Examining the factor structure of the 40-item Defense Style Questionnaire among adults with depression: an individual participant data meta-analysis

The 40-item Defense Style Questionnaire (DSQ-40) is frequently used to measure defense mechanisms in adults with depression because of its brevity and efficient assessment. However, previous research has yielded inconsistent results concerning its factor structure, with one- (overall defensive functioning), three- (mature, neurotic, immature), and four- (mature, neurotic, immature, image-distorting) factor solutions being reported. Therefore, we examined the DSQ-40's latent structure across different samples. Using multigroup confirmatory factor analysis (CFA), we tested the one-, three-, and four-factor solutions using baseline data from three research projects of psychodynamic psychotherapy for depressed adults (N=667; mean age 33.9; standard deviation [SD]=9.3; 79.0% female). Configural invariance across samples was evaluated, and model fit was compared using multiple fit indices. Across samples, the three- and four-factor models outperformed the one-factor model, with the four-factor model providing the best relative fit; however, it fell short of conventional model fit thresholds. Separate CFAs found the four-factor model showing the best relative fit in two samples, and the three-factor model in one, but all models failed to meet acceptable fit criteria. Several defense mechanisms showed weak standardized loadings, high inter-factor correlations, and cross-loadings. These findings indicate that the DSQ-40 lacks a stable factor structure in adults with depression and raise important questions about the instrument's construct validity, suggesting that its subscale scores cannot be assumed to validly represent the theorized defense categorizations in this population. Further psychometric refinement and evaluation are needed to determine the DSQ-40's suitability for both clinical and research applications.