Estimating the accuracy of psychological scales using longitudinal data

Area of Science:

• Psychometrics
• Biostatistics
• Clinical Psychology

Background:

• Psychological screening scales are crucial for diagnosing conditions like major depression.
• Existing methods for accuracy estimation may not fully capture the complex distributions of screening scale data.
• Accurate measurement is vital for effective patient assessment and treatment.

Purpose of the Study:

• To present a novel statistical method for estimating the accuracy of psychological screening scales.
• To model the unique distributional properties of screening scale data, including semicontinuity and varying skewness.
• To apply the proposed method to assess the Symptom Checklist's accuracy in detecting major depression.

Main Methods:

• Utilizing receiver operating characteristic (ROC) curves and associated statistics for accuracy estimation.
• Modeling screening scale outcomes with truncated normal distributions to account for data characteristics.
• Incorporating subject-specific random effects to handle repeated measurements within individuals.

Main Results:

• The proposed method effectively estimates the accuracy of psychological screening scales.
• Truncated normal distributions accurately represent the observed data patterns.
• The Symptom Checklist's accuracy for major depression was estimated using the developed model.

Conclusions:

• The presented method offers a robust approach for evaluating the accuracy of psychological screening instruments.
• The statistical modeling accounts for complex data distributions and repeated measures, enhancing reliability.
• This work provides a valuable tool for researchers and clinicians assessing psychological conditions.