The practice of social science research in Ethiopia: some observations-II
Part II: Obsession with Fancy Models and Some Recommendations
In this part, I discuss my second concern with the practice of social science research and conclude with some suggestions.
The second problem in social science research is the increasing overreliance on fancy models. The use of fancy models is often motivated as a tool to study causation. Also, it is increasingly used to quantify everything, even those inherently qualitative. However, we know that most models are based on strong assumptions, which are hardly met in reality. Moreover, capturing human behavior with models is a tall order, given randomness of human behavior and measurement issues.
Despite the shortcomings of models, it seems that we all have accepted that we should use some model to undertake any kind of research. The proliferation of user-friendly software has also made possible even those with rusty maths and statistics background use fancy models. There is no wrongdoing in using models as long as it helps to answer the research questions. The problem is applied researchers are increasingly judged by the kind of fancy models they use, rather than in their real contributions to improving our understanding of an issue. When asked about their research topic, it is common to witness researchers immediately dive into talking about the fancy models that they are applying. Talking about the latest fancy models might make us look smart. I get that. But we should also know that we are not credited for using a model somebody else has labored to give its birth.
Because of the obsession with causality and models, applied social science researchers, including a good number of applied economists, are increasingly being indulged in some form of p-hacking—massaging data so that a ‘black box’ model could yield results that are easy to explain. This laborious work may add value in terms of publication or career, but it adds little to our understanding of the subject matter at hand.
So, the question is how to make social science research more relevant for Ethiopia? Here, I make two recommendations but would like to request you to add more doable recommendations. First, making it compulsory for graduate students in social science departments including economics, write at least one article without using any model as part of their thesis. Of course, except for the few mathematical/statistically inclined students who could make the most out of models. In doing so, less mathematically prepared researchers including economists would be relieved from working with fancy models, in which they usually have little grip. Researchers’ energy and time could then be rightly placed. They will be focused on asking relevant research questions, and spotting and generating appropriate qualitative and quantitative data that can be used to duly answer the questions.
Second, encouraging graduate students and researchers alike to contribute to popular newspapers and magazines based on their research projects and current issues, and making this count in their career progression. Communicating with the general public via the popular media would force researchers to come out from their hideouts—models and jargons. They will start asking themselves about their meaningful contributions that they can show to the general public. I guess this will eventually bring applied social science researchers closer to what they profess to serve.
