Discovering new ways of thinking with EfD

7th May 2026

When I started my internship, I expected to work closely with a lot of data. However, what I never expected, was for how this opportunity changed the way I think, not just about development but about research itself.

My internship was based on two vastly different contexts. The first was centred around the individual household economy of an informal settlement in Windhoek, Namibia and the second was modelling the coping and response to flooding for rural smallholding farms in Kaikamosing, Uganda.

Namibia: Inequality

Okahandja Park is an informal settlement in northwestern Windhoek, roughly 6km from the city centre. Namibia is one of the most unequal countries in the world, with a Gini coefficient of 59.1, and 43.3% of the population living in multidimensional poverty. Through the analysis of 48 households in Okahandja Park (data from a survey carried out in 2012), I explored the different income sources, food consumption, and the gap between what they earned and the standard of living. I found the contrasts extraordinary. The highest-income household in the study had a surplus of over 128,000 NAD above their standard of living threshold. Another household, with nine members and a negative disposable income of -8,015 NAD, was surviving almost entirely on food transfers. These households existed within the same settlement yet had entirely different economic realities.

Uganda: Flooding

In Kaikamosing, Eastern Uganda, the vulnerability looks different, but the logic is the same. Uganda is a country where 7 in 10 households depend on agriculture for income, and 1 in 3 people rely directly on their own crop production for food. My work here involved modelling the impact of flooding on smallholder farmers, applying flood-driven crop loss rates, 35% for grain crops and 60% for root crops like cassava and sweet potatoes, to real household income data. Households that appeared to have a reasonable surplus before the flood scenario were pushed dramatically towards, or below, their standard of living threshold afterwards. For those already operating on the margins, the modelled losses were devastating. Furthermore, with limited livestock assets, the available coping responses had minimal effect in offsetting the losses to crop income.

Scrutinise data

Another thing I learned from my internship, was not from the data itself, but how it changed my relationship with data analysis. Before this experience, I observed data at face value. However, since working so closely with the data from the household economy approach, plus additional research for the context of each settlement, I learned to really scrutinise that data. It was an opportunity to really question what sort of story the data was telling me, what I was assuming and what was fact. Moreover, looking at the research from Windhoek and Kaikamoising, I had to observe the data within its context and not just make generalisations. This kind of critical engagement with the data is something I was aware of before but never truly practiced in depth.

Lastly, this opportunity taught me that vulnerability is not random or poor luck, but structural. Whether it is from poor wages, infrastructure, health care, or a failed harvest, the households most exposed to disruption are consistently those with the least cushion to absorb it. That realisation deepened my appreciation for Evidence for Development’s mission, because gathering rigorous, context-specific data is not just an academic exercise. It is the foundation for making aid provision more effective, more targeted, and more meaningful to the people who need it most.

Josephine Jackman