© 2012 Rong Su
THE POWER OF VOCATIONAL INTERESTS AND INTEREST CONGRUENCE
IN PREDICTING CAREER SUCCESS
BY
RONG SU
DISSERTATION
Submitted in partial fulfillment of the requirements
for the degree of Doctor of Philosophy in Psychology
in the Graduate College of the
University of Illinois at Urbana-Champaign, 2012
Urbana, Illinois
Doctoral Committee:
Professor James Rounds, Chair
Professor Fritz Drasgow
Professor Lawrence J. Hubert
Professor Brent W. Roberts
Associate Professor Daniel A. Newman
ABSTRACT
The present dissertation examined the incremental validity of vocational interests
beyond cognitive ability and personality for predicting academic achievement and career
success. Analysis of a national longitudinal survey, Project TALENT (Wise, McLaughlin, &
Steel, 1979), showed that interests were the most influential contributor to income, even
within occupational groups and after controlling for occupational prestige. Interests were
also found to be powerful predictors of college grades, college persistence, degree
attainment, and occupational prestige. The effects of interest congruence were examined
using polynomial regression and response surface methodology (Edwards, 2002). In most
cases, interest congruence was positively associated with academic achievement and
career success. Interest congruence in Science was particularly important for better college
grades, great likelihood of persistence, and a higher income. Lastly, a moderated mediation
model for career success was developed and tested. Number of children had negative
effects on income and occupational prestige, for women in particular. Interests mediated
the effect of gender on career success and were found to be fundamental channels through
which gender operates and influences career outcomes. Degree attainment mediated the
effects of interests, ability, and personality on career success and was an essential pathway
for individuals to attain prestigious occupations. Results from the present dissertation have
implications for the research on person-environment fit, measurement of individual
differences, personnel selection, and provide insight for the debate on the causes of
differential career attainment for men and women.
ii
To my parents,
Shixiong Su and Chengping Pei
iii
ACKNOWLEDGEMENTS
It is with deep gratitude that I write down the following words of thanks. First of all,
completion of this dissertation would not have been possible without tremendous support
from my doctoral committee. I am utterly grateful to Professor Fritz Drasgow, Professor
Lawrence J. Hubert, Professor Daniel A. Newman, Professor Brent W. Roberts, and
Professor James Rounds (in alphabetical order of last names). All of them have served as
my role models and mentors through graduate school, have imparted their knowledge to
me in classes and in person, and have greatly inspired me in my research pursuit. They
offered me numerous valuable advices and spared much of their precious time with me
from the incubation of the dissertation idea to the correction of a last typo. Their positive
influence has shaped me and will always be kept in heart.
Special thanks to my advisor, Professor James Rounds. He is the reason that I came
to the United States for graduate studies and is the force that supported me, among many
other students that he unconditionally cared and helped, to achieve everything big or small
in my career. His knowledgeableness, pure interest in science, passion for research,
devotion to work and his students, selflessness, humor, and wisdom are the best gifts that a
student could receive and the best legacy that a student could inherit.
Wholehearted thanks to my boyfriend Rui Xue, whose love, understanding, and
courage has supported me every moment in life, at work, and throughout the dissertation
completion process.
I dedicate this dissertation to the most important and influential people in my life –
my father, Shixiong Su, and my mother, Chengping Pei – with the utmost love and respect.
Thank you for your guidance and faith in me over the past twenty-six years.
iv
TABLE OF CONTENTS
INTRODUCTION..................................................................................................................................................... 1
PERFORMANCE ACROSS THE LIFE SPAN: A BROAD DEVELOPMENTAL PERSPECTIVE .................................... 5
VOCATIONAL INTERESTS: CONCEPTUALIZATION AND MOTIVATIONAL PROPERTIES ..................................... 6
MEASUREMENT OF INTERESTS ....................................................................................................................... 10
PREDICTIVE VALIDITY OF INTERESTS: BEYOND ABILITY AND PERSONALITY ............................................. 11
EFFECT OF INTEREST CONGRUENCE ON PERFORMANCE: THE POLYNOMIAL REGRESSION APPROACH .... 15
INTERESTS MEDIATE GENDER EFFECT ON CAREER SUCCESS ...................................................................... 18
FAMILY CHARACTERISTICS MODERATE GENDER EFFECT ON CAREER SUCCESS ......................................... 20
DEGREE ATTAINMENT MEDIATES THE EFFECT OF INDIVIDUAL DIFFERENCES ON CAREER SUCCESS....... 22
CURRENT STUDY.............................................................................................................................................. 23
METHOD ................................................................................................................................................................ 24
PARTICIPANTS ................................................................................................................................................. 24
MEASURES ....................................................................................................................................................... 26
DATA ANALYSIS ............................................................................................................................................... 37
RESULTS ................................................................................................................................................................ 41
DISCUSSION ......................................................................................................................................................... 58
POWER OF INTERESTS IN PREDICTING ACADEMIC ACHIEVEMENT AND CAREER SUCCESS ........................ 58
INTEREST CONGRUENCE AND PERFORMANCE ............................................................................................... 64
DETERMINANTS OF CAREER SUCCESS FOR MEN AND WOMEN .................................................................... 69
LIMITATIONS AND FUTURE DIRECTIONS ....................................................................................................... 72
v
CONCLUSION ....................................................................................................................................................... 76
REFERENCES ....................................................................................................................................................... 77
TABLES................................................................................................................................................................... 90
FIGURES. .............................................................................................................................................................. 123
vi
INTRODUCTION
“High occupational-interest scores should correlate with measures of superiority.”
– Strong (1943, p.486)
The study of interests has a long and prosperous history dating back to the
beginning of the twentieth century (Parsons, 1909). Early pioneers in vocational interest
research predicted that “the developments with regard to the diagnostic meaning of
interests would prove to be one of the great, if not the greatest, contributions to applied
psychology” (Strong, 1943). Research in this area has since made monumental advances in
interest assessment and interest measures have become an indispensable element for
career counseling. The Self-Directed Search (Holland, Fritzsche, & Powell, 1994), for
example, has been used by more than 30 million people worldwide and has been translated
into more than 25 different languages (Psychological Assessment Resources, 2009); the
ACT Interest Inventory (American College Testing, 1995) is administered to over 1 million
middle school and high school students in the U.S. per year; and the Strong Interest
Inventory (Harmon, Hansen, Borgen, Hammer, 1994) is used by more than 70 percent of
U.S. colleges and universities to assist students in their career planning. The power of
interest measures for guiding individuals to their preferences and passions is well
documented in the literature.
Despite the important role of interests in career guidance and early investigations
on the link between interests and superior school and work performance (e.g., Strong,
1943), the study of interests has been dormant in the personnel selection literature. Rarely
1
have interests been used to predict job performance or to explain career success. One view
that is commonly shared by scholars is that interests play a major role in individuals’
educational and career choices (e.g., choosing a college major, choosing an occupation) and
yet have minimal effect on how well individuals do after they enter a certain major or
occupation. For example, Barrick and Mount (2005) stated that “interests influence the
types of environments we seek and the types of people and activities we prefer…
understanding the person’s interests will not likely enhance our understanding of how we
do work” (pp. 365-366). Ployhart, Schneider, and Schmitt (2006) also claimed that “the
prediction issue in interest measurement is not how well someone will do but which
among several alternative choices they will choose. In a real sense, vocational interest
measurement is about binary choices… (but not) about predicting continua of behavior or
performance” (p. 469). Cognitive ability and personality, on the other hand, are believed to
be the critical dispositional basis for determining how a person behaves and what the
person can achieve once he/she has chosen an environment that is consistent with his/her
interests (e.g., Ployhart et al., 2006).
This view that interests only determine one’s preference and choices but not “how
well we do work” is due to, at least partly, mixed empirical results on the relations between
interest and performance criteria and a lack of comprehensive meta-analytic evidence to
support the interest-performance link. There has been a substantial amount of primary and
meta-analytic studies suggesting the critical role of cognitive ability and personality for
educational achievement (e.g., Kuncel, Hezlett, & Ones, 2004; Kuncel, Ones, & Sackett,
2010), job performance (e.g., Hunter & Hunter, 1984; Barrick & Mount, 1991), and career
success (e.g., Judge, Higgins, Thoresen, & Barrick, 1999; Judge, Klinger, & Simon, 2010).
2
Results from these studies established a consensus about the importance of cognitive
ability and personality for performance and provided strong support for the use of
cognitive ability and personality measures in personnel selection. In contrast, no
comprehensive meta-analytic review on the predictive validity of interests for academic
achievement and job performance was published until very recently. An older meta-
analysis (Hunter & Hunter, 1984), based on a limited number of three studies using the
Strong Interest Inventory, reported a .10 correlation between vocational interests and job
performance. In light of this result, Schmidt and Hunter (1998) concluded that interest
measurement is not a good means of predicting future job performance. Only recently, two
meta-analyses (Van Iddekinge, Roth, Putka, & Lanivich, 2011; Nye, Su, Rounds, & Drasgow,
2012) showed that interests have significant relations with a range of performance criteria
and called for reconsidering vocational interests for personnel selection.
Two questions, however, beg further investigation: First, what are the effects of
vocational interests on performance over and beyond cognitive ability and personality?
Only a limited number of studies to date have examined this question. For example, Van
Iddekinge, Putka, and Campbell (2011) demonstrated with a military sample that interests
have incremental validity beyond cognitive aptitude and personality for predicting job
knowledge, job performance, and continuance intentions. There has yet to be a study to
examine the incremental validity of interests on a variety of performance criteria in both
education and work settings using a large representative sample. The present dissertation
is designed to fill this void. Second, what is the effect of interest congruence (i.e., the
compatibility between an individual’s interests and the characteristics of an academic or
work environment) on performance? Two meta-analyses of interest and performance had
3
inconsistent conclusions regarding this question. Van Iddekinge et al. (2011) found that
congruence indices showed weaker predictive validity for job performance and turnover
compared to interest scale scores. Nye et al. (2012), on the other hand, showed that,
consistent with the person-environment fit theory, the correlation between congruence
indices and performance in both academic and work contexts are stronger than that
between interest scale scores and performance. Despite the divergent results, both studies
acknowledged the limitations of using congruence indices for integrating information from
multiple interest scales and suggested polynomial regression analysis as an alternative and
advantageous method for studying the effect of interest-environment fit (Edwards, 1993,
2002). No study to date has applied the polynomial regression and response surface
methodology outlined by Edwards to studying the effect of interest congruence. The
present dissertation is the first attempt to use such methodology to examine the effect of
interest congruence on a variety of performance criteria in education and work settings.
The current dissertation employs data from Project TALENT to address the above
research questions. Project TALENT was a national longitudinal survey conducted on 5% of
the High School students in the United States in 1960. Funded by the United States Office of
Education and developed by the American Institute of Research, it is one of the largest and
most comprehensive longitudinal studies of individual differences and development ever
conducted in the United States. Over 400,000 students from 1,353 schools across the
country provided responses on measures of cognitive abilities, vocational interests,
personality, as well as detailed demographic and sociocultural background information.
The participants were followed up till the present day and were reassessed on a wide range
of educational, occupational, and personal outcomes. The full Project TALENT data set was
4
only made available for research recently, providing an extraordinary opportunity to study
the relationship between vocational interests (among other individual differences
variables) at formative ages and educational and career outcomes later in life.
In the remainder of the introduction, I define the use of “performance” in the
present dissertation from a broad developmental perspective, briefly discuss the
conceptualization and measurement of interests, and hypothesize the linkage between
interests and educational/career attainment beyond cognitive ability and personality.
Performance across the Life Span: A Broad Developmental Perspective
The term performance, as is defined in the present dissertation, refers to the
accomplishment of individuals’ goal-relevant behaviors at a certain point of their schooling
or work (Campbell, Gasser, & Oswald, 1996) as well as the accumulation of individuals’
achievements throughout the course of their education or career (Judge et al., 1999).
Essentially, performance measures are evaluations of individuals’ accomplishments. The
evaluators could be different – school teachers, job supervisors, peers, or society; the
contexts of evaluation could be different – academic setting or work setting; the target of
evaluation and the scale of the target could vary – it can be a very specific behavior, a piece
of work, work over a period of time, attainment as a result of cumulative efforts, or status
gained through a succession of attainments over an extended time span. However, these
different measures of performance all come down to the accomplishment of a behavior, or
the achievements of an individual.
In the present dissertation, performance is viewed as a multi-faceted construct and
a gradual process that unfolds over an individual’s life span. This broad developmental
perspective of performance expands the traditional concept of job performance, typically
5
assessed using supervisor ratings or other objective measures, to encompass multiple
indicators at multiple stages of a person’s life. In the education setting, these indicators
include school grades, persistence, and degree attainment (Kuncel et al., 2004; Kuncel et al.,
2010); and in the work setting, these indicators include income and occupational status
(Judge et al., 1999; Judge et al., 2010). The following indicators of academic and work
performance are used as criteria in the present dissertation: college GPA, college
persistence, degree attainment, income, and occupational prestige. Terms including
educational achievement, career success, educational or career attainment are used as
alternatives to the term performance in the present dissertation.
Vocational Interests: Conceptualization and Motivational Properties
The present dissertation examines measured dispositional interests (for a review on
the distinction between situational interest and dispositional interest and that between
measured and expressed interests, see Savickas, 1999; Su, Rounds, and Armstrong, 2009).
Dispositional interest is trait-like, reflecting a person’s lasting preferences for behaviors,
situations, contexts in which activities occur, and/or the outcomes associated with the
preferred activities (Rounds, 1995). Dispositional interests are associated with attention,
positive emotional states, and direction toward or away from an object, as well as actual
function of behavior (Savickas, 1999). As such, interest directly reflects a person’s goals,
values, and aspirations – which essentially define one’s identity (Hogan & Blake, 1999). As
an inherent part of self-identity, interest serves as the impetus for individuals to navigate
through and function effectively in their environments by influencing the direction, vigor,
and persistence of individuals’ goal-relevant efforts and behaviors (Nye, et al., 2012; Silvia,
2006).
6
Interest directs individuals’ choices and focuses effort on goal achievement in the
chosen fields. John L. Holland (1959, 1997), in his widely-adopted theory of vocational
interests and career choices, suggested that individuals will seek environments that would
allow them to express their interests. Holland proposed that there are six types of
vocational interests (Realistic, Investigative, Artistic, Social, Enterprising, and Conventional,
referred to collectively as the RIASEC model) and six corresponding types of work
environments. He suggested that individuals would be drawn to academic or work
environments that are compatible with their interests. For example, a student who has
strong Investigative interests may be attracted to a major in astronomy and later choose
the job of a research scientist. Indeed, interest has been repeatedly shown to be a robust
predictor of choice of college majors and occupational membership (e.g., Fouad, 1999;
Holland et al., 1994; Strong, 1943). Moreover, interest can direct individuals’ movement
and progress within a major or an occupation by influencing their preferences for certain
activities. If the aforementioned student simultaneously possesses high Enterprising
interests, which involves influencing people, he or she may be oriented to administrative
and managerial activities within a research organization, set goals and prioritize efforts to
attain the position of a project manager or a department director; whereas an individual
lacking such interests may orient his or her effort solely to research and strive to attain
goals that involve development of ideas and publications.
It is essential to emphasize the point that interest influences the initial choice of an
academic or work environment as well as all succeeding choices of sub-environments.
Interest not only influences what major or occupation a person would choose to enter in
the first place, but also affects how the person would prioritize efforts for goal attainment
7
once in a major or an occupation, how the person would behave on particular tasks, and in
turn, what the person could achieve. The influence of interest is better conceptualized as a
continuous process throughout the span of individuals’ career development. In a sense,
interest is about a series of choices that are closely linked to goal-relevant behaviors and
eventual achievement of interest-congruent goals.
In addition to its directional function, interest also energizes effort and determines
the amount of effort an individual would put forth toward a certain goal. People are more
likely to undertake and then work hard on tasks that they find interesting (Sansone &
Smith, 2000). Prominent educational psychologists such as Felix Arnold and John Dewey
have long acknowledged that interest is the key to attention and effort in education
(Arnold, 1910; Dewey, 1913). Interest has been shown repeatedly as a motivational
resource for learning through focused attention, increased cognitive functioning and
affective involvement, and has been linked to better learning outcomes (e.g., Ainley, Hidi, &
Berndorff, 2002; Hidi, 2000; Renninger, Hidi, Krapp, 1992). In the work setting, similarly,
interest affects the level of attention one exerts on a job task, the extent of engagement at
work, and the likelihood for individuals to take initiative on the job, seek out training
opportunities, or take on extra job responsibilities (Fisher & Noble, 2004; Deci & Ryan,
1985). Indeed, studies have shown that interests are associated with the amount of effort
and initiative exhibited on the job both at the between-individual level and at the withinindividual level (Fisher & Noble, 2004; Van Iddekinge et al., 2011).
Last but not the least, interest sustains effort and determines persistence in the
pursuit of goals. Interest entails an enthusiasm and a consciousness that persists during the
interval between the first encounter of a new percept and final attainment of the object
8
(Herbermann, Pace, Pallen, Shahan, & Wynne, 1913). Thus, interest contributes to the
engagement in a task or commitment to a goal until the objective is achieved (Sansone &
Smith, 2000). The persistence could imply either short-term continuance on a particular
task or perseverance for a long-term goal. When individuals are interested in a task, such as
reading texts on a certain topic, they tend to spend more time doing it; persistence on task,
in turn, predicts better task performance (Ainley et al., 2002). More importantly, interest
can fuel and sustain effort over a long period of time even when goals are distant or when
times/situations are difficult. For example, when individuals have interests compatible
with an occupational field or a college major, they are more likely to persist in the
occupation (e.g., Gottfredson & Holland, 1990; Hunter & Hunter, 1984; Morris, 2003) or
attain a degree in such field (e.g., Allen & Robbins, 2010; Webb, Lubinski, & Benbow, 2002).
To summarize, interest has three motivational properties as it influences (a) the
goals that are pursued and the activities that are chosen (direction), (b) the effort that is
put forth (vigor), and (c) the endurance of energy that is invested in the pursuit
(persistence). All three aspects of the motivational process are direct determinants of
performance (Campbell, 1990; Campbell et al., 1996). Because of these motivational
properties, interest is fundamental to the process of career development and the
attainment of educational and occupational goals. Next, I briefly discuss the measurement
of interests before presenting the theoretical rationale and empirical evidence that
interests should predict educational achievement and career success over and above
cognitive ability and personality.
9
Measurement of Interests
Interests are typically measured at three levels of specificity (Hansen, 1984; Rounds,
1995): At the most specific level are occupational interests, followed by basic interests at
the intermediate level, and the broadest, general interests. Occupational interests are
commonly assessed by empirically-keyed scales measuring similarity of interests between
an individual and incumbents employed in a particular occupation. For example, an
occupational scale Psychologists is supposed to distinguish individuals strongly interested
in a career of psychological research and service from the general population. Basic
interests are measured at the level of generality between interest in a certain occupation
and interest in a broad area. Basic Interest Scales (BIS) characterize shared properties of
activities (e.g., Teaching, Selling, and Writing) and are often implied in the object of interest
(e.g., Mathematics, Physical science, and Engineering). General interests are broad interest
areas, such as Business or Science, and are usually measured with items that encompass a
heterogeneous group of occupations and activities. They sit at the highest level of the
interest hierarchy and are the fundamental dimensions for describing the vocational
interest domain.
Holland’s (1959, 1997) RIASEC interest model is the most widely embraced and
researched taxonomy of general interests. However, models of general interests do not
have to be theoretically derived but can also be developed using a factor analytic approach.
Research has shown that Holland interest types are better construed as independent
factors and a factorial approach to interest measurement is recommended to capture
individual differences in the interest domain (Tay, Su, & Rounds, 2011). The search for
fundamental interest factors began with Thurstone’s (1931) factor analysis of Strong’s
10
occupational scales and was continued by a series of landmark studies including Torr
(1953), Guilford, Christensen, Bond, and Sutton (1954), and Jackson (1977). Despite some
divergence in the factor solution and labeling among these studies, several general interest
factors emerge repeatedly (Liao, Jin, Su, Tay, & Rounds, 2012, August): (a) interest in
Science, (b) interest in Language or Artistic interests, (c) interest in People, (d) interest in
Business, and (e) interest in Mechanics and Nature. The present dissertation derived
general interest factors from the Project TALENT Interest Inventory because general
interest factors provide a parsimonious and comprehensive solution for mapping the
vocational interest domain. Compared to basic interests that are more particular, general
interest factors facilitate the interpretation of the effects of interests and interest
congruence. Furthermore, predictive validity is likely to be impacted by the specificity of
predictor and criterion; matching specificity of predictor and criterion will lead to the
highest predictive validity (Ajzen & Fishbein, 1977; Campbell, 1990). The breadth of
general interest factors is most suitable for predicting broad outcomes such as educational
attainment and career success.
Predictive Validity of Interests: Beyond Ability and Personality
Cognitive ability has received extensive attention in the field of industrial-
organizational psychology because it has been shown to be predictive of job performance
and career success through a number of studies (e.g., Hunter & Hunter, 1984; Judge et al.,
1999; Schmidt & Hunter, 1998). The past two decades have also witnessed a resurgence of
personality traits, as well as their facets and composites, as predictors for performance
(e.g., Barrick & Mount, 1991; Dudley, Orvis, Lebiecki, & Cortina, 2006; Judge & Bono, 2001).
Interests, in contrast, were largely ignored in the personnel selection literature partly due
11
to the assumption that interests only predict educational and career choices and are not
good predictors of performance within a chosen field. Findings from two recent meta-
analyses have questioned this assumption (Nye et al., 2012; Van Iddenkinge et al., 2011).
Van Iddekinge et al. (2011) examined 74 studies and 141 independent samples and found
interests to be predictors for job performance, training performance, turnover intentions,
and actual turnover behavior. The validity estimates from the Van Iddekinge et al. (2011)
study are significantly higher than those reported in a previous meta-analysis (Hunter &
Hunter, 1984). For example, the overall meta-analytic correlation between interests and
job performance is .14; when regression-weighted composites of multiple interest scale
scores were created to predict job performance, the unshrunken and shrunken validities
for the interest composites are .37 and .18, respectively; moreover, among interest scales
of different specificity, occupational interest scales (most specific interest scales) have the
highest predictive validity of .23 for job performance (specific outcome at the occupation
level). Nye et al. (2012) conducted a meta-analysis of 60 studies and found that interests
are related to performance and persistence in both work and academic contexts. Predictive
validities are .20 and .23 for work and academic settings, respectively. More importantly,
when interest congruence was used instead of interest scale scores, predictive validities of
interests for work and academic performance rose to .36 and .32, respectively. The effect
sizes for the interest-performance correlation found in these two meta-analyses are in the
same range as reported correlations between conscientiousness and performance (Barrick
& Mount, 1991; Judge et al., 1999). These meta-analyses provide critical evidence for the
predictive power of interests for performance criteria in both work and academic settings.
12
Despite recent research endeavors to establish the role of interest for predicting
performance, some additional questions have yet to be addressed: How large are the effects
of interests compared to those of cognitive ability and personality when they are used for
predicting the same criteria? More importantly, do interests have predictive validity over
and beyond that of cognitive ability and personality? Lubinski (2000) argued that an
independent approach to identifying key person determinants of achievement, i.e., focusing
on a single class of predictors such as either personality or interests in isolation from each
other, is suboptimal for providing a comprehensive solution for assessments used in
personnel selection and promotion. Studies that take into account of the communalities
among the three individual difference domains are needed.
Theoretically, interest should predict performance beyond cognitive ability and
personality because the mechanism for interest to affect behaviors is different from that for
ability and for personality. Cognitive ability predicts performance because people “can do;”
it contributes to performance primarily through its influence on job-related knowledge
(Campbell, 1990; Schmidt, Hunter, & Outerbridge, 1986). Personality predicts performance
because people “will do” or “typically do;” the effect of personality on performance is
thought to be mediated through motivational variables, such as goal-setting (Barrick,
Mount, & Strauss, 1993) and striving for status and achievement (Barrick, Stewart, &
Piotrowski, 2002). Interest predicts performance because people “want to do;” interests
affect performance through the influence on all three aspects of the motivational process:
direction, vigor, and persistence (Nye et al., 2012; Silvia, 2006). Although interests and
personality are both linked to motivation, they are conceptually distinct constructs.
Personality, as Hogan and Blake (1999) argued, directly reflects a person’s reputation, or
13
how he or she characteristically behave, whereas interest directly reflect one’s identity, or
what they desire and who they aspire to be deep down. In his neo-socioanalytic model of
personality, Roberts (e.g., Roberts & Wood, 2006) also argued that interests belong to the
domain of motives and values, which is separate from the domain of personality traits.
Empirically, overlap between interests and cognitive abilities and personality were found
to be in general modest in size (e.g., Ackerman & Heggestad, 1997; Barrick, Mount, & Gupta,
2003). It is expected that interest will have incremental validity over and beyond cognitive
ability and personality.
Only a limited number of studies to date used interest in conjunction with ability
and personality and compared their effects on performance criteria. Evidence regarding
the incremental validity for interests over and beyond cognitive ability and personality is
even scarcer (merely four studies). Table 1 provides a summary of studies that examined
interests simultaneously with cognitive ability and personality for predicting academic or
work performance. The four studies that examined incremental validity of interest over
ability and personality are noted with an asterisk.
A few conclusions are evident from the studies that examined incremental validity
of interests: First of all, these four studies provided some preliminary evidence that
interests have predictive validity beyond cognitive ability and personality for a range of
specific and broad performance criteria. Second, all four studies used samples that were
from a single occupational field. Participants in Dyer’s (1987) study were 250 nursing
students; Gellatly, Paunonen, Meyer, Jackson, & Goffin (1991) surveyed 59 managers from a
food-service organization; both Black (1999) and Van Iddekinge et al. (2011) focused on
military samples. The generalizability of results from these samples is limited. Third, all
14
four studies used a concurrent design or a design where predictors and criteria were
measured only shortly apart from each other (e.g., 8 weeks in Black, 1999). Such design
may inflate the correlations between predictors and criteria and the results may have
limited generalizability to a context where predictions need to be made for distal
educational or occupational outcomes. Thus, the necessity for a study that examines the
incremental validity of interest over and beyond ability and personality for academic and
work performance with a large, representative sample is warranted. Surprisingly, given
the rich selection of both predictor and criterion variables available from Project TALENT
and the longitudinal design perfectly suited for examining the aforementioned research
question, no such study has ever been conducted with the Project TALENT data. The
present dissertation is designed to examine the incremental validity of interest beyond
cognitive ability and personality using the Project TALENT data set.
Hypothesis 1. Interests have incremental validity for college grades, college
persistence, degree attainment, occupational prestige, and income over cognitive abilities
and personality.
Effect of Interest Congruence on Performance: The Polynomial Regression Approach
According to Holland’s (1959, 1997) theory, individuals are drawn to environments
that are compatible with their vocational interests. More importantly, Holland proposed
that, the congruence between individuals’ interests and their environment influences their
attitudes, behaviors, and performance outcomes: high compatibility between interests and
an environment will lead to higher satisfaction, larger likelihood to persist in the
environment, and greater success. Holland’s theory on interest congruence is consistent
with the propositions and findings from the broader person-environment fit theoretical
15
framework, which link the degree of fit or compatibility between a person and his or her
environment with positive work outcomes (e.g, Kristof-Brown, Zimmerman, & Johnson,
2005). Many researchers have questioned the usefulness and validity of interest scale
scores alone and argued that interest congruence is a superior predictor (e.g., Gottfredson
& Jones, 1993; Hirschi, 2009; Nye et al., 2012; Prediger, 1998; Warwas, Nagy, Watermann,
& Hasselhorn, 2009). It is expected that interest will predict performance criteria only
when it is compatible with or relevant to an occupational field or a college major, but not
when it is incompatible with the occupation or major. For example, a strong interest in
science will be a good predictor for the performance of scientists, and yet is not expected to
be associated with performance in business.
Despite the theoretical foundation and empirical evidence supporting the
association between interest congruence and performance, past studies that examined the
predictive validity of interest congruence compared to interest scale scores did not always
yield favorable results (e.g., Van Iddekinge et al., 2011). Questions about interest
congruence, or broadly speaking, of person-environment fit measures, center on the use of
a single congruence index to quantify the degree of fit between a person and an
environment. Typically, an index for interest congruence is developed by first deriving a
profile for an individual based on the relative strengths of their interests in several areas
and a profile for an occupation based on the environmental characteristics, and next,
computing the difference or similarity between these two profiles (for a full review of
various congruence indices, see Brown and Gore, 1994, or Camp and Chartrand, 1992).
Congruence indices have often been criticized for various reasons (Edwards, 1993; Tinsley,
2000). For example, they reduce a multi-dimensional problem to a single index and
16
obscure important information about the interest factors being examined. Consequently,
results may be difficult to interpret and inappropriate conclusions may be drawn. It is
possibly suboptimal to examine interest congruence using congruence indices and
inconsistent results may result from their limitations.
To remedy the aforementioned problems, Edwards (2002; Edward & Parry, 1993)
has suggested using polynomial regression analysis as an alternative to congruence indices
and response surface methodology as a general strategy to study the issue of congruence in
organizational psychology. The polynomial regression technique is based on the premise
that person and environment measures represent distinct constructs and that any main
effects of person and environment measures and any interaction effects between these
constructs should be tested empirically. Instead of computing a single congruence score
indexing the difference or similarity between person and environment and regressing a
criterion measure onto this index, this technique employs linear or quadratic polynomial
regression equations of individual and organizational characteristics. These equations are
generalizations of congruence indices that represent sums of algebraic, absolute, or
squared difference scores, as they take the same mathematical form as congruence indices
without imposing unwarranted assumptions or constraints on the coefficient estimates. It
provides an advantageous method for studying the effect of person-environment fit.
Unfortunately, no study in the interest literature to date has applied this approach. The
present dissertation applies the polynomial regression analysis and response surface
methodology to examining the effect of interest congruence with an occupational field or a
college major on performance criteria.
17
Hypothesis 2a. Occupational field moderates the relationship between interests and
performance in the work setting, such that the effect of interests on performance will be
stronger in compatible or relevant occupational fields.
Hypothesis 2b. College major moderates the relationship between interests and
performance in the academic setting, such that the effect of interests on performance will
be stronger in compatible or relevant majors.
Interests Mediate Gender Effect on Career Success
In the study of career attainment, the effect of gender is a significant issue. Since the
1970s, differential career attainment of women and men has been the topic of heated
national debates among researchers and educators (e.g., Ceci & Williams, 2007; Ceci,
Williams, & Barnett, 2009; Watt & Eccles, 2008). Women are underrepresented in many
high-status and remunerative occupations, particularly occupations in the fields of science,
technology, engineering, and mathematics (STEM); and even when they are in the same
occupations as men, women are less likely to attain equal levels of position or gain equal
earnings as men (Weinberger & Kuhn, 2010). Despite the improvement of women’s
representation in many occupational fields and the narrowing of the gender wage gap
across generations, the issue of career attainment disparity between women and men still
persists and has been the center of concern in the media and among the public till this date
(e.g., Dillon & Rimer, 2005, Janurary 19; Pinker, 2008; Tierney, 2008, July 15).
Among the individual differences variables that potentially explain women’s
underrepresentation and lesser attainment compared to men in high status and
remunerative occupational fields, the cognitive ability domain, especially abilities related
to mathematics and spatial orientation, has been heavily emphasized, whereas the role of
18
interest has largely been ignored (see Ceci et al., 2009 for a review). Early research on
career attainment disparity between two genders viewed the issue as a problem that
pertains to mathematics achievement. For example, the STEM fields, where the issue of
women’s underrepresentation and underachievement is the most salient, are typically
described as “math-intensive” fields (Ceci et al., 2009). Women’s lower enrollment in the
study of advanced mathematics was thought to be the reason that precludes them from
entering occupations requiring mathematical competence (Chipman, Brush, & Wilson,
1985; Gallagher & Kaufman, 2005). More recently, spatial ability was found to be a salient
predictor for advanced degree attainment and occupational achievement in STEM and was
proposed to be a major contributor for the gender disparity in these fields (e.g., Wai,
Lubinski, & Benbow, 2009). Nonetheless, it has been shown that women and men differ
very little in their average math ability as well as other types of cognitive abilities (e.g.,
Hyde & Linn, 2006; Hedges & Nowell, 1995). Moreover, even though there is evidence that
the variances of male abilities are larger than those of females and that more males than
females were found among mathematically or spatially gifted individuals, a series of
longitudinal studies with intellectually precocious students showed that, even among the
profoundly talented, their career trajectories and attainment still diverged and were driven
mainly by their preferences instead of abilities (Ferriman, Lubinski, & Benbow, 2009;
Robertson, Smeets, Lubinski & Benbow, 2010). Ceci et al. (2009), after reviewing evidence
for a comprehensive range of claimed causes for women’s underrepresentation in science,
concluded that preferences rather than aptitudes constitute the most powerful explanatory
factor for career attainment disparity between two genders.
19
Despite the importance of preferences, little research has examined the power of
vocational interests in explaining the gender achievement gap. In contrast to the cognitive
ability and personality domains, interest was found to be a domain where large gender
differences exist (Su et al., 2009). Su et al. (2009) reported in their meta-analysis an effect
size of d = .93 for gender difference in the Things-People interest dimension, with men
having stronger Things interests and women having strong People interests. Such large
gender differences in interests matter for differential career achievement of women and
men, as shown by a number of studies (e.g., Jacobs, Davis-Kean, Bleeker, Eccles, &
Malanchuk, 2005; Meece, Parsons, Kaczala, Goff, & Futterman, 1982; Wise, Steel, &
Macdonald, 1979). The present dissertation, therefore, examined whether interests
mediate the effect of gender on criteria of career success. The mediation effects of cognitive
ability and personality were also examined in the present dissertation, yet no specific
hypothesis was made.
Hypothesis 3. Interests partially mediate the effect of gender on career success.
Family Characteristics Moderate Gender Effect on Career Success
In addition to the influence of vocational interests, family characteristics, especially
the presence and number of children, constitute another salient factor that affects men and
women’s career attainment. Fulfilling child rearing responsibilities requires a considerable
amount of time and effort. Children present in the household may increase demands in the
family sphere and limit available resources to devote to work, which in turn influence
career attainment outcomes. Many studies have shown the collision between parenthood
and career success (Ceci et al., 2009). However, it has been found that having children has a
particularly serious effect on women’s careers (Leslie, 2007, March). For example, Leslie
20
(2007, March) analyzed four surveys by the National Center for Education Statistics and
showed that the number of children has a negative linear relation with the number of hours
worked per week for women at their academic jobs; in comparison, the relationship was
positive for men. Female academics with children are also less likely to be on the tenure
track (Jacobs & Winslow, 2004; Mason & Goulden, 2004), more likely to drop off the tenure
track for part-time work until their children get older (Leboy, 2007, September 21), more
likely to be at small colleges and/or in adjunct positions (Leslie, 2007, March), and less
likely to be promoted compared to women without children (Ginther & Kahn, 2006). These
discrepant patterns of career attainment between women with and without children have
also been found for women in other occupational fields (e.g., Mroz, 1987), but have not
been observed for men.
Hakim (2005; 2006) used life preferences or work-family balance goals to explain
the differential effect of having (more) children on career attainment for women and men.
Compared to men, women are more likely to endorse goals that try to balance work and
family, whereas men are more likely to set agentic goals that emphasize career. When they
have children, women have greater preference for a “home-centered” or a “balanced”
lifestyle, whereas men take a more career-focused perspective. Such gender differences in
life preferences have been found in the general population as well as in groups with
outstanding intellectual abilities (Ferriman et al. 2009; Lubinski, Benbow, Shea, Eftekhari-
Sanjani, & Halvorson, 2001). Different goals and lifestyles for women and men are likely to
influence their work productivity and may have an accrued effect on their career
attainment over time. This effect may be small when individuals do not have children or
have a small number of children, but is expected to be stronger as the number of children
21
increase. Therefore, the present dissertation examines the extent to which family
characteristics, as indicated by the number of children, moderate the effect of gender on
career success.
Hypothesis 4. Number of children moderates the relationship between gender and
career success, such that the effect of gender on career success is stronger as the number of
children increases.
Degree Attainment Mediates the Effect of Individual Differences on Career Success
Education has long been noted to contribute positively to work outcomes such as
income and occupational status (Card, 2001; Mincer, 1974; Welland, 1976; Willis & Rosen,
1979). In some occupational fields, such as law and medicine, an advanced degree or
specific educational credentials in the field is necessary for entry into the field; in other
occupational fields, level of education is usually associated with pay and promotion. Past
research has also provided substantial evidence that individual differences in cognitive
ability, personality, and interests predict the level of educational attainment. The present
dissertation, therefore, examines whether degree attainment mediates the relationship
between individual differences variables and career success.
Hypothesis 5a. Degree attainment partially mediates the relationship between
interests and career success.
Hypothesis 5b. Degree attainment partially mediates the relationship between
cognitive abilities and career success.
Hypothesis 5c. Degree attainment partially mediates the relationship between
personality and career success.
22
Current Study
To summarize, the primary goal of the present dissertation is to examine the
incremental validity of vocational interests beyond cognitive abilities and personality for
criteria of educational achievement and career success, namely, college GPA, college
persistence, degree attainment, income, and occupational prestige. The incremental
validity of interests was examined using data from Project TALENT (Wise et al., 1979).
Polynomial regression analysis and response surface methodology (Edwards, 2002;
Edwards & Parry, 1993) was applied to examine the effect of interest congruence on the
criteria. It is expected that college major and occupational field moderate the effect of
interests on educational achievement and career success. The predictive validity of
interests is expected to be higher for success in compatible or relevant majors and
occupations.
The relationship between interests, cognitive abilities, personality and educational
achievement is represented in Figure 1. A moderated mediation model is developed that
delineates the relationship between interests, cognitive abilities, personality, and career
success (see Figure 2). The effect of gender on career success is expected to be partially
mediated through interests. Number of children is expected to moderate the direct effect of
gender on career success. Degree attainment is expected to partially mediate the
relationship between interests, abilities, personality and career success.
23
Participants
METHOD
The data come from Project TALENT (Wise et al., 1979), a national longitudinal
study developed by the American Institutes for Research (AIR). The original survey was
conducted in 1960 with 5% of America’s high school students from 1,353 public, private,
and parochial senior high schools throughout the country. Over 440,000 students in grade
9 through 12 participated in two full days or four half days of testing. The battery of tests
and questions were extensive, examining students’ knowledge and aptitudes in a range of
domains, personality, interests in various occupations and activities, aspirations for future
education and vocation, family background, and personal experiences. After the original
testing, the participants were re-contacted via mail three times, at 1st, 5th, and 11th years
after their high school graduation, and were surveyed again about their educational,
occupational, and personal developments. The response rates to the follow-up studies were
51.4 percent for the one-year follow-up, 35.3 percent for the 5-year follow-up, and 25.8
percent for the 11-year follow-up. The present dissertation used participants’ responses
from the original study and the first three follow-up surveys. Table 2 presents the number
of participants at each data collection point, in total, by gender, and by race/ethnicity.
Project TALENT is the only nationally representative longitudinal study of such
large scale. It is an ideal data set for studying how interests, along with abilities and
personality, impact academic, work, and life outcomes later in life.
First of all, it includes a comprehensive set of individual differences variables,
covering the domains of cognitive abilities, personality, and interests. Importantly, the
interest inventory in Project TALENT includes a total of 205 items that measure interests in
24
a wide range of occupations and a variety of activities. Such an extensive assessment of
interests is not available from any other national longitudinal studies, such as the National
Longitudinal Survey (NLS; Borus, 1981; Sproat, Churchill, & Sheets, 1985) and the High
School and Beyond project (HSB; Jones et al., 1983). The availability of exhaustive
measures of interests, cognitive abilities, and personality suits the purpose of examining
incremental validity of interests over and beyond abilities and personality.
Second, Project TALENT is a longitudinal survey that tracked participants from their
high school years till 11 years after high school graduation. The longitudinal design of
Project TALENT has clear advantages over cross-sectional studies, as it allows for studying
the effects of individual differences at an early age on performance outcomes at later life
stages (Tharenou, 1997). As noted previously, the present dissertation takes a broad,
developmental perspective on performance and plans to examine the incremental validity
of interests for a range of performance criteria over time in both academic and work
settings. The three follow-up studies in Project TALENT captured participant’s educational
and career outcomes at multiple critical time points of life (first year in college, one year
after graduation from college, and after years in the work force, respectively) and suited
the goal of the present dissertation for examining performance outcomes over time.
Third, Project TALENT provides one of the largest and most representative samples,
with individuals entering diverse occupations and developing different career trajectories
over time. Such a large sample that represents a sufficient number of individuals in each
occupational field and each college major is necessary for examining the effect of interest
congruence on performance criteria.
25
Measures
Below I first delineate measures for individual differences (i.e., cognitive abilities,
personality, and interests), followed by measures for criteria of educational achievement
and career success, and lastly, measures for proposed moderators and mediators. For each
construct, I describe the variables available in the Project TALENT data set, and then
discuss analyses or transformations conducted with the variables to form the measures
used in the present dissertation. An illustration of data collection point at which each
demographic, predictor, outcome, and moderator variable was measured can be found in
Table 3. Table 4 shows the descriptive statistics and correlations of all the demographic
variables, abilities, personality, interests, and outcomes.
Individual Differences Measures
Cognitive Abilities. The Project TALENT original survey contains a set of scales that
represent different content domains of cognitive abilities, including verbal abilities (e.g.,
Vocabulary scale, Reading Comprehension scale), quantitative abilities (e.g., Arithmetic
Reasoning, Advanced Mathematics), visualization and spatial abilities (e.g., Visualization in
3D, Mechanical Reasoning), memory and retrieval (e.g., Memory for Words, Memory for
Sentences), and perceptual speed and accuracy (e.g., Clerical Checking, Table Reading).
The structure of cognitive ability measures in Project TALENT, as shown through
factor analysis, resembles Carroll’s (1993) three-stratum model of intelligence, where
specific abilities clustered to form the above second-order ability domains, and a higher-
order general ability factor explains a large proportion of the variances among these ability
domains. An alternative and simpler model, the radex organization of cognitive abilities
originally proposed by Guttman (1957) and developed by Snow and colleagues (1978), has
26
been applied to the Project TALENT data (e.g., Wai et al., 2009). The radex model organizes
around three correlated cognitive ability domains: verbal, numerical, and spatial. The
center of the radex represents communality across these three domains or a higher order
construct of g, general intelligence, whereas towards the periphery of the radex lie more
specific abilities. The radex model has been shown to be theoretically and empirically
parallel to the hierarchical model of cognitive abilities (Marshalek, Lohman, & Snow, 1983).
Therefore, composite measures for these three abilities domains – verbal, quantitative, and
spatial abilities – were developed for the dissertation given the conceptual simplicity of this
organization of cognitive abilities, and more importantly, the importance of these three
ability domains for predicting educational achievement and career success (e.g.,
Humphreys, Lubinski, & Yao, 1993; Kuncel, Hezlett, & Ones, 2001; Wai et al., 2009). Unit
weighting was used in constructing the three ability composites so that every ability scale
component had unit variance and no scale was over-weighted.
The verbal ability composite consists of three scales:
1. Vocabulary (30 items that measure general knowledge of words).
2. English Composite (113 items measuring capitalization, punctuation, spelling,
usage, and effective expression in English).
3. Reading Comprehension (48 items measuring the comprehension of written text
covering a broad range of topics).
The math ability composite consists of four scales:
1. Mathematics Information (23 items measuring knowledge of math definitions
and notation).
27
2. Arithmetic Reasoning (16 items measuring the reasoning ability needed to solve
basic arithmetic items).
3. Introductory Mathematics (24 items measuring all forms of math knowledge
taught through the 9th grade).
4. Advanced Mathematics (14 items covering math topics taught in Grade 10 to 12
college preparatory courses, such as algebra, plane and solid geometry, probability, logic,
logarithms, and basic calculus).
The spatial ability composite consists of four scales:
1. Two-Dimensional Spatial Visualization (24 items measuring the ability to
visualize two-dimensional figures when they were rotated or flipped in a plane).
2. Three-Dimensional Spatial Visualization (16 items measuring the ability to
visualize two-dimensional figures after they had been folded into three-dimensional
figures).
3. Mechanical Reasoning (20 items measuring the ability to deduce relationships
between gears, pulleys, and springs as well as knowledge of the effects of basic physical
forces, such as gravity).
4. Abstract Reasoning (15 items constituting a nonverbal measure of finding logical
relationships in sophisticated figure patterns).
Intercorrelations among the ability composite measures were .64, .60, and .76 for
math-spatial, verbal-spatial, math-verbal, respectively (see Table 4). This correlation
pattern resembles what was found in previous studies (e.g., Cooley & Lohnes, 1968; Wai et
al., 2009).
28
Personality. The Project TALENT original survey contains ten personality scales:
Sociability, a 12-item scale that measures preference for spending time with a group or
alone; Social Sensitivity, a 9-item scale that measures sympathy and thoughtfulness in social
situations; Impulsiveness, a 9-item scale that measures tendency to do things or make
decisions hastily; Vigor, a 7-item scale that measures the level of energy; Calmness, a 9-item
scale that measures even-temperedness and self-control; Tidiness, an 11-item scale that
measures orderliness with tasks, surroundings, and personal appearance; Culture, a 10item scale that measures liking and appreciation for aesthetic qualities or values;
Leadership, a 5-item scale that measures inclination for taking charge; Self-confidence, a 12-
item scale that measures sense of self-assurance; and Mature Personality, a 24-item scale
that measures industriousness, perseverance, and dependability.
A validation study (Roberts, personal communication, April 17, 2012) was
conducted to examine the relationship between the Project TALENT personality scales and
measures of the Big Five personality traits (Goldberg, 1999; John, Donahue, & Kentle,
1991), as the Big Five personality factors provide a fundamental and coherent structure for
organizing personality traits and has been shown to be an immensely useful taxonomy for
studying the relationship between personality and many important work outcomes, such
as job performance and career success (e.g., Barrick & Mount, 1991; Tett, Jackson, &
Rothstein, 1991). Five of the ten Project TALENT personality scales were shown to tap into
the Big Five: Mature Personality as well as Tidiness both measure facets of
Conscientiousness; Social Sensitivity measures Agreeableness; Sociability measures
Extraversion; and Self Confidence measures Emotional Stability. None of the Project
TALENT personality scales represented Openness to Experience. Therefore, the above five
29
personality scales are used as indicators for the Big Five to predict performance criteria in
the present dissertation, with scores of the Mature Personality scale and the Tidiness scale
combined to form a composite score of Conscientiousness. The rest of the personality
scales, namely Impulsiveness, Calmness, Vigor, Leadership, and Culture, were not closely
related to the Big Five personality factors. Nonetheless, these scales offer additional
information about participants’ personality and thus are also included as predictors for the
performance criteria. Intercorrelations among the personality scales range from .12
between Impulsiveness and Self Confidence to .65 between Culture and the Conscientiousness
composite (see Table 4).
Interests. The interest inventory from Project TALENT original survey consisted of
two parts with 205 items in total. Part I of the Project TALENT Interest Inventory included
122 items that covered a wide range of occupations (e.g., automobile mechanic, college
president) and asked participants to rate how much they would like that kind of work, on a
scale from 1 to 5 (1 = “dislike very much” and 5 = “like very much”); Part II of the Project
TALENT Interest Inventory asked participants how much they would like each of 83
activities (e.g., work on arithmetic problems, invest money) on the same scale. For the
present dissertation, general interest factors are derived from these items following a top-
down factor analytic procedure recommended by Goldberg (2006). This procedure yields a
factor hierarchy of interests with two broad interest factors on the top followed by an
increasing number of interest factors at lower, more specific levels (see Figure 3 for the
Project TALENT interest factor hierarchy). The seven-factor solution is most consistent
with findings from previous factor analytic studies of interests and was chosen as the best
representation of the Project TALENT interest factor structure.
30
Five of the seven interest factors conform to the findings from early factor analytic
studies by Thurstone (1931), Torr (1953), Guilford et al. (1954).
Things: Interests in working with things, tools, or in the outdoors;
Artistic: Interests in creative expression, including fine and performing arts;
Science: Interests in sciences and research, including natural sciences, medical
sciences, and mathematics;
People: Interests in serving and helping people;
Business: Interest in business activities, including business management, sales,
finances and accountancy.
Two novel factors that were not consistently found in previous studies emerged
from current analysis of the Project TALENT Interest Inventory:
Leadership: Interest in leadership positions in the government or other public
sector bodies. This factor resembles the Leadership factor found in Torr’s (1953) factor
analytic study and, to some extent, the Enterprising factor found by Jackson (1977). A
recent factor analytic study (Liao et al., 2012) has also found a Leadership interest factor
distinct from Business interests – the latter represents interests in the pursuit of economic
outcomes regardless of the role one takes in the corporate setting whereas the former
represents interests in the pursuit of leadership roles regardless of personal economic
outcomes.
Athletic: Interest in various sports, such as baseball, track, swimming, and fishing.
This factor was not reported in previous factor analytic studies of interests and is found
due to the inclusion of particular items in the Project TALENT Interest Inventory. Because
31
this interest factor represents leisure rather than vocational interests, it is dropped from
the analyses for the present dissertation.
Items with factor loadings larger than .40 and cross loadings smaller than .35 were
selected to construct the general interest factors. Mean scores are taken from all eligible
items loaded on each factor to represent participants’ interests on that general interest
factor. Intercorrelations among the general interest factors ranged from -.25 between
Things and People interests to .59 between Business and Leadership interests (see Table
4).
Performance Criterion Measures
1st Year Grade in College. One year after their high school graduation, participants
were asked in the follow-up survey if they had attended college since leaving high school
and, if so, participants were asked to report their letter grade (A, B, C, D, or F) for courses in
different subjects as well as average grade for all courses thus far. Letter grades were then
translated into numeric codes (4, 3, 2, 1, or 0) for data analysis. The average grade variable
was used as an indicator for performance in the 1st year of college.
College Grades. In the 5th year follow-up survey, participants were asked to report
their average letter grade in major and in all subjects (A, A-, B+, till F). Letter grades were
then translated into numeric codes (12, 10, 9, till 1). Both grade in major and overall grade
were used as indicators for performance in college.
College Persistence. In the 11th year follow-up survey, participants were asked if
they attended college continuously until they obtained a bachelor’s degree. Participants
also reported their status as college students (e.g., full-time, part-time) if they continuously
attended college, reported their reasons if they had dropped out, or indicated that they
32
were still enrolled as undergraduate students. For the present dissertation, a criterion
variable of college persistence was developed, with participants who dropped out from
college coded as 0, participants who persisted until they obtained a degree coded as 1, and
participants who were still students excluded from the analyses.
Degree Attainment. An important goal for the Project TALENT study was to
understand the educational development of American students and what early variables
contribute to educational achievement later in life. Thus, in all follow-up studies, a large
proportion of the survey was devoted to assessing participants’ educational experiences,
including their high school education, college attendance, and post-college education.
Based on participants’ answers to the questionnaire, an amount-of-education variable was
coded by the Project TALENT staff, ranging from the lowest educational attainment of
“High school dropout: completed grade 8 but not grade 9” (coded as 1) to the highest of
“Doctoral or law degree (including LL.B. as well as J.D.)” (coded as 12). As an indicator for
their eventual degree attainment, the amount-of-education variable from the 11th year
follow-up was used in the present dissertation.
Income. In all follow-up studies, participants were asked to report their rate of pay
per month, per week, or per hour. The responses were then coded by Project TALENT staff
to estimated annual earnings and converted into a logarithmic annual earning variable.
Logarithmic income is typically used in studies with income as criterion variable because
income values span over several orders of magnitude with outliners at the extremely high
income end. Instead of using linear income values, using logarithmic income is more
appropriate for regression analysis and usually facilitates the interpretation of and
inference from the results. The present dissertation thus used the logarithm of reported
33
annual earning at the 11th year follow-up study as one indicator for career success. To
allow reasonable comparison, only participants who reported that they were working full
time at the 11th year follow-up are included in the analyses with income as the criterion.
Logarithmic income of the participants ranged from 2 to 94.
Occupational Prestige. Occupational prestige refers to the social status of an
occupation as regarded by members of a society and reflects the general standing of
individuals in that occupation within the society. Hauser and Warren (1997) reviewed
studies on a wide range of occupational prestige indices and reported that occupational
prestige can be coded reliably and it appears rather stable across time. The most widely
used and validated measure of occupational prestige, the Duncan Socioeconomic Index
(SEI; Duncan, 1961), was developed based on the classic 1947 North and Hatt prestige
study carried out at the Nation Opinion Research Center (Reiss, 1961) and was mapped
onto the 1950 Census occupational classification. Other indices for occupational prestige,
most notably the Hollingshead Index of Social Position (Hollingshead, 1965), were found to
have reasonably high correlations (above .80) with Duncan’s SEI. Several major updates on
the SEI have been conducted to map SEI scores to later census occupational classifications
(Blau & Duncan, 1967; Nakao & Treas, 1994; Stevens & Featherman, 1981). Specifically,
Stevens and Featherman (1981) used the prestige scores from National Opinion Research
Center surveys in the 1960s (Siegel, 1971) and mapped them to the 1970 census
occupation titles (U.S. Bureau of Census, 1971), with scores ranging from 13.88 (Loom
fixers) to 90.45 (Law teachers). Also, unlike Duncan’s SEI, which were developed primarily
based on white males, Steven and Featherman’s revised SEI provided prestige scores using
regressors pertained to the characteristics of all workers (TSEI2) aside from those
34
developed based on only male workers (MSEI2). Therefore, Stevens and Featherman’s
TSEI2 is used in the present dissertation. The 1970 census occupation codes are available
for all the reported occupational fields from the Project TALENT 11th year follow-up study,
based on which TSEI2 scores are assigned.
Moderator and Mediator Measures
Occupational Characteristics. Each of the three Project TALENT follow-up studies
asked the participants to report the job that they held at the time. Only responses from the
11th year follow-up were used for the present dissertation to examine work outcomes
because, at the time of the first two follow-up studies, a sizable proportion of the
participants were still pursuing advanced education so the sample was not fully
representative of the population that was in the world of work. In the follow-up surveys,
participants provided free responses to two questions on what their jobs were called and
what they did on the job. These answers were then coded by Project TALENT staff into 250
3-digit occupation codes. For the present dissertation, these occupational codes were used
to form occupational groups that had at least 50 people, with only a few exceptions (e.g.,
FBI and Secret Service, N = 22; Optometrist, N = 31). Because environments are defined and
characterized by the kinds of people in it (Schneider, 1987), for each occupational group,
group members’ average score in each of the six interest factors was computed as the
group’s characteristic of how much it would allow job incumbents to express their interest
in that area. To avoid dependence between the characteristics of an occupational group and
the interests of an individual within that group, the occupational characteristic entries for
each participant were calculated as the within-group mean without that participant. Table
35
5 listed the occupational groups and occupational characteristics in Things, Artistic, People,
Science, Business, and Leadership for each group.
Major Characteristics. The Project TALENT 5th year and 11th year follow-up studies
asked the participants to report which of the 43 listed areas was or is their major in college
as an undergraduate student. Similar to the procedure for developing occupational
characteristics, major characteristic entries for each participant were calculated as the
within-major average interest scores without that participant. Table 5 listed the major
characteristics in Things, Artistic, People, Science, Business, and Leadership for each of the
43 majors at the 5th year follow-up, which were used for examining the effect of interest
congruence on overall and major grades in college. Table 6 listed major characteristics at
the 11th year follow-up, which were used for examining the effect of interest congruence on
college persistence. Correlations between 5th year major characteristics and 11th year
major characteristics in corresponding interest areas were high, ranging from .88 to .95.
Degree Attainment. The same variable of degree attainment from the 11th year
follow-up used as a criterion of educational achievement is used as a mediator for
predicting career success.
Family Characteristics. A substantial proportion of the Project TALENT follow-up
questionnaire was designed to obtain information about participants’ personal and family
developments. Participants responded to an extensive range of questions regarding their
marital and family situation as well as intersection between family life and work life, such
as marital status, number of living children, whether they stopped or plan to stop working
for family reasons. As child rearing involves a considerable amount of responsibilities and
has profound impact on education and career pursuits, the total number of children was
36
selected as an objective indicator for family characteristics in the present dissertation. The
total number of children at the 11th year follow-up was used as a moderator for predicting
career success.
Demographic Measures
Four demographic measures were included in the analyses as there are well-
documented effects of these variables on educational achievement and career success
outcomes: gender, race/ethnicity, socioeconomic index, and cohort. These variables were
available from the Project TALENT original study. Gender was coded as male = 0, female =
1. Race/ethnicity was represented using five dummy codes: White/Caucasians,
Black/African Americans, Asian Americans, Native Americans, and Latino(a)/Hispanic (the
“Other” category is omitted). Socioeconomic (SES) index was a composite variable derived
from participants’ responses to a set of question regarding family income, family
possession of certain properties, father’s occupation, and father’s and mother’s education
attainment. The index scores ranged from 58 to 131. Cohort represents the grade (9th, 10th,
11th, or 12th) which participants were in at the original survey. It was coded as a numeric
variable ranging from 9 to 12, with a larger number standing for an older cohort.
Data Analysis
Data analysis for the present dissertation consisted of three parts. First, multiple
regression analysis was used to show the predictive validity of vocational interests, alone,
for performance criteria in education and work contexts, and to demonstrate the
incremental validity of interests beyond cognitive abilities and personality (Hypothesis 1).
Second, polynomial regression analysis and response surface methodology was applied to
understand the effect of interest congruence (Hypotheses 2a and 2b), using the procedures
37
suggested by Edwards (2002; Edwards & Parry, 1993). Third, moderated path analysis was
used to test the moderated mediation model in Figure 2 (Hypotheses 3, Hypothesis 4, and
Hypotheses 5a to 5c), using an integrative analytical framework for moderation and
mediation proposed by Edwards and Lambert (2007). The first two parts of the data
analyses were conducted with both educational achievement and career success outcomes;
the third part of the analyses focused on career success.
Multiple Regression Analysis
To examine the predictive validity of interests, each of the performance criteria was
regressed onto six general interest factors. Regression coefficients and relative importance
weights (Johnson, 2000, 2004) were estimated for each interest factor. Observed multiple
correlation, R2, and adjusted R2 were estimated for each regression model.
To examine the incremental validity of interests over abilities and personality, a
series of regression models were fitted with variables entered into the model in the
following order: (1) demographic variables: gender, race/ethnicity, socioeconomic index;
(2) three cognitive ability composites; (3) nine personality scales; (4) six general interest
factors. All the predictors other than gender and race/ethnicity were standardized before
fitting the models. The same statistics reported for the regression models with interest
predictors alone were computed for the above hierarchical regression models. A
statistically significant change in R2 from the third step to the fourth step provides evidence
for the incremental validity of interests.
Polynomial Regression Analysis and Response Surface Methodology
Polynomial regression analysis was conducted to examine the effect of interest
congruence on performance criteria. The general form of polynomial regression equation
38
for examining the effect of congruence between one interest factor and the corresponding
environmental characteristic can be represented by the following formula:
Z = b0 + b1X + b2Y + b3X2 + b4XY + b5Y2 + e
where Z is the criterion, such as income, X is the individual characteristic, such as the level
of Business interests, and Y is the corresponding environmental characteristic, such as the
extent to which an occupation allows individuals to express their Business interests. This
equation is a generalization of regressions using squared difference scores as predictors.
It can be applied to analysis that compares the profile similarity of individuals and their
environments by including multiple pairs of individual and environmental characteristics
(Edwards, 2002; Edwards & Parry, 1993). In the case of the present dissertation, interest
scores for all six interest factors and major or occupational characteristics in all the
corresponding areas were included.
To visualize and interpret the effect of interest congruence on performance criteria,
the response surface for each pairing of interest score and major/occupation characteristic
was plotted. Stationary points and principal axes were computed using the formulas
provided by Edwards (2002).
Moderated Path Analyses
Moderated path analyses were used to examine the moderated mediation model for
career success, following the integrative analytical framework proposed by Edwards and
Lambert (2007). This analytical framework frames mediation in terms of a path model,
expresses relationships among variables in the model by regression equations, and
incorporates moderation by supplementing these equations with the moderator variable
and its product with the independent variable and the mediator variable. It relies on
39
reduced form equations (Johnston, 1984) to express direct, indirect, and total effects at
selected levels of moderator variables. The moderated mediation model for career success
is presented in Figure 2, with income and occupational prestige as two outcome variables.
Interests, abilities, and personality have direct effects on income and occupational prestige
as well as indirect effects through degree attainment. Gender also has direct effects on
income and occupational prestige, moderated by number of children, as well as indirect
effects through interests. Bootstrapping was performed to estimate bias-corrected
confidence intervals around the coefficients and to test, specifically, (1) if interests mediate
the effect of gender on career success, (2) if the number of children moderates the effect of
gender on career success, and (3) if degree attainment mediates the effects of individual
difference variables on career success.
40
RESULTS
Interests were found to predict performance criteria in both work and academic
settings and have incremental validity beyond cognitive ability and personality. Interests
were especially powerful for predicting income and were also predictive of college grades,
college persistence, degree attainment, and occupational prestige. In most cases, interest
congruence was positively associated with the performance criteria. As hypothesized,
interests, in particular People interests and Things interests, mediated the effect of gender
on career success. In contrast, the indirect effect of gender on career success through
abilities and personality was relatively small. Degree attainment mediates the effects of
individual differences on career success variables. As expected, number of children was
found to negatively impact income and occupational prestige, especially for women. Below
I detail and discuss the results.
Table 8 presents the results from regression analysis of work and academic
performance criteria on six vocational interest factors. Taken together, six interest factors
accounted for twenty-four percent of the variance in income, twenty-one percent in
occupational prestige, and twenty-five percent in degree attainment. These R2 values are
close to those found for ability and personality in predicting career success (.29 and .28,
respectively; Judge et al., 1999), indicating that interests alone are powerful predictors of
career success. Interests also predicted various academic performance criteria. Six interest
factors accounted for seven percent of the variance in the overall grade in college (and four
percent and five percent in first year grade and major grade, respectively) and nine percent
of the variance in college persistence. These results are in the same range as findings from
41
previous studies on the predictive validity of interests for academic performance (Nye et
al., 2011).
Regression coefficients and relative importance weights (RIW; Johnson, 2000, 2004)
in Table 8 showed the effect and relative contribution of each interest factor for each
criterion. The RIW is an index for interpreting results of multiple regressions in terms
of the proportionate contribution from each predictor to R2, after correcting for the effect
of intercorrelations among the predictors (Lorenzo-Seva, Ferrando, & Chico, 2010). The
RIW represents the proportion in the total variance accounted for explained by each
predictor and the RIWs of all the predictors in a regression add up to 100. Notably, Science
and Leadership interests had positive effects on all the criteria. Science interest was
particularly important for college persistence (exp(B) = 1.39), degree attainment (β = .30,
RIW = 35.8), and occupational prestige (β = .30, RIW = 38.9). The effects of Leadership
interest were smaller. The effects of other interest factors varied by criterion. For example,
Artistic interest (which has a component of liking schooling and culture in general) was the
most important contributor to college grades (β = .15, RIW = 44.5 for first year grade, β =
.16, RIW = 33.7 for overall grade, and β = .16, RIW = 39.9 for major grade) and was also
positively associated with college persistence (exp(B) = 1.24) and degree attainment (β =
.16, RIW = 11.0), but led to lower income (β = -.11, RIW = 9.8). People interest (which has a
strong service component) had a markedly negative impact on income (β = -.41, RIW =
60.0). Both People and Things interests were negatively associated with college persistence
(exp(B) = .68 and .72, respectively), degree attainment (β = -.38 and -.26, RIW = 31.6 and
7.3, respectively), and occupational prestige (β = -.29 and -.36, RIW = 18.6 and 19.2,
42
respectively). These results provide evidence for the predictive validity of interests for
work and academic performance.
Even after controlling for ability and personality, interests still made a considerable
contribution to the prediction of the criteria, especially the career success outcomes (see
Table 9). Interests were extremely powerful predictors for income, accounting for sixteen
percent of the variance beyond ability and personality. The RIW may be more reflective of
the relative contribution of interests to income in comparison with ability and personality.
Importantly, interests contributed to 83.3 percent of the total variance accounted for in
income, as compared to 12.0 percent by ability and 4.7 percent by personality. Interests
also had substantial incremental validity over ability and personality for predicting degree
attainment and occupational prestige (ΔR2 = .08 and .07, RIW = 35.4 and 32.9, respectively).
The incremental validity of interests for predicting college grades and persistence was
smaller (ΔR2 = .01 and RIW = 15.6 for first year grade, ΔR2 = .03 and RIW = 26.6 for overall
grade, ΔR2 = .03 and RIW = 28.7 for major grade, and ΔNegelkerke R2 = .03 for persistence).
Note that the ΔR2 for overall grade, major grade, and persistence are in the same range as
findings reported by previous studies on the incremental validity of interests for job
performance (e.g., Gellatly et al., 1991; Van Iddekinge, et al., 2011). Moreover, although
ability was the most powerful predictor for academic performance, the relative
contribution of interests was similar to that of personality. In sum, interests have
incremental validity for predicting career success and academic achievement beyond
ability and personality. Hypothesis 1 is supported. Figure 4 depicts the relative
contribution of ability, personality, and interests to each criterion.
43
It may be argued that interests predicted income solely because interests influenced
the choice of occupational fields that differed in average pay, but not because interests
influenced performance. To test this possibility, two additional analyses were conducted.
First, the incremental validity of interests over abilities and personality in predicting
income within twelve occupational groups was examined. The occupational groups are
available in the Project TALENT data set. A list of occupations within each occupational
group can be found in Appendix F of the Project TALENT Data Bank Handbook (Wise et al.,
1979). It is expected that, if interests predicted income solely through career choice, the
incremental validity of interests for predicting income within occupational groups would
be small or nonexistent. Table 10 shows the sequential R2 change of ability, personality,
and interests for predicting income within each occupational group from hierarchical
regression analysis. As shown, interests had substantial incremental validity over abilities
and personality in predicting income within occupational groups. The average within-
occupational group R2 change of interests was .12, only slightly reduced from the R2 change
of .16 for predicting income across all occupational groups. More importantly, the within-
group incremental validity of interests in predicting income was much higher than that of
ability (.04) and personality (.03).
Second, the incremental validity of interests over abilities and personality in
predicting income was examined after controlling for occupational prestige. If interests
have increment validity over abilities and personality even after controlling for
occupational prestige, it rules out the explanation that interest predicted income solely
because individuals selected themselves into occupations with differential prestige, and in
turn differential income, based on interests (e.g., construction workers versus doctors).
44
Table 11 shows the results from hierarchical regression analysis, with occupational
prestige entered first and abilities, personality, and interests entered subsequently.
Occupational prestige accounted for nine percent of the variance in income. Again, interests
were shown to have sizeable incremental validity over abilities and personality in
predicting income even after controlling for occupational prestige (ΔR2 = .13). The
predictive power of interests was stronger than that of ability (ΔR2 = .05) and personality
(ΔR2 = .02). These results from two supplementary analyses provide counterevidence to
the assumption that interests are only about career choice and that, once individuals enter
their occupations, cognitive ability and personality traits take over and predict
performance, not interests.
Few of the previous studies that examined the incremental validity of interests for
performance controlled for demographic variables. The present dissertation also examined
the incremental validity of interests after controlling for four demographic variables:
gender, race/ethnicity, SES, and cohort. After demographics were taken into account, the
effects of interests (as well as ability, and to some extent, personality) dropped
substantially. As shown in Table 12, the additional variance explained by interests after
controlling for demographic variables, ability, and personality decreased for all the criteria,
particularly for income. Note that gender contributed to 35.3 percent of the total variance
explained in income, and the relative contribution of interests dropped from 83.3 percent
to 45.5 percent (compared to 8.1 percent by ability and 3.0 percent by personality). This
result highlights the strong link between gender and interests and necessitates the study of
interests for understanding the mechanisms through which gender influences career
outcomes. I examined the mediation effect of interests in the last part of the analysis. SES
45
appeared to be an important contributor to degree attainment and occupational prestige
(RIW = 16.5 and 14.5, respectively). The relative contribution of interests to these two
criteria (RIW = 23.7 and 24.3, respectively) was smaller than that of ability but greater than
that of personality. Ability was still the most powerful predictor for college grades,
contributing to about half of the total variance explained. The relative importance of
interests and personality to college grades were similar. Figure 5 depicts the relative
contribution of gender, race/ethnicity, SES, cohort, ability, personality, and interests to
each criterion.
Next, the effect of interest congruence was examined on four of the seven outcomes
– income, overall grade in college, major grade in college, and college persistence – using
polynomial regression analysis. This analysis was not performed on occupational prestige,
because this outcome is essentially tied with occupational fields and there is no variability
of occupational prestige within an occupation. Therefore, it does not make sense to use
interest congruence with occupation to predict occupational prestige. The analysis was not
performed on first year grade in college either, because the majors of participants were
unknown at the time of the first follow-up survey and thus interest congruence with major
could not be calculated. Similarly, degree attainment ranged from high school drop outs to
doctoral or law degree earners, so no one set of environmental characteristics was
available for calculating interest congruence with academic environment. Thus, the effect
of interest congruence was not examined for degree attainment. The polynomial regression
equation for each of the four outcomes took a general form as the following:
46
Z = b0 + b1XThings + b2YThings + b3XThings2 + b4YThings2 + b5XThingsYThings +
b6XArtistic + b7YArtistic + b8XArtistic2 + b9YArtistic2 + b10XArtisticYArtistic +
b11XScience + b12YScience + b13XScience2 + b14YScience2 + b15XScienceYScience +
b16XPeople + b17YPeople + b18XPeople2 + b19YPeople2 + b20XPeopleYPeople +
b21XBusiness + b22YBusiness + b23XBusiness2 + b24YBusiness2 + b25XBusinessYBusiness +
b26XLeadership + b27YLeadership + b28XLeadership2 + b29YLeadership2 + b5XLeadershipYLeadership +
e
where Z is the criterion, X is individuals’ interests on an interest factor, and Y is the ratings
of corresponding environment characteristic. For each interest element, the terms X and X2
represent the linear and quadratic effects of interests on Z, Y and Y2 represent the linear
and quadratic effects of corresponding major or occupational characteristics on Z, and the
interaction term XY represents the extent to which major or occupational characteristics
moderate the effect of interests on the criterion.
Table 13 shows the results of polynomial regression for each criterion. For
comparison, I also regressed each criterion onto three types of congruence indices: (1) a
fully constrained profile difference index (Z = b0 + b1Σ(Xi – Yi)2 + e, i.e., regression
coefficients are constrained to be the same for the squared difference between each
interest factor and corresponding environmental characteristic), (2) a partly constrained
profile difference index (Z = b0 + Σ bi (Xi – Yi)2 + e, i.e., regression coefficients are allowed to
vary for the squared differences), and (3) a profile correlation index. Results show that
polynomial regression equations predicted much better than congruence indices,
accounting for thirty-six percent of the variance in income (as compared to 0.4 percent, 1.4
percent, and 1.7 percent by the three congruence indices, respectively), nine percent of the
variance in overall grade in college (as compared to 0 percent, 0.3 percent, 0.2 percent),
seven percent of the variance in major grade (as compared to 0 percent, 0.2 percent, 0.1
percent). Polynomial regression equation for college persistence also had a Nagelkerke R2
47
of .12, as compared to .01, .01, and .02 by three congruence indices, respectively. These
results demonstrate that polynomial regression equations were more predictive than
congruence indices.
It may be argued that the results for these congruence indices are not comparable
with the results for polynomial regression equations, as the congruence indices do not
include linear effects of interests and major/occupational characteristics (Xi and Yi).
Therefore, I also examined the incremental validity of Xi and Yi over the higher order terms
(X2, Y2, and XY) in polynomial regressions as well as the incremental validity of Xi and Yi
over the three congruence indices if these linear terms had been included. The results are
shown in Table 14. The third, fourth, and fifth columns present the R2 changes of the higher
order terms in polynomial regressions and the congruence indices, the R2 changes added
by Xi and Yi in polynomial regressions and over three congruence indices, and the total R2,
respectively. Results show that, with higher order terms alone, polynomial regression
equations predicted better than congruence indices. However, Xi and Yi had considerable
incremental validity over the congruence indices such that if the linear effects of interests
and major/occupational characteristics were included with the congruence indices, the
total R2s were close to those of polynomial regressions. These results showed that a
polynomial regression equation was more predictive than congruence indices partly
because of its unconstrained form, as suggested by Edwards (1993, 2002), but mostly
because of its inclusion of linear effects of interests and environmental characteristics.
Table 15 lists the unstandardized regression coefficients of the polynomial
regression equations for each criterion. Based on these polynomial regression equations,
the response surface plot of each criterion by each pair of interest factor and corresponding
48
environmental characteristic was drawn (see Figure 6 to Figure 29). The contour plot and
the first and second principal axes of each response surface were mapped on the X-Y plane.
The regression coefficients of the interaction terms (XY) in Table 15 denote the effects of
interest congruence on the criteria. Positive regression coefficients of XY are consistent
with the hypothesis of the present dissertation, meaning that congruent interests with a
major or occupation benefit academic or work performance. To the contrary, negative
regression coefficients of XY mean that congruent interests impair performance.
The effects of interest congruence for income were mostly positive. Congruent
interests in Things, Artistic, Science, Business, and Leadership all had significantly positive
influence on income, supporting Hypothesis 2a. The only exception is congruence interest
in People, which had a negative effect. As shown in Table 15 and Figure 6 to 11, in general,
occupational characteristics (Y) had relatively large effects on income (Z) compared to the
effects of interests (X) and interest congruence. Within the observed range, Science and
Business characteristics of an occupation had curvilinear but mostly positive effects on the
income of individuals within the occupation (Figure 8 and Figure 10, respectively); People
characteristic of an occupation was negatively associated with income (Figure 9); and the
effect of Things characteristic of an occupation was mostly negative (Figure 6). The
relatively large effects of occupational characteristics make it somewhat difficult to tell the
effects of interest congruence from the response surface plots. Nonetheless, the effects of
interest congruence may be detected by examining the Z-X relationship at the highest and
lowest levels of Y. For example, Science interest had a positive linear effect on income at all
levels of Science characteristic of an occupation, yet the positive effect of Science interest
was stronger when the Science characteristic of an occupation was high than when it was
49
low (Figure 8). The effect of Leadership interest on income was positive overall, yet it was
the strongest when the Leadership characteristic of an occupation was high and became
slightly negative when the Leadership characteristic of an occupation was low (Figure 11).
On the contrary, People interest had a negative effect on income at all levels of People
characteristic of an occupation, yet this negative effect was the strongest when the People
characteristic of an occupation was the highest (Figure 9).
The effects of interest congruence were somewhat mixed for college grades.
Consistent with Hypothesis 2b, congruent interests in Science, People, and Business led to
better grades. As shown in the Figure 14 and Figure 20, overall grade and major grade
were the highest when an individual’s Science interest and a major’s Science characteristic
were both high. The grades were the lowest when there was a mismatch, that is, when an
individual’s Science interest was high and yet Science characteristic of a major was low, or
vice versa. Figure 15 depict the effect of interest congruence in People on overall grade.
This response surface is of a concave shape, conforming nicely to the prediction of
Hypothesis 2b: when the level of an individual’s People interest matched the level of People
characteristic of a major, grades were high; when the levels of interest and major
characteristic in People were incompatible, grades were low. The effect of interest
congruence in People on major grade was also positive, but much weaker, as shown in
Figure 21. Figure 16 and Figure 22 illustrate the effect of interest congruence in Business
on overall grade and major grade, respectively. The effects were modest, but again, a
pattern can be seen that grades were higher when the level of interest and that of major
characteristic matched and lower when they were incompatible. Interestingly, the effect of
interest congruence always seemed to be greater on overall grades than that on major
50
grades. Inconsistent with Hypothesis 2b, interest congruence in Things had negative effects
on grades, as shown in Figure 12 and Figure 18. Interest congruence in Artistic also had a
negative effect on overall grade in college (Figure13); and the effect of interest congruence
in Artistic on major grade was insignificant (Figure 19). Interest congruence in Leadership
did not appear to affect college grades (Figure 17 and Figure 23).
As expected, interest congruence significantly affected persistence in college. The
more compatible individuals’ interests in Artistic, Science, People, Business, and
Leadership are with their majors’ characteristics in these areas, the more likely that they
stayed in college rather than dropping out (exp(B) were 1.23, 1.43, 1.13, 1.24, and 1.16,
respectively), supporting Hypothesis 2b. Interest congruence in Things did not have a
significant effect on college persistence. As shown in Figure 25, 26, and 29, when
participants’ interest in Artistic, Science, or Leadership and their majors’ corresponding
characteristics in these areas were both high, the odds ratio that they persisted in college
was the highest. Also note that in Figure 26, when Science characteristic of a major was the
highest and an individual’s Science interest was the lowest, there was almost no chance
that the individual would persist in college. The patterns of the effects of interest
congruence in People and Business were more complex, as shown in Figure 27 and 28. The
odds ratio for persistence was highest when People characteristic of a major is at the
highest level and when an individual’s People interest was about one standard deviation
below average; at higher levels of People interest, the odds ratio that an individual would
persist in college became lower. The odds ratio for persistence was close to zero, when an
individual’s People interest was very high, and yet a major’s People characteristic was very
low. As for interest congruence in Business, when a major’s Business characteristic was the
51
highest, higher Business interest was associated with greater likelihood for persisting in
college; when a major’s Business characteristic was low, the effect of Business interest on
persistence was minimal.
The above results are from polynomial regressions in which terms corresponding to
six interest areas were analyzed simultaneously. The effects of terms in one interest area
(e.g., XThings, YThings, XThings2, YThings2, XThingsYThings) should be regarded as the effects of one
interest area while controlling for the effects of the other five interest areas. Therefore, the
results were affected by the intercorrelations among the six sets of terms. To examine the
extent to which this was the case, I also analyzed polynomial regressions with one of the six
interest areas at a time for each criterion. The polynomial regression equations took a
general form as the following:
Z = b0 + b1Xi + b2Yi + b3Xi2 + b4Yi2 + b5XiYi + ei
where i represented each of the six interest areas. Table 16 lists the unstandardized
regression coefficients from the polynomial regression equations with six interest areas
analyzed separately. Note that the negative effect of interest congruence in Artistic on
overall grade in college became insignificant; the negative effect of interest congruence in
Things on major grade in college also became insignificant; and the effect of interest in
Artistic on major grade in college was found to be positive. On the contrary, when analyzed
separately from other interest areas, the positive effect of interest congruence in Business
on college grades became insignificant. The magnitude of other effects of interest
congruence changed as well, although the direction of most effects stayed positive,
supporting Hypothesis 2a and partly supporting Hypothesis 2b.
52
The last part of the results focus on the proposed path model for two career success
variables, income and occupational prestige. An issue occurred when fitting the regression
equations for the moderated mediation models: the number of children and the product
term of gender and number of children were so highly correlated (r = .74) that severe
multicollinearity was present when both terms were in the equation. Therefore, the
interaction between gender and number of children on career success and their effects on
career success were examined separately from the mediation part of the model. The effect
of number of children on career success was estimated within each gender. Results showed
that, overall, the number of children negatively affected two career success variables (the
regression coefficients of income and occupational prestige on the number of children are .75 and -4.24, respective). The effect of number of children on occupation prestige was
more profound – the number of children was associated with lower occupation prestige for
both men and women, but the negative influence for women was even more severe (B = 3.98 and -5.40 for men and women, respectively). With regard to income, the number of
children had a small positive effect for men (B = .37), yet a strong negative effect for
women (B = -3.43). Figure 30 illustrated the overall and by-gender effect of number of
children on income and Figure 31 illustrated the overall and by-gender effect of number of
children on occupational prestige. Note that although men had higher intercepts of income
and occupational prestige, women and men with no children differed very little in their
income and occupational prestige. When their number of children increased, however, the
gender wage gap and the gender gap in occupational prestige quickly widened. Hypothesis
4 was supported.
53
To examine the mediation path model for career success, regressions for two
outcome variables, income and occupational prestige, as well as regressions for each
proposed mediator variable were analyzed. Using the reduced regression forms, the
indirect effects of gender on career success through individual difference variables and the
indirect effects of individual difference variables on career success through degree
attainment were computed from the regression coefficients. One thousand bootstrapping
samples were produced to construct confidence intervals around the regression
coefficients and to test the significance of the indirect effects. Standardized regression
coefficients from the regression analyses are presented in Table 17. The direct effect of
gender on career success and the indirect effects of gender on career success through
individual difference variables are presented in Table 18. The direct effects of abilities,
personality, and interests on career success and their indirect effects through degree
attainment are presented in Table 19. Due to the large sample size of the present study,
almost all the indirect effects were significant at p < .001 level. Therefore, the results
presented below focus on the indirect effects that were larger than .01.
As predicted in Hypothesis 3, all six interest factors partially mediated the effect of
gender on income, such that gender affected income partially through the interest in
Things, Artistic, Science, People, Business, and Leadership (the indirect effects through six
interest factors are .06, -.02, -.01, -.09, -.01, -.01, respectively). Except for Things interest,
the indirect effects of gender through the other five interest factors were all in the negative
direction, meaning that women had greater interest in most areas that led to lower income
(i.e., Artistic and People interests) and were less interested in areas that led to higher
income (i.e., Science, Business, and Leadership interests). The mediation effect of People
54
interest was the strongest. The indirect effect of gender on income through Things interest
was positive, because women had lower Things interest compared to men, which impaired
income. In contrast, the indirect effects of gender on income through abilities and
personality were minimal. Only the mediation effect of math ability was close to -.01. On
average, women had slightly lower math ability, which led to lower income. Gender had a
sizable negative direct effect of -.38 on income; and with all the indirect effects through
individual difference variables, gender had a total effect of -.47 on income. Table 18 also
presented the raw correlation between gender and income (in the third column) and the
partially correlation between gender and income while controlling for each individual
difference variable (in the fourth column). Again, note that the correlation between gender
and income reduced substantially after People interest was taken into account. The
changes in the correlation between gender and income were small when abilities and
personality were taken into account.
As for occupational prestige, the indirect effects were in general stronger compared
to those for income since degree attainment had a much stronger influence on occupational
prestige (β = .59). Consistent with Hypothesis 3, the effect of gender on occupational
prestige was partially mediated through interests in Things, Artistic, Science, People,
Business, and Leadership (the indirect effects were .18, .02, -.03, -.02, -.01, -.02,
respectively). Things interest was the strongest mediator for the effect of gender on
occupational prestige, because men had much stronger preference for things-oriented
careers (e.g., auto mechanics, construction workers) that were usually of lower prestige.
Women had stronger interest in People that led to lower-prestige occupations. Among the
interests that contributed to occupational prestige, women had greater Artistic interest,
55
and yet lower interests in Science, Business, and Leadership. The effect of gender on
occupational prestige were also mediated by the abilities (indirect effects were .02, -.03,
.01, respectively). The mediation effects of personality were again very small. Gender had a
negative direct effect of -.08 on occupation prestige. However, the total effect of gender on
occupational prestige was slightly positive, largely due to the positive indirect effect
through Things interest. These results testify that interests are important channels through
which gender operates and influences career outcomes.
As shown in Table 19, degree attainment mediated the effects of most individual
difference variables on income and occupational prestige. Again, due to the much stronger
effect of degree attainment on occupational prestige than that on income, the indirect
effects of individual difference variables on occupational prestige were much stronger. For
income, the level of education individuals attained partially mediated the effects of verbal,
math ability, and spatially ability (indirect effects were .02, .02, and -.01, respectively).
Degree attainment also partially mediated the effects of all the interest variables on income
except for Business interest. Verbal and Math abilities as well as Artistic, Science, and
Leadership interests contributed to the level of education achieved, which in turn led to
higher income. On the contrary, Spatial ability as well as Things and People interests were
associated with lower level of education achieved, which in turn led to lower income. The
mediation effects of degree attainment for personality variables were in general small,
except that for the Leadership personality. Leadership personality was positively
associated with degree attainment, which in turn led to higher income (indirect effect was
.01).
56
Because degree attainment was strongly associated with occupational prestige (in
fact, the education level of an occupation was one of the three criteria for the scoring of
occupational prestige), it turned out to be a critical mediator between individual difference
variables and occupational prestige. Most notably, degree attainment partially mediated
the effects of Verbal, Math, and Spatial abilities on occupational prestige (indirect effects
were .13, .15, -.04, respectively). Verbal and Math abilities were positively associated with
the level of education achieved, which in turn led individuals to more prestigious
occupations. Spatial ability, on the contrary, had negative indirect effect on occupational
prestige. Degree attainment also partially mediated the effects of Things, Science, People,
and Leadership interests on occupational prestige (indirect effects were -.10, .07, -.06, .05,
respectively) and fully mediated the effect of Artistic interest on occupational prestige
(indirect effect was .05). Interestingly, the effect of Business interest on neither career
success variables was mediated by degree attainment. Degree attainment was also found to
mediate the effects of all the personality variables on occupational prestige. The strongest
indirect effect was again found for Leadership personality. Hypothesis 5 was supported.
57
DISCUSSION
The current dissertation sought answers to three interrelated research questions.
First of all, do interests have incremental validity over and beyond ability and personality
in predicting academic and work performance, and how large is the predictive power of
interests compared to that of ability and personality? These questions were examined
using Project TALENT, a national longitudinal survey. Second, using polynomial regression
and response surface methodology, this dissertation examined the effect of interest
congruence on performance criteria, or the extent to which major or occupational
characteristics moderate the effects of interests on academic achievement and career
success. Third, this dissertation developed and tested a moderated mediation model for
career success. Specifically, this study examined whether number of children moderated
the effect of gender on career success, whether the effect of gender on career success was
mediated through interests, and whether the effects of individual difference variables on
career success were mediated through degree attainment. Overall, results showed that
interests had incremental validity over ability and personality for predicting work and
academic performance. In most cases, interest congruence was associated with higher
performance. The hypothesized moderated mediation model of career success was also
supported. Below I review key findings for the three research questions and discuss
substantive and methodological implications of each part of the results.
Power of Interests in Predicting Academic Achievement and Career Success
It is time to rethink the common belief that interests only predict choice but not
performance and to reassess the value of vocational interests beyond career guidance.
Results from the present dissertation showed that six interest factors were predictive of
58
performance criteria in work and academic contexts and confirmed the findings from two
recent meta-analyses that linked interests to performance (Nye et al., 2012; Van Iddekinge
et al., 2011). Moreover, results supported the incremental validity of interests beyond
ability and personality and the incremental validity of interests held up after controlling for
demographic variables. Interests appeared to be the most powerful contributor to income
and were important determinants of degree attainment and occupational prestige. The
power of interests in predicting academic performance was as large as that of personality
variables.
The most notable result is the strong association between interests and income.
Interests, as measured in participants’ high school years, alone explained close to one
fourth of the variance in income 11 years after their high school graduation.
After ability and personality were taken into account, interests still contributed an
overwhelming proportion of the total variance explained. Interests may influence income
through many mechanisms. First of all, interests can affect income through selection into
different occupations that vary greatly in average pay level. The strong relationship
between occupational characteristics and income can be seen clearly from the regression
coefficients in Table 9 and the response surface plots for income in Figure 6 to Figure 11.
Individuals may be attracted to different occupational fields that require similar education
and skill level yet differ greatly in average wage. For example, carpenters (high in Things
interest) and hairdressers (high in People interest) both require no education beyond high
school and involve a moderate level of job skill training, yet the latter were paid a lot lower.
Individuals who are strongly interested in People might be drawn to become hairdressers
instead of carpenters and in turn earn lower income. Interests may also affect individuals’
59
career paths by inspiring them to obtain higher levels of education or additional training to
enter highly skilled or professional occupational fields with higher pay (e.g., physicians), as
evidenced by the results on the mediation of degree attainment between interests and
income. Even within an occupational field, interests differentiate individuals’ income by
motivating them to pursue different work activities or attain different positions. For
example, a physicist with a strong Business interest may set the goal and strive to become a
research director or a department head who is paid much more than a physicist with a
strong Things interest and works as a specialized technician. Lastly, interests may influence
income indirectly through the acquisition of job knowledge and the improvement of job
performance. Only the first of the above mechanisms is about interest influencing career
choices, whereas the rest of these mechanisms are about interest motivating individuals to
achieve. Results on the incremental validity of interests for predicting income within
occupational groups and results after controlling for occupational prestige (Table 10 and
Table 11) provide indirect evidence that interests drive not only career choices but also
behaviors and performance after individuals enter certain occupations.
Interestingly, interests appeared to be more powerful predictors for distal outcomes
(i.e., degree attainment, occupational prestige, aside from income) than for more proximal
outcomes (i.e., college grades). The high predictive validity of interests found for degree
attainment, occupational prestige, and income is impressive given that these criteria were
assessed much later in participants’ life (at the 11th year after their high school graduation).
One explanation for this finding may be that these criteria had higher reliability since
grades may vary across institutions and may be subject to more errors in self-report.
Another possible reason is that interests by nature are more predictive of long-term
60
outcomes than short-term outcomes. Recall the three motivational properties of interest:
direction, vigor, and persistence. Interest influences performance through a continuous
process by channeling individuals’ efforts to the direction that they have passion for,
energizing the efforts that individuals put forth, and sustaining their efforts till the goal is
accomplished. As such, the effect of interest is cumulative and may unfold to a greater
extend over a longer time span. For example, an individual with a strong interest in Science
may drop out of college because of financial strain or a semester of poor grades, but may
eventually get back to school, obtain an advanced degree, and seek a career as a research
scientist. Interested individuals are more likely to persevere and succeed in the areas that
they are passionate about despite short-term setbacks.
Another important finding is that demographic variables had strong effects on
performance outcomes, as shown in Table 12 and Figure 5. Gender contributed 35.3
percent of the total variance explained in income, and SES accounted for sizeable
proportions of variance in degree attainment and occupational prestige. The relationships
between demographic variables and career and educational outcomes have been well
documented in the economics and sociology literature. The gender wage gap and the
dynamic of gender wage inequity over time has been one of the most heavily studied labor
market phenomenon (e.g., Weinberger & Kuhn, 2010). Many studies have also reported
moderate correlations between father’s occupation, one of the key indicators of a young
person’s socioeconomic status, and that person’s occupational status and education level
later in life (e.g., Hauser & Warren, 1997). However, demographic variables are frequently
placed in a less important place in models of psychological variables and the mechanism
through which demographic variables affect work outcomes are less well understood.
61
Previous studies on the individual difference determinants of career success (e.g., Judge et
al., 1999) have not examined the effects of gender or socioeconomic status. Only one of the
studies on the incremental validity of interests over ability and personality in predicting
performance included gender and language of the participants as control variables (i.e.,
Black, 1999). Black (1999) found somewhat inconsistent results about the incremental
validity of interest: the interest measure explained additional variance in military training
performance after controlling for demographics, cognitive ability, and personality, but not
in training course completion. Granted, demographic information of participants may not
be available for every study. Nonetheless, researchers need to be aware that the
incremental validity and the relative contribution of psychological variables on career
success are likely to be much smaller after demographics are taken into account. In
addition, it is important to understand the psychological links between demographic
variables and career outcomes, including interests as a major mediating channel.
For the purpose of the current dissertation, the RIWs (Johnson, 2000, 2004) offered
useful information on the relative importance of interests, ability, and personality for each
criterion. ΔR2, or the change in variance accounted for when interests were entered into
the hierarchical regression answered the question “do interests have incremental validity
over and beyond ability and personality in predicting performance?” In situations where
predictors are correlated with each other, however, the use of R2 change and discussion of
incremental validity of some predictors over others may not accurately represent the
importance of each predictor. This is because the order for entering predictors into the
hierarchical regressions may be based on arbitrary assumptions and the hierarchical
procedure can mask the actual contribution of predictors that are entered later. In contrast,
62
RIWs represent the relative contribution of each predictor in the presence of all the other
predictors. As such, comparing the RIW of interests to that of abilities and personality
answered the questions “how strong is the predictive power of interests relative to that of
ability and personality” and provided valuable information with regard to the importance
of interests in addition to the change in R2.
The analyses of the incremental validity and relative contribution of interests
focused on six interest factors collectively. It is important to note that the direction and
magnitude of the effects of interests on academic achievement and career success will vary
by interest factor and how each interest factor is measured. For example, results in Table 8
highlighted the positive contribution of Science interest to every criterion; Artistic interest
appeared to be important for educational achievement. Science interest is marked by a
desire for knowledge and the love of learning that would be valuable for all the criteria; and
the contribution of Artistic interest to educational achievement is because a number of
Artistic interest items in Project TALENT measure the liking of studying and general
culture. Selecting and combining items from these two interest factors is a plausible
approach to maximize the predictive validity of interests for academic performance. These
findings for Science and Artistic interests are similar to Campbell’s (1966) development of
an Academic Achievement scale for the Strong Vocational Interest Blank. The Academic
Achievement scale, saturated with Science and Artistic items, differentiated students who
have potential for high academic achievement and students who have no aspiration for
academic activities (Campbell, 1971, pp. 195-208). Lastly, the People interest items in
Project TALENT are lower-level service types of occupations and activities and the majority
of Things interest items are skilled trades or laborer types of occupations and activities. As
63
such, People interest was found to negatively impact college persistence, degree
attainment, occupational prestige, and income; and Things had negative effects on all the
outcomes other than income. Had these interest factors been measured differently (e.g.,
with professional or highly prestigious occupations), the results may have differed in
magnitude or direction.
Interest Congruence and Performance
Does interest congruence contribute to the prediction of performance? Results from
the current dissertation showed that a single unified answer may not be satisfactory. In
most cases, the answer is “Yes,” supporting Holland’s theory and P-E Fit theory. However,
the power of interest congruence in predicting academic and work performance varied by
criterion; and the importance of each interest factor was different. Interest congruence
appeared to be essential for college persistence. Strong interests in Artistic, Science, and
Leadership were particularly crucial for the persistence of individuals in majors marked by
these characteristics. Interest congruence in Things, Artistic, Science, Business, and
Leadership all led to higher income, although these effects were shadowed by the
overwhelming effects of occupational characteristics. For college grades, the effect of
interest congruence varied by interest factor. Not surprisingly, a strong Science interest
was vital for individuals in a highly scientific major to get good overall academic records
and to excel in their major field of study. People interest and Business interest also
mattered for doing well academically in corresponding majors.
Results from the current dissertation point to the importance of interest congruence
in Science in addition to the power of Science interest alone. Interest congruence in Science
contributed to better grades, greater likelihood to persist in college, as well as a high
64
income; and the effects of interest congruence in Science were relatively large compared to
those of other interests. A strong interest in Science is a necessary condition for success in
scientific majors and careers. This finding has implication for science education and
research on participation and achievement in the STEM fields. More research is needed on
the development of Science interest, the origin of individual differences in Science interest,
and the causal links between Science interest and achievement in scientific majors and
careers. Furthermore, these results suggest that identifying individuals with strong Science
interests and selecting on the basis of Science interest is valuable for relevant fields.
Graduate school selection, for example, may benefit from the use of interest measures in
addition to the assessment of cognitive aptitudes.
As illustrated in the response surface plots, each interest factor (X) and major or
occupational characteristic in each area (Y) functioned differently for each performance
criterion (Z). For this reason, it is not surprising that the congruence indices did not explain
much variance in the criteria (see Table 13). A fully constrained profile difference index
(i.e., Z = b0 + b1Σ(Xi – Yi)2 + e) specifies that the response surface for every criterion should
be of a concave shape like that depicted in Figure 15. Also it specifies that the quadratic
effects of every interest factor and corresponding major or occupational characteristic
should be of the exact same size (i.e., equal to b1). These constraints clearly did not hold in
the examination of six interest factors and corresponding major or occupational
characteristics in the present dissertation. A partly constrained profile difference index
(i.e., Z = b0 + Σ bi (Xi – Yi)2 + e) allows the regression coefficients to vary for each squared
difference term, but still specifies that the response surface should be of a concave shape
and that the quadratic effects of each pair of interest factor and corresponding major or
65
occupational characteristic should be the same (i.e., equal to bi). As constrained forms of
the polynomial regression equation, the profile difference indices were much less
predictive than the unconstrained, more flexible polynomial regression equation. A profile
correlation index, though not directly comparable to a polynomial regression equation,
suffers from the same drawbacks of congruence indices – it reduces the functions of
multiple interest dimensions to one index and combines distinct contribution from
individual interests and that from major or occupational characteristics (Edwards, 1993).
More importantly, all three types of congruence indices discard information about absolute
levels of interests, which contributed to the majority of the variance explained in income
and substantial amount of variance explained in college grades and persistence (see Table
14). Polynomial regression equations integrate information of absolute levels of interests,
major or occupational characteristics, and their interaction, and were shown to have much
stronger predictive validity for the performance criteria in the current dissertation. In this
sense, polynomial regression analysis is a useful and more powerful technique for
examining the effects of interests and interest congruence.
Nonetheless, the polynomial regression technique is not without its limitations.
First, the inclusion of square and product terms is highly likely to cause multicollinearity in
the equation. Though it does not reduce the predictive power of a regression model as a
whole, the presence of multicollinearity in the model may produce invalid results for
individual predictors. The most frequently recommended solution for multicollinearity
resulted from square and interaction terms is to mean center predictors X and Y before
squaring and multiplying them (e.g., Edwards, 2002; Kutner, Nachtsheim, Neter, & Li,
2005). However, mean centering does not solve the problem. As suggested by Echambadi
66
and Hess (2005), though mean centering may work to reduce the correlation between the
predictors, it does not alleviate the negative consequences due to the collinearity problem.
When mean centering fails to work, one or more terms may need to be dropped from the
equation. For example, when Science interest and Science characteristic of a major were
used to predict overall grade in college, simultaneous inclusion of squared Science interests
and the interaction term of Science interest and a major’s Science characteristic caused
such high collinearity that it was unable to perform matrix inversion for calculation of the
regression coefficients. As a result, the squared Science interest term was dropped from the
model in subsequent analyses since it contributed relative little to the model.
Another source of multicollinearity is the intercorrelations among different sets of
terms in a polynomial regression equation. As shown in the current dissertation, when six
interest areas were analyzed separately instead of simultaneously, regression coefficients
differed to various degrees in magnitude and sometimes even in direction. The changes in
results appeared to be more substantial when college grades were used as outcomes. This
type of multicollinearity was not as severe as the multicollinearity caused by higher order
terms and did not hinder the computation of regression coefficients. However, it renders
the estimates of regression coefficients less precise and is likely to multiply the effect of any
bias in model specification. The current dissertation is intended to use polynomial
regression analysis for examining the effects of interests and environmental characteristics
as a whole and to apply the response surface methodology to visualize and understand the
effect of interest congruence. Researchers, however, need to be aware of the influence of
intercorrelations among predictors on the estimates in a polynomial regression equation
and need to exert caution when trying to interpret individual regression coefficients from
67
the equation. Furthermore, polynomial regression approach is best suited for interest
congruence research where the interest measure is developed as independent interest
factors instead of interest types that have a correlation pattern such as Holland type
interest scales.
In addition, as suggested by Edwards (2002), polynomial regression analysis adopts
the standard regression assumption that the independent variables are measured without
error. As the reliability of component measures decreases, coefficient estimates may be
biased. The presence of multicollinearity in a polynomial regression equation, as noted
previously, can multiply the effects of the bias. As an alternative, structural equation
modeling with latent variables (Bollen, 1989) may be a more fruitful method.
This dissertation provided a first step in the application of polynomial regression
analysis and response surface methodology by examining the effects of interest
congruence. The use of the technique in congruence research with interests needs further
improvement and investigation. It is important to note that the idea of using polynomials of
interest factors to study the congruence-criterion link is essentially different from the idea
of using a congruence index in interest research. The two approaches make different
assumptions regarding the structure of interests and use different information of interests:
the polynomial regression approach treats interests as discrete dimensions and uses
information about absolute levels of interests independent from each other, whereas the
congruence index approach in interest research treats interests as interrelated types and
uses information about the shape of interest types holistically. The two approaches also
differ in their interpretation of congruence: the polynomial regression approach interprets
the effect of interest congruence between each pair of interest factor and corresponding
68
major or occupational characteristic, whereas the congruence index approach interprets
the effect of interest congruence between the shape of interests and the major or
occupational interest profile. As such, results from the two approaches may not be directly
comparable. The differences in the two approaches may explain the discrepancy between
the results from the current dissertation, which employed the polynomial regression
technique and found three types of congruence indices to be less predictive, and the results
from the Nye et al. (2012) meta-analysis, which was based on the Holland model and found
interest congruence to be more powerful predictor than interest scores alone. Nonetheless,
it is recommended that the relative contribution of interest level and interest profile be
examined and compared empirically rather than simply discarding one in favor of the other
(Cronbach & Gleser, 1953) no matter which of the two approaches is used.
Determinants of Career Success for Men and Women
What are the determinants of career success? What factors are responsible for the
differential career attainment of men and women? The third part of this dissertation
provided insight to these two questions. Results showed that, among all the individual
difference variables, interests were the most important channels through which gender
operated and influenced career success. The gender wage gap can be partly accounted for
by gender differences in interests, particularly large gender differences in Things and
People interests. Another responsible factor for the differential career success of men and
women was the number of children. The number of children had a particularly pronounced
effect for women’s career. Compared to women with no children or few children, women
with more children were in much less prestigious occupations and earned a substantially
lower income; the gap between women with more children and their male counterparts
69
were also much larger than that between women and men without children (or delayed
having children).
Although income and occupational prestige were both used as indicators for career
success, it is best to discuss them separately. Income and occupational prestige had
somewhat different determinants and the gender gaps in these two outcomes were of
different severity. As shown in Table 17, degree attainment appeared to be the single most
important determinant for occupational prestige. Meeting the threshold of educational
requirement is essential for entering prestigious occupations. Income, in contrast, seemed
to be influenced most strongly by interests and occupational characteristics as shown
previously. Of course, level of education was positively associated with income, yet the
education-income link was much weaker compared to the education-occupational prestige
link. In addition, gender had a sizable direct negative effect on income, whereas the direct
effect of gender on occupational prestige was much smaller. Overall, the gender gap in
income was much greater and was further widened by the gender differences in interests
and as the number of children increased. The gender gap in occupational prestige was
more benign, and after taking into account the indirect effects through interests, women
even had a small advantage.
Results from the present dissertation highlighted the role of interests, in particular
Things and People interests, for the differential career outcomes of men and women.
Previous studies have shown that the gender difference along a Thing-People interest
dimension was among the largest gender differences found in psychological variables (d =
.93; Su et al., 2009). Men have a strong preference for things-oriented careers and women
have a strong preference for people-oriented careers. This dissertation linked the gender
70
difference in interests to career outcomes. Specifically, Things interest was the strongest
mediator of the effect of gender on occupational prestige. Men tended to have higher
interest in Things that in turn led them to less prestigious occupations. The mediation of
People interest constituted the most profound indirect effect of gender on income.
Women’s overall stronger interest in People drew them to service type of occupations that
were not highly paid. It may also be a critical reason why women shy away from STEM
occupations that have higher pay yet are less likely to fulfill their interest in People.
Consistent with the findings from previous studies (e.g., Leslie, 2007, March; Mroz,
1987), this dissertation showed that number of children had differential effects on men’s
and women’s career. Women’s income and occupational prestige were more negatively
associated with the number of children they had. This result, however, was only
correlational. Therefore, it is possible that the number of children was a result rather than
a cause of career success, as heavier job responsibilities in a prestigious occupation might
lead individuals to delay having children; or it is possible that career success and number
of children were both results of a third variable such as level of education, as educated
individuals were more likely to attain more successful careers as well as to choose having
fewer children. More likely, family characteristics of individuals, including number of
children, might have a dynamic relationship and mutually influence with individuals’
career success over the span of their career development. Despite these alternative
explanations, the finding from the present dissertation highlighted the link between the
family and work spheres of individuals. The choice of having children or not and having
how many children is closely associated with individuals’ career pursuit and career
attainment, especially for women. This result, along with the findings on the mediation of
71
interests between gender and career outcomes, suggests that work and life preferences
may be among the most important factors for understanding men and women’s differential
career attainment and are promising directions for future research.
Lastly, the present dissertation provided evidence that degree attainment mediated
the effects of individual difference variables on income and occupational prestige. Note
that, for occupational prestige, many times the indirect effects of individual difference
variables through degree attainment were larger than the direct effects of these variables.
The results showed that degree attainment was an essential pathway for individuals to
realize their abilities and interests and to achieve career success, and in particular, to find
their doorways to prestigious occupations.
Limitations and Future Directions
The current dissertation used data from Project TALENT, a national longitudinal
survey. While this data set has many advantages for addressing the research questions of
this dissertation, including a large sample that represents individuals from diverse majors
and occupations, the longitudinal design, and an extensive pool of predictor variables,
readers are reminded that this data set represents the generation of individuals who were
in high school in 1960. The characteristics of these participants may be different from the
characteristics of current high school students and workforce. Results from the current
dissertation have yet to be cross-validated with a contemporary sample. Also, as previously
discussed, the effects of interests on the criteria may vary depending on how an interest is
measured. The interest measure in Project TALENT needs to be validated with another
well-established contemporary interest measure. Dr. Brent Roberts and his research team
at the University of Illinois are working on the validation of the Project TALENT interest
72
inventory and personality inventory. Longitudinal studies are needed in future research to
continue examining the relationship between interests and success at work and school.
Although no comparable contemporary data set is currently available for cross-
validating results from the present dissertation, some existing findings can be used to
evaluate the likelihood that these results will generate to a contemporary sample. First,
past research has shown that the structure of vocational interests is relatively stable over
time, despite changes in the mean levels of interests (Bubany & Hansen, 2011; Tracey &
Rounds, 1993). Second, gender differences in interests have also been shown to be
relatively consistent across generations (e.g., Hansen, 1988; Su et al., 2009). Su et al. (2009)
found that younger cohorts had smaller gender differences in Artistic interests and
Enterprising interests. However, gender differences in other interest areas were not
moderated by cohort. The large gender differences in the Things-People interest dimension
persisted across generations. Third, a reanalysis of the studies from the Nye et al. (2012)
meta-analysis that spanned over 60 years showed that the predictive validity of interests
for performance was not correlated with year of publication. These findings provide
indirect evidence that results from the present dissertation are likely to generalize to a
contemporary sample.
The latest time point when the participants in the present dissertation were
surveyed was 11 years after their high school graduation, or when participants were
around 29 years old. Career attainment and family characteristics of participants later in
their life were not observed. This means that the data were truncated to a certain extent,
especially for the individuals who entered the world of work at a later time point (e.g.,
participants who completed doctoral degrees) and those who planned to have children at
73
an older age. It is likely that the income of these individuals would grow years after they
entered the work force and become more reflective of their abilities, interests, and
educational attainment, and that the number of children of these individuals would also
increase. In turn, results may be somewhat different if a longer period of observation were
available. For example, the relationship between predictor variables and income may be
indeed stronger; and the relationship between number of children and career success may
be less dramatic. The time frame of this study needs to be taken into consideration when
interpreting the results from the current dissertation.
The current data set does not have information about participants’ job performance
or continuance intention on the job. Results on career success indicators and academic
performance criteria may have limited generalizability to these job performance criteria.
More studies are needed to examine the power of interests and interest congruence in
predicting job performance. A broad developmental perspective of performance adopted in
the present dissertation is a useful approach for studying and comparing the determinants
of performance in different contexts and at different stages in life.
Results of the present dissertation showed that interests are major channels
through which gender operates and influences career outcomes. Other demographic
variables such as SES and race/ethnicity are also influential, but examining their effects is
beyond the scope of this dissertation. Future research needs to examine the psychological
mechanisms through which other demographic variables affect academic achievement and
career success. In addition, given the power of interests for predicting career success and
the strong link between gender and interests, it is important to examine the developmental
process of interests and the origin of such large gender differences in interests. Future
74
research needs to identify the relationship and dynamic between interests and other
important parts of individuals’ identity, such as gender role and racial/ethnic identity, and
to examine how different parts of identity develop and impact individuals’ work outcomes
holistically.
75
CONCLUSION
Interests represent an essential part of individuals’ identity. Interest directs
individuals’ choices, motivates goal-oriented behaviors, energizes and sustains efforts until
goals are achieved. Traditionally, the role of interest for predicting performance had been
largely ignored in favor of cognitive ability and personality. Built on two recent meta-
analyses that supported the interest-performance link, the current dissertation showed
that interests had incremental validity for predicting academic achievement and career
success over and above ability and personality; interest congruence, in most cases, was also
positively associated with performance outcomes; and interests were important channels
through which gender affects career success. The power of interests in predicting
performance is substantial and extends across an individual’s education and career
development.
76
REFERENCES
Ackerman, P. L., & Heggestad, E. D. (1997). Intelligence, personality, and interests: Evidence
for overlapping traits. Psychological Bulletin, 121(2), 219–245. doi:10.1037/00332909.121.2.219
Ainley, M., Hidi, S., & Berndorff, D. (2002). Interest, learning, and the psychological
processes that mediate their relationship. Journal of Educational Psychology, 94(3),
545–561. doi:10.1037/0022-0663.94.3.545
Ajzen, I., & Fishbein, M. (1977). Attitude-behavior relations: A theoretical analysis and
review of empirical research. Psychological Bulletin, 84(5), 888-918.
doi:10.1037/0033-2909.84.5.888
Allen J., & Robbins, S. (2010). Effects of interest-major congruence, motivation, and
academic performance on timely degree attainment. Journal of Counseling
Psychology, 57(1), 23-35. doi:10.1037/a0017267
American College Testing Program. (1995). Technical manual: Revised Unisex Edition of the
ACT Interest Inventory (UNIACT). Iowa City, IA: Author.
Arnold, F. (1910). Attention and interest: A study in psychology and education. New York:
Macmillan.
Barrick, M. R., & Mount, M. K. (1991). The big five personality dimensions and job
performance: A meta-analysis. Personnel Psychology, 44(1), 1-26.
doi:10.1111/j.1744-6570.1991.tb00688.x
Barrick, M. R., & Mount, M. K. (2005). Yes, personality matters: Moving on to more
important matters. Human Performance, 18(4), 359-372.
doi:10.1207/s15327043hup1804_3
Barrick, M. R., Mount, M. K., & Gupta, R. (2003). Meta-analysis of the relationship between
the five-factor model of personality and Holland’s occupational types. Personnel
Psychology, 56(1), 45–74. doi:10.1111/j.1744-6570.2003.tb00143.x
Barrick, M. R., Mount, M. K., & Strauss, J. P. (1993). Conscientiousness and performance of
sales representatives: Test of the mediating effects of goal setting. Journal of Applied
Psychology, 78(5), 715–722. doi:10.1037/0021-9010.78.5.715
77
Barrick, M. R., Stewart, G. L., & Piotrowski, M. (2002). Personality and job performance:
Mediating effects of motivation among sales representatives. Journal of Applied
Psychology, 87(1), 43–51. doi:10.1037/0021-9010.87.1.43
Black, M. S. (1999). The efficacy of personality and interest measures as a supplement to
cognitive measures in the prediction of military training performance. Available from
ProQuest Dissertations and Theses database. (UMI No. MQ40348)
Blau, P. M., & Duncan, O. D. (1967). The American occupational structure. New York: Wiley.
Bollen, K. A. (1989). Structural equations with latent variables. New York: Wiley.
Borus, M. E. (Ed.). (1981). Pathways to the future: A report on the National Longitudinal
Survey of Youth labor market experience in 1979. Columbus, OH: Ohio State
University, Center for Human Resource Research.
Brown, S. D., & Gore, P. A., Jr. (1994). An evaluation of interest congruence indices:
Distribution of characteristics and measurement properties. Journal of Vocational
Behavior, 45(3), 310-327. doi:10.1006/jvbe.1994.1038
Bubany, S. T., & Hansen, J. C. (2011). Birth cohort change in the vocational interests of
female and male college students. Journal of Vocational Behavior, 78(1), 59-67.
doi: 10.1016/j.jvb.2010.08.002
Camp, C. C., & Chartrand, J. M. (1992). A comparison and evaluation of interest congruence
indices. Journal of Vocational Behavior, 41(2), 162-182. doi:10.1016/00018791(92)90018-U
Campbell, D. P. (1966). Manual for the Strong Vocational Interest Blank. Stanford, CA:
Stanford University Press.
Campbell, D. P. (1971). Handbook for the Strong Vocational Interest Blank. Stanford, CA:
Stanford University Press.
Campbell, J. P. (1990). Modeling the performance prediction problem in industrial
organizational psychology. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of
industrial and organizational psychology (Vol. 1, 2nd ed., pp. 687-732). Palo Alto, CA:
Consulting Psychologists Press.
Campbell, J. P., Gasser, M. B., & Oswald, F. L. (1996). The substantive nature of job
performance variability. In K. R. Murphy (Ed.), Individual differences and behavior in
organizations (pp. 258-299). Hillsdale, NJ: Erlbaum.
78
Card, D. (2001). Estimating the return to schooling: Progress on some persistent
econometric problems. Econometrica, 69(5), 1127-1160. doi:10.1111/14680262.00237
Carroll, J.B. (1993). Human cognitive abilities: A survey of factor-analytic studies. Cambridge,
England: Cambridge University Press.
Ceci, S. J., & Williams, W. M. (2007). Why aren’t more women in science? Top researchers
debate the evidence. Washington, DC: American Psychological Association.
Ceci, S. J., Williams, W. M., & Barnett, S. M. (2009). Women’s underrepresentation in science:
Sociocultural and biological considerations. Psychological Bulletin, 135(2), 218–261.
doi:10.1037/a0014412
Chipman, S. F., Brush, L., & Wilson, D. (Eds.). (1985). Women and mathematics: Balancing
the equation. Hillsdale, NJ: Lawrence Erlbaum Associates.
Cooley, W. W., & Lohnes, P. R. (1968). Predicting development of young adults: Project Talent
5-year follow-up studies. Pittsburgh, PA: American Institutes for Research.
Cronbach, L. J., & Gleser, G. C. (1953). Assessing the similarity between profiles.
Psychological Bulletin, 50(6), 456-473. doi:10.1037/h0057173
Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and self-determination in human
behavior. New York: Plenum.
Dewey, J. (1913). Interest and effort in education. Boston: Houghton Mifflin.
Dillon, S., & Rimer S. (2005, January 19). No break in the storm over Harvard president’s
words. The New York Times. Retrieved from
http://www.nytimes.com/2005/01/19/education/19harvard.html?ex_110714893
2&ei_1&en_4f97524a80c3e413
Dudley, N. M., Orvis, K. A., Lebiecki, J. A., & Cortina, J. M. (2006). A meta-analytic
investigation of conscientiousness in the prediction of job performance: Examining
the intercorrelations and the incremental validity of narrow traits. Journal of Applied
Psychology, 91(1), 40–57. doi:10.1037/0021-9010.91.1.40
Duncan, O. D. (1961). A socioeconomic index for all occupations. In A. J. Reiss (Ed.)
Occupations and social status (pp.109-138). New York: Free Press.
79
Dyer, E.D. (1987). Can university success and first-year job performance be predicted from
academic achievement, vocational interest, personality and biographical measures?
Psychological Reports, 61 (2), 655-671. doi:10.2466/pr0.1987.61.2.655
Echambadi, R., & Hess, J. D. (2005). Mean-centering does not alleviate collinearity problems
in moderated multiple regression models. Marketing Science, 26(3), 438-445.
doi:10.1287/mksc.1060.0263
Edwards, J. R. (1993). Problems with the use of profile similarity indices in the study of
congruence in organizational research. Personnel Psychology, 46(3), 641-665.
doi:10.1111/j.1744-6570.1993.tb00889.x
Edwards, J. R. (2002). Alternatives to difference scores: Polynomial regression analysis and
response surface methodology. In F. Drasgow & N. Schmitt (Eds.), Measuring and
analyzing behavior in organizations: Advances in measurement and data analysis (pp.
350-400). San Francisco, CA: Jossey-Bass.
Edwards, J. R., & Lambert, L. S. (2007). Methods for integrating moderation and mediation:
A general analytical framework using moderated path analysis. Psychological
Methods, 12(1), 1-22. doi:10.1037/1082-989X.12.1.1
Edwards, J. R., & Parry, M. E. (1993). On the use of polynomial regression equations as an
alternative to difference scores in organizational research. Academy of Management
Journal, 36(6), 1577-1613. doi:10.2307/256822
Ferriman, K., Lubinski, D., & Benbow, C. P. (2009). Work preferences, life values, and
personal views of top math/science graduate students and the profoundly gifted:
Developmental changes and sex differences during emerging adulthood and
parenthood. Journal of Personality and Social Psychology, 97(3), 517-532.
doi:10.1037/a0016030
Fisher, C. D., & Noble, C. S. (2004). A within-person examination of correlates of
performance and emotions while working. Human Performance, 17(2), 145–168.
doi:10.1207/s15327043hup1702_2
Fouad, N. A. (1999). Validity evidence for interest inventories. In M. L. Savickas & R. L.
Spokane (Eds.), Vocational interests: Meaning, measurement, and counseling use (pp.
193–209). Palo Alto, CA: Davis-Black.
80
Gallagher, A. M., & Kaufman, J. C. (2005). Gender differences in mathematics: An integrative
psychological approach. New York: Cambridge University Press.
Gellatly, I. R., Paunonen, S. V., Meyer, J. P., Jackson, D. N., & Goffin, R. D. (1991). Personality,
vocational interest, and cognitive predictors of managerial job performance and
satisfaction. Personality and Individual Differences, 12(3), 221-231.
doi:10.1016/0191-8869(91)90108-N
Ginther, D. K., & Kahn, S. (2006). Does science promote women? Evidence from academia
1973–2001. Unpublished manuscript. Retrieved from
http://www.nber.org/papers/w12691.pdf
Goldberg, L. R. (1999) A broad-bandwidth, public-domain, personality inventory measuring
the lower-level facets of several Five-Factor Models. In I. Mervielde, I. Deary, F. de
Fruyt, & F. Ostendorf (Eds.), Personality Psychology in Europe (Vol. 7, pp. 7-28).
Tilburg, The Netherlands: Tilburg University Press.
Goldberg, L. R. (2006). Doing it all Bass-Ackwards: The development of hierarchical factor
structures from the top down. Journal of Research in Personality, 40(4), 347-358.
doi:10.1016/j.jrp.2006.01.001
Gottfredson, G. D., & Holland, J. L. (1990). A longitudinal test of the influence of congruence:
Job satisfaction, competency utilization, and counterproductive behavior. Journal of
Counseling Psychology, 37(4), 389-398. doi:10.1037/0022-0167.37.4.389
Gottfredson, G. D., & Jones, E. M. (1993). Psychological meaning of profile elevation in the
Vocational Preference Inventory. Journal of Career Assessment, 1(1), 35-49.
doi:10.1177/106907279300100105
Guilford, J. P., Christensen, P. R., Bond, N. A., Jr., & Sutton, M. A. (1954). A factor analysis
study of human interests. Psychological Monographs, 68(4), 375-413.
doi:10.1037/h0093666
Guttman, L. (1957). Empirical verification of the radex structure of mental abilities and
personality traits. Educational and Psychological Measurement, 17(3), 391-407.
doi:10.1177/001316445701700306
Hakim, C. (2005). Sex differences in work–life balance goals. In D. Houston (Ed.), Work-life
balance in the twenty-first century (pp. 55–79). London: Palgrave Macmillan.
81
Hakim, C. (2006). Women, careers, and work–life preferences. British Journal of Guidance
and Counseling, 34(3), 279–294. doi:10.1080/03069880600769118
Hansen, J. C. (1984). The measurement of vocational interests. In S. D. Brown & R. W. Lent
(Eds.), Handbook of counseling psychology (pp.99–136). New York: Wiley.
Hansen, J. C. (1988). Changing interests of women: Myth or reality? Applied Psychology,
37(2), 133-150. doi: 10.1111/j.1464-0597.1988.tb01132.x
Harmon, L. W., Hansen, J. C., Borgen, F. H., & Hammer, A. L. (1994). Strong Interest Inventory
applications and technical guide. Stanford, CA: Stanford University Press.
Hauser, R. M., & Warren, J. R. (1997). Socioeconomic indexes for occupations: A review,
update, and critique. Sociological Methodology, 27(1), 177-298. doi:10.1111/14679531.271028
Hedges, L. V., & Nowell, A. (1995). Sex differences in mental test scores, variability, and
numbers of high-scoring individuals. Science, 269(5220), 41–45.
doi:10.1126/science.7604277
Herbermann, C. G., Pace, E. A., Pallen, C. B., Shahan, T. J., & Wynne, J. J. (Eds.). (1913). The
psychology of interests. The Catholic encyclopedia (Vol. 8, pp. 75-76). New York:
Gilmary Society.
Hidi, S. (2000). An interest researcher's perspective: The effects of extrinsic and intrinsic
factors on motivation. In C. Sansone & J. M. Harackiewicz (Eds.), Intrinsic and
extrinsic motivation (pp. 309-339). San Diego: Academic Press.
Hirschi, A. (2009). Development and criterion validity of differentiated and elevated
vocational interests in adolescence. Journal of Career Assessment, 17(4), 384-401.
doi:10.1177/1069072709334237
Hogan, R., & Blake, R. (1999). John Holland’s vocational typology and personality theory.
Journal of Vocational Behavior, 55(1), 41–56. doi:10.1006/jvbe.1999.1696
Holland, J. L. (1959). A theory of vocational choice. Journal of Counseling Psychology, 6(1),
35-45. doi:10.1037/h0040767
Holland, J. L. (1997). Making vocational choices: A theory of vocational personalities and
work environments (3rd ed.). Odessa, FL: Psychological Assessment Resources.
Holland, J. L., Fritzsche, B., & Powell, A. (1994). Self-Directed Search: Technical manual.
Odessa, FL: Psychological Assessment Resources.
82
Hollingshead A. B. (1965). Two-factor index of social position. New Haven, CT: Yale Press.
Humphreys, L. G., Lubinski, D., & Yao, G. (1993). Utility of predicting group membership and
the role of spatial visualization in becoming an engineer, physical scientist, or artist.
Journal of Applied Psychology, 78(2), 250–261. doi:10.1037/0021-9010.78.2.250
Hunter, J. E., & Hunter, R. F. (1984). Validity and utility of alternative predictors of job
performance. Psychological Bulletin, 96(1), 72-98. doi:10.1037/0033-2909.96.1.72
Hyde, J. S., & Linn, M. C. (2006). Gender similarities in mathematics and science. Science,
314(5799), 599–600. doi:10.1126/science.1132154
Jackson, D. N. (1977). Jackson Vocational Interest Survey manual. Port Huron, MI: Research
Psychologists Press.
Jacobs, J. E., Davis-Kean, P., Bleeker, M., Eccles, J. S., & Malanchuk, O. (2005). “I can, but I
don’t want to”: The impact of parents, interests, and activities on gender differences
in math. In A. M. Gallagher & J. C. Kaufman (Eds.), Gender differences in mathematics:
An integrative psychological approach (pp. 246–263). New York: Cambridge
University Press.
Jacobs, J. A., & Winslow, S. E. (2004). Overworked faculty: Job and stresses and family
demands. Annals of American Political and Social Scientist, 596(1), 104–129.
doi:10.1177/0002716204268185
John, O. P., Donahue, E. M., & Kentle, R. L. (1991). The Big Five Inventory--Versions 4a and
54. Berkeley, CA: University of California, Berkeley, Institute of Personality and
Social Research.
Johnson, J. W. (2000). A heuristic method for estimating the relative weight of predictor
variables in multiple regression. Multivariate Behavioral Research, 35(1), 1–19.
doi:10.1207/S15327906MBR3501_1
Johnson J. W. (2004). Factors affecting relative weights: the influence of sampling and
measurement error. Organizational Research Methods, 7(3), 283–299.
doi:10.1177/1094428104266018
Johnston, J. (1984). Econometric methods. New York: McGraw-Hill.
Jones, C., Clarke, M., Mooney, G., McWilliams, H., Crawford, I., Stephenson, B., & Tourangeau,
R. (1983). High school and beyond 1980 senior cohort: First follow-up (1982) data file
user's manual. Washington, DC: National Center for Educational Statistics.
83
Judge, T. A., & Bono, J. E. (2001). Relationship of core self-evaluations traits – self-esteem,
generalized self-efficacy, locus of control, and emotional stability – with job
satisfaction and job performance: A meta-analysis. Journal of Applied Psychology,
86(1), 80-92. doi:10.1037/0021-9010.86.1.80
Judge, T. A., Higgins, C., Thoresen, C. J., & Barrick, M. R. (1999). The Big Five personality
traits, general mental ability, and career success across the life span. Personnel
Psychology, 52(3), 621-652. doi:10.1111/j.1744-6570.1999.tb00174.x
Judge, T. A., Klinger, R. L., & Simon, L. S. (2010). Time is on my side: Time, general mental
ability, human capital, and extrinsic career success. Journal of Applied Psychology,
95(1), 92-107. doi:10.1037/a0017594
Kristof-Brown, A. L., Zimmerman, R. D., & Johnson, E. C. (2005). Consequences of
individuals’ fit at work: A meta-analysis of person-job, person-organization, persongroup, and person-supervisor fit. Personnel Psychology, 58(2), 281-342.
doi:10.1111/j.1744-6570.2005.00672.x
Kuncel, N. R., Hezlett, S., & Ones, D. S. (2001). A comprehensive meta-analysis of the
predictive validity of the Graduate Record Examinations: Implications for
graduate student selection and performance. Psychological Bulletin, 127(1), 6281. doi:10.1037/0033-2909.127.1.162
Kuncel, N. R., Hezlett, S., & Ones, D. S. (2004). Academic performance, career potential,
creativity, and job performance: Can one construct predict them all? Journal of
Personality and Social Psychology, 86(1), 148-161. doi:10.1037/00223514.86.1.148
Kuncel, N. R., Ones, D. S., & Sackett, P. R. (2010). Individual differences as predictors of
work, educational, and broad life outcomes. Personality and Individual Differences,
49(4), 331-336. doi:10.1016/j.paid.2010.03.042
Kutner, M. H., Nachtsheim, C. J., Neter, J., & Li, W. (2005). Applied linear statistical models
(5th ed.). New York: McGraw-Hil/Irwin.
Leboy, P. (2007, September 21). Why women leave academic medicine. Inside Higher
Education. Retrieved from www.insidehighered.com/news/2007/09/21/women
84
Leslie, D. W. (2007, March). The reshaping of America’s academic workforce. Research
Dialogue, 87. New York: TIAA-CREF. Retrieved from http://www.tiaa-
crefinstitute.org/ucm/groups/content/@ap_ucm_p_tcp_docs/documents/document
/tiaa02029384.pdf
Liao, H.-Y., Jin, J., Su, R., Tay, L., & Rounds, J. (2012, August). Finding the big factors in
interest: A factor analytic model of basic interests. Paper presented at the 120th
annual convention of the American Psychological Association, Orlando, FL.
Lorenzo-Seva, U., Ferrando, P. J., & Chico, E. (2010). Two SPSS programs for interpreting
multiple regression results. Behavioral Research Methods, 42(1), 29–35.
doi:10.3758/BRM.42.1.29
Lubinski, D. (2000). Scientific and social significance of assessing individual differences:
“Sinking shaft at a few critical points.” Annual Review of Psychology, 51(1), 405–444.
doi:10.1146/annurev.psych.51.1.405
Lubinski, D., Benbow, C. P., Shea, D. L., Eftekhari-Sanjani, H., & Halvorson, B. J. (2001). Men
and women at promise for scientific excellence: Similarity not dissimilarity.
Psychological Science, 12(4), 309–317. doi:10.1111/1467-9280.00357
Marshalek, B., Lohman, D. F., & Snow, R. E. (1983). The complexity continuum in the radex
and hierarchical models of intelligence. Intelligence, 7(2), 107-127.
doi:10.1016/0160-2896(83)90023-5
Mason, M. A., & Goulden, M. (2004). Marriage and baby blues: Redefining gender equity and
the academy. Annals of American Political and Social Scientist, 596(1), 86–103.
doi:10.1177/0002716204268744
Meece, J. L., Parsons, J. E., Kaczala, C. M., Goff, S. B., & Futterman, R. (1982). Sex differences
in math achievement: Toward a model of academic choice. Psychological Bulletin,
91(2), 324–348. doi:10.1037/0033-2909.91.2.324
Mincer, J. (1958). Investment in human capital and personal income distribution. Journal of
Political Economy, 66(4), 281-302. doi:10.1086/258055
Morris, M. A. (2003). A meta-analytic investigation of vocational interest-based job fit, and
its relationship to job satisfaction, performance, and turnover. Available from
ProQuest Dissertations and Theses database. (UMI No. 3089788)
85
Mroz, T. A. (1987). The sensitivity of an empirical model of married women's hours of work
to economic and statistical assumptions. Econometrica, 55(4), 765-799.
doi:10.2307/1911029
Nakao, K., & Treas, J. (1992). The 1989 socioeconomic index of occupations: Construction
from the 1989 occupational prestige scores (GSS Methodological Report No. 74).
Chicago: National Opinion Research Center.
Nye, C. D., Su. R., Rounds, J., & Drasgow, F. (2012). Vocational interests and performance: A
quantitative summary of over 60 years of research. Perspectives on Psychological
Science, 7, 384-403. doi:10.1177/1745691612449021
Parsons, C. K. (1909). Choosing a vocation. Boston: Houghton-Mifflin.
Pinker, S. (2008). The sexual paradox: Men, women, and the real gender gap. New York:
Macmillan.
Ployhart, R. E., Schneider, B., & Schmitt, N. (2006). Staffing organizations: Contemporary
practice and theory (3rd ed.). Mahwah, NJ: Erlbaum.
Prediger, D. J. (1998). Is interest profile level relevant to career counseling? Journal of
Counseling Psychology, 45(2), 204-211. doi:10.1037/0022-0167.45.2.204
Psychological Assessment Resources (2009). Publisher website. Retrieved from
http://www.self-directed-search.com/
Reiss, A. J., Jr. (1961). Occupations and social status. New York: Free Press of Glencoe.
Renninger, K. A., Hidi, S., & Krapp, A. (1992). The role of interest in learning and
development. Hillsdale, NJ: Erlbaum.
Roberts, B. W., & Wood, D. (2006). Personality development in the context of the neo-
socioanalytic model of personality. In D. Mroczek & T. Little (Eds.), Handbook of
personality development (pp. 11-39). Mahwah, NJ: Lawrance Erlbaum Associates.
Robertson, K. F., Smeets, S., Lubinski, D., & Benbow, C. P. (2010). Beyond the threshold
hypothesis: Even among the gifted and top math/science graduate students,
cognitive abilities, vocational interests, and lifestyle preferences matter for career
choice, performance, and persistence. Current Directions in Psychological Science,
19(6), 346-351. doi:10.1177/0963721410391442
86
Rounds, J. (1995). Vocational interests: Evaluation of structural hypotheses. In D. Lubinski
& R. V. Dawis (Eds.), Assessing individual differences in human behavior: New
concepts, methods, and findings (pp. 177–232). Palo Alto, CA: Consulting
Psychologists Press.
Sansone, C., & Smith, J. L. (2000). Interest and self–regulation: The relation between having
to and wanting to. In C.Sansone & J. M. Harackiewicz (Eds.), Intrinsic and extrinsic
motivation (pp. 341-372). San Diego, CA: Academic.
Savickas, M. L. (1999). The psychology of interests. In M. L. Savickas & A. R. Spokane (Eds.),
Vocational interests: Meaning, measurement, and counseling use (pp. 19-56). Palo
Alto, CA: Davies-Black.
Schmidt, F. L., & Hunter, J. E. (1998). The validity and utility of selection methods in
personnel selection: Practical and theoretical implications of 85 years of research
findings. Psychological Bulletin, 124(2), 262–274. doi:10.1037/0033-2909.124.2.262
Schmidt, F. L., Hunter, J. E., & Outerbridge, A. N. (1986). The impact of job experience and
ability on job knowledge, work sample performance, and supervisory ratings of job
performance. Journal of Applied Psychology, 71(3), 432–439. doi:10.1037/00219010.71.3.432
Schneider, B. (1987). The people make the place. Personnel Psychology, 40(3), 437–453. doi:
10.1111/j.1744-6570.1987.tb00609.x
Siegel, P. M. (1971). Prestige in the American occupational structure. Available from
ProQuest Dissertations and Theses database. (UMI No. T-22336)
Silvia, P. J. (2006). Exploring the psychology of interest. New York: Oxford University Press.
Snow, R. E. (1978). Theory and method for research on aptitude processes. Intelligence,
2(3), 225-278. doi:10.1016/0160-2896(78)90019-3
Sproat, K. V., Churchill, H., & Sheets, C. (1985). The National Longitudinal Surveys of labor
market experience: An annotated bibliography of research. Lexington, MA: Lexington.
Stevens, G., & Featherman, D. L. (1981). A revised socioeconomic index of occupational
status. Social Science Research, 10(4), 364-395.doi:10.1016/0049-089X(81)90011-9
Strong, E. K. Jr. (1943). Vocational interests of men and women. Stanford, CA: Stanford
University Press.
87
Su, R., Rounds, J., & Armstrong, P. I. (2009). Men and things, women and people: A metaanalysis of sex differences in interests. Psychological Bulletin, 135(6), 859-884.
doi:10.1037/a0017364
Tay, L., Su, R., & Rounds, J. (2011). People-Things and Data-Ideas: Bipolar dimensions?
Journal of Counseling Psychology, 58(3), 424-440. doi:10.1037/a0023488
Tett, R. P., Jackson, D. N., & Rothstein, M. (1991). Personality measures as predictors of job
performance: A meta-analytic review. Personnel Psychology, 44(4), 703-742.
doi:10.1111/j.1744-6570.1991.tb00696.x
Tierney, J. (2008, July 15). A new frontier for Title IX: Science. The New York Times.
Retrieved from http://www.nytimes.com/2008/07/15/science/15tier.html?_r_2
Tinsley, H. E. A. (2000). The congruence myth: An analysis of the efficacy of the person—
environment fit model. Journal of Vocational Behavior, 56(2), 147-179.
doi:10.1006/jvbe.1999.1727
Tharenou, P. (1997). Managerial career advancement. In C.L. Cooper & I.T. Robertson
(Eds.), International Review of Industrial and Organizational Psychology (pp. 40–93).
New York: John Wiley & Sons.
Thurstone, L. L. (1931). A multiple factor study of vocational interests. Personnel Journal,
10(3), 198-205.
Torr, D. V. (1953). A factor analysis of 40 interest variables. USAF Human Resources
Research Center Research Bulletin, 53-67, iv, 36.
Tracey, T. J., & Rounds, J. B. (1993). Evaluating Holland’s and Gati’s vocational interest
models: A structural meta-analysis. Psychological Bulletin, 113(2), 229–246.
doi:10.1037/0033-2909.113.2.229
U.S. Bureau of Census, (1971). Classified index of industries and occupations: 1970 census of
population. Washington, DC: Government Printing Office.
Van Iddekinge, C. H., Putka, D. J., & Campbell, J. P. (2011). Reconsidering vocational
interests for personnel selection: The validity of an interest-based selection test in
relation to job knowledge, job performance, and continuance intentions. Journal of
Applied Psychology, 96(1), 13-33. doi:10.1037/a0021193
88
Van Iddekinge, C. H., Roth, P. L., Putka, D. J., & Lanivich, S. E. (2011). Are you interested? A
meta-analysis of relations between vocational interests and employee performance
and turnover. Journal of Applied Psychology, 96(6), 1167-1194.
doi:10.1037/a0024343
Wai, J., Lubinski, D., & Benbow, C. P. (2009). Spatial ability for STEM domains: Aligning over
50 years of cumulative psychological knowledge solidifies its importance. Journal of
Educational Psychology, 101(4), 817-835. doi:10.1037/a0016127
Warwas, J., Nagy, G., Watermann, R. & Hasselhorn, M. (2009). The Relations of Vocational
Interests and Mathematical Literacy: On the Predictive Power of Interest Profiles.
Journal of Career Assessment, 17(4), 417-438. doi:10.1177/1069072709339284
Watt, H. M. G., & Eccles, J. S. (2008). Gender and occupational outcomes: Longitudinal
assessments of individual, social, and cultural influences. Washington, DC: American
Psychological Association.
Webb, R. M., Lubinski, D., & Benbow, C. P. (2002). Mathematically facile adolescents with
math–science aspirations: New perspectives on their educational and vocational
development. Journal of Educational Psychology, 94(4), 785–794. doi:10.1037/00220663.94.4.785
Weinberger, C. J., & Kuhn, P. J. (2010). Changing levels or changing slopes? The narrowing
of the U.S. gender earnings gap, 1959-1999. Industrial and Labor relations reviews,
63(3), 384–406.
Welland, J. D. (1976). An Investigation of the Role of Abilities in the Individual Earnings
Function. Available from ProQuest Dissertations and Theses database. (UMI No.
7712872)
Willis, R. J., & Rosen, S. (1979). Education and self-selection. Journal of Political Economy, 87
(5, Part 2), S7-S36. doi:10.1086/260821
Wise, L. L., McLaughlin, D. H., & Steel, L. (1979). The Project TALENT Data Bank Handbook,
Revised. Palo Alto, CA: American Institutes for Research.
Wise, L. L., Steel, L., & MacDonald, C. (1979). Origins and career consequences of sex
differences in high school mathematics achievement (Report to the National Institute
of Education on its Grant NIE-G-78–001). Palo Alto, CA: American Institutes for
Research.
89
Success in dental
school indicated by
GPA
High school
graduates vs.
dropouts
College graduates vs.
dropouts
College graduates
who attended
graduate school vs.
those who did not
Combs &
Cooley
(1968)
Cooley &
Lohnes
(1968)
Cooley &
Lohnes
(1968)
Performance
Criterion
Military training
performance and
successful training
course completion
Chen,
Podshadley,
& Shrock
(1967)
Black
(1999)*
Reference
Project TALENT
grade 12 males
Project TALENT
grade 12 males
Project TALENT
grade 9 males and
females
68 students in
sophomore class
of the University
of California
School of
Dentistry
264 Canadian
military trainees in
an 8-week training
program
Sample
Project TALENT
Interest Inventory
scales
Project TALENT
Interest Inventory
scales
Project TALENT
Interest Inventory
scales
The dentist scale from
the Kuder Preference
Record, Form D
Interest
Measure
Canadian Work
Preference Inventory
(CWPI), assessing
Methodical, Objective,
Innovative, Directive,
and Social types of
interests
90
Project TALENT
ability scales
Project TALENT
ability scales
Project TALENT
ability scales
Dental Aptitude Test
(DAT) scores and the
California
Performance Test
(CPT)
Ability
Measure
Canadian Forces
Aptitude Test (CFAT),
measuring verbal
skills, spatial ability,
and problem-solving
skills
Project TALENT Student
Activities Inventory
Project TALENT Student
Activities Inventory
Project TALENT Student
Activities Inventory
Personality
Measure
Measure of Personal
Attributes (MPA),
assessing work
orientation, dominance,
dependability,
adjustment,
cooperativeness, internal
control and physical
condition
California Psychological
Inventory (CPI)
9 factors were extracted from the
predictor variables and among them two
have significant loadings on dental GPA.
Kuder dentist scale had a positive
loading of .148 though non-significant
loading on Factor I, labeled manual skill
factor.
Interest, ability, personality were
examined separately in this study.
Significant predictors of high school
graduation for males are sports and
STEM interests; significant predictors for
females are social service and
biomedical interests. Labor and skilled
trades interests predict high school
dropout for both genders.
Only canonical correlations were
reported in the study. Science interest
positively predicts college persistence (r
= .21), interest in outdoors/shop
activities was also a significant predictor
(r = .20).
Only canonical correlations were
reported in the study. Cultural interest
positively predicts likelihood to attend
graduate school (r = .16), whereas
interest in outdoors/shop activities was
a negative predictor (r = -.11).
CWPI total score (yet not the subscale
scores) has incremental validity over
CFAT and MPA (ΔR2 = .06) for training
performance; addition of neither CWPI
total score nor subscale scores
accounted for significant larger variance
in success of training course completion
over CFAT and MPA.
Results
Table 1. Studies on the predictive validity of interests, ability, and personality for work and academic performance.
TABLES
Managerial job
performance
measured by
supervisor ratings on
a range of criteria
Academic
performance
indicated by
cumulative GPA &
job performance
assessed by
supervisor ratings
Kanfer, Wolf,
Kantrowitz,
& Ackerman
(2010)
Performance
Criterion
Success in a clinical
psych graduate
program indicated by
GPA, professional
qualifying exam
scores, practicum
evaluations,
internship
evaluations, peer
ratings, & faculty
ratings
Supervisor rating of
nursing performance
Gellatly,
Paunonen,
Meyer,
Jackson, &
Goffin
(1991)*
Dyer (1987)*
Daehnert &
Carter
(1987)
Reference
Table 1 (cont.)
105 college
students enrolled
in a cooperative
school-work
program in
science &
engineering
59 unit managers
from a large foodservice
organization
350 College of
Nursing
sophomore
students
192 student at a
graduate program
in clinical psych
Sample
Unisex Edition of the
American College
Testing Interest
Inventory (UNIACT;
ACT, 1981)
Career Directions
Inventory (CDI;
Jackson, 1986)
Strong-Campbell
Interest Inventory
Form T-325 (SCII)
Interest
Measure
Social Service scale
from the Strong
Vocational Interest
Blank (SVIB)
91
A range of cognitive
ability measures that
assess verbal,
numerical, & spatial
abilities (see note for
details**)
Personnel
Assessment Form
(PAF) measuring
quantative aptitude
and general
information
ACT, High school
GPA, university GPA,
nursing GPA
Ability
Measure
Graduate Record
Examinations (GRE;
verbal, quantitative,
and the psychology
test) and the Miller
Analogies Test (MAT)
NEO-FFI (Costa &
McCrae, 1992)
Personality Research
Form-E (PRF-E; Jackson,
1984)
California Psychological
Inventory
Personality
Measure
Minnesota Multiphase
Personality Inventory
(MMPI)
SCII social, realistic, art, & military
activities interests have incremental
validity for nursing process; teaching
interest for “organize work and use
time,” nature interest for “help other
understand scientific principles or logic
of situation,” social interest for “suggest
creative solutions to problems.”
5 interest factors were derived and a
number of them were related to
performance criteria controlling for
ability and personality: administration
interest predicts ratings on conducting
routine job tasks (r = .27, ΔR2 = .05) &
promotability (r = .28, ΔR2 = .02); social
service interest predicts ratings on
verbal communication (r = .28, ΔR2 =
.07); outdoor interest predicts ratings on
training & managing unite personnel (r =
.31, ΔR2 = .06).
Non-cognitive measures were factor
analyzed and only results for their
composites were reported. Social
Orientation/Communiation (Social
Interest loaded heavily on this factor)
has incremental validity over cognitive
abilities for overall job performance (r =
.26), procedural job performance (r =
.33), and non-technical performance (r =
.39), but not for academic performance.
Only correlations were reported. The
SVIB Social Service scale was not found
to be a significant predictor for the
performance criteria.
Results
Performance ratings
from plant
superintendent &
two assistant
superintendents
College entrance test
scores or Iowa High
School Content
Examination scores
Reed (1941)
Performance
Criterion
Job performance
indicated by
supervisor ratings on
overall & specific
aspects; faculty
ratings on
professional
achievement;
income; and self
ranking
Manager
performance
assessed by district
manager ratings
Poe & Berg
(1952)
Olds (1993)
Muchinsky &
Hoyt (1974)
Reference
Table 1 (cont.)
259 freshmen at
Fort Hays Kansas
State College &
266 high school
students living
near the college
33 production
supervisors in a
steel plant
68 managers and
assistant
managers at a
local restaurant
chain
138 individuals
who graduated
from the College
of Engineering of
Kansas State
University
Sample
Thurstone Vocational
Interest Schedule
Strong Vocational
Interest Blank (SVIB)
Vocational Preference
Inventory (VPI)
Interest
Measure
Strong Vocational
Interest Blank (SVIB):
STEM interests and
social service interests
92
California Short-Form
Test of Mental
Maturity;
Adaptability Test;
Survey of Space
Relations; Test of
Mechanical
Comprehension
Form BB; Survey of
Mechanical Insight
Stenquist Mechanical
Aptitude Test
Thurstone Test of
Mental Alertness
(Thurstone &
Thurstone, 1952)
measuring verbal,
math, & general
abilities
Ability
Measure
Achievement and
Affiliation scales
from the Edwards
Personal Preference
Schedule (EPPS)
Pernreuter Personality
Inventory
Fundamental
Interpersonal Relations
Orientations Behavior
(Schutz, 1977),
Differential Factors
Opinion Survey (Guilford
& Zimmerman, 1954),
Guilford-Zimmerman
Temperament Survey
(Guilford & Zimmerman,
1976)
Bernreuter Personality
Inventory
Personality
Measure
Quantitative and
Linguistic scores from the
American Council on
Education Psychological
Examination (ACE; ETS,
1950)
Only correlations were reported:
Interest in descriptive activities predicts
achievement in English literature (r =
.22), interest m academic activities
predicts average high-school grade (r =
.25) & college aptitude (r = .22)), and
interest in athletic activities predicts
achievement in physical science (r = .25).
Only t-test was conducted comparing
high- and low-performance supervisors.
Among the SVIB scales, only the Aviator
scale significantly distinguished two
groups.
Only correlations were reported. Social
service interests are significant
predictors for faculty ratings on
professional achievement (r = .27), selfranking of performance among
individuals with similar
training/experience (.34) and among
classmates (.32); STEM interests had
correlation coefficients of similar size
but none were significant.
Correlations for vocational interests
were not reported.
Results
Technical knowledge,
interpersonal
knowledge, task
proficiency, effort,
continuation
intentions
Performance
Criterion
First-year GPA, BARS
self-rated
performance, class
absenteeism, intent
to quit, OCBs
418 U.S. Army
soldiers
2771 freshmen
students at 10 U.S.
colleges and
universities
Sample
Work Preference
Survey
Interest
Measure
Derived from a
biodata measure. Two
scales reflect
vocational interest:
learning/intellectual
interest, artistic
interest
Armed Forces
Qualification Test
(AFQT)
Ability
Measure
SAT; ACT
Work Suitability
Inventory (WSI)
Personality
Measure
Derived from a biodata
measure
Only correlations were reported:
intellectual & artistic interests both are
significant predictors for first-year GPA
(r = .13/.21) and BARS self-rated
performance (r = .26/.24), intellectual
interest predicts lower class
absenteeism (r = -.07), and artistic
interest predicts lower intent to quit (r =
-.09).
Interests have incremental validity for all
criteria: ΔR2 = .06 for technical
knowledge, ΔR2 = .07 for interpersonal
knowledge, ΔR2 = .03 for task
proficiency, ΔR2 = .04 for effort, and ΔR2
= .08 for continuation intentions
Results
93
Note. * Four studies that examined the incremental validity of interests beyond abilities and personality
** Eleven tests were administered to provide an assessment of verbal, numerical, and spatial abilities. Three measures were included to assess verbal ability: Extended Range
Vocabulary (a multiplechoice vocabulary test; Educational Testing Service [ETS] Kit; Ekstrom, French, Harman, & Derman, 1976), Multidimensional Aptitude Battery (MAB)
Similarities (a test of verbal knowledge; Jackson, 1985), and MAB Comprehension (a test of cultural knowledge; Jackson, 1985). Four measures were included to assess numerical
ability: Number Series (a test of inductive reasoning from the Primary Mental Abilities battery; Thurstone, 1962), Math Approximation (a test that requires examinee to arrive at an
estimated answer and then choose from among five possible answers; see Ackerman, Beier, & Boyle, 2002), Arithmetic (a test that presents a series of math problems requiring a
variety of operations such as adding two fractions and reducing to the lowest term; Cureton & Cureton, 1955), and Math Knowledge (a wide-range test of mathematical knowledge,
from simple computation to algebra, geometry, and other advanced topics; see Ackerman & Kanfer, 1993). Four tests were included to assess spatial abilities: Paper Folding (a test
that requires the determination of where holes would appear in a piece of paper after it is unfolded; ETS Kit; Ekstrom et al., 1976), Cube Comparison (a test that requires figuring out
whether, after rotation, two cubes are the same or different; ETS Kit; Ekstrom et al., 1976), Spatial Orientation (a test of three dimensional visualisation; see Ackerman & Kanfer,
1993), and Spatial Analogy (a test of analogical reasoning using figural stimuli; see Ackerman & Kanfer, 1993).
Van
Iddekinge,
Putka, &
Campbell
(2011)*
Schmitt,
Oswald, Kim,
Imus,
Merritt,
Friede, &
Shivpuri
(2007)
Reference
Table 1 (cont.)
Table 2. Number of participants at four data collection points, in total, by gender, and by
race/ethnicity.
st
th
th
Base Year
1 year follow-up
5 year follow-up
11 year follow-up
377,016
182,545
122,557
96,757
Male
188,174 (49.9%)
87,090 (47.7%)
60,562 (49.4%)
47,535 (49.1%)
Female
188,841 (50.1%)
95,455 (52.3%)
61,995 (50.6%)
49,222 (50.9%)
White/Caucasians
Total N
147,471 (94.4%)
111,122 (95.2%)
120,266 (94.9%)
86,664 (95.0%)
Black/African Americans
6,549 (4.19%)
4,047 (3.47%)
4,897 (3.86%)
3,071 (3.36%)
Asian Americans
1,022 (0.65%)
836 (0.72%)
866 (0.68%)
508 (0.56%)
Native Americans
415 (0.27%)
291 (0.25%)
296 (0.23%)
317 (0.35%)
Latino(a)/Hispanic
625 (0.40%)
401 (0.34%)
328 (0.26%)
622 (0.68%)
88 (0.06%)
66 (0.06%)
54 (0.04%)
88 (0.10%)
Other
94
Table 3. Measurement time point of demographics, predictors, outcomes, and moderators.
Base Year
Sex
st
1 year follow-up
th
11 year follow-up
X
X
X
Race/Ethnicity
SES
X
Cohort
X
Cognitive abilities
X
Personality
X
Interests
X
1st year grade in college
th
5 year follow-up
X
Overall grade in college
X
Major grade in college
X
College persistence
X
Degree attainment
X
Income
X
Occupational prestige
X
College major
X
X
Occupational field
X
Number of children
X
95
Variable
Sex
White
Black
Asian
Native
Latino(a)
SES
Cohort
Verbal ability
Math ability
Spatial ability
Extroversion
Agreeableness
Impulsiveness
Vigor
Calmness
Culture
Leadership
Emotional Stability
Conscientiousness
Things interest
Artistic interest
Science interest
People interest
Business interest
Leadership interest
Income
Occupational prestige
Degree attainment
First year grade in college
Overall grade in college
Major grade in college
College persistence
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
0.50
0.94
0.04
0.01
0.00
0.00
97.73
10.40
-0.01
-0.01
0.00
6.55
4.55
1.94
3.58
4.15
5.11
1.28
5.03
16.48
2.19
2.80
2.72
2.77
2.84
2.48
48.03
46.83
6.99
3.53
8.71
7.62
0.60
Mean
0.50
0.23
0.20
0.08
0.05
0.06
10.20
1.11
0.91
0.92
0.80
2.94
2.36
1.63
2.13
2.51
2.37
1.35
2.46
7.28
0.85
0.88
0.93
0.92
0.80
1.21
9.02
21.34
1.97
0.95
1.86
1.77
0.49
SD
377015
156170
156170
156170
156170
156170
358030
377015
375167
376572
367891
370070
370070
370070
370070
370070
370070
370070
370070
370070
368026
367988
368021
367961
368031
363234
61360
61613
84883
85855
68829
69229
49453
N
-0.03
0.04
0.00
0.00
0.00
-0.01
0.01
0.07
-0.12
-0.26
0.16
0.25
0.02
0.00
0.08
0.26
0.03
0.02
0.14
-0.63
0.28
-0.23
0.74
-0.12
-0.27
-0.50
-0.10
-0.21
0.09
0.15
0.18
-0.09
1
----0.20
0.03
0.26
0.21
0.23
0.06
0.03
0.02
0.06
0.04
0.01
-0.05
0.03
0.03
-0.03
-0.08
-0.06
-0.09
-0.07
-0.06
0.06
0.08
0.06
0.05
0.04
0.04
0.04
2
----0.19
-0.02
-0.28
-0.22
-0.25
-0.06
-0.03
-0.01
-0.05
-0.03
0.00
0.06
-0.02
-0.02
0.02
0.09
0.06
0.09
0.07
0.06
-0.06
-0.08
-0.07
-0.04
-0.04
-0.03
-0.04
3
96
---0.02
0.00
0.00
0.01
0.02
-0.01
-0.01
-0.01
-0.03
-0.01
-0.02
-0.01
-0.01
-0.02
0.01
0.00
0.03
0.00
0.00
0.01
0.01
0.02
0.03
-0.01
-0.01
-0.01
0.02
4
--0.03
-0.02
-0.02
-0.02
0.00
-0.01
-0.01
0.00
-0.01
-0.01
-0.01
0.00
-0.01
-0.01
0.01
0.01
0.01
0.01
0.00
0.00
0.00
-0.01
0.00
-0.01
-0.01
-0.01
-0.01
5
-0.06
0.00
-0.05
-0.05
-0.04
-0.01
-0.01
-0.01
-0.02
-0.01
-0.01
-0.01
-0.01
-0.01
0.01
0.01
-0.01
0.02
0.01
0.00
-0.02
-0.04
-0.03
-0.02
0.00
-0.01
-0.04
6
0.09
0.44
0.41
0.33
0.15
0.18
0.09
0.18
0.18
0.19
0.15
0.16
0.18
-0.13
0.10
0.14
-0.10
0.01
0.07
0.14
0.36
0.41
0.11
0.11
0.10
0.22
7
0.29
0.22
0.17
0.08
0.14
0.04
0.06
0.15
0.13
0.04
0.10
0.15
0.02
0.07
0.03
0.02
0.11
0.02
-0.09
0.00
-0.01
0.15
-0.02
-0.01
0.02
8
0.76
0.60
0.14
0.25
0.07
0.18
0.26
0.22
0.08
0.21
0.27
-0.16
0.19
0.21
-0.05
0.00
0.04
0.07
0.43
0.46
0.33
0.24
0.30
0.32
9
0.64
0.06
0.14
0.05
0.15
0.22
0.13
0.10
0.18
0.22
-0.02
0.11
0.33
-0.18
0.05
0.12
0.19
0.47
0.53
0.31
0.21
0.27
0.35
10
0.01
0.04
0.02
0.11
0.14
0.02
0.01
0.12
0.10
0.14
0.00
0.24
-0.27
0.00
0.05
0.21
0.29
0.35
0.15
0.08
0.09
0.18
11
0.52
0.24
0.51
0.44
0.46
0.35
0.38
0.46
-0.15
0.10
0.02
0.17
0.13
0.07
-0.02
0.07
0.02
0.00
0.03
0.01
0.00
12
0.22
0.43
0.58
0.62
0.40
0.30
0.62
-0.20
0.26
0.10
0.23
0.10
0.05
-0.07
0.12
0.09
0.10
0.11
0.11
0.05
13
0.25
0.16
0.20
0.26
0.12
0.19
0.01
0.10
0.04
0.00
0.04
0.07
0.00
0.02
0.02
0.01
0.03
0.02
-0.02
14
Table 4. Means, standard deviations, and correlations of demographics, abilities, personality, interests, and outcomes.
0.43
0.43
0.40
0.32
0.52
0.01
0.10
0.16
0.01
0.10
0.11
0.06
0.12
0.12
0.03
0.06
0.04
0.05
15
0.53
0.38
0.44
0.64
-0.08
0.14
0.15
0.08
0.11
0.08
0.03
0.17
0.15
0.11
0.10
0.11
0.10
16
Variable
Sex
White
Black
Asian
Native
Latino(a)
SES
Cohort
Verbal ability
Math ability
Spatial ability
Extroversion
Agreeableness
Impulsiveness
Vigor
Calmness
Culture
Leadership
Emotional Stability
Conscientiousness
Things interest
Artistic interest
Science interest
People interest
Business interest
Leadership interest
Income
Occupational prestige
Degree attainment
First year grade in college
Overall grade in college
Major grade in college
College persistence
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Table 4 (cont.)
0.50
0.94
0.04
0.01
0.00
0.00
97.73
10.40
-0.01
-0.01
0.00
6.55
4.55
1.94
3.58
4.15
5.11
1.28
5.03
16.48
2.19
2.80
2.72
2.77
2.84
2.48
48.03
46.83
6.99
3.53
8.71
7.62
0.60
Mean
0.50
0.23
0.20
0.08
0.05
0.06
10.20
1.11
0.91
0.92
0.80
2.94
2.36
1.63
2.13
2.51
2.37
1.35
2.46
7.28
0.85
0.88
0.93
0.92
0.80
1.21
9.02
21.34
1.97
0.95
1.86
1.77
0.49
SD
377015
156170
156170
156170
156170
156170
358030
377015
375167
376572
367891
370070
370070
370070
370070
370070
370070
370070
370070
370070
368026
367988
368021
367961
368031
363234
61360
61613
84883
85855
68829
69229
49453
N
0.41
0.31
0.65
-0.22
0.33
0.13
0.24
0.12
0.09
-0.08
0.15
0.13
0.12
0.14
0.15
0.08
17
0.30
0.47
-0.03
0.21
0.18
0.06
0.16
0.20
0.05
0.17
0.17
0.11
0.13
0.13
0.10
18
0.40
-0.07
0.10
0.12
-0.01
0.07
0.10
0.04
0.14
0.13
0.08
0.09
0.08
0.07
19
97
-0.13
0.19
0.18
0.14
0.14
0.10
0.01
0.19
0.17
0.17
0.19
0.22
0.14
20
0.03
0.34
-0.25
0.37
0.35
0.24
-0.12
-0.01
-0.08
-0.13
-0.14
-0.03
21
0.41
0.47
0.42
0.40
-0.16
0.15
0.15
0.16
0.17
0.19
0.09
22
-0.02
0.39
0.47
0.18
0.30
0.34
0.10
0.05
0.09
0.18
23
0.28
-0.02
-0.42
-0.13
-0.24
0.06
0.10
0.13
-0.10
24
0.59
0.09
0.09
0.08
0.02
-0.01
-0.01
0.06
25
0.18
0.19
0.24
0.05
0.03
0.03
0.14
26
0.30
0.27
-0.01
0.02
0.01
0.17
27
0.71
0.20
0.26
0.27
0.60
28
0.24
0.32
0.34
0.75
29
0.35
0.47
0.20
30
0.69
0.30
31
0.30
32
Table 5. Ratings of occupational characteristics on six interest areas.
Job title
Government Worker
Proprietor; Contractor; In Business for
Self
Researcher (NEC)
Mathematician
Other Stat. (Including Actuary, etc.)
Things
0.11
Artistic
0.09
Science
0.33
People
-0.24
Business
0.23
Leadership
0.18
0.47
-0.21
0.14
-0.66
0.34
0.30
-0.11
0.07
0.48
-0.52
-0.14
0.29
0.11
-0.09
1.13
-0.50
0.15
0.33
-0.05
-0.04
0.74
-0.38
0.22
0.23
Systems Analyst (Computer)
0.23
-0.07
0.84
-0.64
0.24
0.30
Computer Programmer
0.20
-0.14
0.62
-0.57
0.13
0.20
Computer Specialist (NEC)
0.34
-0.39
0.99
-0.90
0.04
0.25
Scientist or Physical Scientist (NEC)
0.54
-0.21
0.79
-0.77
-0.23
0.07
Chemist
0.22
-0.17
1.19
-0.76
-0.03
0.16
Physicist
0.21
0.25
1.44
-0.94
-0.10
0.42
Geologist
0.41
-0.10
0.95
-0.95
-0.20
0.09
Engineer (NEC)
0.55
-0.40
1.10
-0.92
0.10
0.33
Civil and/or Hydraulic Engineer
0.71
-0.40
0.80
-0.86
0.15
0.22
Electrical and/or Electronic Engineer
0.50
-0.44
1.10
-0.94
-0.09
0.12
Mechanical or Automotive Engineer
0.68
-0.46
0.93
-0.95
0.01
0.24
Aeronautical Engineer
0.45
-0.16
1.13
-0.91
0.08
0.32
Chemical Engineer
0.30
-0.50
1.19
-1.04
0.02
0.21
Architect
0.35
0.04
0.78
-0.94
0.05
0.20
Lab Technician or Research Assistant in
Physical Science, Etc.
Biologist, Zoologist, Botanist,
Paleontologist
Specialist in Agricultural Science
0.59
-0.38
0.49
-0.79
-0.11
0.06
-0.12
0.08
1.07
-0.53
-0.19
0.14
0.64
-0.34
0.60
-0.74
-0.03
0.17
Physician, General Practitioner
-0.10
0.13
1.39
-0.81
0.14
0.63
Dentist
0.27
-0.17
1.11
-0.76
0.43
0.60
Pharmacist
0.13
-0.21
0.96
-0.66
0.20
0.35
Optometrist
0.13
-0.34
0.60
-0.78
0.26
0.28
-0.60
0.37
0.69
0.66
-0.40
-0.18
0.61
-0.10
0.34
-0.60
0.05
-0.02
Dental Hygienist
-0.63
0.30
0.29
0.55
-0.18
-0.23
Specialized Therapist: Miscellaneous
-0.12
0.33
0.54
-0.03
0.01
0.13
Medical and Dental Technologists
(Other)
Medical and Dental Technicians;
Biological and Clinical Lab. Technicians
Research Assistant in Biology
-0.30
0.34
0.88
0.23
-0.14
-0.07
-0.35
0.03
0.25
0.32
-0.23
-0.18
-0.20
0.44
0.99
-0.27
-0.03
0.19
Graduate Nurse (RN)
Optician
Hospital Administrator, Etc.
0.08
-0.04
0.66
-0.44
0.22
0.45
Psychologist
-0.23
0.44
0.79
-0.44
0.18
0.51
Social Worker
-0.24
0.49
0.31
-0.02
0.03
0.29
98
Table 5 (cont.)
Job title
Counseling & Guidance (NonPsychologist)
Economist
Things
-0.05
Artistic
0.28
Science
0.44
People
-0.20
Business
0.16
Leadership
0.32
0.05
0.14
0.74
-0.60
0.45
0.70
Social Scientist (Misc.)
-0.19
0.22
0.59
-0.49
0.12
0.44
Lawyer
-0.05
0.19
0.63
-0.80
0.52
1.04
Public Administrator (NEC)
0.12
0.07
0.40
-0.42
0.17
0.50
City Planner (NEC)
0.19
0.14
0.83
-0.69
0.16
0.45
Law Clerk
0.09
0.14
0.63
-0.71
0.19
0.72
Teacher (NEC)
-0.16
0.36
0.28
0.04
0.15
0.25
Teaching Pre-School Children
-0.55
0.57
0.02
0.67
-0.20
-0.11
Teaching Elementary School
-0.43
0.44
0.12
0.46
0.03
0.03
Teaching High School (NEC)
-0.05
0.28
0.30
0.03
0.17
0.29
Teaching High School Math
0.07
0.02
0.70
-0.28
0.21
0.24
Teaching High School Science
0.14
-0.03
0.88
-0.54
-0.05
0.14
Teaching High School Social Studies
0.04
0.20
0.32
-0.32
0.26
0.62
Teaching High School English
-0.35
0.81
0.28
0.10
0.04
0.28
Teaching High School Foreign
Languages
Teaching High School Commercial
Education
Teaching High School Home Economics
-0.45
0.70
0.31
-0.01
-0.01
0.32
-0.16
0.00
-0.06
0.27
0.45
0.03
-0.47
0.59
0.11
0.84
0.00
0.03
0.86
-0.34
0.41
-0.62
0.07
0.11
Teaching High School Trade &
Industrial Educ.
Teaching High School Physical
Education
Teaching Art (H.S., Elem.Sch., NonSchool)
Teaching Music (High School,
Elementary School, Non-School)
Teaching the Handicapped
0.12
-0.12
0.28
-0.23
0.01
0.11
-0.11
0.76
0.21
-0.03
-0.19
0.02
-0.40
0.89
0.14
0.08
-0.21
-0.07
-0.32
0.55
0.24
0.21
-0.09
0.03
Speech Therapist
-0.35
0.85
0.47
0.27
0.00
0.32
School Administrator (Non-College)
0.13
0.13
0.43
-0.26
0.31
0.47
College or University Teacher (NEC)
0.01
0.45
0.66
-0.44
0.19
0.48
College or University Teacher: Math
0.00
0.16
0.94
-0.43
0.23
0.38
College or University Teacher: Science
0.16
0.15
1.27
-0.68
-0.05
0.44
College or University Teacher: Soc.
Science
College or University Teacher: English
-0.12
0.48
0.81
-0.54
0.37
0.97
-0.38
1.11
0.61
-0.25
0.10
0.59
0.03
0.19
0.51
-0.55
0.32
0.61
Educational Researcher
-0.14
0.43
0.58
-0.15
0.18
0.39
Reading Specialist
-0.54
0.48
0.10
0.28
-0.15
-0.09
0.14
0.38
1.01
-0.61
-0.01
0.37
College or University Administrator
Graduate Assistant (Teaching and/or
Research
99
Table 5 (cont.)
Job title
Educational Aide (Teacher's Aide, etc.)
Librarian and Related Occupations
Clergy
Things
-0.38
Artistic
0.36
Science
-0.07
People
0.71
Business
0.01
-0.44
0.76
0.24
0.24
-0.13
Leadership
-0.04
0.05
0.09
0.11
0.29
-0.70
0.09
0.49
-0.43
0.35
0.30
0.36
-0.23
-0.03
0.01
0.57
0.27
0.04
-0.04
0.33
Writer (NEC)
-0.16
0.61
0.56
-0.41
-0.03
0.34
Journalist, Reporter, Etc.
-0.22
0.49
-0.01
-0.39
0.02
0.42
0.13
0.40
0.14
-0.65
0.21
0.48
Editor
-0.25
0.73
0.29
-0.29
0.01
0.42
Commercial, Fashion or Advertising
Artist, Illustrator
Designer of Consumer Goods (Except
Clothing)
Musician (Instrumental)
-0.11
0.42
-0.05
-0.39
-0.27
-0.13
0.61
-0.44
0.47
-0.89
-0.09
-0.01
-0.16
0.83
0.05
-0.20
-0.12
0.07
0.10
0.61
0.52
-0.56
0.26
0.50
-0.68
0.66
-0.27
0.21
-0.38
-0.32
0.12
0.07
0.07
-0.68
-0.14
0.03
0.11
-0.29
0.14
-0.50
-0.05
0.13
-0.03
0.25
0.30
-0.07
-0.19
0.05
Forestry, Hunting, Trapping, Fishing,
Logging
Farm or Ranch Owner
0.84
-0.51
-0.01
-0.83
-0.17
-0.09
1.03
-0.78
-0.21
-0.76
-0.14
-0.05
Farm or Ranch Manager
0.94
-0.68
-0.11
-0.79
-0.11
-0.07
Farming: Other & Misc. (Gardener,
Nursery Person)
U.S. Armed Forces
0.80
-0.57
-0.14
-0.64
-0.23
-0.06
0.42
-0.17
0.58
-0.80
0.08
0.46
Intelligence Operations (CIA, etc.)
0.32
-0.01
0.39
-0.77
0.05
0.51
Police (Public) NEC
0.64
-0.42
0.12
-0.79
-0.01
0.27
FBI and Secret Service
0.38
-0.38
0.41
-0.90
0.11
0.60
Firefighter
0.84
-0.57
-0.07
-0.81
-0.04
0.05
Other Protective
0.48
-0.12
0.16
-0.52
0.08
0.20
Banking and Finance
0.26
-0.22
0.30
-0.66
0.43
0.45
Investment Consultant, Stock
Consultant
Certified Public Accountant
0.01
-0.05
0.74
-0.73
0.47
0.70
0.29
-0.32
0.49
-0.67
0.65
0.55
Accountant, Auditor, Comptroller
(Except CPA)
Purchasing and Procurement
0.25
-0.34
0.34
-0.55
0.46
0.30
0.38
-0.30
0.26
-0.58
0.31
0.28
-0.12
0.07
0.13
-0.24
0.44
0.35
0.58
-0.32
0.57
-0.80
0.27
0.35
Monk, Brother, or Nun
Religious Worker
Radio-TV Newscaster, Commentator
Theatric Arts
Performer Arts
Misc. Performing Arts; (Except
Performing)
Professional Athlete
Recreation workers
Buyer for Retail Store
Efficiency Expert, Industrial
Engineering, Product Management
(NEC)
100
Table 5 (cont.)
Job title
Advertising
Things
0.04
Artistic
0.30
Science
0.28
People
-0.43
Business
0.20
-0.13
0.55
0.28
-0.22
0.27
0.53
Personnel Administrator
0.11
-0.02
0.37
-0.44
0.27
0.37
Appraiser, Estimator
0.27
-0.09
0.25
-0.40
0.22
0.27
Loan Investigator, Credit Investigator
0.17
-0.08
0.27
-0.41
0.35
0.34
Business Manager, Business
Administrator (NEC)
Manufacturing Management
0.30
-0.17
0.41
-0.60
0.36
0.38
0.48
-0.42
0.50
-0.87
0.21
0.40
Wholesale or Retail Trade
Management; Marketing
Supervisor in a Business (E.G. Night
Manager)
Sales Clerk, Checker, Cashier in a Store
(Store)
Route Person
0.44
-0.29
0.18
-0.60
0.26
0.23
0.70
-0.43
0.22
-0.69
0.11
0.16
-0.18
0.05
-0.16
0.41
0.04
-0.14
0.90
-0.52
-0.20
-0.75
0.03
0.12
Stockbroker
0.13
-0.11
0.52
-0.66
0.53
0.61
Real Estate Sales Person
0.10
-0.08
0.21
-0.39
0.31
0.22
Insurance Sales Person
0.38
-0.16
0.35
-0.66
0.43
0.47
Auto Sales Person
0.62
-0.24
0.08
-0.83
0.39
0.52
Other Sales Person (NEC)
0.43
-0.27
0.35
-0.74
0.32
0.36
Public Relations
Sales Manager
Leadership
0.31
0.41
-0.18
0.31
-0.71
0.39
0.47
Bookkeeper
-0.53
0.15
-0.26
0.76
0.08
-0.33
Teller or Bank Clerk
-0.49
0.13
-0.30
0.68
0.03
-0.29
Cashier (Bank)
-0.19
0.16
-0.02
0.41
0.33
0.07
Misc. Computing & Account Recording
Occupations
Keypunch Operator, Flex writer
Operator, Etc.
Computer and EAM Operator,
Supervisor, Etc.
Secretary (NEC)
-0.35
0.07
-0.24
0.50
0.09
-0.20
-0.54
0.18
-0.25
0.82
-0.06
-0.35
0.22
-0.25
0.25
-0.42
0.10
0.10
-0.64
0.25
-0.36
0.82
0.04
-0.30
Medical or Dental Secretary
-0.71
0.27
-0.15
0.82
-0.17
-0.40
Legal Secretary
-0.67
0.36
-0.35
0.77
-0.01
-0.25
Stenographer, Court Reporter, Etc.
-0.59
0.15
-0.45
0.82
0.08
-0.40
Typist, Clerk-Typist
-0.59
0.12
-0.33
0.79
-0.06
-0.30
Proofreader
-0.51
0.55
0.07
0.30
-0.71
-0.51
Clerk (Misc.)
-0.37
0.06
-0.28
0.60
0.09
-0.20
Office Manager
-0.15
-0.01
-0.04
0.05
0.20
0.03
Phone Operator, PBX Operator
-0.50
0.33
-0.22
0.72
0.00
-0.25
Other "Public Contact" Occupations
-0.23
0.11
-0.18
0.25
-0.06
-0.07
Radio, Telegraph, or Teletype Operator
0.02
-0.19
0.08
0.06
-0.12
-0.07
Mail Carrier
0.75
-0.34
0.11
-0.54
0.21
0.22
101
Table 5 (cont.)
Job title
Misc. Clerical
Things
0.47
Artistic
-0.23
Science
-0.01
People
-0.32
Business
0.13
Electrician (NEC)
0.88
-0.59
0.12
-0.83
-0.10
Electronic Technician
0.72
-0.48
0.44
-0.86
-0.14
0.05
Appliance Repair Person
0.86
-0.59
0.02
-0.84
-0.21
-0.16
Phone Installer and Repair Person
0.80
-0.50
0.14
-0.78
0.04
0.03
Repair and Service Computers and
Punch-Card Equipment
Mechanic (NEC)
0.74
-0.63
0.49
-0.90
-0.28
-0.06
0.99
-0.71
-0.14
-0.83
-0.26
-0.07
Auto Mechanic
1.13
-0.79
-0.29
-0.90
-0.31
-0.23
Airplane Mechanic
1.00
-0.50
0.10
-0.84
-0.13
-0.01
Office Machine Repair
0.77
-0.57
0.19
-0.94
-0.29
-0.11
Industrial Machine Repair
1.06
-0.59
0.03
-0.82
-0.06
0.02
Repair Misc. Small Mechanical Objects
(E. G. Clock)
Machinist
0.57
-0.56
0.23
-0.92
-0.41
0.04
1.01
-0.58
0.06
-0.83
-0.09
0.01
Cabinet Maker
0.83
-0.22
0.06
-0.58
0.17
0.20
Carpenter
1.12
-0.49
-0.01
-0.75
-0.11
0.01
Metal Trades
0.96
-0.49
-0.08
-0.73
-0.09
0.00
Bricklayer, Mason, Roofer, Painter,
Plasterer
Plumber, Pipefitter
0.97
-0.48
-0.06
-0.70
-0.06
-0.04
1.04
-0.58
-0.02
-0.81
-0.01
0.03
Road-building, Earth moving
Equipment
Misc. Building and Construction
0.97
-0.64
-0.21
-0.79
-0.20
-0.08
0.89
-0.44
0.12
-0.73
-0.08
0.18
Mining, Quarrying, Well-Drilling
0.99
-0.59
-0.04
-0.75
-0.11
-0.01
Airplane Pilot
0.39
-0.34
0.59
-0.92
-0.06
0.22
Air Traffic Controller
0.57
-0.45
0.52
-0.82
-0.09
0.15
Merchant Marine Occupations
0.74
-0.42
0.31
-0.89
-0.09
0.25
Railroad Engineer, Conductor,
Firefighter, Switch Person, Etc.
Auto, Bus, or Truck Driver, Etc.
0.77
-0.37
0.04
-0.71
0.04
0.12
0.80
-0.58
-0.17
-0.62
-0.08
-0.01
Printing Trades
0.63
-0.41
-0.16
-0.63
-0.05
-0.08
Surveyor
0.68
-0.47
0.13
-0.86
-0.07
0.13
Drafts Person
0.75
-0.29
0.23
-0.73
-0.01
0.02
Photographer
Leadership
0.11
-0.02
0.26
0.11
0.23
-0.49
-0.07
0.06
Interior Decoration
-0.34
0.52
-0.18
0.04
0.04
0.16
Clothing and Fashion Trades
-0.35
0.27
-0.21
0.78
0.06
-0.21
Dietitian
-0.59
0.21
0.25
0.74
-0.27
-0.25
Airline Steward, Stewardess
-0.64
0.39
-0.12
0.50
-0.13
-0.23
0.03
-0.02
-0.10
0.21
-0.03
-0.12
0.93
-0.49
-0.10
-0.65
0.06
-0.11
Food & Beverage Preparation (Cook,
Baker Bartender, Etc.)
Butcher, Meat Cutter
102
Table 5 (cont.)
Job title
Misc. Services (Personal and Other)
Things
-0.18
Artistic
0.03
Science
-0.24
People
0.40
Business
-0.10
0.65
-0.38
-0.17
-0.56
0.18
0.01
Hairdresser, Manicurist, Cosmetologist
-0.56
0.13
-0.40
0.66
-0.21
-0.38
Practical Nurse (PN)
-0.55
0.10
0.13
0.66
-0.40
-0.35
Nurses' Aide, Medical Aide, Psychiatric
Aide, Etc.
Domestic Service
-0.28
0.09
-0.04
0.44
-0.16
-0.15
-0.35
0.33
-0.22
0.93
-0.13
-0.31
0.62
-0.32
-0.07
-0.33
0.00
0.00
Barber
General Labor, Unspecialized
(Unskilled and Semi-Skilled)
103
Leadership
-0.21
Table 6. Ratings of major characteristics on six interest areas at the 5th year follow-up.
Major
Math
Things
0.06
Artistic
0.00
Science
0.82
People
-0.35
Business
0.17
Chemistry
0.05
-0.14
1.09
-0.65
-0.11
0.21
Physics
0.25
-0.06
1.20
-0.82
-0.11
0.25
Physical science other
0.24
0.00
1.19
-0.74
-0.05
0.30
Anatomy/physiology
-0.30
0.02
0.58
0.03
0.05
0.11
0.03
0.27
1.36
-0.50
0.15
0.66
Zoology
-0.07
0.02
1.06
-0.61
-0.09
0.25
Botany
-0.01
0.07
0.86
-0.63
-0.21
0.09
Biological sciences other
-0.02
0.17
1.03
-0.38
-0.11
0.23
Psychology
-0.25
0.38
0.58
-0.27
0.02
0.23
Sociology
-0.34
0.29
0.25
0.15
-0.06
0.11
History
-0.12
0.37
0.35
-0.36
0.17
0.59
0.11
-0.04
0.64
-0.71
0.47
0.67
Political science or government
or international relations
-0.15
0.22
0.49
-0.63
0.23
0.86
Social sciences other
-0.17
0.44
0.35
-0.13
0.16
0.46
Social work
-0.41
0.43
0.14
0.32
0.01
0.10
English
-0.41
0.84
0.25
0.06
-0.05
0.25
Journalism
-0.11
0.49
0.19
-0.27
0.10
0.46
Foreign languages
-0.42
0.72
0.37
0.10
-0.10
0.15
Fine arts
-0.29
0.81
0.01
-0.03
-0.21
-0.04
Performing arts
-0.44
0.64
0.01
-0.01
-0.14
0.06
Music
-0.34
0.93
0.14
0.02
-0.23
-0.04
0.01
0.33
0.54
-0.63
0.01
0.47
Religion or theology
-0.02
0.33
0.15
-0.13
0.01
0.10
Humanities other
-0.17
0.39
0.31
-0.14
-0.11
0.13
Law/Pre-law
-0.35
0.59
0.15
0.48
0.14
0.14
0.04
0.05
0.76
-0.43
0.12
0.32
Biochemistry
Economics
Philosophy
Medicine/Pre-med
Dentistry/Pre-dentistry
Leadership
0.22
0.07
-0.11
0.73
-0.45
0.26
0.29
Pharmacy
-0.02
-0.26
0.74
-0.60
-0.05
0.20
Nursing
-0.61
0.43
0.69
0.64
-0.43
-0.20
Other health professions
-0.31
0.17
0.68
0.02
-0.11
0.05
Architecture
0.43
-0.04
0.48
-0.83
0.04
0.22
Engineering
0.57
-0.40
0.89
-0.91
0.04
0.24
Computer sciences
0.12
-0.30
0.38
-0.55
-0.17
0.19
Statistics
-0.14
-0.26
0.37
-0.49
0.10
0.16
Elementary education
-0.60
0.44
-0.03
0.64
-0.10
-0.14
0.10
-0.09
0.19
-0.20
-0.01
0.11
Education other
-0.14
0.33
0.17
0.18
0.07
0.08
Library science
-0.51
0.49
0.02
0.42
-0.06
0.13
Physical education
104
Table 6 (cont.)
Major
Accounting
Business and commerce
Home economics
Agriculture or forestry
Things
0.22
Artistic
-0.32
Science
0.36
People
-0.55
Business
0.58
0.10
-0.17
0.18
-0.36
0.40
0.28
-0.57
0.43
-0.02
0.79
-0.11
-0.21
0.76
-0.62
0.17
-0.83
-0.16
0.01
105
Leadership
0.39
Table 7. Ratings of major characteristics on six interest areas at the 11th year follow-up.
Major
Math
Things
0.12
Artistic
-0.03
Science
0.82
People
-0.38
Business
0.17
Chemistry
0.10
-0.02
1.23
-0.60
-0.07
0.26
Physics
0.25
0.02
1.25
-0.88
-0.10
0.30
Physical science other
0.35
-0.04
0.82
-0.62
-0.01
0.23
Anatomy/physiology
-0.26
0.13
0.78
-0.10
-0.22
0.03
Biochemistry
0.19
0.49
1.43
-0.29
-0.03
0.62
Zoology
0.08
0.10
1.12
-0.59
-0.07
0.32
Botany
0.14
0.28
0.78
-0.45
-0.21
0.06
Biological sciences other
0.01
0.11
0.94
-0.40
-0.08
0.19
Psychology
-0.22
0.37
0.54
-0.23
-0.01
0.21
Sociology
-0.23
0.41
0.26
0.05
0.05
0.25
History
-0.10
0.32
0.33
-0.39
0.17
0.61
0.14
-0.07
0.60
-0.71
0.46
0.63
Political science or government
or international relations
-0.07
0.30
0.48
-0.57
0.34
0.90
Social sciences other
-0.02
0.29
0.32
-0.11
0.16
0.40
Social work
-0.36
0.33
0.13
0.36
-0.09
-0.01
English
-0.39
0.85
0.26
0.10
-0.05
0.20
Journalism
-0.14
0.60
0.09
-0.28
0.16
0.40
Foreign languages
-0.42
0.75
0.25
0.12
-0.12
0.11
Fine arts
-0.26
0.71
0.00
-0.03
-0.31
-0.15
Performing arts
-0.42
0.81
-0.05
-0.16
-0.22
0.08
Music
-0.35
0.92
0.15
0.05
-0.23
0.00
0.05
0.36
0.54
-0.59
0.04
0.41
Religion or theology
-0.01
0.39
0.15
-0.12
-0.01
0.05
Humanities other
-0.22
0.56
0.27
-0.07
0.09
0.21
0.14
0.08
0.21
-0.62
0.30
0.68
-0.06
0.18
1.13
-0.57
0.03
0.42
Dentistry/Pre-dentistry
0.17
-0.14
0.77
-0.53
0.35
0.29
Pharmacy
0.09
-0.14
0.94
-0.57
0.17
0.32
Nursing
-0.61
0.38
0.65
0.63
-0.41
-0.19
Other health professions
-0.40
0.26
0.58
0.34
-0.21
-0.16
Architecture
0.44
-0.05
0.49
-0.81
0.06
0.18
Engineering
0.59
-0.38
0.89
-0.90
0.04
0.22
Computer sciences
0.34
-0.31
0.38
-0.60
0.04
0.15
Statistics
-0.01
-0.28
0.75
-0.64
0.18
0.26
Elementary education
-0.55
0.47
0.00
0.66
-0.03
-0.10
Economics
Philosophy
Law/Pre-law
Medicine/Pre-med
Physical education
Leadership
0.22
0.12
-0.07
0.21
-0.19
0.01
0.13
Education other
-0.17
0.29
0.19
0.16
0.04
0.09
Library science
-0.45
0.58
-0.07
0.39
0.16
0.06
106
Table 7 (cont.)
Major
Things
Artistic
Science
People
Business
Accounting
0.26
-0.30
0.34
-0.52
0.54
0.36
Business and commerce
0.16
-0.20
0.21
-0.39
0.38
0.29
-0.54
0.47
0.07
0.86
-0.06
-0.18
0.76
-0.64
0.13
-0.85
-0.16
0.00
Home economics
Agriculture or forestry
107
Leadership
Table 8. Results from regression analysis of work and academic performance outcomes on
interests.
Criterion
Income
Occupational
prestige
Degree attainment
First year grade in
college
Overall grade in
college
Major grade in
college
Predictor
Things interest
β
0.02
RIW
9.6
Artistic interest
-0.11
9.8
Science interest
0.13
8.6
People interest
-0.41
60.0
Business interest
0.16
5.6
Leadership interest
0.05
6.4
Things interest
-0.36
19.2
Artistic interest
0.10
9.9
Science interest
0.30
38.9
People interest
-0.29
18.6
Business interest
0.10
4.1
Leadership interest
0.07
9.2
Things interest
-0.26
7.3
Artistic interest
0.16
11.0
Science interest
0.30
35.8
People interest
-0.38
31.6
Business interest
0.05
3.2
Leadership interest
0.09
11.1
Things interest
-0.11
21.5
Artistic interest
0.15
44.5
Science interest
0.09
23.5
People interest
-0.03
5.0
Business interest
-0.04
1.6
Leadership interest
0.01
3.9
Things interest
-0.15
32.1
Artistic interest
0.16
33.7
Science interest
0.09
15.6
People interest
0.04
13.5
Business interest
-0.08
2.5
Leadership interest
0.03
2.5
Things interest
-0.13
34.0
Artistic interest
0.16
39.9
Science interest
0.04
8.7
People interest
0.02
11.7
Business interest
-0.07
2.8
Leadership interest
0.03
2.9
108
2
R
0.49
R
0.24
Adjusted R
0.24
0.46
0.21
0.21
0.50
0.25
0.25
0.20
0.04
0.04
0.26
0.07
0.07
0.22
0.05
0.05
2
Table 8 (cont.)
Criterion
Persistence in
college
Predictor
Things interest
EXP(B)
0.72
RIW
10.9
Artistic interest
1.24
12.3
Science interest
1.39
36.1
People interest
0.68
22.4
Business interest
1.11
4.4
Leadership interest
1.14
13.9
Note. RIW = Relative Importance Weight.
109
Nagelkerke R
0.09
2
Table 9. Incremental validity of interests over ability and personality for work and
academic performance outcomes.
Criterion
Predictor
Income
Occupational prestige
Degree attainment
First year grade in
college
Overall grade in college
Major grade in college
Persistence in college
2
Cognitive abilities
.27
.07
.07
.07
12.0
Personality
.30
.09
.09
.02
4.7
Interests
.50
.25
.25
.16
83.3
Cognitive abilities
.49
.24
.24
.24
58.9
Personality
.51
.26
.26
.02
8.2
Interests
.57
.33
.33
.07
32.9
Cognitive abilities
.54
.29
.29
.29
57.2
Personality
.56
.31
.31
.03
7.4
Interests
.63
.40
.40
.08
35.4
Cognitive abilities
.35
.13
.12
.13
70.4
Personality
.38
.14
.14
.02
14.1
Interests
.39
.16
.16
.01
15.6
Cognitive abilities
.34
.12
.11
.12
50.4
Personality
.38
.15
.15
.03
23.1
Interests
.42
.18
.18
.03
26.6
Cognitive abilities
.26
.07
.07
.07
44.9
Personality
.31
.09
.09
.02
26.3
Interests
.35
.12
.12
.03
Cognitive abilities
2
ΔR
2
R
Nagelkerke R
.18
Adjusted R
2
R
ΔNagelkerke R
.18
Personality
.19
.02
Interests
.22
.03
Note. RIW = Relative Importance Weight.
110
RIW
28.7
2
Table 10. Incremental validity of interests over ability and personality for income within
occupational groups.
2
2
ΔR
(Interests)
0.03
0.08
0.02
0.17
Business Administration
0.05
0.02
0.10
4
General Teaching and Social Service
0.01
0.01
0.04
5
Humanities, Law, Social and Behavioral Science
0.02
0.04
0.11
6
Fine Arts, Performing Arts
0.08
0.04
0.14
7
Technical Jobs
0.04
0.03
0.14
8
Proprietors, Sales
0.05
0.03
0.19
9
Mechanics, Industrial Trades
0.03
0.03
0.19
10
Construction Trades
0.02
0.01
0.02
11
Secretarial-Clerical, Office workers
0.01
0.01
0.07
12
General Labor, Community and Public Service
0.05
0.04
0.24
0.04
0.03
0.12
Occupational Group
1
Engineering, Physical Science, Mathematics, and Architecture
2
Medical and Biological Science
3
Average within-occupational group ΔR
2
111
ΔR
(Ability)
0.00
2
ΔR
(Personality)
0.02
No.
Table 11. Incremental validity of interests over ability and personality for income after
controlling for occupational prestige.
Step
Predictor
R
R
2
Adjusted R
2
ΔR
2
1
Occupational prestige
.31
.09
.09
.09
2
Abilities
.38
.14
.14
.05
3
Personality
.41
.17
.17
.02
4
Interests
.55
.30
.30
.13
112
Table 12. Incremental validity of interests over demographics, ability, and personality for
work and academic performance outcomes.
Criterion
Income
Occupational
prestige
Predictor
Gender
R
R
2
Adjusted R
2
ΔR
2
Race/Ethnicity
0.3
SES
2.9
Cohort
4.9
Demographic variables (total)
.53
.28
.28
.28
Cognitive abilities
.54
.29
.29
.01
Personality
.54
.29
.29
.01
3.0
Interests
.56
.31
.31
.02
45.5
3.4
Race/Ethnicity
0.7
14.5
Cohort
2.0
Demographic variables (total)
.37
.14
.14
.14
Cognitive abilities
.54
.29
.29
.15
47.8
Personality
.55
.31
.31
.01
7.3
Interests
.60
.36
.36
.06
24.3
Gender
5.8
Race/Ethnicity
0.7
SES
16.5
Cohort
First year grade in
college
8.1
Gender
SES
Degree attainment
RIW
35.3
2.4
Demographic variables (total)
.47
.22
.22
.22
Cognitive abilities
.63
.40
.40
.18
44.6
Personality
.64
.41
.41
.02
6.3
Interests
.67
.45
.45
.04
23.7
Gender
2.9
Race/Ethnicity
0.7
SES
0.9
Cohort
14.5
Demographic variables (total)
.26
.07
.07
.07
Cognitive abilities
.40
.16
.16
.09
55.2
Personality
.42
.18
.18
.02
12.9
Interests
.43
.19
.19
.01
13.0
113
Table 12 (cont.)
Criterion
Overall grade in
college
Major grade in
college
Predictor
Gender
R
R
2
Adjusted R
2
ΔR
2
RIW
7.2
Race/Ethnicity
0.4
SES
1.1
Cohort
4.0
Demographic variables (total)
.21
.04
.04
.04
Cognitive abilities
.39
.15
.15
.11
47.0
Personality
.43
.19
.19
.04
21.3
Interests
.44
.19
.19
.01
18.9
Gender
7.5
Race/Ethnicity
0.6
SES
2.4
Cohort
Persistence in
college
5.1
Demographic variables (total)
.19
.04
.04
.04
Cognitive abilities
.31
.10
.09
.06
41.5
Personality
.35
.12
.12
.03
22.1
Interests
.36
.13
.13
.01
Nagelkerke R
.05
.15
.17
.18
Demographic variables (total)
Cognitive abilities
Personality
Interests
Note. RIW = Relative Importance Weight.
114
2
ΔNagelkerke R
.05
.10
.01
.01
20.5
2
Table 13. Comparison of predictive validity for polynomial regression equations and
congruence indices.
Criterion
Income
Overall grade in
college
Major grade in
college
Persistence in
college
2
Predictor
Polynomial Regression Equation
R
.60
R
.36
Sum of squared differences (fully constrained)
.07
.00
.00
Sum of squared differences (partly constrained)
.12
.01
.01
Profile Correlation
.13
.02
.02
Polynomial Regression Equation
.30
.09
.09
Sum of squared differences (fully constrained)
.00
.00
.00
Sum of squared differences (partly constrained)
.05
.00
.00
Profile Correlation
.05
.00
.00
Polynomial Regression Equation
.26
.07
.07
Sum of squared differences (fully constrained)
.01
.00
.00
Sum of squared differences (partly constrained)
.04
.00
.00
Profile Correlation
.03
.00
.00
Nagelkerke R
.12
Polynomial Regression Equation
Sum of squared differences (fully constrained)
.01
Sum of squared differences (partly constrained)
.01
Profile Correlation
.02
115
Adjusted R
.36
2
2
Table 14. Comparison of predictive validity for polynomial regression equations and
congruence indices with linear effects of interests and major/occupational characteristics.
Criterion
Income
Overall grade in
college
Major grade in
college
Persistence in
college
2
Predictor
Polynomial Regression Equation
2
2
ΔR by Xi , Yi , XiYi
.11
2
ΔR by Xi, Yi
.25
Total R
.36
Sum of squared differences (fully
constrained) with Xi and Yi
Sum of squared differences (partly
constrained) with Xi and Yi
Profile Correlation with Xi and Yi
.00
.32
.32
.01
.31
.32
.02
.31
.33
Polynomial Regression Equation
.04
.05
.09
Sum of squared differences (fully
constrained) with Xi and Yi
Sum of squared differences (partly
constrained) with Xi and Yi
Profile Correlation with Xi and Yi
.00
.08
.08
.00
.07
.08
.00
.07
.07
Polynomial Regression Equation
.04
.03
.07
Sum of squared differences (fully
constrained) with Xi and Yi
Sum of squared differences (partly
constrained) with Xi and Yi
Profile Correlation with Xi and Yi
.00
.06
.06
.00
.06
.06
.00
.06
.06
Nagelkerke R
.12
Polynomial Regression Equation
.09
.03
Sum of squared differences (fully
constrained) with Xi and Yi
Sum of squared differences (partly
constrained) with Xi and Yi
Profile Correlation with Xi and Yi
.01
.09
.10
.01
.09
.10
.02
.08
.10
116
2
2
Table 15. Unstandardized regression coefficients from polynomial regression equations
with six interest areas analyzed simultaneously.
Criterion
Income
Overall grade in college
X
0.01
Y
-13.80
X
-0.06
2
XY*
0.75
Y
12.22
Artistic
-0.65
-4.62
-0.20
0.34
0.00
Science
0.48
6.77
-0.01
0.18
-3.79
People
-2.26
-9.14
0.42
-0.81
-2.49
Business
0.85
10.34
-0.02
0.13
-7.76
Leadership
0.29
1.87
0.04
0.34
-5.38
Things
-0.34
-0.39
0.13
-0.40
0.50
Artistic
0.47
-0.08
0.23
-0.53
0.65
Science
-0.18
1.28
0.47
-0.80
Predictor
Things
People
0.21
-0.43
-0.18
0.51
-0.27
-0.15
0.54
-0.04
0.20
-0.56
0.11
-0.46
-0.05
-0.06
0.33
Things
-0.29
0.14
0.09
-0.16
-0.48
Artistic
0.22
0.60
0.11
-0.05
0.17
Science
0.02
0.19
0.21
-0.38
People
0.06
-0.61
-0.06
0.13
-0.13
-0.07
0.09
-0.03
0.16
-0.49
0.07
-1.84
-0.02
0.04
1.71
Things
-0.02 (0.82)
0.73 (2.07)
-0.05 (0.95)
-0.01 (0.99)
0.31 (1.37)
Artistic
-0.01 (0.99)
0.87 (2.38)
-0.01 (0.99)
0.20 (1.23)
-0.20 (0.82)
Science
0.21 (1.23)
-1.41 (0.24)
-0.05 (0.96)
0.36 (1.43)
0.95 (2.56)
People
-0.34 (0.71)
0.87 (2.37)
-0.07 (0.93)
0.12 (1.13)
0.12 (1.13)
Business
0.13 (1.14)
-0.02 (0.98)
-0.01 (0.99)
0.21 (1.24)
-2.70 (0.07)
Leadership
0.08 (1.09)
2.35 (10.51)
-0.05 (0.95)
0.15 (1.16)
-0.33 (0.72)
Business
Leadership
Major grade in college
Business
Leadership
Persistence in college**
2
Note. *Underlined coefficients for XY were nonsignificant. A significance level was set to .002 to keep the Type I error rate to
the nominal probability level (.05) as a total of 24 tests of significance were conducted.
** For college persistence, numbers in the parentheses are exp(B).
117
Table 16. Unstandardized regression coefficients from polynomial regression equations
with six interest areas analyzed separately.
Criterion
Income
Overall grade in college
Major grade in college
Persistence in college**
X
0.92
Y
7.90
X
-0.54
2
XY*
0.42
Y
-7.48
Artistic
-0.60
-5.72
-0.29
0.31
2.55
Science
0.48
12.95
-0.09
0.35
-7.48
People
-1.92
-5.82
0.01
-0.36
0.47
Business
0.27
10.19
-0.13
0.90
7.37
Leadership
0.64
18.01
-0.12
0.40
-13.00
Things
-0.24
-0.30
0.05
-0.10
0.08
Artistic
0.29
0.35
0.10
-0.06
-0.18
Science
0.08
0.15
0.19
-0.20
People
0.24
-0.03
-0.08
0.11
-0.21
Business
0.04
-0.82
0.01
-0.02
-0.22
Leadership
0.06
-0.58
0.03
-0.01
0.74
Things
-0.19
-0.57
0.04
-0.03
-0.05
Artistic
0.19
0.57
0.07
0.07
0.01
Science
0.12
-0.69
0.10
0.17
People
0.14
0.20
-0.05
0.07
-0.46
Business
0.03
-1.04
0.01
0.01
-0.08
Leadership
0.07
-1.24
0.04
0.00
1.44
Things
0.06 (1.06)
-0.35 (0.71)
-0.17 (0.84)
0.07 (1.07)
-0.69 (0.50)
Artistic
0.05 (1.05)
0.65 (1.91)
-0.05 (0.96)
0.20 (1.22)
-0.81 (0.44)
Science
0.26 (1.30)
-0.03 (0.97)
-0.03 (0.97)
0.20 (1.22)
-0.17 (0.85)
People
-0.24 (0.79)
0.16 (1.18)
-0.14 (0.87)
0.31 (1.36)
-0.58 (0.56)
Business
0.11 (1.12)
0.58 (1.79)
-0.03 (0.97)
0.26 (1.30)
-3.88 (0.02)
Leadership
0.21 (1.23)
0.01 (1.01)
-0.07 (0.94)
0.29 (1.34)
1.40 (4.04)
Predictor
Things
Note. *Underlined coefficients for XY were nonsignificant. A significance level was set to .002 to keep the Type I error rate to
the nominal probability level (.05) as a total of 24 tests of significance were conducted.
** For college persistence, numbers in the parentheses are exp(B).
118
2
DA
-0.21
0.06
0.03
0.02
0.01
0.19
-0.13
0.23
0.25
-0.07
-0.04
-0.05
-0.04
-0.02
0.01
0.02
0.08
-0.02
0.04
-0.17
0.09
0.12
-0.10
0.00
0.09
OP
-0.08
-0.02
0.00
0.00
-0.01
0.02
-0.03
0.06
0.08
-0.01
0.01
0.00
-0.01
-0.01
0.01
-0.02
0.02
-0.01
0.02
-0.19
0.00
0.08
0.03
0.05
0.01
0.59
0.08
-0.20
0.00
-0.01
-0.03
0.37
0.25
VA
-0.12
-0.14
0.02
-0.01
-0.03
0.37
0.19
MA
-0.25
-0.18
0.02
0.01
-0.02
0.28
0.15
SA
0.16
-0.04
-0.01
-0.01
0.00
0.14
EX
0.25
0.00
0.00
0.00
0.00
0.18
AG
0.02
0.00
-0.01
0.00
-0.01
0.09
IM
0.00
-0.02
-0.03
-0.01
-0.01
0.18
VG
0.08
0.00
-0.01
-0.01
-0.01
0.18
CA
0.27
0.03
-0.01
-0.01
0.00
0.20
CU
0.03
0.10
0.00
0.00
0.01
0.17
LP
0.02
0.01
-0.01
0.00
0.00
0.16
ES
0.00
0.01
-0.02
-0.01
-0.01
0.18
CS
-0.63
0.02
0.01
0.01
0.00
-0.13
THI
0.28
0.10
0.01
0.01
0.01
0.12
ART
-0.23
0.10
0.03
0.02
0.01
0.16
SCI
0.74
0.05
0.00
0.01
0.01
-0.08
PPL
-0.13
0.08
0.00
0.01
0.02
0.02
BUS
-0.27
0.09
0.01
0.00
0.01
0.09
LEA
119
Note. Dependent variables of the regressions are listed in the columns and independent variables are listed in the row. OP = Occupational prestige, DA = Degree attainment, VA = Verbal ability, MA =
Math ability, SA = Spatial ability, EX = Extroversion, AG = Agreeableness, IM = Impulsiveness, VG = Vigor, CA = Calmness, CU = Culture, LP = Leadership personality, ES = Emotional stability, CS =
Conscientiousness, THI = Things interest, ART = Artistic interest, SCI = Science interest, PPL = People interest, BUS = Business interest, LEA = Leadership interest.
IV (row)/
Income
DV(column)
Sex
-0.38
Black
-0.01
Asian
0.01
Native
0.00
Latino(a)
-0.01
SES
0.05
Cohort
-0.11
Verbal ability
-0.01
Math ability
0.05
Spatial ability
0.02
Extroversion
0.03
Agreeableness
0.00
Impulsiveness
-0.01
Vigor
0.00
Calmness
0.02
Culture
-0.02
Leadership
0.03
Emotional Stability
0.00
Conscientiousness
0.01
Things interest
-0.08
Artistic interest
-0.08
Science interest
0.05
People interest
-0.11
Business interest
0.10
Leadership interest
0.02
Degree attainment
0.08
Table 17. Standardized regression coefficients from the mediation path model of career success.
Table 18. Direct effect of gender on income and occupational prestige, indirect effect of
gender on income and occupational prestige through individual difference variables,
correlation between gender and income/occupational prestige, and partial correlations
between gender and income/occupational prestige after accounted for each individual
difference variable.
Gender
Income
Correlation
-0.38
-0.50
Partial
Correlation
Occupational
Prestige
Correlation
-0.08
-0.10
Partial
Correlation
Verbal ability
0.00
-0.50
0.02
-0.15
Math ability
-0.01
-0.49
-0.03
-0.05
Spatial ability
0.00
-0.47
0.01
-0.03
Extroversion
0.01
-0.50
0.00
-0.12
Agreeableness
0.00
-0.50
-0.01
-0.14
Impulsiveness
0.00
-0.50
0.00
-0.10
Vigor
0.00
-0.50
0.00
-0.10
Calmness
0.00
-0.50
0.00
-0.12
Culture
Leadership
personality
Emotional Stability
0.00
-0.49
0.00
-0.15
0.00
-0.50
0.00
-0.11
0.00
-0.50
0.00
-0.11
Conscientiousness
0.00
-0.50
0.00
-0.11
Things interest
0.06
-0.46
0.18
-0.23
Artistic interest
-0.02
-0.48
0.02
-0.15
Science interest
-0.01
-0.48
-0.03
-0.04
People interest
-0.09
-0.30
-0.02
-0.01
Business interest
-0.01
-0.49
-0.01
-0.09
Leadership interest
-0.01
-0.47
-0.02
-0.06
Total
-0.48
0.03
Note. Underlined regression coefficients were nonsignificant at p < .001.
120
Table 19. Direct effects of individual difference variables on income and occupational
prestige, indirect effects of individual difference variables on income and occupational
prestige through degree attainment, correlation between each individual difference
variable and income/occupational prestige, and partial correlations between each
individual difference variables and income/occupational prestige after accounted for
degree attainment.
Verbal ability
Income
Correlation
-0.01
0.07
Degree attainment
0.02
Total
0.01
Math ability
0.05
Degree attainment
0.02
Total
0.07
Spatial ability
0.02
Degree attainment
-0.01
Total
0.02
Extroversion
0.03
Degree attainment
0.00
Total
0.03
Agreeableness
0.00
Degree attainment
0.00
Total
0.00
Impulsiveness
Degree attainment
-0.01
-0.01
Vigor
0.00
Degree attainment
0.00
Total
0.02
Degree attainment
0.00
Total
0.02
Degree attainment
Total
-0.02
0.00
Occupational
Prestige
Correlation
0.06
0.43
0.13
Partial
Correlation
0.16
0.19
0.19
0.08
0.06
0.47
0.15
0.16
0.23
0.21
-0.01
0.13
0.29
-0.04
0.07
-0.05
-0.02
0.01
-0.02
0.07
-0.02
0.08
-0.01
-0.07
0.00
-0.10
0.13
-0.03
0.09
-0.03
0.00
-0.01
-0.01
0.02
-0.02
0.01
-0.03
0.06
-0.01
0.03
-0.01
Calmness
Culture
-0.06
0.00
Total
Partial
Correlation
0.12
-0.01
0.05
-0.02
0.03
0.01
-0.01
0.17
0.01
0.08
0.02
-0.09
-0.02
-0.13
-0.02
0.01
-0.01
121
0.15
0.08
Table 19 (cont.)
Leadership personality
Income
Correlation
0.03
0.06
Degree attainment
0.01
Total
0.03
Emotional Stability
0.00
Degree attainment
0.00
Total
0.01
Degree attainment
0.00
Total
0.01
Things interest
-0.08
Degree attainment
-0.01
Total
-0.09
Artistic interest
-0.08
Total
Occupational
Prestige
Correlation
0.02
0.17
0.01
0.04
-0.01
0.01
0.07
Degree attainment
0.01
Total
0.06
People interest
-0.11
Degree attainment
-0.01
Total
-0.12
Business interest
0.10
Degree attainment
0.00
Total
0.10
Leadership interest
0.02
Degree attainment
0.01
Total
0.03
0.14
-0.01
0.02
0.02
0.06
0.21
0.02
0.09
0.04
0.24
-0.19
0.25
-0.12
-0.10
-0.16
-0.29
-0.16
0.00
-0.21
-0.08
0.05
0.06
-0.02
0.01
Science interest
0.05
Partial
Correlation
0.07
-0.01
Conscientiousness
Degree attainment
Partial
Correlation
0.15
0.05
0.06
0.05
0.18
0.08
0.10
0.30
0.07
0.09
0.15
-0.42
0.03
-0.39
-0.13
-0.06
0.05
-0.03
0.09
0.05
0.07
0.09
0.00
0.04
0.05
0.18
0.01
0.12
0.05
0.07
Note. Underlined regression coefficients were nonsignificant at p < .001.
122
0.19
0.03
Personality
Abilities
Interests
123
Degree
Attainment
College
Persistence
College GPA
Figure1. A framework for interests, cognitive abilities, personality and educational achievement.
FIGURES
Personality
Abilities
Interests
Degree Attainment
Number of
Children
Occupational
Prestige
Income
124
Note. Solid lines represent direct effects and doted lines represent indirect effects. Demographic variables other than gender were omitted for
simplicity.
Gender
Figure 2. A framework for interests, cognitive abilities, personality and career success.
Figure 3. Project TALENT interest inventory factor hierarchy.
Note. a. A broad “Things” or “Masculine” factor, including Mechanics, Nature-Agriculture, Military, and Sports items. Physical Sciences, some
Leadership and some Business items also load on this factor (relatively small loadings).
b. A broad “People” or “Feminine” factor, including Artistic, Education, Social Service, Personal Service, Social Sciences, Clerical, and
some Leadership and some Business items.
c. A “Things” or “Masculine” factor, Including Mechanics, Nature-Agriculture, Military, and Sports items.
d. This seems to be a factor with occupations and activities that have higher prestige and that require higher education level, Including
Sciences, Leadership, and Fine Arts items.
e. A “People” or “Feminine” factor, including Clerical, Personal Service, Social Service, Performing Arts, and some Business items.
f. This factor includes Artistic (Fine and Performing Arts), Social Service, Personal Service, and Social Sciences items.
g. The “Sciences” factor includes Physical Sciences, Engineering, Medical Science, and Mathematics items.
h. This factor combines Clerical items from the “People” factor (e) and Business Details items (part of Business).
i. At this level, the Sports items break away from the Things factor.
j. The “Business” factor includes Clerical, Business Details and Business Contact items.
k. A purified Artistic factor.
l. A “People” factor reemerge, combining (traditionally female-oriented) Social and Personal Service items from the “Artistic & Social”
factor above and Clerical items from the “Business” factor above.
m. A purified Business factor.
n. This is a small factor about computation that includes mathematics (from Sciences) and accountancy (from Business) items.
125
Figure 4. Relative contribution of ability, personality, and interests to each criterion.
100%
90%
80%
70%
60%
50%
40%
Interests
30%
Personality
20%
Ability
10%
0%
126
Figure 5. Relative contribution of demographic variables, ability, personality, and interests
to each criterion.
100%
90%
80%
70%
60%
Interests
50%
Personality
40%
Ability
30%
Cohort
20%
SES
10%
Race
Gender
0%
127
Figure 6. Effect of interest congruence in Things on income.
Z = 44.27 + 0.01*X – 13.80*Y – 0.06*X2 + 0.75*XY + 12.22*Y2
First principal axis: Y = -96.72 + 32.66*X
Second principal axis: Y = 0.56 – 0.03*X
128
Figure 7. Effect of interest congruence in Artistic on income.
Z = 44.27 –0.65*X – 4.62*Y – 0.20*X2 + 0.34*XY + 0.00*Y2
First principal axis: Y = -5.93 + 1.73*X
Second principal axis: Y = 24.19 – 0.58*X
129
Figure 8. Effect of interest congruence in Science on income.
Z = 44.27 + 0.48*X + 6.77*Y – 0.01*X2 + 0.18*XY – 3.79*Y2
First principal axis: Y = 0.89 + 0.02*X
Second principal axis: Y = 1945.76 – 41.54*X
130
Figure 9. Effect of interest congruence in People on income.
Z = 44.27 – 2.26*X – 9.14*Y + 0.42*X2 – 0.81*XY – 2.49*Y2
First principal axis: Y = -1.86 – 0.14*X
Second principal axis: Y = -7.87 + 7.37*X
131
Figure 10. Effect of interest congruence in Business on income.
Z = 44.27 + 0.85*X + 10.34*Y – 0.02*X2 + 0.13*XY –7.76*Y2
First principal axis: Y = 0.67 + 0.01*X
Second principal axis: Y = 3640.29 – 121.07*X
132
Figure 11. Effect of interest congruence in Leadership on income.
Z = 44.27 + 0.29*X + 1.87*Y + 0.04*X2 + 0.34*XY – 5.38*Y2
First principal axis: Y = 0.17 + 0.03*X
Second principal axis: Y = -122.31 – 31.74*X
133
Figure 12. Effect of interest congruence in Things on overall grade in college.
Z = 6.94 – 0.34*X – 0.39*Y + 0.13*X2 – 0.40*XY + 0.50*Y2
First principal axis: Y = 15.37 – 2.28*X
Second principal axis: Y = 0.20 + 0.44*X
134
Figure 13. Effect of interest congruence in Artistic on overall grade in college.
Z = 6.94 + 0.47*X – 0.08*Y + 0.23*X2 – 0.53*XY + 0.65*Y2
First principal axis: Y = -4.34 – 2.06*X
Second principal axis: Y = 0.20 + 0.49*X
135
Figure 14. Effect of interest congruence in Science on overall grade in college.
Z = 6.94 – 0.18*X + 1.28*Y + 0.47*XY – 0.80*Y2
First principal axis: Y = 0.77 + 0.27*X
Second principal axis: Y = -4.72 – 3.65*X
136
Figure 15. Effect of interest congruence in People on overall grade in college.
Z = 6.94 + 0.21*X – 0.43*Y – 0.18*X2 + 0.51*XY – 0.27*Y2
First principal axis: Y = -0.62 + 0.83*X
Second principal axis: Y = 2.64 – 1.20*X
137
Figure 16. Effect of interest congruence in Business on overall grade in college.
Z = 6.94 – 0.15*X + 0.54*Y – 0.04*X2 + 0.20*XY – 0.56*Y2
First principal axis: Y = 0.49 + 0.19*X
Second principal axis: Y = -4.83 – 5.30*X
138
Figure 17. Effect of interest congruence in Leadership on overall grade in college.
Z = 6.94 + 0.11*X – 0.46*Y – 0.05*X2 – 0.06*XY + 0.33*Y2
First principal axis: Y = 10.17 – 13.09*X
Second principal axis: Y = 0.71 + 0.08*X
139
Figure 18. Effect of interest congruence in Things on major grade in college.
Z = 8.66 – 0.29*X + 0.14*Y + 0.09*X2 – 0.16*XY – 0.48*Y2
First principal axis: Y = 0.10 – 0.14*X
Second principal axis: Y = -10.85 + 7.37*X
140
Figure 19. Effect of interest congruence in Artistic on major grade in college.
Z = 8.66 + 0.22*X + 0.60*Y + 0.11*X2 – 0.05*XY + 0.17*Y2
First principal axis: Y = -5.74 – 2.67*X
Second principal axis: Y = -1.50 + 0.37*X
141
Figure 20. Effect of interest congruence in Science on major grade in college.
Z = 8.66 + 0.02*X + 0.19*Y + 0.21*XY – 0.38*Y2
First principal axis: Y = 0.22 + 0.26*X
Second principal axis: Y = -5.14 – 3.86*X
142
Figure 21. Effect of interest congruence in People on major grade in college.
Z = 8.66 + 0.06*X – 0.61*Y – 0.06*X2 + 0.13*XY – 0.13*Y2
First principal axis: Y = -1.88 + 0.56*X
Second principal axis: Y = -11.68 – 1.78*X
143
Figure 22. Effect of interest congruence in Business on major grade in college.
Z = 8.66 – 0.07*X + 0.09*Y – 0.03*X2 + 0.16*XY – 0.49*Y2
First principal axis: Y = 0.10 + 0.17*X
Second principal axis: Y = -9.66 – 6.02*X
144
Figure 23. Effect of interest congruence in Leadership on major grade in college.
Z = 8.66 + 0.07*X – 1.84*Y – 0.02*X2 + 0.04*XY + 1.71*Y2
First principal axis: Y = -188.95 + 99.10*X
Second principal axis: Y = 0.54 – 0.01*X
145
Figure 24. Effect of interest congruence in Things on college persistence.
Z = exp(0.72 – 0.20*X + 0.73*Y – 0.05*X2 – 0.01*XY + 0.31*Y2)
First principal axis: Y = -120.23 – 66.38*X
Second principal axis: Y = -1.16 + 0.02*X
146
Figure 25. Effect of interest congruence in Artistic on college persistence.
Z = exp(0.72 – 0.01*X + 0.87*Y – 0.01*X2 + 0.20*XY – 0.20*Y2)
First principal axis: Y = 1.78 + 0.44*X
Second principal axis: Y = -12.39 – 2.30*X
147
Figure 26. Effect of interest congruence in Science on college persistence.
Z = exp(0.72 + 0.21*X – 1.41*Y – 0.05*X2 + 0.36*XY + 0.95*Y2)
First principal axis: Y = -17.13 + 5.72*X
Second principal axis: Y = 0.70 – 0.17*X
148
Figure 27. Effect of interest congruence in People on college persistence.
Z = exp(0.72 – 0.34*X + 0.87*Y – 0.07*X2 + 0.12*XY + 0.12*Y2)
First principal axis: Y = 11.13 + 3.41*X
Second principal axis: Y = -2.77 – 0.29*X
149
Figure 28. Effect of interest congruence in Business on college persistence.
Z = exp(0.72 + 0.13*X – 0.02*Y – 0.01*X2 + 0.21*XY – 2.70*Y2)
First principal axis: Y = 0.00 + 0.04*X
Second principal axis: Y = 282.27 – 25.33*X
150
Figure 29. Effect of interest congruence in Leadership on college persistence.
Z = exp(0.72 + 0.08*X + 2.35*Y – 0.05*X2 + 0.15*XY – 0.33*Y2)
First principal axis: Y = 3.33 + 0.25*X
Second principal axis: Y = 41.29 – 4.02*X
151
Figure 30. Overall and by-gender effect of number of children on income.
60
Income
Y = 50.79 + .37*X
55
50
45
40
Y = 49.13 – .75*X
35
Male
30
Female
25
20
Overall
Y = 47.49 – 3.43*X
15
10
5
0
0
1
2
3
4
5
6
7
8
9
10 11 12
# of Children
Figure 31. Overall and by-gender effect of number of children on occupational prestige.
60
Occupational Prestige
55
50
45
40
35
Male
30
Female
25
Overall
20
15
Y = 55.50 – 3.98*X
10
Y = 52.73 – 5.40*X
5
Y = 53.06 – 4.24*X
0
0
1
2
3
4
5
6
7
8
9
10 11 12
152
# of Children