Quality survey questions are extremely important. While cognitive interviewing can help refine survey questions for endogenous comprehension, questions relating to autonomy and norms require even more careful consideration. In the original Women Empowerment in Agriculture Index (WEAI), the autonomy indicator is collected using a set of questions about motives for decisions. Let us consider the following WEAI question:
Regarding [ASPECT ] I do what I do because I personally think it is the right thing to do. Would you say that this is always true, somewhat true, not very true or never true?
In several instances both respondents and enumerators noticed that these questions regarding motives for decisions were difficult for participants to understand because people are uncertain about why they do what they do. Furthermore when a respondent gives an answer, it is not easy for the enumerator to identify which response code it maps. For example, in some languages it is nearly impossible to distinguish between “somewhat true” and “not very true”. Another concern is that people may not actually feel like they are making decisions in these aspects, especially if they feel they are following cultural norms, or when other factors constrain the decision. The autonomy section is still one of the most challenging parts of the WEAI to implement. A high degree of training and practise is needed for this to be performed well.
Men and women’s motives may be determined by coercive or internalized social pressures to undertake specific actions. Certain questions can help to capture this. For example, whether a woman thinks intimate partner violence (IPV) is justified captures an aspect of women’s own sense of empowerment as beliefs in regressive gender norms can constrain individual agency. The following statements were designed to reflect domains representative of deeply entrenched gender norms in Nepal and to reflect on attitudes towards sensitive concepts and behaviours such as intimate partner violence, son preference, and gender attitudes:
· A woman should tolerate violence in order to keep her family together.
· There are times when women deserve to be beaten.
· A woman should obey her husband in all things.
· Women should leave politics to men.
· Women’s work should be limited to household chores such as cleaning and cooking.
· Education is not valuable for daughters/daughters-in-law.
Questions or statements should be introduced to reflect domains representative of deeply entrenched gender norms in each specific country. As an example, in Pakistan, when measuring women’s economic roles, leadership, the right to keep financial records, and education are contentious issues, and are not easily captured by exogenous framings. Whereas in Lebanon women’s empowerment is deemed to be associated with knowledge about accessing the justice system. In such a case, it is advisable to assess women’s knowledge by asking each respondent a number of questions concerning the rights and laws embodied in the juridical system(s) in Lebanon. This could be followed by a number of questions about individual capacity (i.e., willingness and knowledge to take legal action if required) in order to investigate whether this knowledge is applied. Additionally, distinguishing between the possession of knowledge and the application of knowledge is of particular importance, as it produces indications of whether knowledge is a constraint for implementing certain practices or if there are other constraints that affect this choice.
While logical, many surveys do not start with pre-questions in order to prepare for a complex question. However, this is necessary in order to capture social norms. The choice of such culturally- and context-specific questions should be informed by previous studies conducted in the area/site on gender norms and attitudes. This would reduce the number of questions required over time.
In the literature there is some debate about whether questions about attitude towards women could ever be used to generate an index capable of reflecting critical consciousness. Some concepts are very difficult to capture in one moment of time and can only be determined over time, suggesting the need for panel datasets on women empowerment.
” It is important to remember that measuring something as sensitive as gender-based violence (GBV) can be particularly challenging and requires extra caution. When investigating this topic, the evaluator has to put in place a number of precautions to maximize the disclosure of actual violence and—more importantly—reduce the risk to anyone involved in the survey, such as the women interviewed or the enumerators. Questions on violence should not be asked unless privacy is ensured, and enumerators should reschedule the interview to another time if necessary.
If another person (excluding children not yet able to understand) enters the room, enumerators should be trained to switch the subject and to interrupt the interview if necessary. If, despite repeated attempts, privacy cannot be obtained, this section of the questionnaire should not be implemented. Enumerators should be properly trained in how to deal with cases of violence and should be equipped with a list of services and associations that can provide support to any respondent experiencing domestic violence, if they are asked for help. Finally, emotional support should also be provided to enumerators.”
This blog has focused on improving survey questions in order to capture gender norms, the next blog explores practical (ethnographic-inspired) tips for qualitatively measuring women’s empowerment.
[1] Sources: Lombardini, S., Bowman, K., & R. Garwood (2017); Malapit, H.J.L., Sproule, K., & C. Kovarik (2017); O’Hara, C., & F. Clement (2018)
[1] For further information see: the Ethical and Safety Guidelines for Implementing the DHS Domestic Violence Module: https://www.dhsprogram.com/topics/gender-Corner/upload/DHS_Domestic_Violence_Module_Ethical_Guidelines.pdf
