UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Cross-cultural validity of Holland’s self-directed search with a specially designed measure of entrepreneurial… Masango, Sylvia Janet 1999

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1999-0427.pdf [ 5.18MB ]
Metadata
JSON: 831-1.0089076.json
JSON-LD: 831-1.0089076-ld.json
RDF/XML (Pretty): 831-1.0089076-rdf.xml
RDF/JSON: 831-1.0089076-rdf.json
Turtle: 831-1.0089076-turtle.txt
N-Triples: 831-1.0089076-rdf-ntriples.txt
Original Record: 831-1.0089076-source.json
Full Text
831-1.0089076-fulltext.txt
Citation
831-1.0089076.ris

Full Text

C R O S S - C U L T U R A L V A L I D I T Y OF H O L L A N D ' S S E L F - D I R E C T E D S E A R C H W I T H A S P E C I A L L Y D E S I G N E D M E A S U R E OF E N T R E P R E N E U R I A L I N T E N T I O N S by S Y L V I A J A N E T M A S A N G O Bachelor of Education (Technical/ Vocational), Huddersfield Polytechnic, United Kingdom, 1989 A THESIS S U B M I T T E D IN P A R T I A L F U L F I L L M E N T OF T H E R E Q U I R E M E N T S F O R T H E D E G R E E OF M A S T E R OF A R T S in T H E F A C U L T Y OF G R A D U A T E STUDIES Department of Educational and Counselling Psychology, and Special Education We accept this thesis as conforming to the required standard T H E U N I V E R S I T Y OF BRIT ISH C O L U M B I A June 1999 © Sylvia Janet Masango, 1999 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of ^ n / w c j t ^ J/^^Cp fytdj. (ojy The University of British Columbia -S^^CLcc^C IS <yL^C~c^J-t Vancouver, Canada Date DE-6 (2788) 11 A B S T R A C T The most notable attempt to identify and organize vocational interests has been Holland's model of vocational interests and personality types. Holland's theory hypothesizes six occupational interests and work environments: Realistic, Investigative, Artistic, Social, Enterprising, and Conventional. These are arranged in a hexagonal model through which the relationships and interactions among types and environments are defined using the Self-Directed Search (SDS). A review of the literature on the construct validity of Holland's SDS demonstrated support and generalizability of the model among the American population. Adequacy of the model in non-US cultures has not been established. Similarly, empirical evidence on the relation between Holland's SDS and a measure of Entrepreneurial Intentions (EI) is lacking. The current study was designed to (a) investigate the cross-cultural validity of Holland's SDS in terms of whether SDS generalizes to a sub-group of the Zimbabwe culture and (b) examine the relation between SDS and a newly developed measure of Entrepreneurial Intentions. The sample included 378 national diploma students from Zimbabwe representing five academic majors. Two hypothesized models defined the relationship between observed variables and the latent hypothetical constructs based on Holland's theory of vocational interests and work environments. According to the theory, adjacent types have a stronger relationship than alternate types, which, in turn, have a stronger relationship than opposite types. Confirmatory factor analysis (CFA) was computed to establish the hexagonal ordering of the R I A S E C types and the construct validity of the SDS in a Zimbabwean cultural setting. Support for the R I A S E C ordering was obtained on four of the six adjacent types. Results for the whole group showed strong correlations between the following adjacent types, Realistic and Investigative r = .70, Artistic and Social r = .51, Social and Enterprising r = .61, and Enterprising and Conventional r = .68. Low correlations were obtained between Investigative and Artistic r = .06 and between Realistic and Conventional r = -.15. Results o f an exploratory analysis with two subgroups of college majors showed a similar pattern of hexagonal ordering. Results of the goodness-of-fit indices with the whole group showed inadequate fit of the implied model to sample data. In addition, goodness of fit indices with the two subgroups of college majors showed a similar inadequate fit. Comparison of means by gender between Holland's (1994)-college sample and the Zimbabwean College sample showed differences. Mean differences for the male group ranged from -1.20 to 15.86 and that of the female group ranged from .58 to -12.45. Confirmatory factor analysis with the whole group was used to test the relationship between SDS subscales (Social and Enterprising) and Entrepreneurial Intentions subscales. Chi-square ^ 2 (335 , /V = 290) = 783.25,/? < .001, A D G F = .81, and R M S E A = .07 were obtained. Results on the specified models showed a poor model fit. Given the results of the current study, it appears accommodation of the cultural value dimension may be needed i f Holland's model is to provide a better assessment of occupational interests and work environments in Zimbabwe. In light of this, future research should focus at modifying Holland's theory in the context of the Zimbabwean cultural setting. TABLE OF CONTENTS Page A B S T R A C T - i i T A B L E OF C O N T E N T S iv L IST OF T A B L E S vi i i L IST OF F I G U R E S ix A K N O W L E D G E M E N T S x C H A P T E R 1. I N T R O D U C T I O N 1 Background — 3 The Problem - 5 Purpose of the S t u d y — 6 Rationale 8 Contribution 9 2. R E V I E W OF L I T E R A T U R E - 10 General Vocational Interests Theories — 10 Career Development Theory — 11 Work Adjustment Theory 11 Theory of Reasoned Act ion 12 Holland's Theory of Personality Types 12 Structural Setting of the Hexagonal Model 14 Randomized Test- - 16 Confirmatory Analysis - 16 Cluster Analysis 17 Construct Validity of the Self-Directed Search -20 Person-Environment Congruence - 24 Relations between SDS and Entrepreneurship 27 Theory of Entrepreneurship 27 Entrepreneurial Aspirations 28 Characteristics of an Entrepreneur 29 Measures of Entrepreneurship 30 Summary -.— 31 3. M E T H O D O L O G Y 34 Measures — - - 34 Published Measure 34 Developed Measure 35 Procedure 35 Scale Specification —36 Development of EI Questionnaire - — — 36 Pretest — 38 Item Writing and Assessment — 39 Reliability - — 40 Validity Assessment 41 Participants — — --41 Data Analysis - - - 42 Hypothesized Models 43 Goodness-of Fit Indices 46 Summary 47 4. R E S U L T S - — — — 51 Sample Size and Missing Data — 51 Preliminary Analysis 52 Reliabil ity Analysis — - 52 Test-Retest —- — — 53 Validity Assessment 54 Content-Related Validity 54 Criterion-Related Validity 55 Construct-Related Validity — — —56 Analysis of Mean Differences 57 Confirmatory Factor Analysis 60 Whole Group Analysis —61 Sub-group Analysis 65 Analysis of SDS and EI Subscales 71 Summary 75 5 D I S C U S S I O N 76 Integrative Summary of Results — 76 Limitations of Study — 84 Contributions and Conclusions 84 R E F E R E N C E S - - - 87 L IST OF A P P E N D I C E S — - - — 99 A Letters 100 B Questionnaires - 105 C Missing Data - 1 1 5 D Item-Total Correlation T a b l e s — — - 117 SDS and EI Scales and Correlation Matrices Table of Specification LIST OF TABLES Page Table 1 Goodness-of-Fit Indices 47 Table 2 Response Rates by Sex and College Major — 5 2 Table 3 Internal Consistency Reliabil ity Estimates 53 Table 4 Test-Retest Correlation Coefficient for EI 54 Table 5 Criterion-Related Validity - 56 Table 6 Comparison of Means — 5 8 Table 7 Mean Differences 59 Table 8 T-values for Parameter Estimates (Whole Group) 62 Table 9 Summary of the Goodness-of-fit results -66 Table 10 T-values for Parameter Estimates (Bus, and Educ. Sub-group) 67 Table 11 T-values for Parameter Estimates (Engineering and Science) 68 Table 12 T-values for Parameter Estimates (SDS and EI) 71 Table 13 Intercorrelations Between SDS and EI Subscales 74 ix L I S T O F F I G U R E S Page Figure 1 The Hexagonal Model — 4 Figure 2 Models of R I A S E C Personality Type— - 14 Figure 3 Dimensional Orientations of the R I A S E C Model — 26 Figure 4 Hypothesized Measurement Model of R I A S E C 49 Figure 5 Hypothesized Measurement Model of SDS and EI — 5 0 Figure 6 R I A S E C Intercorrelations (Whole Group) 61 Figure 7 Whole Group Model (Lambda Coefficients) - 64 Figure 8 R I A S E C Intercorrelations (Business and Education) 65 Figure 9 R I A S E C Intercorrelations (Engineering and Science) 65 Figure 10 Business and Education Model (Lambda Coefficients) 69 Figure 11 Engineering and Science Model (Lambda Coefficients) 70 Figure 12 SDS / EI Model (Lambda Coefficients) — — 73 X A C K N O W L E D G E M E N T S First and foremost, I would like to express my sincere thanks to my beloved husband, Rangai, for all that he had to endure during the long separation. Though thousands of miles away from each other for almost three years, your love never faltered. Your patience, understanding, and unwavering support served as the driving force. Accordingly, to you Rangai, I dedicate this work as a token of my appreciation. To my mother, Wanzirai, I say thank you for instilling in me the virtues of perseverance and hard work. Y o u have been a wonderful mentor. In the same vein, I would like to thank my sister in-law, Verna Utete, for providing expert advice on scale development and my brother in-law, Tichaona, my sisters, Sarah, Winfred, and Rosemary for their encouragement, moral support, and belief in my ability. Special thanks go to Rosemary for single handedly managing the family project, during the three-year period that I have been away. I would also like to express my sincere thanks to my wonderful friends, Annerose and B i l l Sims, for their help and support during my period of study in Canada. I would like to thank Dr. Nand Kishor, my supervisor, for the professional guidance and encouragement he provided throughout this study. I wish to express my sincere thanks and appreciation for the support that he provided at the most critical moments. M y sincere thanks go to my committee members, Dr. Kadriye Ercikan and Dr. Wi l l iam McKee for their support and professional input. Finally, I would like to express my sincere appreciation to Dr Mar ia Trache for assisting me with my data analysis and Maxwel l Chimhina, my research assistant, for his assistance in the data collection. 1 C H A P T E R 1 I N T R O D U C T I O N Self-estimates of vocational interests and their measures have been the subject of research since the early 1940s. The measures, which are widely used for career planning in colleges/schools, industry and government agencies, are generally acknowledged as valuable aids to educational guidance and identification of vocational pursuits. Nunnally (1978) noted that scores in interest measures are, by and large, predictive of an individual's vocational preferences, degree of satisfaction with occupations, and possibility of changes in occupations. Interests, in the context of vocational preferences, refer to an individual's preference for particular vocational pursuits or work activities. Gable (1986) described interests in terms of targets, direction, and intensity. Wherein targets of interests are activities, the direction could be described as interested/disinterested or liked/disliked, and the intensity could be high or low. Thus, interests with high intensity compel an individual to seek out the preferred activity. To date, however, there has not been consensus on the psychometric properties of the vocational measures and their validity within and across cultures (Athanasou & Cooksey, 1993; Furnham, Toop, Lewis, & Fisher, 1995). Generally, an individual's disposition is considered to be the product of a characteristic interaction among a variety of cultural and personal forces including peers, parents, social class, culture, and the physical environment (Holland, Fritzsche, & Powel l , 1994). In line with this notion, Rounds and Tracey (1996) reported a growing interest in the problems of applying and generalizing models and measures generated on one population, across cultures and ethnic backgrounds. The assumption is that a person's interests and competencies create a 2 particular personal disposition that leads the individual to acquire a personality type. Accordingly, an individual wi l l seek an occupational environment that corresponds to his/her orientation. Indeed, the concern throughout the world has been the need for a better fit between personal dispositions and job demands. The need for vocational measures that can best demonstrate this relationship cannot be overemphasized. In keeping with this notion, Fitzgerald and Rounds (1989) underscored the importance of studies that addressed career activities considering the fact that technological, demographic, and structural changes to work emphasize the ongoing need for adaptation and adjustment throughout adult life. In the same vein, Osipow (1983) reported that the varieties of work and the ways individuals choose their work, more especially in Western culture, have attracted the interest of behavioral scientists. This is due to the importance work holds in human behavior. Knowledge of personal characteristics corresponding with job demands can be an important source of information, not only to the individual, but also to industry, institutions of learning, policy makers. Sadly, the pursuit of such interests by behavioral scientists has not been as widespread in Afr ica, especially in Zimbabwe, where diagnosis of career orientation would go a long way in determining the degree to which an individual's personal qualities match the environmental demands of an occupation and vice versa. According to a theory of reasoned action (Ajzen & Fishbein, 1980), it is explained that "to predict and understand an individual's behavior, it is important to identify the determinants of intentions and measure the behavior of interest" (pp. 27). The theory elaborates that a person's intentions are a function of two basic determinants, one personal in nature and the other reflecting social influence. 3 Background During the past 30 years, the most persistent and notable attempts to identify and organize vocational interests and personality traits have emerged from Holland's model of vocational interests and personality types (Holland, 1985a). Empirical evidence has shown Holland's vocational interest inventory as one of the most widely used measures of vocational interests (Dumenci, 1995; Ferreira & Wood, 1995; Gottfredson, 1978; Upperman & Church, 1995). The measure has been acclaimed as a standard model for conceptualizing vocational interests and environments. More specifically, Holland's (1985a) theory of vocational interest assumes that individuals make occupational choices, which place them in environments that are compatible with their predominant personality characteristic. The theory asserts that people search for conducive environments that let them exercise their skills and abilities, express their attitudes and values, and take on problems and roles they find stimulating and satisfying, and avoid responsibilities they find distasteful and formidable. Central to Holland's (1985a) theory is the idea that people can be categorized meaningfully on the basis of their vocational interests as one of six personality types: • Realistic (R) - Has mechanical ability (e.g., auto-mechanic, farmer, surveyor, and electrician). Scores high on traits such as concreteness. Described as asocial. • Investigative (I) - Has mathematical and scientific ability (e.g., biologist, chemist, and physicist). Scores high in situations involving critical thinking. Described as analytical. • Artistic (A) - Has artistic ability (e.g., designer, architect, and musician). Scores high in artistic and self-expression. Described as imaginative and intuitive. Prone to expressing emotions. 4 • Social (S) - Has interpersonal skills (e.g., teacher and counselor). Described as empathetic and sociable. Avoids situations involving intellectual problem solving. • Enterprising (E) - Has persuasive and verbal skills for manipulating and dominating people (e.g., sales person, T V producer, and entrepreneur). Aspires to power and status. Described as, adventurous, ambitious, extroverted, sociable, and self-confident. • Conventional (C) - Typified by a great concern for rules and regulations, prefers structure and order (e.g., accountant, banker, tax expert, and clerk). Described as efficient, inflexible, unimaginative, and methodical. Has clerical and arithmetic ability. The six personality types are referred to as R I A S E C , and a hexagon is used to depict relationships among the six types (Figure 1). Figure 1. The hexagon model interpreting person-environment relations (Holland, 1985). The typology of six personality types is theoretically meant to differentiate among vocational preferences. Holland (1985a) argued that individuals are most successful when they operate in environments that are congruent with their personality type, because such environments provide opportunities that are more compatible with their aspirations. Assessment and classification of vocational interests and patterns is achieved through . the administration of one of Holland's vocational inventories, the Self-Directed Search (SDS; 5 Holland, 1985b). The SDS is a questionnaire developed to measure the six R I A S E C traits outlined in Figure 1. These are measured by comparing them against four content domains: Activit ies, Competencies, Occupational Preferences, and Self-Estimates of abilities. Hence, the SDS was designed to (a) explore the entire range of occupational possibilities, (b) identify new options or confirm a choice already made, and (c) identify college majors congruent with a student's interests. The Prob lem Although Tokar and Swanson (1995) noted that a large number of studies found some evidence of validity for Holland's trait characterization of the types, the evidence to date lacks consistency and robustness as noted by Broughton, Trapnell, and Boyes (1991). Dumenci (1995) also found from the literature that construct validity studies have provided mixed support for the structure of the SDS when exploratory or confirmatory procedures were used. For example, on administering SDS and 13 ability measures to a group of women undergraduates, Lowman and Wil l iams (1987) noted that the results obtained from the multitrait multi-method analysis failed to support Holland's theoretical model. Similarly, in a factor analytic study of the construct validity of Holland's SDS test, Rachman, Amernic, and Aranya (1981) found four factors associated with Realistic, Investigative, Artistic, and Conventional. The Social and Enterprise subscales, in combination, formed one factor. On the other hand, confirmatory factor analysis of the six-factor model showed a rejection (Boyle & Fabris, 1992), or moderate fit (Khan & A l v i , 1991). In the same vein, Rounds and Tracey (1996) found that the SDS was not supported, contrary to views about the alleged construct validity of SDS across cultures. Thus, there is need for further investigations of the construct validity of Holland's hexagonal model, especially in 6 cultural settings other than North American. The psychometric properties of the SDS across cultures needed further investigation i f the inferences of outcomes from the hexagon model were to be generalized across cultures and settings. Purpose of the Study The purpose of the current study was to investigate the construct validity of Holland's (1994) SDS in a non-western cultural setting (Zimbabwe). As stated earlier, it had not been empirically demonstrated that the SDS had utility for describing the occupational interests of vocational and technical students in Zimbabwe, a group as yet unexamined in the vocational interest literature. It was possible that Holland's model might lack validity in the Zimbabwean cultural setting. Therefore, to maximize the benefits given the costs, an additional purpose of the study was to determine i f a reliable and valid measure of Entrepreneurial Intentions (EI) could be developed based on 2 of the 6 subsets of the SDS (Enterprising and Social). The entrepreneurship scale was designed to measure an individual's entrepreneurial intentions. The quest for such a measure in Zimbabwe cannot be underestimated. The major concern in both the public and private sectors has been the mismatch between an individual's competencies and the program of study or occupation one was pursuing. Over and above this, plans are afoot to turn some of the colleges into "technical commercial colleges" specializing in entrepreneurship. To this end, the present study may yield a practical benefit as well. Hence, should Holland's model hold in this culture, it would provide valuable information for the validation of the EI scale. Based on Holland's (1985a) theoretical definitions, it was predicted that Enterprising and Social subsets of Holland's R I A S E C should be strongly correlated with characteristics depicting an entrepreneur. Grant (1996) defined entrepreneurial intentions as one's 7 judgements about the likelihood of owning a business. A n entrepreneur is one who identifies and exploits an opportunity (Peterson, 1985) and develops a niche in the market to satisfy some need (Garfield, 1986). Entrepreneurial characteristics include persistence, risk-taking, achievement oriented, and a proactive personality. Bateman and Grant (1993) described a "proactive personality" as one who is unconstrained by situational forces and one who effects environmental change. Individuals with proactive personalities identify opportunities and act on them. Thus, a visionary and action-oriented personality is core to business success. In this regard, the psychometric properties of Holland's SDS needed further investigation. Psychometric evidence, especially construct validity, has been acknowledged as proof of the adequacy, appropriateness, and relevance of a test or measure. The key issue as posited by Cronbach and Meehl (1955) is the significance of validity. Cronbach and Meehl explained that construct validation is involved whenever a test is to be interpreted as a measure of some attribute. In keeping with this assumption, Messick (1980) emphasized the fact that the adequacy of a test, as a measure of the characteristic it is interpreted to assess, was answerable on scientific grounds, by appraising the psychometric evidence in terms of measurement properties, especially construct validity. Aspects to be considered whenever a test is proposed for a specific purpose include, "is the test any good as a measure of the characteristics it is interpreted to assess?" This question, according to Messick (1989), is a technical and scientific one and may be answered by appraising evidence bearing on the test's psychometric properties, especially construct validity. The second question is, "should the test be used for the proposed purpose in the proposed way" (pp. 29). This is an ethical question and its answer requires an evaluation of the potential consequences of the testing in terms of 8 social values. In particular, he stated that the scope of a measure should go beyond clinical and counseling usage. Rat ionale Despite numerous studies on Holland's SDS, relatively little research has been undertaken in Afr ica. Possible ramifications of the inferred score meaning have not been sufficiently explored. In this regard, investigating uniformity and differences of tests across cultural groups and settings, more especially in Afr ica, would provide valuable information. In addition, such information should help establish the generalizability of test interpretation and use. Messick (1975) underscored the fact that "the extent to which a measure displays the same properties and patterns under different circumstances, then, is an important empirical question" (pp. 956). The present study endeavored to establish the construct validity of the SDS across cultures. In addition, a newly developed measure of entrepreneurship was used as index to establish the extent to which SDS subsets (social and enterprising) were related to the entrepreneurial intention subscales. Holland's theory assumes that vocational interests are an important expression of personality. In view of this assumption, subsets of Holland's R I A S E C were hypothesized to correspond meaningfully with the measure of Entrepreneurship. Interestingly, empirical evidence on the relationship between entrepreneurial personality traits and Holland's hexagonal model is lacking. Hence, the development and validation of a measure of entrepreneurship in terms of its measurement characteristics, based on Holland's SDS, was (a) expected to contribute to knowledge of important theoretical issues applied within the education domain and (b) to provide information with potential to impact on vocational and technical education in Zimbabwe. 9 Contr ibut ion This research has contributed to our knowledge and understanding in two parallel ways. First, the results are likely to advance our knowledge about the generality and validity of Holland's SDS. It may also contribute to the theoretical understanding of personality type and occupational choice. Over and above this, further examination of two of Holland's six personality traits and work environments (Social and Enterprising) using a specially designed measure of entrepreneurial intentions provided additional information on the validity of Holland's R I A S E C model. Secondly, the study made methodological contributions. The use of the measurement model to investigate cross-cultural validity of the SDS facilitated in the examination of the covariation among a set of observed variables in order to gather information on their underlying constructs. 10 C H A P T E R 2 R E V I E W O F L I T E R A T U R E Introduction The review of literature in this chapter was to provide the rationale for a theory-based approach to construct validation of Holland's (1985b) Self-Directed Search across cultures. Rachman, Amernic, and Aranya (1981) asserted that "to establish construct validity one requires a theory which can provide information about the structure of the measure and about the relationship between the measure and previously developed and validated measures" (p.426). Included in this review was a description of four principal theories of vocational psychology, with particular emphasis placed upon Holland's (1985a) theory of personality types and work environment. Next, a number of research articles that dealt with (a) the structural setting of Holland's hexagonal model, (b) construct validity of the SDS, and (c) person-environment congruence were examined to ascertain the hexagonal model's invariance across cultures. Finally, related theory and research in entrepreneurial characteristics and traits uniquely associated with entrepreneurship was examined as a guide towards developing a measure of entrepreneurship. General Vocat ional Interest Theories The basis for vocational psychology, as noted by Tracey and Rounds (1993), dates back to about a century ago when Parsons (1909) proposed a tripartite model of choosing a vocation: (a) self analysis of one's abilities, aptitudes, interest, ambitions, and resources, (b) occupational analysis of work requirements, tasks, and opportunities, and (c) the use of "true reason" to relate the self and occupational analyses. In line with this notion, Thurstone, 11 Paterson and Darley, Kuder , and Strong (cited in Tracey & Rounds, 1993), were instrumental in the development of occupational and vocational interest inventories. Hansen and Campbell, in their study, (cited in Tracey & Rounds, 1993), posited that interest inventories have been used to help students, employees and employers make occupational and vocational decisions as well as determine the most suitable approaches to organizing human work environments. Similarly, theories on vocational and occupational interests were developed. Kerlinger (1973) referred to a theory as "a set of interrelated constructs (concepts), definitions, and propositions that represent a systematic view of phenomenon by specifying relations among variables, with the purpose of explaining and predicting the phenomenon" (p.4). Evidence from literature showed that theories in the area of career development were diverse. Accordingly, Osipow (1990) identified four major theoretical approaches that appeared to dominate thinking about careers. Thus, consistent with Osipow's findings, four theories pertinent to the current study were examined. Career Development Theory (Super, 1990) Super's developmental theory was formulated on the premise of life stages in human development and their impact on personality and vocational choice. According to Super, Starishevsky, Mat l in, and Jordan (cited in Super, 1990), the theoretical framework was based on the assumption that people are differentially qualified for occupations and their interests and abilities fall into similar patterns. The assumption was that vocational and occupational preferences are indicative of personality type. W o r k Adjustment Theory (Lofquist & Dawis, 1983) The theory advances personality type dimensions to include flexibility, activeness, and reactiveness. Flexibil i ty is associated with work history involving a wide range of situations, 12 jobs, and people (social type). Activeness refers to leadership, organizational skills and achievements, including a history of taking initiative (enterprising type). Reactiveness refers to rule abiding, fol lowing through assignments, preference for highly structured situations (conventional type). These constructs categorized people according to personality traits. Theory of Reasoned Action (Ajzen & Fishbein, 1980) The theory subscribes to the view that human beings are generally rational in the way they make use of information. The theory elaborates the fact that personality traits underlie and influence the behavior in question. Personal/attitudinal components and social/normative components are noted as key determinants of one's behavior impacting on one's beliefs, perceptions, and intentions. Ajzen and Fishbein (1980) underscored the fact that behavior is determined by an interaction between personality and environment. Thus, behaviors can vary in action, context, and time elements. For example, one's personality trait might induce the individual to portray certain types of behaviors, such as, domineering, aggressive, and altruistic behaviors. Theory of Personality Types and Work Environment (Holland, 1985a) Holland's theory hypothesizes six occupational types or themes arranged in a hexagonal model: Realistic, Investigative, Artistic, Social, Enterprising, and Conventional ( R I A S E C - see Figure 1, Chapter 1). The assumption, according to the theory is that, correlations for the adjacent types (R l , IA, A S , S E , E C , and CR) are greater than correlations for the alternate types (RA, IS, A E , S C , E R , and CI) which, in turn, are greater than correlations for the opposite type (RS, IE, and AS) . This three-factor model describes the circumplex ordering of the types with equal spacing around the circle. 13 Accordingly, Holland's theory is widely considered one of the most influential theories and occupational taxonomies in vocational psychology (Borgen, 1986; Brown & Brooks, 1990; Osipow, 1983). Central to the theory is a set of assumptions postulating that: 1. People can be characterized by their resemblance to each of the six R I A S E C personality types. A type being a cluster of personal attributes against which the individual can be measured. The theory implies that the profile of a person's resemblance to each of the six personality types - R I A S E C - is indicative of a person's vocational assets and liabilities (Holland, 1985a; Walsh & Holland, 1992). Hence, a person's interests and competencies create a particular personality disposition that leads to thinking, perceiving, and acting in a special way. 2. The environments in which people live can be characterized by each of the six traits, meaning that for each personality type, there is a parallel model environment. The assumption is that people tend to surround themselves with those who share the same interests, competencies, and general views about life and the world of work. 3. Person-environment congruence is associated with satisfaction, productivity, creativity, personal and vocational stability (Walsh & Holland, 1992). Hence, the hexagonal model provides a framework for assessing the degree of consistency in a personality pattern and the extent of congruency between a personality type and its environment (Rachman, Amernic, & Aranya, 1981). Congruence refers to the degree to which an individual's personal quality matches the occupational environment of one's preferred or planned field. In line with the theory, Holland (1985b) developed, among other measures, the Self-Directed Search (SDS); a questionnaire designed to measure the six vocational traits. The 14 SDS assesses the person's resemblance to each type through the R I A S E C subscales based on four content domains: Activit ies, Competencies, Occupations, and Self-Estimate of one's capabilities. The typology provides a parsimonious representation of data. Typology approaches accurately identify traits relevant to person-environment fit by describing the process that produced the relevant behavior (Gay, 1998). However, although Holland's theory of vocational interest has been the dominant model for the past 20 years, it appears there has not been universal support for the generalizability of the R I A S E C structure to various ethnic groups as evidenced in the studies that have investigated Holland's R I A S E C model (Tracey & Rounds, 1993). Gati (1991b) advocated convincingly for a hierarchical structure of interest (see Fig. 2). R I I I I I I R I A S E C Holland's Circumplex Tracey and Rounds' Alternative Three-Class Partitioning Figure 2. Representation of 3 models of R I A S E C (Tracey & Rounds, 1993) Structural Setting of the Hexagonal Model Although the hexagonal structure is acclaimed to be culturally invariant (Hansen, 1987; Holland, 1985a), this assertion has, however, not gone unchallenged. Gati (1982) argued that Holland's circular order structure did not adequately account for the relations among R I A S E C types. In particular, Trapnell (1989) reiterated that the hexagonal 15 dimensionality as outlined in Holland's theory, should be restricted to the universe of Holland's specified occupations, not to the universe of all possible vocational interests. In the same vein, Fouad and Dancer (1992) investigated similarity in the structure of occupational interest and the fit of a hexagonal circular model to data from a sample of 557 male professionals and student engineers from U.S. and Mexico. In essence, the authors compared the interests of student engineers and professionals in the United States and Mexico with a view to finding out whether Holland's hexagon, more especially its calculus assumption, was valid across cultures. Nonmetric multidimensional scaling was used in the analysis. Similarity of the structure between U.S. and Mexican professionals and student engineers was supported. The correlation ranged from r = .32 to r = .87 (median r = .65). Moderate support was found for the calculus assumption. The authors found that their data generally supported a circular arrangement, but did not support the equilateral hexagon model. The authors' use of a confirmatory multidimensional scaling allowed for goodness-of fit which, as noted by Chartrand (1992), is preferable to visual examination in determining hexagonal fit. What is worth noting is that Fouad and Dancer's (1992) sample did not represent the demographic composition of professionals and student engineers of the two countries, making generalization to that population difficult. In yet another study, Tracey and Rounds (1993) conducted a meta-analytic review of studies that evaluated the structural fit of Holland's (1985b) and Gati 's (1979, 1982) structural models (see F ig. 2). Included, were studies from 1970 to 1989 accessed from journals, research reports, and dissertations cited in published studies. Sample sizes ranged from 44 to 2,621. 16 The R I A S E C correlation matrix was considered the sample unit. Data set of 104 R I A S E C correlation matrices (77 U.S. and 27 international matrices) were integrated using three methods of statistical analysis: (a) Hubert and Arable's (1987) randomized test of hypothesized order relations which allowed a direct test of Holland and Gati 's order predictions, (b) confirmatory factor analysis, and (c) cluster analysis. Randomized Test The randomized test was conducted using R A N D O R D (Tracey, 1991) a F O R T R A N program. Holland and Gati 's order models accounted for mean correlations respectively of .63, .64 for the overall sample. However, results of the analysis of variance conducted on the corresponding indices across nationality (United States vs other countries) revealed support for Gati 's model for both U.S. and Non-U.S. groups F ( l , 101) = 0.32,p > .05, except for sex and age where the model was found to fit female matrices better than male matrices and 18 -22 year olds better than those below 18 and above 22. Holland's order model yielded the best fit to the data for U.S. matrices but not for the non-US sample. Confirmatory Analysis L I S R E L (Joreskog & Sorbom, 1986) was used to examine and establish possible constraints on the values among the factor correlations. The three-parameter and three-group partition models were used to describe the relation among the six personality types. Holland's circumplex structure was the most parsimonious and showed adequate fit to the data with a P G F T [(parsimonious goodness of fit index), (Mulaik et al., 1989)] of .94. The fit of the circumplex model to the 104 matrices confirmed the assumption of equality of correlations among adjacent types. Thus, support was found for the equilateral hexagon model. 17 Cluster Analysis Carol l and Arabie's I N D C L U S (as cited in Tracey & Rounds, 1993) was used to examine the discrete structure of the R I A S E C types. The best fit for the data was the six-linked-cluster solution: [(R, I), (I, A ) , (A, S), and (S, E) , (E, C) , (C, R)], R2 = .63. Tracey and Rounds (1993) noted a pattern where a new clustering of RI , A , S E C emerged on R I A S E C matrices drawn from the cross-cultural literature (Fig. 2). The clustering was considered logical in that an E type is related to an S type through involvement with people and likewise, an E type is related to an C type through business (accountant, bank teller, and so forth). Overall, Holland's order model was found to have a better fit than Gati 's across the 77 U S R I A S E C matrices, but lacked support across the 27 international R I A S E C matrices. Thus, the poor fit of Holland's order model to non U.S. matrices raised doubts regarding (a) the adequacy of using U.S. R I A S E C measures with non-U.S. populations, (b) the adequacy of the circumplex model in non-U. S. cultures, and (c) the adequacy of translated versions (non-English) R I A S E C measures. Interestingly, of the 104 matrices, only 27 matrices were from international countries and of these 15 matrices were from Australia. Excluded were Afr ica and Eastern Europe. The need for a more representative sample is apparent, notwithstanding the fact that the poor fit of Holland's order model to non-U.S. matrices made the validity of the typology across diverse cultures questionable. Expanding on the previous review, but in a more recent study, Rounds and Tracey (1996) performed another structural meta-analytic review to (a) examine the cross-cultural equivalence of Holland's (1985a) model of vocational interests across U.S ethnic groups and samples from international countries and (b) establish the generalizability and universality of 18 R I A S E C personality types by evaluating the extent to which constructs developed to characterize personality in one culture could be applied across diverse cultures. Rounds and Tracey (1996) argued that conclusions of reviews on the validity and generalizability of Holland's model across cultures were, in some cases, based on too few studies. For example, Hansen's (1987) conclusions were based on eight studies, and Holland (1985a) cited seven studies to support his conclusion. Further, the authors noted that primary studies that investigated the fit of Holland's order model across cultures had serious limitations that confounded conclusions made. These included methodological problems, such as, (a) inappropriate and widely varied methods to test Holland's hexagonal structure, (b) "Eye ball ing" the correlation matrix or spatial representation of the data, and (c) conducting the binomial test of predicted order proposed by Wakefield and Doughtie (1973), thus, violating the assumption of independent order relations among occupational types. Subsequently, Hubert and Arabie (1987) demonstrated, through a randomization test, the inappropriate use of Wakefield and Doughtie's (1973) binomial test which disregards the independent order relations among occupational types. Holland and Gati 's models and the alternate three-class partition were evaluated for data model fit (see Fig. 2). The R I A S E C correlation matrix was used as the sampling unit. Studies from 1970-1992 were considered. Ninety-six (96) R I A S E C matrices from 19 countries excluding Afr ica and East Europe were synthesized. Included were the matrices from the 1993 review with the 73 Anglo-American matrices serving as benchmark. The statistical analysis used to evaluate the cross-cultural generality and universality of the three R I A S E C models included Hubert and Arable's (1987) randomization test, multivariate analysis ( M A N O V A ) and post hoc paired t test comparisons of the 19 correspondence index (CI). Similar conclusions to the 1993 meta-analysis were obtained. Holland's circular order model was not supported. Holland's model had a perfect data-model fit for the 73 U.S. matrices, drawn from Anglo-American sample. Further analysis by country showed a good model fit of Holland's model with matrices from Iceland, Japan and Israel, but had a poor model fit on the total sample across cultures. These conclusions were consistent with Rounds and Zevon's (1983) research on Holland's circular order structure with non-Caucasian groups. The findings of these two meta-analytic reviews (Rounds & Tracey, 1996; Tracey & Rounds, 1993) demonstrated that Holland's (1985a) hypothesis that the R I A S E C model is culture invariant may not be accurate. The psychometric properties and validity of the SDS circular order structure, when applied outside the United States and across diverse cultures, shows mixed conclusions. The exclusion or non-representation of studies from Afr ica in both meta-analytic reviews raises questions about the utility of the typology in those countries. In the same vein, Glidden-Tracey and Parraga (1996) investigated the structure of R I A S E C types on 98 Bol iv ian University students. The Spanish version of Holland's (1973, form-E) cicumplex SDS, Gati 's (1979) three-group partition and Tracey and Rounds' (1996) alternate three-class partition were evaluated for cross-cultural universality. Hubert and Arable's (1987) randomized test of hypothesized order was applied in the analysis. The correspondence index (CI) for each of the models was Holland's circumplex model, CI= -0.11. Only 32 of the 72 model predictions were confirmed. This indicated lack of congruence between data and model; Gati 's model, CI = .22. Thus, 22 of the 36 model predictions were confirmed. Tracey and Rounds' model, CI = -. 23. Only 17 of the 44 model predictions were confirmed. Further test of the circumplex by gender yielded similar results. 20 A three-way multidimensional scaling (MDS) was conducted to further examine the structural fit between Holland's model and the Bol iv ian data. The variance accounted for ( V A R ) was 9 9 9 only i? = .20 and subscales of male and female accounted for, R = .14 and R =26 respectively. A poor data-model fit of the entire three R I A S E C model structures for the Bol iv ian sample was noted. Instead, a four-type model seemed to characterize the vocational interest of the Bol iv ian sample better than the six-personality type, with the interest subscales ordered as R I S C E A . Notwithstanding the fact that the findings raised some interesting questions about the basic psychometric information needed to evaluate the adequacy of a scale for a given population, more especially non-U.S. groups, the study had some limitations, as acknowledged by the authors. These included, sample size and selection criteria, non-stipulation of college major fields, and the use of a weaker and inappropriate form E to measure vocational interest of university students. Construct Va l id i ty of the Self-Directed Search Messick (1980) posited that questions of the adequacy of a test, as a measure of a characteristic it is interpreted to measure, were answerable by appraising the test's psychometric evidence and the appropriateness of its use. Chartrand (1992) expressed similar views by stating that basic psychometric information was needed to evaluate the adequacy of a scale for a given population. Bernstein, Garbin, and Teng (1988) referred to construct validity in terms of "how well a scale fulfil ls properties ascribed to it by a relevant theory" (p.379). Holland's typology has been shown to have considerable validity, especially with the Anglo-American population. However, indications from literature were that the construct 21 validity of the R I A S E C model has not been subjected to critical scrutiny, with regards to the psychometric properties of SDS (Campbell, 1985). Chartrand and Camp (1991) stated that, generally, conclusive information concerning vocational instruments' construct validity is lacking. In the same vein, Meier (1991) explained that construct validity, particularly discriminant validity, continued to be an elusive ideal for many vocational scales. In keeping with this notion, Rachman, Amernic, and Aranya (1981) investigated the construct validity of the SDS instrument and the hexagonal model based on Holland's theory. A focus sample o f 1206 Canadian professional accountants (CA's) participated in the study. The confirmatory factor analysis provided support for the validity of Holland's SDS. The results showed that the SDS measures six factors: Realistic, Investigative, Artistic, Conventional with Enterprising and Social combined to form a single factor, and a sixth general interest factor emerging. Rachman et al. (1981) concluded that the structure of the SDS as a whole was quite clear, and that most of the items which were supposed to measure R I A S E C personality types, tended to form unidimensional scales. However, the authors did not provide an indication of the fit of the models in terms of goodness of fit indices. The only demographic information given was about the participants' profession (Accountants). Khan and A l v i (1991) investigated the validity of Holland's model in a non-western culture by fitting SDS data to the model using confirmatory factor analysis. Gender differences in the fit of the data to the model was also investigated. A sample of 376 Pakistan University students majoring in engineering, medicine, fine arts, and commerce was administered an adapted version of Holland's (1977) SDS. Data were analyzed using unweighted least squares technique. Partial support for the construct validity of the SDS was noted. For the most part, there was a generally good fit for 22 the hypothesized factor structure for males - A G F I = .91. The model failed to provide a good fit for the data for females - a negative coefficient of determination was observed R2 = -.586. Strong similarity was noted between Enterprising and Conventional factors, r = .90, and between Social and Enterprising factors, r = .77. The validity and universality of the hexagonal model, as evidenced by contradictory conclusions remains questionable. Boyle and Fabris (1992) assessed the adequacy of the R I A S E C model using a combination of exploratory, congeneric and confirmatory factor analysis. A sample of 401 Australian apprentices and qualified plumbers participated in the study. A matrix model of 30 x 30 was specified. L I S R E L maximum-likelihood parameter estimation and goodness of fit indices provided an indication of the adequacy of the model. GFI was found to be 0.83, A G F I was 0.75 and R M S R was 0.08. Goodness of fit results suggested some inadequacy of the six-factor R I A S E C model. The goodness of fit indices from the congeneric factor analysis provided partial support for the validity of five of the six R I A S E C dimensions. The authors , stressed the need for revising the SDS instrument.by eliminating items accounting for the least amount of shared variance associated with the R I A S E C . The authors concluded that the findings provided only partial support for the validity o f Holland's R I A S E C model o f personality types. Results from exploratory factor analysis (EFA) supported fully the validity of the Artistic theme while both the Enterprising and Conventional dimensions combined to form a single factor. Results from C F A failed to support the model with A G F I = 0.8. The congeneric factor analysis results showed that the Realistic theme was not supported. Overall, the findings indicated the need for the R I A S E C model and the SDS instrument to be reviewed in order to improve the existing levels of validity. 23 In the same vein, Oosterveld (1994) investigated the adequacy of the SDS by fitting 10 factor model on Holland's (1985b, 1987) correlation matrices reported in the SDS manual. Data analysis was by means of weighted least square technique. Trait factor loadings of the subscales generally exceeded .50. The fit of the model was generally satisfactory: GFI = 0.96; A G F I = 0.95 and R M S R = 0.06. Incidentally, the parameter estimates and the fit measures for the male sample data were not shown, though referred to as being similar to the results for the female sample. On the whole, the results supported the validity of the instrument for the American population. But, as Tracey and Rounds (1993) posited, contradictory conclusions have been given on the validity of the SDS instrument when applied to groups other than those on which it was validated. Dumenci (1995) investigated the construct validity of the SDS. A sample of 700 college students representing a cross section of academic majors participated in the study. The author used Widaman's (1985) structural measurement approach of evaluating convergent and discriminant validity and method variance. Twenty (20) hierarchically nested models were specified a priori. Model T4M4 was considered parsimonious and the best fitting model representing the structure of the SDS for both males and females. Comparative fit index (CFI) and Tucker-Lewis Index (TLI) showed moderate model fit with CFI = .89, TL I = .87 for males, CFI = .86, TLI = .83 for females. Strong relations were noted between enterprising and conventional subscales r = .59 for males and r = .56 for females. Dumenci (1995) concluded that data of the college sample supported both the convergent and discriminant validity of the SDS. The author, however, sounded a word of caution on the generalizability of findings. 24 Oosterveld (1994) stated that, as a rule of thumb, the GFI and A G F I should be larger than 0.95 and R M S R close to zero. Bentler and Bonell (1980) stated, as acceptable, a model fit index of above .90. Cuttance and Ecob (1987) indicated that acceptable models have an A G F I index of 0.8 or higher. However, Dumenci's (1995) fit indices were well below the stipulated levels. Considering that 92% of the total sample were Anglo-American, the results were consistent with Tracey and Rounds' (1993, 1996) conclusions that Holland's SDS was best suited to the Anglo-American population. Person-environment congruence Osipow (1987) posited that the idea of a good match between people and their work remains at the core of career psychology. Finding the job niche best suited to one's characteristics should result in job performance, satisfaction, and productivity. Accordingly, Holland's theory postulates that people can be characterized by their resemblance to each of the R I A S E C types and corresponding environments. Person-environment (P E) congruence is associated with satisfaction, performance capability, and personal and vocational stability. Expanding on this view, Assouline and Meir (1987) conducted a meta-analytic review of the person-environment congruence. A total of 41 congruence studies yielding 77 correlation coefficients were accessed from published and unpublished sources (e.g. journals, books, dissertations, and so forth). Sample sizes for the studies varied from 12 to 7706 (median 120). The authors hypothesized that environmental reference, to which an individual's personality type was expected to be congruent and the measuring methods used should account for variance between person-environment fit. Synthesis of 53 of the 77 correlations showed mean correlation with the criterion of satisfaction to be .21. Further analysis of the 53 correlations broken down by measuring 25 method and environmental reference yielded mean correlation range of .30 to .40 between congruence and satisfaction. General lack of relationship was noted between personality and intentions and between personality type and achievement, with mean correlations as low as r = - .02 and r = .10 respectively. Hence, although there was a moderate relation between person-environment fit and satisfaction, the conclusion was that neither achievement nor stability was significantly associated with person-environment congruence hypothesis. In a related study, Camp and Chartrand (1992) compared the relative merit of congruence indices in predicting work satisfaction from occupational congruence. Data were collected from a sample of 125 undergraduate females enrolled on a career-planning course. Using 13 different indices proposed as measures for person-environment fit for Holland's (1973, 1985a) theory, the authors analyzed data using first-letter agreement, two-letter code agreement, and three-letter code agreement. Results showed that the congruence indices were not interchangeable, a clear indication that choice of congruence index had important implications for testing the viability of Holland's propositions. For example, correlation coefficients between congruence indices and the achievement and satisfaction criteria were generally weak, ranging from r = .00 to r = .26 for achievement and r = -.07 to r = .15 for satisfaction. The authors noted further that few of the congruence indices purported to measure Holland's person-environment fit conformed to the dictates of the theory. Limitations with regards to sample representation and narrow focus in academic majors were noted. Similarly, Tokar and Fischer (1998), expanding on the body of research on relations between personality and interest, examined the correspondence between Holland's vocational personality typology and five-factor model of personality. A sample of 490 employees drawn from a wide range of occupational settings (174 occupational titles) participated in the study. 26 Specifically, the authors investigated the relation between the five-factor model (i.e. neuroticism, extraversion, and openness to experience, agreeableness, and consciousness) and the dimensions underlying R I A S E C using Prediger's (1982) bipolar dimensions of data/ideas and things/people as shown in Figure 3. According to Prediger, (a) the data pole is characterized by a preference for impersonal ordered tasks, whereas the idea pole is characterized by a preference for interpersonal tasks involving abstraction, curiosity, and originality; (b) the things end is characterized by a preference for nonpersonal tasks involving machinery and working with one's hands, whereas the people end is characterized by interpersonal tasks. Data Ideas People Figure 3. Prediger's (1982) dimensional orientations of Holland's (1985a) R I A S E C model Strong relations were found between R I A S E C and the five-factor personality variables with intercorrelations of r = -.89 and r = -.88 for men and women respectively. The findings were significant when the congruence index was examined using a first-letter agreement index, but not significant when computed using two-letter code. Studies that have compared 27 indices of congruence have had different outcomes depending upon the congruence index used (Camp and Chartrand, 1992). Relations between Holland's SDS and Entrepreneurship According to Holland's (1985a) theory of personality types and work environments, people flourish in their work environments when there is congruence between personality type and the characteristics of the environment. Thus, Enterprising personality is characterized by the ability to manipulate others for personal or organizational goals and Social is characterized by interpersonal competencies. According to Peterson (1985), an entrepreneur is one who identifies and exploits an opportunity and develops a niche in the market to satisfy a need. In a similar vein, Schumpter (cited in Ripsas, 1998) defined an entrepreneur as a risk-bearer, a discoverer of new opportunities, and an innovator. In this regard, it is assumed that enterprising and Social subsets of Holland's R I A S E C should positively correlate with characteristics depicting an entrepreneur. In keeping with this view, a measure of entrepreneurship based on the Holland types was developed to provide an index for establishing the extent to which subsets of the SDS can be differentiated by the measure of entrepreneurial intentions. Theory of entrepreneurship Lack of an integrated theory and definition of entrepreneurship in the literature was noted as shown by divergent views on the criteria and classification of an entrepreneur. Theories on entrepreneurship include the fol lowing: 1. Psychosocial development theories according to Weber (cited in Cunningham & Lischeron, 1991) assert that one's upbringing and experience facilitate acquisition of 28 strong achievement motivation. Role models are said to inculcate a business culture, more especially, i f parents owned a business. 2. A theory of achievement motivation (McClel land, 1961) asserts that the strength of achievement-related motives influence behavior. Motive, as it refers to one's disposition to strive for anticipated benefit, is seen as a determinant of persistence in an activity. Hypothesized that the achievement motive is the mainspring of entrepreneurial activity. A n individual's level of aspirations and persistence forms the hallmark of purposeful behavior. 3. Fielder's theory o f leadership effectiveness (cited in Cunningham & Lischeron, 1991), provides a theoretical framework for understanding leadership effectiveness as an important attribute for success in an organization, more especially an entrepreneurial venture. Visionary and action-oriented style of leadership is core to business success. This includes dynamic interaction, consistency, imagination, inspiration, insight, and foresight. 4. From a theory of reasoned action (Ajzen & Fishbein, 1980), three key attitudes that predict intentions are attitude toward the act, social norms, and perceived behavioral control. In line with Bandura's theory of self-efficacy, perceived behavior control refers to the perception that one can execute the target behavior. In line with this, "Entrepreneur's ideas and intentions form the initial strategic template of new organizations and are important underpinnings of new venture development" (Bird, 1988, p.442). Entrepreneur ia l Aspirat ions Scott and Twomey (1988) carried out a survey to identify factors that influence entrepreneurial aspirations among students. Data from a sample of 436 undergraduates from the United States, United Kingdom, and the Republic of Ireland accessed from 29 previous studies were used in the analysis. Factors identified as instrumental to an individual's desire for a business start-up included the following: • Predisposing factors, which, through one's upbringing, experience, personality and perceptions, lead to acquiring strong achievement motivation. • Triggering factors - these are situational and include the effects of looking for work and the prospect of unemployment. • Possessing a business idea - provides the drive to fulf i l l a cherished desire. Characterist ics of an Entrepreneur McCle l land (1987) investigated the characteristics of successful entrepreneurs in Malawi , India, and Ecuador with a view to developing a means of selecting promising entrepreneurs and examining the extent to which they showed the personal characteristics uniquely associated with entrepreneurship. They employed a specially designed technique called the 'Behavioral Event Interview (BEI). Competencies found to be characteristic of successful entrepreneurs included the following: • Proactive personality - unconstrained by situational forces and able to effect environmental change, opportunity-seeking and problem-solving behaviors, and aggressive posturing relative to competitors. • Achievement oriented, demonstrating skills of organization and time management. • Competitive, innovative and skilled at offsetting risks rather than allowing it to be an impediment to progress. • Initiative, foresight and resourcefulness. 3 0 Measures of Entrepreneurship Bateman and Grant (1993) developed a proactive personality measure with the intent of using it to identify individuals with entrepreneurial characteristics. Specifically, their focus was on the measurement and correlates of proactive behavior as a personal disposition. A n initial 47-item scale was developed and subjected to a series of validation tests. Data were collected from three samples of university students. A 17-item unidimensional scale was developed from the initial 47 items. The "B ig Five" (neuroticism, extraversion, openness, agreeableness, and conscientiousness) measure was used to establish the psychometric characteristics of the proactive scale and its relationships with personality subscales of the B ig Five. Coefficient alphas of .89, .87, and .87 were obtained for samples 1, 2, and 3 respectively. The proactive scale was found to be positively correlated with the following personality constructs of the B ig Five: consciousness (r = A3,p < .001) and extraversion (r = .25,p < .01). In a related study, Grant (1996) investigated the extent to which a proactive personality was related to entrepreneurial intentions. O f particular interest was the examination of the degree of congruence between the proactive personality scale and entrepreneurial intentions. Entrepreneurial intentions were defined as one's judgements about the likelihood of owning a business (Grant, 1996). Proactive personality was defined as "one who is relatively unconstrained by situational forces and who effects environmental change" (Bateman & Grant, 1993, p. 43). Data were collected from a sample of 181 university students. A 17-item proactive personality scale (Bateman & Grant, 1993) was used. Included in the scale were such items as, "I excel at identifying opportunities" and "No matter what the odds, i f 31 I believe in something I wi l l make it happen." The results showed positive correlation between entrepreneurial intentions and proactive personality scale (r = .48, p< .01). The author concluded that proactive personality, gender, education, and parental role models impacted on entrepreneurial intentions. . Chapter Summary Although Holland's theory of vocational interest has generated considerable research, no consensus has yet emerged on the generalizability and universality of the hexagonal model across cultures. There has been empirical support on the generalizability of the model among the Anglo-American groups (Glidden-Tracey & Parraga, 1996; Tracey & Rounds, 1993, 1996). Research indicated some degree of variability with conflicting outcomes and perceptions concerning the hexagonal model's cross-cultural invariance. Additionally, there were flaws in the methodological approaches of most studies. These included (a) sample sizes, where some studies had sample sizes as small as 21; (b) use of inappropriate and widely varied methods to test Holland's hexagonal structure; and (c) visual examination of the hexagonal fit (Chartrand ,1991; Rounds & Tracey, 1996). The chapter provided some important approaches to the analysis of Holland's hexagonal model across cultures. That is, multidimensional scaling (MDS) , Hubert and Arable's (1987) randomized test of hypothesized order relations, and cluster analysis dominated the studies on the hexagonal structure. Studies that investigated the validity of the SDS used correlational, factor, and discriminant analysis. Fouad and Dancer (1992) argued that knowledge of the structural equivalence across cultures was critical in determining whether Holland's (1985a) theoretical 32 framework generalized beyond the United States. Although moderate support for the circumplex assumption was reported, doubts have been raised regarding the adequacy of the circumplex model in non-U. S. Cultures (Glidden-Tracey & Parraga, 1996; Tracey & Rounds, 1993, 1996). It has also been found that the R I A S E C arrangement was not always reproduced across cultural groups, gender, and minority (Fouad & Dancer, 1992; Glidden-Tracey & Parraga, 1996; Tracey & Rounds, 1993, 1996). Thus, the cultural invariance of the model remains an open question. Subsequently, literature on construct validity of Holland's SDS demonstrated that construct validity of an instrument in one culture should not be taken as a passport for use in another culture (Berry, 1989). In addition, empirical evidence showed that SDS measurement properties have not been comprehensively assessed in cross-cultural settings, more especially, in Afr ica. Accordingly, empirical information on cross-cultural validity of the SDS and the extent of its utility in the Afr ican cultural setting has not been substantiated. Knight (1997) argued that it was naive to assume a scale developed in the United States would function with equal efficacy across cultures. Studies on person-environment congruence have provided conflicting and ambiguous conclusion. Schwartz (1991) warned that person-environment congruence might be confounded with different factors related to different personality types. Over and above this, varying methods used to calculate congruence and analyze data made it difficult to draw meaningful conclusions. Camp and Chartrand (1992) concluded that congruence indices used to measure person-environment fit did not all measure the construct equally wel l , and that some indices did not incorporate Holland's circumplex assumption. Cumulatively, the conclusions from the 33 studies that investigated person-environment fit across cultures have added support to the growing body of evidence that challenges some of Holland's congruence conceptions. Sti l l inconclusively resolved, were questions on the validity of the SDS cross-culturally. A s posited by Messick (1989): (a) Is SDS a good measure of the characteristics it is interpreted to assess? This raises issues of psychometric properties in terms of construct validity and (b) should the measure be used for the proposed purpose? This is an ethical question, and raises issues of potential consequences of the testing in terms of social values. Over and above this, an extensive review of the literature on Holland's hexagonal model and entrepreneurship revealed a lack of empirical evidence on the relation between subscales of Holland's SDS and a measure of entrepreneurship. This proposition, it appears, remains unanswered. Literature in the area of entrepreneurship identified four traits as characteristics common to individuals with entrepreneurial intentions. These characteristics are, proactive personality, risk-taking, need for achievement, and locus of control (Bateman & Grant, 1993; McCle l land, 1987; Scott & Twomey, 1988). 34 C H A P T E R 3 M E T H O D O L O G Y This chapter discusses the approaches used to examine the construct validity of Holland's SDS and the methodology used to develop the entrepreneurial scale. The relation between two of the SDS subscales (social and enterprising) and a newly developed measure of Entrepreneurial Intentions (EI) in a Zimbabwean cultural setting is discussed. Secondly, a description of the target population from which a representative sample was drawn and the methods used to obtain the sample are discussed. This is followed by procedures employed in data collection. Finally, data analytic techniques used in the study are discussed. Measures Two measures were used in this study. These were Holland's (1994) SDS, which is a published measure, and Entrepreneurial Intentions (EI), a new measure, which was developed for this study. Publ ished Measure SDS is a self-administered instrument specifically designed to measure a person's resemblance to each of the six personality types and work environments measured through six subscales, (i.e., Realistic, Investigative, Artistic, Social, Enterprising, and Conventional). Each scale consists of four components: (a) Activities, (e.g., I can fix mechanical things); (b) Competencies, (e.g., I can change a car's oi l or tire); (c) Occupational Preferences, (e.g., Auto Mechanic), and (d) Self-estimate of Abil i t ies, (e.g., (a) Mechanical Abi l i ty, (b) Manual Skil ls). The Activities and Competencies domains contain a total of 66 items each, 11 per trait. The Preferences domain contains a total of 3 5 154 items, 14 per trait. Responses to items for Activities, Competencies, and Preferences domains are dichotomously scored, that is, " l i ke" vs "dis l ike" and "yes" vs "no." The Self-estimate domain (1 and 2) is scored on a seven-point Likert-type scale wi th l = low and 7 = high. The SDS instrument has a total of 228 items (Appendix B). Developed Measure The EI scale, a self-report measure, has four subscales: (a) Proactive Personality (e.g., I usually make swift decisions to attain my goals) (12 items); (b) Risk-Taking (e.g., I don't like taking chances) (13items); (c) Need for Achievement (NAch) (e.g., I am good at influencing the course of events) (13 items); and (d) Locus of Control (LoC) (e.g., Achiev i ng my set objectives depends on how hard I work) (16 items). This makes a total of 54 items (Appendix B). The items are scored on a six-point Likert-type scale with 1 = strongly disagree and 6 = strongly agree. Two types of reliability analysis, namely internal consistency reliability and test-retest reliability using a two-week interval between tests were performed on the entrepreneurship scale (Crocker & Alg ina, 1986). Details about the development of the scale are presented later. Procedure Introductory letters were sent to principals of the three out of ten institutions that participated in the study. Participants were identified through the colleges' records office. Four hundred (400) participants identified through a stratified random sampling procedure from the five academic fields (business management, education, liberal arts, natural sciences, and engineering) were individually requested for their willingness to participate in the study prior to administering the SDS (228 items) and Entrepreneurial Intentions (54 items) measures. This procedure took into account the heterogeneous 36 nature of the sample. As echoed by Leedy (1997), "population consists of definite strata, each of which is distinctly different, but the units within the stratum are as homogeneous as possible" (p.214). Correspondences to principals and participants are included in Appendix A . Scale Specification Entrepreneurial Intentions instrument was developed to identify the dimensional structure o f SDS. A s already alluded to, literature on entrepreneurship and Holland's SDS were examined in order to explore the nomological network associated with the construct of entrepreneurial intentions (Cronbach & Meehl , 1955). Holland's theory hypothesizes six vocational /occupational types: Realistic, Investigative, Artistic, Social, Enterprising, and Conventional (R IASEC) and that there exists a natural match between these interest types and corresponding work environments with the same label. Consequently, subsets of Holland's SDS (social and enterprising) were expected to correspond meaningfully with the measure of entrepreneurship. Development of the Entrepreneurial Intentions (EI) Questionnaire I designed the EI questionnaire with the intent of (a) identifying individuals with entrepreneurial characteristics and (b) examining the relationship between SDS subscales (social and enterprising) and EI subscales. The first step based on Holland's (1985a) theory of personality type and work environments, involved a search of the current literature in the area of entrepreneurship. This search was aimed at exploring the key elements that motivate an individual to start up a new venture. EI as a personality trait describes an individual's latent behavior towards an enterprising culture. Thus, consistent with literature (Bateman & Grant, 1993; Bird, 1988; Grant, 1996; McCle l land, 1987; 37 Scott & Twomey, 1988; Ward, 1994;), four traits identified as characteristics common to successful individuals are: • Proactive Personality - proactive behavior is a personal disposition that impels individuals to behave proactively towards their situation. A proactive person is one who is relatively unconstrained by situational forces and effects environmental change (Bateman & Grant, 1993). Proactive people scan for opportunities, show initiative, take action, and persevere until they reach closure by bringing about change. • Risk Taking - includes the notion of assuming the risk of a business with regards to consequence of uncontrolled change (Ripsas, 1998). Risk taking involves seasoned judgement that allows decisions to be made in a timely way. • Need for Achievement (NAch) - is characterized by personal responsibility, setting goals for oneself, and a strong desire for feedback (McClel land, 1965). A person with high need for achievement is wil l ing to work toward distant goals and responds positively to competition (Jackson, 1989). • Locus of Control (LoC) - concerns a person's expectancy for reinforcement. A n individual with an internal control orientation believes that reinforcement is contingent on his /her own behavior whereas an individual with an external control orientation believes that reinforcement is contingent on luck, chance, or powerful others (Rotter, Chance & Phares, 1972). The next step in the development of the above subscales entailed describing in detail the content of each domain. For example, Risk-taking included subdomains such as (a) ability to manage risk situations, (b) ability to make appropriate decisions necessary for 38 goal attainment, and (c) pursuance and discovery of new opportunities irrespective of resources. Having thus identified the subdomains, a table of specification was drawn up to guide in item development and to ensure that the entrepreneurship subscales adequately sampled the relevant areas. The deductive/theory-based approach to item development was adopted as opposed to the inductive or data driven approach. Base on literature (e.g., Gable & Wolf, 1993), 28 items were initially written to assess for sub domains of enterpreneurship. Next, a Q-sort method was used to classify the items into the four sub domains. This resulted in a total of 20 items (5 items for each subscale). Benson and Hocevar (1985) argued that negatively and positively worded items should be balanced on personality, attitude, and other rating scales in order to disrupt possible response sets. This approach was adopted during the construction of items. The items were formatted on a 6-point Likert-type scale with 1 being (strongly disagree) and 6 being (strongly agree). Nunnally (1978) noted that the reliability of the scale should increase as a function of the number of steps employed. For example, a 5-point to a 7-point scale was considered most reliable. Pretest Pretest involved the administration of 20 items to 150 participants representative of those for whom the study was designed. Gable (1986) stated that estimation of the number to be involved should at least be 6 times the number of items. Preliminary analyses involved computing item-total scale correlation for each subscale. A minimal item-total scale correlation of .3 for item inclusion was adopted for each subscale. Low reliabilities of .40 (proactive personality), .18 (risk taking), .19 (need for achievement), and .05 (locus of control) were obtained using Cronbach's (1951) coefficient alpha (r a). This could have been, partly, due to inadequate sampling of items from the domain. 39 Kerlinger (1986) posited that i f an instrument is not reliable enough, add more items of equal kind and quality. In light of this, additional items parallel to the first set in each subscale were developed. The Spearman-Brown formula was used as a guide in coming up with additional items needed for each subscale using the formula: Where A^= the total number of parallel items combined, Pxx, = the 'stepped-up reliability, intended reliability which, in this study, was set at alpha .7 (Al len & Yen, 1979). The Spearman-Brown formula expresses the reliability of a test, Pxx, in terms of the reliability of parallel components of the test (Al len & Yen 1979; Crocker & Alg ina, 1986). Prr = The reliability of the shorter item version. Thus, based on this approach, Locus of Control subscale, for example, would have needed 45 additional items. Notwithstanding this, a decision in terms of the need for developing quality items tapping the construct of interest was also taken into account. This resulted in the following number of items per subscale being considered adequate enough to tap the construct of interest for each subscale: proactive personality scale 12 items, risk taking and need for achievement scales 13 items each, and locus of control 16 items, making a total of 54 items (Appendix B). Item Writing and Assessment of Parallelness Further review of literature provided rationale for the development of additional items. Thus, items were based on attributes pertinent to one's inclination towards entrepreneurship. Gable and Wol f (1993) explained that items that are added to an existing scale to enhance reliability should clearly parallel the best items on the existing 40 scale. Thus, items tapping one's internal locus of control included such items as, "I feel that success in a given venture is due to one's ability" or "Achieving my set objectives depends on how hard I work." As elaborated by Lefcourt (1976), such items involve cognitive processing, autonomy, self-confidence, and resistance to influence attempts. In the same vein, the format of the items and the instrument were evaluated and revised to meet the requirements of clarity, conciseness, parallelness, and ease of administration. Expert judgement was sought to evaluate the items with respect to the extent to which the items described the construct and met technical criteria with regards to appearance of bias, level of readability, and so forth (see Appendix B). This was achieved by having the items vetted by: • Head of the Department of Business Education - Harare Polytechnic in Zimbabwe • Aff inity Group Tours - Dr. Robert Gallacher • Graduate students in the Faculty of Education (ECPS, C U S T ) • Faculty of Commerce (Entrepreneurship Section). Their comments and suggestions resulted in further revision of items to ensure clarity, comprehension, and suitability with regards to the construct was achieved. After this process, the EI questionnaire (54 items) was ready for field-testing. Reliability Nunnally (1978) reiterated that since reliability is an important issue in the use of any measurement method, investigations of reliability should be made when new measures are being developed. Consequently, two types of reliability analyses were performed on the final version of the entrepreneurship scale. Namely, internal 41 consistency reliability using Cronbach (1951) coefficient alpha (ra) and test-retest reliability coefficient. Validity Assessment Several types of validity analyses were performed on the entrepreneurship scale. To examine the concurrent-related validity, the relationship between the EI scale and Holland's SDS was computed. In concurrent validity, the new measure is correlated with an established measure of that trait. Participants The target population from which a representative sample was drawn for f ield-testing was students of 10 technical and vocational colleges in Zimbabwe. The sample was drawn from three colleges (Bulawayo Polytechnic, Masvingo Technical College and Belvedere Technical Teachers' College) considered representative of the target population. The three colleges project a national outlook in that their student population is drawn from throughout the country. In addition, they offer courses common to all the ten colleges. Concerning the generalizability of the findings, Ga l l , Borg, and Gal l (1996) elaborated that results are generalizable to the accessible population, then to the target population especially when comparative information is taken into account. Rounds and Tracey (1993) questioned the logic of vocational interest measures using homogeneous samples in construct validity studies. Consistent with this, the six R I A S E C occupational types formed the basis for determining the appropriateness of the sample. The sample was, thus, drawn from five major academic fields: education, business, liberal arts, natural sciences, and engineering. 42 With regards to sample size, Leedy (1997) suggested that sample size depends largely on the degree to which the sample approximates the qualities and characteristics of the target population. Gay (1996) suggested that when selecting the sample size, "the larger the population size, the smaller the percentage of the population needed to get a representative sample" (p. 125). That is, for a population beyond 5,000, a sample size o f 400 should be adequate. MacCal lum, Browne, and Sugawara (1996) echoed similar sentiments with regards to power. In keeping with this and cognizant of the fact that, from literature review a median sample size of 390 was noted, a sample of 400 participants was considered representative enough of the target population for the current study. Data Analysis Confirmatory factor analysis (CFA) was the procedure used to investigate the cross-cultural validity of Holland's (1994) SDS and the extent to which subsets of the SDS (Enterprising and Social) were differentiated by the measure of EI. This approach to data analysis facilitated the examination of the covariation among a set of observed variables in order to gather information on their underlying constructs (Byrne, 1998). The measurement model (CFA) specifies the relations of the observed measures to their posited underlying constructs, with the constructs allowed to intercorrelate freely (Anderson & Gerbing, 1988) The measurement model is expressed algebraically as: x=AJ+S y = Ayrj + s (2) 43 where x and y are observed variables, E, and r/ are latent factors, and 8 and s error of measurement. The factor loadings A.s provide information about the extent to which the observed variable is able to measure the latent variable (Schumacker & Lomax, 1996). Hypothesized Models Holland's theory of personality types and work environment hypothesizes six occupational themes. Based on this theoretical model, the study aimed at testing the hypothesis that: 1. The hexagonal model (R IASEC) was invariant across cultures. 2. Two SDS subscales wi l l relate strongly to subscales of a newly developed measure of entrepreneurial intentions (EI). Based on the first model (Figure 4), confirmatory factor analysis is computed for the whole group to test the assumption that SDS is culture invariant. In addition, an exploratory analysis by college majors (i.e., business and education majors combined and engineering and science majors combined) wi l l be carried out to investigate i f the pattern from the whole group analysis generalizes to each of the specific groups. The model has 6 latent variables (Realistic, Investigative, Artistic, Social, Enterprising, and Conventional) measured by 30 observed variables. Ben- Porath, Berry, Buss and Royce, Irvine and Carrol, and Prediger (cited in Tracey, Wanatabe, & Schneider, 1997), stated that before any measure can be applied across cultures, its construct validity must be established in each culture. Hence, cross-cultural validity of Holland's model is investigated with the whole group and by college majors because patterns might be different from the whole group analysis. This is consistent with Farh, Leong, and Law 's (1998) study on cross-cultural validity of Holland's model in Hongkong where analysis 44 by college majors was conducted to test i f Holland's model holds for the whole group as well as by subgroups. Holland's circumplex model (hexagonal ordering) predicts that correlations for the six adjacent types (R l , IA, A S , SE, E C , and CR) should be greater than correlations among alternate types (RA, A E , E R , IS, SC , and CI) which, in turn, should be greater than correlations among opposite types (RS, IE, and AS) . This assumption is tested based on the whole group and by college majors. This procedure helps to establish i f there is a pattern in the R I A S E C ordering from the whole group analysis. Glidden-Tracey and Parraga (1996) employed a similar approach in assessing the structure of R I A S E C in terms of the hexoganal ordering in Pakistan. The second hypothesized model (Figure 5), tests the assumption that there wi l l be a strong relationship between two SDS subscales (Social and Enterprising) and EI subscales (Proactive Personality, Risk-Taking, Need for Achievement, and Locus of Control). Twenty-eight (28) observed variables measure the six latent variables. Confirmatory factor analysis is computed for the whole group. The following methodological issues were taken into account prior to using the measurement models: • Sample Size - Effects of sample size on goodness-of-fit indices has been extensively addressed in S E M literature (Bollen, 1990; Byrne, 1998). For example, the chi-square goodness of fit test is a direct function of sample size where % is affected by large sample size. Byrne (1998) purported that the analysis of covariance structures is grounded in large samples. In line with this, a sample of 378 for the current study was considered large enough for S E M analysis. 45 • Model Identification - concerns whether a single unique value for each and every free parameter can be obtained from the observed data (Hoyle, 1996; Schumacker & Lomax, 1996). Accordingly, a model can be (a) under-identified i f one or more parameters cannot be uniquely determined, (b) just-identified i f there is just enough information for the parameters to be estimated, and (c) overidentified when there is more than one way of estimating a parameter. The underlying premise, therefore, is to determine whether constraints placed on the model are sufficient to determine unique estimates of the structural coefficients. Consequently, model identification depends on the specification of parameters as fixed, free, or constrained (Joreskog & Sorbom, 1993; Schumacker & Lomax, 1996). Model identification can be determined through the formula: \{p + qh> + q + i)-t (3) where p + q is the number of observed variables analyzed and t is the total number of independent parameters estimated. • Model Estimation - The aim of S E M is to minimize the difference between Z (implied covariance matrix) and S (sample covariance matrix). Model estimation is the process of obtaining an implied covariance matrix E that is as close as possible to the observed covariance matrix S. Amos program offers several choices of methods of estimation. These include, generalized least squares (GLS) , unweighted least squares (ULS) , and maximum likelihood (ML) . Empirical evidence shows that M L is by far the most frequently used because the estimates produced by this method are robust. Anderson and Gerbing (1988) noted that, under the assumption of multivariate normal distribution of the observed variables, M L estimators have the desirable 4 6 asymptotic or large sample properties of being unbiased, consistent, and efficient. Hence, M L estimation was used in this study. Goodness-of-Fit Indices Goodness-of-fit is determined simultaneously with estimation. The extent to which the hypothesized models adequately describe the sample data was estimated by means of the maximum-likelihood (ML) parameter estimation (Joreskog & Sorbom, 1993). One component of a "good" model, as explained by Schumacker and Lomax (1996), is the fit between the sample covariance matrix - S and the estimated population covariance matrix - £ . Overall, goodness-of-fit indices provide the basis for making decisions about the fit of a model or the relative fits of different models. However, empirical evidence on model fit indices point to the fact that this area is in a state of flux (Loehlin, 1992). Joreskog (1993) advised that (a) for test of exact fit, researchers should use chi-square and reject the model i f /?-value is less than 0.05 and (b) for test of close fit, to use root mean square error of approximation ( R M S E A ) and accept the model i f R M S E A is < 0.05. Based on the literature (e.g., Byrne, 1998; Joreskog, 1993; Schumacker & Lomax, 1996), the chi-square goodness-of-fit index (% ), Goodness-of-fit index (GFI), Adjusted Goodness of Fit Index (AGFI) , Root Mean Square residual (RMR) , and Root Mean Square Error Adjusted ( R M S E A ) were chosen for use in this study. Table 1 shows the fit indices and the associated critical values. 47 Table 1 Goodness-of-Fit (GOF) Criteria and Acceptable Fit Interpretation Goodness-fit-indices Acceptable level Interpretation/Comments Chi -square Tabled % 2 value (p > a) Compares obtained % 2 value with tabled value for given df. A significant %2 value relative to the df indicates that the observed and estimated matrices differ. Chi-square is sensitive to sample size. GFI 0 (no fit) to 1 (perfect Measures the amount of variance/covariance in 1 - y2 t r ( S - I ) fit) S that is predicted by the reproduced matrix _. Value of .90 and above reflects a good fit. AGFI 0 (no fit) to 1 (perfect Adjusts for the number of degrees of freedom in 1 -fit) the model with .90 and above considered a good model fit. RMR <.05 Indicates the closeness of E to S matrix. Values of .05 or less are desirable. R M S E A <.05 Value less than .05 indicates a good fit. Source: Shumacker and Lomax (1996). Summary This chapter outlined the steps that were used in investigating cross-cultural validity of Holland's SDS and the procedures taken in constructing and validating the Entrepreneurship scale used in identifying the dimensional structure of SDS and the construct validity of the SDS subsets. Two measurement models were hypothesized a priori based on Holland's theory of personality type and work environments. 48 Confirmatory factor analysis was used to establish the construct validity of Holland's SDS. The research design employed the measurement model to test the theory-based causal assumption among the latent traits. Confirmatory factor analysis made it possible for one to examine the relationships between a set of observed variables and the underlying latent traits. Model fit was determined through the use of the chi-square statistic, Goodness of fit measures, root mean square residual ( R M R ) , and root mean square error of approximation ( R M S E A ) . Figure 4. Measurement Model for SDS 50 (dh-Msoc5 soc3 @~-+\soc2 sod @P~~+\entr1 @r*\ entr§ @r*\ entr8 @h^\ entr7 tfT&-+\ intr6 pr1 pr2 \+~® pr3 K@ pr5 rsk6\+® *7 \c8 nhc9}^~~^ nac1l\+—4l) nac1^~~4l^ nacl3r*^£iy Figure 5. Measurement Model: SDS and El 51 C H A P T E R 4 R E S U L T S In this chapter, item analysis results for the entrepreneurial intentions (EI) questionnaire are reported. The results of the reliability and validity tests for the EI are presented including results of mean differences between Holland's (1994) college sample (used as the normative sample) and the Zimbabwean college sample. Data analysis on construct validity of Holland's (1994) SDS across cultures is presented. Included is the examination of the relation between SDS subscales (social and enterprising) and a newly developed measure of entrepreneurial intentions. The results of data analysis through confirmatory factor analysis (CFA) are given. Sample Size and Miss ing Data In total, 400 national diploma students representing five academic majors were administered the SDS along with the EI scale. O f the 400 who participated, 388 (97%) returned the questionnaires. Data from 10 of the returned questionnaires for SDS and 20 for EI were unusable and, thus, excluded from the analysis. The final sample was 378 (94.5%) for SDS and 368 (92%) for EI. The sample consisted of 207 (55%) males and 171 (45%) females ranging in age from 18 years to 40 years with a mean age of 23.59 years. The sample was carefully selected to provide a broad cross-section of occupational environments as defined by Holland (1985a) and to enable generalizability to the target population. The response rate by sex and college major is shown in Table 2. Due to missing data for individual items, the total number of cases varied within each analysis (see Appendix C). Listwise deletion was applied to exclude cases with missing values causing the sample size to fluctuate from analysis to analysis. 52 Table 2 Response Rates by Sex and College Major College Major Male % Female % Total Teacher Education 31 53.4 27 46.6 58 Business Education 78 50.3 77 49.7 155 Liberal Arts 18 29.51 43 70.49 61 Natural Science 34 63.38 18 34.62 52 Engineering 44 89.80 5 10.20 49 375 (3 cases unclassified) Preliminary Analysis Item-total scale correlation coefficients were computed for each of the EI subscales. Tables DI - D4 (Appendix D) show the corrected item-total correlation for each subscale. Gable and Wo l f (1993) posited that item-total correlation represents the correlation of the particular item with the other items on the scale. Good affective measures, they pointed out, have average inter-item correlations in the .30 to .40 range, therefore, items correlating less than .2 should be targeted for review. They further elaborate that, deleting items with low item/scale correlations enhances the alpha level. In this study, item total-scale correlations ranged from .03 to .4. Using a minimum item-total scale correlation of .3 for item inclusion (Gable & Wolf, 1993), seven items for proactive personality scale with rit values below .3 were removed (Table DI ) . Similarly, seven items for need for achievement (Table D2), eleven items for locus of control (Table D3), and eleven items for risk taking (Table D4), were excluded from the analysis. Reliability Analysis Coefficient alpha reliabilities were computed using Cronbach's coefficient alpha to test the robustness of the EI subscales. The subscale reliabilities were moderate, 53 ranging from r a .40 to .65. Table 3 lists the coefficient alpha reliability ra , mean inter-item correlation rix, and item total scale correlations r\t for the EI and SDS subscales. Table 3 Internal Consistency Reliability Estimates: Entrepreneurial Intentions (EI) Questionnaire n Number Of Items ra ru ru Range EI Subscale Proactive Personality 332 5 .61 .24 .33- .41 Risk Taking 314 2 .40 .26 .30- .31 Need for Achievement 336 6 .65 .24 .33- .45 Locus of Control 362 5 .62 .26 .33-.42 SDS Subscales Realistic 354 38 .78 .42 .31- .72 Investigative 352 38 .76 .44 .39- .70 Artistic 348 38 .82 .50 .52- .74 Social 354 38 .72 .38 .38- .65 Enterprising 356 38 .72 .41 .52- .72 Conventional 357 38 .79 .52 .56- .71 Note. ra = Reliability coefficient, ru = Mean inter-item correlation, rjt = Item-total scale correlation T e s t - R e t e s t A sub sample of 186 drawn from the initial sample was administered the test-retest after a two-week interval. Test-retest reliability yields a reliability estimate, rxx' = Pxx' (Crocker & Algina, 1986). The test-retest reliability coefficients are shown in Table 4. These indicate a moderate level of stability. Crocker and Alg ina (1986) reiterated that few, i f any, standards exist forjudging the minimum acceptable value for a test-retest reliability estimate. They further pointed to the fact that test-retest reliabilities for personality interest measures are often low compared to those for aptitude tests. 54 Table 4 Test-retest Correlation Coefficients for EI Subscales Proact 1 Risk 1 NAch 1 LoC 1 Mean ~SD Sample (N= 368 for Pre-Test; N=\S6 for Test-Retest) Proact 2 .59 24.10 3.59 Risk 2 .31 9.78 1.87 NAch 2 . .54 27.25 4.56 LoC 2 . . .54 23.65 4.11 Mean 23.61 9.34 26.25 23.27 SD 3.91 2.20 5.10 4.80 Note. 1 = first administration (e.g. Proact 1) and 2 = Test-retest (e.g. Proact. 2). Validity Assessment Messick (1995) posited that validity is not a property of the test, but rather of the meaning of the test scores. One validates not a test, but an interpretation of data arising from a specific procedure. In a similar vein and based on a theoretical rationale, evidence of the validity of EI and SDS was assessed in a variety of ways. Content-Related Validity Content validity with regards to content relevance, representativeness, and sampling adequacy of the content was ensured through the preparation of a table of specifications of the domain of EI (Appendix F). Review of current literature on entrepreneurship provided information about the nature and scope of the content domain of the EI construct. A s mentioned earlier, content-related validity focuses on the content of the test items and their relation to the intended domain. In light of this, an initial draft was developed and items were critically reviewed for their relevance and edited before 55 being pilot-tested. Review of items by experts referred to in greater detail in Chapter 3, facilitated in the development of additional items of equal kind and quality. Based on an a priori assumption that items should correlate at least .3 with the scale total, items with rA less than .3 were removed from the final scale. Cr i ter ion-Related Val id i ty Gable and Wo l f posited that criterion-related validity addresses the question about the relationship between scores on the measurement and scores on the instrument based on some external criterion. Accordingly, to examine the concurrent-related validity, the relationship between EI and the criterion (SDS) was analyzed with respect to the theory underlying the EI instrument and the criterion (SDS). The statistic employed was the correlation coefficient. As shown in Table 5, the EI subscales demonstrated weak correlations with the two SDS subscales (social and enterprising). According to Holland's theory of personality type and work environments, the EI subscales should have a strong relationship with social and enterprising personality types and work environments. Gable and Wo l f (1993) reiterated that with a correlation of .90 one could confidently conclude the existence of criterion-related validity. 56 Table 5 Intercorrelations Between EI Subscales and Criterion-Related validity Measures Proact Risk Nach L o C Realist Invest Artisti Social Enterp Conve S a m p l e ( - 7 V = 3 6 8 ^ E I Subscales Proact .40 .62 .37 .10 .11 .13 . .22 .26 .09 Risk .44 .21 .06 .08 .03 .05 .14 . .04 Nach - .33 .07 .08 .16 .22 .34 .17 L o C ~ .00 .02 .04 .06 .13 .07 SDS ubscales Realist .59 .20 .21 .10 -.01 Invest -- .08 .31 .19 .18 Artisti ~ .42 .32 .05 Social - .53 .32 Enterp .59 Conve Construct-Related Validity Construct-related validity addresses the question: "To what extent do certain explanatory concepts (constructs) explain covariation in the responses to the items on the instrument (Gable and Wolf, 1993)?" Evidence of construct-related validity was gathered by administering the SDS and EI instruments to a representative sample for which the instruments were designed. Analysis techniques used to establish construct validity included correlations of EI subscales with SDS subscales (Table 5) and confirmatory factor analysis. In addition, mean comparisons between Holland's 1994 college sample 5 7 and the Zimbabwean college sample were computed to establish i f there were differences between means of these two groups. The results are shown in Tables 6 and 7. Analysis of M e a n Differences To answer the question, is there a difference between the means of the two groups and i f so, is the difference a "real" difference or has it arisen by chance? First, the standard error of the differences between the means was computed to establish the dispersion of the differences. This was calculated using the formula: Where SEM ( and SEM are the standard errors of the normative sample and the observed sample respectively (Kerlinger, 1986). To evaluate the extent of the differences, the t ratio was computed using the formula: The measured differences between MA (SDS) and MB, expressed in standard error unit, showed large differences between the two samples. Kerlinger (1986) posited that the important question to ask is, "how far away from the hypothesized mean of zero must a deviation be to be significant" (p. 210)? With regards to this study and using a simple rule o f thumb (Kerlinger (1986), a critical ratio of 1.98 should be considered significant about the .05 level. A s shown in Table 7, the deviations away from a hypothesized mean of zero ranged from .58 to -12.45. (4) 58 Table 6 Comparison of Means and Standard Deviations between Holland's (1994) College Sample and the Study's College Sample by Gender College Students: Holland (1994 Females Males (« = 475) (n = 344) Scale M Activities Zimbabwe's College Students Females Males (n= 171) (n = 2071 SD M SD M SD M SD R 2.55 2.67 4.93 3.45 3.20 2.87 5.61 3.07 I 3.71 3.09 4.28 3.28 5.10 3.49 6.69 3.35 A 5.52 3.03 4.38 3.17 5.33 3.29 5.42 3.42 S 7.19 2.62 5.25 3.08 6.78 2.79 7.00 2.87 E 6.04 2.97 6.65 2.83 8.72 2.20 9.14 1.88 C 4.15 3.50 3.31 3.34 7.73 3.29 7.91 3.27 Competencies R 3.78 2.83 6.57 3.15 2.23 2.16 5.21 3.06 I 5.42 3.18 6.29 3.15 4.81 2.73 6.59 2.71 A 4.79 2.64 4.02 2.69 5.57 2.56 5.33 3.01 S 8.39 2.49 7.29 2.90 7.93 2.62 8.13 2.84 E 6.18 2.91 6.61 2.87 6.92 2.92 7.64 2.62 C 6.83 2.98 5.24 3.02 6.60 3.13 6.86 2.71 R 1.20 2.09 3.36 3.58 3.70 3.47 6.50 3.92 I 2.97 3.33 3.47 3.52 4.48 4.44 6.19 4.77 A 4.23 3.69 3.76 3.78 5.49 4.19 6.21 4.89 S 6.81 4.09 4.54 4.04 5.48 4.27 6.11 4.72 E 4.39 3.69 4.82 3.66 7.71 4.01 9.04 4.11 C 2.63 3.73 2.21 3.31 7.84 5.13 8.38 5.30 R 14.17 8.32 24.31 10.60 16.47 8.40 26.30 10.28 I 19.77 9.95 22.56 10.73 21.63 10.12 28.40 11.30 A 22.04 10.60 19.21 11.02 23.52 10.80 23.96 12.78 S 33.54 9.37 27.43 10.33 30.89 9.25 31.49 10.44 E 24.98 10.19 27.11 9.87 33.04 9.37 35.29 9.29 C 22.64 10.65 18.22 9.73 32.10 12.83 32.26 12.14 Holland's (1994) results show group means and standard deviation by gender for Activit ies, Competencies, Occupations and Summary. Means and standard deviations for Self-Estimates (1 and 2) are not given separately but are included in the Summary results. This approach was adopted for comparability purposes. 59 Table 7 Mean Differences between Normative Sample and Observed Sample SEM1 SEM2 Mean SEM t SEMj SEM2 Mean SEM t Females Diff. Males Diff Activities R .12 .22 -.65 .25 -2 .6 . .19 .21 -.68 .28 -2.43 I .14 .27 -1.39 .30 -4.6 .18 .23 -2.41 .29 -8.31 A .14 .25 .19 .29 .66 .17 .24 -1.04 .29 -3.59 S .12 .21 .14 .24 .58 .17 .20 -1.75 .26 -6.73 E .14 .17 -2.68 .22 -12.18 .15 .13 -2.49 .20 -12.45 C .16 .25 -3.58 .30 -11.93 .18 .23 -4.60 .29 -15.86 Competencies R .13 .17 1.55 .21 7.38 .17 .21 1.36 .27 5.03 I .15 .21 .61 .26 2.35 .17 .19 -.30 .25 -1.20 A .12 .20 -.78 .23 -3.39 .15 .21 -1.31 .26 -5.04 S .11 .20 .46 .23 2.00 .16 .20 -.84 .26 -3.23 E .13 .22 -.74 .26 -2.85 .15 .18 -1.03 .23 -4.48 c .14 .24 .23 .28 .82 .16 .19 -1.62 .25 -6.48 Occupations R .10 .27 -2.50 .29 -8.62 .19 .27 -3.14 .33 -9.52 I .15 .34 -1.51 .37 -4.08 .19 .33 -2.72 .38 -7.16 A .17 .32 -1.26 .36 -3.5 .20 .34 -2.45 .39 -6.28 S .19 .33 1.30 .38 3.42 .22 .33 -1.57 .40 -3.93 E .17 .31 -3.32 .35 -9.49 .20 .29 -4.22 .35 -12.06 c .17 .39 -5.21 .43 -12.12 .18 .37 -6.17 .41 -15.05 Summary R .38 .64 -2.30 .74 -3.11 .57 .71 -1.99 .91 -2.19 I .46 .77 -1.86 .90 -2.07 .58 .79 -5.84 .98 -5.96 A .49 .83 -1.48 .96 -1.54 .59 .89 -4.75 1.07 -4.44 S .43 .71 2.65 .83 3.19 .56 .73 -4.06 .92 -4.41 E .47 .72 -8.06 .86 -9.37 .53 .65 -8.18 .84 -9.74 c .49 .98 -9.46 1.10 -8.60 .52 .84 -14.04 .99 -14.18 Note. SEM= Standard Error of the Means; t > 2.0 are significant at/? < .05 6 0 Conf i rmatory Factor Analysis In trying to establish the cross-cultural validity of the SDS, confirmatory factor analysis ( C F A ) was computed for (a) the whole group and (b) by college majors. O f primary importance in structural equation modeling is the extent to which a hypothesized model (£) "fits" the sample data (S) (Byrne, 1998). The goodness-of-fit indices as well as the individual parameters were examined. Identification of parameters to be estimated was consistent with the underlying theory. The theoretical construct underlying the SDS postulates 6 latent factors (Realistic, Investigative, Artistic, Social, Enterprising, and Convention). A hexagon defines the relationships among the R I A S E C types. Holland's theory postulates that the intercorrelations among the R I A S E C types are based on their proximity on the hexagon. Thus, correlations for the adjacent types (R l , IA, A S , S E , E C , and CR) are greater than correlations of alternate types (RA, IS, A E , SC , and CI) which, in turn, are greater than correlations of the opposite types (RS, IE, and AS) . Similarly, confirmatory factor analysis was computed for the whole group to test the assumption that there wi l l be a strong relationship between two SDS subscales (Enterprising and Social) and EI subscales (Proactive Personality, Risk-Taking, Need for Achievement, and Locus of Control). The goodness-of-fit indices as well as the individual parameters were examined. Amos Version 3.6 was the statistical package used for the confirmatory factor analysis. The SDS measurement model had 30 observed variables and 6 latent variables and the measurement model for the two SDS subscales and EI subscales had 28 observed variables and 6 latent traits. 61 Whole Group Sample Results (Figure 6) show support of the R I A S E C ordering on four of the six adjacent types. Low correlations were obtained on two of the adjacent types. Thus, data of the Zimbabwean sample seem to support a four-factor ordering as opposed to a six-factor ordering. R I C Figure 6. Interrelationships among the 6 SDS Latent Variables for the Whole Group Parameter estimates as shown in Table 8 were within the admissible range of t-value > 2.0 at the .05 alpha level (Byrne, 1998). The squared multiple correlations (R2), serves as reliability index and represent the proportion of variance in each item accounted for by the respective latent variable. This information is useful in identifying particular items fitting less well into the factor solution (Gable & Wolf, 1993). For example items 1 and 5 under factor 1 show squared multiple correlations of R2 = .72 and R2 - .11 respectively. 6 2 Table 8 T-values for Parameter Estimates from Whole Group Sample Path Estimate SE /-value R (unstandardized) Lambda (X) 1,1 1.00 - - .72 2,1 .68 .06 11.49 .38 3,1 1.04 .07 14.31 .54 4,1 .53 .04 13.75 .51 5,1 .22 .04 5.78 .11 6,2 1.00 - - .64 7,2 .70 .06 12.18 .46 8,2 1.11 .09 11.89 .44 9,2 .47 .04 12.61 .49 10,2 .34 .04 8.66 .25 11,3 1.00 - - .67 12,3 .89 .06 14.50 .58 13,3 1.34 .09 15.23 .64 14,3 .46 .04 10.65 .34 15,3 .46 .04 10.89 .35 16,4 1.00 - - .56 17,4 .90 .08 11.27 .48 18,4 1.53 .13 11.62 .51 19,4 .42 .05 8.37 .26 20,4 .25 .03 7.39 .20 21,5 1.00 - - .41 22,5 1.55 .16 10.02 .47 23,5 2.23 .23 9.86 .49 24,5 .88 .10 9.09 .36 25,5 .76 .08 9.37 .39 26,6 1.00 - - .54 27,6 .90 .08 11.19 .43 28,6 1.83 .15 12.50 .53 29,6 .74 .06 12.49 .53 30,6 .66 .05 12.87 .57 Note, r-values > 2.0 are significant at p < .05; SE = Standard error of estimates. A chi-square - ^ 2 ( 3 9 0 , # = 339) = 1782.45, p< .001 was obtained. Results of the other goodness-of-fit indices were, GFI = .69, A G F I = .63, R M R = 1.09, and R M S E A = .10, an indication that the hypothesized model was not adequately supported by the 63 sample data. The results showed a poor model fit given that acceptable fit levels for the chi-square fit index should be p > .05 and R M S E A should be < .05 GFI and A G F I indices should be at least .90 and above (Byrne, 1998; Shumacker & Lomax, 1996). Figure 7 shows the factor loadings for each of the items. The residuals or theta delta for each observed variable represent the proportion of error variance in each of the items, that is, that part of the total item variance which is not explained by the factor solution (Schumacker & Lomax, 1996). Figure 7. Measurement Model: Whole Group 65 Business and Education Majors and Engineering and Science Majors As shown in Figures 8 and 9, results by college majors showed a similar pattern to the results for the whole group with regards to the R I A S E C ordering. However, results by college majors showed a stronger than expected relationship between some of the alternate and opposite types, a situation not reflected in the whole group results. Figure 8. Interrelationships among the 6 SDS Latent Variables for Business and Education Majors). R E S Figure 9. Intercorrelations among the 6 SDS Latent Variables for Engineering and Science Majors 66 Goodness-of-fit results for Business and Education majors were, chi-square ^ 2 (390,TV = 195) = 1150.31,p < .001; GFI = .67; A D G F = .61; R M R = 1.29; and R M S E A = .11. The results show a poor model fit. Goodness-of-fit results for Engineering and Science majors were, chi-square X2(390,N = 91) = 945.99,p < .001 indicating a poor model fit. Similarly, GFI = .58, A G F I = .50, R M R = 1.39, and R M S E A = .13 showed a large discrepancy between the implied model _ and sample data. Joreskog (1993) stated that a good model fit should have a %2 with p > .05 and R M S E A = < .05. The implied model did not adequately fit the sample data for the business and education majors and engineering and science majors. Table 9 shows summary of results. Table 9 Summary of Results of the Goodness-of-Fit Indices Hypothesized Measurement x2 Models N p value GFI AGFI RMR RMSEA Comments SDS Model • Whole Group 339 1782.45 (p<.001) .69 .63 1.09 .10 Poor Model Fit • Business and Education 195 1150.31 .67 .61 1.29 .11 Poor Model Fit Majors (p<.001) • Engineering. & Science Majors 91 945.99 (p<.001 .58 .50 1.39 .13 Poor Model Fit SDS & EI Model • SDS & EI Subscales 290 783.25 (p<.00\ .84 .81 .41 .07 Mediocre Model Fit 67 Parameter estimates by college majors are shown in Tables 10 and 11. R2 serves as reliability index shows the amount of variance for each observed variable explained by its latent variable. The measurement models (Figure 11 and 12) show standardized factor loadings and the residuals associated with each observed variable. Table 10 T-values for Parameter Estimates of Model from the Education and Business majors Path Estimate (unstandardized) SE /-value R2 Lambda (k) U 1.00 .71 2,1 .585 .08 7.48 .30 3,1 1.097 .10 11.35 .67 4,1 .418 .05 8.91 .40 5,1 .22 .05 4.39 .11 6,2 1.00 - - .19 7,2 1.37 .17 8.22 .48 8,2 2.34 .29 7.98 .48 9,2 .94 .11 8.28 .49 10,2 1.00 - _ .41 11,3 1.00 - - .66 12,3 .71 .07 9.62 .48 13,3 1.34 .12 10.99 .62 14,3 .47 .06 8.25 .37 15,3 .40 .06 6.95 .27 16,4 1.00 - - .06 17,4 2.45 .39 6.37 .47 18,4 4.23 .71 5.99 .37 19,4 1.24 .25 4.93 .21 20,4 1.00 - .33 21,5 1.00 - - .33 22,5 1.72 .27 6.31 .38 23,5 2.45 .41 5.98 .32 24,5 1.37 .20 6.81 .49 25,5 1.02 .15 6.81 .49 26,6 1.00 - - .44 27,6 .99 .13 7.77 .46 28,6 1.54 .21 7.30 .39 29,6 .70 .09 7.99 .49 30,6 .62 .07 8.27 .54 Note, /-values > 2.0 are significant atp < .05; SE = Standard error of estimates Table 11 T-values for Parameter Estimates of Model from the Engineering and Science Majors Path Estimate (unstardized) SE J-value R1 Lambda (A,) 1,1 1.00 .79 2,1 .82 .12 6.88 .47 3,1 1.02 .15 6.74 .45 4,1 .57 .08 3.50 .52 5,1 .29 .08 3.50 .14 6,2 1.00 - - .19 7,2 1.44 .29 5.02 .38 8,2 2.45 .48 4.65 .32 9,2 1.11 .19 5.72 .53 10,2 1.00 - - .48 11,3 1.00 - - .56 12,3 .94 .13 7.29 .67 13,3 1.42 .20 6.97 .60 14,3 .44 .09 5.02 .31 15,3 .62 .10 6.40 .51 16,4 1.00 - - .13 17,4 2,66 .43 6.26 .61 18,4 2.88 .54 5.30 .41 19,4 1.01 .22 4.52 .29 20,4 1.00 - .43 21,5 1.00 - - .38 22,5 1.95 .32 6.09 .72 23,5 1.84 .38 4.86 .38 24,5 .78 .17 4.65 .34 25,5 .78 .15 5.12 .43 26,6 1.00 - - .44 27,6 .64 .18 3.57 .20 28,6 1.63 .33 4.95 .44 29,6 .61 .13 4.75 .39 30,6 .61 .12 5.10 .48 Note, ^-values > 2.0 are significant atp < .05; SE = Standard error of estimates Figure 10. Measurement Model: Bus & Educ. Majors re 11. Measurement Model: Eng. & Sc. Majors 71 Analysis of SDS and E I Subscales Confirmatory factor analysis was used to test the relationship between SDS subscales (Social and Enterprising) and EI subscales for the whole group. A s shown in Table 12, parameter estimates were generally within the admissible range of /-value > 2.0 at the .05 alpha level (Byrne, 1998). Table 12 T-value Estimates of Target Model (SDS subscales and EI subscales) Path Estimate (unstandardized SE /-value R1 Lambda (A,) 1,1 1.00 .14 2,1 1.97 .33 5.94 .41 3,1 .93 .21 4.55 .13 4,1 1.54 .27 5.62 .30 5,1 1.38 .26 5.33 .23 6,2 1.00 - - .45 7,2 .68 .12 5.69 .29 8,3 1.00 - - .17 9,3 1.38 .22 6.40 .47 10,3 .93 .17 5.63 .25 11,3 1.02 .20 5.00 .17 12,3 .96 .17 5.55 .24 13,3 .81 .16 4.99 .16 14,4 1.00 - - .30 15,4 1.42 .25 5.61 .25 16,4 1.23 .26 4.75 .15 17,4 1.36 .22 6.11 .36 18,4 1.68 .31 5.48 .23 1,1 1.00 - - .31 2,1 1.36 .20 6.77 .27 3,1 3.68 .53 6.96 .29 4,1 3.30 .38 8.59 .65 5,1 2.512 .34 7.37 .34 6,2 1.00 - - .43 7,2 1.19 .13 9.37 .43 8,2 2.12 .27 7.87 . .29 9,2 2.24 .21 10.86 .69 10,2 1.00 .13 7.81 .28 Note, /-values > 2.0 are significant at/? < .05; SE = Standard error of estimates 72 Intercorrelations among the latent variables are shown in Figure 12 and Table 13. Locus of control exhibited low correlations with the two SDS subscales (social and enterprising). Out of range correlations were obtained between the two Entrepreneurial Intentions subscales - Proactive Personality and Need for Achievement r = 1.08. This could be due to possible nesting of the two variables. A chi-square ^ 2 ( 3 3 5 , N = 290)= 7 8 3 . 2 5 , < .001 indicated inadequate fit to sample data. Results of other goodness-of-fit indices were, GFI = .84, A D G F I = .81, R M R = .41, and R M S E A = .07. MacCal lum, Browne, and Sugawara (1996) stated that R M S E A < .05 indicate a good fit and that values as high as .08 represent reasonable errors of approximation in the population. Hence, consistent with MacCal lum et al. the model indicated mediocre fit. Lambda coefficients or factor loadings show the strength of relationship (Figure 12) between the factor and the observed variable. 73 Figure 12. Measurement Model: SDS and El 74 Correlations among the latent variables are shown in bold in Figure 12. Because of possible difficulties in identifying the correlation corresponding to the latent variables, these correlations are shown in Table 13. These results should not be confused with the criterion-related results in Table 5. Table 13 Intercorrelations among SDS subscales and EI Subscales TV =290 Proact Risk N A c h L o C Social Entrepr E I Subscales SDS Subscales Proact Risk N A c h L o C Social .76 1.08 .57 .42 .71 .27 .61 .46 .37 .09 Enterpr .52 .24 .50 .06 .67 75 Summary This chapter provides the results of the cross cultural validity of Hol land's SDS in a Zimbabwean cultural setting. In addition, the Chapter provides results on the extent to which SDS subscales (social and enterprising) are differentiated by a newly developed measure of entrepreneurial intentions. To begin with, sample demographics and characteristics are reported. Preliminary analysis of the EI scale involved computing item-total scale correlation coefficient. Using a minimum item total-scale (/*,.,) correlation of .3 for item inclusion, items with (rit) values < .3 were excluded. Reliabil ity for the EI scale was established through internal reliability estimates and test-retest stability analysis. Moderate results were obtained. Intercorrelations between EI subscales and criterion-related validity measures (SDS subscales) showed low to medium results. Similarly, the process of establishing the cross cultural validity of SDS and relationship of the SDS subsets with EI included correlations of EI with SDS, comparison of means and standard deviations between Holland's 1994 college sample and the Zimbabwean college sample. The results showed differences between the groups. The analysis generally showed that data of the Zimbabwean college sample do not support Holland's assumption that SDS is culture invariant. Chi-square results showed poor model fit across the hypothesized model. However, R I A S E C ordering was reproduced on four of the six adjacent types. 76 C H A P T E R F I V E D I S C U S S I O N A N D C O N C L U S I O N S This chapter begins with a summary and discussion of the results in context of past research. Implications of Holland's SDS in an Afr ican cultural setting are discussed including the relation between SDS subsets (social and enterprising) and EI measure. Finally, limitations of the study and directions for future research are discussed Integrative Summary of Results and Discussion The purpose of the current study was twofold: 1. To investigate the cross-cultural validity of Holland's SDS in Zimbabwe. 2. To examine the extent of relationship between SDS subscales (social and enterprising) and subscales of a newly developed measure of Entrepreneurial Intentions (EI). Two models were hypothesized a priori based on Holland's theory. Confirmatory factor analysis was computed for the whole group to test the assumption that SDS was culture invariant. A n exploratory study of two subgroups was also conducted to observe i f the pattern would generalize across sets of college majors. Six latent factors: Realistic, Investigative, Artistic, Social, Enterprising, and Conventional measured by 30 observed variables were postulated by Holland's (1985a) theory. Schumacker and Lomax (1996) posited that one component of a "good" model is the fit between the sample covariance matrix - S and the estimated population covariance matrix - _. Thus, the primary interest in structural equation modeling is the extent to which a hypothesized model adequately describes the sample data (Byrne, 1998). In this regard, examination of model fit included assessment of: 77 (a) Parameter estimates, that is, their feasibility in terms of correct sign, size, and consistency with the underlying theory; appropriateness of standard errors (e.g., not too large and not close to zero); and statistical significance - the /-statistic which operates as a z-statistic in testing that the estimate is statistically different from zero. (b) Measurement model in terms of the extent to which it is adequately represented by the observed measures. R reported for each observed variable serve as reliability indicator representing the proportion of the variance in each item accounted for by its respective latent variable. (c) Model as a whole through goodness-of-fit statistics Generally, the results on all the models showed a poor model fit, although all parameters exhibited the correct sign and were within the admissible range of /-value > 2.0 at .05 alpha level These results suggest that the hypothesized model was not adequately supported by the sample data. Similar results were obtained on sub-sample models by college majors. The evidence are that the Zimbabwean data does not support the assumption that SDS is culture invariant. The results are consistent with Farh, Leong, and Law's (1998) findings, where, in their study on cross-cultural validity of Holland's model in Hong Kong, they obtained a chi-square - ^ f 2 ( l2 ,N = 1674) = 736.51,/? < .01 with all model fit indices below the acceptable range and R M S E A = .19. Tracey and Rounds (1993, 1996), in a meta-analytic review of 72 studies, concluded that the psychometric properties and validity of the SDS order structure, when applied outside the United States and across diverse cultures, remains questionable. A s echoed by Tracey, Wanatabe, and Schneider (1997), these different structures could exist at either item level or at the scale level. Methodological issues such as 78 equivalence of measurement (e.g. do occupations serve the same function?) and conceptual equivalence (e.g. do different cultures attach the same meaning to a similar concept?) are quite evident in the results of the Zimbabwean sample. Examination of the hexagonal ordering for the whole group, contrasted with the American normative sample reported in the Vocational Preference Inventory (VPI) manual, showed notable exceptions to the American norms where the correlation between Investigative and Artistic from data of the Zimbabwean whole group sample was r = .06, as opposed to r = .43. Four of the intercorrelations (i.e., Realistic and Investigative; Artistic and Social; Social and Enterprising; and Enterprising and Conventional) of the six Holland types were comparable for the American and Zimbabwean sample (Figure 6). In addition, through visual examination of the pattern of correlations, results of the business and education majors showed a correlation between the opposite types Realistic and Social (r = .40) which was higher than in the whole group. A similar pattern was found among engineering and science majors for the correlation between Social and Realistic (r = .49). Both of these correlations were larger than those in the whole-group results. These findings are consistent with those of Farh et al's (1998) where they found low corrections between investigative and artistic and between realistic and conventional and higher-than-expected correlations between Realistic and Social with data of the Hong Kong sample. These researchers' findings raised questions about the cross-cultural invariance of Holland's SDS. Where the present research findings diverge from the findings of these previous researchers, however, is that correlations of the adjacent types are not only low but negative, for example, correlation between Realistic and Conventional i s r = - . 1 5 , r = - .14 (Figures 6 and 9 respectively). 7 9 Holland (1994) postulated that a hexagonal model could estimate the degree of congruence (agreement) between a person and an occupation (environment). Farh, Leong, and Law (1998) argued that, from a cross-cultural perspective, Holland's theory and its associated measuring instruments were indigenous to the United States, thus, generalizability of Holland's theory to a non-U.S. cultural context depends on the similarity of that culture to the United States. Leong (1997) argued for the adoption of etic (external value) and emic (culture-specific variables) approaches i f Holland's model was to generalize across cultures. In a similar vein, Leong (1997), addressing issues of cultural validity and equivalence of measurement, talked about the problem of conceptual equivalence in terms of whether two cultures attach the same meaning to a similar concept, or even have the same concepts. Farh et al. (1998) found social career to be the most popular among the science majors, accounting for 41%. Closer examination revealed that science majors who wanted to pursue social careers did not do so because of their high social interest, but because of their concern for future jobs. I would like to assume that this might also be one possible explanation for this particular finding with the Zimbabwean sample. In addition, items such as, manual skills, understanding of others, and work outdoors that are contained in the SDS questionnaire wi l l not discriminate between groups because culturally, these activities are a part of one's upbringing. Over and above this, Holland's assumption that people look for jobs or pursue vocations congruent to their personality type may not hold in the Zimbabwean setting where career opportunities are limited and unemployment is high. It is therefore not 80 surprising that people look for jobs or vocations that have better prospects and opportunities rather than those that match their personality type. The second hypothesized model tested the assumption that there wi l l be a strong relationship between two SDS subscales (Enterprising and Social) and EI subscales (Proactive Personality, Risk-Taking, Need for Achievement, and Locus of Control). Testing of the relationship between SDS subscales and EI subscales (Figure 5) was through confirmatory factor analysis. The model had 28 observed variables and 6 latent variables. First, various statistical analysis approaches were used to establish the reliability and validity of inferences on the new measure (EI). Preliminary analysis involved computing item-total scale correlation for each subscale. Items with total-scale correlation < .3 were dropped. Gable and Wo l f (1993) pointed out that deleting items with low item-scale correlations enhances the alpha level. Similarly, Bateman and Grant (1993) initially generated 47 items for their 17-item proactive personality scale and excluded those items that did not tap the construct of interest. Thus, 18 items out of the initial 54 items were used in the analysis. Internal reliability estimates using Cronbach alpha and test-retest results, after a two-week interval, were moderate. Reliabil ity was also assessed through confirmatory factor analysis. In model testing, R2 indicates the reliability index of each observed variable to its latent variable. The reliability indices ranged from .13 to .47. Moderate reliability coefficients could be attributable to an incidence of error in the sampling of items from the domain, variability in individual responses over time, and possible existence of situational factors influencing the responses to items (Gable & Wolf, 1993). Interestingly, of the 365 participants with valid 81 information on the demographics on "work experience", only 10 (2.6%) had some work experience, 355 (93.7%) had never worked. Similarly, of the 362 participants with valid information on "parent owned businesses", only 34 (9%) had parents operating business, 328 (86.5%) did not. Thus, I am tempted to speculate that this situation might have had some effects on the response pattern. Notwithstanding this, empirical evidence also shows that affective measures have generally lower reliability levels than do cognitive measures, because cognitive skills tend to be more consistent and stable than most affective characteristics (Gable & Wolf, 1993). In addition, the reliability estimates and test-retest coefficients were, generally, consistent with some of the published scales. Faver (1982) reported reliability coefficients of r a = .46 for need for achievement scale. Bateman and Grant (1993) reported coefficient alpha ra = .87, r = .29, and test-retest reliability coefficient r = .72 for their proactive personality scale. Levison (1974) reported coefficient alpha ra - .82 for his locus of control scale. Similarly, Holland (1971) reported internal consistency coefficients (KR-20) of .53 to .85, test-retest reliability coefficients of .31 to .87) for the SDS using a 3-week time interval. He acknowledged these reliabilities as moderate. Holland (1994) noted that "a comparison of the internal consistency coefficients suggests that the 1994 edition appears more reliable than the 1985 edition, which, in turn, appears slightly more reliable than the 1977 edition. Hence, the item analysis over time may have refined the scales and subscales of the SDS to a point at which all scales and subscales of the SDS have moderate to high (.70 or above) internal consistency" (p. 23). Over and above this, the inclusion of negatively worded items might have affected item equivalence. Benson and Hocevar (1985), Marsh (1987) posited that 82 negatively and positively worded items should be balanced, to disrupt possible response sets. Contrary to this view, Holland (1994) pointed Out that, unlike positively worded items, negatively worded items have problems of interpretation. Further review o f the literature revealed that negation of an item may affect item equivalence, in that participants tend to respond inconsistently to a mixed format. In addition, Clark, Chase, and Wason, in their study (cited in Gable & Wolfe, 1993), noted that what is at stake is whether the negative and positive items are measuring the same construct. Content-related validity was ensured through the preparation of a table of specifications of the domain of EI (Appendix F). A review of literature on entrepreneurship provided information about the nature and scope of the domain of EI construct (Grant, 1996; McClel land, 1987). Expert judgement was sought to review the items with respect to the extent to which the items described the intended latent trait and met technical criteria in terms of level of readability and appearance of bias. Construct validity was assessed through correlations, comparison of means between Holland's 1994-college sample and the Zimbabwean College sample, and confirmatory factor analysis. Intercorrelations among EI subscales and criterion-related validity measures (SDS subscales) were low to moderate ranging from r = .05 to r = .32. Risk Taking and Locus of Control had low correlations with the two SDS subscales. Low to moderate reliabilities of the EI subscales might have had an effect. Leong (1997) noted that most career psychology theories have tended to emphasize person variables and ignore contextual and environment variables, that is, contextual variables involving the constructs of individualism and collectivism as a major value orientation guiding human behavior. 83 Triandis (1989) echoed similar sentiments by stating that individualists - "independent se l f - give priority to personal goals over gOals of collectives. Collectivists "interdependent se l f - on the other hand, tend to subject their personal goals to collective goals. Thus, the possibility of the participants' responses being confounded by such cultural variables cannot be ruled out. Comparison of means by gender between Holland's 1994-college sample (used as the normative group) and the Zimbabwean College sample was made to establish whether the means of these two groups were similar or different. The results showed differences ranging from .58 to -12.45 between means of the two female groups and -1.20 to -15.86 between means of the two male groups. Kerlinger (1986) noted that t> 2 is significant at the .05 alpha level. The measurement model (Figure 5) was used to examine the relationship between SDS subscales (Social and Enterprising) and EI subscales. This study is the first to explore directly the relation between SDS subsets (social and enterprising) with a measure of entrepreneurship. Moderate intercorrelations were obtained between EI subscales and SDS subscales (Table 11). Locus of control subscale showed weak correlation r = .06 and r = .09 with enterprising and social respectively. Parameter estimates were within the accepted range and R for the 28 observed variables ranged from .13 to .69. Chi-square ^ 2 ( 3 3 5 , # = 290) = 783.25,p < .001 was obtained. Results of other fit indices were, GFI = .84, A D G F I = .81, and R M S E A = .07. Consistent with MacCal lum, Browne, and Sugawara (1996), the model indicated a mediocre fit. Out of range values were obtained between proactive personality scale and need for achievement scale. Results exhibited unreasonable correlation of r = 1.08. A critical 84 reexamination of the data file failed to detect any unusual patterns. Further review of literature on structural equation modeling pointed to the difficulty in detecting errors of this nature due to the combined complexities associated with both the data and the specified model. Byrne (1998) explained that such problems could be linked to parameter, values at which the iteration process begins. Limitations of this Study The intent of this study was to investigate the cross-cultural validity of Holland's SDS and to establish the relation between SDS subscales (social and enterprising) and a newly developed measure of EI. A sample of 400 college students was considered large enough and representative enough to the target population. Unfortunately, although this number was initially obtained, sample size fluctuations due missing data lowered the sample to about 339. The analysis of covariance structures, as purported by Byrne (1998), are grounded in a large sample theory and, as such, large samples are critical to the obtaining of precise parameter estimates, as well as to the tenability of asymptotic distributional approximations. The second limitation of the study is related to the timing in which data were collected. Data were collected in mid February; two weeks after the students had entered college fresh from high school. In addition, student unrest occurred during the time data were being collected (1999, February 26, Zimbabwe Mirror; 1999, March 3, Chronicle-Zimbabwe). In light of this, career instability might have confounded the true picture. Contribution and Conclusions The results from data of the Zimbabwean sample failed to support the construct validity of Holland's (1994) SDS. These findings are likely to advance knowledge about 85 the cross-cultural invariance of Holland's model more especially in the Zimbabwean cultural setting where it has not yet been empirically demonstrated that the SDS has utility for describing the occupational interests of vocational and technical students. As posited by Leong and Brown (1995), it is only through the testing of the cultural validity of existing western-centered theories and models that knowledge of the boundary conditions of these models can be increased. This study is the first to explore directly the relation between SDS subscales and a newly developed measure of entrepreneurial intentions. Career issues pertinent to Zimbabwe include unemployment, economic factors, career barriers, and school to work transition. Hence, the development and validation of a measure of entrepreneurship is expected to contribute to knowledge of important theoretical issues applied within the education domain and to provide information with potential impact on vocational and technical education in Zimbabwe. Stead and Watson (1998) noted that "while it can be stated with reasonable certainty that most career measures in international use are also used in South Afr ica, it is equally true that there is a paucity of research that has assessed the relevance and validity of such usage" (p. 294). Given the results of the current study, it appears accommodation of the cultural value dimension may be needed i f Holland's model is to provide a better assessment of occupational interests and work environments in Zimbabwe. In light of this, future research should focus at modifying Holland's theory in the context of the Zimbabwean cultural setting. In addition, this research agenda could be extended to include a more diverse sample from the one used in this study. This would provide a valuable means of assessing whether the SDS could be adapted to suit the Zimbabwean cultural setting. 86 Stead and Watson (1998) argued that simply testing more blacks does not necessarily address the cultural context and economic realities that exist within the country. Leong (1997) stated that limitations could be accounted for by culture-specific considerations. Cultural specificity is concerned with concepts, constructs, and models that are specific to certain cultural groups and that play a role in explaining and predicting behavior. Similarly, Afr ican culture, as posited by Stead and Watson (1998), values the oral approach. Hence, future research should consider using a mixture of quantitative and qualitative approaches e.g. interviews. Such an approach could provide context rich research. 87 References Ajzen, I., & Fishbein, M . (1980). A theory of reasoned action. Englewood Cl i f fs, N J : Prentice-Hall. A l len, M . J . , & Yen, W. M . (1979). Introduction to measurement theory. Monterey, California: Brooks/Cole. Anderson, J . C , & Gerbing, D. W. (1988). Structural equation modeling in practice: A review and recommended two-step approach. Psychological Bulletin, 103 (3), 411-423. Athanasau, J . A . , & Cooksey, R. W. (1993). Self-estimates of vocational interests. Australian Psychologist. 28 (2). 118-127. Bagozz, R. P. (1993). Assessing construct validity in personality research: Applications to measures of self-esteem. Journal of Research in Personality. 27 (1), 49-87. Bateman, J . S., & Grant, J . M . (1993). The proactive component o f organizational behavior. Journal of Organizational Behavior, 14 (2), 103-118. Benson, J . , & Hocevar, D. (1985). The impact of item phrasing on the validity of attitude scales for elementary children. Journal of Educational Measurement, 22 (3), 231-240. ' Bentler, P. M . , & Bonetti, D .G. (1980). Significance tests and goodness of fit in the analysis of covariance structures. Psychological Bulletin. 88, 588-606. Bernstein, I. H., Garbin, C. P., & Teng, G. K. (1988). Appl ied multivariate analysis. N Y : Springer-Verlag. Berry, J . W. (1989). Imposed etics-emics-derived etics: The operationalization of a compelling idea. International Journal of Psychology, 24, 721 -73 5. Bird, B. (1988). Implementing entrepreneurial ideas: The case for intention. Academy of Management Review, 13 (3), 442-453. Bol len, K. A . (1990). Overall fit in covariance structure models: Two types of sample size effects. Psychological Bulletin, 107, 256-259. Bol len, K. A . , & Long, J . S. (Eds.), (1993). Testing structural equation models. Newbury Park, C A : Sage. Borgen F. H. (1986). New approaches to the assessment of interests. In W. B. Walsh & S. H. Osipow (Eds.), Advances in vocational psychology: V o l 1. The assessment o f interests (pp. 31-54). Hil lside, N J : Erlbaum. Boyle, G. J . , & Fabris, S. (1992). Confirmatory and congeneric factor analysis of Holland's self-directed search. Personality and Individual Differences 13, 1077-1084. Brown, D., & Brooks, L. (1990). Career choice and development (2 n d ed.). San Francisco: Jossey-Bass. Byrne, B. M . (1995). One application of structural equation modeling from two perspectives: Exploring the E Q S and L I S R E L strategies. In R. H. Hoyle (Ed.), Structural equation modeling: Concepts, issues, and applications (pp. 138-157). Thousand Oaks, C A : Sage. Byrne, B. M . (1998). Structural equation modeling with L I S R E L , P R E L I S , and SIMP LIS: Basic concepts, applications, and programming. Mahwah, New Jersey: Lawrence Erlbaum Associates. Camp, C. C , & Chartrand, J. M . (1992). A comparison and evaluation of interest congruence indices. Journal of Vocational Behavior, 41, 162-182. 89 Campbell, D. T., & Fiske, D. W. (1959). Convergent and discriminant validation by the multitrait-multimethod matrix. Psychological Bulletin, 56, 81-105. Campbell, N . J . (1985). The Self-Directed Search. Test Critiques, 2, 697-706. Chartrand, J . M . (1992). Research and application using Holland's typology: Reactions from a scientist-practitioner perspective. Journal of Vocational Behavior, 40, 194-200. Chartrand, J . M . , & Camp, C. C. (1991). Advances in the measurement of career development constructs: A 20-year review. Journal of Vocational Behavior, 39, 1-39. Crocker, L., & Alg ina, J . (1986). Introduction to classical and modern test theory. N Y : Holt, Rinehart, and Winston. Cronbach, L. J . , & Meehl , P. E. (1955). Construct validity in psychological tests. Psychological Bulletin, 52, 281-302. Cunningham, J . B., & Lischeron, J . (991). Defining entrepreneurship. Journal of Small Business Management, 29, 45-61. Cuttance, P. & Ecob, R. (1987). Structural modeling by example: Applications in educational, sociological, and behavioral research. New York: Cambridge. Dawis, R., & Lofquist, L. (1984). A psychological theory of work adjustment. Minneapolis: University of Minnesota Press. Dumenci, L. (1995). Construct validity of the self-directed search using hierarchical nested structural models. Journal of Vocational Behavior, 47, 21-34 Farh, J . , Leong, F. T. L., & Law, K. S. (1998). Cross-cultural validity of Holland's model in Hong Kong. Journal of Vocational Behavior, 52, 425-440. 9 0 Faver, C. A . (1982). Achievement orientation values, and women's employment. Journal of Vocational Behavior, 20, 67-80. Ferreira, J . A . A . , & Wood, A . B. (1995). The development and validation of a Holland-type Portuguese vocational interest inventory. Journal of Vocational Behavior. 46, 119-130. Fitzgerald, L. F., & Rounds, J . B. (1989). Vocational behavior: A critical analysis. Journal of Vocational Behavior. 35. 105-163. Fouad N . A . , & Dancer, L. S. (1992). Cross-cultural structure of vocational interests. Journal of Vocational Behavior, 40, 129-143. Gable, R. K. (1986). Instrument development in the affective domain. Boston: Kluwer-Nijhoff. Gable, R. K., & Wolf, M . B. (1993). Instrument development in the affective domain: Measuring attitudes and values in corporate and school settings (2 n d ed.). Boston: Kluwer Nijhoff. Ga l l , M . D., Borg, W. R., & Gal l , J . P. (1996). Educational research: A n introduction (6 t h ed.). White Plains, N Y : Longman. Garfield, C. A . (1986). Peak performers: The heroes of American business. In B. Bi rd (Ed.), Implementing entrepreneurial ideas: The case for intention (pp. 451). Academy of Management Review, 13 (3), 442-453. Gati, I. (1979). A hierarchical model for the structure of vocational interests. Journal of Vocational Behavior, 15, 90-106. Gati, I. (1982). Testing models for the structure of vocational interests. Journal of Vocational Behavior, 21. 164-182. Gati , I. (1991). The structure of vocational interests. Psychological Bulletin, 109, 309-324. Gay, L. R. (1996). Educational research: Competencies for analysis and application (6 t h ed.). Upper Saddle River, N J : Prentice Hal l . Glidden-Tracey, C. E., & Parraga, M . I. (1996). Assessing the structure of vocational interests among Bol iv ian university students. Journal of Vocational Behavior, 48 (1), 96-106. Grant, J . M . (1996). The proactive personality scale as a predictor of entrepreneurial intentions. Journal of Small Business Management, 34 (3), 42-49. Hansen, J . C. (1987). Cross-cultural research on vocational interests. Measurement and Evaluation in Counseling and Development, 19, 163-176. Hoelter, J . W . (1983). The analysis of covariance structures: Goodness-of-fit indices. Sociological Methods & Research, 11, 325-344. Hofstede, G. (1980). Culture's consequence: International differences in work-related values. Beverly Hi l ls , C A : Sage. Holland, J . L. (1971). A theory-ridden, computerless, impersonal vocational guidance system. Journal of Vocational Behavior, 1, 167-176. Holland, J . L. (1977). Manual for the Vocational Preference Inventory. Palo Al to, C A : Consulting Psychologist Press. Holland, J . L. (1985a). Making vocational choices: A theory of vocational personalities and work environments (2 n d ed.). Englewood Clif fs, N J : Prentice-Hall. Hol land, J . L. (1985b). The self-directed search: Professional manual. Odessa, F L : Psychological Assessment Resources. Holland, J . L. (1987). Some speculation about the investigation of person-environment transactions. Journal of Vocational Behavior, 31, 337-340. Holland, J . L. (1994). Self-directed search: SDS Form R (4 t h ed.). Odessa, F L : Psychological Assessment Resources. Holland, J. L., Fritzsche, B. A . , & Powell , A . B. (1994). The self-directed search manual (1994 ed.). Odessa, F L : Psychological Assessment Resources. Hoyle, R. H. (1996). Structural equation modeling: Concepts, issues, and applications. Thousand Oaks, California: Sage. Hu , L. T., & Bentler, P. M . (1995). Evaluating model fit. In R. H. Hoyle (Ed.), Structural equation modeling: Concepts, issues, and applications (pp. 76-99). Thousand Oaks, C A : Sage. Hubert, L., & Arabie, P. (1987). Evaluating order hypothesis within proximity matrices. Psychological Bulletin. 102. 172-178. Jackson, D. N . (1989). Personality research form manual, (3 r d ed.). Port Huron, M I : Research Psychologists Press, Inc. Joreskog, K. G. & Sorbom, D. (1981). L I S R E L V : Analysis of linear structural relationships by the method of maximum likelihood. In R. H. Hoyle (Ed.), Structural equation modeling: Concepts, issues, and applications (pp. 39). Thousand Oaks, C A : Sage. Joreskog, K. G . (1993). Testing structural equation models. In K. A . Bol len & J . S. Long (Eds.), Testing structural equation models (pp. 294-316). Newsbury Park, C A : Sage. Joreskog, K. G . , & Sorbom, D. (1986). L I S R E L v i : Analysis of linear structure relationships by the method of maximum likelihood- Mooresvil le, IN: Scientific Software. Kerlinger, F. N . (1986). Foundations of behavioral research. (3 r d ed.). N Y : Holt, Rinehart and Winston. Khan, S. B., & A l v i , S. A . (1991). Holland's typology: A study in non-western culture. Journal of cross-cultural Psychology, 22, 283-292. Knight, G . A . (1997). Cross-cultural reliability and validity of a scale to measure f irm entrepreneurial orientation. Journal of Business Venturing, 12, 213-225. Leedy, P. D. (1997). Practical research: Planning and design (6 t h ed.). Upper Saddle River, N J : Prentice-Hall. Lefcourt, H. M . (1976). Locus of control: Current trends in theory and research. New York: John Wiley and Sons. Leong, F. T. (1997). Cross-cultural career psychology. Journal of Counseling Psychology, 44, 355-359. Levenson, H. (1974). Act iv ism and powerful others: Distinction within the concept of internal-external control. Journal of Personality Assessment, 38, 377-383. Loehl in, J . C. (1992). Latent variable models: A n introduction to factor, path, & structural analysis. Hil lside, N J : Lawrence Erlbaum Associates. Long, J . S. (1983). Covariance structure models: A n introduction to L I S R E L . Beverly Hi l ls , C A : Sage. Lowman, R. L., & Wil l iams, R. E. (1987). Validity of self-ratings of abilities and competencies. Journal of Vocational Behavior, 31, 1-13. MacCal lum, R. C , Browne, M . W., & Sugawara, H. M . (1996). Power analysis and determination of sample size for covariance structure modeling. Psychological Methods. 1. 130-149. Marsh, H.W. (1987). The factorial invariance of responses by males and females to a multidimensional self-concept instrument: Substantive and methodological issues. Multivariate Behavioral Research. 22, 457-480. McCle l land, D. (1961). The achieving society. Princeton, N J : VanNostrand. McCle l land, D. C. (1987). Characteristics of successful entrepreneurs. Journal of Creative Behavior. 21 (3). 219-233. Meir , S. (1991). Vocational behavior, 1988-1990: Vocational choice, decision-making, career development interventions, and assessment. Journal of Vocational Behavior. 39. 131-181. Messick, S. J . (1980). Test validity and the ethics of assessment. American Psychologist, 30 (11), pp. 1012-1027. Messick, S. J . (1989). Validity. In L. L inn (Ed.), Educational measurement (3 r d ed., pp. 13-146). Messick, S. J . (1995). Standards of validity and the validity of standards in performance assessment. Educational Measurement: Issues and Practice, 14, 5-8 Mulaik, S. A . , James, L. R., Van Alstine, J . , Bennett, N. , Linda, S., & Sti lwell, C. D. (1989). Evaluation of goodness-of-fit indices for structural equation models. Psychological Bulletin. 105, 430-445. Nunnally, J . C. (1978). Psychometric theory (2 n d ed.). New York: M c Graw-Hi l l . 95 Oosterveld, P. (1994). Confirmatory factor analysis of the self-directed search test, a multitrait-multimethod approach. Personality and Individual Differences, 17 (4), 565-569. Osipow, S. H. (1983). Theories of career development (3 r d ed.). Englewood Cl i f fs, N J : Prentice-Hall. Palmer, M . (1971). The application of psychological testing to entrepreneurial potential. California Management Review, 13 (3), 32-38. Parsons, F. (1909). Choosing a vocation. In W. B. Walsh & S. H. Osipow (Eds.), Career decision making (pp. 2). Hil lside, N J : Lawrence Erlbaum Associates. Peter, J P. (1979). Reliability: A review of psychometric basics and recent marketing practices. Journal of Marketing Research, 16 (1), 6-17. Peterson, R. (1985). Raising risk-takers. Metropolitan Toronto Business Journal, 75 (7), 30-34. Prediger, D. J . (1982). Dimensions underlying Holland's hexagon: Missing link between interests and occupations? Journal of Vocational Behavior, 21, 259-287. Rachman, D., Amernic, J . , & Aranya, N . (1981). A factor-analytic study of the construct validity of Holland's self-directed search. Educational and Psychological Measurement, 41, 425-436. Ripsas, S (1998). Towards an interdisciplinary theory of entrepreneurship. Small Business Economics, 10, 103-115 Rotter, J . B., Chance, J . , & Phares, E. J . (1972). Applications of a social learning theory of personality. New York: Holt, Rinehart, and Winston. Rounds, J . B., Zevon, M . A . (1983). Multidimensional scaling research in vocational psychology. Appl ied Psychological Measurement, 7. 491-510. Rounds, J . , & Tracey, T. J . (1996). Cross-cultural structural equivalence o f R I A S E C models and measures. Journal of Counseling Psychology, 43, 310-329. Schartz, R. H. (1991). Achieving orientation of personality type: A variable to consider in tests of Holland's congruence-achievement and other hypothesis. Journal of Vocational Behavior, 38, 225-235. Schneider, B. (1987). E = f(P, B) : The road to a radical approach to person-environment fit. Journal of Vocational Behavior, 31, 353-361. Schumacker, R. E., & Lomax, R. G. (1996). Beginner's guide to structural equation modeling. Mahwah, N J : Lawrence Erlbaum Associates. Scott, M . G. , & Twomey, D. F. (1988). The long-term supply of entrepreneurs: Students' career aspirations in relation to entrepreneurship. Journal of Small Business Management, (October), 1-5. Spokane, A . R. (1985). A review of research on person-environment congruence in Holland's theory of careers. Journal of Vocational Behavior, 26, 306-343. Stead, G . B., & Watson, M . B. (1998). Career research in South Afr ica: Challenges for the future. Journal of Vocational Behavior, 52, 289-299. Super, D. E. (1990). A life-span, life-space approach to career development. Journal of Vocational Behavior. 36. 282-298 Tokar, D. M . , & Fischer, A . N . (1998). More on R I A S E C and the five-factor model of personality: Direct assessment of Prediger's (1982) and Hogan's (1983) dimensions. Journal of Vocational Behavior, 52, 246-259. Tokar, D. M . , & Swanson, J. L. (1995). Evaluation of correspondence between Holland's vocational personality typology and the five-factor model of personality. Journal of Vocational Behavior, 46, 89-108. Tracey, J . G . , Wanatabe, N. , & Schneider, P. L. (1997). Structural invariance of vocational interests across Japanese and American cultures. Journal of Counseling Psychology, 44, 346-354. Tracey, T. J . (1991). R A N D O R D : A program for evaluating Holland and Gati's order hypothesis using the randomization test of hypothesized order relationship (computer program). University of Illinois, IL: Urbana-Campaign. Tracey, T. J . , & Rounds, J . (1993) Evaluating Holland and Gati's vocational interest models: A structural meta-analysis. Psychological Bulletin, 113, 229-246. Trapnell, P. D. (1989). Structural validity in the measurement of Holland's vocational typology: A measure of Holland's types scaled to an explicit circumplex model. Unpublished master's thesis, University of British Columbia, Vancouver, Brit ish Columbia, Canada. Triandis, H. C. (1989). The self and social behavior in differing cultural contexts. Psychological Review, 96, 506-520. Upperman, P. J . , & Church, A . J . (1995). Investigating Holland's typological theory with army occupational specialties. Journal of Vocational Behavior, 47, 61-75. Wakefield, J . A . , Doughtie, E. B. (1973). The geometry of relationship between Holland's personality typology and the vocational preference inventory. Journal of Counseling Psychology, 20, 513-518. 98 Walsh, W. B., & Holland, J . L. (1992). A theory of personality types and work environments. In W. B. Walsh, K. H. Craik, & R. H. Price (Eds.), Person-environment psychology: Models and perspectives (pp. 35-70). Hil lsdale, N J : Erlbaum. Ward, E. A . (1994). Construct validity of need for achievement and locus of control scales. Educational and Psychological Measurement, 54, 983-992. LIST OF A P P E N D I C E S 100 Appendix A Letters Letter sent to each participant p. 101 Letter sent to Principals of participating colleges p. 103 101 Letter sent to each participant Dear Participant, Cross-cultural Validity of Holland's Self-Directed Search (SDS) in Zimbabwe and the Relation of SDS with a Specially Designed Measure of Entrepreneurial Intentions (El). We are writing to invite you to participate in a research project that is being conducted at your college. The genera! focus of the study is to explore the relationship between your current area of study and your personality traits, competencies, and self-estimates to occupational clusters based on Holland's Self-Directed Search and a measure of entrepreneurship. The study is directed by Dr. Nand Kishor, professor of Educational Psychology and Special Education at the University of British Columbia, in collaboration with Sylvia Masango, a graduate student at UBC. The study will be in partial fulfillment of the requirements of Sylvia's Master of Arts in Measurement, Evaluation, and Research Methodology (MERM). The research is designed to explore the individual's self-understanding in terms of personal development, vocational potential, vocational alternatives considered, and satisfaction with current vocational aspiration. Holland's Self- Directed Search (SDS) which is based on the assumption that people can be categorized as one of six personality types: Realistic, Investigative, Artistic, Social, Enterprising, and Conventional will be used as well as a newly developed measure of entrepreneurship. The six SDS scales estimate a person's resemblance to each of the personality types. SDS has been used extensively in North America and Europe to classify individuals according to their personality traits and vocational preferences. The model has enabled individuals assess the degree of congruence between their personality types and their current vocational pursuits. However, empirical evidence on the cross-cultural validity of the SDS in Africa, more especially in Zimbabwe is lacking. The current study aims at gathering important information that will add to knowledge about occupational preferences, competencies, and abilities. For a study of this nature, it is important to enlist the views and inputs of your students. Your college has been identified as one of the institutions that are instrumental to this study. We kindly request for permission to undertake the study at your institution. We also ask for access to your student records to enable us sample, 103 Letter sent to Principals of participating colleges The Principal, Cross-cultural Validity of Holland's Self-Directed Search (SDS) in Zimbabwe and the Relation of SDS with a Specially Designed Measure of Entrepreneurial Intentions (El). We are writing to invite for the participation of your students in a research project that is being conducted at your college. The study is directed by Dr. Nand Kishor, professor of Educational Psychology and Special Education at the University of British Columbia, in collaboration with Sylvia Masango, a graduate student at UBC. The study is in partial fulfillment of the requirements of Sylvia's Master of Arts in Measurement, Evaluation, and Research Methodology (MERM). The research is designed to explore the individual's self-understanding in terms of personal development, vocational potential, vocational alternatives considered, and satisfaction with current vocational aspiration. Holland's Self- Directed Search (SDS) which is based on the assumption that people can be categorized as one of six personality types: Realistic, Investigative, Artistic, Social, Enterprising, and Conventional will be used as well as a newly developed measure of entrepreneurship. The six SDS scales estimate a person's resemblance to each of the personality types. SDS has been used extensively in North America and Europe to classify individuals according to their personality traits and vocational preferences. The model has enabled individuals assess the degree of congruence between their personality types and their current vocational pursuits. However, empirical evidence on the cross-cultural validity of the SDS in Africa, more especially in Zimbabwe is lacking. The current study aims at gathering important information that will add to knowledge about occupational preferences, competencies, and abilities. For a study of this nature, it is important to enlist the views and inputs of your students. Your college has been identified as one of the institutions that are instrumental to this study. We kindly request for permission to undertake the study at your institution. We also ask for access to your student records to enable us sample, according to the set criteria, participants to the study. It is hoped that information from this study will contribute to knowledge of those skills and competencies that accurately 105 Appendix B Questionnaires Review of Items by Experts p. 106 Entrepreneurial Intentions Questionnaire p. 110 106 To Whom it May Concern Re: Vetting of Items on Entrepreneurial Intentions. The questionnaire on entrepreneurship is intended for students of technical and vocational education - post secondary. The intent is to identify those individuals who have an inclination towards owning a business. Based on the literature, four traits (Proactive Personality, Risk Taking, Need for Achievement, and Locus of Control) were identified as characteristic of an entrepreneur. Thus, the questionnaire is made up of four subscales and the items for each subscale are supposed to tap the respective trait/construct. In this regard, I kindly ask you to assist me by vetting the items for each of the four subscales in terms of: • Clarity • ' Comprehension • Readability • Wording • Whether items for each subscale are parallel • Whether items tap the construct in question In addition, I kindly ask you to make suggestions for improvement. I welcome additions to the items especially the 1 s t two subscales. Note: It would be most appreciated if I can get feedback by Feb. 8 t h 107 Proactive Personality Items Comments/Suggestions 1 I am able to seize opportunities ahead of others. 2 I am good at making swift decisions that are appropriate for goal attainment. 3 I am always searching for new ideas. 4 I have the ability to spot opportunities with chances of success. 5 I feel that in any given situation change is desirable. 6 I tend to take calculated moves in what I do. 7 I like taking the lead in initiating new projects. 8 I am always on the lookout for new opportunities. 9 I feel comfortable with the way things are. 10 I commit myself to new opportunities cautiously. 11 I consider the need to change as failure of one's planning process Suggested additional items 108 Risk Taking Items Comments/Suggestions 1 I avoid costly decisions when planning my future. 2 In decision-making situations, I prefer to explore possible options cautiously. 3 I like going for high-risk projects with chances of high returns. 4 When an opportunity arises, I don't hesitate ploughing in all my resources. 5 1 have a preference for low-risk projects. 6 When I fail in what I set out to achieve, I always start again with a new idea. 7 I prefer activities that give a sense of security. 8 I can withstand opposing views when promoting an idea. 9 I have a powerful urge to compete. 10 I have the ability to withstand setbacks in pursuit of my goals. 11 I generally avoid taking risk. 12 I try to compromise to obtain necessary approvals for my project Suggested Additional Items 109 Need for Achievement Items Comments/Suggestion 1 I am able to effectively utilize resources at my disposal 2 I prefer to plan for what I intend accomplishing ahead. 3 I am good at forging my own idea inspite of opposing views from others. 4 I am a powerful source at influencing events. 5 I consider interpersonal relationships a fundamental resource. 6 I possess a business idea. 7 I find the current economic changes in the country stimulating opportunity recognition. 8 I have the ability to work tirelessly in pursuit of my intended goal. 9 I am more powerful at oral than written communication. 10 I am more concerned with why something will work than why it will not work. 11 I am naturally capable of presenting my idea clearly and to the point. 12 In promoting an idea, I always adopt a salesperson approach. 13 I have the tendency to interrupt before others have completed what they have to say. 14 If I believe in an idea, I make it happen. 15 I prefer situations where I am able to control the outcome. 110 Locus of Control Items Comments/Suggestions 1 I feel that success in a given venture is due to chance. 2 Achieving my set objectives depends on how hard I work. 3 My achievement in life is determined by my own actions. 4 No matter how hard I work, luck is not on my side. 5 If I lose in a competition, I contribute it to unfair play. 6 When I fail to perform well in a test it is because I didn't work hard enough. 7 I believe that my success in life is dependent on those in authority. 8 Getting to be a leader depends on my ability. 9 My ideas never come to fruition because I don't work hard enough. 10 Failure to succeed is due to factors beyond my control. 11 I believe that getting good grades depends on whether the teacher likes you. 12 When something good happens to me it is because I earned it. 13 No matter what I do the outcome mainly depends on luck. 14 I feel there is no way of protecting my personal interest from bad luck. 15 Failure to achieve my plans is due to my fault. I l l Entrepreneurial Intentions Questionnaire Please complete the demographics and then respond to the questionnaire by indicating the extent to which you agree or disagree with each statement using the following scale: 1. Strongly disagree 2. Moderately disagree 3. Disagree slightly 4. Agree slightly 5. Moderately agree 6. Strongly agree We thank you in advance for your participation Age (yrs) Sex M F Current marital status (circle): 1 Never Married; 2 Married; 3 Divorced; 4 Separated; 5 Widowed; 6 Cohabiting Highest level in school Completed Current level of Study Current program Currently working: Yes No if yes, please state type of job Spare time activities ; Parents owned a business Yes No Please respond to the questions below by cycling your answer. Do not skip any items. 1 I A lways s e i z e opportunit ies when they present themse lves . 2 I pursue projects that I be l ieve have high returns even if nobody has tried before. 3 I a m able to effectively utilize phys ica l r esources at my d i sposa l to improve my c h a n c e s for s u c c e s s . 4 I feel that s u c c e s s in a given venture is due to one ' s ability 5 I usual ly m a k e swift dec i s ions to attain my goals . 6 I don't like taking c h a n c e s . 7 I usual ly plan a h e a d for what I intend accompl i sh ing . 8 Ach iev ing my set object ives d e p e n d s on how hard I work. 112 cn' 3" Q) O CQ ^ - i CQ CD CD 9 ^ (/> O Q) Q. CQ CD —t —i CD D) CD <p_ CO • CO > CQ « CQ ^ > S CQ O - i - i Q-CD CD CD c±CO CD CD 3 *< 0 jjf 0 >< > W 3 § <D CQ 9 I am always searching for new ideas. 10 I commit myself to new opportunities cautiously. 11 I usually forge ahead with my own idea inspite of opposing views from others. My achievement in life is determined by my own actions. I often act on impulse. I find uncertainty stressful. Failure to achieve my plans is due to my fault. I am good at influencing the course of events. No matter how hard I work, luck is not on my side. When I say I'll do something, I do it. I have a preference for low-risk projects. I feel there is no way of protecting my personal interest from misfortunes. I consider interpersonal relationships a fundamental resource. If I lose in a competition, I attribute it to unfair play. I believe change is a good thing. I stick to my idea even if it means loosing my job. No matter what I do, the outcome mainly depends on luck. I am good at originating ideas. 27 When I fail to perform well in a test, it is because I didn't work hard enough. I usually take the lead in initiating new projects. I can withstand setbacks in pursuit of my goals. I usually work tirelessly in pursuit of my intended goal. I believe that my success in life depends on those in authority. I usually try out new ideas. I go for activities that give a sense of security. 34 I am more concerned with why something will work than with why it will not work. 35 Getting to be a leader depends on my ability. 36 I feel comfortable with the way things are. 37 I consider myself the "gambling" type. 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 28 29 30 31 32 33 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 113 O CO _ ? _ CO O CO > > _ c/> • Q) O CQ ^ =, CQ ro ro Q) D - CQ «g CQ - j S 3 _<Q _ ro ^ CD *< ro ro CQ O —< Q. CD CD  ro 53 CD > CO CQ =r (0 CQ 38 I a m naturally c apab l e of present ing my idea c lear ly and to the point. 39 If I fail in life, I a m to b lame. 40 W h e n I fail in what I set out to ach ieve , I a lways start aga in with a new idea. 41 I have a powerful urge to compete . 42 In promoting an idea, I a lways adopt a s a l e spe r son approach . 43 My ideas never c o m e to fruition because I don't work hard enough to rea l ize them. 44 I like cha l leng ing s i tuat ions. 45 W h e n an opportunity ar ises , I don't hesitate invest ing in it. 46 I have the tendency to interrupt before others have comple ted what they have to say . 47 Fai lure to s u c c e e d is due to factors beyond my control . 48 I c an usual ly c o m e up with a plan ahead of others. 49 I wou ld take a c h a n c e than miss it. 50 If I bel ieve in an idea , I m a k e it happen . 51 I bel ieve that getting good grades depends on whether the teacher l ikes you or not. 52 In dec is ion-making situat ions, I exp lore poss ib le opt ions caut iously . 53 I a m powerful at oral communica t ion . 54 W h e n someth ing good happens to me, it is b e c a u s e I ea rned it. 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 Percentage of Missing Data Appendix C p. 115 115 Percentage of Missing; Data: Entrepreneurial Intentions Questionnaire Observed Variable Number of Observations Percent of Observations Missing Missing Yl 4 1.1 Y2 None -Y3 6 1.6 Y4 4 1.1 Y5 2 0.5 Y6 6 1.6 Y7 1 0.3 Y8 1 0.3 Y9 None -Y10 2 0.5 Yl l 1 0.3 Y12 2 0.5 Y13 2 0.5 Y14 2 0.5 Y15 7 1.9 Y16 8 2.2 Y17 6 1.6 Y18 2 0.5 Y19 8 2.2 Y20 7 • 1.9 Y21 4 1.1 22 4 1.1 Y23 3 0.8 Y24 5 1.4 Y25 2 0.5 Y26 3 0.8 Y27 3 0.8 Y28 6 1.6 Y29 7 1.9 Y30 13 3.5 Y31 2 0.5 Y32 3 0.8 Y33 8 2.2 Y34 2 0.5 Y35 4 1.1 Y36 6 1.6 Y37 None -Y38 1 0.3 Y39 None -Y40 1 0.3 Y41 13 3.5 Y42 11 3.0 Y43 3 0.8 Y44 5 1.4 Y45 9 2.4 Y46 3 0.8 Y47 2 0.5 Y48 5 1.4 Y49 3 0.8 Y50 12 3.3 Y51 2 0.5 Y52 3 0.8 Y53 None -Y54 2 0.5 117 Appendix D Item-Total Correlation Table DI Item-Total Correlation for Proactive Subscale p. 118 Table D2 Item-Total Correlation for Need for Achievement Subscale p. 118 Table D3 Item-Total Correlation for Locus of Control Subscale p.l 19 Table D4 Item-Total Correlation for Risk-Taking Subscale p. 120 118 Table D l Proactive Subscale: Item-Total Correlation (ru) - 12 Items No. Item Description r 1 I always seize opportunities when they present themselves. .14 5 I usually make swift decisions to attain my goals. .18 9 I am always searching for new ideas. .39 13 I often act on impulse. .04 18 When I say I'll do something, I do it. .20 23 I believe change is a good thing. .23 28 I usually take the lead in initiating new projects .41 32 I usually try out new ideas. .36 36Re I feel comfortable with the way things are. -.09 40 When I fail in what I set out to achieve, I always start again with a new idea. .23 44 I like challenging situations. .41 48 I usually come up with a plan ahead of others. .33 Table D2 Need for Achievement Subscale: Item-Total Correlation (rit) - 13 Items No. Item Description ru 3 I am able to effectively utilize physical resources at my disposal to improve my chances for success. .20 7 I usually plan ahead for what I intend accomplishing. .27 11 I usually forge ahead with my own idea inspite of opposing views from others. .27 16 I am good at influencing the course of events. .33 21 I consider interpersonal relationships a fundamental resource. .23 26 I am good at originating ideas. .41 30 I usually work tirelessly in pursuit of my intended goal. .28 34 I am more concerned with why something will work than why it will not work. .17 38 I am naturally capable of presenting my idea clearly and to the point. .45 42 In promoting an idea, I always adopt a salesperson approach. .36 46 I have the tendency to interrupt before others have completed what they have to say. -.10 50 If I believe in an idea, I make it happen. .39 53 I am powerful at oral communication. .38 119 Table D3 Locus of Control Subscale: Item-Total Correlation ( r „ ) - 16 Items No. Item Description ru 4 I feel that success in a given venture is due to one's ability. .25 8 Achieving my set objectives depends on how hard I work. .39 12 M y achievement in life is determined by my own actions. .38 15 Failure to achieve my plans is due to my fault. .26 17Re No matter how hard I work, luck is not on my side. .25 20Re I feel there is no way of protecting my personal interest from .25 misfortunes. 22Re If I lose in a competition, I attribute i f to unfair play. .16 25Re No matter what I do, the outcome mainly depends on luck. .28 27 When I fail to perform well in a test, it is because I didn't work .34 hard enough. 31Re I believe that my success in life depends on those in authority. .29 35 Getting to be a leader depends on my ability. .42 39 If I fail in life, I am to blame. .41 43 M y ideas never come to fruition because I don't work hard enough to realize them. .10 47Re Failure to succeed is due to factors beyond my control. .07 51Re I believe that getting good grades depends on whether the teacher likes you or not. .22 54 When something good happens to me, it is because I earned it. .24 120 Table D4 Risk-Taking Subscale: Item-Total Correlation ( 4 ) - 13 Items No. Item Description r« 2 I pursue projects that I believe have returns even i f nobody has tried before. .22 6Re I don't like taking chances. .05 lORe I commit myself to new opportunities cautiously. -.03 14Re I find uncertainty stressful. -.07 19Re I have a preference for low-risk projects. .17 24 I stick to my idea even i f it means loosing my job. .08 29 l e a n withstand setbacks in pursuit of my goals. .03 33Re I go for activities that give a sense of security. .06 37 I consider myself the "gambling" type. .23 41 I have a powerful urge to compete. .15 45 When an opportunity arises, I don't hesitate investing in it. .30 49 I would take a chance than miss it. .36 52Re In decision-making situations, I explore possible options cautiously. -.03 121 Appendix E Description of Holland's (1994) Occupational Types p. 122 Description of SDS Subscales and Observed Variables p. 123 Description of EI Subscales and observed variables p. 124 SDS Correlation Matrix, Means, and SD for the Whole Group p. 125 SDS Correlation Matrix, Means, and SD for the Bus and Educ Group p. 126 SDS Correlation Matrix, Means, and SD for the Eng and Science Group p. 127 SDS and EI Correlation Matrix, Means, and SD P. 128 122 Description of Holland's Occupational Types Occupational Type Personality Type Occupational Environment Realistic Fosters technical competencies Conforming Characterized by ordered and achievements. Individuals Frank and systematic are generally asocial and lack Materialistic manipulation of objects, human relations but display Natural tools, machines, and mechanical ability. Jobs include Practical animals. automobile mechanic, farmer, and electrician. Investigative Fosters scientific and Analytical Characterized by demands mathematical abilities. Cautious that entail symbolic, Promotes a scholarly Critical systematic, creative atmosphere. Individuals see the Introverted investigation of physical world in complex, abstract, and Methodical and cultural phenomena. original ways. Jobs include Rational biologist, chemist, & geologist. Artistic Encourages expressive, Imaginative Characterized by free, original, intuitive, Impulsive unsystematized activities nonconforming, and artistic Intuitive and competencies to create abilities. Occupations include Nonconforming art forms or products. writer, actor/actress, and Original musician. Social Fosters social competencies. Cooperative Entails the manipulation of Encourages interpersonal Sociable others to inform, train, cure, relations. Occupations include Helpful and develop or enlighten. religious worker, counselor, and Idealistic teacher. Understanding Enterprising Encourages people to see Adventurous Entails the manipulation of themselves as aggressive, Domineering others to attain popular, self-confident, and Impulsive organizational or self-possessing leadership and Ambitious interest goals. speaking ability. Occupations Optimistic include salesperson, entrepreneur, and manager. Conventional Encourages clerical Conforming Entails the explicit, . competencies. Sees the world in Careful ordered, and systematic conventional, stereotyped, and Conservative manipulation of data, such constricted ways. Jobs include Orderly as, record keeping and data bookkeeper, banker, and Conscientious processing. clerical. 123 Description of SDS subscales and observed variables Subscale Variable Item # Question/Statement Realistic A R 1 Activities/Realistic: 11 Composite Items CR 2 Competencies/Realistic: 11 Composite Items OR 3 Occupations/Realistic: 14 Composite Items SRA 4 Self-Estimate (a)/Realistic: 1 Item SRB 5 Self-Estimate (b)/Realistic 1 Item Investigative A l 6 Activities/Investigative: 11 Composite Items CI 7 Competencies/Investigative: 11 Composite Items OI 8 Occupations/Investigative: 14 Composite Items SIA 9 Self-Estimate (a)/Investigative: 1 Item SIB 10 Self-Estimate (b)/Investigative 1 Item Artistic A A 11 Activities/Artistic: 11 Composite Items C A 12 Competencies/Artistic: 11 Composite Items OA 13 Occupations/Artistic: 14 Composite Items SAA 14 Self-Estimate (a)/Artistic: 1 Item SAB 15 Self-Estimate (b)/Artistic 1 Item Social AS 16 Activities/Social: 11 Composite Items CS 17 Competencies/Social: 11 Composite Items OS 18 Occupations/Social: 14 Composite Items SSA 19 Self-Estimate (a)/Social: 1 Item SSB 20 Self-Estimate (b)/SociaI 1 Item Enterprising A E 2 l ' Activities/Enterprising: 11 Composite Items CE 22 Competencies/Enterprising: 11 Composite Items OE 23 Occupations/Enterprising: 14 Composite Items SEA 24 Self-Estimate (a)/Enterprising: 1 Item SEB 25 Self-Estimate (b)/Enterprising 1 Item Conventional A C 26 Activities/Conventional: 11 Composite Items CC 27 Competencies/Conventional: 11 Composite Items OC 28 Occupations/Conventional: 14 Composite Items SCA 29 Self-Estimate (a)/Conventional: 1 Item SCB 30 Self-Estimate (b)/Conventional 1 Item 124 Description of EI Subscales and observed variables Subscale Variable Item# Question/Statement Proactive (EI) Pr 1 9 I am always searching for new ideas. Pr2 28 I usually take the lead in initiating new projects. Pr3 32 I usually try out new ideas. Pr4 44 I like challenging situations. Pr5 48 I can usually come up with a plan ahead of others Risk-Taking Rsk6 45 When an opportunity arises, I don't hesitate investing in it Rsk7 49 111 11. I would take a chance than miss it. Need for Nac 8 16 I am good at influencing the course of events. Achievement Nac9 26 I am good at originating ideas. Nac 10 38 I am naturally capable of presenting my idea clearly and to the point. Nac 11 42 In promoting an idea, I always adopt a salesperson approach. Nac 12 50 If I believe in an idea, I make it happen. Nac 13 53 I am powerful at oral communication. Locus of Loc 14 8 Achieving my set objectives depends on how hard I Control work. Loc 15 12 My achievement in life is determined by my own action. Loc 16 27 When I fail to perform well in a test, it is because I didn't work hard enough. Loc 17 35 Getting to be a leader depends on my ability. Loc 18 39 If I fail in life, I am to blame. Social (SDS) Soc 1 16 Activities/Social: 11 Composite Items Soc 2 17 Competencies/Social: 11 Composite Items Soc 3 18 Occupations/Social: 14 Composite Items Soc 4 19 Self-Estimate (a)/Social: 1 Item Soc 5 20 Self-Estimate (b)/Social 1 Item Enterprising Entr 6 21 Activities/Enterprising: 11 Composite Items Entr7 22 Competencies/Enterprising: 11 Composite Items Entr 8 23 Occupations/Enterprising: 14 Composite Items Entr 9 24 Self-Estimate (a)/Enterprising: 1 Item Entr 10 • 25 Self-Estimate (b)/Enterprising 1 Item CJl LO CO CO O i 3C O I O I O I O I O I O I O 1 I—1 O O I I O O O O I O O O O O O O O O O O O O P - 1 M h O H * r O r ^ M M U i O M O L O O C O H - ' U J f - ' O O O O O h-> - J CO - J ^ > J C 0 s l ( J i U l l D ( J \ v ) O P O ^ P O ^ O ^ v ) 0 ^ ^ f f l v l ^ K ) W P P O O O ^ ^ W i X i U l P O O i U 3 W C O P - > M O i < j n O M C O i X > ^ b O O r o r O U l O i ( J i P J P - ' P J C O O r o NJ co O i u i • • • • O I O I O I O O O I O I 1—* I O O O I O O O O I O O O O O O O O O O O O P - 1 I-1 M kD P-> h-> W O J O C O M N J O N J O r O O C n O O O O h - ' O O O l - > O O O r O N J N j r O M ' N j r O C O > - > r O ( J 1 0 i O • J I O ^ Q O W C D U C O ^ W M M O H I D W H J i U l W C O l D I O O K I C f l H ^ C O L O U l ^ W C O O O P- 1 CO CJI (Jl H W ^ C D C T i C n C D O J C O H C D - v J O W 0 i 0 i P J M C P 0 0 P J M C ? l C 0 r O M C ^ M k D M M ^ M C 0 P - ' O CO V£> P-> P-> O I O I O O O O O O h - 1 I I O O O I O O O O O O O O O O O O O O O O P - 1 CO <Ti U • ( J i O C 0 O N J P - ' N J O r 0 P J O O O N J O O O O C 0 O P - ' O O r O P - > P - i r 0 C 0 C T i C 0 0 i r 0 0 i O rO CO i n W - J l £ > U l D - J U i C n H O ^ O i U l W C D W C T i U ^ ^ H W ^ O . E - H a i U l ^ L O C n - J O ^ m a i o i \ ) u ) ^ o o w o c o oo l O i v u i a i w i v i D W U i w - J o u i ^ c o o H O ^ a i o O O i UD O I O I O I O I M I I I l O I I O O O O I O t l O O O O O O O M M U l . C H ^ W l J i H ^ O O O O O O O O O O O O O O O O O O N J O i r O C O O J C O O C T j ^ l O - J C J i f J i ^ J i O U ) M (Jl ^ U N l C O W O f f i f O C O C D L O O O N l > l U i i C > ^ W O 1X3 r—1 P J P J ( J l - J P - l l X > C 0 U l O - J CO 0^  N J U l ^ ^ H C D ^ l D l D C r i H H - J ^ H C X I J l C D - J O 0 KD O I O I O I M O I I O l O I O O O O O O O O O O O O O O O h - 1 Co c n • • . U l O O i O O i O O O O O O O O O O P J O P J O O P J O O O O r O P J O P J 0 CJl 00 W U i J i - J L O W O a i V- 1 O i OS. C ^ C 0 H * - W ( J i J i K ) ^ L n ( J l C 0 K ) f f i O a i ^ U ) O 01 KD I D C O J ^ O U i f O O - J - J N J CTi U U l C T i U i C J v J U ^ W O C D H ^ C n ^ ^ H - J O NJ Cn O I O I 1—1 O O O I I O O O I O O O O O I O O O O O O O P - * CO - J O i O (Jl O O O O P-1 O O O O P - ' O P - ' P J P - 1 C O O O O O O C O ( J i C n C n O l£> - J - J CTi U v ) O U3 IX> CO LuJ C i ~ J W L n K ) H U ) W H O s J W O C O g U i > l H O 1X1 U l ~ J A . O CO <X) CO r o NJ 0 3 C O O D O ^ C O U i ( J i H > C . k D C O X j ^ ^ V £ ) ^ 0 O i CJl O O M I O O O O I O O O I O O O O I I O O I O O O r - 1 -vj CO . . . . . . . . . CO P-1 cTl O O O O M P-» O O O P - > O O O O O O O O O O O 0 i U i 0 i O W Cn O W <X> CO CO CTt (Ti J i J i s J O U i a i U i ^ W c n t O W C J C D r O r O O I—1 CO 6 CO O IO U1 -O (Jl 05. h-i L O H C T i W U l M - J C n ^ W C f i r O O W M m O •X) CO h-» I I O O O O O O O O O O O O O O O O O O O O r - 1 P-> CO • • • • - . O O O r o I-" P-1 O O C0P- 1P->OP J UiP->C0OOC0P J P- 1 P->0iO M oi. o o os* oo o r o M o o ^ i t . c o m O H ^ ( T i ^ - J H L n o i c n o o O - J O U i NO —1 - 0 CO W O C n P ^ C A O C O W H C O t O H U C O O KD - J NJ O i I I I O O O I I O O I O O O O I O I O l O M CO U i O P-1 O O O O O O O O O P J P J O P J O O O O O 0 i O U l 1—1 - J U l O 0* U l H N I H H C Q W I D ^ H W O W W H O co os. p-» r o i — 1 P-* u i a\ r o WCTICOPOUIOIH-JUIWIOCO^O O i O O O O O O O O O O O O O O O I I I I I I P ^ CO - J • • • • • • • j — 1 H-> i—1 i-» r o i—1 o P J O O O O O O O O P J O O P - , O M 00 O i O cTi U l Oi" O i OA P O l ^ c n M H J i O U l ^ J i ^ C n O 00 NJ P-» CO 00 U l KD CO ^ O O f i W - J C D O > W O U l r O W O OS. U i I I I I I O O O O O O O O O O O O O O P - 1 U l <£> O O O O O O O M r O P P P W W W J i U l O l ^ O KD c o u i c£> a\ cTi o o i i \ J o o c o r o r o o a " i - J u > u i * £ > o NJ 00 cTi CO CTi NJ CO P U i s l U O P U K T i O ' i U i W I v O ro OJ o o o o o o O O O O O O O O O O O O P - 1 h-» • • • • • • * CO - J O O O I—1 O P-* r o U W P M M U M M U i ^ c T i O - J 00 —1 CO (Ti •£> P-* O U I J i H O ^ W v l s J O ^ O O r o c o —3 co CTI co r o r o H-» U)V1WPMIOUIWCBUIUIO CO O O O O O O O O O O O O O O O O O P - * r o cn • • . . . . . O O h-* O O O W r O H W P U i r O W U l ^ O N0 (Ti Oi. KD CO O U l CO -O i n u i P W P H W H O W O pj 0^ co r o cTt o —J c o r o o o o c o r o u i - j r o L o o CO \D cn r o i i i i i i o O O I O O O O O O M r o oo t—* r—1 r o p j P-J o o o p - > o o p - ' r o p - > p j M o P-> o p- 1 o <Ti r o en ^ r o a i D c n ^ r o ^ c o o - J O P-1 O i P-J O VD <Ti CO W P O i U i O - J W N J P O <X> CO O O O O O O O O O O O O O O O P - 1 01 U i • • • • • • • p-» o o o OP-> p- 1 p j r o o c o p - ' p - ' p - ' o o U l CO O KD CO CO rO CO O C O C J i C O O U i P - ' U l - J O U l P-1 O rO - J CO 0^  O i O C O U i a i C D H C T i ^ C T l O I—' U i O O O O O O O O O O O O O O P - 1 -J r o o o P-1 P-> r o o t-^ r o r o c o r o c o c r i c n o KD O U l -O U l J i O i KD NJ C O C O ^ P m P O O KD 00 O O CO CS CO ^ H C O - J C O ^ r J i O P-1 U l O O O O O O O O O O O O O P - 1 r o oo • • • • • . . 0 tTi r o p-> o P-1 p- 1 c o r o r o u i u >t u o o O i -J O i KD r o u i cn m NJ -J c o o P-1 *X> P-1 O i CO O i - J O i O i O i i X J C O t T i N J O KD O O O O O O O O O O O O O P J c o r o . * • • • • . O M CO O rO P-" H» O i NO NJ NJ C J O r o —J co r o p- 1 oo o o cn <x> cn c o O i r o o P-1 U l CO O i CO P-1 IX) O i O <X> —] CO o —J O i KD O O O O O O O O O O O O P-1 01 co o P-* o o p- 1 p j r o i—1 NJ o r o o <X> 00 CO - J O *X> t- 1 P-< O i CO O i O o O i o m o O i p - » ' C 0 p-* u i - j r o r o o KD KD o o o o o o o O O O M M O i • • • • • • . . . . . r o r o o PJ o co r o H-> co p-* o CO —1 O i O i C O - O - J - J O i U D N J U l O O CO >X> O - J CO tX> NJ (Ti C O C l C J i O P-1 U l O O O O O O O O O W cn O i c o r o c o c o o i L o r o o i O i O o r o CO U) U i (Ti (Ti ID U l o (Ti KD KD CO U l PJ tO H* - J O NJ U i O O O O O O O O P > 00 O i . . . . . . . . . p-1 u i c o p - » p - » w r o o i 0 i 0 i o O O i C J i C O O i C O C T i O O 73 o Cd CO H CO PH CD o > o > > CO o o CO CO CO > > o CO > o o o o n CO o > o CO a co > o CO a PH 3 II n CO > O CO c O tfl II 2 CO t 1 ! co S <X) f-3 S 2 > o II o o n o tfl S ! o f 1 PI o pa o a r o (Ti CO CO CO CO CO (Ti Ul C/J b ui g» is ™ to w NO t o 0 — O — • —• o — O O.' o p O O p o P P p o o p 4 ^ ^ £ ^ o u i X i i o ; ! i b o N O •' _. o •' <=> •' .' o .' •' P g o P o .• o 7 •' •• ^ " ^ S - i S o ^ - - ^ b 2 P o P o o P 2 o r ~ - o o o N p ; - r - o o . o t o P ° u o ^ P 0 - P 0 .. _ u . f .• 0 *. P 0 o g | S - o - | p S £ § 2 ° i 2 § P - g — ~c i l to to -o .u l£ " . — NO , NO tr . oo o o I— NO ^ ~ ON p o o Ul ^ -J OO P o o '_ O hj o P b P o — Ul Ul o ° b P o — — Ul o w b p M O W \o o\ •—1 O ^ M K) — — - J t o O s ] Ul I—' •P. O K) W oo to o O — O SO t O O OO \o W K) ^ — o ~- oo — to — — — 00 4s» O Vi OO 0 0 S O S) U) w OO O 4^ e - t O OO N J 1 as • — ON ^ . — b o • .• • O ' i • _ U) O o O i— o to P „ • . b S ' b P r o o ^ P o b £ • ' w f e S S b ^ p g g p p g i 3 o £ p g g £ g o ^ i © o o 0 u i o • p g ° = > p - p g g t o - J ~- u, .' O.' K ) ~ * . 0 0 O i o C ^ O ^ P ^ k j U -Ul — Ui to o <*> b p oo ON o o b P — to to Ul O *• b P Ul o >— to o b p NO — tO Ul o "> o ui KJ 00 o o ° kl p o — J> O O M o — — IO v l W W OO O t o Z~ o o o 2 u> -—) — o ° ° P 00 — o o ^ M U i 0 O U J so b P P P t^ J H- O 4^ O — -t^  , 4^ U J N> b o o o J j b b o W s i OO NJ 1 SO 4 -^ U> g P P P K3 o o o ' U) Oi oo .' ui 4^ u» o o o o K> 4^ 4 .^ ^3 - 4 * to w o o o o o o UJ i—• KJ KJ O s j w s j <3 u i c> o p p p io •— UJ s i O OO U i o OO OO 4> O o o o p b to 4^ to s i OS U l o .' o o o ° b - to ^ s i OS W H—* UJ 4> OJ p p p p ; W U J U l U l i ^ O OO w P O O O i o U i U t — N) W SO NJ — p p o o o o — ^ w u i t o a i b ^ b ? " s l O O W O O j i O - * O U i K) - 4 i O O © O O O i— NJ >— t O 4^ U J b O0 U l - o oo o ^ O OO U i - o o o o o — N> UJ 4^ b OO O N> UJ o w w w o o o o — to UJ UJ b O U i 4 -^ O oo to J> O o o — b ~ b ro oo o s i UJ o o © © — p p p ! U i — U J s i V© SO U l — U l to o to i— oo — U i ON s i U) i— UJ — , UJ ON 4*. • t O s i Ui 2 o o  O s i s i s i b - 0 0 *• o .' .' •' K) O O O ^ Ul O £ O p p ! -o. o\ iji 1 _ U) U) ^  1 fipp-W W o P O K) O w 2 *• 4^ *• o o P o u> u> 1^1 <£• V\ o *• w o P o k) k) SO ^ UJ , — ui ui • oPoCPr J> -~ UJ O -fck o oo 0 ° O O g oo S P ui o 00 — K) O P o — "> ° Ul 0 - ^ - k o „ ' o o < Ul Ui t o ° ON k> o P r-4s. OO O 1 OJ o -^ ° ! ON UJ I SO o s ) O _ o UJ 4^ s i U l ^ OO "-/I ^ (sj — . ^ P S P S I S O - ' s ; - ^ s y y g p ! s : § . . J> ~ s — O SO O ON P U l SO po 0C ^ ON NO ' , P o , ° - I ' ^ ° — ° . S * * w bi P . ^ J > 00 Ul NO J> 1 o o ;"" i o , o ' M <-i • ON O Ul A w o - w -K> O J> Ul NO O 0 p oo 0 OO „ O u, ^ P k. P b ° O KJ — NO — o • - w ^ W M O , O Ui Ul C/3 > CO > o > > o o C/3 <z> C/3 > C/3 C/3 CD o m o w 0 3 to > C/l m CO > n o o o n C/3 n > C/3 n 03 O 55 > o o > C/3 ON m Z H £ o o tn r •n o !» H tn CO c C/3 tn C/3 a tn a c o > H o z: C/3 c CO o o G : ^ P S -I j$ ON o o Ul Kj Ul O ON \~ ~J O O P 1 0 u, § OO o o o o tO C/3 4a. 2 m O 'H-4a. 4a. VO 4a. VO .' O O O o '— b o w M a A ^ ui 4a. o vo — y ij O O 8 ° ° KQ l— <^ oo --J S £ § Q o - 4a. to — Ov Ui p ' M p po o oo _ Ul o o o b o : s w ' to . o vo , b o to ' O to > o — i ui o . OO V i VO £ u> '<— o 4a. Ov Ul £X o <=> , b ^ P b O _ Ov •—1 . 4^ 4a. 0 ^ * 0 to P P to 2 3 g 2 -o ~ o° ^ 0 0 ^ ^ P o o P 1 0 ki LL ov — -o oo J £ ~J - y - - io O i o ~ O — O Q Ui O OS P »- »o N! w P P P u o —1 P P ^ P S vfj w p p p g — ° o ^ MgvoP oo o° U) — o , to b P P g o — — p to OO VO ^  o o -o o ;_. o o to o b b ^ ui o •— ^ 0 VO Ul o • o o to - b b g° Ul Ui oo vo b •' •' ui w O O Ul Ul — o * o _ o o 4a. ,— O 00 Ov to 0 O Ul ° - s Ov Ov o 1 o Ov k ° ~ i P O Ui — V u, O to o ui S - o -OS — OV IO , o o - o b P P io ui — o to 4* 4a. to ov O °v O 0 ivPPiil U l ui to 00 Ov Ul Ul o 0 0 ° o to P P to ov 4*. to oo 4a. Ov VO O ^1 OO -o o \ - b ~o — o VO O O O ov — b VO Ov o 00 Ul O O U to * — - 1 o to O O -o b b io Ul 00 o 1 ifa. O O O to bv — 1 Ul Ul o I ° - Ul o o o Ul ^- ui ~ J o * to Ul to _> O Ul to '_ ov Ul O o oo o O O Ul 4v U l « O Ul o 1 - L); O O ov — to vo VO 00 H-« u b O — o 4* O O ui to ui vo to Ui to to O * ° o ;— p p '— 4a. ui ui oo ui to — oo o - ° o b p p — 4*. Ul ~J Ui tO tO OV o * * o 4a. p p 4a O O W W ON "J S) — -J o w ^ o to P P 4=> Q> Ui Ul OO Ui NJ to ui o *> - -P P o to Ul Ul o OW4VO 0 to ov io P r oo 4a. o Ul oo o ,-N 4a. O o o to — ov to VO 4a. o o to b to vo o oo o o to to Ui to to ov O .' to o ui — vo Ov o P r-. os o P ui £ to vo 00 * o P 1 5 o P 4* * • zZ ui — 4a-o P v*> K g £ o P to o 4a. >0 o P 4a. U> o P to — Ul o P to o uS vo 00 -> o P ui "1 to ~o vo 0 0 o P to ov io vo o P ui £ & ^ o P Ov *; to ^ to Ov o P Ul OV 4a. — o P o P Ul *" Ul ^ —1 OV o ~ s ui g OV O O o b b it £ .' o o b Ul oo o o 2 io '-O O O b to b O O O O O O O O O O £ C vS Ul Ul o to o — to vo Ul o it o P to — ^ oo O OV to P — \o — o -° to P — U) Ui OS o ^ o P oo — o ^ • ;— ° U J ho KO o w p U. KJ Ul U J o 0 0 KJ P U J U ) co to O O O O O ' o ; S o S ' o P —1 P o P o « * o P b oo O « ^ o o ^ • o ° *• P 4^ vo to G vo to !g § P 2 VO — ov Ui oo to — — to — to 4a- VO OV .' O o io — ui Kj ui O o ° — to o Ov * Ul „ Ov P I p 2 o o ; IO to VO 5 S — OV p p S ov 00 0 P o to Ui "H o VO g o ^ O OO io ov o o o C O >— - J — o o — KJ UI ON U J C O i o o > k) U ) i U J o ' o o O N 4> o o 4 i 4*. to — to -~J OS O — to OS 4*. Os O O Ui to oo — U) ~~J o o 00 U) J> o o o p p ^ U J U J OS OS OS o ^ U J 1 to H— U J O " i- M U l s i W O o o o o — o tO Ui U J o CO 4> O - J O o -o -o o o o o o — 4^ tO O — Ui CO o Ui — Ul O O O — i—1 u> o ~ J 4> O 4^ WO ) o o 3 U J ' j i 4 i tO p p , ui j> ; -O O i U J to , o — 4o- O -J o Ui o — o U J o Ui o 1-b P b 4a. to ui -O 4a. O O 0 0 o Ul P 4a. H W O O Ov Ul o * o 4a. P to to ^— —I OO OO Ul n *• « 1 — Ul ' . 4a. to 1 I to ov ' J Ui O ' 4a. O > OV O O to o Ov S Ul vo ui p g p O — ' •O Ul ' ui vo ; p ~ ' ui b OO O 4a. O ov g o g P o o b ° to P o ui b P o to 1^. to VO So P o * • 4a. s °= oo 0 P o ui u, o to o _ to b Rig u P U O IO Ul VO o to o — o — o '— OV — Ul IO 4a. 4a. o ° -u iPb Ul Ul o VO 4a. O o ^ 4a. ."-* to o to o — o Ul O Ov O o o C/5 o > o > on o C/5 o C/3 on C/l > on o m o m on m > on m 03 > o o o o o on o > on o CO D K - 2 ui F 1 o > on so D o H m tn z o 2 Cd m o Q Cd z o m on c CO o o c ^ 1 o o o — CO o N O I O 4=-o NO 00 to ^O I O ^O NO s> N O tO * -4 ^ C N NO 4 * NO ON tO 4 * ON bo -o. b O N b Ul to NO 4*. O N p p 4 * U J NO 4=-o © to K J O N O 00 *-o o b — NO 00 o o p p 4*. O N O O p O b b NO oo NO O N O O b b ON U J ON NO o o to to — o Ji. — O O b '— NO o 4 * — o o to u> — Ui NO o o b to as O N p p to to ~ J ON 4 * NO • • b b O N U I p o p b — — Ul N O U J O N to O N o o o — — N O to O N 4*-o o o b — b O N — J 4 ^ w ui o o o o b to b 4 * 4 * -o to to Ul .' O .' S >° P. O N ^ to - Ul *• o o o b ;_ b w o\ u - O H. U J ^ ° p ^ 5 $ P o P ° 8 S P o .' o i i o K J U. „ P o P 2 to 2 oo £ -P o P o £\ o U J I O so to O N P o P 0 ^ , 0 = s = ° o ° o p p p b to O N U J U J — o U J ~ J O N o o — b o o b to to O N NO U J o p b — — O N *. to o o b -o -j 4 * NO .' o — O kj U J O N „ O — u> O .' O a© p P O N © u, m O P O O L- b 4=. ~ J o -o. o o b — ui ui U> -O O O b J > N O O N -O p p U J O N — O N O O O '— to U J J > to o o p p p o p p p o o p p p o p — 7s1 0 0 0 00 — 4^ W Ul M 0 0 0 K J — -o to to U i O NO O .' o o ;_ — u; u> - J ° ON to — ~j to U J Ul •— to 00 J > — ON 0 0 0 0 0 0 o o o b o o b U J s> p p to o o b 00 o o b o O N p "~ *• o © - ^ ul ^  b p o to i— _ to o o 4*. Ui tO N O o o b b U J to O N J > O .' b © S to 4 ^ O o P 0 0 O N 00 — sj o P gs o P S -o p p s £ o P to to J > ^ co o o P U J w o P U J to O N - J U J — ' O P S £ o P ^ N^ U J o^ o w o r' P, J > 0 to o o o  r-' o O U J S J U i ^ o P b o U J O N Ui P P o to O NO 1 o b b £ ^ 0 o b 0 £ <-/! O 0 s c *• ~ J o O to • NO — to p s ° b r § gp 00 O p g o , ^ b ui oo 4 ^ O N O o to to 2 4 * O N U J o o b b to — 4a. Ui o o b b O N — 4 * O N O O b — J > Ul 00 U J 0 o b — O N 4=-U J to .' O O p p to n P o o ^ to ^ o 1^ P o ^ Ul o b p ^ Si I b P 00 U J Ul J > o b p — to U J -o o M b p U J to U J ~ J , o b — O N o — o b o o to rt — 00 t--> NO 00 w 00 o P o 10 — ; ON U J — o P o O b ui P N to O O N o P o iNi o b — S1 ui NO O N 4^ O P p U J U J O N 00 os P P P ~ £ s Ul NO o P o io r; b - J N ^ 0 0 NO NO W P P Ul 1 — U J to ~ -O. Ul O p b — 00 U J ~ - U J O p to '— O N O O N 00 O O b to to o J > - J o o b b ui O N NO U J P P — O N Ul ON o o b to Ul J > ON to o o — to — 4> 00 NO U J to U J O N — U J U J Ul to J > U J IO o o p o Ul 00 Ul ON J > U J U J Ul o b _ S £ o P o to - : « o ^ J > O N O N O P O b b NO P U J O P O U J to U J Ul o os O Ui o o P -to — b to o o IO O o 0 r* to o Ui o — to U J so J > to .' o fcj O N OO K J p p O N o _ b b 00 5 Ui o b o o o o U J tO J > to ~ J SO O O 1 i— to so o 00 00 o o J > to O U J U J to O o to to so o U J Ul o o to — U J U ) -^ i to — ~ J NO NO O N o o K J to -O U J to J > o — to b — o 00 o to U J i 1^ so ' tO ON ' P P ! io J> i SO U J I 1^ to 1 P P : to to ' to 00 ' 00 J > O O ' 1^— to 00 1 O Ul o o to U J J > o U J U J P r~ ui b ui o NO O - U i - M W I J O ? ^ - J S . - O * - I O — J > O S > , N O U i U l t O O U l — o 0 0 0 0 0 0 — io — to to U J '— b NO OO 00 ON -O NO O O U J J > U J 00 to O O p O O p — to to — io ~ b U J o 00 to to O 4 i o to to o 0 0 0 0 — to U J to b Ui 6 A NO O SO M Ui A O O p o — to — o•J si O 00 O N O O — U J b Ui o U J o o o o " o o o o - p s 0 ^ So § K? O ON o 4 i to J > NO 00 o w I s — o b o o o o o 90 > -0 -^5 N O J > z ?° H ^ W 73 tn co GO 75 - 0 2 : > o 00 > n NO z > n o z > z > o z > r o n r o o r o n r o o r o o 00 CO o n co o n to CO o o U J CO o n J > co o n Ul tn z H g ON m z H CO a CO z D tn K J 0 0 z „ o ^ !^ tr) s > CO c H o D tn r Table of Specification Appendix p. 130 130 Table of Specification Construct Evaluative attributes Characteristics 1 Proactive • Realize a vision Optimistic Personality • Abi l i ty to seize opportunities. Innovative • Opportunity-seeking Foresight • Make things happen A Doer • Stand out from the crowd Intuitive • Alert to opportunities 2 Risk Taking • Abi l i ty to rock the boat Impulsive » Unconstrained by situational Adventurous forces Resilient • Experiment with new ideas Persistent » Take risks for personal goal » Manage risk situations 3 Need for » Desire for self-determination Sociable Achievement » Demonstrates organizational Enthusiastic skills and time management Resourceful » Desire for feed-back Competitive 4 Locus of Control < » Desire to be in charge of own Strong-willed destiny Bel ief in oneself » Internal reinforcement Strong personality » Accept responsibility High self-esteem 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0089076/manifest

Comment

Related Items