Ruiz, M., Hopwood, C., Edens, J.F., Morey, L.C., & Cox, J. (2018). Initial development of pathological personality traits domain measures using the Personality Assessment Inventory (PAI). Personality Disorders: Theory, Research, and Treatment.
This study set out to create measures of the five personality disorder trait domains outlined in Section III of the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (American Psychiatric Association, 2013) from the Personality Assessment Inventory items (Morey, 2007). Rasch rating scale model analyses and classical test theory analyses were applied to existing data sets (N = 3,877; community, clinical, offender, college) to identify relevant items. Five scales were created that had acceptable unidimensionality and generally conformed to Rasch model expectations. The ability of the items to cover the underlying construct and their differential item function by sex were acceptable, though a few of the proposed scales had weaknesses in these areas. Internal consistency was acceptable for all scales and the factor structure was generally consistent with expectations, but some scales had concerning cross-loadings. Preliminary analyses demonstrated validity of the scales in relation to history of mental health treatment/current symptoms, substance abuse, and, for one scale, violent rearrests. There were small-to-moderate associations with noncorresponding traits, suggesting a degree of saturation with general personality impairment. The relevance of the proposed scales for the assessment of Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition personality disorder is discussed.
12/25/2021 04:05:34 am
thanks for sharing actually you are right, an emphasis of PDTRT moving forward is to publish studies with larger sample sizes given the problems associated with smaller samples including lack of statistical power, lower precision and poorer stability of effect sizes (e.g., Schönbrodt & Perugini, 2013), and the increased odds that statistically significant effect sizes from small samples may be substantial overestimates. Studies submitted with very small samples are likely to be rejected without review (even when reporting patient data).
Leave a Reply.