Method Matters: Underreporting of Intimate Partner Violence in Nigeria and Rwanda

Claire Cullen

doi:10.1596/1813-9450-9274

Abstract

This paper analyzes the magnitude and predictors of misreporting on intimate partner and sexual violence in Nigeria and Rwanda. Respondents were randomly assigned to answer questions using one of three survey methods: an indirect method (list experiment) that gives respondents anonymity; a direct, self-administered method that increases privacy; and the standard, direct face-to-face method. In Rwanda, intimate partner violence rates increase by 100 percent, and in Nigeria, they increase by up to 39 percent when measured using the list method, compared with direct methods. Misreporting was associated with indicators often targeted in women's empowerment programs, such as gender norms and female employment and education. These results suggest that standard survey methods may generate significant underestimates of the prevalence of intimate partner violence and biased correlations and treatment effect estimates.

Highlights

It is challenging to accurately measure sensitive attitudes, behaviors, and experiences such as political preferences, prejudice, risky sex, and intimate partner violence (IPV)
There is suggestive evidence of a positive relationship between IPV and women’s education when IPV is measured directly, but a negative relationship when using the list method. These results suggest that the standard, direct methods currently used in most surveys risk generating significant underestimates of IPV prevalence
If the difference in prevalence between the list and direct methods is a proxy for fear, shame or social desirability bias, these results suggest that questions about emotional and physical IPV are more sensitive than sexual violence, at least in Nigeria where the experiment had more statistical power

Summary

Introduction

It is challenging to accurately measure sensitive attitudes, behaviors, and experiences such as political preferences, prejudice, risky sex, and intimate partner violence (IPV). In Rwanda and Nigeria, we randomly assigned women to answer questions about their experience of emotional, physical and sexual violence using one of three methods: an indirect method (list experiment) and two direct survey methods: face-to-face questions asked by an enumerator, or audio computer-assisted self-interview (ACASI) on an electronic tablet. Compared to the direct face-to-face method, the list experiment method provides additional confidentiality as respondents do not directly disclose their experience of violence. Instead, they give a response that allows researchers to calculate average prevalence across the whole sample, essentially making respondents’ answers anonymous (Blair and Imai, 2012). The ACASI method affords more privacy than face-to-face methods as respondents answer questions themselves by listening to pre-recorded questions in a headset and selecting their answer on a tablet

Objectives

Methods

Results

Conclusion