• Abstract
• Prior Evidence
• Setpoint Model Of Interests
• Interests Are Hierarchal
• General Discussion
• Results
• Contextualization Of Vocational Interests

Abstract

Growing evidence on the predictive validity of vocational interests for job performance calls for greater consideration of interest assessment in organizations. However, a consensus on the fundamental dimensions of interests that are aligned with the contemporary world of work is still lacking. In the current research, we developed an organizing framework of vocational interests and empirically validated an 8-dimension model (SETPOINT: Health Science, Creative Expression, Technology, People, Organization, Influence, Nature, and Things). We propose that interests are structured hierarchically, with preferences for: 1. specific work activities at the lowest level (assessed using interest items) 2. basic interests for homogeneous classes of activities at the intermediate level (assessed using basic interest scales) 3. broad-band interest dimensions describing general tendencies of individuals to be drawn to or motivated by broad types of work environments at the top. To derive broad-band interest dimensions, it is necessary to base it on a comprehensive range of content-specific basic interest constructs. In Study 1, we conducted an extensive review of existing basic interest scales and developed a new assessment of basic interests with 41 homogeneous scales across two samples. In Study 2, we demonstrated the structural validity of the proposed dimensional model using second-order confirmatory factor analysis and exploratory structural equation modeling with a large, diverse sample of working adults and supported its predictive validity for occupational membership in new and traditional sectors of work. We discuss implications from the current findings for building interest theory, using interest assessment for organizational research, and evaluating interest structure with appropriate methods.

Prior Evidence

Meta-analyses conducted by two independent research teams (Nye, Su, Rounds, & Drasgow, 2012, 2017; Van Iddekinge, Roth, Putka, & Lanivich, 2011) have linked vocational interests and the fit between individual interests and their environments to various criteria of job performance. Interests have also been shown to have incremental validity over cognitive ability and personality traits in predicting job performance (Van Iddekinge, Putka, & Campbell, 2011) and career success (Rounds & Su, 2014; Stoll et al., 2017). This burgeoning evidence on the validity of interests for predicting performance behaviors, along with long-established findings that interests drive educational and occupational choices (Kuder, 1977; Lubinski, 2000; Strong, 1943), highlights the importance of interests for work-related outcomes and calls the field toward greater attention to and consideration of interest assessment in personnel selection and beyond (Van Iddekinge, Putka, et al., 2011).

The lack of consensus on the dimensional structure of vocational interests hinders the communication and accumulation of research findings and impedes the advancement of interest theory.

The assumption of correspondence between interest types and occupational clusters serves as the foundation for career guidance

Setpoint Model Of Interests

Eight-dimension interest model, which we titled SETPOINT:

• Health Science
• Creative Expression
• Technology
• People
• Organization
• Influence
• Nature
• Things

Interests Are Hierarchal

We have established that interests are structured hierarchically, with preferences for specific activities at the lowest level, basic interests at the intermediate level representing core mental schemata that individuals use to classify activities, and broad-band interest dimensions at the top describing overall tendencies of an individual to be drawn to or motivated by general types of environments.

General Discussion

Interest research has enjoyed a long history dating back to the dawn of the 20th century, and interest inventories have been widely used for guiding individuals’ career choices. Nonetheless, the changing nature of work and growing needs for using interest assessment in organizational research necessitates an updated understanding and a clear consensus about the fundamental dimensions of interests. The current article contributes to the literature in several ways: (1) First, we have established that interests are structured hierarchically, with preferences for specific activities at the lowest level, basic interests at the intermediate level representing core mental schemata that individuals use to classify activities, and broad-band interest dimensions at the top describing overall tendencies of an individual to be drawn to or motivated by general types of environments. We clarified the confusion in the literature about analytical methods for evaluating interest structure and demonstrated that interests are best represented using a higher order CFA model or an ESEM model. (2) Second, we have highlighted the need for building broad-band interest dimensions from a comprehensive set of content-specific, homogeneous basic interest constructs that fully reflect the world of work. The new basic interest measure, CABIN, provides a foundation for deriving broad-band interest dimensions and a great stand-alone assessment for organizational researchers and practitioners to use. (3)Third, we have demonstrated that the proposed SETPOINT model best represents the interest domain of the 21st-century labor force and is effective at predicting occupational membership, particularly in three fast-growing sectors (healthcare, STEM, and green occupations). (4) Next we discuss the meanings of and theoretical implications from the identified interest dimensions, potential applications of the new dimensional model of interests and basic interest assessment in the organizational setting, and methodological considerations in the investigation and evaluation of interest structure.

Results

A summary of descriptive statistics (means and standard deviations) for the 41 new basic interest scales and their intercorrelations are reported in Table 3.

The goodness-of-fit indices for the baseline model, the proposed eight-dimension model, and the alternative six-dimension model from CFA and ESEM are presented in Table 4. The baseline model fitted well to the data. All the items had high loadings on corresponding basic interest factors (range .75.95). Reliabilities were high for all the scales (s .90.97). These results indicate that CABIN has good structural validity and the items are excellent indicators of the basic interest constructs. Given the comprehensive range of basic interests from our crosswalk, these findings imply that the current selection of constructs represents a good first-order collection for testing second-order broad-band interest dimensions. Table 4 shows that the eight-dimension model fits the data better than does the six-dimension model, although both had adequate fit. Given the specification of the two models being nonnested, we could not test for their statistical difference. However, according to the standard of CFI change (.002), there was a practical difference between the two models. AIC and BIC also indicated that the proposed model fitted the data better and was statistically more parsimonious than was the alternative model. ESEM analyses provided stronger support for the proposed model. When secondary loadings were allowed for basic interest constructs to load on other interest dimensions beyond the specified primary dimension, the proposed model showed excellent fit to the data, whereas the alternative model showed inadequate fit. Factor loadings from second-order CFA for the two models are presented in Tables 5 and 6, and loadings from ESEM for the two models are presented in Tables 7 and 8. Overall, these findings support the structural validity of the SETPOINT model of interests. We are particularly interested in the validity of the new model for predicting occupational membership. Table 9 summarizes the results from logistic regression analyses. McFadden’s R2 showed that the SETPOINT model had excellent or close-to-excellent fit for all six prediction models. Both McFadden’s R2 and the OOR showed that the SETPOINT model was superior at predicting occupational membership across the board, particularly for the three fast-growing work sectors. Specifically, interest in Health Science was the strongest predictor of occupational membership in healthcare (b 1.67, p .001), meaning that an individual who scored 1 point higher on the Health Science dimension was 5.31 times more likely to be a healthcare professional. Lower interest in the Influence dimension was also predictive of being a healthcare professional. Occupational membership in STEM was predicted by interest in Technology (b 1.90, p .001). An individual who scored 1 point higher on the Technology dimension was 6.69 times more likely to be a STEM professional. Interest in the nature dimension strongly predicted occupational membership in green occupations (b 1.86, p .001), with a 1-point increase on the dimension translating to 6.42 times the likelihood to be in a green occupation. In comparison, occupational membership in education, manual and skilled trades, and office and administrative jobs was marked by higher interest in People (b 1.28, p .001), Things (b .80, p .001), and Organization (b .76, p .001), respectively. Comparing the logistic regression results for the two models, the eight-dimension model offers much more straightforward solutions for explaining and predicting occupational membership. As previously discussed, interests in healthcare, STEM, and green occupations are not well represented by extant interest models, because most were developed before the 1980s and do not fully reflect today’s world of work. As a result, occupational membership in healthcare and STEM fields could be understood only by using a combination of multiple interest dimensions from the six-dimension model. Occupational membership in the green sector could only be predicted negatively by Conventional interests in the six-dimension model. For the three traditional work sectors, the SETPOINT model also provides a clearer correspondence between interest dimensions and occupational membership compared to the six-dimension model.

Contextualization Of Vocational Interests

Interests are contextualized and describe individuals’ affective reactions to and cognitive appraisals of objects and activities in external environments (Rounds & Su, 2014; Su et al., in press). Contextualization sets vocational interests apart from other individual difference variables such as personality traits. This unique property of interests motivated the current research and is further implied from the findings. The SETPOINT model of interests reflects the changes in the world of work. Compared to existing interest models, such as Holland’s RIASEC types, it better captures interests in emerging industries and occupations, evolving nature of jobs and work tasks, and expanding job requirements. To interpret the eight interest dimensions, it is important to pay attention to the description of the dimensions in Table 2 as well as the basic interest scales that compose the dimensions and their factor loadings in Table 5, discussed next.