An Item-Level Psychometric Analysis of the Personality Assessment Inventory Clinical Scales in a Psychiatric Inpatient Unit

• Published in: Assessment (Odessa, Fla.) Vol. 16; no. 4; pp. 373 - 383
• Main Authors: Siefert, Caleb J., Sinclair, Samuel J., Kehl-Fie, Kendra A., Blais, Mark A.
• Format: Journal Article
• Language: English
• Published: Los Angeles, CA SAGE Publications 01-12-2009
• Subjects: Adult; Ceiling Effects; Correlation; Female; Hospital Units; Humans; Inpatients; Internal Consistency; Item Analysis; Male; Measurement Techniques; Patients; Personality Assessment; Personality Assessment Inventory; Personality Inventory; Personality Measures; Psychometrics; Reliability; inpatient assessment; Multi-item Multi-trait Analysis Program; psychometric properties; personality assessment; psychiatric inpatients; Personality Assessment Inventory (PAI)
• ISSN: 1073-1911; 1552-3489
• DOI: 10.1177/1073191109333756

Multi-item multiscale self-report measures are increasingly used in inpatient assessments. When considering a measure for this setting, it is important to evaluate the psychometric properties of the clinical scales and items to ensure that they are functioning as intended in a highly distressed clinical population. The present study examines scale properties for a self-report measure frequently employed in inpatient assessments, the Personality Assessment Inventory (PAI). In addition to examining internal consistency statistics, this study extends prior PAI research by considering key issues related to inpatient assessment (e.g., scale distinctiveness, ceiling effects). Coefficient alphas, interitem correlations, and item— scale relationships suggest that the PAI clinical scales and subscales are internally consistent. Items for respective clinical scales generally showed significantly higher item—scale correlations with their intended scale (as compared with their item—scale correlation with scales they were not intended to measure). In addition, scales’ coefficient alpha scores were higher than their interscale correlations. Taken as a whole, these results support the hypothesis that PAI scales were measuring relatively distinct constructs in this inpatient sample. Findings are discussed with regard to the implications for scale interpretation in inpatient assessment, functioning of individual scales and subscales, and functioning of specific items. Limitations of the present study and directions for future research are discussed.