It was an ambitious goal: rigorously mainstreaming gender into our eight year, eight country (Afghanistan, DRC, Nepal, Pakistan, Sierra Leone, South Sudan, Sri Lanka, and Uganda) research project, the Secure Livelihoods Research Consortium.
Here’s what we’ve learned: what went well, what didn’t, and what we’d do differently with hindsight.
Leadership commitment and expertise matters.
The SLRC leadership team started with an explicit commitment to rigorous gender mainstreaming. The leadership team was well versed in gender theories, frameworks, data collection, management and analyses. They had past research records of producing strong gender analysis. This set the right tone for the entire project.
Country teams with strong, senior level gender expertise did the best gender research.
Unsurprisingly, we found that the level of gender analysis, and analysis overall, was outstanding within teams that had both women and men researchers who were well versed in gender and feminist theories, methods, data collection and analysis.
Gender analysis needs to be non-negotiable.
We had to work to persuade everyone in SLRC that gender mattered. Everybody agreed about the importance of having mixed research teams and ensuring we interviewed both men and women in our fieldwork. Yet not everyone bought into gender mainstreaming in our analyses. Some argued, often quite convincingly, that we would be better to focus on quality than quantity – that we would produce more insightful and useful analysis if we dedicated our gender expertise to specific projects and activities, rather than spreading it too thinly across all our work. In the end we did both: some high quality specific gender focused work, and a disaggregation of our data by gender where it mattered.
Ensure gender peer review for every working paper and report.
Every paper we published underwent gender peer review by people with gender expertise on the subject matter and or geographic area of focus. The peer review process enhanced the overall outputs of each paper and the SLRC as a whole.
Make delivering on gender mainstreaming everyone’s business.
We had some successes and some disappointments. In several cases, men and women with gender expertise from different national and international teams paired up on papers to produce very strong results. Yet, our gender task team – mandated to ensure that gender was embedded in the research process, analysis and uptake in every country – was overwhelmingly dominated by women. We were frustrated that we didn’t manage to get a better gender balance in our own gender team.
Building gender expertise in research teams is critical but time-consuming.
While we had strong gender expertise on nearly every team, researchers with less expertise were interested to learn how to better incorporate gender from study design through results write up. The SLRC held workshops, appointed gender focal points, and sought to build the capacity of any researcher trying to improve their gender analytical skills. This is a model that is worth replicating. However, we learned that it can take longer than anticipated for researchers to build the skills necessary to move their gender analysis beyond disaggregating the data by sex.
For best results, go all in for gender mainstreaming research.
The SLRC’s gender mainstreaming approach sought to tackle all stages of research: applying theory to define research questions, developing research questions, methodologies, data collection teams, data collection, data analysis and production of outputs. We managed this better in some stages than in others: we embedded gender in our research methodologies; tried to include men and women in every research team; held women-only focus groups; disaggregated data by gender/sex and in our statistical analysis sought explanations where we found significant differences in men and women’s responses or experiences. Yet, we could have done better at integrating a gender lens in the very early stages of research. In retrospect, we didn’t always define our research themes by drawing on gender theory adequately, or incorporate gender into our research questions enough and sometimes found ourselves working hard to make up for that later.
Learn from missed opportunities.
Not everything went as we hoped – and it’s important to reflect on, and learn from, such experiences. Most notably, we missed the opportunity of comparing female and male headed households across the five countries involved in the panel survey. This was due to two factors: first, we couldn’t work out which was optimal – setting a universal definition of household headship, or letting country teams define it in ways appropriate for their local context. We chose the latter but then regretted this decision. The problem was that many countries allowed respondents to self-report on headship so we could neither make comparisons between countries, nor actually measure headship with any accuracy within a country. For all we knew, different respondents defined headship using different criteria.
Second, because we didn’t provide enough guidance or instruction, in some of the countries, we failed to capture headship on the survey forms in both waves so we can’t see it changing over time. The few countries that were consistent in their approach within country did come up with important and significant findings regarding the conditions of female headed households compared to male headed households. But overall we lost an important opportunity to be able to say something about how female headed households fare compared to male headed households within and among conflict affected and fragile countries. With a third wave of the survey planned for 2018, we hope to be able to learn from and rectify this in the future.