Chet Robie

Introduction: The purpose of this study was to compare the structural validity of Holland’s circular/hexagonal model in two versions of a commonly used Realistic, Investigative, Artistic, Social, Enterprising, Conventional (RIASEC) occupational interest measure – one that used items with a neutral response option ( unsure) and one that used items without a neutral response option. Method: These comparisons were made using a sample of 1,025 undergraduate university business majors. A two-group Cosine Function Model (CFM) implemented in standard structural equation modeling software was used to investigate the circumplex fit across the two versions of the assessment. Results: CFM analyses suggested a high level of equivalence across versions such that (1) the correlation matrix of each group shows a good fit to the RIASEC circumplex, and (2) the two correlation matrices are essentially identical. Conclusion: The neutral response option (or lack thereof) did not seem to affect model fit. The generalizability of these results should be explored in future studies.