Optimum split-half reliabilities for the Rorschach: projective techniques are more reliable than we think.

Area of Science:

• Psychometrics
• Psychological Assessment
• Personality Testing

Background:

• Assessing the internal reliability of psychological tests is crucial for valid interpretations.
• Traditional methods for estimating split-half reliability may not account for distributional properties, potentially leading to inaccurate results.
• The Rorschach test, a widely used projective technique, requires robust reliability estimates for its variables.

Purpose of the Study:

• To introduce and evaluate a new technique for optimizing split-half reliability estimates.
• To examine the impact of this technique on the internal reliability of major Rorschach variables.
• To provide guidance on applying this method to other projective instruments and address limitations of existing approaches.

Main Methods:

• A novel optimization technique was developed to refine split-half reliability calculations.
• The technique was applied to specific Rorschach variables across two distinct participant samples.
• Internal reliability indexes were computed and compared before and after applying the optimization method.

Main Results:

• The optimization technique produced substantial and comparable increases in internal reliability for several Rorschach variables.
• The improvements were consistent across two diverse samples, suggesting generalizability.
• The study identified specific distributional anomalies that can affect odd-even reliability coefficients.

Conclusions:

• The proposed technique offers a valuable enhancement for split-half reliability estimation in psychometric research.
• This method can improve the reliability of Rorschach variables and potentially other projective tests.
• Researchers should be cautious when using standard odd-even reliability coefficients without considering underlying data distributions.