Uncovering potential interviewer-related biases in self-efficacy assessment : a study among chronic disease patients

Background: Self-efficacy refers to an individual's belief in their ability to accomplish specific tasks and achieve goals, and plays an essential role in achieving positive outcomes in a wide range of domains. Central to the measurement of any form of self-efficacy is the assessment without bias, also in case of an interview situation. Methods: Outpatients with macular edema, an eye disease, participated in this questionnaire-based cross-sectional study. The study assessed self-efficacy using the General Self-Efficacy Scale (GSE) in German. Interviewers read questionnaires aloud to patients. Differential item functioning (DIF) was investigated using likelihood-ratio χ2 tests for interviewer, sex, age, education, working status, income, diagnosis, and health-status. Results: The analysis included N = 556 patients. Median age was 68.4 (IQR: 62.0 – 76.0) years and mean overall GSE score 32.8 (SD: 4.81). No DIF was detected for interviewer. However, DIF was found in item 1 for education (uniform DIF, NCDIF_{no degree vs. degree} = 0.042; easier with degree vs. none), in item 1 and 3 for income (item 1: non-uniform DIF, NCDIF_{<€ 1,125 vs.≥€ 1,125≤€ 1,950} = 0.050 / NCDIF_{< € 1,125 vs.≥€ 1,950} = 0.099; item 3: uniform DIF, NCDIF_{<€ 1,125 vs.≥€ 1,125—≤€ 1,950} = 0.024 / NCDIF_{< € 1,125 vs.≥€ 1,950}: 0.095; both easier with higher income), in item 2 for working status (uniform DIF, NCDIF_{retired vs. other} = 0.017; easier if working) and in item 3 for sex (non-uniform DIF, NCDIF_{male vs. female} = 0.043; easier for women in low ability, harder for them from medium ability on). Conclusions: Given that no DIF was detected concerning interviewers, our findings indicate that an objective assessment of self-efficacy in a face-to-face interview may be feasible, provided that interviewers receive appropriate training. Since DIF effects concerning other patients characteristics found were small, the GSE may provide a relatively bias free way to assess self-efficacy in an interview setting.