I’ve encountered many obstacles during my time at the Urban Institute, ranging from poorly managed data, to weird technical glitches, to poorly engaged supervisors, to paywalls blocking my access to important academic obstacles. I overcame all of these obstacles through mix of determination, ingenuity, intellectual humility, and (when all else fails) asking for help. This is the same method I used to conquer my biggest problem: the fact that my data made no sense.
The data I’ve been working with during the past 6 weeks is a survey data-set of slums in the developing world. The purpose of the project that collected the survey was to identify ways in which environmental degradation disproportionately impacted slum-dwellers, an in particular female slum-dwellers. Simple, right? Not really: the data seemed to belie any of the obvious relationships we would expect to see between climate change and other variables.
However, by conducting more research on the studied populations, I came to understand how the different backgrounds of respondents could influence their perspectives and therefore the way they responded to the survey. Taking this into account, I was able to rule out many of the variables I was looking at and focus on the ones that would respond least to the biases of the respondents. Using this method, I was able to find more significant results concerning our research questions!