This Data for Health Gender Chat discussion focused on the challenges of collecting accurate and complete gender data and how to conduct better analyses using proxy indicators for gender norms to improve health outcomes.

Common Challenges in Gender Data Analysis and Strategies to Overcome Them

Missing data

One way to overcome missing gender data is with the use of proxy indicators. For example, if a survey didn’t measure the gender norms around intimate partner violence, a proxy indicator could be constructed using a measure of women working outside the home to see how common it is and discover any correlations with rates of intimate partner violence. Using several proxies for the same gender norm variable can be quite valuable when conducting a gendered data analysis.

Validity of data

Another limitation lies in the validity of the data based on how the data was collected, in what setting and if the respondent felt confident and safe enough to provide accurate data. Data collection methods (mobile and in-person) need to match the interviewer with the respondents’ characteristics because there can be gender bias if a male interviewer is talking to a female or other gender interviewee. Training interviewers about bias, sensitivity, respect, and management of the interview situation is critical.
For example, sexual activity can be perceived as a negative behavior for single women in some cultures, as well as behaviors such as smoking, working outside the home, etc. Respondents may not want to answer questions truthfully if they feel they are being judged by a biased interviewer. Understanding country-specific context and how gender norms affect reporting of behaviors is important.

Complete data

Many data collectors face the constant challenge of attaining complete, quality data. One solution is to opt to use multiple data sources such as when investigating gender-based violence, police records and death statistics. There is still a challenge in using multiple datasets, however: the heterogeneity of the data and data definitions may vary among sources.

Decision-Making Flowchart for Constructing Gender Norm Proxies

This decision-making flowchart is designed as an aid throughout the process of gender data analysis. This chart is useful to question data availability regarding gender norms and how the available data on gender norms (or their proxies) can be used for exploring health outcomes and health outcome determinants. (Click here for more information on how to use proxy indicators.)

As shown in the first step of flowchart, it’s critical to begin by looking at the existence of any data about gender norms, and then to consider using proxy norms in situations when direct measure of gender norms is not available. It’s also important to question how the data was collected, by whom, and how accurately it measures the health outcome of interest. This is of special interest because oftentimes analyses report differences in the minimum gender data collected and disaggregated (biological sex and age) and then when there are gaps in health outcomes, there is no information about the reasons why.
If there are existent data on gender norms, then we can think about how these gender norms are distributed by groups in the data and correlate them with the health outcomes (such as intimate partner violence (negative) or vaccination (positive). However, it is more common that surveys don’t directly measure gender norms.

Further Resources on Gender Data Analysis

Sex and Gender Differences Research Design for Basic, Clinical, and Population Studies: Essentials for Investigators
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7263836/
Video: Gender Analysis in Health Data
This video provides an overview of what gender analysis in health data means and methods to apply gender analyses to large health datasets, with a focus on the differences between sex disaggregation and gender analysis using gender equity indicators.
https://genderhealthdata.org/resource/seminar-recording-gender-analysis-in-health-data/