Corrections For Spurious Influences On Correlations Between Mmpi Scales

Area of Science:

• Psychometrics
• Psychological Measurement

Background:

• Correlations between measures with shared items are inflated due to overlap.
• This spurious component affects personality scales like the MMPI and CPI.
• Existing adjustment methods are insufficient.

Purpose of the Study:

• To address spurious correlations caused by overlapping items in psychometric measures.
• To apply Bashaw and Anderson's (1967) reliability solution for accurate correlation analysis.
• To demonstrate the impact of item overlap on correlation structures.

Main Methods:

• Review and comparison of statistical and experimental adjustment methods.
• Application of Bashaw and Anderson's (1967) reliability solution.
• Analysis of the "Minnesota normals" MMPI data.
• Principal components analysis to examine correlation structure changes.

Main Results:

• 32 out of 36 correlations were altered after adjustment.
• Six correlations were radically changed, and two reversed sign.
• Principal components analysis indicated significant shifts in the correlation structure.

Conclusions:

• Item overlap significantly distorts correlations in personality scales.
• The reliability solution effectively corrects for spurious correlations.
• Accurate correlation analysis is crucial for understanding psychometric relationships.