Fieldworker-Induced Social Desirability Bias in wife beating justification reporting: An insight from large scale survey in India

Saurabh Singh, International Institute for Population Sciences (IIPS)

This study investigates the Fieldworker-Induced Social Desirability Bias in justification of wife beating and impact of Fieldworker as well as third-person presence on women's responses regarding the justification of wife beating. The study utilized data from the individual file and the fieldworker characteristics file of the NFHS-5. Logistic regression, Propensity Score Matching (PSM) and Bayesian cross-classified multilevel modelling were performed. The presence have found less likelihood if respondent and FW were from same place of residence (OR=0.84, 95% CI: 0.75 - 0.93), have same religion (OR=0.82, 95%CI: 0.72 - 0.92), have same native language (OR=0.86, 95%CI: 0.73 - 0.99), and FW have higher level of education (OR=0.8, 95%CI: 0.67 - 0.96). PSM results show if third person were present during the interview the justification of beating for going out without telling husband, refuses to have sex with husband and burns food become 4.3% (95% CI: 1.5-7.2), 6.1% (95% CI: 3.7-8.5) and 5.4% (95% CI: 2.7-8.1) higher. The intraclass correlation coefficient for wife beating justification at the fieldworker level ranged between 29%-33%. It was observed that variation due to fieldworker for non-sensitive outcomes was negligible. This study highlights the role of third-person presence and Interviewer-Induced Social Desirability Bias in influencing responses.

Keywords: Data and Methods, Multi-level modeling , Bayesian methods

See extended abstract.