The Construct Validity of Three Entry Level Personality Inventories Used in the UK: Cautionary Findings from a Multiple-Inventory Investigation

This paper reports psychometric analyses into the convergent and divergent validity of three popular entry-level measures of occupational personality in the UK and Continental Europe. A sample of 504 individuals completed all three measures: the British version of the Hogan Personality Inventory (HPI), the Occupational Personality Questionnaire (OPQ Version FS5.2), and the Business Personality Indicator (BPI). In addition, independent ratings of the conceptual loading of primary source scales onto the Five Factor Model (FFM) were obtained (n = 66). Data were used in a three-stage analytical procedure directed at examining psychometric and construct validity. Results are reported for descriptive statistics (means, standard deviations, Cohen's d), internal consistency reliability (Cronbach's alphas), and exploratory factor analyses. Findings into the construct validity of first-order scales (i.e. primary source scales) and second-order scales (i.e. FFM loadings) are presented in detail, including multitrait-multimethod (MTMM) analyses of convergent and divergent validity. For some scales, the observed variability in our sample suggested significant range restriction/enhancement. It was found that scale reliabilities were generally lower than those typically reported by the test publishers, and that published factor structures for these measures could not be replicated by the authors for this sample of individuals. Further independent construct validity research into occupational personality inventories is encouraged based upon our proposed model of single-, dual-, and multiple-inventory construct validation studies. Practically, our findings suggest that when IWO psychologists or personnel professionals aim to select/ screen job applicants for a particular personality trait those who are selected may vary depending on (i) which personality inventory is used, (ii) the actual variability in the applicant sample tested, and (iii) reliability of the scales under consideration. As such, this study sounds a note of caution. Future research is called for to replicate these findings.