Examination of adherence to Holland's vocational interest structure of temporary cash assistance recipients.

摘要

The objective of this study was to determine whether Vocational Preference Inventory (VPI), a career assessment inventory designed to assist individuals identify jobs and career possibilities (Armstrong & Anthoney, 2009), would produce a good fit of Holland's interest structure model for welfare recipients. Specifically, the effectiveness of Holland's congruence notion of six interest-based categories: Realistic, Investigative, Artistic, Social, Enterprising, and Conventional (RIASEC) as classifiers of occupations for these individuals were examined using Confirmatory Factor Analysis (CFA) and Rasch techniques. The six career interest types are represented by a hexagon where relationship between them is inversely proportional to the distance between them (Bullock, Andrews, Braud, & Reardon, 2010; Deng, Armstrong, & Rounds, 2007), closest pairs have highest correlation followed by alternate pairs and lastly opposite pairs.;Our broad hypotheses tested the relationships among the interest pairs (adjacent, alternate, and opposite), reliability, and the interrelations of jobs that load onto each of these six subscales from which 11 specific hypotheses were derived. Data was drawn from two samples 2506 and 818 individuals who took the original VPI and individuals who took the VPI along with pictures respectively. For the sample that took original VPI, only one of Holland's order predictions was supported whereby, adjacent pairs had greater correlations than opposite pairs. For the sample that took VPI along with pictures, all adjacent and alternate pairs had greater correlation than opposite pairs, consistent with Holland's order prediction and a support of our second and third hypotheses. Consequently, CFA findings revealed the model with pictures was a better fit than the model with original VPI. Having pictures, improved the model by 23% i.e. 67% of sample data with original VPI and 90% of sample data with pictures conformed to Holland's order prediction respectively. Furthermore, similar findings were supported by Rasch fit statistics where there were more fitting jobs (43) from the sample that took VPI with pictures compared to 30 fitting jobs from the sample that took the original VPI. Although the presence of pictures was found to enhance model fit, none of our samples data entirely replicated Holland's model prediction.