Mode of Administration Effects in Psychopathology Assessment: Analyses of Gender, Age, and Education Differences in Self-Rated Versus Interview-Based Depression
Respondents may answer sensitive questions differently depending on the mode of assessment (e.g., questionnaire, interview). The possibility of differences in these mode effects across groups is an important assessment consideration given the implications for test-score validity and bias. Although differences in mode effects are highly relevant to issues in clinical assessment, research is limited and generally fails to make use of psychometrically sound methods. The present research argues for greater attention to demographic differences in mode effects with regard to the assessment of psychopathology. Analyses of differences in mode effects in depression assessment are presented for gender, age, and education based on data from 440 community adults. Depression symptom ratings from a structured clinical interview were regressed on a self-report state depression composite score; analyses tested for differences of intercept and slope across gender, age, and education. Neither slope nor intercept differences were found for gender, but intercept differences were obtained for both age and level of education. These results indicate that younger and better educated respondents receive somewhat lower than expected ratings of depression in interviews, given their level of self-rated symptoms. These findings suggest that use of a single mode of assessment over- or underpredicts depression in certain participants. More generally, they demonstrate the value of multimethod assessment of psychopathology and justify further inquiry into mode-effect differences in clinical assessment.