Over the past 10 years, job interview training has emerged as an area of study among adults with schizophrenia and other serious mental illnesses who face significant challenges when navigating job interviews. The field of mental health services research has limited access to assessments of job interview skills with rigorously evaluated psychometric properties.
We sought to evaluate the initial psychometric properties of a measure assessing job interview skills via role-play performance.
As part of a randomized controlled trial, 90 adults with schizophrenia or other serious mental illnesses completed a job interview role-play assessment with eight items (and scored using anchors) called the mock interview rating scale (MIRS). A classical test theory analysis was conducted including confirmatory factor analyses, Rasch model analysis and calibration, and differential item functioning; along with inter-rater, internal consistency, and test-retest reliabilities. Pearson correlations were used to evaluate construct, convergent, divergent, criterion, and predictive validity by correlating the MIRS with demographic, clinical, cognitive, work history measures, and employment outcomes.
Our analyses resulted in the removal of a single item (sounding honest) and yielded a unidimensional total score measurement with support for its inter-rater reliability, internal consistency, and test-retest reliability. There was initial support for the construct, convergent, criterion, and predictive validities of the MIRS, as it correlated with measures of social competence, neurocognition, valuing job interview training, and employment outcomes. Meanwhile, the lack of correlations with race, physical health, and substance abuse lent support for divergent validity.
This study presents initial evidence that the seven-item version of the MIRS has acceptable psychometric properties supporting its use to assess job interview skills reliably and validly among adults with schizophrenia and other serious mental illnesses.
NCT03049813.