Symposium: The Role of Gender in Data Collection through Household Surveys
Survey research allows us to enter the homes of millions of people around the world to uncover some of the most intimate aspects of people’s daily lives. It is used to address some of the most challenging societal issues we face, including forced labour, forced marriage, drug use, gender-based violence and other hidden crimes. To measure the extent of these issues, people are often asked a series of sensitive questions by a stranger who comes to their door. It comes as no surprise that some information is more readily shared than other types of information. We’ve seen the challenges of this first-hand in our attempts to measure modern slavery, particularly, forced sexual exploitation. Conducting this research requires careful integration of gender inclusive language and measurement standards in order to ensure that these resource-intensive and valuable measurement efforts are optimized and accurately reflect reality.
In modern slavery research, surveys represent a measurement gold standard when sampling strategies are built upon official statistics for stratified, probabilistic and random sampling. Survey research represents a significant investment in time and financial resources, and is a rare opportunity to reach a broader population with questions related to modern slavery. Optimizing these opportunities benefits the entire field because it allows us to reach potentially unidentified survivors and victims of modern slavery that may otherwise never be contacted. There are two main aspects of survey research and gender that will be discussed in this piece: gender considerations around enumerators or interviewers, and gender considerations around respondents.
Gender and survey enumerators
Enumerators or survey research interviewers are a critical link in our global research chains. Enumerators are the sole researchers coming into direct contact with our research respondents, and everything about their personal characteristics, presentation and representation of the groups with which they are affiliated are known to affect whether survey respondents participate in the research project or whether their answers are influenced by the enumerator. Mismatched gender, among other sociodemographic characteristics, has been shown to affect unit and item nonresponse in surveys.
Two main theories explain the impact of interviewer gender on responses. Social distance theory posits that when respondents and interviewers are the same gender, the response rate is higher and responses themselves are more accurate. For example, research suggests that interviewers obtain a greater number of responses to sensitive questions if they are the same gender as the respondent. Social desirability theory suggests that respondents will try and provide response that will be perceived favourably by the interviewer. Studies have shown that when female interviewers in the US and Mexico ask questions about gender equality, they are more likely to receive egalitarian responses. In the Arab states, responses to the same question were affected by both gender and religion (as denoted by the interviewer wearing religious dress).
There are also some interesting preliminary findings from one round of the Arab Barometer, which show that interviews conducted by female interviewers tended to be approximately 4 minutes longer on average. More research is needed to ascertain whether the longer interview duration indicates a greater willingness among respondents interviewed by a female to discuss the issues or if female interviewers take more time to build rapport or ask questions. Additionally, it is unclear if either of these factors has an impact on the quality and accuracy of the information being collected.
In order to account for gender specifically affecting social desirability bias in survey administration settings, several efforts have advocated for gender-matching, but even this approach is not complete. There are other substantial aspects of an enumerator’s background and presentation including cultural, socioeconomic and educational characteristics, that can also influence survey respondents and their willingness to participate and to provide honest responses.
Gender and survey respondents
Importantly, the respondents that participate in our surveys are the source of all the insights that can be derived from this research. In addition to encouraging their participation, it is critical that we protect all survey respondents. Conducting research in the private homes of respondents is also fraught with many challenges, not only preserving safety but also access and considering who is permitted to speak on behalf of the household. Compounding this complexity is the reality that in many situations asking about forced marriage and about family members in forced labour may place respondents in potential situations of discomfort, if not outright risk, if their abusers live within the same household. Ethical standards and protections are often put in place to ensure that survey respondents are supported by local referrals and experts, as well as equipped with the necessary information to discuss their situation further if desired.
There are still serious challenges that must be dealt with, especially in cultures with strong patriarchal traditions where women may not be interviewed without a chaperone or witness. The presence of a third-party, which is sometimes a child or another family member who is curious about the process and wants to listen to the interview, is rarely coded or considered in final research analysis, yet this is likely to have a substantial impact on the respondent’s willingness to report honestly. For example, in the Arab Barometer research mentioned above, when asked “Husbands should have final say in all decisions concerning the family”, there was greater agreement when a third party was at the interview.
The impact of gender should also be considered in the cognitive testing part of the survey process to ensure that our language is inclusive and reflective of the societies we are studying. Ensuring that we provide equitable opportunity for male and female respondents to participate and that both understand the questions, can recall the information sought and are comfortable responding is critical. Gendered constructions of occupations, marital status and even the actual survey language itself underpin our ability to accurately measure the prevalence of all forms of modern slavery.
Gender inclusivity and considerations are critical in the survey research we undertake in the modern slavery field and beyond. For such significant investments, we have an obligation to ensure that our methods of data collection are not a barrier to participation for any members of a community, especially the world’s most vulnerable women and girls.
This article has been prepared by Jacqueline Joudo Larsen as a contribution to Delta 8.7. As provided for in the Terms and Conditions of Use of Delta 8.7, the opinions expressed in this article are those of the author and do not necessarily reflect those of UNU or its partners.
 Joudo Larsen, J & Okech, D (2021) Prevalence Estimation: Methods Brief – Method 01: Household Surveying. Global Fund to End Modern Slavery. Available from: https://www.gfems.org/news/2021/2/10/resource-prevalence-estimation-methods-brief