The Common Item In Measurement: Effects On Structure

Area of Science:

• Psychometrics
• Personality Psychology
• Quantitative Psychology

Background:

• The influence of shared items on personality structure analysis is a critical consideration in psychometric research.
• Previous studies have raised concerns about potential biases introduced by common items in personality assessments.

Purpose of the Study:

• To investigate the impact of common items on the inferred personality structure using the California Psychological Inventory (CPI).
• To compare personality structures derived from original scales, reduced scales (excluding common items), and common items alone.

Main Methods:

• Smallest space analysis was employed to analyze three correlation matrices: original CPI scales, reduced scales, and common items.
• Visual comparison, interpoint distance correlations, and configurational similarity analyses were used to assess structural congruence.

Main Results:

• A high degree of similarity was observed between the personality structures derived from the original scales, the item-overlap-free structure, and the structure of the common items.
• The analyses indicated that common items tend to amplify the intrinsic structure present within the scales.

Conclusions:

• The presence of common items does not appear to introduce significant distortion in personality structure analysis.
• Common items likely enhance the visibility of an existing underlying structure rather than creating artificial patterns.
• The 'common-item problem' is unlikely to pose a serious threat to the validity of personality measurement.