equal numbers of White, Black, Hispanic, and Asian students wer e sampled from each college resulting in a total of 3 ,924 participants. Results indicated having a family income of more than $75,000 a year significantly predicted college GPA for White ( = .06, p < .05) and Hispanic students ( = .09, p < .05). Overall higher SES levels tend to predict GPA and persistence; however, this relationship may differ by type of institution. In particular, Robbins et al. (2006) did not find a significant relatio nship between SES and GPA or retention for college students at 2-year institutions. In light of the studies indicating a relationship between SES and academic achievement in college (Fischer, 2007; Robbins et al, 2004; Robbins et al., 2006) and college ret ention (Robbins et al., 2004; Robbins et al., 2006), I planned to control for SES in my model. The Chronosystem Age is often a predictor of college success. Because of the importance of age, this study will examine the chronosystem time variable of age. It is reasonable to assume that as students age and progress through college they will become more independent. Studies have shown that older students tend to receive higher course grades (Gupta et al. 2006) and higher GRE scores (Awad, 2007). Gupta et al. examined predictors of success in 451 undergraduates in an upperlevel mathematics course at a public university in the Northeast. Results indicated that age was a

PAGE 63

63 significant predictor of final course grade with older students receiving higher grades. A wad (2007) examined 313 African American students from a historically Black college in the Northeast to identify variables that predict GPA and GRE scores Students completed the GRE verbal subtest. Results showed that the only significant predictor of GRE scores was age ( = .18, p < .05), with older students having higher scores. On the basis of these studies show ing a relationship between a ge and academic achievement ( Awad, 2007; Gupta et al., 2006), I excluded all students over the age of 26. Purpose of the Study T he p urpose of this study wa s to test a model of college success and retention based on a bioecological perspective (Bronfenbrenner & Morris, 1998). This new bioecological systems model of college retention focuses on the ecological systems (family, peer, college, and the selfsystem) that predict college success and the psychological processes (identity, perceived academic self efficacy, and achievement goals) that mediate the relationship between the ecological systems and college success. See Figure 1 for an illustration of the conceptual model. The hypotheses I propos ed to test based on my review of the literature are represented in Table 1 -1. Specifically, parental autonomy support, instructor relatedness, on-campus friend relatedness, sense of belonging, extracurricular activities, and entity beliefs have a direct relationship to intention to persist. Parental involvement, instructor relatedness, and sense of belonging directly affect academic self efficacy which in turn, directly affects intention to pers ist. Parental autonomy support, parental relatedness, sense of belonging, and conscientiousness directly affect commitment making which also directly affects intention to pers ist. Parental autonomy support, sense of belonging, and conscientiousness directl y affect identification with commitment which directly affects intention to persist. Parental involvement, parental autonomy support, instructor relatedness, extracurricular activities, entity beliefs, and

PAGE 64

64 conscientiousness directly affect academic achievement which affects intention to persist. Commitment making, identification with commitment, per ceived academic self efficacy mastery goals, and performance-avoid goals directly affect academic achievement. Sense of belonging and entity beliefs affect performance av oid goals, which directly affect academic achievement. Parent relatedness sense of belonging, and entity beliefs directly affect mastery goals which directly affect academic achievement. Perceived academic self efficacy directly affects mastery goals. Parental autonomy support, off-campus friend relatedness, instructor relatedness, on campus friend relatedness, sense of belonging, and extracurricular activities directly affect social adjustment which in turn, affects intention to persist. Perceived a cademic self efficacy, commitment making, identification with commitment, mastery goals, and performance avoid goals dir ectly affect social adjustment. This model addresses weaknesses in the literature by combining the sociological and psychological perspectives that are commonly used in the literature on college retention. In addition, this model uses Bronfenbrenner and Morriss (1 998) bioecological theory as a framework for examining college retention. Unlike many previous studies, I tested the entire conceptual model, instead of isolated sections. Research Hypotheses Hypothesis 1. Academic achievement, social adjustment, perceived academic self efficacy, commitment making, identification with commitment, parental autonomy support, relatedness to in structors relatedness to oncampus friend s, sense of belonging, extracurricular activities, and e ntity beliefs predict intention to persist. Hypothesis 2. Perceived academic self efficacy, commitment making, identification with commitment, mastery goals, performance -avoid goals, parental involvement, parental autonomy support, relatedness to instructors extracurricular activities, e ntity beliefs and conscientiousness predict academic achievement. Hypothesis 3. Perceived academic self efficacy, commitment making, identification with commitment, mastery goals, performance avoid goals, parental autonomy support,

PAGE 65

65 relatedness to offcampus friend s, relatedness to instructors, relatedness to on campus friend s, extracurricular activities, and sense of belonging predict social adjustment. Hypothesis 4. Perceived academic self efficacy, relatedness to parent s, sense of belonging, and e ntity beliefs predict mastery goals. Hypothesis 5. Perceived academic self efficacy sense of belonging, and e ntity beliefs predict performance-avoid goals. Hypothesis 6. Parental involvement, relatedness to instructors, and sense of belonging predict perceived academic self efficacy. Hypothesis 7. Parental autonomy support, relatedness to parent s, sense of belonging, and conscientious ness predict commitment making. Hypothesi s 8. Parental autonomy support, sense of belonging, and conscientiousness predict identification with commitment. Theoretical Significance Researchers have proposed several models of college retention (Bean & Eato n, 2000; Tinto, 1975, 1993), but they have not conducted adequate tests of these models. Most studies of college retention focus on relatively few variables and lack a theoretical focus that integrates the variables into a coherent framework. My dissertati on is the first study to use Bronfenbrenner and Morriss (1998) bioecological theory as a theoretical framework for creating and testing a conceptual model of student retention in college. This model includes variables from five social contexts (person, family, peers, school, and culture) that have been shown to predict GPA and persistence to graduation. The theoretical significance of this dissertation is that it will advance the study of college retention by providing a cohesive structure for the identification and testing of relationships among the variables that are likely to contribute to students retention in college. Practical Significance The failure of a significant proportion of students who enter college to complete their programs is a serious problem. Estimates suggest that almost 50% of students do not complete their college programs within 5 years ( National Center for Higher Education Management

PAGE 66

66 Systems, 2007). The practical significance of my dissertation is that it might provide researchers and higher education administrators with information on variables that might increase college students timely completion of their degrees. The value of the bioecological perspective is that it suggests variables from several social contexts the person, the home, the school, the culture, and peers that might improve students persistence to graduation. Programs that combine influences from each of these social contexts might be more effective than programs that focus on only one context, typically the school. By examining a more comprehensive set of predictor variables, I may be able to identify variables from these different contexts that contribute to retention of students to graduation. I hope that my study will offer insights for the design of experimen tal studies that show that innovative programs that combine factors from multiple social contexts are more effective in increasing college students retention than programs that focus on a single social context.

PAGE 67

67 Figure 1 -1 Proposed Bioecological Model of College Student Retention

PAGE 68

68 Table 1 -1 Proposed relationships in the model Commitment Making Identification with Commitment Mastery Goals PerformanceAvoid Goals Perceived Academic Self Efficacy GPA Social Adjustment Intention to Persist Parental Invol vement Parental Relatedness Parental Autonomy Support * Off Campus Friend Relatedness Instructor Relatedness On Campus Friend Relatedness Sense of Belonging * Ex tracurricular Activities Conscientiousness Entity Beliefs * Commitment Making Identification with Commitment Mastery Goals Performance Avoid Goals Perceived Ac ademic Self Efficacy * GPA Social Adjustment

PAGE 69

69 CHAPTER 2 METHOD The purpose of this study was to test a new model of college retention that focused on the relationships among the family, peer, and college microsystem s, the self system, the psychological processes of identity, goal orientations, and perceived academic selfefficacy, and the outcome variables, academic achievement, social adjustment, and intention to persist. Participants Participants were 299 college students from a large southeastern university. At the university, students were recruited from educational psychology classes and received course credit for their participation. All participants complete d the survey online. Instruments Predictor Variables Identity Identity was measured using the Commitment Making and Identification with Commitment subscales of the Dimensions of Identity Development Scale (DIDS; Luyckx et al., 2008). This 25item instrument measures the five identity dimensions, Commitmen t Making, Identification with Commitment, Exploration in Breadth, Exploration in Depth, and Ruminative Expl oration. Participants responded the items using a 5 point Likert scale ranging from 0 ( strongly disagree ) to 4 ( strongly agree ). Sample items for Com mitment Making include Know what I want to do with my future, and Made a choice concerning some of my plans for the future. Sample items for Identification with Commitment include Future plans give me self confidence, and Sense that the direction I want to take in life will really suit me. In the Luyckx et al. (2008) study a sample of 263 Caucasian freshmen college students from a university in Belgium completed the 25item measure. This sample was 72.6% female, and the

PAGE 70

70 mean age was 19.14 years. Cronbachs alphas were .86 for Commitment Making, .86 for Identification with Commitment, .81 for Exploration in Breadth, .79 for Exploration in Depth, and .86 for Ruminative Exploration. Luyckx et al. (2008) conducted a confirmatory factor analysis to determ ine if the data would fit a five-factor model consistent with the proposed dimensions. The data did fit the fivefactor model with a Satorra Bentler scaled chi -square of 658.93, RMSEA = .07, and CFI = .94. Luyckx et al. (2008) conducted a cluster analysis of the identity statuses. Six clusters emerged: Achievement, Diffused Diffusion, Carefree Diffusion, Ruminative Moratorium, Foreclosure, and Undifferentiated. The Achievement cluster scored high on Commitment Making and Identification with Commitment, mod erately high on Exploration in Breadth and Exploration in Depth, and low on Ruminative Exploration. The Diffused Diffusion cluster scored low on Commitment Making and Identification with Commitment, intermediate on Exploration in Breadth and Exploration in Depth, and high on Ruminative Exploration. The Carefree Diffusion cluster scored low on Commitment Making, Identification with Commitment, Exploration in Breadth, and Exploration in Depth, and intermediate on Ruminative Exploration. The Ruminative Explora tion cluster scored low on Commitment Making and Identification with Commitment, and high on Exploration in Breadth, Exploration in Depth, and Ruminative Exploration. The Foreclosure cluster scored high on Commitment Making and Identification with Commitme nt, and low on Exploration in Breadth, Exploration in Depth, and Ruminative Exploration. The Undifferentiated cluster scored intermediate on all dimensions except Exploration in Breadth, which was moderately high. As evidence of construct validity, parti cipants scores on the DIDS clusters were related to their scores on measures of self esteem, depressive symptoms, anxiety symptoms, self -

PAGE 71

71 reflection, and self-rumination. Selfesteem was assessed using the Rosenberg Self Esteem Scale (RSES; Rosenberg, 1965 ). Depressive symptoms were assessed using 12 items of the Center for Epidemiologic Studies Depression Scale (CES-D; Radloff, 1977). Anxiety symptoms were assessed using the revised Symptom Checklist (SCL-90-R; Arrindell & Ellema, 1986). Selfreflection and self -rumination were assessed using the Rumination-Reflection Questionnaire (RRQ; Trapnell & Campbell, 1999). For selfesteem, the Achievement and Foreclosure clusters scored the highest and the Diffused Diffusion and Ruminative Moratorium clusters scored the lowest. For depression and anxiety symptoms, the Achievement, Foreclosure, and Carefree Diffusion clusters scored the lowest and the Diffused Diffusion and Ruminative Moratorium clusters scored the highest. For self-reflection, the Achievement and Ruminative Moratorium clusters scored the highest and the Carefree Diffusion cluster scored the lowest. For self rumination, the Ruminative Moratorium and Diffused Diffusion clusters scored the highest and the Achievement, Carefree Diffusion, and Foreclosure clusters scored the lowest. These results are consistent with the expected relationships between the identity clusters and these measures. Particularly relevant for this study is that the Achievement and Foreclosure clusters, which are both high in Commitment Making and Identification with Commitment, scored high on selfesteem and low on anxiety and depression. In my university sample I obtained Cronbachs alphas of .86 for Commitment Making and .89 for Identification with Commitment. Goal orientation Go al orientation was measured using two of the three Personal Achievement Goal Orientations subscales from the Patterns of Adaptive Learning Scales (PALS; Midgley et al., 2000). These t wo subscales are Mastery Goal Orientation and Performance-Avoid Goal Orie ntation. Response options are measured on a 5-point Likert scale that ranged from 0 ( not at all true ) to 4 ( very true ). The Mastery Goal Orientation subscale contains six items including I

PAGE 72

72 like class work that I'll learn from even if I make a lot of mista kes and an important reason why I do my class work is because I like to learn new things. The Performance-Avoid Goal Orientation subscale consists of six items including it's very important to me that I don't look stupid in my class and an important reason I do my class work is so that I dont embarrass myself. Ross, Shannon, Salisbury-Glennon, and Guarino (2002) examined the reliability and validity of the scores of a college sample on the PALS. Participants were 184 undergraduate students in educa tion classes at a southeastern university. The sample was 74% female with a mean age of 22. Cronbachs alphas for students scores on Mastery Goal Orientation, and Performance -Avoid Goal Orientation were .78 and .84. A confirmatory factor analysis revealed that the items did load on a three-factor model as hypothesized. For my university sample I obtained Cronbachs alphas of .85 for Mastery Goal Orientation and .89 for Performance -Avoid Goal Orientation. Perceived academic self efficacy Eight questions from the Self Efficacy for Learning and Performance subscale of the Motivated Strategies for Learning Questionnaire (MSLQ; Pintrich et al., 1991) were used to measure perceived academic self efficacy in a specific classroom. For this study, t his subscale was modified to reflect perceived academic self efficacy for all college courses. This subscale measures expectancy for success and perceived self efficacy. Responses options for the subscale range from 0 ( not at all true of me ) to 7 ( very true of me). Sa mple items include I believe I will receive excellent grades in my courses and Im certain I can mas ter the skills being taught in my classes. With a sample of 380 Midwestern college and community college students, Pintrich et al. obtained a C ronbach alpha of .93. With regard to predictive validity, the students scores on the scale were significantly correlated with final course grades ( r = .41). Pintrich et al. also

PAGE 73

73 conducted a confirmatory factor analysis and found that all eight items loaded on a factor they named Self Efficacy for Learning and Performance Recently, Cordell-McNulty and Ashton (2006) obtained a Cronbach alpha of .92 for the scores of 364 students on this scale at a college in the Northeast, and Cordell-McNulty and Ashton (2008) obtained a Cronbach alpha of .94 for the scores of 172 students at a community college in the Southeast. In this study I obtained a Cronbachs alpha of .93 for the university sample. Parental involvement Parental involvement was measured by the Academic Support s ubscale of the Parental Support for College measure (Mounts, 2004). This 40item scale measures three types of parental support: academic, financial, and social support. Items response options on a 4-point Likert scale range from 0 ( strongly disagree ) to 3 ( strongly agree ). The Academic Support subscale consists of 12 items. Sample ite ms for academic support include My parents/guardian asked me about my class work (homework, exams, grades, and classes), and My parents/guardian gave encouragement when classes were tough. Mounts reported internal consistency data for two diverse samples of college freshmen. The first sample consisted of 183 students, 52% male, 56% White, 23% African-American, 13% Asian, 6% Latino, and 2% Multiracial. T he second sample was composed of 400 students, 62% female, 58% White, 22% African -American, 7% Latino, 6% Asian, 5% Multiracial, and 2% other For students scores on the Academic Support subscale, a Cronbach alpha of .84 was obtained in the first sample, and a Cronbach alpha of .79 was obtained in the second sample. As evidence of convergent validity for the Academic Support subscale, Mounts (2008) reported significant positive correlations of .47 between academic support and maternal acceptance and of .27 between academic s upport and parental acceptance. Parental acceptance was measured using the Childs Report of Parenting Behavior Inventory (CRPBI). As evidence

PAGE 74

74 of discriminant validity for the Academic Support subscale, Mounts reported negative correlations of -.18 with the Beck Depression Inventory (BDI), and -.08 with the Beck Anxiety Inventory (BAI). These results provide evidence of the convergent and discriminant validity of the participants scores on the academic support subscale because it was positively correlated with parental and maternal acceptance and negatively correlated with depression and anxiety. For the university sample in this study I obtained a Cronbach alpha of .83. Parental autonomy support Parental autonomy support was measured by the Parental Auto nomy Support Scale (Duchesne et al., 2007; Ratelle et al., 2005). This measure is designed to assess college students perceptions of their parents supporting their autonomy in their decision to pursue college studies. This 8 item scale was adapted from Paulson et al. (1994) and Robinson et al. (1995). Item response options are scored on a 5point Likert scale ranging from 0 ( totally disagree ) to 4 ( totally agree ). Sample items include my parents gave me a lot of freedom with respect to my choice of college program, and my parents let me express myself during the decision process of choosing a college program. For the scores of 262 Canadian college students, Rat elle et al. obtained a Cronbach alpha of .88. For the scores of 498 Canadian college students (279 females, and 219 males) Duchesne et al. obtained a Cronbach alpha of .87. I obtained a Cronbach alpha of .89 for the university sample in this study Relatedness Students sense of relatedness to their parents, off-campus friends, instructors, and oncampus friends was measured by 16 items from a scale developed by Furrer and Skinner (2003). These items measured relatedness to four social partners: parents, instructors, classmates, and friends. The stem for each item is When Im with . The four items for each stem are I feel accepted, I feel like someone special, I feel ignored, and I feel unimportant. Response

PAGE 75

75 options ranged from 0 ( strongly disagree ) to 6 ( strongly agree ). Cordell -McNulty and Ashton (2008) used this measure in a sample of 146 community college students and obtained a Cronbach alpha of .82 for participants scores Cronbach alphas for the scores on the subscales were .88 for relatedness to parent, .82 for relatedness to instructors, .81 for relatedness to friends, and .84 for relatedness to classmates. For the university sample in this study I obtained Cronbach alphas of .88 for relatedness to parent .86 for relatedness to instructor, .88 for relatedness to oncampus friend s and .87 for relatedness to offcampus frien ds. Sense of belonging on campus Students sense of belonging on campus was measured by the Sense of Belonging subscale of the Perceived Cohesion Scale (PCS; Bollen & Hoyle, 1990). The three items of this measure are I see myself as a part of the campus community, I feel a sense of belonging to the campus community, and I feel that I am a member of the campus community. Response options on t his subscale range on an 11 point Likert scale from 0 ( strongly disagree ) to 10 ( strongly agree ). Bollen and Hoyle conducted a factor analysis of the responses of 102 undergraduates at a private liberal arts college and found that all three items loaded on one latent variable they identified as S ense of Belonging. In a study of 272 Latino undergraduates, Hurtado and Carter (1997) obtained a Cronbach alpha of .94 for the participants scores on the 11 items, and a factor analysis indicated that all three items loaded on the Sense of B elonging factor Mounts (2004) administered the sense of belonging subscale to 319 freshmen students and obtained a Cronbach alpha of .89 for African American students and a Cronbach alpha of .91 for White students. I obtained a Cronbach alpha of .96 for the university sample in this study Theory of intelligence Participants theory of intelligence was measured using the Theory of Intelligence Scale (Dweck, Chiu, & Hong, 1995). This 3item instrument measures implicit beliefs about

PAGE 76

76 intelligence with response options on a 6-point Likert scale ranging from 0 ( strongly agree ) to 5 ( strongly disagree ). The items are you have a certain amount of intelligence, and you cant really do much to change it, your intelligence is something about you that you cant change very much, and you can learn new things, but you cant really change your ba sic intelligence. Participants scores are averaged with a higher score indicating an incremental belief. Dweck et al. (1995) reported internal reliabilities ranging from .94 to .98 across six studies for the T heory of I ntelligence S cale. They also reported a 2 week test retest reliability coefficient of .80. Factor analyses were conducted for five different data sets, and the three items loaded consistently on one factor. For the university sample in this study I obtained a Cronbach alpha of .93. Personal ity Personality was measured by the Conscientiousness subscale of the Big Five Inventory (BFI; John et al., 1991). The BFI is a 54item instrument that measures the five personality dimensions, extraversion, openness, conscientiousness, agreeableness, and neuroticism. Items are rated on a 5 point Likert scale from 0 ( disagree strongly ) to 4 ( agree strongly ). There are nine items for the Conscientiousness subscale. Sample items for Conscientiousness include does a thorough job and perseveres until the task is finished. John and Srivastava (1999) surveyed a sample of 462 undergraduates at the University of California, Berkeley, and obtained a Cronbachs alpha of .82 for students scores on the Conscientiousness subscale. Providing support for convergent validity for the Conscientiousness subscale, John and Srivastava (1999) compared the BFI and two other personality measures, the Trait Descriptive Adjectives (TDA; Goldberg, 1992) and the NEO Five Factor Inventory (NEOFFI; Costa & McCrae, 1992). The uncorr ected pairwise convergent validity estimate for students scores on the Conscientiousness subscale of the BFI and the Conscientiousness subscale of the TDA was .81, and for scores on the Conscientiousness subscale of the BFI and the Conscientiousness

PAGE 77

77 subsc ale of the NEO -FFI, .79. The corrected pairwise convergent validity estimate for scores on the Conscientiousness subscale of the BFI and the Conscientiousness subscale of the TDA was .94, and for the Conscientiousness subscale of the BFI and the Conscientiousness subscale of the NEO-FFI, .96. John and Srivastava (1999) also conducted a confirmatory factor analysis to test the convergent and discriminant of the BFI, the TDA, and the NEOFFI. The CFA estimated the latent factors of the Big Five personality characteristics. The standardized validity coefficient obtained from the CFA for the Conscientiousness subscale of the BFI was .92. I obtained a Cronbach alpha of .82 for the university sample in this study Socioeconomic s tatus Socioeconomic status was determined using the Four Factor Index of Social Status (FFISS; Hollingshead, 1975). The FFISS is a multidimensional measure that takes into consideration the persons education, occupation, gender, and marital status to determine socioeconomic status. The FFISS consists of the sum of the two weighted factor scores, the education factor and the occupation factor. Education scores range from 1 ( less than seven th grade completed) to 7 ( graduate professional training completed). Occupation scores range from 1 ( e.g ., farm laborers, lowlevel service workers ) to 9 (e.g., higher executives, proprietors of large businesses, major professionals ). The FFISS is calculated by multiplying the occupation scale score by 5 and the education scale score by 3 and adding these scores together. If a person is married then the scores for each spouse are averaged together (Hollingshead, 1975). As evidence of validity, Ho llingshead (1975) reported a correlation of .93 between the occupation group scores of the FFISS and the National Opinion Research Centers occupational prestige scores. Hollingshead showed that the number of school years completed was positively correlated with occupation score on the FFISS for males ( r = .84, p < .001) and females ( r = .85, p < .001) using data from the 1970 census. In addition median earned income was positively

PAGE 78

78 correlated with occupation scores on the FFISS for males ( r = .78, p < .001) and females ( r = .67, p < .001) using data from the 1970 census (Hollingshead, 1975). Gottfried (1985) compared scores on the FFISS to the Revised Duncan (TSEI2) Socioeconomic Index and the Seigel 1965 NORC Prestige Scale using data from parents of 130 infants recruited from birth announcements. Gottfried reported a correlation of .79 between the FFISS and the Revised Duncan Socioeconomic Index and a correlation of .73 ( p < .001) between the FFISS and the Seigel Prestige Scale. Demographic information For this study, students gender, ethnicity, age, and extracurricular activities were obtained from their self -reports. Students were also asked to report SAT or ACT scores, and high school GPA and to give permission for validation of those scores from the registrar. Outcome Variables Social adjustment Social adjustment was measured by the social adjustment subscale of the Student Adaptation to College Questionnaire (SACQ; Baker & Siryk, 1989 ). The SACQ is designed to assess student adjustment to college. The social adjustment subscale measures students success in coping with the interpersonalsocietal demands in the college experience. This 20item subscale is measured on a 9 -point Likert scale ranging from 0 ( applies very closely to me) to 8 ( doesnt apply to me at all ). The social adjustment subscale can be divided into four clusters: general, other people, nostalgia and social environment. The general cluster measures the extent and success of social functioning in general. The other people cluster measures involvement and relationships with people on campus. The nostalgia cluster measures dealing with being away from home and the social environment cluster measures satisfaction with the social environment (Baker & Siryk, 1989).

PAGE 79

79 The SACQ Manual reports 14 studies at two colleges where the SACQ was administered to first and secondsemester freshmen. Alpha coefficien ts for students scores on the social adjustment subscale ranged from .83 to .91. The manual also includes 11 other university data sets. The alpha coefficients for students scores on the social adjustment subscale range from .73 to .91. For criterionrel ated validity, a social activities checklist that included items about campus involvement, attendance of campus events, and use of recreational facilities was used. There was a significant relationship between the social activities checklist and the social adjustment subscale, but not any of the other subscales (Baker & Siryk, 1989). Wick and Shilkret (1986) reported a significant relationship between the amount of extracurricular activity and the social adjustment subscale ( r = .47, p < .01). Savino, Reuter-Krohn, and Costar (1986) found significant negative correlations between the number of visits the student made home and the social adjustment subscale. The significant correlations between social adjustment and participation in social and extracurricular activities and the number of visits the student made home provide evidence o f the criterion related validity of participants scores on the SACQ. With regard to construct validity, Caro (1985) reported significant negative correlations between the social adjustment subscale and emotional reliance on other persons ( r = -.23, p < .01), lack of social selfconfidence ( r = -.43, p < .01), Social Avoidance and Distress Scale ( r = .52, p < .01), and the Revised UCLA Loneliness Scale ( r = -.66, p < .01). Harri s (1988) also reported a significant negative correlation between the Revised UCLA Loneliness Scale and the social adjustment subscale ( r = -.79, p < .01). Saracoglu (1987) reported significant positive correlations between the social adjustment subscale and the SelfEsteem Inventory ( r = .63, p < .01), the SelfEfficacy Scale ( r = .37, p < .05), and social selfefficacy (r = .58, p < .01).

PAGE 80

80 Flescher (1986) found a significant positive correlation between the social adjustment subscale and the full Mental H ealth Inventory ( r = .45, p < .01), the positive affect subscale ( r = .40, p < .01), and the psychological wellbeing subscale ( r = .30, p < .05). There were also significant negative correlations between social adjustment and the anxiety subscale ( r = -.38, p < .05), the depression subscale ( r = -.36, p < .05), and the psychological distress subscale ( r = .50, p < .01) The significant positive correlations between positive mental health constructs and negative correlations between negative mental health constructs provide evidence of the construct validity of the participants scores on the social adjustment subscale. In another example of the construct validity of the social adjustment subscale, Caro (1985) reported a significant positive correlation bet ween social adjustment and perceived social support from friends ( r = .36, p < .01). Hogan (1986) also reported a significant positive correlation ( r = .31, p < .01 ). Savino et al. (1986) found significant positive correlations between social adjustment an d social support (total scale) during S emester 1 ( r = .30, p < .01) and S emester 2 ( r = .40, p < .01). The significant correlations between social support measures and social adjustment support the construct validity of participants scores on the social adjustment subscale. Baker and Siryk (1989) also reported correlations between the SACQ subscales and attrition after 1 year at Clark University. For six of the eight data sets there was a significant negative correlation between the social adjustment subsc ale and first -year attrition, suggesting the importance of social adjustment to college retention. Overall the psychometric data on the social adjustment subscale of the SACQ support its use in the study of college retention. For the university sample in t his study I o btained a Cronbach alpha of .90.

PAGE 81

81 Academic achievement Students were asked to report their overall GPA and were also asked for their permission to obtain their GPA from the registrar Only 120 students gave permission to have their GPAs obtained from the registrar. Because this number was insufficient to test the proposed model, students selfreports of their GPA were used as the measure of academic achievement. In a meta analysis of self reported grades, class ranks, and test scores, From a m eta-analysis of 12 studies of the validity of self-reported GPA in 12,089 college students, Kuncel, Cred, and Thomas (2005) found that the accuracy of self-reported GPAs was sufficiently high ( r = .90) to warrant use in educational research. Intention to persist Intention to persist w as measured using a single item adapted from Hausmann et al. (2007), How confident are you that you will complete your intended degree at your current institution? Students responded on a 7point Likert scale ranging from 1 ( not at all confident ) to 7 ( very confident ). Cabrera, Nora, and Castaneda (1993) tested a structural equation model of college retention with 2,459 freshmen at a large southern university where persistence was measured by students reenrolling the next fall and found that intention to persist had the strongest direct effect on actual persistence (.4 9 ). Analyses I use d structural equation modeling to estimate the relationships in my model. The model contained eight endogenous variables perceived academic self efficacy, mastery goals, performance avoid goals, commitment making, identification with commitment, academic achievement, social adjustment, and intention to persist The model contain ed 10 exogenous variables, parental involvement, parent relatedness, parental autonomy support, off-campus

PAGE 82

82 friend relatedness, instructor relatedness, on-campus friend relatedness, sense of belonging, extracurricular activities, conscientiousness, and entity beliefs

PAGE 83

83 CHAPTER 3 RESULTS The purpose of this study was to test a new model of college retention that focused on the relationships among the family, peer, and college microsystems, the selfsystem, the psychological processes of identity, goal orientations, and perceived academic selfefficacy, and the outcome va riables, academic achievement, social adjustment, and intention to persist. In this chapter the demographic characteristics of the sample, descriptive statistics, analysis of the hypothesized model, revisions to the model, and the results of testing the research hypotheses are described. Descriptive Statistics The university sample consisted of 299 participants. The sample was predominantly female (81%) and Caucasian (66%). Demographic data are presented in Table 3 1. The age range for this sample was 18-45. For data analysis I only included participants up to 26 years of age because 26 is considered the last year of emerging adulthood according to Arnett (2000, 2006). The mean SES score on the FFIS S was 31.61, indicating a predominant ly middle class sample. Means and standard deviations for all the variables are reported in Table 3-2. The mean self reported college GPA was 3.33. The correlation matrix is presented in Table 3 -3. Analysis of the Proposed M odel The hypothesized model was estimated using MPlus. Because the sample was predominant ly female, the data were weighted by gender according to the national college enrollment by gender in 2005. In the fall of 2005, 43% of all enrolled students were male and 57% were female ( National Center for Educa tion St atistics, 2008). Gender was weighted .43 for mal es and .57 for females. To control for causes that were left out of the model, high school GPA, gender, and ethnicity were included as predictors for every endogenous variable. SAT

PAGE 84

84 scores were not used as a control variable in the test of the original or the revise d model because scores ranged from 830 to 2140, indicating that some students were reporting more than the total of their verbal and quantitative scores. Consequently, it was impossible to determine which scores students were reporting. Instead of the SAT, high-school GPA was used as the control for prior ability. Preliminary analyses of the SES data and student comments about the measure suggested that students may not have reported accurate information on which to base their SES, so SES was dropped from further analyses. Because there were too few participants to include each ethnicity separately in the analyses, students were split into two groups as follows: (a) Caucasian and Asian students, and (b) African American, Hispanic, Native American, and students who classified themselves as other. The goodness of fit test result was 2 (36) = 159.22, p = .00, indicating that the model did not fit the data. Other goodness of fit indices included the comparative fit index (CFI) = .38, the Tucker Lewis index (TLI) = -.08, and the root mean square error of approximation (RMSEA) = .21. All of these indices indicated poor fit. An analysis of the modification indices showed that some of the relationships could be improved. The largest modification index was between C ommitment Making and I dentification with Commitment (68.04). I allowed these errors to correlate. I ran the analys is again and the model still had poor fit, 2 (34 ) = 91.73, p = .00, CFI = .71, TLI = .46, RMSEA = .09. The largest modification index was 34.27 between perceived self efficacy and I dentification with C ommitment. I add ed a path from I dentification with C ommitment to perceived self efficacy but th e model still had poor fit, 2 (34 ) = 58.68, p = .00 5, CFI = .88, TLI = 77, RMSEA = .0 6. The largest modification index was 11.05 between perceived self -efficacy and conscientiousness so I added a path from conscientiousness to perceived self efficac y. T his time the model fit was better but still not very good. The goodness of fit test result was 2 (33 ) = 47.34, p = .0 5, CFI =

PAGE 85

85 .92, TLI = .86, RMSEA = .04. The largest modification index was 4.93 between perceived self efficacy and performance -avoid goals, so I added a path fr om self efficacy to performance -avoid goals. The model was run again and the goodness of fit indices were 2Research Hypotheses (33 ) = 44.07, p = .0 9, CFI = .95, TLI = .89, RMSEA = .04, indicating a good f it. See Table 3 4 for a list of the total, direct, and indirect effects in the revised model. The following research hypotheses were tested, with high-school GPA, gender, and ethnicity controlled. These variables are not mentioned in the statement of each hypothesis. However, when the effects of those variables were significant, they are included in the description of the effects for the hypothesis. Hypothesis 1 was, academic achievement, social adjustment, perceived academic self efficacy, commitment making, identification with co mmitment, parental autonomy support, instructor relatedness, on-campus friend relatedness, sense of belonging, extracurricular activities, and e ntity beliefs predict intention to persist. The only significant direct path to intention to persist was from pe rceived self efficacy ( .29, p = .000). However, conscientiousness has an indirect relationship to intention to persist through perceived self= .08, p = .001) and through identification with commitment and perceived self= .02, p = .005). A significant indirect effect to sense of belonging was mediated through identification with commitment and self efficacy ( = .01, p = .01), and a significant indirect effect of identification with commitment on intentions to persist was mediated through self efficac y ( .05, p = .000). Hypothesis 2 was perceived academic self efficacy, commitment making, identification with commitment, mastery goals, performance -avoid goals, parental involvement, parental autonomy support, instructor relatedness, extracurricular ac tivities, e ntity beliefs and

PAGE 86

86 conscientiousness predict academic achievement. Results indicated significant direct paths to academic achievement from entity beliefs ( = .14, p = .04), performance -avoid goals ( -.20, p = .0005), self efficacy ( .17, p = .005), and ethnicity group ( = .37, p = .002). In addition significant indirect effects of entity beliefs on achievement were mediated through performanceavoid goals ( = -.04, p = .02) and from selfefficacy that were mediated through performanceav oid goals ( .03, p = .02). F our significant indirect effects on academic achievement were also found: from conscientiousness mediated through perceived self efficacy ( = .05, p = .02), from conscientiousness mediated through identification with commitment and perceived self efficacy ( = .009, p = .02), from conscientiousness mediated through perceived self efficacy and performance-avoid goals ( = .008, p = .04), and from conscientiousness mediated through identification with commitment, perceived self efficacy, and performance-avoid goals ( = .002, p = .04). In addition there were two significant indirect effects on academic achievement from identification with commitment. One was mediated through perceived self efficacy ( .03, p = .008) and the ot her was mediated through perceived self efficacy and performance avoid goals ( .005, p = .02). Hypothesis 3 was perceived academic self efficacy, commitment making, identification with commitment, mastery goals, performance-avoid goals, parental autonomy support, off campus friend relatedness, instructor relatedness, oncampus friend relatedness, extracurricular activities, and sense of belonging predict social adjustment. There were several significant direct paths to social adjustment including from parental autonomy support ( = .23 p = .005), oncampus friend relatedness ( = .14, p = .01), sense of belonging ( = .34, p = .001), extracurricular activities ( = .21, p = .001 ), performance -avoid goals ( -.17, p = .0005), and self efficacy ( .13, p = .003). There were al so significant paths from gender ( = .11, p = .03)

PAGE 87

87 and high school GPA ( = .14, p = .02). Academic self efficacy also had a significant indirect effect on social adjustment that was mediated through performance-avoid goals ( .02, p = .02). In addition, there were two significant indirect effect of identification with commitment. One was mediated through perceived self efficacy ( .02, p = .006) and the other was mediated through perceived self efficacy and performance-avoid goals ( .004, p = .02). Hypothesis 4 was perceived academic self efficacy, parent relatedness sense of belonging, and e ntity beliefs predict mastery goals. There were significant paths from entity beliefs ( = -.17, p = .006 ) and perceived self efficacy ( .15, p = .0005). T here was also a significant path from gender ( = .16, p = .004). In addition there was an indirect effect of sense of belonging that was mediated through identification with commitment and perceived academic self efficacy ( = .006, p = .013). Hypothesis 5 was perceived academic self efficacy sense of belonging, and e ntity beliefs predict performance-avoid goals. There were significant paths from entity beliefs ( = .18, p = .006) and perceived self efficacy ( -.14, p = .007). There was also a significant direct effect from gender ( = -.13, p = .03). In addition, there was a significant indirect effect from sense of belonging that was mediated through identification with commitment and perceived academic self efficacy ( = -.006, p = .027). Hypothesis 6 was p arental involvement, instructor relatedness, and sense of belonging predict perceived academic self efficacy. There were significant paths from instructor relatedness ( = .14, p = .0 5), conscientiousness ( = .29, p = .000), and identification with commitment ( .19, p = .000). There was also a significant path from ethnicity group ( = .11, p = .05). In addition a significant indirect effect of sense of belonging was found that was mediated through identification with commitment ( = .04, p = .004). There was also an indirect

PAGE 88

88 effect of conscientiousness that was mediated through identification with commitment ( = .06, p = .0005). Hypothesis 7 was parental autonomy support, parent relatedness, sense of belonging, and conscientiousness predict commitment making. Significant paths were from parental autonomy support ( = .15, p = .02), sense of belonging ( = .14, p = .04) and conscientiousness ( = .29, p = .000). Hypothesi s 8 was parental autonomy support, sense of belonging, and conscientiousne ss predict identification with commitment. Significant paths were from sense of belonging ( = .23, p = .001) and conscientiousness ( = .31, p = .000).

PAGE 89

89 Table 3 -1 Demographic data Measure % N Gender Male 19% 58 Female 81% 240 Ethnicity Caucasian 66% 196 African American 16% 49 Hispanic 9% 27 Asian 3% 9 Native American 0.3% 1 Other 5% 16 Class Freshmen 19% 58 Sophomore 17% 52 Junior 31% 93 Senior 28% 84 Other 4% 12

PAGE 90

90 T able 3 -2 Means and standard deviations of variables Variable N M SD Min Max Commitment making 298 2.93 0.74 0.00 4.00 Identification with commitment 297 2.86 0.77 0.00 4.00 Academic self efficacy 296 4.66 0.91 0.25 6.00 Mastery goals 297 2.52 0.71 0.33 4.00 Perform ance Avoid goals 293 1.80 0.94 0.00 4.00 Parental involvement 289 1.59 0.56 0.00 2.75 Parental relatedness 296 5.34 0.98 1.00 6.00 Instructor relatedness 299 3.58 1.24 0.25 6.00 On Campus friend relatedness 298 4.63 1 .14 0.75 6.00 Off-Campus friend relatedness 297 5.18 1.02 0.50 6.00 Parental autonomy support 294 3.40 0.74 0.00 4.00 Sense of belonging 297 1.96 0.74 0.00 3.00 Conscientiousness 295 25.29 5.94 10.00 36.00 Entity beliefs 296 2.13 1.29 0.00 5.00 Extracurricular activities 285 2.10 1.83 0.00 10.00 Social adjustment 298 5.60 1.34 1.06 8.00 College GPA 251 3.33 0.45 1.58 4.00 Intention to persist 297 5.51 0.94 0.00 6.00 High school GPA 276 3.99 0.64 1.65 6.70 SAT total score 192 1223.86 174.26 830.00 2140.00 ACT total score 144 25.66 3.89 14.00 38.00 SES 249 31.61 7.70 11.00 55.00 Age 298 20.87 2.97 18.00 45.00

PAGE 91

91 Table 3 -3 Correlation m atrix CM IC MG P AG SE PI PR PA OFR IR OCR SB CM 1.00 IC ****.79 1.00 MG ***.24 ***.26 1.00 P AG .12 .09 .09 1.00 SE *.17 ***.23 ****.33 .17 1.00 PI -.02 .04 .05 .10 -.06 1.00 PR .03 .09 .09 -.02 -.01 ***.44 1.00 PA **.21 *.17 *.18 .18 .04 .07 ***.26 1.00 OFR **.21 ****.29 **.21 .03 .08 *.18 ***.27 .12 1.00 IR -.01 .07 **.21 .04 **.23 **.24 .13 .00 **.22 1.00 OCR .01 .14 .13 -.07 .12 ***.24 ****.30 .05 ****.44 ***.24 1.00 SB .13 ** .22 .13 .03 .06 *.15 **.21 .12 .09 .13 ****.29 1.00 EX .04 .06 .01 .00 .01 .04 .06 .06 .05 .08 .04 ****.32 CO **.23 ****.31 ***.26 -.20 ****.29 .07 *.18 **.20 **.20 **.20 **.19 **.23 EN -.12 -.03 **-.20 **.21 *-.14 -.00 -.00 -.11 -.11 -.00 .04 .05 HSGPA .03 .06 .15 .11 .02 .10 .01 .04 .04 .07 .08 .02 Gender .05 .03 **.21 .11 .08 .05 .05 .03 .00 *.15 .03 .12 Group -.01 -.05 -.01 .06 .08 .06 .03 .12 **.22 *.16 .09 -.10 SA *.14 *.17 *.15 **-.23 *.18 .12 ***.25 ****.29 **.22 .12 ****.35 ****.53 GPA .10 .14 .04 .17 **.21 .02 .06 .07 .10 .07 .01 .11 INPS .04 .01 .13 .01 ****.27 .08 *.16 .04 .05 .03 .03 .00 Note CM = commitment making; IC = identification with commitment; MG = mastery goals; PAG = performanceavoid goals; SE = academic self -efficacy; PI = parental involvement; PR = parent relatedness; PA = parental autonomy support; OFR = off-campus friend relatedness; IR = instructor relatedness; OCR = on campus friend relatedness; SB = sense of belonging; EX = extracurricular activities; CO = conscientiousness; EN = entity beliefs; HSGPA = high school GPA; Group = ethnicity group; SA = social adjustment; INPS = intention to persist. *p < .05. **p < .01. ***p < .001. ****p < .0001.

PAGE 92

92 Table 3 -3 Continued EX CO EN HSGPA Gender Grou p SA GPA INPS EX 1.00 CO *.15 1.00 EN .01 .13 1.00 HSGPA .09 .05 *.14 1.00 Gender .14 *** .24 .04 .16 1.00 Group .10 .06 .06 .00 .01 1.00 SA ****.33 ***.27 .05 .10 .03 .02 1.00 GPA *.15 *.14 .11 09 .06 ****.31 *.18 1.00 INPS .04 .07 .08 .01 .05 .11 *.14 *.14 1.00 Note EX = extracurricular activities; CO = conscientiousness; EN = entity beliefs; HSGPA = high school GPA; Group = ethnicity group; SA = social adjustment; INPS = intention to pe rsist. *p < .05. **p < .01. ***p < .001. ****p < .0001.

PAGE 93

93 Table 3 -4 Total, direct, and indirect effects in the revised model Variable Effect CM IC MG P AG SE SA GPA INPS Group Total .02 .08 .10 .01 .10 .04 ****.32 .09 Direct .02 .08 .11 .02 *.11 .07 ***.37 .09 Indirect --.01 .01 .01 .03 .05 .00 Gender Total .07 .07 **.17 .14 .04 **.13 .03 .02 Direct .07 .07 **.16 .13 .03 *.11 .01 .01 Indirect --.01 .01 .01 .02 .02 .03 HSGPA Total .05 .04 .02 .00 .04 **.13 .08 .03 Direct .05 .04 .03 .01 .03 *.14 .09 .01 Indirect --.01 .01 .01 .01 .01 .02 PI Total --.01 .01 .04 .01 .05 .02 Direct ----.04 -.04 -Indirect --.01 .01 -.01 .01 .02 PR Total .09 -.02 --.07 .10 .13 Direct .09 -.02 -----Indirect -----.07 .10 .13 PA Total *.15 .09 --.02 **.17 .04 .09 Direct *.15 .09 ---**.23 .03 .04 Indirect ----.02 .06 .07 .05 OFR Total ------.01 Direct -----.10 --Indirect ------.01 IR Total --.02 .02 *.14 .02 .02 .05 Direct ----*.14 .00 .01 -Indirect --.02 .02 -.02 .03 .05 OCR Total -----**.14 -.15 Direct -----**.14 -.17 In direct -------.02 SB Total *.14 ***.23 .07 .01 .01 ****.39 .10 .03 Direct *.14 ***.23 .07 .01 .05 ***.34 -.19 Indirect --.00 .00 ***.04 .05 .10 .16 Note CM = commitment making; IC = identification with commitment; MG = mastery goals; PAG = performanceavoid goals; SE = academic self efficacy; PI = parental involvement; PR = parent relatedness; PA = parental autonomy support; OFR = offcampus friend relatedness; IR = instructor relatedness; OCR = on campus friend relatedness; SB = sense of belonging; EX = extracurricular activities; CO = conscientiousness; EN = entity beliefs; HSGPA = high school GPA; Group = ethnicity group; SA = social adjustment; INPS = intention to persist. *p < .05. **p < .01. ***p < .001. ****p < .0001.

PAGE 94

94 Table 3 -4 Continued Variable Effect CM IC MG P AG SE SA GPA INPS EX Total -----***.21 .13 .01 Direct -----***.21 .13 .03 Indirect -------.04 CO Total ****.29 ****.31 **.05 .05 ****.35 .05 *.15 *.14 Direc t ****.29 ****.31 --****.29 -.05 -Indirect --**.05 .05 ***.06 .05 .10 *.14 EN Total --** .17 **.18 -.02 .12 .00 Direct --** .17 **.18 --*.14 .01 Indirect -----.02 .02 .01 CM Total -----.82 1.1 1.16 Direct -----.82 1.1 1.16 Indirect --------IC Total --**.03 ** .03 ****.19 .80 1.18 1.45 Direct ----****.19 .77 1.15 1.16 Indirect --**.03 ** .03 -**.03 **.03 .29 MG Total -----.05 .08 -Direct -----.05 .08 -Indirect --------AG Total -----*** .17 *** .20 -Direct -----*** .17 *** .20 -Indirect --------SE Total --***.15 ** .14 -***.15 **.19 ****. 33 Direct --***.15 ** .14 -**.13 **.17 ****.29 Indirect -----.02 .02 .04 SA Total -------.14 Direct -------.14 Indirect --------GPA Total -------.10 Direct -------.10 Indirect --------Note CM = commitment making; IC = identification with commitment; MG = mastery goals; AG = performanceavoid goals; SE = academic self efficacy; PI = parental involvement; PR = parent relatedness; PA = parental autonomy support; OFR = offcampus friend relatedness; IR = instructor relatedness; OCR = on campus friend relatedness; SB = sense of belonging; EX = extracurricular activities; CO = conscientiousness; EN = entity beliefs; HSGPA = high school GPA; Gro up = ethnicity group; SA = social adjustment; INPS = intention to persist. *p < .05. **p < .01. ***p < .001. ****p < .0001.

PAGE 95

95 Table 3-5 Effect s proposed in original model and effects in revised model Variable CM IC MG AG SE GPA SA INPS PI PR PA OFR IR OCR SB EX CO + EN CM IC + MG AG SE + GPA SA Note : CM = commitment making; IC = identification with commitment; MG = mastery goals; AG = performanceavoid goals; SE = academic self efficacy; PI = parental involvement; PR = parent relatedness; PA = parental autonomy support; OFR = off-campus friend r elatedness; IR = instructor relatedness; OCR = on campus friend relatedness; SB = sense of belonging; EX = extracurricular activities; CO = conscientiousness; EN = entity beliefs; HSGPA = high school GPA; Group = ethnicity group; SA = social adjustment; IN PS = intentions to persist. = hypothesized relationship that was significant in revised model. = hypothesized relationship that was not significant in revised model. + = significant relationship that was not proposed in the original model.

PAGE 96

96 CHAPTER 4 DISCUSSION The purpose of this study wa s to test a model of college success and retention based on a bioecological perspective (Bronfenbrenner & Morris, 1998). This new bioecological systems mo del of college retention focused on the ecological systems of f amily, peer, college, and the self that predict college success and the psychological processes (identity, perceived academic self efficacy, and achievement goals) that mediate the relationship between the ecological systems and college success. This model addresse d weaknesses in the research literature by combining the sociological and psychological perspectives that are commonly used in the college retention literature. A preliminary analysis indicated a poor fit of the model to the data. On the basis of the modification indices, I added three paths and one correlated error. In the next section I summarize the results for each of the microsystems. Finally, I identify weaknesses in the study and discuss directions for future research and implications for theory and practice. The Bioecological S ystems The Family Microsytem Within the family microsytem I hypothesized that parental involvement predicts perceived academic self efficacy and GPA ; relatedness to parents predicts commitment making and mastery goals and parental autonomy support predicts commitment making, identification with commitment, GPA social adjustment, and intention to persist. However, only the hypothesized effect s of parental autonomy support on commitment making ( = .15, p = .02) and social adjustment ( = .23, p = .005) were confirmed That finding that parental autonomy support is related to social adjustment supports the findings of Soenens et al. (2007). In addition, t he finding that parental autonomy support is related to commitment making supports the findings of Luyckx Goossens, et al. (2006), and Luyckx, Soenens, Vansteenkiste, et al. (2007). However,

PAGE 97

97 parental autonomy support does not predict identification with commitment as found in Luyckx Goossens, et al. and Luyckx, Soenens, Vansteenkiste, et al. The participants in these two studies were Belgian college students most of whom live d at home. Parental autonomy support might be more important to identity when students live with their parents as m ost of the students in my university sample do not live at home (87%). In sum, only the parental variable of autonomy support is related to other variables in the model, specifically identity and social adjustment, even if college students do not live at home. The Peer Microsystem I hy pothesized that relatedness to offcampus friend s predict s social adjustment. However, results indicate that relatedness to off-campus friends is not related to social adjustment. In contrast, relatedness to oncampus friend s is a significant predictor of social adjustment, suggesting that on-campus friends may matter to social adjustment, but off-campus friends may not. The College Microystem I hypothesized that the college microsystem variable of relatedness to instructors predicts perceived academic sel fefficacy, GPA, and social adjustment; oncampus relatedness to friends predicts social adjustment and intention to persist; sense of belonging predict s commitment making, identification with commitment, mastery goals, performance -avoid goals, perceived a cademic self efficacy, social adjustment, and intention to persist, and e xtracurricular activities predict GPA social adjustment, and intention to persist. Results indicate that r elatedness to instructors is only significantly related to perceived academic self efficacy ( = .14, p = .05). This result with university students is consistent with CordellMcNulty and Ashtons (2008) result with community college students, suggesting that instructors may play a role in increasing the motivation of college students.

PAGE 98

98 As noted in the previous section, r elatedness to friends predicts social adjustment ( = .14, p = .01) as expected. This result supports the findings of Buote et al. (2007) and Friedlander et al. (2007) that friends play a role in the social adjustment of college st udents. Of all the variables in the college microsystem sense of belonging has the largest number of significant relationships to other variables in the model. Sense of belonging is a significant predictor of commitment making ( = .14, p = .04), identifi cation with commitment ( = .23, p = .001), and social adjustment ( = .34, p = .001). That sense of belonging is directly related to the psychological process variables of commitment making and identification with commitment lends support to Rodgers and S ummers s (2008) model, in which sense of belonging is hypothesized to predict psychological process variables. In addition, sense of belonging is indirectly related to the psychological process variables, mastery goals p = .013), performance av oid goals ( = -.006, p = .027) and academic self = .04, p = .004). through identification with commitment. The finding that sense of belonging is direct ly related to social adjustment is consistent with the findings of Mounts (2004) and Ostr ove and Long (2007 ). Sense of belonging is also related to intention to persist through identification with commitment and perceived academic self efficacy ( = .01, p = .01). The number of students extracurricular activities are positively associated wi th social adjustment ( = .21, p = .001). This result is consistent with finding s of Bohnert et al. (2007) who found a negative relationship between the number of hours a week students engaged in extracurricular activities and loneliness and a positive relationship between the number of hours students spent in extracurricular activities and the quality of their friendships. However, participation in extracurricular activities is not related to GPA or intention to persist in contrast to findings of Fischer ( 2007) and Cordell-McNulty and Ashton (2008) who reported significant

PAGE 99

99 relationships between p articipation in extracu rricular activities and GPA. It is notable that those two studies focused on students in the first 2 years of college, whereas most of the st udents in this study (59%) were in their last 2 years of college. Perhaps involvement in extracurricular activities in the later years of college is less important to achievement than it is in the first 2 years. The Self System I hypothesized that (a) the self system variable of conscientiousness predicts commitment making, identification with commitment, perceived self efficacy, and GPA; (b) the self system variable of entity beliefs predicts mastery and performance avoid goals, and academic achievement, and (c) the self system variable of academic achievement predicts academic achievement and intention to persist. Conscientiousness is directly related to the psychological process variables of commitment making ( = .29, p = .000) identification with commitment ( = .31, p = .000) and perceived academic self efficacy ( = .29, p = .000). The significant relationship between conscientiousness and commitment making and identification with commitment supports the findings of Luyckx, Soenens, and Goossens (2006). The finding that conscientiousness predicts perceived academic self -efficacy was not hypothesized in the original model but was added based on the modification indices. In addition, conscientiousness is in directly related to GPA through four psycholog ical process variables: perceived academic self efficacy ( = .05, p = .02), identification with commitment and perceived academic self efficacy ( = .009, p = .02), perceived academic self efficacy and performance-avoid goals ( = .008, p = .04), and identification with commitment, perceived academic self effica cy, and performanceavoid goals ( = .002, p = .04). These indirect relationships suggest path s through which conscientiousness affects GPA and they lend support to the res ults of Nguyen et al. (2005) and Wolfe and Johnson (1995), who also found that conscientiousness is related to GPA

PAGE 100

100 Consistent with my hypotheses and the findings of Robins and Pals (2002), entity beliefs are positively related to performance avoid goals ( = .18, p = .006) and negatively related to mastery goals ( = -.17, p = .006) ; h owever, contrary to expectation and the theoretical literature and previous research, entity beliefs (e.g., Aronson, 200) are positively related to GPA ( = .14, p = .04). In this study, t he correlation between entity beliefs and high-school GPA is nonsignificant, suggesting that this relationship may not be a reliable finding. E ntity beliefs are also indirectly related to GPA through performance-avoid goals ( = -.04, p = .02). This indirect relationship of entity beliefs with GPA is in the hypothesized di rection. Previous achievement, as measured by high school GPA, is directly related to social adjustment ( = .14, p = .02). However, high school GPA was not related to college GPA, as has been typically found in previous research ( e.g., El i as & MacDonald, 2007; Fischer, 2007; Mattson, 2007; Robbins et al., 2004; Robbins et al., 2006). This finding could be related to the restriction in range in high school and college GPAs in this university sample. The Macrosytem Gender is significant ly negatively related to performance avoid goals ( = -.13, p = .03), and positively related to mastery goals ( = .16, p = .004), and social adjustment ( = .11, p = .03) with males less likely to have performance-avoid goals and more likely to have mastery goals, and to be socially adjusted. Ethnicity predicts perceived academic self efficacy ( = .11, p = .05) and GPA ( = .37, p = .002). Students classified as Caucasian and Asian report higher perceived academic self efficacy and GPAs compared to students in other ethnic groups. The finding that ethnicity is related to college GPA is consistent with the results of Cordell-McNulty and Ashton (2008) and Robbins et al. (2006).

PAGE 101

101 Psychological Process Variables Identity I hypothesized that commitment making and identification with commitment predict academic achievement, social adjustment, and intention to persist in college students; however none of these relationships are significant. I dentification with commitment is significantly related to perceived academic self efficacy ( .19, p = .000) a path added to the original model based on the modification indices, and significantly related to social adjustment through perceived academic self efficacy ( .02, p = .006) and through perceived academic self efficacy and performan ce avoid goals ( .004, p = .02). Like its relationship to social adjustment, identification with commitment ha s two significant relationships with GPA mediated through perceived academic self p = .008) and through perceived academic s elf efficacy and performance p = .02). For intention to persist, there is only one significant indirect relationship which is mediated through perceived academic self efficacy p = .000). These results show that identificati on with commitment is related to all three of the outcome variables, but these relationships are completely mediated by academic self efficacy and performance -avoi d goals. Achievement Goals I hypothesized that mastery and performance-avoid goals predict academic achievement and social adjustment. However, results indicate only performanceavoid goals are related to academic achievement ( -.20, p = .0005) and social adjustment ( -.17, p = .0005). The finding that performance-avoid goals are negatively related to academic achievement replicates the findings of Hsieh et al. (2007). The finding that performance-avoid goals predict socia l adjustment supports the negative relationship hypothesized in the Bean and Eaton (2000) model. It is noteworthy that mastery goals are not positively related to social adjustment or GPA. The

PAGE 102

102 research on the relationship between mastery goals and achievem ent is mixed, and most studies reporting positive relationships between mastery goals and achievement have not included a control for prior ability. Perceived Academic Self Efficacy I hypothesized that perceived a cademic self efficacy predict s mastery goa ls, performance avoid goals, academic achievement, social ad justment, and intention to persist. Results suppor ted these hypotheses: Perceived academic self efficacy is related to mastery goals ( .15, p = .0005), academic achievement ( .17, p = .005), social adjustment ( .13, p = .003), and intention to persist ( .29, p = .000). In addition, perceived academic self efficacy is negatively related to performance -avoid goals ( -.14, p = .007). This relationship was not included in the original mo del but was added on the bas is of the modification indices. The finding that perceived academic self efficacy predicts academic achievement supports the results of Bembenutty (2007), Elias and MacDonald (2007), Multon et al. (1991), and Robbins et al. (2004). The finding that perceived academic self efficacy predicts intention to persist lends support to the findings of Robbins et al. In addition, the finding that perceived academic self efficacy predicts s ocial adjustment is consistent with research by DeW itz and Walsh (2002). That perceived academic self efficacy predict s ma stery goals is consistent with the finding of Hsieh et al. (2007). However, unlike this study, Hsieh et al. did not find a significant relationship between perceived academic self effic acy and performance -avoid goals. Outcome Variables I hypothesized that social adjustment and academic achievement predict intention to persist ; however, neither variable predicts intention to persist. This result may be due to a ceiling effect i n the meas ure of intent to persist ( M = 5.51 on a 6.00 scale).

PAGE 103

103 Limitations of the Study This study has several limitations that should be addressed in future studies: First, I did not measure actual persistence, only intention to persist. Although other researchers have found that intention to persist predicts persistence (e.g., Cabrera et al., 1993), measuring actual persistence would increase the ecological validity of the study. Second, GPAs, SAT, and ACT scores were based on students self-reports. Although obta ining official GPAs and test scores to ensure their accuracy is desirable, many students refuse to give researchers permission to obtain their official test scores from the registrar. St rategies to overcome students reluctance need to be explored. Third, because the relationships obtained in this study are correlational, conclusions regarding the direction of causality are speculative and require validation through experimental research. Implications of the Findings for Future Research Of particular note, perceived academic self efficacy is the only variable in the model with a direct relationship to the three outcome variables: intention to persist, social adjustment, and GPA. In addition academic self efficacy mediated many of the indirect effects in the model as well. These results are consistent with Banduras (1997) conception of perceived self efficacy and its importance for academic success. Bandura identified four sources of perceived selfefficacy: (a) mastery experience, (b) vicarious experiences, (c) verbal persuasion, and (d) physiological and affective states. Future research designed to increase students retention in college should investigate strategies based on those four sources. In terms of predicting outcomes, performance-avoid goals is the second most influential variable, predicting students social adjustment and GPA. This finding suggests a need to develop strategies to help students develop more positive achievement goals. It was surprising that mastery goals, commitment making, and i dentification with commitment are not significant predictors of any of the outcome variables. However, identification with commitment is indirectly related to all three of the

PAGE 104

104 outcome variables. In sum, perceived academic self effica cy and performance-avoi d goals are the strongest predictors of the three college outcome s: intent to persist, GPA, and social adjustment. Conclusions The new bioecological model of college retention tested in this study provides researchers and higher education administrators w ith a structured way of identifying strategies from multiple sources with diverse relationships to students GPA, social adjustment, and intention to persist that could be adopted simultaneously to improve the likelihood of increasing college retention. Fo r example, with one exception (offcamp us friends in the peer microsystem) significant relationships to at least one of the outcome variables were found in every mic rosystem of the students lives. Of the parental variables examined in this study, parental autonomy support has a direct relationship with social adjustment. Research focusing on helping parents understand the benefits of providing autonomy support during freshm an orientation or through parent newsletters may be helpful in increasing retention With regard to the college microsystem, several approaches to increasing retention are suggested from this study. S tudents relatedness to instructors has a direct relationship to students perceived academic efficacy. A dministrators could investigate whe ther providing instructors with strategies to help students feel accepted in their classrooms increases students social adjustment, GPA, and intention to persist through its relationship with perceived academic self efficacy. Participation in extracurricular activities has a significant relationship to social adjustment. R e search is warranted to determine whether activities encouraging students to participate in extracurricular activities increases retention Sense of belonging to the institution has the s trongest direct relationship of all the variables in the study to social adjustment. Strategies to increase identity with the university should be tested as a potentially important way to influence students social adjustment. In the self system,

PAGE 105

105 conscient iousness has the largest number of indirect relationships to GPA and intention to persist of any variable in the study. S trategies to increase students conscientiousness may have effects on multiple psychological processes related to ac hievement and inten t to persist, including mastery goals, performanceavoid goals, and perceived academic self efficacy. The results of this study emphasize the importance of the relationships between psychological process variables (in particular, perceived academic self ef ficacy and performance avoid goals ) and the outcome variables of social adjustment, GPA, and intention to persist. These results suggest that r esearchers and higher education administrators might focus on developing interventions to increase academic self efficacy and decrease performance-avoid goals as a means of increasing students retention One way to increase perceived self efficacy that is suggested in the model tested in this study is to increase students feelings of relatedness to their instructors One way to decrease performanceavoid goals suggested by the model is to decrease entity beliefs and help students foster a more incremental view of intelligence. It is my hope that these findings will help advance research on the problems of college retention and ultimately lead to increasing the graduation rate of college students.

PAGE 106

106 LIST OF REFERENCES Adams, G. R., Bennion, L., & Huh, K. (1989). Objective Measure of Ego I dentity status: A reference manual. Logan: Utah State University. Adams G. R., Shea, J., & Fitch, S. A. (1979). Toward the development of an objective assessment of ego identity status. Journal of Youth and Adolescence, 8, 223-237. Alford, S. M. (2000). A qualitative study of the college social adjustment of Black students f rom lower socioeconomic communities. Journal of Multicultural Counseling and Development, 28, 2-15. Armsden, G. C., & Greenberg, M. T. (1987). The Inventory of P arent and P eer A ttachment: Individual differences and their relationship to psychological wellbeing in adolescence. Journal of Youth and Adolescence, 16, 427-453. Arnett, J. J. (2000). Emerging adulthood: A theory of development from the late teens through the twenties. American Psychologist, 55, 469-480. Arnett, J. J. (2006). Emerging adulthood: Understanding the new way of coming of age. In J. J. Arnett & J. L. Tanner (Eds.), Emerging adults in America: Coming of age in the 21st century (pp. 3-19). Washington, DC: American Psychological Association. Aronson, J., Fried, C. B., & Good, C. (2002). Reducing the effects of stereotype threat on African American college students by shaping theories of intelligence. Journal of Experimental Social Psychology, 38, 113-125. Arrindell, W. A., & E llema, J. H. M. (1986). Klachtenlijst (SCL90) [Complaint List SCL -90.]. Lisse. The Netherlands: Swets & Zeitlinger. Asher, S. R., & Wheeler, V. A. (1985). Childrens loneliness: A comparison of rejected and neglected peer status. Journal of Consulting and Clinical Psychology 53, 500-505. Astin, A. W. (1997). How good is your institutions retention rate? Research in Higher Education, 38, 647-658. Awad, G. H. (2007). The role of racial identity, academic self-concept, and selfesteem in the prediction of academic outcomes for African American students. Journal of B lack Psychology, 33, 188-207. Baker, R. W., & Siryk, B. (1984). Measuring adjustment to college. Journal of Counseling Psychology, 31, 179-189. Baker, R. W., & Siryk, B. (1999,1989). Manual of the SACQ. Los Angeles: Western Psychological Services. Balistr eri, E., Busch-Rossnagel, N. A., & Geisinger, K. F. (1995). Development and preliminary validation of the Ego Identity Process Questionnaire. Journal of Adolescence, 18, 179-192.

PAGE 107

107 Bandura, A. (1997). Self efficacy: The exercise of control. New York: Freeman Barber, B. K. (1996). Parental psychological control: Revisiting a neglected construct. Child Development, 67, 3296-3319. Barnes, H., & Olson, D. H. (1992). Parent-adolescent communication. In H. I. McCubbin, H. Barnes, A. Larsen, M. Muxen, & M Wilson (Eds.), Family inventories (2 nd ed., pp. 2944). Saint Paul, MN: Family Social Science. Baumeister, R. F., & Leary, M. R. (1995). The need to belong: Desire for interpersonal attachments as a fundamental human motivation. Psychological Bulletin, 117, 497-529. Bean, J. P., & Eaton, S. B. (2000). A psychological model of college student retention. In J. M. Braxton (Ed.), Reworking the student departure puzzle (pp. 48-61). Nashville, TN: Vanderbilt University Press. Beck, A. T., Epstein, N., Brown, G., & Steer, R. A. (1988). An inventory for measuring clinical anxiety: Psychometric properties. Journal of Counseling and Clinical Psychology, 56, 893897. Beck, A. T., Ward, C. H., Mendelson, M., Mock, J., & Erbaugh, J. (1961). An inventory for measuring depression. Archives of General Psychiatry, 5, 561-571. Bembenutty, H. (2007). Selfregulation of learning and academic delay of gratification: Gender and ethnic differences among college students. Journal of Advanced Academics, 18, 586616. Bers, T. (2005). Parents of traditionally aged community college students: Communications and choice. Research in Higher Education, 46, 413-436. Berzonsky, M. D. (1992). Identity style and coping strategies. Journal of Personality, 60, 771787. Berzonsky, M. D., & Kuk, L. S. (2000). Identity status, identity processing style, and the transition to university. Journal of Adolescent Research, 15, 81-98. Betz, E. L., Betz, N. E., & Menne, J. W. (1989). Manual for the C ollege Student Satisfaction Questionnaire. Unpublished manual. Columbus, OH: Ohio State University. Bohnert, A. M., Aikins, J. W., & Edidin, J. (2007). The role of organized activities in facilitating social adaptation across the transition to college. Journal of Adolescent Research, 22, 189208. Bollen, K. A., & Hoyle, R. H. (1990). Perceived cohesion: A conceptual and empirical examination. Social Forces, 69, 479-504.

PAGE 108

108 Boyd, V. S., Hunt, P. F., Kandell, J. J., & Lucas, M. S. (2003). Relationship between identity processing style and academic success in undergraduate students. Journal of College Student Development, 44, 155-167. Braxton, J. M., & Hirschy, A. S. (2005). Theoretical developments in the study of college student departure. In A. Seidman (Ed.), College student retention: Formula for student success (pp. 61-87). Westport, CT: Praeger. Braxton, J. M., & Hirschy, A. S., & McLendon, S. A. (2004). Understanding and reducing college student departure. No. 3 ASHE -ERIC Higher Education Research Report Series. San Francisco: Jossey Bas s. Braxton, J. M., Sullivan, A. V. S., & Johnson, R. M., Jr. (1997). Appraising Tintos theory of college student departure. In J. C. Smart (Ed.), Higher education: Handbook of theory and research, Vol. 12 (pp. 107-164). New York: Agathon Press. Bronfenbrenner, U. (1979). The ecology of human development: Experiments by nature and design. Cambridge, MA: Harvard University Press. Bronfenbrenner. U., & Morris, P. A. (1998). The ecology of developmental processes. In W. Damon (Series Ed.) & R. M. Lerner (Vol. Ed.), Handbook of child psychology: Vol. 1 (5 th ed., pp. 993-1028). New York: Wiley. Buote, V. M., Pancer, S. M., Pratt, M. W., Adams, G., Birnie-Lefcovitch, S., Polivy, J., & Wintre, M. G. (2007). The importance of friends: Friendship and adjustment among 1 st year university students. Journal of Adolescent Research, 22, 665-689. Byars -Winston, A. M., & Fouad, N. A. (2008). Math and science social cognitive variables in college students: Contributions of contextual factors in predicting goals. Journal of C areer Assessment, 16, 425-440. Cabrera, A. F., Nora, A., & Castaneda, M. B. (1993). College persistence: Structural equations modeling test of an integrated model of student retention. Journal of Higher Education, 64, 123-139. Caro, J. E. (1985). An empirical investigation of attachment in adulthood and its relationship to adjustment in college. Unpublished doctoral dissertation, Syracuse University, Syracuse, NY. Clancy, S. M., & Dollinger, S. J. (1993). Identity, self, and personality: Identity status and the five -factor model of personality. Journal of Research on Adolescence, 3, 227-245. Cordell-McNulty, K., & Ashton, P. (2006, August). College students perceptions of parental involvement and their academic selfefficacy. Poster presented at the meeting of the American Psychological Association, New Orleans, LA.

PAGE 109

109 Cordell-McNulty, K., & Ashton, P. (2008, March). Relationship of perceived social support to self -efficacy and achievement in community college students. Paper presented at the meeting of the Ame rican Educational Research Association, New York Costa, P. T., & McCrae, R. R. (1992). Revised NEO Personality Inventory (NEO PI -R) and NEO Five Factor Inventory (NEO -FFI) professional manual Odessa, FL: Psychological Assessment Resources. Cutrona, C. E. (1989). Ratings of social support by adolescents and adult informants: Degree of correspondence and prediction of depressive symptoms. Journal of Personality and Social Psychology, 57, 723-730. Cutrona, C. E., Cole, V., Colangelo, N., Assouline, S. G., & Russell, D. W. (1994). Perceived parental social support and academic achievement: An attachment theory perspective. Journal of Personality and Social Psychology, 66, 369-378. Cutrona, C. E., & Russell, D. (1987). The provisions of social relationships and adaptation to stress. In W. H. Jones & D. Perlman (Eds.), Advances in personal relationships (Vol. 1, pp. 37-68). Greenwich, CT: JAI Press. Deci, E. L., & Ryan, R. M. (2000). The what and why of goal pursuits: Human needs and the self determination o f behavior. Psychological Inquiry, 11(4), 227-268. De Witz, S. J., & Walsh, W. B. (2002). Self-efficacy and college student satisfaction. Journal of Career Assessment, 10, 315-326. Dixon Rayle, A. (2006). Mattering to others: Implications for the counseling relationship. Journal of Counseling and Development, 84, 483-487. Dixon Rayle, A., & Chung, K. (2007). Revisiting college students mattering: Social support, academic stress, and the mattering experience. Journal for College Student Retention: Research, Theory, Practice, 9, 21-37. Duchesne, S., Ratelle, C. F., Larose, S., & Guay, F. (2007). Adjustment trajectories in college science programs: Perceptions of qualities of parents and college teachers relationships. Journal of Counseling Psychology, 54, 62-71. Dweck, C. S. (1999). Self -theories: Their role in motivation, personality, and development. Philadelphia: Taylor & Francis. Dweck, C. S., Chiu, C., & Hong, Y. (1995). Implicit theories and their role in judgments and reactions: A world from two perspe ctives. Psychological Inquiry, 6, 267-285. Elias, S. M., & Loomis, R. J. (2000). Using an academic selfefficacy scale to address university major persistence. Journal of College Student Development, 41, 450-454.

PAGE 110

110 Elias, S. M., & MacDonald, S. (2007). Using past performance, proxy efficacy, and academic self efficacy to predict college performance. Journal of Applied Social Psychology, 37, 2518-2531. Elliot, A. J., & Church, M. (1997). A hierarchical model of approach and avoidance achievement motivation. Jo urnal of Personality and Social Psychology, 72, 218-232. Epstein, J. L., Connors, L. J., & Clark Salinas, K. (1993). High school and family partnerships: Questionnaires for teachers, parents, and students. Baltimore, MD: Center on School, Family, and Community Partnerships. Erdley, C. A., & Dweck, C. S. (1993). Childrens implicit personality theories as predictors of their social judgments. Child Development, 64, 863-878. Erikson, E. H. (1968). Identity youth and crisis. New York: W.W. Norton. Ferry, T R., Fouad, N. A., & Smith, P. L. (2000). The role of family context in a social cognitive model for career related choice behavior: A math and science perspective. Journal of Vocational Behavior, 57, 348-364. Fischer, M. J. (2007). Settling into campus life: Differences by race/ethnicity in college involvement and outcomes. Journal of Higher Education, 78, 125-161. Flescher, M. (1986). Unpublished data from Skidmore College, Saratoga Springs, NY. Fr eeman, T. M., Anderman, L. H., & Jensen, J. M. (2007). Sense of belonging in college freshmen at the classroom and campus levels. Journal of Experimental Education, 75, 203220. Friedlander L. J., Reid, G. J., Shupak, N., & Cribbie, R. (2007). Social support, selfesteem, and stress as predictors of adjustment to university among firstyear undergraduates. Journal of College Student Development, 48, 259-274. Furrer, C., & Skinner, E. (2003). Sense of relatedness as a factor in childrens academic engagement and performance. Journal of Educational Psychology, 95 (1), 148-162. Gaskill, P. J., & Woolfolk Hoy, A. (2002). Selfefficacy and self regulated learning: the dynamic duo in school performance. In J. Aronson (Ed.), Improving academic achievement: I mpact of psychological factors on education (pp. 185-208). San Diego: Academic Press. Goldberg, L. R. (1992). The development of markers for the BigFive factor structure. Psychological Assessment, 4, 26-42. Goldberg, L. R. (1999). A broad-bandwidth, public domain, personality inventory measuring the lower level facets of several five factor models. In I. Mervielde, I. Deary, F. De Fruyt, & F. Ostendorf (Eds.), Personality psychology in Europe, Vol. 7 (pp. 7-28). Tillburg, The Netherlands: Tillburg University Press.

PAGE 111

111 Goodenow, C. (1993). The psychological sense of school membership among adolescents: Scale development and educational correlates. Psychology in the Schools, 30, 79-90. Gottfried, A. W. (1985). Measures of socioeconomic status in child development research: Data and recommendations. Merrill Palmer Quarterly, 31, 85-97. Grolnick, W. S. (2003). The psychology of parental control. Mahwah, NJ: Erlbaum. Grolnick, W. S., Ryan, R. M., & Deci, E. (1991). Inner resources for school achievement: Motivational mediators of childrens perceptions of their parents. Journal of Educational Psychology, 83, 508-517. Gupta, S., Harris, D. E., Carrier, N. M., & Caron, P. (2006). Predictors of student success in entry level undergraduate mathematics courses. College Student Journal, 40, 97-108. Hagerty, B. M. K., Lynch-Sauer, J., Patusky, K., Bouwsema, M., & Collier, P. (1992). Sense of belonging: A vital mental health concept. Archives of Psychiatric Nursing, 6, 172-177. Harris, B. C. (1988). How the structure and function of social networks relate to loneliness and college adaptation. Unpublished masters thesis, George Washington University, Washington, DC. Hausmann, L. R. M., Schofield, J. W., & Woods, R. L. (2007). Sense of belonging as a predictor of intentions to persist among African Americans and White first year college students. Research in Higher Education, 48, 803-839. Hoekstra, H. A., Ormel, J., & de Fruyt, F. (1996). NEO Persoonlijkheids Vragenlijsten: NEOPI R, NEO FFI. Handleiding. Lisse, The Netherlands: Swets & Zeitlinger. Hofer, B. K., Kennedy, E., & Hurd, K. (August, 2006). From other regulation to self regulation: Parental contact and influence during the transition to college. Paper presented at the meeting of the American Psychological Association, New Orleans. Hoffman, J. A. (1984). Psychological separation of late adolescents from their parents. Journal of Counseling Psychology, 31, 170-178. Hogan, M. P. (1986). Unpublished data from University of Maryland, College Park. Hollingshead, A.B. (1975). Four factor index of social status Unpublished manuscript, Yale University, Department of Sociology, New Haven, CT. Hoover-Dempsey, K. V., & Sandler, H. M. (1995). Parental involvement in childrens education: Why does it make a difference? Teachers College Record, 97, 311-331. Hsieh, P., Sullivan, J. R., & Guerra, N. S. (2007). A closer look at college students: Selfefficacy and goal orientation. Journal of Advanced Academics, 18 (3), 454-476.

PAGE 112

112 Hurtado, S., & Carter, D. (1997). Effects of college transition and perceptions of the campus racial climate on Latino college students sense of belonging. Sociology of Education, 70, 324-345. John, O. P., Donahue, E. M., & Kentle, R. L. (1991). The Big Five InventoryVersions 4a and 54. Berkeley, CA: University of California, Berkeley, Institute of Personality and Socia l Research. John, O. P., & Srivastava, S. (1999). The big five trait taxonomy: History, measurement, and theoretical perspectives. In L. A. Pervin & O. P. John (Eds.), The handbook of personality (2nd ed., pp. 102-138). New York: Guilford Press Kuncel, N. R., Cred, M., & Thomas, L. L. (2005). The validity of self-reported grade point averages, class ranks, and test scores: A metaanalysis and review of the literature. Review of Educational Research, 75 63-82. Larose, S., Tarabulsy, G., & Cyrenne, D. (2005). Perceived autonomy and relatedness as moderating the impact of teacher -student mentoring relationships on student academic adjustment. Journal of Primary Prevention, 26, 111-128. Luyckx, K., Goossens, L., Soenens, B., Beyers, W., & Vansteenkiste, M. (2005). Identity statuses based on four rather than two identity dimensions: Extending and refining Marcias paradigm. Journal of Youth and Adolescence, 34, 605-618. Luyckx, K., Goossens, L., Soenens, B., & Beyers, W. (2006). Unpacking commitment and exp loration: Preliminary validation of an integrative model of late adolescent identity formation. Journal of Adolescence, 29, 361-378. Luyckx, K., Schwartz, S. J., Berzonsky, M. D., Soenens, B., Vansteenkiste, M., Smits, I., & Goosens, L. (2008). Capturing ruminative exploration: Extending the fourdimensional model of identity formation in late adolescence. Journal of Research in Personality, 42, 58-82. Luyckx, K., Soenens, B., & Goossens, L. (2006). The personalityidentity interplay in emerging adult women: Convergent findings from complementary analyses. European Journal of Personality, 20, 195-215. Luyckx, K., Soenens, B., Goossens, L., & Vansteenkiste, M. (2007). Parenting, identity formation, and college adjustment: A mediation model with longitudinal data. Identity, 7, 309-330. Luyckx, K., Soenens, B., Vansteenkiste, M., Goossens, L., & Berzonsky, M. D. (2007). Parental psychological control and dimensions of identity formation in emerging adulthood. Journal of Family Psychology, 21, 546-550. Mahoney, J. L., Cairns, B. D., & Farmer, T. W. (2003). Promoting interpersonal competence and educational success through extracurricular activity participation. Journal of Educational Psychology, 95, 409-418.

PAGE 113

113 Marcia, J. E. (1980). Identity in adolescence. In J. Adelson (Ed.), Handbook of adolescent psychology (pp. 159-187). New York: Wiley. Marcia, J. E. (1993a). The ego identity status approach to ego identity. In J. E. Marcia, A. S. Waterman, D. R. Matteson, S. L. Archer, & J. L. Orlofsky (Eds.), Ego identity: A handbook for psychosocial research (pp. 3-21). New York: SpringerVerlag. Marcia, J. E. (1993b). Prologue. In J. Kroger (Ed.), Discussions of ego identity (pp. xiii-xv). Hillsdale, NJ: Erlbaum. Marcoen, A., Van Cotthem, K., Billiet, K., & Beyers, W. (2002). Dimensies van subjectief welbevinden bij ouderen [Dimensions of subjective well -being in elderly people]. Tijdschrift voor Gerontologie en Geriatrie 33, 156. Mattson, C. E. (2007). Beyond admission: Understanding pre-college variables and the succ ess of at -risk students. Journal of College Admission, 196, 8-13. McCarron, G. P., & Inkelas, K. K. (2006). The gap between educational aspirations and attainment for first-generation college students and the role of parental involvement. Journal of College Student Development, 47, 534-549. McCrae, R. R., & John, O. P. (1992). An introduction to the five-factor model and its applications. Journal of Personality, 60, 175-215. McGregor, H., & Elliot, A. J. (2005). The shame of failure: Examining the link betw een fear of failure and shame. Personality and Social Psychology Bulletin, 31, 218-231. Meeus, W., & Dekovic, M. (1995). Identity development, parental and peer support in adolescence: Results of a national Dutch survey. Adolescence, 30, 931-944. Mendelson, M. J., & Aboud, F. E. (1999). Measuring friendship quality in late adolescents and young adults: McGill Friendship Questionnaires. Canadian Journal of Behavioural Science, 31, 130-132. Midgley, C., Maehr, M. L., Hruda, L. Z., Anderman, E., Anderman, L., & Freeman, K. E., et al. (2000). Manual for the Patterns of Adaptive Leaning Scales. Ann Arbor, MI: University of Michigan. Midgley, C., Maehr, M. L., & Urdan, T. C. (1993). Patterns of Adaptive Learning Survey (PALS) Ann Arbor, MI: University of Michigan. Moller, A. C., Elliot, A. J., & Friedman, R. (2008). When competence and love are at stake: Achievement goals and perceived closeness to parents in an achievement context. Journal of Research in Personality, 42, 1386-1391. Moore, J., Lovell, C. D., McGann, T., & Wyrick, J. (1998).Why involvement matters: A review of research on student involvement in the collegiate setting. College Student Affairs Journal, 17(2), 4-17.

PAGE 114

114 Mounts, N. S. (2004). Contributions of parenting and campus climate to freshmen adjus tment in a multiethnic sample. Journal of Adolescent Research, 19, 468-491. Mounts, N. S. (2007) [Reliability and validity data for Parental Support for College measure]. Unpublished raw data. Multon, K. D., Brown, S. D., & Lent, R. W. (1991). Relation of self efficacy beliefs to academic outcomes: A meta analytic investigation. Journal of Counseling Psychology, 38, 30-38. National Center for Education Statistics. (2005a ). Postsecondary institutions in the United States: Fall 2003 and degrees and other awards conferred: 2002-03. Retrieved July 15, 2008, from http://nces.ed.gov/fastfacts/display.asp?id=72 National Center for Educatio n Statistics. (2005b). Student effort and progre ss: Postsecondar y persistence and progress Retrieved December 1, 2005, from http://nces.ed.gov/programs/coe/2004/section3/indicator19.asp National Center for Education Statistics. (2007). Enrollment in postsecondary institutions, Fall 2005; graduation rates, 1999 and 2002 cohorts; and financial statistics fiscal year 2005. Retrieved February 3, 2009, from http://nces.ed.gov/pubs2007/2007154.pdf National Center for Education Statistics. (2008). Digest of education statistics, 2007. Retrieved June 11, 2009, from http://nces .ed.gov/fastfacts/display.asp?id=98 National Center for Higher Education Management Systems. (2007). Graduation rates: Six-y ear graduation rates of bachelor's students 2006. Retrieved March 24, 2008, from http://www.higheredinfo.org/dbrowser/index.php?level=nation&mode=map&state=0&sub measure=27 Neuville, S., Frenay, M., Schmitz, J., Boudrenghien, G., Noel, B., & Wertz, V. (2007). Tintos theoretical perspective and expectancy value paradigm : A confrontation to explain fr eshmens academic achievement. Psychologica Belgica, 47, 31-50. Nguyen, N. T., Allen, L. C., & Fraccastoro, K. (2005). Personality predicts academic performance: Exploring the moderating role of gender. Journal of Higher Education Policy and Management, 27, 105-116. Ochse, R., & Plug, C. (1986). Cross-cultural investigation of the validity of Eriksons theory of personality development. Journal of Personality and Social Psychology, 50, 1240-1252. Ostrove, J. M., & Long, S. M. (2007). Social class and belonging: Implications for college adjustment. Review of Higher Education, 30, 363-389. Parker, J. G., & Asher, S. R. (1993). Friendship and friendship quality in middle childhood: Links with peer group acceptance and feelings of loneliness and social dissatisfaction. Developmental Psychology 29, 611-621.

PAGE 115

115 Pascarella, E. T., & Terenzini, P. T., (1980). Predicting persistence and voluntary dropout decisions from a theoretical model. Journal of Higher Education, 61, 60-75. Paulson, S. E., Marchant, G. J., & Rothlisberg, B. A. (1994, October). Constructs underlying students perceptions of parents, teachers, and schools. Paper presented at the Midwestern Educational Research Association, Chicago, IL. Pekrun, R., Elliot, A. J., & Maier, M. A. (2006). Achievement goals and discrete achievement emotions. Journal of Educational Psychology, 98, 583-597. Pintrich, P. R., Smith, D. A. F., Garcia, T., & McKeachie, W. J. (1991). A manual for the use of the Motivated Strategies for Learning Questionnaire (MSLQ). Ann Arbor, MI: National Center for Research to Improve Postsecondary Teaching and Learning. Ra dloff, L. S. (1977). The Center for Epidemiologic StudiesDepression Scale: A self report depression scal e for research in the general population. Journal of Applied Psychological Measurement, 1, 385-401. Ratelle, C. F., Larose, S., Guay, F., & Sencal, C. (2005). Perceptions of parental involvement and support as predictors of college students persistence i n a science curriculum. Journal of Family Psychology, 19(2), 286-293. Robbins, S. B., Allen, J., Casillas, A., Peterson, C. H., & Le, H. (2006). Unraveling the differential effects of motivational and skills, social, and self management measures from trad itional predictors of college outcomes. Journal of Educational Psychology, 98, 598616. Robbins, S. B., Lauver, K., Le, H., Davis, D., Langley, R., & Carlstrom, A. (2004). Do psychosocial and study skill factors predict college outcomes? A metaanalysis. Psychological Bulletin, 130, 261-288. Robins, R. W., & Pals, J. L. (2002). Implicit selftheories in the academic domain: Implications for goal orientation, attributions, affect, and selfesteem change. Self and Identity, 1, 313336. Robinson, C. C., Mandleco, B. L., Frost Olsen, S., Bancroft-Andrews, C., McNeilly, M. K., & Nelson, L. (1995, March). Authoritative, authoritarian, and permissive parenting practices: Psychometric support for a new measure. Paper presented at the m eeting of the Society for Resea rch in Child Development, Indianapolis, IN. Robinson, T. N., III. (1996). A revision of the institutional integration model: A redefinition of persistence and the introduction of developmental variables. Unpublished doctoral dissertation, The Ohio State University, Columbus. Robinson, T. N., III. (2003). Identity as a mediator of institutional integration variables in the prediction of undergraduate persistence intentions. Journal of Adolescent Research, 18, 324.

PAGE 116

116 Rodgers, K. A., & Summers, J. J. (2008). African American students at predominantly white institutions: A motivational and self -systems approach to understanding retention. Educational Psychology Review, 20, 171-190. Rosenberg, M. (1965). Society and adolescent self-image. Princeton, NJ: Princeton University Press. Ross, M. E., Shannon, D. M., Salisbury-Glennon, J. D., & Guarino, A. (2002). The Patterns of Adaptive Learning Survey: A comparison across grade levels. Educational and Psychological Measurement, 62, 483-497. Saracoglu, B. N. (1987). Adjustment of learning disabled students to university as related to selfesteem and self efficacy. Unpublished masters thesis, York University, North York, Ontario, Canada. Savino. F., Reuter-Krohn, K., & Costar, D. (1986). Unpublished data from the University of Wisconsin, Madison. Schludermann, E. H., & Schludermann, S. M. (1988). Childrens Report on Parent Behavior (CRPBI -108, CRPBI-30) for older children and adolescents (Technical Report). Winnipeg, Canada: Dept of Psychology, University of Manitoba. Schunk, D. H., Pintrich, P. R., & Meece, J. L. (2008). Motivation in education: Theory, research, and applications. Upper Saddle River, NJ: Pearson Prentice Hall. Silk, J. S., Morris, A. S., Kanaya, T., & Steinberg, L. (2003). Psychological control and au tonomy granting: Opposite ends of a continuum or distinct constructs? Journal of Research on Adolescence, 13, 113-128. Smith, P. L., & Fouad, N. A. (1999). Subjectmatter specificity of self efficacy, outcome expectancies, interests, and goals: Implication s for the social cognitive model. Journal of Counseling Psychology, 46, 461-471. Soenens, B., Vansteenkiste, M., Lens, W., Luyckx, K., Goossens, L., Beyers, W., & Ryan, R. M. (2007). Conceptualizing parental autonomy support: Adolescent perceptions of promotion of independence versus promotion of volitional functioning. Developmental Psychology, 43, 633-646. Solberg, V. S., OBrien, K., Villareal, P., Kennel, R., & Davis, B. (1993). Selfefficacy and Hispanic colle ge students: Validation of the College Se lf -E fficacy Instrument. Hispanic Journal of Behavioral Sciences, 15, 80-95. Tinto, V. (1975). Dropout from higher education: A theoretical synthesis of recent research. Review of Educational Research, 45, 89-125. Tinto, V. (1993). Leaving college: Rethinking the causes and cures of student attrition (2nd ed). Chicago: University of Chicago Press.

PAGE 117

117 Tinto, V. (2005). Moving from theory into action. In A Seidman (Ed.), College student retention: Formula for student success (pp. 317-333). Westport, CN: Greenwood eBooks. Retrieved July 31, 2006 from http://ebooks.greenwood.com.lp.hscl.ufl.edu/reader.jsp?x=C8193&p=334&bc=EC8193 Trapnell, P. D., & Campbell, J. D. (1999). Pr ivate self -consciousness and the FiveFactor model of personality: Distinguishing rumination from reflection. Journal of Personality and Social Psychology, 76, 284. Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief m easures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54, 1063-1070. Wick, S., & Shilkret, R. (1986). Unpublished data from Mount Holyoke College, South Hadley, MA. Winston, R. B., Jr., & Miller, T. K. (1987). Student D evelopment Task and Lifestyle I nventory manual. Athens, GA: Student Development Associates. Wolfe, R. N., & Johnson, S. D. (1995). Personality as a predictor of college performance. Educational and Psychological Measurement, 55, 177-185. Zime t, G. D., Dahlem, N. W., Zimet, S. G., & Farley, G. K. (1988). The M ultidimensional Scale of P erceived S ocial Support. Journal of Personality Assessment, 52, 30-41. Zuckerman, M., Gagne, M., & Nafshi, I. (2001). Pursuing academic interests: The role of implicit theories. Journal of Applied Social Psychology, 31, 2621-2631.

PAGE 118

118 BIOGRAPHICAL SKETCH Kristi Cordell-McNulty was born and raised in south central Pennsylvania, just north of the Mason -Dixon line. She attended Shippensburg University and received a bachelors degree in computer science. After completing an internship and deciding computer science was not the right career path, she received her masters degree in psychology from Shippensburg University. Kristi was then a warded a Fellowship to study for t he Ph.D. in the University of Floridas educational and developmental psychology co-major Ph.D. program. After receiving her Ph.D., Kristi assumed the position of A ssistant Professor at Angelo State University in San Angelo, TX.