Citation
Impulsivity in children

Material Information

Title:
Impulsivity in children a factor analytic study
Creator:
Cohen, Joel Lawrence, 1952-
Publication Date:
Language:
English
Physical Description:
ix, 69 leaves : ; 28 cm.

Subjects

Subjects / Keywords:
Academic achievement ( jstor )
Achievement tests ( jstor )
Clinical psychology ( jstor )
Factor analysis ( jstor )
Intelligence quotient ( jstor )
Mathematics ( jstor )
Psychometrics ( jstor )
Questionnaires ( jstor )
Reflectance ( jstor )
Standard score ( jstor )
Child psychology ( lcsh )
Clinical Psychology thesis Ph. D ( lcsh )
Dissertations, Academic -- Clinical Psychology -- UF ( lcsh )
Psychological tests for children ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis--University of Florida.
Bibliography:
Bibliography: leaves 66-68.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Joel Lawrence Cohen.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
023287755 ( ALEPH )
06363625 ( OCLC )

Downloads

This item has the following downloads:


Full Text










IMPULSIVITY IN CHILDREN: A FACTOR ANALYTIC STUDY













BY

JOEL LAWRENCE COHEN


A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY


UNIVERSITY OF FLORIDA


[1979



































Copyright 1979

by

Joel Lawrence Cohen













ACKNOWLEDGMENTS


The author would like to express his deep appreciation to

Dr. Wiley C. Rasbury for his tireless efforts as chairman of his doctoral committee. Dr. Rasbury's many helpful suggestions, particularly with regard to ways for improving the author's writing style, contributed significantly to the successful completion of the present study. In addition, the author would also like to express his heartfelt thanks to Dr. Cynthia D. Belar, who also served on his doctoral committee. Dr. Belar's emotional support throughout the course of the completion of this study greatly aided the author, particularly at those times when his motivation and energy level began to waver. The author would also like to thank the other members of his doctoral committee: Dr. Hugh C. Davis, Dr. Jacquelin Goldman, and Dr. Larry Loesch, for their many helpful comments and suggestions. And finally, the author would also like to thank his wife, Mary, who amongst other things, typed the final copy of this manuscript. But more importantly, the author would like to thank his wife for putting up with six years of his work related irritability, piles of papers all around their home, and the placement of work directed towards the completion of his doctoral degree, before less pragmatic but potentially more enjoyable pursuits. Without her assistance and support, none of this would have been possible or desirable.














TABLE OF CONTENTS

ACKNOWLEDGEMENTS .................................................i1 i

LIST OF TABLES ................................................... vi

ABSTRACT ......................................................... vii

ONE INTRODUCTION ........................................... 1

Purpose ........... ................................ 1
Literature Review ................................. 2
Rationale ......................................... 6
Overview and Hypotheses ........................... 8

TWO METHOD ................................................. 11

Subjects .......................................... 11
Instrument ........................................ 13
Procedure ......................................... 13
Factor Analyses ................................... 16

THREE RESULTS ................................................ 18

Factor Analyses ................................... 18
Ma les ..... ................................... 18
Females ...................................... 19
Factor Scores ..................................... 22
Factor Score Reliability .......................... 23
The Relationship of Factor Scores to Subject
Variables ....................................... 26
The Relationship of Factor Scores to ISC Performance and Teacher Ratings of Impulsivity.....28 Validity of the ISC ............................... 33

FOUR DISCUSSION ............................................. 36










REFERENCES ....................................................... 66

BIOGRAPHICAL SKETCH.............................................. 69

APPENDICES

A REVISED IMPULSIVITY SCALE FOR CHILDREN (RISC) AND
IMPULSIVITY SCALE FOR CHILDREN (ISC) ................ 49

B TEACHER RATING SCALE FOR IMPULSIVITY .................. 56

C FACTOR MATRICES FOR MALES AND FEMALES ................. 57














LIST OF TABLES


TABLE NO. PAGE

1 Mean Age, Otis-Lennon IQ, and Standard Scores
from the Metropolitan Achievement Test .............. 12

2 Mean Age, Otis-Lennon IQ, and Standard Scores
from the Metropolitan Achievement Test for
the Reliability and Validity Sub-samples ............ 15

3 Significant Rotated Factor Loadings for Male
Factors 1 and 2 ..................................... 20

4 Significant Rotated Factor Loadings for Female
Factors 1 and 3 ..................................... 21

5 T-test Comparisons of Reliability Sub-sample to
Total Subject Pool .................................. 24

6 Reliability Coefficients for Factor Scores for
Males and Females ................................... 25

7 The Correlational Relationships of Factor Scores to
Subject Variables ................................... 27

8 Multiple Correlations of Impulsivity and Reflectivity
Factor Scores with Subject Variables ................ 29

9 T-test Comparisons of Validity Sub-sample to Total
Subject Pool ........................................ 31

10 The Correlational Relationships of Factor Scores to
ISC Performance and Teacher Ratings of Impulsivity..32

11 The Correlational Relationships of ISC Performance to Subject Variables ................................ 35










Abstract of Dissertation Presented to the Graduate Council of the University of Florida in Partial Ful fillIIment of the
Requirements for the Degree of Doctor of Philosophy IMPULSIVITY IN CHILDREN:
A FACTOR ANALYTIC STUDY

By

Joel Lawrence Cohen

August, 1979

Chairman: Wiley C. Rasbury, Ph.D. Major Department: Department of Clinical Psychology

Impulse control problems have been implicated in academic

performance deficits in children. However, to date, there has not been available a sufficiently reliable and valid method for assessing those aspects of impulsivity which, on an overt behavioral level, contribute to children being labelled as "impulsive." The purpose of this study was to develop a questionnaire to tap a potentially situation specific behavioral dimension of impulsivity. The questionnaire used in the present study was a revised version of the Impulsivity Scale for Children.

Both male and female students in fourth, fifth, and sixth grade elementary school classrooms participated in the study. The Revised Impulsivity Scale (RISC) was group adninistered to each individual class of subjects. The total sample studied included 450 children. Subsequently, a randomly selected sub-sample of approximately 100 vii








subjects was re-administered the RISC approximately ten weeks after its initial administration. A second randomly selected sub-sample of approximately 100 subjects completed the original version of the ISC. Teachers of children in this latter sample were also asked to rate each child's behavior according to the degree to which it was characterized by impulsivity. Where available, several pieces of demographic data were collected for each child participating in the study. These included age, race, Otis-Lennon Mental Abilities Test IQ, and standard scores on the Total Reading and Total Math portions of the Metropolitan Achievement Tests.

The questionnaire responses for males and females were analyzed separately. They were subjected to a principal components factor analysis using an oblique method of rotation. These analyses yielded two negatively correlated behavioral dimensions for both males and females; an "impulsive behavior" dimension and a "reflective behavior" dimension.

Factor scores were computed on both dimensions for each subject. An examination of the relationship between factor scores on both dimensions for those subjects who were re-administered the scale ten weeks after its initial administration, indicated that both dimensions were highly reliable for both males and females in this group. The observed relationship between factor scores on both viii








dimensions and teacher ratings of impulsivity, strongly supported the validity of the factor scores for males. For the females, however, while this relationship Was significant for the impulsivity dimension, it only approached significance for the reflectivity dimension.

The relationship between factor scores on both dimensions and available demographic data was also examined. Contrary to expectations, it was found that less reflective male and female subjects had higher achievement test scores than more reflective subjects. In addition, for females, it was also found that better achievement test performance was associated with higher scores on the impulsivity dimension. Several possible explanations for these findings were discussed, as were methods for testing them out in subsequent investigations using these scales.

Several methodological issues were discussed pertaining to the potential generalizability of these findings. In addition, several potential uses for the impulsivity and reflectivity scales were offered. However, it was noted that the individual scales were in need of refinement prior to their use as either research or clinical tools. A number of methods for revising the scales were offered.













CHAPTER I
INTRODUCTION

Purpose

The Impulsivity Scale for Children (ISC: Sutton-Smith and

Rosenberg, 1959) was developed as a self-report measure of a broad dimension of impulsive behavior. It was expected that the availability of an economical test for measuring this construct would allow practitioners such as psychologists and teachers to have at their

disposal, a method for conceptualizing impulsive behavior with a significant degree of consensual validation. To the extent that impulsive behavior could be implicated in poor academic performance, then the ability to objectively isolate very impulsive children would serve as a first step in developing appropriate interventions. However, despite the potential utility of such a scale, it has received remarkably little attention in the psychological literature to date. As such, the reliability and validity of the ISC, as a measure of impulsive behavior, have not been firmly established.

The purpose of the present study was to examine both the

reliability and validity of a revised form of this scale. The proposed revisions included the addition of several scale items which theoretically should have aided in the delineation of this scale as a measure of

1






2

impulsive behavior. The revised scale's long term stability was also examined for a normative population of nine to twelve-year-old elementary school children. Factor analytic methods were used to assess the factorial structure of the revised version of the scale. Scores on the revised version of the scale were compared to scores on the original version of the ISC, for a sub-sample of the total population studied. And finally, comparisons between scale scores and teacher ratings of impulsive behavior were used to assess the validity of both versions of the scale.


Literature Review

As it was originally developed, the ISC was a nineteen item

self-report scale for measuring a broad dimension of impulsive behavior. Each item could be responded to either "yes" or "no". This original scale was itself based upon the combination of two earlier versions of the scale, one based upon the responses of a group of college freshmen, and one based upon responses obtained from a sample of fourth, fifth, and sixth grade students. All of the items on the original version of the scale were keyed in the affirmative for impulsivity. In an answer to the criticism that the scale tapped the tendency to respond in the affirmative to self-descriptive statements regardless of content, and not impulsive behavior, Hirschfield (1965) modified the scale to include, in random order, both affirmative and negative






3

items keyed for impulsivity. Response categories were also changed from yes-no to true-false. A more recent, but less frequently used modification of the scale by Bjorklund and Butter (1973), involved the inclusion of five additional true-false questions to serve as a "Lie" scale. In this particular study, subjects giving three or more positive responses to the Lie items were excluded from further investigation.

To date, only two studies have reported any reliability data for the ISC. Sutton-Smith and Rosenberg (1959) reported a testretest reliability of .85 for the scale. More recently, Kendall and Finch (1978) reported a short term (one month) dependability coefficient of .56 and a long term (three months) stability coefficient of .33.

Validation studies have relied almost exclusively upon correlations between ISC scores and teacher rankings or ratings of impulsive behavior. For example, Sutton-Smith and Rosenberg (1959) reported that fifteen of eighteen correlations relating teacher rankings or ratings of impulsive behavior to impulsivity scale scores achieved significance at the .05 level or better. These correlations ranged from -.33 to -.73. Hirschfield (1965) also correlated teacher




The method of ranking/rating impulsivity determined the directionality of the correlations (e.g., 1=least impulsive and l0=most impulsive vs. 1=most impulsive and 10=least impulsive).






4

rankings of impulsive behavior with ISC scores, in a sample of fifth and sixth grade students. Ranked difference correlations were computed comparing teachers' rankings of their students within each class and the students' ranked test scores. Five of the resulting

fifteen correlations reached significance at the .05 level. These significant correlations ranged in magnitude from .73 to .87. Using teacher ratings as opposed to rankings of impulsivity, Bjorklund and Butter (1973) found little relationship between teacher ratings of specific behaviors and ISC scores, in a sample of fourth grade students. More recently, Bentler and McClain (1976) correlated ISC scores with both teacher and peer ratings of impulsive behavior for a sample of sixth grade students. The resulting correlations were .45 and .35 respectively. And finally, Vacc and Mercurio (1978) found a correlation of .56 between teacher ratings of impulsivity and ISC scores in a sample of fifth grade students.

The available reliability data for the scale areobviously very limited. Furthermore, the available validity data are not at all consistent. This may, in part, reflect methodological problems with the validation studies carried out thus far. Sample sizes have varied from a mean of 28.5 for each classroom examined (Sutton-Smith and Rosenberg, 1959) to an N of 144 (Bjorklund and Butter, 1973). With respect to the latter, it is of interest to note that, while having a modest sample size, and making use of teacher ratings of specific






5

behaviors (as opposed to the more global impulsive behavior ratings or rankings used in other studies), this study also yielded the least support for accepting the ISC as a valid measure of impulsive behavior. However, this finding may also have been an artifact of the more stringent statistical procedures used in this study, relying as it did, on teacher ratings vs. rankings of impulsive behavior. This latter explanation does lose some support in light of the work by Bentler and McClain (1976) and Vacc and Mercurio (1978), both of whom relied upon teacher ratings of impulsivity.

Obviously, additional research is needed to more firmly establish the psychometric credibility of the ISC. However, it is important to note that, with respect to its reliability, there are several potential sources of measurement error within the scale which could act to limit its reliability. These sources of error include the presence of scale items which are very global and non-situationally specific. In addition, many of the items are not very descriptive or operationally defined, and this may contribute to their being misunderstood by the children responding to them. And finally, the use of dichotomous rating categories can act to limit the reliability of the scale. Little attention has been paid to the issues of dependability vs. stability of ISC scores. This issue reflects a more basic concern with hypothesizing about whether impulsive behavior, defined in a broad sense, can be viewed as a state variable which






6

might be expected to spontaneously diminish with time, or a trait variable which would not be expected to change over time, unless specific interventions were designed to promote change.


Rationale

Impulsive behavior has been identified as being undesirable, particularly as it relates to poor academic performance in school. For example, based upon a large body of empirical research, Virginia Douglas and her associates (Douglas, 1974) have argued that the academic performance deficits frequently found in hyperactive children, are attributable to deficits in the areas of impulse control and attention. The realization of the potentially aversive conseq!lences of impulsive behavior has led to the development of a number of different types of interventions designed to enable impulsive children to learn to exert more conscious control over their behavior. For example, a number of investigators have proposed the use of instructional programs designed to teach cognitive self-mediation of behaviors, as a potentially effective method for modifying impulsive behavior (e.g., Kendall and Finch, 1976, 1978; Meichenbaum and Goodman, 1971). Others have recommended that impulsive children be instructed in the use of more effective problem solving and search strategies (e.g., Egeland, 1974; Heider, 1971). The use of these types of interventions depends upon the practitioner's ability to






7

effectively isolate children whose behavior is deemed to reflect impulsivity. One accepted method for doing this revolves around a conceptualization of impulsivity as a cognitive style variable, along with reflectivity. Kagan and his associates have proposed the use of the Matching Familiar Figures Test for classifying children as being either reflective or impulsive in their approach to problem solving situations (Kagan, Rosman, Day, Albert and Phillips, 1964; Kagan, 1965a, 1965b, 1965c). The Matching Familiar Figures Test is a match-to-sample cognitive task. Children are presented with a standard figure and usually, four, six or eight figures differing from one another in one or more details. The task for the child is to choose the alternative which exactly matches the standard. With few exceptions, the Matching Familiar Figures Test has been the method of choice for isolating impulsive subjects in research studies designed to examine the relative efficacy of various methods for modifying impulsive behavior. However, as several authors have noted, there are many methodological problems with the Matching Familiar Figures Test (e.g., Ault, Mitchell and Hartmann, 1976; Egeland and Weinberg, 1976), as well as broader conceptual problems with the impulsivity-reflectivity cognitive style dimension (e.g., Block, Block and Harrington, 1974). In addition there is also a body of psychological literature which strongly suggests that what is tapped by the Matching Familiar Figures Test may, in fact, not reflect those behaviors which teachers use as criterion measures






8

for classifying a child as being impulsive (e.g., Bjorklund and Butter, 1973; Vacc and Mercurio, 1978). Hence, the need for a scale specifically designed to tap the types of behaviors suggested as reflecting impulsive behaviors at home, in school, and in interpersonal situations. One such scale is the Impulsive Classroom Behavior Scale (Weinreich, 1975), a nine item teacher rating scale which was specifically developed for use in research involving the modification of impulsivity. The search for a more economical, less time consuming measure of impulsivity led to the development of the ISC. However, the review of the psychometric status of the scale suggests that we are not, at this point in time, able to state emphatically that the ISC is a reliable, valid measure of the construct of impulsivity. If its psychometric credibility could be established, then the ISC could hypothetically serve a multitude of purposes. In addition to allowing the practitioner to define observable impulsive behavior for the purposes of research and/or intervention, the focus of the scale on observable behaviors could have implications for delineating the specific foci of intervention plans.


Overview and Hypotheses

Except for the original study by Sutton-Smith and Rosenberg (1959), and the subsequent work by Hirschfield (1965), much of the research attempting to shed some light on the psychometric credibility of the ISC,






9

has not had the study of the scale's reliability and/or validity as its primary focus.

Given some of the problems with the ISC as it was originally developed, it was felt that the scale had to be revised prior to any attempts to establish the scale's reliability and validity. It was the establishment of the reliability and validity of this revised version of the scale which was given primary emphasis in this study. The scale's long term stability (ten weeks) was assessed, and factor analytic procedures, with their emphasis upon the delineation of the factorial composition of scales, was used to assess its basic structure. The relationship of the revised version of the scale (RISC) to the original ISC was also assessed by examining the correlation between scores on both scales, for a subset of the total population studied.

If the original version of the scale was used, one would predict that factor analysis would yield only one general factor. This is due to the scale's dichotomous rating categories, ambiguous items, and lack of situational specificity. In an effort to more clearly delineate the factor structure of the scale, and generate information relevant to the conceptualization of impulsivity, the original ISC was revised via the addition of twenty-nine marker items. These included items reflecting conduct problems, items reflecting behavioral self-control, and a set of six items randomly selected from the scale's original item pool, phrased in the negative for impulsivity. The






10

latter were included as a gross indicator of whether subjects were carefully attending to the task. The rating categories for the scale were also modified to allow the child a choice of five responses for each item.

Based upon these considerations, the following predictions with respect to the factor analytic data were made:

(1) upon analysis, an impulsive behavior factor would emerge, reflecting loss of affective and behavioral controls, poor planning,

and restlessness.

(2) in addition, a second factor would emerge reflecting behavioral

self-control.

(3) while there is some indication in the literature, that males score as being more impulsive than females on the ISC, it was predicted that, in terms of content, there would be no difference in the factor structure of the RISC for males and females.














CHAPTER II
METHOD

Subjects

The data for the study were obtained from a sample of 450

children, 230 females and 220 males, between the ages of nine and twelve. All were enrolled in either fourth, fifth, or sixth grade elementary school classes in Gainesville, Florida. This final subject pool was chosen from approximately 1100 children. The only criterion

for selection was the return of a signed parental consent form allowing the child to participate. Other than this criterion, the attempt was made to choose a subject population which, as closely as possible, approximated an unselected sample, in order to insure the presence of a high degree of variability on the impulsivity scale.

In Table 1, the mean age, and where available, the mean OtisLennon Mental Abilities Test IQ, and mean achievement test scores (i.e., total reading and total math standard scores from the Metropolitan Achievement Tests) are provided, for the total sample, and for the males and females separately. Includinq students from the P.K. Yonqe Laboratorv School.














TABLE 1

Mean Age, Otis-Lennon IQ, and Standard Scores From
the Metropolitan Achievement Test


Age


Metropolitan Achievement Test IQ Total Reading Total Math


Total Sample (N=450)


Males (N=220)


Females (N=230)


10.91 114.90 (SD=.93)*(SD=l5.22)

10.92 115.42 (SD=.94) (SD=15.99)

10.91 114.39 (SD=.93) (SD=14.45)


79.71
(SD=15.46)

79.42
(SD=l6.16)

79.99 (SD=14.79)


87.78
(SD=13.47)

87.50 (SD=14.59)

88.06
(SD=12.30)


SD - standard deviation







13

Instrument

The most contemporary version of the ISC (Bjorklund and Butter, 1973), containing twenty-five true-false items was used as the original version of the scale (see Appendix A). All items were keyed in the affirmative for impulsivity given Hirschfield's (1965) finding that impulsive children were no more likely than nonimpulsive children to respond in the affirmative to self-descriptive questionnaire items. With the addition of the twenty-nine marker items previously mentioned (see Appendix A), the total Revised Impulsivity Scale for Children (RISC) included 54 items, sequenced in random order. For the revised scale, subjects were asked to rate each item on a five point scale ranging from l=never happens, to 5=always happens.


Procedure

The RISC was group administered by the principal investigator to each individual class of subjects. At each administration, each child was given a copy of the scale and then asked to read the directions silently as they were read aloud to them. The directions used were as follows:

there will be fifty-four items on this questionnaire.
Read each item to yourself, as I read it aloud. Then, using the scale to the right of each item, decide how
often you engage in that particular behavior. A "I"
means never, a "3" means sometimes, and a "5" means
always. After you have decided upon your answer, circle
the appropriate number to the right of each item. (Appendix A)






14

The "2" and "4" response choices were also explained to the subjects in terms of their relationship to the other three possible choices. That is, subjects were told that:

a "2" response represented more than a "1" response
(never), but less than a "3" response (sometimes): a "4"
response represented more than a "3" response (sometimes),
but less than a "5" response (always). (Appendix A)

Items were read to the subjects in order to allow for the pacing of the scale administration. By allowing approximately 15-30 seconds per question, the total time required to administer the scale was approximately 15-20 minutes. The children were encouraged to answer each question as truthfully as they could. The questionnaires were numbered in order to maintain the subject's anonymity. They were also assured that their performance on the scale had nothing to do with their academic capabilities.

Following the administration of the RISC to the total subject pool, two sub-samples were selected. This was accomplished using the procedures outlined for generating random samples in the manual for the Statistical Package for the Social Sciences (SPSS: Nie, Hull, Jenkins, Steinbrenner and Bent, 1975). This process was used to generate two random samples of approximately 100 subjects each. Table 2 includes the mean age, IQ and achievement test scores for each total sub-sample as well as for the males and females within each sub-sample.














TABLE 2

Mean Age, Otis-Lennon IQ, and Standard Scores from the
Metropolitan Achievement Test for the
Reliability and Validity Sub-samples


Metropolitan Achievement Test


IQ Total Reading


Total Math


Reliability Sub-sample (N=92)


Males (N=41)


Females (N=51)


Validity Sub-sample (N=94)


Males (N=52)


Females (N=42)


10.90 * 115.55 (SD=1.03) (SD=13.64)


10.96 (SD=.96)

10.84 (SD=1 .09)


11.00
(SD=.88)

11.05 (SD=.90)

10.93
($D=.86)


115.00
(SD=14.51)

115.92 (SD=13.18)


116.34
(SD=14.39)

116.13 (SD=14. 77)

116.64 (SD=14. 10)


79.17
(SD=14.36)

77.76
(SD=15.48)

80.20 (SD=13.60)


83.00
(SD=12.47)

83.10
(SD=12.29)

82.86 (SD=12.94)


89.35
(SD=12.55)

89.28 (SD=14.78)

89.40 (SD=l0.86)


91.55
(S=l1 .82)

91.18 (SD=12.32)

92.07 (SD=ll .28)


SD - standard deviation






16

Subjects in the reliability sub-sample (N=92) were re-administered the RISC under the same conditions used previously, approximately ten weeks after its initial administration. Subjects in the validity subsample (N=94) were administered the original version of the ISC. Directions for this administration were as follows:

there will be twenty-five items on this questionnaire.
Read each item to yourself as I read it aloud. Then
decide whether it is true about you or false about
you. If it is true or mostly true, circle the T; if
it is false or not usually true about you, then circle
the F. (Appendix A)

The teachers of subjects in the validity sub-sample were also asked to rate each subject on a 1-11 scale of impulsivity (see Appendix B), where "1" indicated that the child's behavior was not at all characterized by impulsivity, and "11" indicated that the child's behavior was characterized by extreme impulsivity.


Factor Analyses

Given the exploratory nature of this investigation, and the suggestions in the literature that sex differences may represent an important variable in this area, the questionnaire data were analyzed for males and females separately.

The factor analyses were carried out in the following manner. A 54 X 54 pearson product moment correlation matrix was computed. This

matrix was then submitted to a principal factors analysis,






17

as outlined in the SPSS Manual (Nie et al., 1975). This mode of factor analysis employed an iteration procedure for improving communality estimates which were then used as the main diagonal elements of the correlation matrix. Rotation was carried out using the direct oblimin method of oblique rotation. The obliqueness criterion was set at zero. This procedure allowed for correlation amongst the generated factors but did not bias the analyses in this direction. The resulting rotated factor structures were then examined for the extent to which they approximated simple structure (i.e., most all of the loadings on a given factor were small, with only a few being of significant size; a given variable should have significant loadings on only one factor and, the pattern of loadings for a given factor

across all of the variables should be different from one factor to the next). Several guidelines for the interpretation of factor structures were used (Comrey, 1973). These included the criteria that for a loading to be significant, it had to be greater than �.40, and that at least six items had to load significantly on a given factor for it to be interpreted.














CHAPTER III
RESULTS

Factor Analyses

For both the males and the females in the subject pool, factor analysis was originally performed with no limits placed upon the number of factors to be extracted and then rotated. However, after this initial anlaysis, which yielded nineteen rotated factors for each group, it was decided to limit a subsequent analysis for each group to the extraction and rotation of five factors. It is this analysis for both the males and the females studied, which is reported here. This decision was based upon Comrey's (1973) discussion of factor extraction, and the fact that the percentage of variance accounted for by each successive factor beyond five, in conjunction with the

lack of significant loadings on those factors, suggested that it was highly unlikely that they would be interpretable. Males

The rotated factor structure for males yielded two interpretable

factors, accounting for 67.4% of the total variance of the questionnaire. These two factors were found to be negatively correlated (r=-.30) with no overlap of items having significant loadings. Items loading

18






19

significantly on these two factors are presented in Table 3. In addition, both the factor pattern matrix and the factor structure matrix for males are presented in Appendix C.

Factor 1 for males, accounted for 51.5% of the total variance. Items loading significantly on this factor were indicative of a

high energy level, restlessness, and episodic loss of affective control. This factor was labelled Impulsivity. Factor 2 for males accounted for 15.9% of the total variance. Items loading significantly on this factor were concerned with the thoughtful planning of activities, and self-directed behavior. Thus, this factor was

labelled Reflectivity.

Females

The rotated factor solution for the females in the subject pool also yielded two negatively correlated (r=-.24) interpretable factors which, in this case, accounted for 60.3% of the total variance. Items having significant loadings on these two factors are presented in Table 4; the rotated factor pattern and factor structure matrices are presented in Appendix C.

The interpretable factor structure for females less closely

approximated Comrey's (1973) criteria for simple structure, than did the factor structure for the males. There was no overlap of items loading significantly on the two factors, but the items loading






2(
TABLE 3 Significant Rotated Factor Loadings for Male Factors 1 and 2


Factor #1
Item #


Description


23 I like to wrestle and
horse around. 10 I am restless.


Factor Loadings
1 2 3 4 5


.53


-.10


.49 -.03


-.09 -.02 -.06 .30

.15 .11 .09 .28


24 I like to just blow off
steam.
11 I have trouble sitting
still.
36 I get so angry sometimes I
want to kick, scream and
throw things.
6 I like to keep moving
around.
38 I like to dare kids to
do things.
30 I must admit, I'm a pretty
good talker.
41 I usually say the first
thing that comes into
my head.
Factor #2
Item #


.48 .01 -.06


.46 .46


.43 .42 .42


.40 -.03


15 I finish all of my homework -.15
before I go out to play.
20 I usually plan what I am .06
going to do before I do it.
5 When I do my homework, I take -.003 my time to make sure I do
the best job I can.
52 I keep my room at home neat -.16
and clean.
49 I go to bed without being -.003
told that I have to.
8 I like to bother other child- .26
ren at school (negative).
17 In school, I speak only when -.23
I am spoken to by the
teacher.


-.03

-.05


.02

-.10


.08 -.02 .24


.26 .08


.15 .31


.01 -.08 -.10 .24


- .02

-.06


.13 -.03 .20


-.22


.07 -.08 -.1I


.02 -.06 .14 -.02


-.19

-.03


.51 .50 .50


.47

.44

-.42

.40


-.27 .31

-.01 .20 .003 .16





.05 .30 .01 .30 .04 .25


-.02 -.08 -.04 .26


.07 .19


.09 -.12 -.03 .27 .16 .09 -.02 .25








TABLE 4
Significant Rotated Factor Loadings for Female Factors 1 and 3


Description


I have trouble sitting still.


Factor Loadings 1 2 3 4 .49 -.04 .03 -.00


4 I like to show off in front of my friends.
23 I like to wrestle and
horse around.
21 I get angry quickly.

24 I like to just blow off
steam.
30 I must admit, I'm a
pretty good talker.


.47 .20


.45


.42 -.02


.41


-.04

-.19

.01


.03 .26


.04 .27 .32


-.17


-.06 .21


.12 -.01 -.02 -.01 .18


.40 -.01


-.33


.20 -.09 .32


Factor #2
I tem #


31 When things get quiet, I
like to stir up a little
fuss. (negative)
20 I usually plan what I am
going to do before I do
it.
38 I like to dare kids to do
things. (negative)
29 I can read for half an hour
or more without getting
up from my seat.
5 When I do my homework, I
take my time to make sure
I do the best job I can. 15 I finish all of my homework
before I go out to play.


.37 -.08 -.46


.13 .39


.05 -.01 .46 -.03 -.09 .22


.11 -.10


-.01


-.45 -.08


.11 .44 -.06


-.02 -.16 .44


-.17


-.14 .42


.21 .27 .27 .27


-.15 .36


.16 -.00 .26


Factor #1
Item #

11


.24









significantly were less factor pure than those for the males, in certain instances.

Factor 1 for females, accounting for 48% of the total variance was very similar to Factor 1 for the males, although not as robust in terms of the number of items loading significantly and the percentage of variance accounted for by the factor. It included items indicative of loss of emotional control, and restlessness, and again, would appear to be reflective of an underlying Impulsivity dimension. Factor 3 for females, accounting for 12.3% of the total variance also bore a great deal of content similarity to Factor 2 for the males. Items loading significantly were concerned with methodical responding and the maintenance of control, suggesting that it was also indicative of an underlying Reflectivity dimension.


Factor Scores

Having identified two fairly strong factors for the males and

two somewhat less robust but similar factors in terms of content for the females, factor score estimates were computed for each subject within each group. The method for computing factor scores was similar to the preferred procedure described by Comrey (1973), wherein items with significant loadings on a given factor are first standardized to the same mean and standard deviation, and then weighted based upon the loading of that item on the factor, prior to being summed to give







23

a total score. In this case, the item weightings represented an integral value approximately proportional to the factor loading of that item. The major advantage of this method for computing factor scores is that it allows those items having the highest loadings on

a given factor to have the greatest effect in estimating the factor scores.


Factor Score Reliability

The availability of a second set of questionnaire responses for a randomly selected sub-sample of the total subject population, allowed for an assessment of the reliability of the factor scores for both the Impulsivity and Reflectivity factors. Factor scores were computed for subjects in the reliability sample using their responses at the ten week re-administration of the RISC and the weighting coefficients derived from the factor analyses of the total samples of males and females. It was felt that the same weighting coefficients could be used for the reliability sub-sample because

t-tests comparing that group with the total sample indicated no significant differences in either age, IQ, or achievement test scores (see Table 5). As such, it was assumed that this sub-sample was representative of the larger group. Table 6 presents the reliability coefficients for the factor scores for the males and females separately.








TABLE 5

T-Test Comparisons of Reliability Sub-sample to Total Subject Pool


Variable N of Cases Mean Stand. Dev. Deg. of Freedom T Value 2-Tailed Probability for T-test


Total Sample 447 10.91 0.93 123.39 0.14 0.891

Reliability 92 10.90 1.03
Sub-sample

IQ
Total Sample 310 114.90 15.22
102.54 -0.34 0.732 Reliability 66 115.55 13.64
Sub-sample

TREAD*
Total Sample 316 79.71 15.46
105.32 0.28 0.783 Reliability 69 79.17 14.36
Sub-sample

TMATH**
Total Sample 316 87.78 13.47
105.03 -0.93 0.355 Reliability 69 89.35 12.55
Sub-sample
*TREAD - Standard score for the total reading portion of the Metropolitan Achievement Tests.
**TMATH - Standard score for the total math portion of the Metropolitan Achievement Tests.













TABLE 6

Reliability Coefficients for Factor Scores
For Males and Females


Sex
Males (N=41)


Females (N=51)


Factor Scores:

Impul si vi ty


Refl ecti vi ty


*p<.O01


. 72*


.68*


.67*


.64*








As can be seen, both the impulsivity and reflectivity dimensions were highly reliable for both the male and female subgroups within the reliability sub-sample. However, this finding must be tempered by the fact that the relatively small sample sizes may have acted to artificially inflate the correlation coefficients.


The Relationship of Factor Scores to Subject Variables

Factor scores were correlated with available demographic data for all subjects. Table 7 presents the resulting correlations for males and females.

For both the males and females respectively, the results

indicated that older subjects tended to be less reflective (Males: r=-.25, p<.Ol; Females: r=-.39, p<.OOl). However, only in the case of the females was there a corresponding increase in impulsivity for the older children (r=.23, p<.OOl).

An interesting finding involved the relationship between subjects' standard scores for the Total Reading and Total Math portions of the Metropolitan Achievement Tests. While there was no significant relationship found for either group, between factor scores

for the Impulsivity and Reflectivity dimensions and Otis-Lennon Mental Abilities Test IQ, in both groups it was found that less reflective subjects had higher achievement test scores (Males: Total Reading, 4=-.21, p<.Ol. Total Math, r=-.27, p<.91; Females: Total














TABLE 7

The Correlational Relationships of Factor
Scores to Subject Variables


MALES
(N=220)

Subject Variable Metropolitan Achievement Test
Total Reading Total Math Age IQ Standard Score Standard Score Factor Score:

Impulsivity .04 -.01 -.07 -.04

Reflectivity -.25 -.11 -.21* -.27"


p<.0l


FEMALES
(N=230)

Subject Variable Metropolitan Achievement Test
Total Reading Total Math Age IQ Standard Score Standard Score Factor Score:

Impulsivity .23* .04 .22"* .21"* Reflectivity -.39* .07 -.20** -.25*


*p<.001
**p<.Ol






28

Reading, r=-.20, p<.Ol; Total Math, r=-.25, p<.Ol). Furthermore, in the female subgroup, it was also found that better achievement test performance was also associated with higher scores on the Impulsivity dimension (Total Reading, r=.22, p<.Ol; Total Math, r=.21, p<.O1).

Given the fact that both dimensions were found to be negatively correlated with one another for males and females, it was decided to examine, on at least a descriptive level, the multiple correlation of factor scores on both dimensions with each of the available subject variables. This was accomplished using the multiple regression procedure outlined in the SPSS Manual (Nie et al., 1975). The results of these analyses are presented in Table 8. As can be seen, while for both males and females, the multiple correlation of both factors with age (males: .31; females: .35), Total Reading Score (males: .28; females: .25) and Total Math Score (males: .35; females: .27) from the Metropolitan Achievement Tests, are substantial, only in the case of "age" for the females did the factor scores taken jointly account for a substantial proportion of the variance (i.e., 20%) in a subject variable.


The Relationship of Factor Scores to ISC
Performance and Teacher Ratings of Impulsivity

The availability of a sub-sample of subjects who had completed the original version of the ISC and were also rated for impulsivity












TABLE 8

Multiple Correlations of Impulsivity and Reflectivity
Factor Scores with Subject Variables MALES (N=154)*


Adjusted


Standard


Multiple R R Square R Square Error Age .31 .10 .09 0.90339 IQ .12 .01 .00 15.98204 Metropolitan Achievement Test
Total Reading
Standard Score .28 .08 .07 15.44168
Total Math
Standard Score .35 .12 .11 13.57607

FEMALES
(N= 56)**
Adjusted Standard
Multiple R R Square R Square Error Age .45 .20 .19 0.82477 IQ .09 .01 .00 14.48784 Metropolitan Achievement Test
Total Reading
Standard Score .25 .06 .05 13.73290
Total Math
Standard Score .27 .07 .06 11.85801
*Represents that portion of the total sample of males (N=220) having data points for each of the variables included in the analyses.

**Represents that portion of the total sample of females (N=230) having data points for each of the variables included in the analyses.







30

by their classroom teachers allowed for an examination of the validity of factor scores based on the Impulsivity and Reflectivity dimensions, the relationship between RISC performance and performance on the original version of the ISC, and the validity of the original version of the ISC. T-tests comparing this validity sub-sample with the total sample, on the basis of age, IQ, and achievement test

scores found no significant differences on any of these variables except for the total Math portion of the Metropolitan Achievement Tests (see Table 9). Hence, it was felt that in general, findings pertaining to this sub-sample would be generalizable to the sample as a whole.

Table 10 presents the correlations between the Impulsivity and

Reflectivity factor scores, performance on the ISC, and teacher ratings of impulsivity, for males and females separately. Again, these findings must be interpreted with caution given the sample size limitations. Despite this fact, there were strong indications as to the validity of the factor scores for both males and females. This was based upon the observed relationship between factor scores on both dimensions, and teacher ratings of impulsive behavior. For both

groups, it was found that subjects scoring higher on the impulsivity dimension were also rated as being more impulsive by their teachers (Males: r=.29, p<.05; Females: r=.29, p<.05). In addition, male subjects with higher scores on the reflectivity dimension were also rated as being less impulsive by their teachers (r=-.32, p<.05). This relationship approached significance for females.









T-test Comparisons of


TABLE 9

Validity Sub-sample to Total Subject Pool


Variable N of Cases


Mean


Stand. Dev. Deg. of Freedom


2-Tailed Probability T Value for T-test


Age
Total Sample

Validity
Sub-sample

IQ
Total Sample

Val i di ty
Sub-sample TREAD*
Total Sample

Validity
Sub-samiple TMATH**
Total Sample


447


Val i di ty 67 Sub-sample
*TREAD - Standard score for
**TMATI- Standard score for


10.91


10.99 114.90 116.34



79.71 83.00 87.78 91.55


0.93 0. 88



15.22 14.39



15.46 12.47



13.47 11 .82


139.94


100.53


113.52


105.65


-0.84


-0.74


-1.88


-2.31


the total reading portion of the Metropolitan Achievement Tests. the total math portion of the Metropolitan Achievement Tests.


0.405


0. 462


0.063


0.023













TABLE 10

The Correlational Relationships of Factor Scores to ISC Performance and Teacher Ratings of Impulsivity


MALES (N=52) ISC Score Teacher Rating


Factor Score:
Impulsivity .58** .29* Reflectivity -.55"* -.32"

*p<.05
**p<.001

FEMALES (N=42) ISC Score Teacher Rating


Factor Score:
Impulsivity .67*** .29*

Refl ecti vi ty -.61*** -.24**

*p<.05
**p<.07
***P< .001







33

Since the RISC used as its base, the ISC as it was originally developed by Sutton-Smith and Rosenberg (1959), their relationship to one another was examined for subjects in the validity sub-sample. As can be seen from an examination of Table 9, the expected relationship of RISC scores on both dimensions to ISC scores was found for both males and females in the validity sub-sample. That is, males found to have higher factor scores on the Impulsivity dimension of the RISC were also found to have ISC scores more indicative of Impulsivity (r=.58, p<.O01). It was also found that more reflective subjects, based on RISC performance, had lower ISC scores (r=-.55, p<.O01). For the females, high RISC impulsivity factor scores were also associated with high ISC scores indicative of impulsivity (r=.67, p<.O01). In addition, reflective females, based on their RISC performance, were found to have lower ISC scores (r=-.61, p<.O01).


Validity of the ISC

As was previously mentioned, to date, attempts to validate the ISC have relied upon examining the correlation of ISC scale scores to either teacher ratings or teacher rankings of impulsivity. This relationship was examined in the present study for all of the subjects in the validity sub-sample. The resulting correlations indicated that while the ISC was a valid measure of impulsivity for the males in the sub-sample (r=.27, p<.05), and for the sub-sample as a whole






34

(r=.26, p<.Ol), it was far less adequate a measure of the construct for the females in this group (r=.05, p<.40).

As was the case with the RISC factor scores, ISC performance for subjects in the validity sub-sample were correlated with available demographic data. Table 11 presents the resulting correlations for males and females. As can be seen, none of these relationships approached significance for either group.













TABLE 11

The Correlational Relationships of ISC Performance
to Subject Variables


MALES
(N=52)
Metropolitan Achievement Test Total Reading Total Math Age IQ Standard Score Standard Score


ISC Performance .10 -.20 -.19 -.15




FEMALES
(N=42)
Metropolitan Achievement Test Total Reading Total Math Age IQ Standard Score Standard Score


ISC Performance .11 -.06 -.04 -.11














CHAPTER IV
DISCUSSION


The study of impulse control problems in children has been impeded by a lack of appreciation for the potential heterogeneity of this construct. This is despite the fact that impulsivity has been looked to as a major contributor to academic difficulties in children. The psychological literature to date has suggested that the

available methods for assessing impulsivity in children have not adequately tapped those aspects of the dimension which reflect, for

example, the types of behaviors used by teachers as criteria for labelling a child as being impulsive. The present study was undertaken in an effort to attempt to delineate, via questionnaire data, a potentially situation specific behavioral dimension of impulsivity.

The factor analysis of a revised self-report questionnaire designed to measure impulsivity, yielded two negatively correlated behavioral dimensions for both males and females. Both dimensions tended to be somewhat more robust for the males than for the females in the sample studied. The first dimension was characterized by a number of impulsive characteristics, and was named accordingly. The second dimension, indicative as it was of the pre-planning of activities,

36






37

methodical responding, and affective control, was named "Reflectivity" for both groups.

Further analyses indicated that, for a randomly selected subsample of the total subject pool, factor scores derived from subjects' questionnaire responses for each dimension, were highly reliable. Through the correlation of factor scores with teacher ratings of impulsivity, the validity of the factor scores for both dimensions was strongly supported for the males in another randomly selected subsample of subjects. The validity of the factor scores was also supported for the females in this group although less substantially for the Reflectivity factor, than was the case for the males.

The implications of these findings are many as are the possibilities for further research. However, prior to further interpretation, it must be noted that there are several methodological issues which serve to restrict the generalizability of these findings and hence, must be dealt with in subsequent research designed to refine this scale. First, as already mentioned, the relatively small number of subjects in the reliability and validity sub-samples act to limit the extent to which we can accept the findings pertaining to those groups. Further investigations attempting to examine the utility of the RISC should consider increasing the total sample size as well as the size of samples used for establishing the scale's reliability and validity. In addition, with respect to establishing the scale's






38

reliability, subsequent studies should consider examining its short term dependability (e.g., one week) as well as its long term stability (e.g., ten weeks).

Two related concerns involve the composition of the subject

pool. The attempt was made to, as closely as possible, approximate an unselected sample in order to enhance the probability of there being a wide degree of variability on the impulsivity dimension. In analyzing the results, the major subject variable attended to, apart from demographic variables such as age, IQ, and achievement test scores, was subjects' sex. Katzenmeyer and Stenner (1977), in their discussion of factor structure comparisons, pointed out the importance of attending to differences in factor structure across criterion groups within samples (i.e., with criterion groups based upon sex, race, etc.). The present study did not examine, at any level, the impact of race, and the associated variable of social class upon questionnaire performance and resulting factor structure. With respect to the latter, prior research has suggested that lower class children are in fact, more impulsive, as measured by the MFFT, than are middle class children (e.g., Heider, 1971; Schwebel, 1966). Race was not studied here primarily because of the relatively small number of non-white children in the total sample (N=51). This variable, as well as social class should, however, be examined within future attempts to refine this scale, and in so doing, possibly broaden its






39

applicability in terms of subject populations for which it could be of some use.

Related to this issue is the already touched upon issue of factor structure comparisons. Given the rather exploratory nature of this study, the decision was made to analyze the results for males and females separately. The resulting analyses indicated that while there was content similarity across the two dimensions for both groups, the factor structure for males was somewhat more robust in terms of the absolute number of items loading significantly on the factors, and the percentage of variance accounted for by the factors, than that for the females. The desire to broaden the applicability of the scale could argue for attempts to compare the factor structures across criterion groups in order to examine the degree to which they share common variance. The goal in carrying out such an analysis would be to see if it were possible to develop a single scale for each dimension which would have validity across criterion groups. However, because of the seminal nature of this study, this was left for future investigations, following the refinement of the factors in terms of the addition and deletion of scale items in order to increase the percentage of variance accounted for by the individual factors. Subsequent investigations may, at some point, attempt to examine the extent to which factor structures are comparable across different criterion groups (i.e., males vs. females; blacks vs. whites; or a






40

combination of the two). A number of empirical methods for establishing the invariance of factor structures have been reported (e.g., Harman, 1967; Rummel, 1970). On an empirical basis, because it provides an estimate of invariance with confidence intervals associated with those estimates, the coefficient of invariance described by Katzenmeyer and Stenner (1975, 1977) may be best suited for this purpose.

Given the attempt to examine the comparability of factor structures across criterion groups, the empirical question arises as to what, if anything, a composite scale (e.g., based upon male and female performance combined) would add to prediction to a particular criterion (e.g., in the case of impulsivity, possibly academic performance) over and above that which is achieved based upon a scale derived from an individual criterion group. This is a question open to empirical investigation through the use of multiple regression techniques.

Several potential uses for the impulsivity and reflectivity

scales derived from the factor analysis of RISC performance, immediately present themselves. Both scales would appear to have direct research applicability. Their ability to identify impulsive and reflective subjects for study should, in future investigations, allow for the further delineation of the correlates of this dimension..,The development of a reliable, valid method for identifying impulsive and reflective subjects is a necessary first step in any attempt to






41

conduct research as to the origins of these types of behaviors, the consequences of these types of behaviors, the types of environmental contingencies which could affect them, as well as uncovering possible mediating personality variables.

While the temptation is to hypothesize about direct clinical

applications for these scales, the present findings, especially with respect to the relationship of factor scores to achievement test data, suggest that at this point, such an attempt would be premature. Rather, these findings suggest a number of potential avenues for further research, particularly with respect to the examination of the relationship between RISC scores and a number of other environmental/organismic variables, several of which will be discussed below.

However, all of these are potential uses for the scales, which will depend for their achievement upon the refinement of the scales to the point where they could be considered reliable and valid measures of impulsivity and reflectivity. Of necessity, the next step in what may be viewed as a step-wise research program examining the utility of the RISC would be to refine the individual scales. The goal here would be to increase their reliability and validity, increase the percentage of variance accounted for by the individual factors, while also broadening the spectrum of behaviors accounted for by the dimensions.






42

The results of the present investigation indicate that a

number of the scale items, with negligible loadings on either of the interpretable factors, may be deleted during subsequent uses of the scale. In addition, given the quality of the items loading significantly on the two interpretable dimensions, it is also likely that new items could be constructed which would theoretically tap the same domain as those items being retained. These new items would include both general items indicative of impulsive and reflective behavior (e.g., item 10, "I am restless," for impulsivity; item 20, "I usually plan what I am going to do before I do it," for reflectivity), and more domain specific items such as those suggesting impulsive and reflective behavior in the classroom (e.g., item 5, "When I do my homework, Itake my time to make sure that I do the best job that I can," for reflectivity). The inclusion of the latter type of item stems from the observation that, particularly with respect to those items loading significantly for the males in the total subject pool, several were indicative of behaviors specific to academic settings and activities.

The prospect of developing domain specific items merits further attention. It may be that in future investigations it will be possible to tap impulsive and reflective behavior in a number of different situations. These may include situations in the home and in interpersonal settings as well as those in school. To the extent that






43

this was achieved, it might then be possible to develop what would amount to an impulsivity-reflectivity profile for a given child. This would address itself to the manner and extent to which these types of behaviors manifest themselves across a wide variety of situations. There is in fact some theoretical precedent for considering this approach. Block et al. (1974), in their investigation into the utility of a measure of cognitive impulsivity, suggested that the impulsivity dimension may be sufficiently broad to require a number of assessment tools, each sensitive to the impact of specific situational variables. This may be extended to include the impact of organismic variables on the quality and quantity of impulsive and reflective behavior. In both cases, the specific variables, and their potential impact are open to empirical investigation.

In discussing methods for refining the scales, it is necessary to deal with the somewhat perplexing finding that for both males and females in this study, higher achievement test scores were associated with low reflectivity factor scores. Furthermore, for females, higher achievement test scores were also associated with high impulsivity factor scores. This is of interest in light of the fact that a number of prior investigators (e.g., Douglas, 1974) have maintained that for certain populations (e.g., hyperactive children), impulsivity was associated with academic performance deficits. In






4A

addition, the thrust of the impulsivity modification literature (E.G., 'leichenhaum and Goodman, 1971) has been to reduce impulsive responding and increase reflective responding in an effort to increase academic Performance levels.

To this point in the Psychological literature, the majority of the research attempting to relate impulsivity to academic performance deficits have focused on selected subject populations, such as hyperactive children. The focus of the present investigation upon the performance of normal children, highlights the need for extreme caution in terms of extending the conceptualization of impulsivity based on one population to a qualitatively different population of subjects. 4hat this finding may in fact suogest, is that imoulsivity represents an individual difference dimension wherein its manifestations and the variables affecting it, differ across various oooulations.

Emohasizing the concept of impulsivity as an individual difference dimension, one possible explanation for this findinq refers to the possibility that impulsivity and reflectivity as they are tapped by the RISC, represent domains apart from the types of behaviors necessary for adequate functioning on cognitive tasks such as the 'Ietropolitan Achievement Tests. The fact that scores on both dimensions correlate significantly with teacher ratings of impulsivity suggests that the dimensions may be interoretable as being indicative of management or conduct problems which do not necessarily have to have aversive consequences for academic performance in all children. One method for examining this possibility would be to investigate the relationship between performance on the RISC impulsivity and reflectivity dimensions and the child's status on the Behavior Problem Checklist (Quay and Peterson, 1967). In particular, an examination of the relationship






45

between scores on both RISC dimensions, and placement on the Behavior Problem Checklist conduct problem dimension could potentially give support to the hypothesis that, at least for some children, impulsivity and reflectivity as measured by the RISC may reflect management/conduct problems not necessarily related to academic performance.

That body of psychological literature attempting to relate behavioral impulsivity and cognitive impulsivity as measured by Kagan's Matching Familiar Figures Test (Kagan et al., 1964) suggests another possible explanation for this finding. A number of investigators have examined the relationship between these two variables (e.g., Ault, Crawford and Jeffrey, 1972; Bentler and McClain, 1976; Bjorklund and Butter, 1973; Nadeau, 1968). In general, the results of these

investigations have suggested that cognitive impulsivity, at least as it is measured by the Matching Familiar Figures Test was not related to a broader, more generalized tendency towards impulsive behavior. While these results are not conclusive, particularly because of methodological problems with the Matching Familiar Figures Test, and more generalized problems with the impulsivity/reflectivity cognitive style dimension, they do suggest that it is possible for a child to be behaviorally impulsive but not cognitively impulsive, or vice versa. To the extent that, as Kagan and his associates have maintained, cognitive impulsivity relates to performance on cognitive






46

tasks, the outcome mentioned above could account for the present finding. In any case, a subsequent investigation using the RISC could begin to generate data to respond to this hypothesis by administering the RISC and the Matching Familiar Figures Test (and possibly the Behavior Problem Checklist) to a sufficiently large population of subjects, and then correlating the results.

There is yet a third possible explanation for these findings. It may be that the children included in this study were not very accurate self-reporters due to confusion with regards to how the items had to be responded to, or a lack of attention to item content. This possibility receives some support from an examination of those items included in the scale, which had both positive and negative forms. The correlation of comparable items yielded the expected significant negative correlation in only three instances out of six for the males, and in only two instances out of six for the females. Given Hirshfield's (1965) empirically supported contention that the use of a negative form of a normally positive item on the ISC does not significantly increase its cognitive complexity, one could hypothesize that these findings suggest a lack of attention to item content, a lack of motivation, or confusion. However, the magnitude of the factor loadings, and the overall consistency in the data, cast doubt upon this latter possibility.






47

A more plausible explanation, again related to the issue of

accuracy in the subjects' self-reporting, relates to the possibility that there exists, a discrepancy between the childrens' perceptions of their own behavior, and others' perceptions of the childrens' behavior. That is, clinically, it is not uncommon to see a fair amount of dissonance between a child's self-reported behavior, and that of adults rating the child's behavior.

One method for testing out this possibility would be to administer the scale, in a subsequent study, not only to another sample of children, but also to their teachers. The teachers would be asked to complete the questionnaire based on how they felt the children

would respond if their perceptions of their own behavior were objectively accurate. Additionally, the teachers could be asked to complete a versions of the RISC which would use the same items in terms of content, but which would assess the teachers' perceptions of the children's behavior.



In summary, the RISC was developed as a self-report instrument for measuring a broad dimension of impulsive behavior. The results of the present investigation yielded two behavioral dimensions, as measured by the RISC, for both males and females: an "impulsivity" dimension and a "reflectivity" dimension. While there are clear






48

implications for the use of such scales, much additional work is needed to replicate the findings of the present study, as well as to further refine both the impulsivity and reflectivity dimensions.














APPENDIX A
REVISED IMPULSIVITY SCALE FOR CHILDREN (RISC)

NAME

There will be fifty-four items on this questionnaire. Read each item to yourself, as I read it aloud. Then, using the scale to the right of each item, decide how often you engage in that particular behavior. A "l" means never, a "3" means sometimes, and a "5" means always. After you have decided upon your answer, circle the appropriate number, to the right of each item.


*1. I don't think I am as happy as
other people.

*2. I play hooky sometimes.

*3. I'm not known as a hard and
steady worker.

***4. I like to show off in front of my friends.

****5. When I do my homework, I take my
time to make sure that I do the
best job that I can.

*6. I like to keep moving around.

***7. I don't like going to school.

***8. I like to bother other children at school.

****9. I usually have enough time to do
everything I want to do everyday.

*10. I am restless.


Never
1


Sometimes
2 3


Always
4 5


1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5









Never
I


***lI. I have trouble sitting still.

**12. I don't like being "it" when
we play games of that sort.

**13. I don't like to shoot with bows
and arrows.

****14. My school grades are important to
me.

****15. I finish all of my homework before
I go out to play.

***16. I don't like to tell on my friends when they misbehave.

****17. In school, I speak only when I am
spoken to by the teacher.

*18. Whenever there is a fire-engine
going someplace, I like to follow
it.

*19. Sometimes, I can hardly stop from
throwing snowballs at people I
see walking by.

****20. I usually plan what I am going to
do, before I do it.

***21. I get angry quickly.

***22. I often break things that belong
to me and my friends.

*23. I like to wrestle and horse
around.

*24. I like to just blow off steam.


*25. I'll try anything.


Sometimes
2 3


Always
4 5


1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5










*26. It's hard to stick to the rules
if you're losing the game.

****27. I don't have to be reminded to
do my chores at home.

*28. I often act on the spur of the
moment, without stopping to think.

****29. I can read for half an hour or
an hour without getting up from
my seat.

*30. I must admit, I'm a pretty good
talker.

*31. When things get quiet, I like to
stir up a little fuss.

*32. My home life is not always happy.

***33. I don't like it when the teacher pays attention to other children
and not me.

*34. I don't think you should always
have to do what you are told.

*35. I like to go with lots of other
kids, not just one.

***36. I get so angry sometimes, that I want to kick, scream, and throw
things.

*37. I get into tricks at Halloween.

*38. I like to dare kids to do things.

****39. In class, I make sure I know the
answer before I raise my hand to
answer a question.


Never
1


Sometimes
2 3


Always
4 5


1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5










*40. I like throwing stones at targets.

*41. I usually say the first thing
that comes into my head.

*42. I like being "it" when we play
games of that sort.

**43. I never play "hooky".

**44. I am not restless.

*45. I make friends quickly.

***46. I don't like to listen to the teacher when she tells me what
to do.

*47. I like to shoot with bows and
arrows.

**48. I think I am as happy as other
people.

****49. I go to bed without being told
that I have to.

****50. When I have to, I take turns
at school.

*51. It's fun to push people off
the edge, into the pool.

****52. I keep my room at home, neat
and clean.

***53. I use words that my parents say
I shouldn't use.

**54. I don't like to dare kids to do things.


Never


Sometimes
3 3


Always
5

5


1 2 3 4 5


1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5






53


* Items from the original Impulsivity Scale for Children (ISC).
** Behavioral self-control marker items.
*** Conduct problem marker items.
**** Market items keyed in the negative.







54
IMPULSIVITY SCALE FOR CHILDREN


Name

There will be twenty-five items on this questionnaire. Read each item to yourself as I read it aloud. Then, decide whether it is true about you or false about you. If it is true or mostly true, circle the "T"; if it is false or not usually true about you, then circle the "F". 1. I like to dare kids to do things. T F 2. I am restless. T F 3. I play hooky sometimes. T F 4. I like to wrestle and horse around. T F 5. I must admit, I'm a pretty good talker. T F 6. I like to shoot with bows and arrows. T F 7. I like throwing stones at targets. T F 8. My home life is not always happy. T F 9. When things get quiet, I like to stir up T F
a little fuss.

10. I make friends quickly. T F 11. I like to just blow off steam. T F 12. I don't think I am as happy as other people. T F 13. I'll try anything. T F 14. I'm not known as a hard and steady worker. T F 15. It's fun to push people off the edge, into T F
the pool.

16. Sometimes, I can hardly stop from throwing T F
snowballs at people I see walking by.

17. I usually say the first thing that comes into T F
my head.






55

18. I like to keep moving around. T F 19. I like to go with lots of other kids, T F
not just one.

20. I don't think you should always have to do T F
what you are told.

21. I get into tricks at halloween. T F 22. I often act on the spur of the moment, T F
without stopping to think.

23. It's hard to stick to the rules, if you're T F
losing the game.

24. I like being "it", when we play games T F
of that sort.

25. Whenever there is a fire engine going T F
someplace, I like to follow it.















APPENDIX B
TEACHER RATING SCALE FOR IMPULSIVITY


DATE:

TEACHER'S NAME:


Impulsivity in children is the tendency of a child to respond quickly and without reflection or delay, to be restless, to indulge in horseplay, lose control of his feelings, break the rules, enter activities with overwhelming vigor, and generally to lose control in acceptable or unacceptable ways.

Using the scale below, rate each child listed in terms of the degree to which his/her everyday classroom behavior reflects impulsivity as it is defined above.


Child's Name:
1.
2.
3.
4.
5.
6.
7.
8.
9.
10. 11.
12.
13.
14. 15. 16.
17. 18. 19. 20.


Not at all
impulsive
1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2


Extremely
impulsive
11 11 11
II 11 11 11 11 11 I1 11 11 11 11 11 11 11 I1 11 1I















APPENDIX C
FACTOR MATRICES FOR MALES AND FEMALES

Factor Pattern For Males


Item #

1

2

3

4

5

6

7

8

9 10 11

12


Factor 1

-0.02174 0.26010

0.11762 0.24038

-0.00253 0.43490

-0.13078 0.26286 0.00039 0.49112 0.46231 0.06102 0.11216 0.02005

-0.15365 0.02974

-0.22714 0.29160

0.17188 0.056 36 0. 331 72 0.20261


Factor 2

0.04875

-0.00859 0.08017

-0.24888 0.50008 0.01690

-0.19421

-0.42288 0.29109

-0.03184

-0.03068

-0.11236 0.05513 0.33296 0.51213

-0.15976 0.40387

-0. 00344 0. 15306 0.50440

-0.21581 0.26601


Factor 3

0.47510 0.08655 0.37175

-0.10774

-0. 02558

-0.00160 0.05836 0.09428

-0. 22662 0.14706 0.25531

-0. 00551

-0.02528 0.03745 0.14181 0.01960 0. 15824

-0.10436 0. 17556

-0.19497 0.08082

0.06989


Factor 4

0.04602

-0.25755

-0.21757 0.13047 0.02297 0.13434

-0.08483

-0.12127 0.04382 0.11250 0.08362 0.03489

-0.21775

-0.13913

-0.02094 0.44739 0.09380

-0.14209

-0.02634 0.10535
-0.19006 0.02925


Factor 5

0.00751

-0.06270 0.07228

-0.02529 0.03539

-0.03269

-0.32717

-0.02616 0.09067 0.09450 0.15493

-0.09116 0.42177

-0.03135 0.04737

-0.06134
-0.01848

-0.38111

-0.39621 0.01228 0. 34984 0.06996








Item #

23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50


Factor 1

0.53136 0.48098 0.33423

0.36732 0.07419 0. 32436

-0.04434 0.42219

0.37587

-0. 00853 0.19610

0.06216 0. 30899 0.45583 0.19202 0.42293 0.02145 0.35151 0.40296

-0.06027 0.03116 0. 02809 0. 11630

-0.07761 0.03712 0.09015

-0.00284
-0. 04304


Factor 2

-0. 09763 0. 01455

-0.05066

-0.066 18

-0.02083

-0.10206

0.13231 0.06903

-0.23135

-0.15921

-0.02988

0.02197 0.06324

-0.05452 0.05881

-0.10468 0.32615

-0.05585

-0.03045 0.04107

-0.24630

-0. 01589 0.31775

-0.18657

-0.02814 0.24435 0.43911 0. 31530


Factor 3

-0. 09488

-0.05593 0.01142 0.12873

-0.15589

-0.01714

-0.10958

-0.08074

-0.03494 0.33270 0.27329

0. 34858 0. 14775 0.00868 0.04166

-0.06467 0.05805

-0.08420

0.02161

-0.16050

-0.47267

0. 04490

-0.18037 0.05459

-0.07711

-0.54022

0.06776
-0.07724


Factor 4

-0.01852

0.07640 0.04504

-0.18250

-0.17855

-0.07890 0.30207

-0.11046

-0.19968 0.33198

-0.03452

-0.03398 0.04947

-0.07544

-0.19690

-0.21613 0.18946

-0.08516

-0.05601 0.08953 0.33403 0.05349

-0.11379

-0. 39604 0.03280

-0.03480 0.04405

0.30707


0.20514 -0.14024 -0.05028 -0.28720 -0.38801


Factor 5

-0.06259

-0.01506

-0.17866

-0.03761 0.06318 0.05960

0.08051

-0.01078

-0.23199

-0.04522 0.10450

-0.13750

-0.09109

-0. 10362

-0. 50649

-0.27185 0.09677

-0.07119 0.00320 0.11320

-0.02472

-0.06839

-0. 24970 0. 14786

-0.16639

-0.09972 0.06587
0.13467








Factor 1

-0.16499 0. 241 74
-0.01298


Factor 2

0. 46950

-0.15164

0.00314


59
Factor 3

-0.01631
-0.07950

-0.07939


Factor Structure for Males


Item #
1

2 3 4 5 6 7 8 9
10 11
12 13 14 15 16 17 18 19 20 21


0.26600 -0.32219 0.11409 -0.02499 -0.01429


Item #

52 53 54


Factor 4

-0. 08206
-0.10096 0.30777


Factor 5
-0.03659

-0.25445 0.08433


Factor 1 Factor 2 Factor 3

0.00063 0.01378 0.46640 0.33975 -0.11673 0.12446 0.15824 0.00799 0.38453 0.28330 -0.30700 -0.06494
-0.16803 0.50925 -0.07423 0.41018 -0.11019 0.03569 0.03111 -0.20926 0.06846 0.43015 -0.52036 0.16597
-0.14093 0.32696 -0.25631 0.46924 -0.17382 0.19494 0.44212 -0.16767 0.30102 0.10932 -0.14051 0.01035 0.03398 0.06923 -0.01239
-0.03943 0.31178 0.01376
-0.29971 0.54992 0.07840 0.00282 -0.15490 0.01999 0.34649 0.45932 0.09330 0.40497 -0.13969 -0.06685 0.24686 0.03027 0.18152 0.13883 0.51327 -0.24125 0.35732 -0.28560 0.14087


Factor 4

0.03440
-0.31892
-0. 24794 0.07067 0.05339 0.04409
-0.09093
-0.20316 0.07401 0.00998

-0.01356 0.01097
-0.21073

-0.12892 0. 03511 0.42818 0. 15442
-0.22142
-0.08497 0.12932
-0.25126


Factor 5

0.01996
-0.14421 0. 03889
-0.10941 0. 10496
-0.12914
-0.32668

-0.15580 0. 13362
-0.02442 0.04094
-0.11914 0. 38827
-0.00018 0.15223

-0.06265 0. 09688
-0.46161
-0.42036 0.07386 0. 22698









Item #

23 24 25 26 27 28 29 30

31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50


Factor 1

0.57012 0. 45899 0.38527 0. 44686 0.08567 0. 35465

-0.17678 0.41867 0.53952 0. 01 588 0.21414 0.13178 0.31736 0.51398 0.34370 0.55888

-0.13282 0. 39464 0. 42496

-0.13507

-0.00563 0. 04326 0.08781 0.02854 0.07192

-0.00605

-0.15246

-0.24125


Factor 2

-0.25692

-0.12189

-0.17332

-0.20296

-0.02928

-0.19358 0.18289

-0.05688

-0.38232

-0.17650

-0.10210

-0.04992

-0.05274

-0.20959

-0.08131

-0.27316 0.33740

-0.16710

-0.15550 0.09430
-0.19642

-0.03489 0.26031

-0.16974

-0.05262 0.25319 0.44478 0.37001


Factor 3

-0.03066

-0.01148

-0.04920 0.17982

-0.13950 0. 02 833

-0.13916

-0.03990

0.03427 0.33373 0.29691 0.35485

0.17159 0.06366 0. 06600

-0.001 80 0.02127

-0.03945 0. 06764

-0.17466
-0. 45956 0. 04743

-0. 19281 0.07960

-0. 07119
-0.55228 0.02385

-0.12446


Factor 4

-0.13276

-0.02015

-0.03771

-0. 26888

-0.18470

-0.14650 0.32761

-0.19078

-0. 30204

0. 30903

-0.08089

-0.06810

-0.02211

-0.17859

-0.26611

-0.32265 0. 20610

-0.16124

-0.14096 0.11754 0.33202

0. 04084

-0.12884

-0. 38314

0.01651

-0.02464 0.06941

0.34416


Factor 5

-0.20720

-0.12648

-0.26501

-0.14883

-0.03184

-0.03877 0. 12845

-0.11180

-0. 36784

-0.04569 0.04875

-0.15363

-0.15628

-0. 22785

-0.55819

-0.40308 0. 14687

-0.17007

-0.10364 0. 13988

-0.04265

-0. 07436

-0.24165 0.11709

-0.17692

-0.08849

0. 12815 0. 20711









Factor 2

-0.26443 0.51105
-0.25607 0.04231


Factor 3

-0. 00272
-0.07397

-0.03531
-0.09382


Factor 4

-0. 35870

-0.02477
-0.17115 0.31901


Factor Pattern for Females


Item #
1
2 3 4 5 6 7 8 9 10 11
12 13 14 15 16 17 18 19 20 21


Factor 1
0.22870
-0.02815 0.01363 0. 47484
-0.01661 0. 36791 0. 07200 0.30574 0. 02766 0.33825 0. 48894 0.11286
-0.11029
-0.01317

-0.16870 0.09125
-0.19453 0.06282 0.03846 0.04985 0.42109


Factor 2
-0.20494 0. 02032
-0. 33632 0.19541
-0.15741
-0.06327 0.06780
-0.14309 0. 08821
-0.03231
-0.04213
0.05195 0.29965
0.07723
-0.13633 0. 34882
-0.19878
-0.19936
-0.10152

-0.00983
-0.01846


Item #
51 52 53 54


Factor 1
0.39636
-0.28115

0.36233
-0.10585


Factor 5
-0. 47490

0.06231
-0. 34023
0.10734


Factor 3

0.07888
-0. 24397

-0.08022
-0.03928 0. 44032 0.07616
-0.15109

-0.10053 0.10477

0.08035 0.03491 0.04171 0.16941 0.16441 0.42459
-0.02760
0. 26432

-0.05332 0.02712 0.45713 0.00675


Factor 4

0.03035
-0.01009
-0.01897 0.02818 0. 34878 0.06403
-0.11551
-0.12740 0.07710
-0.20033 0.00057
-0. 18852
-0.06370 0.35234 0.16490
-0.19549
-0.15730
-0.00173 0. 09804
-0.02880
-0.17482


Factor 5
-0.04825

-0.08177
0.06419 0.02969
-0.14730 0.05123

0.13544 0.06352 0.14193 0.07644 0.03334
0.12794 0.18790

-0.16462
-0.00067 0.15443
0.09939 0.25507 0.50142
-0.08535
-0.06196









Factor 4 Factor 5


I tem

22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49


Factor 1

0. 24004 0.44995 0.41075 0.25270 0.19763 0.01658 0. 30072

-0.00655 0.39638 0. 369 75 0.27889 0.25271 0.06124 0. 163T5

0.32959

-0.10587 0.11465 0. 04470 0.03959 0.21285

-0.10521 0.16601 0.02899 0.02137 0. 16764

-0.05027

-0.11510

-0.12157


Factor 2

-0.23041

0. 10895 0.12157 0.08886

-0.29771

-0.07086 0. 12232 0. 10841

-0.01433

-0.08341

-0.24774

-0.06908

-0.36717

-0.16305

-0.14391

-0.25296

-0.09891 0. 1 8052 0.03773

-0.24037

0.10749 0. 58980 0. 17892
-0.09568

-0. 33978 0. 14346 0.22291

-0.09626


Factor 3

-0.00367

-0.18837

-0.00515

-0.29003

-0.31339

-0.03879

-0.24378

0. 44263

-0.32777

-0.45874 0.04431 0.18417 0.05573 0.04185

-0.10878

-0.37035

-0.45154 0. 38456 0.07425

-0.15579 0.12345

-0.07511

-0.12364

0.04449

-0.10698

-0.00772

-0.07934 0.12953


-0.04469 0.03524

-0.02312 0. 15000 0.02788 0.11885

-0.18821

-0.06201 0. 20070 0.11980

-0.34365

-0.02737

-0.09626 0.21717

-0.01567 0.02475

0.07546 0.06187 0.00321

-0.11641 0. 13844

-0.02232 0.19710 0.45226

-0.12603

-0.21177 0.41454 0.06900


0. 16887 0.26897

-0.00871 0. 10849

-0.00058 0.11606

0.16435 0.26955

-0.09225

0.13108 0.06823

-0.07742

0. 24827

-0.10134 0.07271 0.23111 0. 20914 0. 06836 0. 58375

0.06900 0. 16438 0.02092 0. 10805 0. 13365 0.08866 0.43337 0.05362 0.04075








Factor 1

0.06989
0.01391

-0.14165 0. 18993 0.02922


Factor 2

0.20827

-0.01532
-0.22973
0.03448 0. 42750


Factor 3

0.27351

-0.29909 0.35125
-0. 35369
-0.00357


Factor 4

0.16287

-0.06231 0.17379 0.12509
-0.02530


Factor Structure for Females


Factor 1
0.22252 0.01139 0.09569 0.46121
-0.17928 0.36160 0. 16305 0.37981 0.01278
0.36507 0.49375 0.14792
-0. 14179
-0.14461

-0. 2744 0.10988
-0.23526 0. 16090
0.14717
-0.07625


Factor 2
-0.23269

-0.00167
-0.33975 0.13162
-0.11994
-0.10056
-0. 08247

-0.19196 0.10721
-0.07354
-0.10298 0. 04301 0. 34056 0.09183
-0.07458 0.33941
-0.16657
-0.19098
-0.05944 0.01021


Factor 3

0.01801
-0. 22350

-0.121 36
-0.14227 0.48993
-0.02044

-0. 20675
-0.20874 0.08968
-0.03652
-0.09310
-0.02071 0. 18309 0.23425 0.47161
-0.06694 0.29707
-0.12474

-0.06036 0. 45498


Factor 4

0.00309
-0.02748

-0.04343
-0.02743 0.39171 0.02219

-0.14549
-0.18231 0.08357
-0.23756
-0.06007
-0.19979
-0.02683 0.37753 0.22194
-0.1)9994
-0. 11254
-0.02940 0.07843 0.01169


Item #

50 51 52 53 54


Factor 5

0.06498 0. 46464

-0.15895 0. 28344 0.06866


Item #

1 2 3 4 5 6 7 8 9 10 11
12 13 14 15 16 17 18 19 20


Factor 5

-0.02669
-0.04732 0.05284 0. 15897 0.24363 0.11556 0.17337 0.10450 0. 13670
0.14307 0.13518 0.15626 0.16233

-0.19695
-0.12229 0.21356
-0.19777 0. 26142 0.49480
-0.14694








Factor 3 Factor 4 Factor 5


Item #

21 22 23 24 25 26 27 28 29 30 31 32 33 34
35 36 37 38 39 40

41 42 43 44 45 46 47 48


Factor 1

0. 42948 0.31596 0. 53782 0.39641 0.31757 0.31093 0. 04730 0. 40418

-0.06057 0.43268 0.50802

0.35953 0.20292

0.16556 0.12522 0.39408 0.06860 0.27643
-0.06582 0.14831 0.31349
-0.12975 0.11194 0.03511
-0.00207 0. 27546 0.05658
-0.16495


Factor 2

-0.08697
-0.25104 0.05714 0. 06390 0.04726

-0.34942
-0.06184 0.06910 0.16296
-0.09285
-0.15309

-0.28999

-0.09669
-0.35450
-0. 18180

-0.19151
-0.24691
-0.12877 0.21195 0.08632

-0.28000 0.15018 0. 56240 0.18217
-0 .(7Y6641
-0.36831 0.17712 0.25312


-0.10473
-0.11121
-0.32934
-0.09704

-0.34757

-0.38161
-0.05513
-0.35248 0. 40349
-0.39133
-0.56459

-0.08725 0.12683
-0.03631 0.02694

-0.21341
-0.39826
-0.51305 0. 38261
-0.02526
-0.24855

0.14472
-0.07606
-0.11497 0.05490
-0.20037
-0.07426
-0.00216


-0.22520
-0.08823
-0.041 73

-0.06933 0.09098
-0.03884 0.10710
-0.24883
-0.02092 0.12162 0.02246

-0.38530
-0.04117
-0.11947

0.19756

-0.07461
-0.01457 0.00761
0.09921
-0. 001'.68

-0.16922 0.16329
-0.02760 0. 1 8553 0. 44661
-0.17291
-0.21238 0.42828


0.03563 0. 20608 0.40905 0.09590 0. 21530 0. 06886 0.11695 0.28655
0.20814 0. 04326 0.27791 0.11323
-0.05444
0.22558

-0.09033
0.15333 0. 24460
0.29690 0.03043 0.58392 0.12541

0.12598 0.11979 0.14377 0.11130 0. 11909 0.44075 0.04737






65

Item # Factor 1 Factor 2 Factor 3 Factor 4 Factor 5
49 -0.13944 -0.06381 0.15211 0.09181 -0.01726 50 -0.03050 0.23147 0.27795 0.18741 0.05002 51 0.20208 -0.00420 -0.38381 -0.10643 0.51588 52 -0.25381 -0.18996 0.41061 0.22092 -0.27069 53 0.32046 0.00997 -0.43038 0.06074 0.38234 54 -0.00914 0.42794 0.00858 -0.01434 0.11180















REFERENCES


Ault, R.L., Crawford, D.E., and Jeffrey, W.E. Visual scanning
strategies of reflective, impulsive, fast-accurate and slowinaccurate children on the Matching Familiar Figures Test.
Child Development, 1972, 43, 1412-1417.

Ault, R.L., Mitchell, C., and Hartmann, D.P. Some methodological
problems in reflection-impulsivity research. Child Development,
1976, 47, 227-231.

Bentler, P.M., and McClain, J. A multitrait-multimethod analysis of
reflection-impulsivity. Child Development, 1976, 47, 218-226.

Bjorklund, D.F., and Butter, E.J. Can cognitive impulsivity be
predicted from classroom behavior? The Journal of Genetic
Psychology, 1973, 123, 185-194.

Block, J., Block, J.H. and Harrington, D.M. Some misgivings about
the Matching Familiar Figures Test as a measure of reflectionimpulsivity. Developmental Psychology, 1974, 10, 611-632.

Comrey, A.L. A first course in factor analysis. New York:
Academic Press, 1973.

Douglas, V.I. Sustained attention and impulse control: Implications
for the handicapped child. In J.A. Swets and L.L. Elliot (Eds.)
Psychology and the handicapped child. Washington, D.C.: U.S.
Department of Health, Education and Welfare, DHEW Publication
Number (OE) 73-05000, 1974.

Egeland, B. Training impulsive children in the use of more efficient
scanning techniques. Child Development, 1974, 45, 165-171.

Egeland, B., and Weinberg, R.A. The Matching Familiar Figures Test:
A look at its psychometric credibility. Child Development, 1976,
47, 483-491.

Harman, H. H. Modern factor analysis. Chicago: The University of
Chicago Press, 1967.









Heider, E. R. Information processing and the modification of
an "impulsive conceptual tempo". Child Development, 1971, 42,
1276-1281.

Hirschfield, P. Response set in impulsive children. The Journal
of Genetic Psychology, 1965, 107, 117-126.

Kagan, J. Individual differences in the resolution of response
uncertainty. Journal of Personality and Social Psychology,
1965, 2, 154-160. (a)

Kagan, J. Reflection-impulsivity and reading ability in primary
grade children. Child Development, 1965, 36, 609-628. (b)

Kagan, J. Impulsive and reflective children: Significance of
conceptual tempo. in J.D. Krumboltz (Ed.), Learning and
the educational process. Chicago: Rand McNally, 1965. (c)

Kagan, J., Rosman, B. L., Day, D., Albert, J., and Phillips, W.
Information processing in the child: Significance of analytic
and reflective attitudes. Psychological Monographs, 1964,
78(0), Whole No. 578.

Katzenmeyer, W. G. and Stenner, A. J. Strategic use of random subsample replication and a coefficient of factor replicability.
Education and Psychological Measurement, 1975, 35, 19-29.

Katzenmeyer, W. G. and Stenner, A. J. Estimation of the invariance
of factor structures across sex and race with implications for
hypothesis testing. Education and Psychological Measurement,
1977, 37, 111-119.

Kendall, P. C., and Finch, A. J., Jr. A cognitive-behavioral
treatment program for impulse control: A case study. Journal
of Consulting and Clinical Psychology, 1976, 44, 852-857.

Kendall, P.C., and Finch, A.J., Jr. A cognitive-behavioral treatment
for impulsivity: A group comparison study. Journal of Consulting and Clinical Psychology, 1978, 46(1), 110-118.

Meichenbaum, D., and Goodman, J. Training impulsive children to
talk to themselves: A means of developing self-control.
Journal of Abnormal Psychology, 1971, 77, 115-126.









Nadeau, G. H. Cognitive style in pre-school children: A factor
analytic study. Unpublished doctoral dissertation, University
of Minnesota, 1968.

Nie, N. H., Hull, C. H., Jenkins, J. G., Steinbrenner, K., and
Bent, D. H. Manual for the statistical package for the social
sciences. New York: McGraw-Hill, 1975.

Quay, H. C., and Peterson, P. R. Manual for the behavior problem
checklist. Champaign, Ill.: Children's Research Center,
University of Illinois, 1967.

Rummel, R. J. Applied factor analysis. Evanston: Northwestern
University Press, 1970.

Schwebel, A. I. Effects of impulsivity on performance of verbal
tasks in middle and lower class children. American Journal
of Orthopsychiatry, 1966, 36, 12-21.

Sutton-Smith, B., and Rosenberg, B. G. A scale to identify impulsive
behavior in children. The Journal of Genetic Psychology, 1959,
95, 211-216.

Vacc, N. A., and Mercurio, S. Two measures and teacher ratings of
impulsivity. Behavioral Disorders, 1978, 2(2), 114-115.

Weinreich, R. J. Inducing reflective thinking in impulsive,
emotionally disturbed children. Unpublished master's thesis,
Virginia Commonwealth University, 1975.














BIOGRAPHICAL SKETCH

Joel Cohen was born in New York City on June 23, 1952. He attended PS 69, and Joseph Pulitzer Junior High. From 1966-1969 he attended Newtown High School. He attended the State University of New York at Buffalo from 1969-1973, where he obtained a Bachelor of Arts degree. Mr. Cohen began graduate school at Northern Illinois University, in Dekalb, Illinois, from 1973-1976, when he received his master's degree in clinical psychology. He completed his clinical internship at the Shands Hospital, University of Florida, in 1978. He then transferred to the University of Florida clinical psychology program in order to complete his work towards receiving his doctoral degree. He hopes to receive his doctoral degree in August, 1979. Subsequently, he plans to accept a position as a staff psychologist at the newly opened Medical Center at the State University of New York at Stony Brook.














I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy.



Wiley C. qasbury, Ph.D.
Associate Professor, Clinical',Psychology


I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy.




Cynthia D. Belar, Ph.D.
Assistant Professor, Clinical Psychology


I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy.




Hugh C. Nvis, Vh.D.
Professor, Clinical Psychology














I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy.



Jacquelin Goldman, Ph.D. Professor, Clinical Psychology


I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy.



Larry oesch, Ph.D.
Assodlate Professor, Counselor Education


This dissertation was submitted to the Graduate Faculty of the College of Health Related Professions and to the Graduate Council, and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy.



Howard K. Suzuki, Ph.D.J Dean, College of Health Related Professions



Harry H. Sisler, Ph.D. Dean, Graduate School




Full Text

PAGE 1

IMPULSIVITY IN CHILDREN: A FACTOR ANALYTIC STUDY BY JOEL LAWRENCE COHEN A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA M979

PAGE 2

Copyright 1979 by Joel Lawrence Cohen

PAGE 3

ACKNOWLEDGMENTS The author would like to express his deep appreciation to Dr. Wiley C. Rasbury for his tireless efforts as chairman of his doctoral committee. Dr. Rasbury's many helpful suggestions, particularly with regard to ways for improving the author's writing style, contributed significantly to the successful completion of the present study. In addition, the author would also like to express his heartfelt thanks to Dr. Cynthia D. Belar, who also served on his doctoral committee. Dr. Belar's emotional support throughout the course of the completion of this study greatly aided the author, particularly at those times when his motivation and energy level began to waver. The author would also like to thank the other members of his doctoral committee: Dr. Hugh C. Davis, Dr. Jacquelin Goldman, and Dr. Larry Loesch, for their many helpful comments and suggestions. And finally, the author would also like to thank his wife, Mary, who amongst other things, typed the final copy of this manuscript. But more importantly, the author would like to thank his wife for putting up with six years of his work related irritability, piles of papers all around their home, and the placement of work directed towards the completion of his doctoral degree, before less pragmatic but potentially more enjoyable pursuits. Without her assistance and support, none of this would have been possible or desirable. i i i

PAGE 4

TABLE OF CONTENTS ACKNOWLEDGEMENTS iii LIST OF TABLES vi ABSTRACT vi i ONE INTRODUCTION 1 Purpose I Literature Review 2 Rationale 6 Overview and Hypotheses 8 TWO METHOD 11 Subjects 11 Instrument 13 Procedure 13 Factor Analyses 16 THREE RESULTS 18 Factor Analyses 18 Males 18 Females 19 Factor Scores 22 Factor Score Reliability 23 The Relationship of Factor Scores to Subject Variables 26 The Relationship of Factor Scores to ISC Performance and Teacher Ratings of Impulsivity 28 Validity of the ISC 33 FOUR DISCUSSION 36

PAGE 5

REFERENCES 66 BIOGRAPHICAL SKETCH 69 APPENDICES A REVISED IMPULSIVITY SCALE FOR CHILDREN (RISC) AND IMPULSI VITY SCALE FOR CHILDREN (ISC) 49 B TEACHER RATING SCALE FOR IMPULSIVITY 56 C FACTOR MATRICES FOR MALES AND FEMALES 57 v

PAGE 6

LIST OF TABLES TABLE NO. PAGE 1 Mean Age, Otis-Lennon IQ, and Standard Scores from the Metropolitan Achievement Test 12 2 Mean Age, Otis-Lennon IQ, and Standard Scores from the Metropolitan Achievement Test for the Reliability and Validity Sub-samples 15 3 Significant Rotated Factor Loadings for Male Factors 1 and 2 20 4 Significant Rotated Factor Loadings for Female Factors 1 and 3 21 5 T-test Comparisons of Reliability Sub-sample to Total Subject Pool 24 6 Reliability Coefficients for Factor Scores for Males and Females 25 7 The Correl ational Relationships of Factor Scores to Subject Variables 27 8 Multiple Correlations of Impulsivity and Reflectivity Factor Scores with Subject Variables 29 9 T-test Comparisons of Validity Sub-sample to Total Subject Pool 31 10 The Correlational Relationships of Factor Scores to ISC Performance and Teacher Ratings of Impul si vi ty . . 32 11 The Correlational Relationships of ISC Performance to Subject Variables 35 vi

PAGE 7

Abstract of Dissertation Presented to the Graduate Council of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy IMPULSI VITY IN CHILDREN: A FACTOR ANALYTIC STUDY By Joel Lawrence Cohen August, 1979 Chairman: Wiley C. Rasbu ry, Ph.D. Major Department: Department of Clinical Psychology Impulse control problems have been implicated in academic performance deficits in children. However, to date, there has not been available a sufficiently reliable and valid method for assessing those aspects of impulsivity which, on an overt behavioral level, contribute to children being labelled as "impul si ve . " The purpose of this study was to develop a questionnaire to tap a potentially situation specific behavioral dimension of impulsivity. The questionnaire used in the present study was a revised version of the Impulsivity Scale for Children. Both male and female students in fourth, fifth, and sixth grade elementary school classrooms participated in the study. The Revised Impulsivity Scale (RISC) was group administered to each individual class of subjects. The total sample studied included 450 children. Subsequently, a randomly selected sub-sample of approximately 100

PAGE 8

subjects was re-administered the RISC approximately ten weeks after its initial administration. A second randomly selected sub-sample of approximately 100 subjects completed the original version of the ISC. Teachers of children in this latter sample were also asked to rate each child's behavior according to the degree to which it was characteri zed by impulsivity. Where available, several pieces of demographic data were collected for each child participating in the study. These included age, race, Otis-Lennon Mental Abilities Test IQ, and standard scores on the Total Reading and Total Math portions of the Metropolitan Achievement Tests. The questionnaire responses for males and females were analyzed separately. They were subjected to a principal components factor analysis using an oblique method of rotation. These analyses yielded two negatively correlated behavioral dimensions for both males and females: an "impulsive behavior" dimension and a "reflective behavior" di mens i on . Factor scores were computed on both dimensions for each subject. An examination of the relationship between factor scores on both dimensions for those subjects who were re-admi ni stered the scale ten weeks after its initial administration, indicated that both dimensions were highly reliable for both males and females in this group. The observed relationship between factor scores on both vi i i

PAGE 9

dimensions and teacher ratings of impulsivity, strongly supported the validity of the factor scores for males. For the females, however, while this relationship was significant for the impulsivity dimension, it only approached significance for the reflectivity dimension. The relationship between factor scores on both dimensions and available demographic data was also examined. Contrary to expectations, it was found that less reflective male and female subjects had higher achievement test scores than more reflective subjects. In addition, for females, it was also found that better achievement test performance was associated with higher scores on the impulsivity dimension. Several possible explanations for these findings were discussed, as were methods for testing them out in subsequent investigations using these scales. Several methodological issues were discussed pertaining to the potential general i zabi 1 i ty of these findings. In addition, several potential uses for the impulsivity and reflectivity scales were offered. However, it was noted that the individual scales were in need of refinement prior to their use as either research or clinical tools. A number of methods for revising the scales were offered.

PAGE 10

CHAPTER I INTRODUCTION Purpose The Impulsivity Scale for Children (ISC: Sutton-Smi th and Rosenberg, 1959) was developed as a self-report measure of a broad dimension of impulsive behavior. It was expected that the availability of an economical test for measuring this construct would allow practitioners such as psychologists and teachers to have at their disposal, a method for conceptualizing impulsive behavior with a significant degree of consensual validation. To the extent that impulsive behavior could be implicated in poor academic performance, then the ability to objectively isolate very impulsive children would serve as a first step in developing appropriate interventions. However, despite the potential utility of such a scale, it has received remarkably little attention in the psychological literature to date. As such, the reliability and validity of the ISC, as a measure of impulsive behavior, have not been firmly established. The purpose of the present study was to examine both the reliability and validity of a revised form of this scale. The proposed revisions included the addition of several scale items which theoretically should have aided in the delineation of this scale as a measure of 1

PAGE 11

2 impulsive behavior. The revised scale's long term stability was also examined for a normative population of nine to twel ve-year-ol d elementary school children. Factor analytic methods were used to assess the factorial structure of the revised version of the scale. Scores on the revised version of the scale were compared to scores on the original version of the ISC, for a sub-sample of the total population studied. And finally, comparisons between scale scores and teacher ratings of impulsive behavior were used to assess the validity of both versions of the scale. Literature Review As it was originally developed, the ISC was a nineteen item self-report scale for measuring a broad dimension of impulsive behavior. Each item could be responded to either "yes" or "no". This original scale was itself based upon the combination of two earlier versions of the scale, one based upon the responses of a group of college freshmen, and one based upon responses obtained from a sample of fourth, fifth, and sixth grade students. All of the items on the original version of the scale were keyed in the affirmative for impulsivity. In an answer to the criticism that the scale tapped the tendency to respond in the affirmative to self-descriptive statements regardless of content, and not impulsive behavior, Hirschfield (1965) modified the scale to include, in random order, both affirmative and negative

PAGE 12

3 items keyed for impulsivity. Response categories were also changed from yes-no to true-false. A more recent, but less frequently used modification of the scale by Bjorklund and Butter (1973), involved the inclusion of five additional true-false questions to serve as a "Lie" scale. In this particular study, subjects giving three or more positive responses to the Lie items were excluded from further i nvesti gation . To date, only two studies have reported any reliability data for the ISC. Sutton-Smith and Rosenberg (1959) reported a testretest reliability of .85 for the scale. More recently, Kendall and Finch (1978) reported a short term (one month) dependability coefficient of .56 and a long term (three months) stability coefficient of .33. Validation studies have relied almost exclusively upon correlations between ISC scores and teacher rankings or ratings of •k impulsive behavior. For example, Sutton-Smith and Rosenberg (1959) reported that fifteen of eighteen correlations relating teacher rankings or ratings of impulsive behavior to impulsivity scale scores achieved significance at the .05 level or better. These correlations ranged from -.33 to -.73. Hirschfield (1965) also correlated teacher The method of ranking/rating impulsivity determined the directionality of the correlations (e.g. , l=least impulsive and 10=most impulsive vs. l=most impulsive and 10=least impulsive).

PAGE 13

4 rankings of impulsive behavior with ISC scores, in a sample of fifth and sixth grade students. Ranked difference correlations were computed comparing teachers' rankings of their students within each class and the students' ranked test scores. Five of the resulting fifteen correlations reached significance at the .05 level. These significant correlations ranged in magnitude from .73 to .87. Using teacher ratings as opposed to rankings of impulsivity, Bjorklund and Butter (1973) found little relationship between teacher ratings of specific behaviors and ISC scores, in a sample of fourth grade students. More recently, Bentler and McClain (1976) correlated ISC scores with both teacher and peer ratings of impulsive behavior for a sample of sixth grade students. The resulting correlations were .45 and .35 respectively. And finally, Vacc and Mercurio (1978) found a correlation of .56 between teacher ratings of impulsivity and ISC scores in a sample of fifth grade students. The available reliability data for the scale are obviously very limited. Furthermore, the available validity data are not at all consistent. This may, in part, reflect methodological problems with the validation studies carried out thus far. Sample sizes have varied from a mean of 28.5 for each classroom examined ( Sutton-Smi th and Rosenberg, 1959) to an N of 144 (Bjorklund and Butter, 1973). With respect to the latter, it is of interest to note that, while having a modest sample size, and making use of teacher ratings of specific

PAGE 14

5 behaviors (as opposed to the more global impulsive behavior ratings or rankings used in other studies), this study also yielded the least support for accepting the ISC as a valid measure of impulsive behavior. However, this finding may also have been an artifact of the more stringent statistical procedures used in this study, relying as it did, on teacher ratings vs. rankings of impulsive behavior. This latter explanation does lose some support in light of the work by Bentler and McClain (1976) and Vacc and Mercurio (1973), both of whom relied upon teacher ratings of impulsivity. Obviously, additional research is needed to more firmly establish the psychometric credibility of the ISC. However, it is important to note that, with respect to its reliability, there are several potential sources of measurement error within the scale which could act to limit its reliability. These sources of error include the presence of scale items which are very global and non-si tuational ly specific. In addition, many of the items are not very descriptive or operationally defined, and this may contribute to their being misunderstood by the children responding to them. And finally, the use of dichotomous rating categories can act to limit the reliability of the scale. Little attention has been paid to the issues of dependability vs. stability of ISC scores. This issue reflects a more basic concern with hypothesizing about whether impulsive behavior, defined in a broad sense, can be viewed as a state variable which

PAGE 15

6 might be expected to spontaneously diminish with time, or a trait variable which would not be expected to change over time, unless specific interventions were designed to promote change. R ati onale Impulsive behavior has been identified as being undesirable, particularly as it relates to poor academic performance in school. For example, based upon a large body of empirical research, Virginia Douglas and her associates (Douglas, 1974) have argued that the academic performance deficits frequently found in hyperactive children, are attributable to deficits in the areas of impulse control and attention. The realization of the potentially aversive consequences of impulsive behavior has led to the development of a number of different types of interventions designed to enable impulsive children to learn to exert more conscious control over their behavior. For example, a number of investigators have proposed the use of instructional programs designed to teach cognitive self-mediation of behaviors, as a potentially effective method for modifying impulsive behavior (e.g., Kendall and Finch, 1976, 1978; Meichenbaum and Goodman, 1971). Others have recommended that impulsive children be instructed in the use of more effective problem solving and search strategies (e.g., Egeland, 1974; Heider, 1971). The use of these types of interventions depends upon the practi tioner' s ability to

PAGE 16

7 effectively isolate children whose behavior is deemed to reflect i mpulsivity. One accepted method for doing this revolves around a conceptualization of impulsivity as a cognitive style variable, along with reflectivity. Kagan and his associates have proposed the use of the Matching Familiar Figures Test for classifying children as being either reflective or impulsive in their approach to problem solving situations (Kagan, Rosman, Day, Albert and Phillips, 1964: Kagan, 1965a, 1965b, 1965c). The Matching Familiar Figures Test is a match-to-sampl e cognitive task. Children are presented with a standard figure and usually, four, six or eight figures differing from one another in one or more details. The task for the child is to choose the alternative which exactly matches the standard. With few exceptions, the Matching Familiar Figures Test has been the method of choice for isolating impulsive subjects in research studies designed to examine the relative efficacy of various methods for modifying impulsive behavior. However, as several authors have noted, there are many methodological problems with the Matching Familiar Figures Test (e.g., Ault, Mitchell and Hartmann, 1976; Egeland and Weinberg, 1976), as well as broader conceptual problems with the impul si vi tyrefl ecti vi ty cognitive style dimension (e.g., Block, Block and Harrington, 1974). In addition there is also a body of psychological literature which strongly suggests that what is tapped by the Matching Familiar Figures Test may, in fact, not reflect those behaviors which teachers use as criterion measures

PAGE 17

8 for classifying a child as being impulsive (e.g., Bjorklund and Butter, 1973; V acc and Mercurio, 1978). Hence, the need for a scale specifically designed to tap the types of behaviors suggested as reflecting impulsive behaviors at home, in school, and in interpersonal situations. One such scale is the Impulsive Classroom Behavior Scale (Weinreich, 1975), a nine item teacher rating scale which was specifically developed for use in research involving the modification of impulsivity. The search for a more economical, less time consuming measure of impulsivity led to the development of the ISC. However, the review of the psychometric status of the scale suggests that we are not, at this point in time, able to state emphatically that the ISC is a reliable, valid measure of the construct of impulsivity. If its psychometric credibility could be established, then the ISC could hypothetically serve a multitude of purposes. In addition to allowing the practitioner to define observable impulsive behavior for the purposes of research and/or intervention, the focus of the scale on observable behaviors could have implications for delineating the specific foci of intervention plans. O verview a nd Hypot he ses Except for the original study by Sutton-Smith and Rosenberg (1959), and the subsequent work by Hirschfield (1965), much of the research attempting to shed some light on the psychometric credibility of the ISC,

PAGE 18

9 has not had the study of the scale's reliability and/or validity as its primary focus. Given some of the problems with the ISC as it was originally developed, it was felt that the scale had to be revised prior to any attempts to establish the scale's reliability and validity. It was the establishment of the reliability and validity of this revised version of the scale which was given primary emphasis in this study. The scale's long term stability (ten weeks) was assessed, and factor analytic procedures, with their emphasis upon the delineation of the factorial composition of scales, was used to assess its basic structure. The relationship of the revised version of the scale (RISC) to the original ISC was also assessed by examining the correlation between scores on both scales, for a subset of the total population studied. If the original version of the scale was used, one would predict that factor analysis would yield only one general factor. This is due to the scale's dichotomous rating categories, ambiguous items, and lack of situational specificity. In an effort to more clearly delineate the factor structure of the scale, and generate information relevant to the conceptualization of impulsivity, the original ISC was revised via the addition of twenty-nine marker items. These included items reflecting conduct problems, items reflecting behavioral self-control, and a set of six items randomly selected from the scale's original item pool, phrased in the negative for impulsivity. The

PAGE 19

10 latter were included as a gross indicator of whether subjects were carefully attending to the task. The rating categories for the scale were also modified to allow the child a choice of five responses for each item. Based upon these considerations, the following predictions with respect to the factor analytic data were made: (1) upon analysis, an impulsive behavior factor would emerge, reflecting loss of affective and behavioral controls, poor planning, and restlessness. (2) in addition, a second factor would emerge reflecting behavioral sel f-control . (3) while there is some indication in the literature, that males score as being more impulsive than females on the ISC, it was predicted that, in terms of content, there would be no difference in the factor structure of the RISC for males and females.

PAGE 20

CHAPTER II METHOD Subjects The data for the study were obtained from a sample of 450 children, 230 females and 220 males, between the ages of nine and twelve. All were enrolled in either fourth, fifth, or sixth grade elementary school classes in Gainesville, Florida.* This final subject pool was chosen from approximately 1100 children. The only criterion for selection was the return of a signed parental consent form allowing the child to participate. Other than this criterion, the attempt was made to choose a subject population which, as closely as possible, approximated an unselected sample, in order to insure the presence of a high degree of variability on the impulsivity scale. In Table 1, the mean age, and where available, the mean OtisLennon Mental Abilities Test IQ, and mean achievement test scores (i.e., total reading and total math standard scores from the Metropolitan Achievement Tests) are provided, for the total sample, and for the males and females separately. Includinq students from the P.K. Yonqe Laboratory School. 11

PAGE 21

12 TABLE 1 Mean Age, Otis-Lennon IQ, and Standard Scores From the Metropolitan Achievement Test Metropolitan Achievement Test Age IQ Total Reading Total Math Total Sample (N=450) 10.91 (SD=.93) 114.90 * ( S D= 1 5 . 22 ) 79.71 (SD=1 5.46) 87.78 ( SD=1 3.47) Males (N=220) 10.92 ( SD= .94) 115.42 ( SD= 1 5 .99) 79.42 (SD=1 6 .16) 87.50 (SD=14.59) Females (N=230) 10.91 ( SD= .93) 114.39 (SD=1 4. 45) 79.99 (SD=14. 79) 88.06 (SD=1 2.30) "k SD standard deviation

PAGE 22

13 Instrument The most contemporary version of the ISC (Bjorklund and Butter, 1973), containing twenty-five true-false items was used as the original version of the scale (see Appendix A). All items were keyed in the affirmative for impulsivity given Hi rschfield' s (1965) finding that impulsive children were no more likely than nonimpulsive children to respond in the affirmative to self-descriptive questionnaire items. With the addition of the twenty-nine marker items previously mentioned (see Appendix A), the total Revised Impulsivity Scale for Children (RISC) included 54 items, sequenced in random order. For the revised scale, subjects were asked to rate each item on a five point scale ranging from l=never happens, to 5=always happens. Procedure The RISC was group administered by the principal investigator to each individual class of subjects. At each administration, each child was given a copy of the scale and then asked to read the directions silently as they were read aloud to them. The directions used were as follows: there will be fifty-four items on this questionnaire. Read each item to yourself, as I read it aloud. Then, using the scale to the right of each item, decide how often you engage in that particular behavior. A "1" means never, a "3" means sometimes, and a "5" means always. After you have decided upon your answer, circle the appropriate number to the right of each item. (Appendix A)

PAGE 23

14 The "2" and "4" response choices were also explained to the subjects in terms of their relationship to the other three possible choices. That is, subjects were told that: a "2" response represented more than a "1" response (never), but less than a "3" response (sometimes): a "4" response represented more than a "3" response (sometimes), but less than a "5" response (always). (Appendix A) Items were read to the subjects in order to allow for the pacing of the scale administration. By allowing approximately 15-30 seconds per question, the total time required to administer the scale was approximately 15-20 minutes. The children were encouraged to answer each question as truthfully as they could. The questionnaires were numbered in order to maintain the subject's anonymity. They were also assured that their performance on the scale had nothing to do with their academic capabilities. Following the administration of the RISC to the total subject pool, two sub-samples were selected. This was accomplished using the procedures outlined for generating random samples in the manual for the Statistical Package for the Social Sciences (SPSS: Nie, Hull, Jenkins, Steinbrenner and Bent, 1975). This process was used to generate two random samples of approximately 100 subjects each. Table 2 includes the mean age, IQ and achievement test scores for each total sub-sample as well as for the males and females within each sub-sample.

PAGE 24

15 TABLE 2 Mean Age, Otis-Lennon IQ, and Standard Scores from the Metropolitan Achievement Test for the Reliability and Validity Sub-samples Age IQ Metropolitan Achievement Test Total Reading Total Math Rel i abi 1 i ty Sub-sample (N=92) 10.90 (SD=1 .03) * 115.55 (SD=1 3.64) 79.17 (SD=1 4. 36) 89.35 ( S D= 1 2 . 55 ) Males (N=41) 10.96 (SD= .96) 115.00 ( SD=1 4.51 ) 77.76 ( S D= 1 5 .48) 89.28 ( SD=1 4 . 78) Females (N=51) 10.84 (SD=1 .09) 115.92 ( SD=1 3.18) 80.20 ( SD=1 3 . 60 ) 89.40 ( SD=1 0.86) Val idi ty Sub-sample (N=94) 11.00 (SD= .88) 116.34 (SD=1 4. 39 ) 83.00 (SD-12.47) 91.55 (SD=11 .82) Males (N=52) 11.05 (SD= .90) 116.13 (SD=14. 77) 83.10 ( SD=1 2.29) 91.18 (SD-12.32) Females ( N=42 ) 10.93 ( S D= . 86 ) 116.64 (SD=14. 10) 82.86 (SD=1 2.94) 92.07 (SD=11 .28) SD standard deviation

PAGE 25

16 Subjects in the reliability sub-sample ( N = 9 2 ) were re-administered the RISC under the same conditions used previously, approximately ten weeks after its initial administration. Subjects in the validity subsample ( N = 9 4 ) were administered the original version of the ISC. Directions for this administration were as follows: there will be twenty-five items on this questionnaire. Read each item to yourself as I read it aloud. Then decide whether it is true about you or false about you. If it is true or mostly true, circle the T; if it is false or not usually true about you, then circle the F. (Appendix A) The teachers of subjects in the validity sub-sample were also asked to rate each subject on a 1-11 scale of impulsivity (see Appendix B), where "1" indicated that the child's behavior was not at all characteri zed by impulsivity, and "11" indicated that the child's behavior was characterized by extreme impulsivity. Factor Analyses Given the exploratory nature of this investigation, and the suggestions in the literature that sex differences may represent an important variable in this area, the questionnaire data were analyzed for males and females separately. The factor analyses were carried out in the following manner. A 54 X 54 pearson product moment correlation matrix was computed. This matrix was then submitted to a principal factors analysis,

PAGE 26

17 as outlined in the SPSS Manual (Nie et al., 1975). This mode of factor analysis employed an iteration procedure for improving communality estimates which were then used as the main diagonal elements of the correlation matrix. Rotation was carried out using the direct oblimin method of oblique rotation. The obliqueness criterion was set at zero. This procedure allowed for correlation amongst the generated factors but did not bias the analyses in this direction. The resulting rotated factor structures were then examined for the extent to which they approximated simple structure (i.e., most all of the loadings on a given factor were small, with only a few being of significant size; a given variable should have significant loadings on only one factor and, the pattern of loadings for a given factor across all of the variables should be different from one factor to the next). Several guidelines for the interpretation of factor structures were used (Comrey, 1973). These included the criteria that for a loading to be significant, it had to be greater than ±.40, and that at least six items had to load significantly on a given factor for it to be interpreted.

PAGE 27

CHAPTER III RESULTS Factor Analyses For both the males and the females in the subject pool, factor analysis was originally performed with no limits placed upon the number of factors to be extracted and then rotated. However, after this initial anlaysis, which yielded nineteen rotated factors for each group, it was decided to limit a subsequent analysis for each group to the extraction and rotation of five factors. It is this analysis for both the males and the females studied, which is reported here. This decision was based upon Comrey's (1973) discussion of factor extraction, and the fact that the percentage of variance accounted for by each successive factor beyond five, in conjunction with the lack of significant loadings on those factors, suggested that it was highly unlikely that they would be interpretable. Males The rotated factor structure for males yielded two interpretable factors, accounting for 67.4% of the total variance of the questionnaire. These two factors were found to be negatively correlated (r=-.30) with no overlap of items having significant loadings. Items loading 18

PAGE 28

19 significantly on these two factors are presented in Table 3. In addition, both the factor pattern matrix and the factor structure matrix for males are presented in Appendix C. Factor 1 for males, accounted for 51.5% of the total variance. Items loading significantly on this factor were indicative of a high energy level, restlessness, and episodic loss of affective control. This factor was labelled Impul si vi ty . Factor 2 for males accounted for 15.9% of the total variance. Items loading significantly on this factor were concerned with the thoughtful planning of activities, and self-directed behavior. Thus, this factor was labelled Reflectivity . Females The rotated factor solution for the females in the subject pool also yielded two negatively correlated ( r=. 24 ) interpretable factors which, in this case, accounted for 60.3% of the total variance. Items having significant loadings on these two factors are presented in Table 4; the rotated factor pattern and factor structure matrices are presented in Appendix C. The interpretable factor structure for females less closely approximated Comrey's (1973) criteria for simple structure, than did the factor structure for the males. There was no overlap of items loading significantly on the two factors, but the items loading

PAGE 29

20 TABLE 3 Significant Rotated Factor Loadings for Male Factors 1 and 2 Factor #1 Description Factor Loadings h 2 Item # 1 2 3 4 5 23 I like to wrestle and horse around. .53 -.10 -.09 -.02 -.06 .30 10 I am restless. .49 -.03 .15 .11 .09 .28 24 I 1 ike to just blow off steam. .48 .01 -.06 .08 -.02 .24 n I have trouble sitting still . .46 -.03 .26 .08 .15 .31 36 I get so angry sometimes I want to kick, scream and throw things. .46 -.05 .01 -.08 -.10 .24 6 I 1 i ke to keep movi ng around. .43 .02 -.02 .13 -.03 .20 38 I like to dare kids to do things. .42 -.10 -.06 -.22 -.27 .31 30 I must admit, I'm a pretty good talker. .42 .07 -.08 -.11 -.01 .20 41 I usually say the first thing that comes into my head. .40 -.03 .02 -.06 .003 .16 Factor #2 Item # 15 I finish all of my homework before I go out to play. -.15 .51 .14 -.02 .05 .30 20 I usually plan what I am going to do before I do it. .06 .50 -.19 .11 .01 .30 5 When I do my homework, I take -.003 .50 -.03 .02 .04 .25 my time to make sure I do the best job I can. 52 I keep my room at home neat and clean. -.16 .47 -.02 -.08 -.04 .26 49 I go to bed without being told that I have to. -.003 .44 .07 .04 .07 .19 8 I like to bother other children at school (negative). .26 -.42 .09 -.12 -.03 .27 17 In school , I speak only when -.23 .40 .16 .09 -.02 .25 I am spoken to by the teacher.

PAGE 30

21 TABLE 4 Significant Rotated Factor Loadings for Female Factors 1 and 3 Factor #1 Descri ption Item # 11 I have trouble sitting still 4 I like to show off in front of my friends. 23 I like to wrestle and horse around. 21 I get angry qui ckly. 24 I like to just blow off steam. 30 I must admi t, I ' m a pretty good talker. Factor #2 Item # 31 When things get quiet, I 1 i ke to stir up a 1 i ttle fuss, (negative) 20 I usually pi an what I am going to do before I do it. 38 I like to dare kids to do things, (negative) 29 I can read for half an hour or more without getting up from my seat. 5 When I do my homework, I take my time to make sure I do the best job I can. 15 I finish all of my homework before I go out to play. Factor Loadings h_ 1 2 3 4 5_ .49 -.04 .03 -.00 .03 .24 .47 .20 -.04 .03 .03 .26 .45 .11 -.19 .04 .27 .32 .42 -.02 .01 -.17 -.06 .21 .41 .12 -.01 -.02 -.01 .18 .40 -.01 -.33 .20 -.09 .32 .37 CO o 1 -.46 .12 .13 .39 .05 -.01 .46 -.03 -.09 .22 .11 -.10 LD ^1" 1 1 O CO .21 .27 -.01 .11 .44 -.06 .27 .27 -.02 -.16 .44 .35 -.15 .36 -.17 -.14 .42 .16 -.00 .26

PAGE 31

22 significantly were less factor pure than those for the males, in certain instances. Factor 1 for females, accounting for 48% of the total variance was very similar to Factor 1 for the males, although not as robust in terms of the number of items loading significantly and the percentage of variance accounted for by the factor. It included items indicative of loss of emotional control, and restlessness, and again, would appear to be reflective of an underlying Impulsivity dimension. Factor 3 for females, accounting for 12.3% of the total variance also bore a great deal of content similarity to Factor 2 for the males. Items loading significantly were concerned with methodical responding and the maintenance of control, suggesting that it was also indicative of an underlying Refl ecti vi ty dimension. Factor Scores Having identified two fairly strong factors for the males and two somewhat less robust but similar factors in terms of content for the females, factor score estimates were computed for each subject within each group. The method for computing factor scores was similar to the preferred procedure described by Comrey (1973), wherein items with significant loadings on a given factor are first standardized to the same mean and standard deviation, and then weighted based upon the loading of that item on the factor, prior to being summed to give

PAGE 32

23 a total score. In this case, the item weightings represented an integral value approximately proportional to the factor loading of that item. The major advantage of this method for computing factor scores is that it allows those items having the highest loadings on a given factor to have the greatest effect in estimating the factor scores . Factor Score Reliability The availability of a second set of questionnaire responses for a randomly selected sub-sample of the total subject population, allowed for an assessment of the reliability of the factor scores for both the Impulsivity and Reflectivity factors. Factor scores were computed for subjects in the reliability sample using their responses at the ten week re-administration of the RISC and the weighting coefficients derived from the factor analyses of the total samples of males and females. It was felt that the same weighting coefficients could be used for the reliability sub-sample because t-tests comparing that group with the total sample indicated no significant differences in either age, IQ, or achievement test scores (see Table 5). As such, it was assumed that this sub-sample was representati ve of the larger group. Table 6 presents the reliability coefficients for the factor scores for the males and females separately.

PAGE 33

24 CO o o Q_ u cd _o Z3 CO ro 4-> O d) QE 03 00 1 JO Z3 co >> JO 03 CD C£ CO c: o 00 •I — s~ 03 Q. E o C_J CO CD >> 4-> JO -M 03 00 JD O SQ_ "O CD I OJ CD 13 03 O'. CO C\J CO o co co li~) LO CO O T3 CD CD S~ Lien CD Q > CD O “O sz 03 *3* O CO o I CO cv 03 CO CO C\J LO C\J o C\J CO LO o co 03 CO o CVI LO CXI LO LO CO LO i — i — i — LO CO ^3" 1 o o LO , — OD sz 03 03 03 LO r— 03 • • . • • , . CD o o LO 03 03 5T 1 — 1 — 1 — i — co CO CD CO 03 C_3 4— o OsJ 03 o CO LO LO LO CO 03 LO CD CD CD i — >> CD i — CD r— ,1 CL 4-> i — CL +-> r— CL E • i — CL E •r— CL £ ii CD 03 r — E 03 i — E 03 ii r— CO •r — 03 CO • r— 03 CO 1! JO JO (/) JO CO i j 03 1 — 03 1 1 — 03 i , — I •i — 03 •i — JO 03 JO * 03 .1 s_ +-> i — Z3 -4— > i — Z3 o +-> i 1 03 O CD co O CD CO O !i > CD 1— Cd 1— Cd LxJ 1 — 03 21 Cd 1 — >•> CD 4-> r— •ICL I— E •r— 03 JO co 03 I •rJD . — =3 CD CO Cd LO CO CD r a. 03 CO 03 4-> O 105.03 -0.93 Reliability 69 89.35 12.55 Sub-sampl e *TREAD Standard score for the total reading portion of the Metropolitan Achievement Tests. **TMATH Standard score for the total math portion of the Metropolitan Achievement Tests.

PAGE 34

25 TABLE 6 Reliability Coefficients for Factor Scores For Males and Females Sex Males (N=41) Females (N=51) Factor Scores : Impul si vi ty .72* .67* Reflecti vi ty .68* .64* * p <.001

PAGE 35

As can be seen, both the impulsivity and reflectivity dimensions were highly reliable for both the male and female subgroups within the reliability sub-sample. However, this finding must be tempered by the fact that the relatively small sample sizes may have acted to artificially inflate the correlation coefficients. The Relationship of Factor Scores to Subject Variables Factor scores were correlated with available demographic data for all subjects. Table 7 presents the resulting correlations for males and females. For both the males and females respectively, the results indicated that older subjects tended to be less reflective (Males: r=-.25, p< . 01 ; Females: r=-.39, p<.001). However, only in the case of the females was there a corresponding increase in impulsivity for the older children (r=.23, p<.001). An interesting finding involved the relationship between subjects' standard scores for the Total Reading and Total Math portions of the Metropolitan Achievement Tests. While there was no significant relationship found for either group, between factor scores for the Impulsivity and Reflectivity dimensions and Otis-Lennon Mental Abilities Test IQ, in both groups it was found that less reflective subjects had higher achievement test scores (Males: Total Reading, 4=-. 21, p< . 0 1 ; Total Math, r=-.27, p< . 9 1 ; Females: Total

PAGE 36

27 TABLE 7 The Correlational Relationships of Factor Scores to Subject Variables MALES (N=220) Subject Variable Metropol i tan Achievement Test Age IQ Total Reading Standard Score Total Math Standard Score Factor Score: Impul si vi ty .04 -.01 -.07 -.04 Reflecti vi ty -.25 -.11 -.21* -.27* p< .01 FEMALES (N=230) Subject Variable Metropol i tan Achievement Test Aqe IQ Total Reading Standard Score Total Math Standard Score Factor Score: Impul si vi ty .23* .04 22** .21** Refl ecti vi ty -.39* .07 -.20*.* -.25** *p< . 001 **p<.01

PAGE 37

28 Reading, r=-.20, p<.01 ; Total Math, r=-.25, p< . 01 ) . Furthermore, in the female subgroup, it was also found that better achievement test performance was also associated with higher scores on the Impulsivity dimension (Total Reading, r=.22, p< .01 ; Total Math, r= . 21 , p< .01 ) . Given the fact that both dimensions were found to be negatively correlated with one another for males and females, it was decided to examine, on at least a descriptive level, the multiple correlation of factor scores on both dimensions with each of the available subject variables. This was accomplished using the multiple regression procedure outlined in the SPSS Manual (Nie et a 1 . , 1975). The results of these analyses are presented in Table 8. As can be seen, while for both males and females, the multiple correlation of both factors with age (males: .31; females: .35), Total Reading Score (males: .28; females: .25) and Total Math Score (males: .35; females: .27) from the Metropolitan Achievement Tests, are substantial, only in the case of "age" for the females did the factor scores taken jointly account for a substantial proportion of the variance (i.e., 20%) in a subject variable. T he Relationship of Factor Scores to ISC Performance and Teacher Ratings of Impulsivity The availability of a sub-sample of subjects who had completed the original version of the ISC and were also rated for impulsivity

PAGE 38

29 TABLE 8 Multiple Correlations of Impul si vi ty and Reflectivity Factor Scores with Subject Variables MALES (N=154)* Adjusted Standard Multiple R R Square R Square Error Age .31 .10 .09 0.90339 IQ .12 .01 .00 15.98204 Metropolitan Achievement Test Total Reading Standard Score .28 .08 .07 15.44168 Total Math Standard Score .35 .12 .11 13.57607 FEMALES (N=l 56) ** Adjusted Standard Multiple R R Square R Square Error Age .45 .20 .19 0.82477 IQ .09 .01 .00 14.48784 Metropolitan Achievement Test Total Reading Standard Score .25 .06 .05 13.73290 Total Math Standard Score .27 .07 .06 11.85801 ^Represents that portion of the total sample of males (N=220) having data points for each of the variables included in the analyses. **Represents that portion of the total sample of females (N-230) having data points for each of the variables included in the analyses.

PAGE 39

30 by their classroom teachers allowed for an examination of the validity of factor scores based on the Impulsivity and Reflectivity dimensions, the relationship between RISC performance and performance on the original version of the ISC, and the validity of the original version of the ISC. T-tests comparing this validity sub-sample with the total sample, on the basis of age, IQ, and achievement test scores found no significant differences on any of these variables except for the total Math portion of the Metropolitan Achievement Tests (see Table 9). Hence, it was felt that in general, findings pertaining to this sub-sample would be general i zable to the sample as a whole. Table 10 presents the correlations between the Impulsivity and Reflectivity factor scores, performance on the ISC, and teacher ratings of impulsivity, for males and females separately. Again, these findings must be interpreted with caution given the sample size limitations. Despite this fact, there were strong indications as to the validity of the factor scores for both males and females. This was based upon the observed relationship between factor scores on both dimensions, and teacher ratings of impulsive behavior. For both groups, it was found that subjects scoring higher on the impulsivity dimension were also rated as being more impulsive by their teachers (Males: r=.29, p<.05; Females: r=.29, p<.05). In addition, male subjects with higher scores on the reflectivity dimension were also rated as being less impulsive by their teachers (r=-.32, p<.05). This relationship approached significance for females.

PAGE 40

31 -O 4-> 03 tO i e ! CL j u CD CD LO O CM CO o CO co o 03 o o Cl O) 13 03 CO O l co CO (J CD * r— 3 JO 3 OO 03 4-> O CD >> 03 4— O to d O to •r— 503 CL E o o to CD O o CD CD d CT) CD Q c r> 00 CD CO CO CM or CO r^ 1 Q 03 CO CM co LU JD • . • . — 1 3 • o o LO •=3* LO CM co CO oo “C i — i — i — i — • — C 03 03 03 CO CO LO o o CM LO i — 03 o 1 — o CO a 03 03 03 CO r^ o 03 . « . . • CD o o <3CO 03 CO 1 — 1 — 1 — CO CO to CD to 03 CJ> 4O •M03 O r— CO co co co co co co 1 £= 03 E 03 >> pc 03 i — OO 4-> 03 OO +J 03 OO •+-> 03 OO •r— to • i— to •r— to J ^ r— ~a i r— X3 1 r — ~a 1 * r— •i — 03 •i — JO 03 • r— JO * 03 •i — JO * 03 s+J i—~ =3 4-> l 3 o 4-> r 3 nr +-> ro O 03 oo O 03 OO O 03 CO hO > CD h> 21 1 — > LU h> < f— 03 CL hh105.65 -2.31 0.023 Validity 67 91.55 11.82 Sub-sample *TREAD Standard score for the total reading portion of the Metropolitan Achievement Tests. **TMATHStandard score for the total math portion of the Metropolitan Achievement Tests.

PAGE 41

32 TABLE 10 The Correlational Relationships of Factor Scores to ISC Performance and Teacher Ratings of Impulsivity MALES (N=52) ISC Score Teacher Rating Factor Score: Impul si vi ty .58** .29* Reflectivity -.55** -.32* *p<. 05 **p<.001 FEMALES ( N =42 ) ISC Score Teacher Rating Factor Score: Impul si vi ty fa-]*** .29* Refl ecti vi ty _ _ g] *** 24** *p<.05 **p<.07 ***P<.001

PAGE 42

33 Since the RISC used as its base, the ISC as it was originally developed by Sutton-Smith and Rosenberg (1959), their relationship to one another was examined for subjects in the validity sub-sample. As can be seen from an examination of Table 9, the expected relationship of RISC scores on both dimensions to ISC scores was found for both males and females in the validity sub-sample. That is, males found to have higher factor scores on the Impulsivity dimension of the RISC were also found to have ISC scores more indicative of Impulsivity (r=.58, p< .001 ) . It was also found that more reflective subjects, based on RISC performance, had lower ISC scores (r=-.55, p<.001). For the females, high RISC impulsivity factor scores were also associated with high ISC scores indicative of impulsivity (r=.67, p<.001). In addition, reflective females, based on their RISC performance, were found to have lower ISC scores (r=-.61, p< . 001 ) . Validity of the ISC As was previously mentioned, to date, attempts to validate the ISC have relied upon examining the correlation of ISC scale scores to either teacher ratings or teacher rankings of impulsivity. This relationship was examined in the present study for all of the subjects in the validity sub-sample. The resulting correlations indicated that while the ISC was a valid measure of impulsivity for the males in the sub-sample (r=.27, p<.05), and for the sub-sample as a whole

PAGE 43

34 (r=.26, p< . 01 ) , it was far less adequate a measure of the construct for the females in this group (r=.05, p<.40). As was the case with the RISC factor scores, ISC performance for subjects in the validity sub-sample were correlated with available demographic data. Table 11 presents the resulting correlations for males and females. As can be seen, none of these relationships approached significance for either group.

PAGE 44

TABLE 11 The Correlational Relationships of ISC Performance to Subject Variables MALES ( N = 52 ) Age IQ Metropol i tan Total Reading Standard Score Achievement Test Total Math Standard Score ISC Performance .10 -.20 -.19 -.15 FEMALES (N=42) Age IQ Metropol i tan Total Reading Standard Score Achievement Test Total Math Standard Score ISC Performance .11 -.06 -.04 -.11

PAGE 45

CHAPTER IV DISCUSSION The study of impulse control problems in children has been impeded by a lack of appreciation for the potential heterogeneity of this construct. This is despite the fact that impulsivity has been looked to as a major contributor to academic difficulties in children. The psychological literature to date has suggested that the available methods for assessing impulsivity in children have not adequately tapped those aspects of the dimension which reflect, for example, the types of behaviors used by teachers as criteria for labelling a child as being impulsive. The present study was undertaken in an effort to attempt to delineate, via questionnaire data, a potentially situation specific behavioral dimension of impulsivity. The factor analysis of a revised self-report questionnaire designed to measure impulsivity, yielded two negatively correlated behavioral dimensions for both males and females. Both dimensions tended to be somewhat more robust for the males than for the females in the sample studied. The first dimension was characterized by a number of impulsive characteristics, and was named accordingly. The second dimension, indicative as it was of the pre-planning of activities, 36

PAGE 46

37 methodical responding, and affective control, was named "Reflectivity" for both groups. Further analyses indicated that, for a randomly selected subsample of the total subject pool, factor scores derived from subjects' questionnaire responses for each dimension, were highly reliable. Through the correlation of factor scores with teacher ratings of impulsivity, the validity of the factor scores for both dimensions was strongly supported for the males in another randomly selected subsample of subjects. The validity of the factor scores was also supported for the females in this group although less substantially for the Reflectivity factor, than was the case for the males. The implications of these findings are many as are the possibilities for further research. However, prior to further interpretation, it must be noted that there are several methodological issues which serve to restrict the general i zabi 1 i ty of these findings and hence, must be dealt with in subsequent research designed to refine this scale. First, as already mentioned, the relatively small number of subjects in the reliability and validity sub-samples act to limit the extent to which we can accept the findings pertaining to those groups. Further investigations attempting to examine the utility of the RISC should consider increasing the total sample size as well as the size of samples used for establishing the scale's reliability and validity. In addition, with respect to establishing the scale's

PAGE 47

38 reliability, subsequent studies should consider examining its short term dependability (e.g., one week) as well as its long term stability (e.g. , ten weeks) . Two related concerns involve the composition of the subject pool. The attempt was made to, as closely as possible, approximate an unselected sample in order to enhance the probability of there being a wide degree of variability on the impulsivity dimension. In analyzing the results, the major subject variable attended to, apart from demographic variables such as age, IQ, and achievement test scores, was subjects' sex. Katzenmeyer and Stenner (1977), in their discussion of factor structure comparisons, pointed out the importance of attending to differences in factor structure across criterion groups within samples (i.e., with criterion groups based upon sex, race, etc.). The present study did not examine, at any level, the impact of race, and the associated variable of social class upon questionnaire performance and resulting factor structure. With respect to the latter, prior research has suggested that lower class children are in fact, more impulsive, as measured by the MFFT, than are middle class children (e.g., Heider, 1971; Schwebel , 1966). Race was not studied here primarily because of the relatively small number of non-white children in the total sample (N=51). This variable, as well as social class should, however, be examined within future attempts to refine this scale, and in so doing, possibly broaden its

PAGE 48

39 applicability in terms of subject populations for which it could be of some use. Related to this issue is the already touched upon issue of factor structure comparisons. Given the rather exploratory nature of this study, the decision was made to analyze the results for males and females separately. The resulting analyses indicated that while there was content similarity across the two dimensions for both groups, the factor structure for males was somewhat more robust in terms of the absolute number of items loading significantly on the factors, and the percentage of variance accounted for by the factors, than that for the females. The desire to broaden the applicability of the scale could argue for attempts to compare the factor structures across criterion groups in order to examine the degree to which they share common variance. The goal in carrying out such an analysis would be to see if it were possible to develop a single scale for each dimension which would have validity across criterion groups. However, because of the seminal nature of this study, this was left for future investigations, following the refinement of the factors in terms of the addition and deletion of scale items in order to increase the percentage of variance accounted for by the individual factors. Subsequent investigations may, at some point, attempt to examine the extent to which factor structures are comparable across different criterion groups (i.e., males vs. females; blacks vs. whites; or a

PAGE 49

40 combination of the two). A number of empirical methods for establishing the invariance of factor structures have been reported (e.g., Harman, 1967; Rummel , 1970). On an empirical basis, because it provides an estimate of invariance with confidence intervals associated with those estimates, the coefficient of invariance described by Katzenmeyer and Stenner (1975, 1977) may be best suited for this purpose. Given the attempt to examine the comparability of factor structures across criterion groups, the empirical question arises as to what, if anything, a composite scale (e.g., based upon male and female performance combined) would add to prediction to a particular criterion (e.g., in the case of impulsivity, possibly academic performance) over and above that which is achieved based upon a scale derived from an individual criterion group. This is a question open to empirical investigation through the use of multiple regression techniques. Several potential uses for the impulsivity and reflectivity scales derived from the factor analysis of RISC performance, immediately present themselves. Both scales would appear to have direct research applicability. Their ability to identify impulsive and reflective subjects for study should, in future investigations, allow for the further delineation of the correlates of this dimension.^^The development of a reliable, valid method for identifying impulsive and reflective subjects is a necessary first step in any attempt to

PAGE 50

41 conduct research as to the origins of these types of behaviors, the consequences of these types of behaviors, the types of environmental contingencies which could affect them, as well as uncovering possible mediating personality variables. While the temptation is to hypothesize about direct clinical applications for these scales, the present findings, especially with respect to the relationship of factor scores to achievement test data, suggest that at this point, such an attempt would be premature. Rather, these findings suggest a number of potential avenues for further research, particularly with respect to the examination of the relationship between RISC scores and a number of other environmental/organismic variables, several of which will be discussed below. However, all of these are potential uses for the scales, which will depend for their achievement upon the refinement of the scales to the point where they could be considered reliable and valid measures of impulsivity and reflectivity. Of necessity, the next step in what may be viewed as a step-wise research program examining the utility of the RISC would be to refine the individual scales. The goal here would be to increase their reliability and validity, increase the percentage of variance accounted for by the individual factors, while also broadening the spectrum of behaviors accounted for by the dimensions .

PAGE 51

42 The results of the present investigation indicate that a number of the scale items, with negligible loadings on either of the interpretable factors, may be deleted during subsequent uses of the scale. In addition, given the quality of the items loading significantly on the two interpretable dimensions, it is also likely that new items could be constructed which would theoretically tap the same domain as those items being retained. These new items would include both general items indicative of impulsive and reflective behavior (e.g., item 10, "I am restless/' for impulsivity; item 20, "I usually plan what I am going to do before I do it," for reflectivity), and more domain specific items such as those suggesting impulsive and reflective behavior in the classroom (e.g., item 5, "When I do my homework, I take my time to make sure that I do the best job that I can," for reflectivity). The inclusion of the latter type of item stems from the observation that, particularly with respect to those items loading significantly for the males in the total subject pool, several were indicative of behaviors specific to academic settings and acti vi ti es . The prospect of developing domain specific items merits further attention. It may be that in future investigations it will be possible to tap impulsive and reflective behavior in a number of different situations. These may include situations in the home and in interpersonal settings as well as those in school. To the extent that

PAGE 52

43 this was achieved, it might then be possible to develop what would amount to an impulsivity-reflectivity profile for a given child. This would address itself to the manner and extent to which these types of behaviors manifest themselves across a wide variety of situations. There is in fact some theoretical precedent for considering this approach. Block et al. (1974), in their investigation into the utility of a measure of cognitive impulsivity, suggested that the impulsivity dimension may be sufficiently broad to require a number of assessment tools, each sensitive to the impact of specific situational variables. This may be extended to include the impact of organismic variables on the quality and quantity of impulsive and reflective behavior. In both cases, the specific variables, and their potential impact are open to empirical investigation. In discussing methods for refining the scales, it is necessary to deal with the somewhat perplexing finding that for both males and females in this study, higher achievement test scores were associated with low reflectivity factor scores. Furthermore, for females, higher achievement test scores were also associated with high impulsivity factor scores. This is of interest in light of the fact that a number of prior investigators (e.g., Douglas, 1974) have maintained that for certain populations (e.g., hyperactive children), impulsivity was associated with academic performance deficits. In

PAGE 53

/M addition, the thrust of the impulsivity modification literature (E.fi. , Meichenbaum and Goodman, 1971) has been to reduce impulsive responding and increase reflective responding in an effort to increase academic nerformance levels. To this point in the Dsychological literature, the majority of the research attempting to relate impulsivity to academic performance deficits have focused on selected subject pooulations, such as hyperactive children. The focus of the present investigation upon the performance of normal children, highlights the need for extreme caution in terms of extending the conceptualization of impulsivity based on one population to a qualitatively different population of subjects. What this finding may in fact suogest, is that impulsivity represents an individual difference dimension wherein its manifestations and the variables affecting it, differ across various oooulations. Emphasizing the concept of impulsivity as an individual difference dimension, one possible explanation for this finding refers to the possibility that impulsivity and reflectivity as they are tapped by the RISC, represent domains apart from the types of behaviors necessary for adequate functioning on cognitive tasks such as the Metropolitan Achievement Tests. The fact that scores on both dimensions correlate significantly with teacher ratings of impulsivity suqgests that the dimensions may be i nteroretable as being indicative of management or conduct problems which do not necessarily have to have aversive consequences for academic performance in all children. One method for examining this possibility would be to investigate the relationship between performance on the RISC impulsivity and reflectivity dimensions and the child's status on the Behavior Problem Checklist (Oua.y and Peterson, 1967). In particular, an examination of the relationship

PAGE 54

45 between scores on both RISC dimensions, and placement on the Behavior Problem Checklist conduct problem dimension could potentially give support to the hypothesis that, at least for some children, impulsivity and reflectivity as measured by the RISC may reflect management/conduct problems not necessarily related to academic performance. That body of psychological literature attempting to relate behavioral impulsivity and cognitive impulsivity as measured by Kagan's Matching Familiar Figures Test (Kagan et al . , 1964) suggests another possible explanation for this finding. A number of investigators have examined the relationship between these two variables (e.g., Ault, Crawford and Jeffrey, 1972; Bentler and McClain, 1976; Bjorklund and Butter, 1973; Nadeau, 1968). In general, the results of these investigations have suggested that cognitive impulsivity, at least as it is measured by the Matching Familiar Figures Test was not related to a broader, more generalized tendency towards impulsive behavior. While these results are not conclusive, particularly because of methodological problems with the Matching Familiar Figures Test, and more generalized problems with the impulsivity/reflectivity cognitive style dimension, they do suggest that it is possible for a child to be behaviorally impulsive but not cognitively impulsive, or vice versa. To the extent that, as Kagan and his, associates have maintained, cognitive impulsivity relates to performance on cognitive

PAGE 55

46 tasks, the outcome mentioned above could account for the present finding. In any case, a subsequent investigation using the RISC could begin to generate data to respond to this hypothesis by administering the RISC and the Matching Familiar Figures Test (and possibly the Behavior Problem Checklist) to a sufficiently large population of subjects, and then correlating the results. There is yet a third possible explanation for these findings. It may be that the children included in this study were not very accurate self-reporters due to confusion with regards to how the items had to be responded to, or a lack of attention to item content. This possibility receives some support from an examination of those items included in the scale, which had both positive and negative forms. The correlation of comparable items yielded the expected significant negative correlation in only three instances out of six for the males, and in only two instances out of six for the females. Given Hirshfield's (1965) empirically supported contention that the use of a negative form of a normally positive item on the ISC does not significantly increase its cognitive complexity, one could hypothesize that these findings suggest a lack of attention to item content, a lack of motivation, or confusion. However, the magnitude of the factor loadings, and the overall consistency in the data, cast doubt upon this latter possibility.

PAGE 56

47 A more plausible explanation, again related to the issue of accuracy in the subjects' self-reporting, relates to the possibility that there exists, a discrepancy between the childrens' perceptions of their own behavior, and others' perceptions of the childrens' behavior. That is, clinically, it is not uncommon to see a fair amount of dissonance between a child's self-reported behavior, and that of adults rating the child's behavior. One method for testing out this possibility would be to administer the scale, in a subsequent study, not only to another sample of children, but also to their teachers. The teachers would be asked to complete the questionnaire based on how they felt the children would respond if their perceptions of their own behavior were objectively accurate. Additionally, the teachers could be asked to complete a versions of the RISC which would use the same items in terms of content, but which would assess the teachers' perceptions of the children's behavior. In summary, the RISC was developed as a self-report instrument for measuring a broad dimension of impulsive behavior. The results of the present investigation yielded two behavioral dimensions, as measured by the RISC, for both males and females: an "impul si vi ty" dimension and a "reflectivity" dimension. While there are clear

PAGE 57

48 implications for the use of such scales needed to replicate the findings of the further refine both the impulsivity and much additional work is present study, as well as to reflectivity dimensions.

PAGE 58

APPENDIX A REVISED IMPULSIVITY SCALE FOR CHILDREN (RISC) NAME There will be fifty-four items on this questionnaire. Read each item to yourself, as I read it aloud. Then, using the scale to the right of each item, decide how often you engage in that particular behavior A "1" means never, a "3" means sometimes, and a "5" means always. After you have decided upon your answer, circle the appropriate number, to the right of each item. *1. I don't think I am as happy as other people. Never 1 2 Someti mes 3 4 Always 5 *2 . I play hooky sometimes. 1 2 3 4 5 *3. I'm not known as a hard and steady worker. 1 2 3 4 5 ***4. I like to show off in front of my friends. 1 2 3 4 5 When I do my homework, I take my time to make sure that I do the best job that I can. 1 2 3 4 5 *6. I like to keep moving around. 1 2 3 4 5 I donÂ’t like going to school. 1 2 3 4 5 ***g I like to bother other children at school . 1 2 3 4 5 I usually have enough time to do everything I want to do everyday. 1 2 3 4 5 *10. I am restless. 1 2 3 4 5 49

PAGE 59

***1 1 . ** 12 . ** 13 . ****14 . ****1 5 . ***16 . **** ] 7 . *18. *19. **** 20 . *** 21 . *** 22 . *23. *24. *25. 50 I have trouble sitting still. I don't like being "it" when we play games of that sort. I don't like to shoot with bows and arrows. Never Sometimes 1 2 3 1 Always 4 5 My school grades are important to 1 2 me. I finish all of my homework before 1 2 I go out to play. I don't like to tell on my friends 1 2 when they misbehave. 4 5 4 5 4 5 In school, I speak only when I am 1 spoken to by the teacher. Whenever there is a fire-engine 1 going someplace, I like to follow it. Sometimes, I can hardly stop from 1 throwing snowballs at people I see walking by. I usually plan what I am going to 1 do, before I do it. 4 5 4 5 I get angry quickly. 12 345 I often break things that belong 12 345 to me and my friends. I like to wrestle and horse around. 1 2 I like to just blow off steam. I'll try anything. 3 3 4 4

PAGE 60

51 *26. It's hard to stick to the rules if you're losing the game. ****27. I don't have to be reminded to do my chores at home. *28. I often act on the spur of the moment, without stopping to think. ****29. I can read for half an hour or an hour without getting up from my seat. *30. I must admit, I'm a pretty good talker. Neve r Sometimes 1 2 3 3 4 3 4 3 4 *31. When things get quiet, I like to stir up a little fuss. *32. My home life is not always happy. ***33. I don't like it when the teacher pays attention to other children and not me. *34. I don't think you should always have to do what you are told. *35. I like to go with lots of other ki ds , not just one . ***36. I get so angry sometimes, that I want to kick, scream, and throw things. *37. I get into tricks at Halloween. *38. I like to dare kids to do things. ****39. In class, I make sure I know the answer before I raise my hand to answer a question. 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 3 3 3 4 4 4 4 4 Always 5 5 5 5 5 5 5 5 5 5 5 5 5 5

PAGE 61

52 Never Sometimes Always *40. I like throwing stones at targets. 12 345 *41. I usually say the first thing that comes into my head. 1 *42. I like being "it" when we play games of that sort. 1 2 4 5 **43. I never play "hooky 1 1 2 4 5 **44. I am not restless. 1 2 4 5 *45. I make friends quickly. 1 2 3 4 5 ***46. I don't like to listen to the teacher when she tells me what to do. 1 2 3 4 5 *47. I like to shoot with bows and arrows. 1 2 3 4 5 **48. I think I am as happy as other people. 1 2 3 4 5 ****49. I go to bed without being told that I have to. 1 2 3 4 5 ****50. When I have to, I take turns at school . 1 2 3 4 5 *51. It's fun to push people off the edge, into the pool. 1 2 3 4 5 ****52. I keep my room at home, neat and clean. 1 2 3 4 5 ***53. I use words that my parents say I shouldn' t use. 1 2 3 4 5 **54. I don't like to dare kids to do things . 1 2 3 4 5

PAGE 62

53 * ** Items from the original Impulsivity Scale for Children (ISC). Behavioral self-control marker items. Conduct problem marker items. Market items keyed in the negative.

PAGE 63

54 IMPULSI VITY SCALE FOR CHILDREN Name There will be twenty-five items on this questionnaire. Read each item to yourself as I read it aloud. Then, decide whether it is true about you or false about you. If it is true or mostly true, circle the "T"; if it is false or not usually true about you, then circle the "F". 1. I like to dare kids to do things. T 2. I am restless. T 3. I play hooky sometimes. T 4. I like to wrestle and horse around. T 5. I must admit, I'm a pretty good talker. T 6. I like to shoot with bows and arrows. T 7. I like throwing stones at targets. T 8. My home life is not always happy. T 9. When things get quiet, I like to stir up T a little fuss. 10. I make friends quickly. T 11. I like to just blow off steam. T 12. I don't think I am as happy as other people. T 13. I'll try anything. T 14. I'm not known as a hard and steady worker. T 15. It's fun to push people off the edge, into T the pool . 16. Sometimes, I can hardly stop from throwing T snowballs at people I see walking by. I usually say the first thing that comes into T my head. F F F F F F F F F F F F F F F F F 17.

PAGE 64

55 18. I like to keep moving around. T F 19. I like to go with lots of other kids, T F not just one. 20. I don't think you should always have to do T F what you are told. 21. I get into tricks at halloween. T F 22. I often act on the spur of the moment, T F without stopping to think. 23. It's hard to stick to the rules, if you're T losing the game. 24. I like being "it", when we play games T of that sort. Whenever there is a fire engine going T someplace, I like to follow it. F F 25. F

PAGE 65

APPENDIX B TEACHER RATING SCALE FOR IMPULSIVITY DATE : TEACHER'S NAME: Impulsivity in children is the tendency of a child to respond quickly and without reflection or delay, to be restless, to indulge in horseplay, lose control of his feelings, break the rules, enter activities with overwhelming vigor, and generally to lose control in acceptable or unacceptable ways. Using the scale below, rate each child listed in terms of the degree to which his/her everyday classroom behavior reflects impulsivity as it is defined above. Child's Name: 1. Not at all impul si ve 1 2 3 4 5 6 7 8 9 10 Extremely impul si ve 11 2. 1 2 3 4 5 6 7 8 9 10 11 3. 1 2 3 4 5 6 7 8 9 10 11 4. 1 2 3 4 5 6 7 8 9 10 11 5. 1 2 3 4 5 6 7 8 9 10 11 6. 1 2 3 4 5 6 7 8 9 10 11 7. 1 2 3 4 5 6 7 8 9 10 11 8. 1 2 3 4 5 6 7 8 9 10 11 9. 1 2 3 4 5 6 7 8 9 10 11 10. 1 2 3 4 5 6 7 8 9 10 11 11 . 1 2 3 4 5 6 7 8 9 10 11 12. 1 2 3 4 5 6 7 8 9 10 11 13. 1 2 3 4 5 6 7 8 9 10 11 14. 1 2 3 4 5 6 7 8 9 10 11 15. 1 2 3 4 5 6 7 8 9 10 11 16. 1 2 3 4 5 6 7 8 9 10 11 17. 1 2 3 4 5 6 7 8 9 10 11 18. 1 2 3 4 5 6 7 8 9 10 11 19. 1 2 3 4 5 6 7 8 9 10 11 20. 1 2 3 4 5 6 7 8 9 10 11 56

PAGE 66

APPENDIX C FACTOR MATRICES FOR MALES AND FEMALES Factor Pattern For Males Item # Factor 1 Factor 2 1 -0.02174 0.04875 2 0.26010 -0.00859 3 0.11762 0.08017 4 0.24038 -0.24888 5 -0.00253 0.50008 6 0.43490 0.01690 7 -0.13078 -0.19421 8 0.26286 -0.42288 9 0.00039 0.29109 10 0.49112 -0.03184 11 0.46231 -0.03068 12 0.06102 -0.11236 13 0.11216 0.05513 14 0.02005 0.33296 15 -0.15365 0.51213 16 0.02974 -0.15976 17 -0.22714 0.40387 18 0.29160 -0.00344 19 0.17188 0.15306 20 0.05636 0.50440 21 0.33172 -0.21581 22 0.20261 0.26601 Factor 3 Factor 4 Factor 5 0.47510 0.04602 0.00751 0.08655 -0.25755 -0.06270 0.37175 -0.21757 0.07228 -0.10774 0.13047 -0.02529 -0.02558 0.02297 0.03539 -0.00160 0.13434 -0.03269 0.05836 -0.08483 -0.32717 0.09428 -0.12127 -0.02616 -0.22662 0.04382 0.09067 0.14706 0.11250 0.09450 0.25531 0.08362 0.15493 -0.00551 0.03489 -0.09116 -0.02528 -0.21775 0.42177 0.03745 -0.13913 -0.03135 0.14181 -0.02094 0.04737 0.01960 0.44739 -0.06134 0.15824 0.09380 -0.01848 -0.10436 -0.14209 -0.38111 0.17556 -0.02634 -0.39621 -0.19497 0.10535 0.01228 0.08082 -0.19006 0.34984 0.06989 0.02925 0.06996 57

PAGE 67

58 Item # Factor 1 Factor 2 23 0.53136 -0.09763 24 0.48098 0.01455 25 0.33423 -0.05066 26 0.36732 -0.06618 27 0.07419 -0.02083 28 0.32436 -0.10206 29 -0.04434 0.13231 30 0.42219 0.06903 31 0.37587 -0.23135 32 -0.00853 -0.15921 33 0.19610 -0.02988 34 0.06216 0.02197 35 0.30899 0.06324 36 0.45583 -0.05452 37 0.19202 0.05881 38 0.42293 -0.10468 39 0.02145 0.32615 40 0.35151 -0.05585 41 0.40296 -0.03045 42 -0.06027 0.04107 43 0.03116 -0.24630 44 0.02809 -0.01589 45 0.11630 0.31775 46 -0.07761 -0.18657 47 0.03712 -0.02814 48 0.09015 0.24435 49 -0.00284 0.43911 50 -0.04304 0.31530 51 0.20514 -0.14024 Factor 3 Factor 4 Factor 5 -0.09488 -0.01852 -0.06259 -0.05593 0.07640 -0.01506 0.01142 0.04504 -0.17866 0.12873 -0.18250 -0.03761 -0.15589 -0.17855 0.06318 -0.01714 -0.07890 0.05960 -0.10958 0.30207 0.08051 -0.08074 -0.11046 -0.01078 -0.03494 -0.19968 -0.23199 0.33270 0.33198 -0.04522 0.27329 -0.03452 0.10450 0.34858 -0.03398 -0.13750 0.14775 0.04947 -0.09109 0.00868 -0.07544 -0.10362 0.04166 -0.19690 -0.50649 -0.06467 -0.21613 -0.27185 0.05805 0.18946 0.09677 -0.08420 -0.08516 -0.07119 0.02161 -0.05601 0.00320 -0.16050 0.08953 0.11320 -0.47267 0.33403 -0.02472 0.04490 0.05349 -0.06839 -0.18037 -0.11379 -0.24970 0.05459 -0.39604 0.14786 -0.07711 0.03280 -0.16639 -0.54022 -0.03480 -0.09972 0.06776 0.04405 0.06587 -0.07724 0.30707 0.13467 -0.05028 -0.28720 -0.38801

PAGE 68

59 Item # Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 52 -0.16499 0.46950 -0.01631 -0.08206 -0.03659 53 0.24174 -0.15164 -0.07950 -0.10096 -0.25445 54 -0.01298 0.00314 -0.07939 0.30777 0.08433 Factor Structure for Males tern § Factor 1 Factor 2 1 0.00063 0.01378 2 0.33975 -0.11673 3 0.15824 0.00799 4 0.28330 -0.30700 5 -0.16803 0.50925 6 0.41018 -0.11019 7 0.03111 -0.20926 8 0.43015 -0.52036 9 -0.14093 0.32696 10 0.46924 -0.17382 11 0.44212 -0.16767 12 0.10932 -0.14051 13 0.03398 0.06923 14 -0.03943 0.31178 15 -0.29971 0.54992 16 0.00282 -0.15490 17 -0.34649 0.45932 18 0.40497 -0.13969 19 0.24686 0.03027 20 -0.13883 0.51327 21 0.35732 -0.28560 22 0.26600 -0.32219 Factor 3 Factor 4 Factor 5 0.46640 0.03440 0.01996 0.12446 -0.31892 -0.14421 0.38453 -0.24794 0.03889 -0.06494 0.07067 -0.10941 -0.07423 0.05339 0.10496 0.03569 0.04409 -0.12914 0.06846 -0.09093 -0.32668 0.16597 -0.20316 -0.15580 -0.25631 0.07401 0.13362 0.19494 0.00998 -0.02442 0.30102 -0.01356 0.04094 0.01035 0.01097 -0.11914 -0.01239 -0.21073 0.38827 0.01376 -0.12892 -0.00018 0.07840 0.03511 0.15223 0.01999 0.42818 -0.06265 0.09330 0.15442 0.09688 -0.06685 -0.22142 -0.46161 0.18152 -0.08497 -0.42036 -0.24125 0.12932 0.07386 0.14087 -0.25126 0.22698 0.11409 -0.02499 -0.01429

PAGE 69

60 -em # Factor 1 Factor 2 23 0.57012 -0.25692 24 0.45899 -0.12189 25 0.38527 -0.17332 26 0.44686 -0.20296 27 0.08567 -0.02928 28 0.35465 -0.19358 29 -0.17678 0.18289 30 0.41867 -0.05688 31 0.53952 -0.38232 32 0.01588 -0.17650 33 0.21414 -0.10210 34 0.13178 -0.04992 35 0.31736 -0.05274 36 0.51398 -0.20959 37 0.34370 -0.08131 38 0.55888 -0.27316 39 -0.13282 0.33740 40 0.39464 -0.16710 41 0.42496 -0.15550 42 -0.13507 0.09430 43 -0.00563 -0.19642 44 0.04326 -0.03489 45 0.08781 0.26031 46 0.02854 -0.16974 47 0.07192 -0.05262 48 -0.00605 0.25319 49 -0.15246 0.44478 50 -0.24125 0.37001 Factor 3 Factor 4 Factor 5 -0.03066 -0.13276 -0.20720 -0.01148 -0.02015 -0.12648 -0.04920 -0.03771 -0.26501 0.17982 -0.26888 -0.14883 -0.13950 -0.18470 -0.03184 0.02833 -0.14650 -0.03877 -0.13916 0.32761 0.12845 -0.03990 -0.19078 -0.11180 0.03427 -0.30204 -0.36784 0.33373 0.30903 -0.04569 0.29691 -0.08089 0.04875 0.35485 -0.06810 -0.15363 0.17159 -0.02211 -0.15628 0.06366 -0.17859 -0.22785 0.06600 -0.26611 -0.55819 -0.00180 -0.32265 -0.40308 0.02127 0.20610 0.14687 -0.03945 -0.16124 -0.17007 0.06764 -0.14096 -0.10364 -0.17466 0.11754 0.13988 -0.45956 0.33202 -0.04265 0.04743 0.04084 -0.07436 -0.19281 -0.12884 -0.24165 0.07960 -0.38314 0.11709 -0.07119 0.01651 -0.17692 -0.55228 -0.02464 -0.08849 0.02385 0.06941 0.12815 -0.12446 0.34416 0.20711

PAGE 70

61 item § Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 51 0.39636 -0.26443 -0.00272 -0.35870 -0.47490 52 -0.28115 0.51105 -0.07397 -0.02477 0.06231 53 0.36233 -0.25607 -0.03531 -0.17115 -0.34023 54 -0.10585 0.04231 -0.09382 0.31901 0.10734 Factor Pattern for Females tern # Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 1 0.22870 -0.20494 0.07888 0.03035 -0.04825 2 -0.02815 0.02032 -0.24397 -0.01009 -0.08177 3 0.01363 -0.33632 -0.08022 -0.01897 0.06419 4 0.47484 0.19541 -0.03928 0.02818 0.02969 5 -0.01661 -0.15741 0.44032 0.34878 -0.14730 6 0.36791 -0.06327 0.07616 0.06403 0.05123 7 0.07200 0.06780 -0.15109 -0.11551 0.13544 8 0.30574 -0.14309 -0.10053 -0.12740 0.06352 9 0.02766 0.08821 0.10477 0.07710 0.14193 10 0.33825 -0.03231 0.08035 -0.20033 0.07644 11 0.48894 -0.04213 0.03491 0.00057 0.03334 12 0.11286 0.05195 0.04171 -0.18852 0.12794 13 -0.11029 0.29965 0.16941 -0.06370 0.18790 14 -0.01317 0.07723 0.16441 0.35234 -0.16462 15 -0.16870 -0.13633 0.42459 0.16490 -0.00067 16 0.09125 0.34882 -0.02760 -0.19549 0.15443 17 -0.19453 -0.19878 0.26432 -0.15730 0.09939 18 0.06282 -0.19936 -0.05332 -0.00173 0.25507 19 0.03846 -0.10152 0.02712 0.09804 0.50142 20 0.04985 -0.00983 0.45713 -0.02880 -0.08535 21 0.42109 -0.01846 0.00675 -0.17482 -0.06196

PAGE 71

62 tern # Factor 1 Factor 2 22 0.24004 -0.23041 23 0.44995 0.10895 24 0.41075 0.12157 25 0.25270 0.08886 26 0.19763 -0.29771 27 0.01658 -0.07085 28 0.30072 0.12232 29 -0.00655 0.10841 30 0.39638 -0.01433 31 0.36975 -0.08341 32 0.27889 -0.24774 33 0.25271 -0.06908 34 0.06124 -0.36717 35 0.1 6315 -0.16305 36 0.32959 -0.14391 37 -0.10587 -0.25296 38 0.11465 -0.09891 39 0.04470 0.18052 40 0.03959 0.03773 41 0.21285 -0.24037 42 -0.10521 0.10749 43 0.16601 0.58980 44 0.02899 0.17892 45 0.02137 -0.09568 46 0.16764 -0.33978 47 -0.05027 0.14346 48 -0.11510 0.22291 49 -0.12157 -0.09626 Factor 3 Factor 4 Factor 5 -0.00367 -0.04469 0.16887 -0.18837 0.03524 0.26897 -0.00515 -0.02312 -0.00871 -0.29003 0.15000 0.10849 -0.31339 0.02788 -0.00058 -0.03879 0.11885 0.11606 -0.24378 -0.18821 0.16435 0.44263 -0.06201 0.26955 -0.32777 0.20070 -0.09225 -0.45874 0.11980 0.13108 0.04431 -0.34365 0.06823 0.18417 -0.02737 -0.07742 0.05573 -0.09626 0.24827 0.04185 0.21717 -0.10134 -0.10878 -0.01567 0.07271 -0.37035 0.02475 0.23111 -0.45154 0.07546 0.20914 0.38456 0.06187 0.06836 0.07425 0.00321 0.58375 -0.15579 -0.11641 0.06900 0.12345 0.13844 0.16438 -0.07511 -0.02232 0.02092 -0.12364 0.19710 0.10805 0.04449 0.45226 0.13365 -0.10698 -0.12603 0.08866 -0.00772 -0.21177 0.43337 -0.07934 0.41454 0.05362 0.12953 0.06900 0.04075

PAGE 72

51 52 53 54 tem 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 63 Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 0.06989 0.20827 0.27351 0.16287 0.06498 0.01391 -0.01532 -0.29909 -0.06231 0.46464 -0.14165 -0.22973 0.35125 0.17379 -0.15895 0.18993 0.03448 -0.35369 0.12509 0.28344 0.02922 0.42750 -0.00357 -0.02530 0.06866 Factor Structure for Females Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 0.22252 -0.23269 0.01801 0.00309 -0.02669 0.01139 -0.00167 -0.22350 -0.02748 -0.04732 0.09569 -0.33975 -0.12136 -0.04343 0.05284 0.46121 0.13162 -0.14227 -0.02743 0.15897 -0.17928 -0.11994 0.48993 0.39171 0.24363 0.36160 -0.10056 -0.02044 0.02219 0.11556 0.16305 -0.08247 -0.20675 -0.14549 0.17337 0.37981 -0.19196 -0.20874 -0.18231 0.10450 0.01278 0.10721 0.08968 0.08357 0.13670 0.36507 -0.07354 -0.03652 -0.23756 0.14307 0.49375 -0.10298 -0.09310 -0.06007 0.13518 0.14792 0.04301 -0.02071 -0.19979 0.15626 -0.14179 0.34056 0.18309 -0.02683 0.16233 -0.14461 0.09183 0.23425 0.37753 -0.19695 -0 . 274'44 -0.07458 0.47161 0.22194 -0.12229 0.10988 0.33941 -0.06694 -0.1)9994 0.21356 -0.23526 -0.16657 0.29707 -0.11254 -0.19777 0.16090 -0.19098 -0.12474 -0.02940 0.26142 0.14717 -0.05944 -0.06036 0.07843 0.49480 -0.07625 0.01021 0.45498 0.01169 -0.14694

PAGE 73

64 tem # Factor 1 Factor 2 21 0.42948 -0.08697 22 0.31596 -0.25104 23 0.53782 0.05714 24 0.39641 0.06390 25 0.31757 0.04726 26 0.31093 -0.34942 27 0.04730 -0.06184 28 0.40418 0.06910 29 -0.06057 0.16296 30 0.43268 -0.09285 31 0.50802 -0.15309 32 0.35953 -0.28999 33 0.20292 -0.09669 34 0.16556 -0.35450 35 0.12522 -0.18180 36 0.39408 -0.19151 37 0.06860 -0.24691 38 0.27643 -0.12877 39 -0.06582 0.21195 40 0.14831 0.08632 41 0.31349 -0.28000 42 -0.12975 0.15018 43 0.11194 0.56240 44 0.03511 0.18217 45 -0.00207 -0.(76641 46 0.27546 -0.36831 47 0.05658 0.17712 48 -0.16495 0.25312 Factor 3 Factor 4 Factor 5 -0.10473 -0.22520 0.03563 -0.11121 -0.08823 0.20608 -0.32934 -0.04173 0.40905 -0.09704 -0.06933 0.09590 -0.34757 0.09098 0.21530 -0.38161 -0.03884 0.06886 -0.05513 0.10710 0.11695 -0.35248 -0.24883 0.28655 0.40349 -0.02092 0.20814 -0.39133 0.12162 0.04326 -0.56459 0.02246 0.27791 -0.08725 -0.38530 0.11323 0.12683 -0.04117 -0.05444 -0.03631 -0.11947 0.22558 0.02694 0.19756 -0.09033 -0.21341 -0.07461 0.15333 -0.39826 -0.01457 0.24460 -0.51305 0.00761 0.29690 0.38261 0.09921 0.03043 -0.02526 -0-0016:8 0.58392 -0.24855 -0.16922 0.12541 0.14472 0.16329 0.12598 -0.07606 -0.02760 0.11979 -0.11497 0.18553 0.14377 0.05490 0.44661 0.11130 -0.20037 -0.17291 0.11909 -0.07426 -0.21238 0.44075 -0.00216 0.42828 0.04737

PAGE 74

65 Item # Factor 1 Factor 2 49 -0.13944 -0.06381 50 -0.03050 0.23147 51 0.20208 -0.00420 52 -0.25381 -0.18996 53 0.32046 0.00997 54 -0.00914 0.42794 Factor 3 Factor 4 Factor 5 0.15211 0.09181 -0.01726 0.27795 0.18741 0.05002 -0.38381 -0.10643 0.51588 0.41061 0.22092 -0.27069 -0.43038 0.06074 0.38234 0.00858 -0.01434 0.11180

PAGE 75

REFERENCES Ault, R.L., Crawford, D.E., and Jeffrey, W.E. Visual scanning strategies of reflective, impulsive, fast-accurate and slowinaccurate children on the Matching Familiar Figures Test. Child Development , 1972 , 4F3, 1412-1417. Ault, R.L., Mitchell, C., and Hartmann, D.P. Some methodological problems in reflection-impul si vi ty research. Child Development , 1976, 47, 227-231 . Bentler, P.M., and McClain, J. A mul ti trai t-mul timethod analysis of reflection-impulsivity. Child Development , 1976, 47, 218-226. Bjorklund, D.F., and Butter, E.J. Can cognitive impulsivity be predicted from classroom behavior? The Journal of Genetic Psychology , 1973, 123 , 185-194. Block, J., Block, J.H. and Harrington, D.M. Some misgivings about the Matching Familiar Figures Test as a measure of reflectionimpulsivity. Developmental Psychology , 1974, 1_0, 61 1-632. Comrey, A.L. A first course in factor analysis . New York: Academic Press, 1973. Douglas, V.I. Sustained attention and impulse control: Implications for the handicapped child. In J.A. Swets and L.L. Elliot (Eds.) Psychology and the handicapped child . Washington, D.C.: U.S. Department of Health, Education and Welfare, DHEW Publication Number (OE) 73-05000, 1974. Egeland, B. Training impulsive children in the use of more efficient scanning techniques. Child Development , 1974, 45, 165-171. Egeland, B., and Weinberg, R.A. The Matching Familiar Figures Test: A look at its psychometric credibility. Child Development , 1976, 47, 483-491. Harman, H. H. Modern factor analysis . Chicago: The University of Chicago Press, 1967. 66

PAGE 76

67 Heider, E. R. Information processing and the modification of an "impulsive conceptual tempo". Child Development , 1971, 42, 1276-1281. Hirschfield, P. Response set in impulsive children. The Journal of Genetic Psychology , 1965, 107 , 117-126. Kagan, J. Individual differences in the resolution of response uncertainty. Journal of Personality and Social Psychology , 1965, 2, 154-160. (a) Kagan, J. Reflection-impulsivity and reading ability in primary grade children. Child Development , 1965, 36_, 609-628. (b) Kagan, J. Impulsive and reflective children: Significance of conceptual tempo, in J.D. Krumbol tz (Ed.), Learning and the educational process . Chicago: Rand McNally, 1965. (c) Kagan, J., Rosman, B. L. , Day, D., Albert, J., and Phillips, W. Information processing in the child: Significance of analytic and reflective attitudes. Psychological Monographs , 1964, 78(1), Whole No. 578. Katzenmeyer, W. G. and Stenner, A. J. Strategic use of random subsample replication and a coefficient of factor replicability. Education and Psychological Measurement , 1975, 35_, 19-29. Katzenmeyer, W. G. and Stenner, A. J. Estimation of the invariance of factor structures across sex and race with implications for hypothesis testing. Education and Psychological Measurement , 1977, 37, 111-119. Kendall, P. C., and Finch, A. J., Jr. A cognitive-behavioral treatment program for impulse control: A case study. J ournal of Consulting and Clinical Psychology , 1976, 44, 852-857. Kendall, P.C., and Finch, A.J., Jr. A cognitive-behavioral treatment for impulsivity: A group comparison study. Jou rnal of Con sul ting and Clinical Psychology , 1978, 46(1), 110-118. Meichenbaum, D., and Goodman, J. Training impulsive children to talk to themselves: A means of developing self-control. Journal of Abnormal Psychology , 1971, 77, 115-126.

PAGE 77

68 Nadeau, G. H. Cognitive style in pre-school children: A factor analytic study. Unpublished doctoral dissertation, University of Minnesota, 1968. Nie, N. H., Hull, C. H., Jenkins, J. G. , Steinbrenner , K., and Bent, D. H. Manual for the statistical package for the social sciences. New York: McGraw-Hill, 1975. Quay, H. C., and Peterson, P. R. Manual for the behavior problem checklist. Champaign, 111.: Children's Research Center, University of Illinois, 1967. Rummel , R. J. Applied factor analysis . Evanston: Northwestern University Press, 1970. Schwebel , A. I. Effects of impulsivity on performance of verbal tasks in middle and lower class children. American Journal of Orthopsychiatry , 1966, 36, 12-21. Sutton-Smi th , B., and Rosenberg, B. G. A scale to identify impulsive behavior in children. The Journal of Genetic Psychology , 1959, 95, 211-216. Vacc, N. A., and Mercurio, S. Two measures and teacher ratings of impulsivity. Behavioral Disorders , 1978, 2 ( 2 ), 114-115. Weinreich, R. J. Inducing reflective thinking in impulsive, emotionally disturbed children. Unpublished master's thesis, Virginia Commonwealth University, 1975.

PAGE 78

BIOGRAPHICAL SKETCH Joel Cohen was born in New York City on June 23, 1952. He attended PS 69, and Joseph Pulitzer Junior High. From 1966-1969 he attended Newtown High School. He attended the State University of New York at Buffalo from 1969-1973, where he obtained a Bachelor of Arts degree. Mr. Cohen began graduate school at Northern Illinois University, in Dekalb, Illinois, from 1973-1976, when he received hi s master ' s degree in clinical psychology. He completed his clinical internship at the Shands Hospital, University of Florida, in 1978. He then transferred to the University of Florida clinical psychology program in order to complete his work towards receiving his doctoral degree. He hopes to receive his doctoral degree in August, 1979. Subsequently, he plans to accept a position as a staff psychologist at the newly opened Medical Center at the State University of New York at Stony Brook. 69

PAGE 79

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. iAi y Al Wiley C. fjasbury, Ph.D. 1 Associate 'Professor, Clinical' Psychology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. ' ^ Cynthia D. 3elar, Ph.D. Assistant Professor, Clinical Psychology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Hugh C. ^avi s , Ph.D. Professor, Clinical Psychology

PAGE 80

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. ^ Jacquel i n Goldman, Ph.D. Professor, Clinical Psychology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Larry^oesch , Ph.D. Associate Professor, Counselor Education This dissertation was submitted to the Graduate Faculty of the College of Health Related Professions and to the Graduate Council, and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. Howard K. Suzuki, Ph.D.^' Dean, College of Health Related Professions Harry H. Sisler, Ph.D. Dean, Graduate School