University of Southern California Dissertations and Theses

The reliability and construct validity of the Dimensions of Self-Concept (DOSC)---Form W: Measure for an Air Force sample

(USC Thesis Other)

The reliability and construct validity of the Dimensions of Self-Concept (DOSC)---Form W: Measure for an Air Force sample

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content THE RELIABILITY AND CONSTRUCT VALIDITY OF THE DIMENSIONS OF SELF-CONCEPT (DOSC)— FORM W MEASURE FOR AN AIR FORCE SAMPLE by Barbara McNeely Foraker A Dissertation Presented to the FACULTY OF THE SCHOOL OF EDUCATION UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF EDUCATION May 1993 UMI Number: DP27510 All rights reserved INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted. In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion. Oisssrtafion PubLsbny UMI DP27510 Published by ProQuest LLC (2014). Copyright in the Dissertation held by the Author. Microform Edition © ProQuest LLC. All rights reserved. This work is protected against unauthorized copying under Title 17, United States Code ProQuest LLC. 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, Ml 48106- 1346 This dissertation, written by Barbara McNeely Foraker under the direction o f the Chairperson o f the candidate's Guidance Committee and approved by all members o f the Committee, has been presented to and accepted by the Faculty o f the School o f Education in partial fulfillment of the requirements fo r the degree o f Doctor o f Education. February 3, 1993 Dean Guidance Committee f y f d L t f t r . .. Chairperson DEDICATION To Gregory W. Foraker, my partner, husband, and friend with lasting love for the endless hours of patience, help, and encouragement which provided the opportunity for me to achieve this lifelong goal. 11 ACKNOWLEDGMENTS I would like to express my deepest gratitude to Dr. William B. Michael, Chairman of my committee, for his generous time and dedicated guidance in the development of this work. He is totally devoted to the growth of all his students. I am thankful for my committee members. Dr. Patrick Rooney for providing inspiration and insights during the process, and to Dr. William Millington for his kind and positive encouragement. I am sincerely thankful for the contributions of Karen St. John who gave freely of her time, professional expertise, and friendship throughout the process. I am indebted to Diane Owens for her understanding and support of this endeavor. I am grateful to my Father, Joseph T. McNeely, who taught me to pursue my dreams and believe in my abilities. I am also grateful to my Mother, Eleanor R. McNeely, who inspired me with a love for books and education. 1 will never forget the support, encouragement, and companionship of the Hawaii students who will be lifelong friends and colleagues as a result of this shared journey we have taken together. Ill TABLE OF CONTENTS Page DEDICATION................................................. ii ACKNOWLEDGMENTS........................................... i ü LIST OF TABLES............................................ V Chapter 1. THE PROBLEM..................................... 1 Background Purpose of the Study Importance of the Study Rationale for Selecting the DOSC-Form W Description of the DOSC-Form W Dimensions Research Findings on Validation of the DOSC Research Questions Conceptual Assumptions Delimitations 2. METHODS AND PROCEDURES ................. 28 Sample The Test Instrument Demographic Variables Data Analyses Methodological Assumptions Limitations 3. ANALYSIS AND INTERPRETATION OF RESULTS . . . 39 Analysis of Results Discussion 4. SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS. . 54 Summary Conclusions Recommendations REFERENCES....................................................66 IV LIST OF TABLES Table Page 1. Means, Standard Deviations, and Intercorrelations of Six Factor Subscales of the DOSC-Form W Including Internal- Consistency Estimates of Reliability Along the Principal Diagonal (Decimal Points Omitted) (N = 2 1 2 ) ...............................41 2. Rotated Orthogonal Factor Matrix of 18 Subtests: Varimax Solution (Decimal Points Omitted)................................... 43 3. Rotated Oblique Factor Matrix of 18 Subtests: Promax Solution (Decimal Points Omitted)................................... 44 4. Interfactor Correlations in the Oblique (Promax) Solution (Decimal Points Omitted) . 46 5. Goodness-of-Fit Results for All Models (N = 212).............................. 48 6. Correlation Coefficients of Scores on Each of Six DOSC-Form W Subscales With Each of Five Demographic Variables (210 < N < 212) (Decimal Points Omitted) .............. 51 V CHAPTER I THE PROBLEM Background Self-concept, as a theoretical construct, has been extensively reviewed in the literature. In 1974, Wylie undertook a considerable review of the literature which revealed more than 200 studies involving self-concept. In 1989, she repeated this review and discovered 455 studies (Wylie, 1989). A recent computer search conducted by this investigator disclosed several thousand references relating to the construct of self-concept. However, the majority of the research in 1974 as well as that in 1989, was focused on substantive work rather than on methodological research. ShaveIson, Hubner, and Stanton (1976) reiterated the lack of methodological research and the importance of self-concept construct validation. In examining self-concept as a multifaceted construct. ShaveIson and Bolus (1982) concluded that "general self-concept can be interpreted as distinct from but correlated with academic self-concept" (p. 16). Several researchers have addressed the obscure nature of defining self-concept as a construct, while recognizing its important relationship to academic achievement (LaBenne & Greene, 1969; Wylie, 1974). Self-concept research, in academic settings, has been conducted mainly with children or adolescents. Coopersmith (1967) developed the Self-Esteem Inventory (SEI) in an effort to identify the conditions which lead to development of high self-esteem in children. Over a five-year period, from 1983-1988, Marsh and colleagues (cited in Wylie, 1989) published a number of studies indicating self-concept as a theoretical construct in preadolescents, adolescents, and late adolescents using his Self Description Questionnaire (SDQ) Forms I, II, and III, respectively. More recently. Marsh and Byrne (1990) have correlated gender and multiple dimensions of self- concept for Australian high school students and Canadian university students. Caracosta and Michael (1986) have examined academic self-concept for a sample of university students by using the Dimensions of Self-Concept, Form H. However, until recently, very little has been done in the area of researching self-concept with employed adults. In the past few years, Crowder and Michael (1989a, 1989b, 1991) have developed, revised, and validated a multidimensional self-concept measure, the Dimensions of Self-Concept-Form W (DOSC-Form W), for workers in an employment setting. This research has demonstrated the potential application of a self-concept measure for employed adults in high technology. A basic assumption made by Crowder and Michael (1991) was that ". . . the constructs central to self-concept in an academic setting involving a commitment to learning and achievement are not essentially different from those associated with the self-concept in a demanding work setting . . ." (p. 448). The DOSC-Form W was developed and validated with samples of working adults in a large electronics/computer company (Crowder & Michael, 1989a, 1989b, 1991). A need exists for additional empirical research conducted in diverse employment settings, which could provide further information on the construct validation of the DOSC-Form W. A variety of employment settings is available for examination. Purpose of the Study The primary purpose of this study was to sample a population of military personnel in an effort to estimate the internal-consistency reliability of each of the six factor subscales of the DOSC-Form W (Level of Aspiration, Level of Anxiety, Job Interest and Satisfaction, Leadership and Initiative, Identification vs. Alientation, and Level of Job Stress) and to obtain evidence regarding the construct validity of each of these factor subscales. A secondary purpose was to ascertain what relationships, if any, existed between the standing on each of the six factor subscales of self- concept and each of several selected demographic variables (e.g. level of formal education, number of years of military duty, age, gender, and military rank). Importance of the Study Previous research in the area of self-concept has demonstrated that an important potential may exist for the development of self-concept theory (ShaveIson & Bolus, 1982) and that insight may be gained concerning the relationship between self-concept and achievement (ShaveIson & Bolus, 1982). However, as indicated by Wylie (1989), further empirical data are needed to substantiate the results of this research. Studies conducted by Crowder and Michael (1989a, 1989b, 1991) involving the development, revision, and validation of the DOSC-Form W have contributed to the knowledge of self-concept with employed adults. As stated earlier, the importance of the proposed inquiry was to add additional data regarding the reliability and construct validity of a measure of self-concept in employed adults who comprised a military population. The information gained through this study was thought to have the potential to provide assistance in counseling individuals in vocational transition with regard to optimal occupational placement in both the military and civilian work environments. Rationale for Selecting the DOSC-Form W For a period of more than fifteen years, Michael and colleagues (Caracosta & Michael, 1986; Michael & Smith, 1976; Michael, Smith, & Michael, 1984, 1989; Michael, Denny, Ireland-Galman, & Michael, 1986-1987; Michael, Denny, Knapp-Lee, & Michael, 1984) have carried out extensive research in the area of academic self-concept. As a result, three forms of a standardized measure of academic self-concept entitled Dimensions of Self-Concept (DOSC) were developed. Form E was constructed for children in grades four through six; Form S, for children and adolescents grades seven through twelve ; and Form H, for college students. In each form, five factor scales portraying five hypothesized constructs were delineated as: Level of Aspiration, Anxiety, Academic Interest and Satisfaction, Leadership and Initiative, and Identification versus Alienation. More recently, Crowder and Michael (1989a, 1989b, 1991) have investigated the application of these constructs in developing a fourth form of the DOSC, Form W, for adults in the workplace. Their findings have provided substantial evidence for the reliability and construct validity of the DOSC-Form W. However, further research is needed to add additional empirical data in support of the reliability and construct validity of this measure. A key assumption in the development of the DOSC-Form W was that "... the motivational features underlying learning and achievement in the school setting do not differ appreciably from the motivational characteristics central to job performance and achievement for adults in employment situations" (Crowder & Michael, 1989a, p. 20). Crowder and Michael (1989a) stated: Consistent with this assumption, the experimental form was developed within essentially the same rationale as that employed by Michael and Smith (1976, pp. 522-523), who proposed five basic constructs as central to academic self-concept and who then translated these constructs into the five factor scales already enumerated for the DOSC. The present rationale, however, was extended to include one additional construct: job stress. Thus, consisting of six factor scales, the experimental test for the workplace was entitled Dimensions of Self-Concept, Work Form (DOSC, Form W). The detailed rationale underlying the conceptualization of the six constructs (Level of Aspiration, Anxiety, Job Interest and Satisfaction, Leadership and Initiative, Identification versus Alienation, and Job Stress) which were translated into the six factor scales was as follows: An unrealistic level of aspiration— either too high or too low— was hypothesized to be related to the probable subsequent occurrence of anxiety. Employees who set unrealistically high levels of expectation could well become discouraged, depressed, and fearful (anxious) of loss of status and of possible criticism (symbolic punishment) from supervisors, peers, or others having significant work relationships. Employees who set unrealistically low levels of expectation might already be anxious and fearful of possible failure. Setting low-level goals could generate a certain degree of immediate security (preservation of self esteem) but at the expense of later development of positive attitudes toward learning new jobs, technologies, or procedures and toward opportunities for positive recognition and leadership roles. Highly anxious employees are likely to lose interest in the job, to fail to acquire a sense of satisfaction with their work, to forego opportunities for leadership roles in the work setting, and eventually to develop feelings of alienation and rejection accompanied by a manifestation of hostility toward the work unit supervisor, or even the company as an institution. On the other hand, employees relatively free of anxiety who are successful in light of realistic levels of aspiration attain success that engenders work interest and feelings of satisfaction with the work experience. Such satisfaction can be anticipated to lead to greater self-confidence and to numerous opportunities to exercise initiative and to assume leadership responsibilities, which in turn are reinforcing mechanisms for attaining even greater work satisfaction and interest, for assuming new leadership roles, and for evolving positive identification with the work unit or company. In other words, frequent success leads to further success; repeated failure, to a greater sense of failure, frustration, and alienation. Accompanying the realization of varying degrees in job performance is the probable occurrence from time to time of changing levels of job stress. In its recurring day-to-day manifestations, job stress can be viewed in terms of perceptions held by the worker that he or she is not meeting his or her own expectations or those of significant others. Thus job stress can be thought to arise in conjunction with conflicts created in meeting these expectations that are closely related to one's level of aspiration. Intensification or prolongation of stress can be expected to lead to the manifestation of anxiety and of fear and possibly even terror associated with an anxiety reaction. In short, the six constructs proposed can be viewed as occurring not only sequentially but also recursively in an interactive fashion (pp. 20-22). Description of the DOSC-Form W Dimensions Corresponding to each of the six constructs just discussed, a matching factor subscale of 15 self-report items was developed to represent an operationalization of the relevant construct (Crowder & Michael, 1991). A quoted description of each factor subscale [but referred to as factor scale in the quotation to follow] along with two representative items follows: Level of Aspiration (Factor Scale 1). This factor is a manifestation of patterns of behavior that portray the degree to which achievement levels and work activities of employees are consistent with their perceptions of their potentialities in terms of work aptitude or of past and current attainments. Two typical items are as follows: 1. 1 strive to be the highest ranked employee in my group. 2. 1 work very hard because 1 want to be promoted. Anxiety TFactor Scale 2). This second factor reflects behavior patterns and perceptions associated with emotional instability, a lack of objectivity, and a heightened or exaggerated concern about performance measures, employee ranking, written or oral presentations, and the preservation of self-esteem in relation to work performance. Underlying this dimension is often the presence of a failure syndrome that indicates a marked discrepancy between the stabilized perceptions of what an employee believes she/he can achieve satisfactorily and her/his idealized perception of her/his expectations concerning what her/his supervisors maintain that she/he can do. This attitudinal pattern can become generalized to a self-concept indicative of feeling oneself to be an unworthy, guilt-ridden, individual with a possible need for self-punishment or even self-destruction. Two sample items are as follows: 1. I become quite worried about how I am doing on the job. 2. I find it difficult to participate in group meetings even though I have some good ideas. Job Interest and Satisfaction (Factor Scale 3). This third dimension portrays the sheer love of working and the pleasure and satisfaction gained by employees in working, in doing new projects, and in tackling new assignments— an affective state much like that realized by the dedicated scholar who 8 gains tremendous satisfaction in working in the library, in reading great books, in writing research papers, and in conceptualizing new theories or explanations for observed phenomena— an intrinsic motivation involving working for its own sake. Two representative items are as follows: 1. Work energizes me. 2. Some of my projects are so enjoyable that I hate to see them come to an end. Leadership and Initiative CFactor Scale 4). This fourth factor appears to represent those behavior patterns and perceptions that are associated with star-like qualities, in which an employee has an opportunity to demonstrate his/her mastery of knowledge, to help others, to give direction to group activities, to become the respected expert whom others consult, to put forth (hopefully diplomatically) sound suggestions for work or project activities reflecting the consensus of other employees in a group, to exhibit a willingness to take the initiative in starting a project or assignment— either an individual or group endeavor— and in following it through to successful completion, and to take pride without display of conceit of one's capabilities to do a job quickly and well. Two selected items are as follows : 1. I am appointed to lead special projects. 2. Other members of my work group look to me as a leader. Identification vs. Alienation fFactor Scale 5), This fifth dimension is intended to represent the extent to which an employee perceives that he/she has been accepted a part of the work group and has been regarded by his/her supervisors and peers as a significant person who is respected for his/her own personal worth and integrity as a human being, in contract to a feeling of being isolated or rejected in the work environment— a feeling manifested by hostility toward the department or company and its members— fellow employees, supervisors, department managers and other significant others at work; alienation embodies considerable resentment if not even defiance of regulations and rules of the company. Two chosen items are as follows : 1. I feel as if I am really a part of my work group. 2. Supervisors are reasonable and fair in the way that they rate their employees. Job Stress fFactor Scale 6). This sixth dimension is intended to represent the extent to which an employee can cope with changing, inconsistent, and/or ambiguous work assignments, increasing work requirements, changing and/or open- ended work environments, deadlines, and unexpected events all of which contribute to a potential conflict situation. Two illustrative items are as follows: 1. There seems to be a sense of urgency about everything. 2. I have more projects than I can effectively do. (Crowder & Michael, 1989a, pp. 22-23). Research Findings on Validation of the DOSC Several studies have been conducted to establish the validity of the different forms of the DOSC (Forms E, S, H, and W). This section contains an overview of the research findings to provide evidence of the construct, concurrent, and predictive validity of the DOSC forms. Construct Validity Michael and Smith (1976) developed three forms of an instrument to measure five school-related constructs of self-concept for students. The purpose of this effort was to provide data on the dimensions of self-concept which could be used by teachers for planning activities to enhance the self-concepts of their students. The three forms consisting of essentially parallel items, were developed as self-report measures. The three initial forms were designed for elementary school (ES), 10 grades four through six; junior high school (JHS), grades seven through nine; and senior high school (SHS), grades ten through twelve. Each form consists of five scales of hypothesized school-related constructs of self-concept : (a) Level of Aspiration, (b) Anxiety, (c) Academic Interest and Satisfaction, (d) Leadership and Initiative, and (e) Identification versus Alienation. Based upon the preliminary findings of item analysis outcomes, several items were revised (Fernandes, Michael, & Smith, 1978). In addition, forms JHS and SHS were combined into a single form (S) for the secondary level (Michael, Smith & Michael, 1978). The following conclusions were derived from the data analyses of these studies: 1. The internal and factorial validity of the five hypothesized dimensions of self-concept for the three forms of the DOSC received substantial support. 2. The Anxiety factor dimension was invariant across all samples. This finding led to the conclusion that anxiety was a pervasive component of academic self- concept at all grade levels studied. Michael, Kim, and Michael (1984) performed a subsequent construct validation study of the DOSC with Form S (secondary school level) to investigate whether the five hypothesized constructs (Level of Aspiration, Anxiety, Academic Interest and Satisfaction, Leadership 11 and Initiative, and Identification versus Alienation) translated into five DOSC subscales, would demonstrate the same factor structure for a college sample as it had for students in grades 7 through 12 (Michael, Smith, & Michael, 1984). The DOSC-Form S, consisting of five 14- item factor subscales that were fractionated into 20 subtests of either three or four items with four subtests per factor subscale was administered to a sample of 2 34 eighth-grade students and a sample of 193 community college students. Results of the study revealed, for the matched factors of the subscales, coefficients of congruence ranging from .72 to .96, with three of the five coefficients being in excess of .90. The comparability of the factor structure for the two samples indicated that the DOSC-Form S could be used to measure academic self-concept for both adolescents and college- age adults. To complement the elementary and secondary forms (Form E and S, respectively) of the DOSC (Michael & Smith, 1976; Michael, Smith, & Michael, 1984) and to remedy the lack of a test of academic self-concept for students in higher education, Michael, Denny, Knapp-Lee, and Michael (1984) designed a preliminary form of the DOSC, Form H, to measure the five factors of academic self-concept for college students. In developing Form H, 12 the researchers applied the same rationale or theory as was previously employed by Michael and Smith (1976). An initial scale of 125 items (25 items for each scale) was constructed and field-tested with college students. The results of the data analyses led to a revised preliminary form of 100 items (20 items per subscale), which was administered to another sample of university students. On the basis of the psychometric results, the following conclusions were formulated: 1. Both the initial and preliminary research forms exhibited satisfactory levels of reliability. 2. The preliminary research form revealed encouraging construct validity as depicted by the fact that the five constructs, within the parameters of the theoretical framework, when translated into the five subscales received substantial empirical support in terms of the factor structure demonstrated (Michael, Denny, Knapp-Lee, & Michael, 1984). In the continuing development of DOSC-Form H, Michael, Denny, Ireland-Galman, and Michael (1986-1987) reviewed the new information obtained on the preliminary research form (Michael, Denny, Knapp-Lee, & Michael, 1984), further examined the surviving items, and made minor editorial revisions which led to a third form administered to a sample of 181 community college 13 students. Results for the investigation using this new 80-item version of the DOSC-Form H, revealed an empirical factor structure consistent with the five hypothesized factor subscales. In addition, as evidenced by the item- analysis data, a relatively high degree of homogeneity of the subscales was also obtained. These results led to the conclusion that positive support for the construct validity of the DOSC-Form H, existed. The findings of this study led to the design of a published college-level version identified as Form H (Michael, Michael, & Denny, 1985). In a follow-up study of the Form H, Caracosta and Michael (1986) conducted a factor analysis of the intercorrelations of 20 subtests (four subtests per factor subscale) and six subtests of another instrument. The statistical outcomes provided empirical support for each of the five hypothesized dimensions (constructs) represented by the five factor subscales of the DOSC and thus support for the theory of academic self-concept associated with these constructs. Predictive Validity Lehn, Vladovic, and Michael (1980) studied the intercorrelations among a standardized reading test, the five subscales of the DOSC-Form S, and course grades in 14 an effort to predict academic success for 548 eleventh- grade students. They reached three main conclusions: 1. With the exception of Asians, a standardized reading test is a more nearly accurate (valid) predictor of academic success than is an academic self-concept measure. 2. Regardless of ethnicity, two subscales of DOSC- Form S reflecting constructs of Level of Aspiration and Academic Interest and Satisfaction afforded evidence of being valid predictors of school achievement. The Level of Aspiration subscale exhibited the greatest validity for all groups with coefficients ranging between .30 and .55. 3. A standardized reading test and an academic self-concept measure in combination would be a more nearly accurate predictor of academic success than would either measure by itself. Omizo, Hammett, Loffredo, and Michael (1981) explored the predictive validity of the DOSC-Form S relative to achievement and sex differences for a sample of 296 Mexican-American seventh grade students. They measured Total Language, Total Mathematics, and composite scores on the Iowa Tests of Basic Skills (Hieronymus & Lindquist, 1971, 1974). The results indicated that the Level of Aspiration subscale demonstrated the greatest 15 validity with coefficients of .58 (p < .01), .21 (p < .05), and .62 (p < .01), respectively, with the three previously mentioned criterion measures. The Anxiety subscale yielded coefficients of -.23 (p < .05), .39 (p < .01), and -.47 (p < .01), respectively. Additional findings revealed that males as compared to females had higher Level of Aspiration subscale scores and lower Anxiety subscale scores. In a series of investigations, Darakjian and Michael (1982, 1983) and Darakjian, Michael, and Knapp-Lee (1984, 1985) conducted a five-year longitudinal study involving an initial sample of 240 Anglo junior high school students from middle income families. They examined the predictive validity of the subscales of the DOSC-Form S and the subtests from the SRA Achievement Series (Naslund, Thorpe, & Lefever, 1972) relative to secondary school grade point average. Findings indicated that the subscales of the DOSC-Form S relative to the criteria of the secondary school grades revealed declining predictive validity coefficients. However, the Level of Aspiration subscale resulted in significant predictive validities of student success for each of four time segments. The median validity coefficients of the Level of Aspiration subscale were .49, .46, .42, and .37. 16 All values were significant beyond the .01 level (Darakjian, Michael, and Knapp-Lee, 1985). Denny (1984) drew the following conclusions from a review of several validation studies: 1. At least three factor subscales corresponding to three hypothesized dimensions of the DOSC (Level of Aspiration, Leadership and Initiative, and Identification vs. Alienation) have received substantial empirical support regarding their criterion-reflected validity. 2. The Level of Aspiration subscale has demonstrated the greatest criterion-related validity. Gold and Michael (1985) studied the correlation between academic self-concept and potential burnout in a sample of 109 graduate students beginning their first semester of practice teaching at the elementary school level. The emphasis of the investigation was to provide information regarding the concurrent validity of the five factor subscales of the DOSC-Form S and three criterion scales of the Maslach Burnout Inventory (MBI) (Maslach & Jackson, 1981). Findings exhibited convincing evidence of concurrent validity for the subscales of the DOSC-Form S in relation to three dimensions of burnout reflected by the MBI criterion subscales. In addition, substantial support was provided that those student teachers who demonstrated higher scores on the DOSC indicating a 17 positive self-concept tended to reflect scores on the MBI associated with minimal tendency toward burnout behaviors. The investigation of the correlations of each of five factor subscales of the DOSC-Form H, with the Total Score of the Academic Self-Concept Scale (ASCS) (Reynolds, 1981; Reynolds, Ramirez, Magrina, & Allen, 1980) was conducted by Halote and Michael (1984) to furnish information regarding two distinct approaches in assessing academic self-concept. For a college sample of 202 primarily Hispanic students, the following conclusions were ascertained from the analyses of the data: 1. The DOSC-Form H is a multidimensional academic self-concept measure with at least three hypothesized factors that could be verified. 2. The ASCS is appropriately depicted as a unidimensional measure of academic self-concept rather than a multidimensional one. 3. For the sample studied the DOSC-Form H showed higher concurrent validity coefficients with self-report college achievement indicators than did the ASCS Total Score. 18 4. The Level of Aspiration subscale of the DOSC demonstrated convincing evidence as a potentially valid predictor of college achievement. 5. The modest correlations between the ASCS and the DOSC-Form S subscale as well as differences in empirical factor structure between the DOSC factor scales and the ASCS Total Score indicated that the two measures are not comparable in what they are measuring (Halote & Michael, 1984). An examination of the construct and concurrent validity of the DOSC-Form H and the Intellectual Achievement Responsibility Questionnaire (lAR) (Crandall, Katkovsky, & Crandall, 1965) was investigated by Caracosta and Michael (1986) for a sample of 239 undergraduate university students. The focus of the study was a factor analysis of the intercorrelations of 20 4-item subtests of the five scales (4 subtests per scale) of the DOSC-Form H, and the six subtests from the lAR, a locus of control measure. Statistical evidence supported the following conclusions: 1. Empirical results revealed five distinct factor dimensions consistent with a theory of academic self- concept proposed by Michael and Smith (1976). 19 2. The lAR is a unidimensional measure of locus of control construct which is independent of the five hypothesized constructs of the DOSC. As has been demonstrated by the foregoing research, many investigations have been conducted which view self- concept as an important variable in learning and academic achievement. However, the majority of the studies has focused on children or adolescents in an academic setting. According to Crowder and Michael (1989a), "Virtually no research has been done with reference to the self-concept as an important variable for fully mature and working adults in their employment setting" (p. 20). They endeavored to delineate a multidimensional theory of self-concept in an employment setting and to develop an experimental form of a self-concept measure which could be used in industrial and business settings. The expansion of the self-concept theory to a work setting is based upon the assumption that "... constructs central to self-concept in an academic setting . . . are not essentially different from those associated with the self-concept in a demanding work setting ..." (Crowder & Michael, 1991, p. 448). This experimental form entitled Dimensions of Self- Concept, Work Form (DOSC-Form W) was developed based upon the rationale used by Michael and Smith (1976) who 20 posited five constructs central to academic self-concept and translated these into five factor scales of the DOSC. Therefore, the DOSC-Form W (Crowder & Michael, 1989a) has evolved from the three earlier forms of standardized academic self-concept measures developed for measuring non-cognitive factors associated with self-esteem or self-concept in the school setting. In DOSC-Form W, a sixth dimension was added to the five dimensions already measured in DOSC-Forms E, S, and H— namely that of Job Stress. The experimental version of the DOSC-Form W consisted of 30 self-report items for each of the six factor subscales (Crowder & Michael, 1989a). This 180- item instrument was administered to a sample of 201 employees in a major electronics/computer company. Item analyses were conducted in which each item of the DOSC- Form W was correlated with a total score of the factor scale for which the item was intended to belong as well as with the total score of each of the remaining five factor scales of which the item was hypothesized not to be a member. Results indicated internal-consistency estimates of reliability varied between .84 and .91 for the six subscales. An exploratory factor analysis revealed that four of the hypothesized constructs received substantial empirical support. Although the 21 construct validity of the subscales of the experimental form showed promise, the correlation of .76 obtained between the Level of Aspiration and Job Interest and Satisfaction factor subscales led to the modification of some items. Additionally, other items were moved from one factor subscale to another. A revised from of the DOSC-Form W was developed and administered to a new sample of 200 employees working in an electronics/computer company (Crowder & Michael, 1989b). This revised form consisted of six 20-item factor subscales. Within each factor subscale four subtests each consisting of five items were formed. Both exploratory and confirmatory factor analyses of a correlation matrix of 24 subtests each of five items derived from the six-factor subscales were completed. Internal-consistency estimates of reliability (alpha coefficients) varied between .88 and .92 for the six subscales. A major conclusion of the study was that the validity of the six constructs associated with the six subscales of the DOSC-Form W received substantial empirical support. Based upon the encouraging evidence for construct validity and the convincing level of reliability of each of the six factor subscales, Crowder and Michael (1991) proceeded to devise a shortened form of the DOSC-Form W 22 that would still provide an effective degree of discriminant validity and sufficient subscale reliability. This second revision consisted of 90-items (15 items for each of the six scales). Internal-consistency coefficients (alpha values) ranged between .86 and .92 for the six subscales. The intercorrelations among the six scales varied between -.40 and .53, with a median coefficient of .07. These relatively low intercorrelations among the six subscales lent support to the discriminant validity of the instrument. For purposes of factor analyses three subtests each consisting of five items from each of the six 15-item factor subscales were formed. The resulting 18 subtests formed from the six factor subscales were intercorrelated and factor analyzed. Construct validity for each factor subscale was evidenced in the results of the orthogonal and oblique exploratory factor analyses that afforded an exceptionally high degree of simple structure consistent with what one would hypothesize for highly homogeneous subscales reflecting substantial discriminant validity. In addition, a confirmatory maximum likelihood factor analysis provided strong support for the presence of six distinct dimensions as evidenced by both orthogonal 23 and oblique solutions that were reinforced by goodness-of- fit results. On the basis of statistical outcomes, the following conclusions were reached: 1. The 90-item form of the DOSC-Form W demonstrated quite satisfactory levels of reliability for all six subscales. 2. The relatively low intercorrelations among the six subscales lent support to the discriminant validity of the instrument. 3. Exploratory factor analyses exhibited clear-cut six-factor solutions with a high degree of simple structure that demonstrated support for the validity of the six constructs. Based upon the empirical evidence presented, it would appear that the 90-item DOSC-Form W has shown promise as a reliable and construct-valid instrument for relatively high-level personnel in the work place. Research Questions Consistent with the stated purposes of this investigation the following five research questions were posed: 1. To what degree did each of the six DOSC-Form W subscales portraying different dimensions of self-concept in an occupational environment reveal evidence of being internally consistent? 24 2. What were the intercorrelations among the six factor subscales of the DOSC-Form W? 3. To what extent did each of the constructs associated with each of the six factor subscales of the DOSC-Form W receive empirical support of its validity on the basis of exploratory factor analyses? 4. In terms of evidence provided by indexes of goodness-of-fit associated with confirmatory maximum likelihood factor analyses, which hypothesized first- order and higher-order factor models served to explain the greatest proportions of covariance among subtests derived from the six factor subscales of the DOSC-Form W? 5. What was the degree of relationship, if any, between each of several selected demographic variables (e.g. level of formal education, number of years of military duty, age, gender, and military rank) and standing on each of the six DOSC-Form W factor scales? Conceptual Assumptions The following conceptual assumptions were implicit in this study: 1. The theory of academic self-concept, as delineated by Michael and Smith (1976), provided a conceptual framework from which to interpret the multidimensional characteristics of the DOSC-Form W. 25 2. The same rationale employed by Michael and Smith (1976) regarding constructs central to academic self- concept could be translated and extended for application to adults in a work setting (Crowder & Michael, 1991). 3. The findings of the current study could be applied to future investigations of the multidimensional construct of self-concept. Delimitations The following delimitations were present in the investigation : 1. Only active-duty military personnel took part in the study. 2. Subjects were a minimum of 21 years of age. Organization of the Remainder of the Study Chapter II presents the methods and procedures employed in the investigation. Topics covered are: (a) the characteristics of the sample; (b) the instrument and associated test variables; (c) demographic variables; (d) the procedure for administration of the DOSC-Form W; (e) methods employed in data analyses; (f) methodological assumptions; and (g) limitations. In Chapter III the findings of the study are analyzed within the context of the five research 26 questions posed in this chapter. A brief discussion of the results for each question is then provided. Chapter IV includes a summary of the findings, conclusions, and recommendations for additional research. 27 CHAPTER II METHODS AND PROCEDURES This chapter provides information concerning (a) the characteristics of the sample, (b) the instrument and associated test variables, (c) the procedure for administration of the DOSC-Form W, (d) demographic variables, (e) methods employed in the data analyses, (f) methodological assumptions, and (g) limitations. Sample The subjects were drawn from a population of approximately 4,100 military personnel working at a military installation on the island of Oahu in the State of Hawaii. The sample was non-random based upon the situational factors which provided an opportunity for individuals voluntarily to participate in the study. Subjects were volunteers for a Career Change Strategies Seminar. Eligibility to attend the seminar was limited by only two factors; the subject must (a) have been on active duty and (b) have been employed with the United States Air Force in Oahu, Hawaii. Attendance was not restricted by ethnic group, religious orientation, | gender, age, educational level, or employee status levels (a continuum from the most junior nonprofessional employees to most senior professional employees). Anyone from the population was eligible to volunteer for the seminar. 28 I The entire sample consisted of 212 Air Force active- duty military personnel (179 males and 33 females) on the island of Oahu in the State of Hawaii. Prior to responding to the 90-question form of the DOSC-Form W self-report questionnaire, each subject was asked to complete what was called the Information Sheet which provided demographic information. Participants ' varied in age from 21 to 56 with 46, 87, 71, and 7 subjects respectively falling within the 21-29, 30-39, , 40-49, and 50-56 age intervals (one subject did not I indicate age). Military rank ranged from Airman First : Class to Colonel with 13 3 and 79 subjects, respectively, I falling within the enlisted (nonprofessional) and officer ; (professional) categories. The number of years of military duty fell between 2 and 32 years with 22, 55, 29, 55, 42, and 9 subjects respectively having served 2- 6, 7-11, 12-16, 17-21, 22-26, and 27-32 years, respectively. With respect to level of formal education, 75, 35, 51, and 49 subjects, respectively, indicated having completed high school, having earned an associate degree, having achieved a bachelor's degree, and having participated in postgraduate study (completion of a master's degree or doctorate). Two respondents did not supply information concerning level of formal education attained. 29 The Test Instrument The current form of the Dimensions of Self-Concept, Form W (DOSC-Form W), as discussed earlier, evolved from two earlier forms developed by Crowder and Michael (1989a, 1989b). This revised form is a 90-item self- report measure of self-concept for adults in an employment setting (Crowder & Michael, 1991). As previously indicated, the DOSC-Form W consists of 90- items which are divided into 15 items for each of the following six construct scales : Level of Aspiration; Anxiety; Job Interest and Satisfaction; Leadership and Initiative; Identification versus Alienation ; and Job Stress. Additionally, three subtests consisting of five items were taken from each of the six factor subscales. Thus, 18 subtests were formed the scores on which were used for subsequent data analysis. Crowder and Michael (1989a) adapted the first five factor subscales based upon the delineations of the factors from the DOSC Technical Manual (Michael, Smith, & Michael, 1989). The sixth subscale. Job Stress, was added as a consequence of further research by Crowder and Michael (1989a). The instrument is designed in the form of a Likert- type scale with differential weights assigned to the responses. The individual is asked to respond to each question by selecting one of the five response 30 alternatives: Never. Seldom. About Half the Time. Very Often. and Always. The respective scoring weights (1, 2, 3, 4, or 5) assigned to the responses indicate level of agreement. The higher the level of agreement with the responses, the greater the possession of the hypothesized construct factor. Sample Items For each of the six factor subscales, a sample item is provided as follows : Factor Subscale Level of Aspiration Level of Anxiety Job Interest and Satisfaction Leadership and Initiative Identification versus Alienation Level of Job Stress Sample Item I strive to be rated in the highest category on my performance evaluation. I become nervous when I am called on to participate in a staff meeting. The projects that I work on are fun and interesting. I can convince my peers to go along with my ideas. Supervisors are attentive to their employee's needs. I have more projects assigned to me than I can effectively do. Administration of the DOSC-Form W The DOSC-Form W was designed to be a self- administering, self-report instrument requiring approximately 30 to 50 minutes to complete. Subjects 31 volunteered to complete the Information Sheet, which supplied demographic information, as well as the DOSC- Form W responses. Participants were instructed that there were no right or wrong answers and that their responses would be kept confidential. They were informed to choose the response that most nearly accurately reflected their personal attitude, feeling, or opinion. Selections were indicated by filling in the appropriate response space on the questionnaire form. Scoring The assessments were scored by the author, and results of the DOSC-Form W were anonymous. The items are arranged so that every seventh item is related to the same factor as follows: Factor Item Numbers Level of Aspiration 1, 7, 13, 19, 25 31, 37, 43, 49, 55 61, 67, 73, 79, 85 Level of Anxiety 2, 8, 14, 20, 26 32, 38, 44, 50, 56 62, 68, 74, 80, 86 Job Interest and 3, 9, 15, 21, 27 Satisfaction 33, 39, 45, 51, 57 63, 69, 75, 81, 87 Leadership and 4, 10, 16, 22, 28 Initiative 34, 40, 46, 52, 58 64, 70, 76, 82, 88 Identification 5, 11, 17, 23, 29 versus Alienation 35, 41, 47, 53, 59 65, 71, 77, 83, 89 32 Level of Job Stress 6, 12, 18, 24, 30 36, 42, 48, 54, 60 66, 72, 78, 84, 90 As mentioned previously each item on the DOSC-Form W provides for five responses: Never, Seldom. About Half the Time, Verv Often, and Always. with differential weights of 1, 2, 3, 4, and 5 assigned, respectively. The potential range of scores on each of the six dimensions can vary from a minimum of 15 to a maximum of 75. The higher the respondent's score on any scale the greater the degree of possession of the hypothesized construct. Demographic Variables Altogether, five demographic variables on the Information Sheet were considered in the analysis of the data : Level of Formal Education (with scoring points indicated) High School (1 to 4 years) = 1 College (1 to 4 semesters) = 2 College (3 to 4 years) = 3 Postgraduate Study = 4 (including a master's degree or doctorate) Number of Years of Military Duty (the number entered as a datum) Age (number of years entered as a datum) Gender (F = 1, M = 2) 33 Military Rank (commissioned officers coded as 2; enlisted personnel, as 1) Data Analyses Data from the Information Sheet and from the DOSC- Form W were entered into the computer and analyzed using ; the SPSS/PC Plus; Statistical Package for the Social Sciences (Norusis, 1990), the SAS/STAT Guide for Personal Computers (SAS Institute, 1990), and the LISREL VII (Joreskog and Sorbom, 1989) software programs. In addition to the calculation of the means and standard I I deviations for scores on each of the measures, I correlation coefficients were computed between scores on each of the 15 possible pairs of the six DOSC-Form W subscales, between scores on each of the 153 possible pairings of five-item subtests, and between scores on each of the six factor subscales and scores assigned to each of the five demographic variables. To examine the internal validity of each subscale, item analyses were preformed by correlating the scores on each item of the DOSC-Form W with (a) a total score of the subscale to which that item was hypothesized to belong and (b) a total score of each of the remaining five subscales for which that item was hypothesized not to be a member. To estimate the internal-consistency 34 reliability of each subscale coefficient alpha was employed. Within each of the six 15-item subscales, three subtests (as indicated earlier) each comprising five items were devised such that the first five items in the subtest constituted the first measure; the next five items in the subtest constituted the second measure ; and i the last five items, the third measure. Exploratory Factor Analyses I For a correlation matrix of the scores on the 18 ' resulting measures, both an orthogonal (varimax) and an ^ oblique (promax) exploratory factor analysis were conducted in an effort to obtain evidence of construct validity. In each analysis, six principal factor axes ; i (consistent with the hypothesized structure) were rotated | using the procedure described in the SAS/STAT Guide for * Personal Computers (SAS Institute, 1990). I Confirmatory Factor Analyses | Use of a confirmatory maximum likelihood factor analysis within the framework of the LISREL VII general | I computer program (Joreskog & Sorbom, 1989) provided the ' means for evaluating how effectively in relation to a | null factor model of no factor structure each of three substantive models served to explain the covariance among the 18 subtests. The three factor models were as 35 follows: (a) a general or one-factor model for which it was assumed that the DOSC-Form W was a uni-dimensional instrument, (b) an oblique two-factor model in which the Level of Anxiety and Level of Job Stress subtests were considered to reflect a negative affect and the remaining 12 subtests associated with the four factor subscales were assumed to portray a positive affect, and (c) an oblique six-factor model corresponding to the hypothesized multidimensional factor structure of the DOSC-Form W that had been designed to provide an operational definition of six constructs. In addition to I the generation of factor solutions, indexes of goodness- | ! of-fit of each of the three substantive models as well as | for the null model were obtained. An index reveals the j I extent to which a given factor model is capable of i I explaining the covariance within the matrix of correlated | variables. The function of these indexes of goodness-of- ] I fit are explained within the context of the findings j pertinent to the fourth research question. ! Methodological Assumptions ! I The following methodological assumptions were | present in this investigation: 1. Exploratory and confirmatory factor analyses | would provide information necessary to yield confirming | evidence of the multidimensionality of the DOSC-Form W. i i 36 ' 2. The data were accurately recorded, scored, and I analyzed. 3. The subjects in the sample were sufficiently j representative of adults in the military work setting so I as to permit at least a limited basis for generalization I of the findings. I i 4. The subjects completing the questionnaire understood the instructions, and provided answers that I I accurately reflected their perceptions and opinions in response to each item. Limitations I The extent to which the findings of this study could I I be generalized could have been influenced by one or more of the following limitations: 1. The use of Career Seminar participants in place I i of random selection of military personnel might have ! affected the external validity. It should be noted that I some subjects would have been more motivated to attend the seminar than others because of the nature of the increased reduction of personnel as a result of the current military down size. Those individuals with a higher probability of attending the seminar would be those near to retirement (20 years with the service) or those facing a reduction in force (layoff) situation. However, no restriction was made on the number of years 37 r of employment (even those who were voluntarily leaving the Air Force before retirement could attend the seminar). 2. Responses to some of the items may have been influenced by subjects who are being terminated or forced to retire and/or who may have been angry with the company and supervisors. 3. Because of the lower statistical representation of women in the Air Force, the number of male subjects exceeded the number of female subjects by a greater extent than would exist in a normal population of civilian workers. 38 CHAPTER III ANALYSIS AND INTERPRETATION OF RESULTS i I In this chapter the statistical outcomes are I reported within the framework of each of the five research questions posed in Chapter I. Then, the results of this inquiry are interpreted with reference to those of earlier investigations in which the DOSC-W has been employed. Consideration is also given concerning the extent to which there is some congruence of the findings of this study with those arising from investigations employing other forms of the DOSC. Analysis of Results In the five subsections to follow, the major statistical findings are summarized. Supporting evidence is provided in tables. Internal Consistencv of the DOSC-Form W Subscales CResearch Ouestion 1^ Two forms of evidence were obtained regarding the i I internal consistency of each of the six factor subscales ; of the DOSC-Form W. As mentioned earlier each item was j correlated with a total score on the subscale of which it was a member as well as with the total score of each of the five remaining subscales. A second indication of 39 ! internal consistency was achieved through calculation of ! j coefficient alpha. I In the instance of every one of the 90 items, the correlation of a given item with the total score of the subscale of which it was a member was higher than the corresponding correlation with the total score on any one of the remaining five subscales. In other words, 90 hits out of 90 outcomes were realized. Typically, the correlations of items with the total scores of their own subscales ranged from approximately .40 to .80. All coefficients were corrected for any spurious overlap. In the instance of the coefficients of internal consistency the range of values was from .87 to .92, as can be seen in the entries in the principal diagonal of Table 1. Specifically, for the Level of Aspiration, Level of Anxiety, Job Interest and Satisfaction, Leadership and Initiative, Identification versus Alienation, and Level of Stress factor subscales, the respective alpha coefficients were .87, .88, .92, .92, .92, and .90. All subscales consisted of 15 items. Intercorrelations Among the Six Factor Subscales rResearch Ouestion 2) As can be seen in the entries in Table 1, the intercorrelations ranged from -.34 to .60. The two highest intercorrelations of .60 and .56 occurred between 40 TABLE 1 I I I Means, Standard Deviations, and Intercorrelations of Six Factor Subscales of the DOSC-Form W Including Internal-Consistency Estimates of Reliability Along the Principal Diagonal (Decimal Points Omitted) (N = 212) 1 DOSC Subscales (1) (2) (3) (4) (5) (6) M SD 1. Level of Aspiration (87) 07 60 40 39 08 58.7 8 . 6 I ^ Level of Anxiety 07 (88) -13 -34 -17 42 33 . 3 9.0 ! 3 . Job Inter est and Satisfaction 60 -13 (92) 50 56 00 56.1 9.4 4. Leadership Initiative 40 -34 50 (92) 52 03 50.5 9.9 1 5. 1 Identifica tion vs. Alienation 39 -17 56 52 (92) -22 54.2 10.2 6. Level of Job Stress 08 42 00 03 -22 (90) 37.5 9.0 41 the Level of Aspiration and Job Interest and Satisfaction | subscales and between Job Interest and Satisfaction and | Identification versus Alienation subscales, respectively. ; The two lowest intercorrelations from the standpoint of j algebraic values were -.3 4 between the Level of Anxiety | and Leadership and Initiative subscales and -.2 2 between \ the Identification versus Alienation and Level of Job ' Stress subscales. From the standpoint of algebraic values the median | coefficient was .08. In terms of absolute values the | I I median coefficient was .28. It is apparent from the coefficients in Table 1 that the two subscales reflecting negative affect (Level of Anxiety and Level of Job I Stress) revealed negative or approximately zero I correlations with the remaining four subscales that reflect positive aspects of self-concept. Among these I four subscales portraying a positive affect ; intercorrelations varied between .39 and .60. The Level of Anxiety and Level of Job Stress measures registered a coefficient of .42. Empirical Support for the Validity of DOSC-Form W Constructs Based on Exploratorv Factor Analyses (Research Ouestion 3^ In Tables 2 and 3 the results of exploratory orthogonal (varimax) and oblique (promax) solutions are 42 TABLE 2 Rotated Orthogonal Factor Matrix of 18 Subtests: Varimax Solution (Decimal Points Omitted) Hypothesized Factor Subscales Subtest Number Factor I II III IV V VI h^ I. Level of 3 86 15 13 08 13 03 84 Aspiration 2 84 05 33 13 07 -02 84 1 71 -05 33 25 21 11 73 II. Leve1 of 2 03 85 -04 -20 02 22 82 Anxiety 3 12 83 -07 07 -23 28 85 1 04 76 -02 -38 09 11 75 III. Job Inter 2 27 -01 86 19 22 02 89 est and 1 21 00 85 21 23 -03 87 Satisfaction 3 33 -16 78 14 26 04 83 IV. Leadership 2 12 -28 11 86 13 10 87 and 3 14 -10 30 81 23 03 82 Initiative 1 23 -09 18 81 35 01 86 V. Identifica 1 12 06 19 13 87 — 16 85 tion vs. 3 23 -12 28 20 81 -14 85 Alienation 2 08 -08 25 36 80 — 05 85 VI. Level of 1 12 05 08 -02 -01 92 88 Job Stress 2 -04 22 —04 04 -14 89 86 3 -01 34 00 11 -18 82 83 43 TABLE 3 Rotated Oblique Factor Matrix of 18 Subtests: Promax Solution (Decimal Points Omitted) Hypothesized Factor Subscales Subtest Number Factor I II III IV V VI h^ I. Level of 3 90 17 39 24 28 08 84 Aspiration 2 91 05 56 31 28 03 84 1 82 — 06 59 45 40 13 73 II. Level of 2 04 90 -11 -28 -10 35 82 Anxiety 3 12 88 -12 —05 -30 44 85 1 03 81 -10 -43 -03 21 75 III. Job Inter 2 47 -09 93 41 48 -02 87 est and 1 53 —08 94 41 48 04 89 Satisfaction 3 56 -21 90 39 50 03 83 IV. Leadership 2 27 — 36 34 92 31 09 87 and 3 35 -22 52 89 44 04 82 Initiative 1 42 -21 46 89 53 01 86 V. Identifica 1 30 —05 43 29 91 -21 85 tion vs. 3 29 -20 52 53 89 -11 85 Alienation 2 41 -23 55 40 91 -21 85 VI. Level of 1 17 26 10 08 -11 92 88 Job Stress 2 -02 40 -08 05 -28 92 86 3 02 50 — 04 10 -28 88 83 44 given. For the varimax solution it is readily apparent that a high degree of simple structure was realized and that for every set of three subtests derived from a factor subscale intended to represent a given construct the loadings were quite high— never any lower than .71 with most values falling between .75 and .89. Loadings I for any of one of the identified factors were quite low on subtests intended to represent other factor dimensions. To a less marked degree than that observed in the I varimax solution, comparable results were obtained in the I promax solution. The simple structure was clearly ' present, but somewhat less pronounced than that found in the orthogonal factor matrix. Nevertheless, the results were quite clear in Table 3 that substantial I support for the validity of each of the hypothesized I I constructs was obtained just as such evidence had I occurred in conjunction with Table 2. It is interesting to note in Table 4 the interfactor correlations among the six identified oblique factors. These correlations did not differ greatly from those arising from intercorrelations of factor scores. The configurations of intercorrelations in Tables 1 and 4 are quite similar. Such an outcome would be expected when factor subscales are quite homogeneous and highly reliable. 45 TABLE 4 Interfactor Correlations in the Oblique (Promax) Solution (Decimal Points Omitted) Factors Identification I II III IV V VI I Level of Aspiration ----- 06 53 33 34 08 II Level of Anxiety 06 — -13 -24 -17 37 III Job Interest and Satisfaction 53 -13 42 51 02 IV Leadership and Initiative 33 -24 42 — 36 10 V Identification vs. Alienation 34 -17 51 36 -22 VI Level of Job Stress 08 37 02 10 -22 — — 46 Outcomes of Confirmatory Factor Analyses Concerning Proportions of Covariance Explained by Alternative Factor Models fResearch Question 4) In the confirmatory maximum likelihood factor analyses the solutions provided factor loadings on the hypothesized dimension or dimensions corresponding to a given factor model that in virtually every instance were statistically significant beyond the .001 level. In the general factor model, the three subtests associated with the Level of Anxiety factor subscale did not have loadings of sufficient magnitude to be statistically significant. In Table 5, goodness-of-fit indexes are furnished for each of the three substantive factor models as well as for the null model of no factor structure. It should be noted in the interpretation of these indexes that typically the closer the degree of fit of the model (in terms of accounting for larger amounts of covariance among the subtests variables or equivalently in terms of reproducing accurately the correlation matrix) the smaller the value for chi-square, chi-square/df, and root-mean square (RMSR) and the larger the values for the goodness-of-fit (GFI), adjusted goodness-of-fit (AGFI), and DELTA. Additional information regarding the first 47 I za I O C X u - i t 3 O c O Z C u Q ) [ o t f l 4 J ^ • 01 >J O (U (Q Z H > O v O 00 00 00 00 .o e g u 0 3 ( 0 3 1 . 1 o O V X J3 Cb 4 - 1 0 1 T3 4 - 1 o U l u u O 0 1 ' —J 0 3 O o s r® 0 1 Cb 4 - 1 X < 4 - 1 < M O S - I C o O X , 0 3 « J o < 4J Cb k, V c/3 Ü U - l < u > 0 3 X c a o > Cb 0 1 e g b. ( U 0 3 3 t/3 u 0 1 ) - < cr < u < 0 V 0 3 0 1 f - 4 1 —4 C c > x) )-l -C 3 0 } o V c < Z u O -J 0 3 o o ill! — i J ; î 1 | "3 :3 « > < 0 ^ i w V O ^ ë 1 j W C T Ï i ! i i 4 , - ^ w a Ü 0 ) 0 3 J 73 c o -g T - * o c 0 3 44 1 - 1 0 3 00 C J 0 3 ( 4 U u <n o —4 3 « 4 4 o > 0 3 00 p - 4 c / 3 <" p - 4 ( U o 0 3 75 4 3 < 4 - i c X < a X 0 3 C C / 3 0 1 C 44 4 3 0 1 3 u X 0 ) 4 3 4 U o 1 - 4 -g l u C / 3 -3 0 ) c 0 3 \o 0 3 4 4 X <X C O O 0 3 a 1 - 3 U c 3 —4 m o X: X ( 4 U o 0 3 <n 0 3 o o > r —1 V 3 4 1 - 3 4 3 0 1 0 3 ( 0 X < 4 - i 73 X « c 0 3 4 4 > 0 3 g 0 ) 0 3 4 3 c 5 3 J 3 J 0 3 > cr < -g 0 3 rH < J \ 4 4 c O' l u XI <o 0 3 0 3 O o > u X u X c O c 0 3 c < o g < 4 U 0 3 o o T - 4 ( 4 U t“ 4 O 4 3 ■U o e g X > x: V C J 2 c Q c c o < 4 U 0 3 a > 00 0 3 > c " r 4 4 3 X X o -J 5 c < -J 2 0 3 c 0 3 c 0 3 ÜÎ 0 3 "O ( 0 X 0 1 0 1 0 1 75 > 3 C C < o 0 3 4t C 0 1 0 1 o\ in o 0 3 X) 4 - 1 o 0 3 0 3 > X C 4 3 < g X o " o i l u c 0 3 C J c / 3 3 4 4 3 o > 4 - 1 3 4 4 4 t B IX X 0 ) O c 4 3 u o U lu ( Q 0 3 X 4 3 3 4 4 U O c l u 0 3 X 43 i u 4J o 75 o O c X c U * r 4 o —1 c X 1 3 0 3 ( D 75 > x: X o (4U 13 c 41 H H O 0 3 3 J 48 five cited indexes may be found in Joreskog and Sorbom (1989a, 1989b) and for the DELTA index in Bentler and Bonnet (1980). The DELTA coefficient is obtained by taking the difference between the chi-square of the null model (representing no factor structure) and the chi- I square of the substantive model (the general factor, two- j factor, or six-factor oblique model) and then by dividing I this difference by the magnitude of chi-square for the I null model. I The results set forth in Table 5 clearly indicate that the six-factor oblique model corresponding to the six hypothsized constructs represented by the six I subscales of the DOSC-Form W provided the closest degree I I of fit of any one of the models and thus served to explain the largest proportion of covariance in the ’ matrix of 18 subtests. This information suggests that I ! substantial empirical support was obtained for the [ I hypothesized multidimensional structure of the DOSC-Form W that had been administered to a military population. I I It is apparent that neither the unidimensional model nor ' the two-factor oblique model came even close in approximating the effectiveness of the six-factors oblique model. 49 Relationship Between Each of Five Demographic Variables and Standing on DOSC-W Factor Subscales (Research Question 5) Table 6 sets forth the coefficients of correlation between scores on each of the six DOSC-W subscales and status with respect to each of five demographic variables. In general, coefficients were low as only two obtained a value of .30. The only factor subscale exhibiting a consistent positive relationship of statistical significance was that of Leadership and Initiative. For this measure coefficients of .30, .28, .30, and .23 were found with respect to the demographic variables of Level of Formal Education, Number of Years of Military Duty, Age, and Military Rank— all coefficients being significant beyond the .001 level. In the instance of gender, a coefficient of -.11 occurred. This coefficient was not statistically significant. Discussion The factor analytic results of this investigation were highly comparable to those reported by Crowder and Michael (1989a, 1989b, 1991) who administered three versions of the DOSC-Form W. In the current analysis as well as in the three reported by Crowder and Michael, the six-factor oblique structure afforded the most 50 TABLE 6 Correlation Coefficients of Scores on Each of Six DOSC-Form W Subscales With Each of Five Demographic Variables (210 < H < 212) (Decimal Points Omitted) Demooraphic Variables DOSC-Form W Subscales Level of Formal Educ. ® (7) No. Years of Military Duty (8) Age (9) Gender (F = 1, M = 2) (10) Military Rank ° (11) 1, Level of Aspiration -15* -06 -08 00 -10 2. Level of Anxiety -22** -13 -15* -01 -23*** 3. Job Inter est and Satisfaction -12 14* 12 -07 -12 4. Leadership and Initiative 30*** 28*** 30*** -11 23*** 5. Identifica tion vs. Alienation 11 12 12 -10 06 6. Level of Job Stress 09 12 16* -05 09 ^ Level of education coded as 1 = high school (1-4 years), 2 = college (1-4 semesters), 3 = college (3-4 years), and 4 = postgraduate study (including completion of a master's degree or doctorate). ^ Commissioned officers coded as 2 and enlisted personnel as 1. * < .05 * * D < .01 *** Ü < .001 51 psychologically meaning solution. In the report by Crowder and Michael (1991) both exploratory and confirmatory factor analyses reinforced the conclusion that the DOSC-Form W is indeed a multidimensional instrument in which the validity of each of the six hypothesized constructs has received substantial support. This conclusion also lends a certain measure of confidence regarding the adequacy of the theory of self- concept set forth by Michael and Smith (1976) in an academic setting as well as the application of this theory to adult workers in a demanding work setting. Moreover, the addition of the construct of Level of Job Stress that was added to the DOSC-Form W appears to complement quite nicely the original theory of academic self-concept proposed by Michael and Smith (1976). In fact, the construct of Level of Job Stress appeared in all analyses involving use of Form W to be relatively independent of the construct of Level of Anxiety. When the factor analytic results obtained for the DOSC-Form W are compared with those in which other forms of the DOSC have been administered to college and secondary school samples as described in the previous review of the literature there is a high degree of consistency. Again and again whether exploratory or confirmatory factor analytic methodology is employed the 52 same factors are replicated across studies. Thus, one may conclude at least somewhat tentatively that a relatively high degree of invariance exists across rather diverse samples in the manifestation of the constructs of the various forms of the DOSC. It is recommended at this point that the DOSC-Form W be tried in a number of employment settings to ascertain which particular subscales would show predictive validity of criterion measures in the workplace. It is time for criterion-related validation studies to occur, now that it appears that the construct validity has been reasonably well established. Accompanying such criterion-related undertakings should be the establishment of extensive norms for the subscales so that they can be effectively employed in both selection and counseling contexts of working environments. 53 CHAPTER IV SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS Summary Background In the past few years, Crowder and Michael (1989a, 1989b, 1991) have developed, revised, and validated a multidimensional self-concept measure, the Dimensions of Self-Concept-Form W (DOSC-Form W), for workers in an employment setting. This research has demonstrated the potential application of a self-concept measure for employed adults in high technology. A basic assumption made by Crowder and Michael (1991) was that "... the constructs central to self-concept in a academic setting involving a commitment to learning and achievement are not essentially different from those associated with the self-concept in a demanding work setting ..." (p. 448). The DOSC-Form W was developed and validated with samples of working adults in a large electronics/computer company (Crowder & Michael, 1989a, 1989b, 1991). A need exists for additional empirical research, conducted in diverse employment settings, which could provide further information on the construct validation of the DOSC-Form W. A variety of employment settings is available for examination. 54 : Purpose of the Study The primary purpose of this study was to sample a population of military personnel in an effort to estimate the internal-consistency reliability of each of the six factor subscales of the DOSC-Form W (Level of Aspiration, Level of Anxiety, Job Interest and Satisfaction, Leadership and Initiative, Identification vs. Alienation, and Level of Job Stress) and to obtain evidence regarding I i the construct validity of each of these factor subscales, j A secondary purpose was to ascertain what relationships, if any, existed between the standing on each of the six j factor subscales of self-concept and each of several j selected demographic variables (e.g. level of formal j education, number of years of military duty, age, gender, j and military rank). Importance of the Study Studies conducted by Crowder and Michael (1989a, 1989b, 1991) involving the development, revision, and validation of the DOSC-Form W have contributed to the knowledge of self-concept with employed adults. The importance of the proposed inquiry was to add additional data pertaining to the reliability and validity of a measure of self-concept in employed adults who comprised a military population. The information gained through this study was thought to have the potential to provide 55 assistance in counseling individuals in vocational transition with regard to optimal occupational placement in both the military and civilian work environments. Research Questions Consistent with the stated purposes of this investigation the following five research questions were I posed: I 1. To what degree did each of the six DOSC-Form W I subscales portraying different dimensions of self-concept in an occupational environment reveal evidence of being internally consistent? 2. What were the intercorrelations among the six ! factor subscales of the DOSC-Form W? I I 3. To what extent did each of the constructs j associated with each of the six factor subscales of the DOSC-Form W receive empirical support of its validity on the basis of exploratory factor analyses? I 4. In terms of evidence provided by indexes of 1 goodness-of-fit associated with confirmatory maximum I I likelihood factor analyses, which hypothesized first- order and higher-order factor models served to explain 1 the greatest proportions of covariance among subtests derived from the six factor subscales of the DOSC-Form W? 5. What was the degree of relationship, if any, between each of several selected demographic variables 56 (e.g. level of formal education, number of years of military duty, age, gender, and military rank) and standing on each of the six DOSC-Form W factor scales? Method Sample. The subjects were drawn from a population of approximately 4,100 military personnel working at a military installation on the island of Oahu in the State of Hawaii. The sample was non-random based upon the situational factors which provided an opportunity for individuals voluntarily to participate in the study. Subjects were volunteers for a Career Change Strategies Seminar. Eligibility to attend the seminar was limited by only two factors; the subject must (a) have been on active duty and (b) have been employed with the United States Air Force in Oahu, Hawaii. Attendance was not restricted by ethnic group, religious orientation, gender, age, educational level, or employee status levels | (a continuum from the most junior nonprofessional I I employees to most senior professional employees). Anyone from the population was eligible to volunteer for the seminar. The entire sample consisted of 212 Air Force active- : duty military personnel (179 males and 33 females) on the island of Oahu in the State of Hawaii. Prior to responding to the 90-question form of the DOSC-Form W ' I I self-report questionnaire, each subject was asked to I complete what was called the Information Sheet (as reproduced in Appendix A) which provided demographic I information. Participants varied in age from 21 to 56 with 46, 87, 71, and 7 subjects respectively falling within the 21-29, 30-39, 40-49, and 50-56 age intervals j I (one subject did not indicate age). Military rank ranged i from Airman First Class to Colonel with 13 3 and 79 subjects, respectively, falling within the enlisted | (nonprofessional) and officer (professional) categories. The number of years of military duty fell between 2 and ! I 32 years with 22, 55, 29, 55, 42, and 9 subjects ^ respectively having served 2-6, 7-11, 12-16, 17-21, 22- 26, and 27-3 2 years, respectively. With respect to level of formal education, 75, 35, 51, and 49 subjects, I respectively, indicated having completed high school, | I having earned an associate degree, having achieved a | bachelor's degree, and having participated in | postgraduate study (completion of a master's degree or doctorate). Two respondents did not supply information j concerning level of formal education attained. ! Test instrument. As previously indicated, the DOSC- Form W consists of 90-items which are divided into 15 items for each of the following six construct scales: Level of Aspiration; Level of Anxiety; Job Interest and 58 Satisfaction; Leadership and Initiative; Identification versus Alienation; and Level of Job Stress. Additionally, three subtests consisting of five items were taken from each of the six factor subscales. The instrument is designed in the form of a Likert- type scale with differential weights assigned to the responses. The individual is asked to respond to each question by selecting one of the five response | I alternatives: Never. Seldom. About Half the Time. Very j Often. and Always. The respective scoring weights (1, 2, I 3, 4, or 5) assigned to the responses indicate level of | I agreement. The higher the level of agreement with the I responses, the greater the possession of the hypothesized j construct factor. I Demographic variables. Altogether, the demographic variables on the Information Sheet were considered in the I analysis of the data : Level of Formal Education j (with scoring points indicated) i High School (1 to 4 years) = 1 College (1 to 4 semesters) = 2 College (3 to 4 years) = 3 Postgraduate Study = 4 (including a master's degree or doctorate) Number of Years of Military Duty (the number entered as a datum) 59 Age (number of years entered as a datum) Gender (F = 1, M = 2) Military Rank (commissioned officers coded as 2; enlisted personnel, as 1) Data analyses. In addition to the calculation of j means, standard deviations, and intercorrelations of I I scores on the six subscales of the DOSC-Form W, I ' comparable information was obtained for the battery of 18 I subtests (each consisting of five items) formed from the ! I six subscales (three subtests per subscale). Scores on these 18 subtests were intercorrelated and subsequently factor analyzed. Both exploratory orthogonal (varimax) and oblique (promax) factor solutions were obtained I following the extraction of six principal factors in harmony with the six hypothesized constructs for the DOSC-Form W. Confirmatory maximum likelihood factor analyses involving a null model of no-factor structure, a general (one-factor) model, a two-factor oblique model comprising Level of Anxiety and Level of Job Stress versus all four other hypothesized factors, and a six- factor oblique model corresponding to the six hypothesized constructs were carried out along with the determination of indexes of goodness-of-fit. To obtain information concerning the internal consistency of each of the six factor subscales two approaches were employed. First, alpha coefficients 60 were derived for each subscale. Second, every item in a given subscale was correlated with the total score of the subscale to which it was intended to belong (with the correction for spurious overlap being made) as well as with the total score of each of the remaining five subscales of which it was not a member. Selected Findings Corresponding to each research question in the order in which it was posed, the following statistical outcomes resulted: 1. In the DOSC-Form W subscales of Level of Aspiration, Level of Anxiety, Job Interest and Satisfaction, Leadership and Initiative, Identification vs. Alienation, and Level of Job Stress, the internal- consistency estimates of reliability (alpha coefficients) were .87, .88, .92, .92, .92, and .90, respectively: moreover, every single item was more highly correlated with the total score of the subscale of which it was a member than with the total score of any one of the remaining five scales. 2. Intercorrelations among the six-factor subscales varied between -.34 and .60 (median absolute value of all coefficients being .28); the two highest intercorrelations of .60 and .56 occurred between the Level of Aspiration and Job Interest and Satisfaction, 61 and between Job Interest and Satisfaction and Identification vs. Alienation subscales, respectively. 3. The orthogonal (varimax) and oblique (promax) exploratory factor solutions afforded exceedingly high and moderately high degrees, respectively, of simple structure; in other words, in each factor solution, the three five-item subtests from each of the six factor subscales intended to represent a hypothesized construct provided higher loadings on one rotated factor than on any one of the other five rotated factors as well as higher loadings on the factor it identified than did any one of the subtests in the other five sets of three subtests intended to define one of the five other factors. 4. In the confirmatory maximum likelihood factor analyses the six-factor oblique model hypothesized to represent the six constructs of the DOSC-Form W furnished indexes of goodness-of-fit that accounted for a higher proportion of covariance among the 18 five-item subtests than did either a general (one-factor) or a two-factor oblique model. 5. Intercorrelations among the six factors obtained in the oblique (promax) solution afforded a pattern quite similar to that of intercorrelations among factors in the confirmatory factor analytic solution involving six 62 oblique factors. In the promax solution intercorrelations among factors ranged from -.24 to .53 j with the median algebraic and absolute values being .10 and .22, respectively. The two highest intercorrelations j i of .53 and .51 occurred between the Level of Aspiration j and Job Interest and Satisfaction factors and between Job Interest and Satisfaction and Identification versus Alienation factors, respectively. In the confirmatory six-factor oblique solution, the intercorrelations among factors (corrected for attenuation) varied between -.38 and .69 with the median algebraic and absolute values being .09 and .38, respectively. Once again, the two highest intercorrelations of .69 and .62 existed between the Level of Aspiration and Job Interest and Satisfaction I factors and between Job Interest and Satisfaction and Identification versus Alienation factors, respectively. 6. As the only subscale to exhibit statistically significant correlation coefficients with any one of the five demographic variables, the Leadership and Initiative measure demonstrated coefficients of .30, .28, .30, and .23 (all p < .001), respectively, with Level of Formal Education, Number of Years of Military Duty, Age, and Military Rank. The coefficient of -.11 for Gender was not statistically significant. 63 Conclusions The following conclusions evolved from the data analyses : 1. All six 15-item factor subscales of the DOSC- Form W appear to be highly reliable and quite homogeneous. 2. In general, the relatively low to modest intercorrelations among the six-factor subscales of the DOSC-Form W suggest the presence of at least a moderate degree of discriminant validity. 3. Striking evidence from both exploratory factor analyses and confirmatory maximum likelihood factor analyses exists for the construct validity of each of the six subscales of the DOSC-Form W. 4. With the possible single exception of the Leadership and Initiative subscale, demographic variables show little if any relationship with the subscales of the DOSC-Form W. 5. It would appear that self-concept among the individuals in the military is a multidimensional construct just as it is in either an academic setting or in a working environment associated with a high technology industry. 64 Recommendations The following recommendations arise from the data i Ï outcomes of this inquiry: I 1. The DOSC-Form W is sufficiently reliable and valid to permit its use in the counseling of military personnel— especially those who are nearing completion of their military commitment and who are seeking information about themselves in relation to future work assignments in the civilian sector. ^ 2. The current investigation should be replicated in other military settings to ascertain whether additional support can be found for the reliability and construct validity of the DOSC-Form W. i I 3. Additional investigations should be undertaken j to ascertain whether any one of the subscales of the j I DOSC-Form W shows criterion-related validity— that is, I validity in forecasting future success in an environment ! of work in the civilian sector. ! 65 REFERENCES I ! Bentler, P. M., & Bonett, D. G. (1980). Significance tests and goodness of fit in the analysis of covariance structures. Psychological Bulletin, M(3), 588-606. Caracosta R., & Michael, W. B. (1986). The construct and concurrent validity of a measure of academic self-concept and one of locus of control for a sample of university students. Educational and Psychological Measurement. 46, 735-744. I Coopersmith, S. (1967). The Antecedents of Self-esteem. j San Francisco: W. H. Freeman. I Crandall, V. C., Katkovsky, W., & Crandall, V. J. 1 (1965). Children's beliefs in their control of I reinforcements in intellectual achievement behaviors. Child Development. 36, 90-109. j Crowder, B., & Michael, W. B. (1989a). The measurement of self-concept in an employment setting. « Educational and Psvchological Measurement. 49. 19- 31. I Crowder, B., & Michael, W. B. (1989b). The construct validity of a revised form of self-concept measure for employees in a work setting. Educational and i Psychological Measurement. 49, 421-428. ■ Î I Crowder, B., & Michael, W. B. (1991). The development ! and validation of a short form of a multidimensional self-concept measure for high technology employees. Educational and Psvchological Measurement. 51. 447- 454. Darakjian, G. P., & Michael, W. B. (1982). Comparative validities of standardized academic self-concept scales and achievement test measures and of teacher ratings of citizenship and effort in forecasting performance of junior high school students. Educational and Psychological Measurement, 42, 629-641. 66 Darakjian, G. P., & Michael, W. B. (1983). The long term comparative predictive validities of standardized measures of achievement and academic self-concept for a sample of secondary school students. Educational and Psychological Measurement, 43, 251-260. Darakjian, G. P., Michael, W. B., & Knapp-Lee, L. (1984). The predictive validity of subscales of an academic self-concept measure administered nine semesters prior to acquisition of criterion data reflecting school achievement. Educational and Psychological Measurement. 44. 715-720. I i Darakjian, G. P., Michael, W. B., & Knapp-Lee, L. (1985). The long-term predictive validity of an academic self-concept measure relative to criteria of secondary school grades earned over eleven semesters. Educational and Psychological Measurement, 45, 397-400. Denny, B. L. (1984). The development and validation of I an academic self-concept measure for college students. Unpublished doctoral dissertation, I University of Southern California. I Fernandes, L. M., Michael, W. B., & Smith, R. A. i (1978). The factorial validity of three forms of ; the dimensions of self-concept measure. Educational and Psychological Measurement. 38. 537-545. I j Gold, Y ., & Michael, W. B. (1985). Academic self- concept correlates of potential burnout in a sample of first-semester elementary-school practice teachers: A concurrent validity study. Educational and Psychological Measurement. 45, 909-914. ■ Halote, B., & Michael, W. B. (1984). The construct and I concurrent validity of two college-level academic I self-concept scales for a sample of primarily I Hispanic community college students. Educational ‘ and Psychological Measurement. 44. 993-1007. Hieronymus, A. N., & Lindquist, E. F. (1971). Iowa Tests of Basic Skills. Boston: Houghton-Miff1in. Hieronymus, A. N., & Lindquist, E. F. (1974). Iowa Tests of Basic Skills: Manual for administrators. supervisors, and counselors. Boston : Houghton- Mifflin. 67 ; ____ I r ----------------------------------------------------------------------------------------------------------- - ; Joreskog, K. G., & Sorbom, D. (1989). LISREL VII. (2nd 1 éd.). Chicago, IL: SPSS Inc. I Joreskog, K. G., & Sorbom, D. (1989a). LISREL 7 : A i guide to the program and applications (2nd éd.). Chicago, IL: SPSS Inc. Joreskog, K. G., & Sorbom, D. (1989b). LISREL 7: User's reference guide (1st ed.). Mooresville, IN: Scientific Software, Inc. LaBenne, W. D., & Greene, B. I. (1969). Educational implications of self-concept theory. Pacific Palisades, CA: Goodyear. Lehn, T., Vladovic, R., & Michael, W. B. (1980). The short-term predictive validity of a standardized reading test and of scales reflecting six dimensions of academic self-concept relative to selected high school achievement criteria for four ethnic groups. Educational and Psychological Measurement. 40., 1017- 1031. Marsh, H. W., & Byrne, B. (1990). The differentiated additive androgyny model: Relations between masculinity, femininity and multiple dimensions of self-concept. (ERIC Document Reproduction Service No. ED 328562) Maslach, C ., & Jackson, S. E. (1981). Maslach Burnout Inventory: Manual. Palo Alto, CA: Consulting Psychologists Press. Michael, W. B., Denny, B., Ireland-Galman, M., & Michael, J. J. (1986-1987). The factorial validity of a college-level form of an academic self-concept scale. Educational Research Quarterly. 11(1), 34-39. Michael, W. B ., Denny, B., Knapp-Lee, L., & Michael, J. J. (1984). The development and validation of a preliminary research form of an academic self- concept measure for college students. Educational and Psychological Measurement. 44, 373-381. 68 Michael, W. B., Kim, I. K., & Michael, J. J. (1984). The factorial validity of the Dimensions of Self- Concept (DOSC) measure for a sample of eighth-grade children and for one of community college adults. Educational and Psychological Measurement. 44. 413, 421. Michael, W. B., Michael, J. J., & Denny, B. L. (1985). Dimensions of Self-Concept Form H-Colleae. San Diego: EdITS. Michael, W. B., & Smith, R. A. (1976). The development and preliminary validation of three forms of a self-concept measure emphasizing school- related activities. Educational and Psychological Measurement ^ 36. 521-528. Michael, W. B., Smith, R. A., & Michael, J. J. (1984). Dimensions of Self-Concept (DOSC): A self- report inventory of five school-related factors of self-concept. Forms E and S: A technical manual. San Diego, CA: EdITS. Michael, W. B., Smith, R. A., & Michael, J. J. (1989). Dimensions of Self-Concept fDOSC): A technical manual (revised). San Diego, CA: EdITS. Naslund, R. A., Thorpe, L. P., & Lefever, D. W. (1972). SRA Achievement Series : SRA Assessment Survey. Multilevel Edition. Green Level. Grades 6-7^ Form E (Reading. Language Arts. Mathematics. Social Studies. Science. Use of Sources. Chicago : Science Research Associates. Norusis, M. J. (1990). SPSS/PC Plus: Statistical Package for the Social Sciences (version 4.0). Chicago: SPSS. Omizo, M. M. Hammett, V. L. , Loffredo, D. A., & Michael, W. B. (1981). The Dimensions of Self- Concept (DOSC) as predictors of academic achievement among Mexican-American junior high school students. Educational and Psychological Measurement. 41. 835-842. Reynolds, W. M. (1981, April). Measurement of academic self-concept in college Students. Paper presented at the annual meeting of the American Educational Research Association, Los Angeles. 69 Reynolds, W. M., Ramirez, M. P., Magrina, A., & Allen, J. E. (1980). Initial development and validation of the Academic Self-Concept Scale. Educational and Psychological Measurement. 40., 1013-1016. SAS Institute. (1990). SAS/STAT Guide for Personal Computers (version 6, 1st ed.). Cary, NC: Author. Shavelson, R. J., & Bolus, R. (1982). Self-concept: The interplay of theory and methods. Journal of Educational Psychology. 74(1), 3-17. Shavelson, R. J., Hubner, J. J., & Stanton, G. C. (1976). Self-concept: Validation of construct interpretations. Review of Educational Research. 46, 407-441. Wylie, R. C. (1974). The Self-Concept: Vol. I. A review of methodological considerations and measuring instruments frev, ed.). Lincoln, NE: University of Nebraska Press. Wylie, R. C. (1989). Measures of self-concept. Lincoln, NE: University of Nebraska Press. 70

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The reliability and construct validity of scores on the Dimensions of Self-Concept (DOSC)---Form W: Measure for a Navy sample

PDF

The 'Widmar' (1981) decision and the status of religious liberty on the public college and university campus: Antecedents, effects, and prospects

PDF

The development and validation of an academic self-concept measure for college students

PDF

The construct and concurrent validity of two college-level academic self-concept measures for a sample of primarily Hispanic community college students

PDF

Impact of an academic assistance program on the educational expectations and occupational aspirations of African American and Hispanic middle school parents for their children

PDF

Perceptions of family child care providers, users, and directors toward family child care provider training in the United States Army Europe

PDF

The demographics, social characteristics, and motivation of suburban home school families

PDF

A comparative study of the educational goals of students, administrators, and faculty at selected historically Black colleges and universities

PDF

The criterion-related and construct validity of scores on an academic self-concept measure for a sample of community college students

PDF

A comparison of the continuing education needs of civilian women and active duty Air Force women

PDF

The construct and concurrent validity of a measure of academic self-concept and one of locus of control for a sample of university students

PDF

The administrative and supervisory activities of junior high school principals and assistant principals in California

PDF

An analysis of student perceptions of the instructional process at the Academic Instuctor School of the United States Air Force University

PDF

The development and validation of a self-concept measure for working adults

PDF

Modification and validation of instruments for measurement of collective bargaining maturity/immaturity in school districts

PDF

A survey of the administration of Pacific Air Forces base education services program

PDF

Utility of demographic and psychological measures for the prediction of Army separation

PDF

Changes in academic self-concept, self-confidence, and grade point averages of migrant Mexican and Mexican-American community college students who served as mentors

PDF

The acceptance of Community College of the Air Force technical credit by two-year colleges, four-year colleges, and universities

PDF

An evaluation of goal congruence of two Air Force Reserve Officer Training Corps detachments as perceived by their cadre and cadet corps

Asset Metadata

Creator Foraker, Barbara McNeely (author)

Core Title The reliability and construct validity of the Dimensions of Self-Concept (DOSC)---Form W: Measure for an Air Force sample

Degree Doctor of Education

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag Education,OAI-PMH Harvest,Psychology,Social Sciences

Language English

Contributor Digitized by ProQuest (provenance)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c30-450821

Unique identifier UC11229397

Identifier DP27510.pdf (filename),usctheses-c30-450821 (legacy record id)

Legacy Identifier DP27510.pdf

Dmrecord 450821

Document Type Dissertation

Rights Foraker, Barbara McNeely

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA