Causal Relationships and the Tools of Applied Econometrics - Workshop at Makerere University

I have a keen interest in broadening participation in social sciences, especially via the scholarship of bright young people in Uganda. There is much to gain from Ugandan insights and perspectives. One contribution towards extending these insights to reach a global social scientific discourse is through offering trainings on methodologies of knowledge generation that influence global conversations. One such methodology is the methodology emanating from the tools of applied econometrics.

To ensure that Ugandan insights can be translated through these methods, students need to engage with them critically. Students have to learn to navigate the tools without over-extending their use or allowing the tool to dictate the questions that need to or can be asked.

This is an attitude I brought into a series of workshops at Makerere University where I engaged masters students in the School of Economics and the School of Statistics to develop a sense of ownership over the use of the tools used to infer causal relationships in quantitative data. Let me describe this approach with an example.

I ask students to think of an intervention that they think has some impact on educational outcomes. Then I ask them to write down the names of their friends and consider how such an intervention would affect their test scores with and without the intervention. This implicitly introduces students to the potential outcomes framework in, e.g., Angrist and Pischke, without explicitly mentioning it as such (at least not initially, so as not to scare students with jargon). Then, we collectively decide on a rule that determines who selects into the educational intervention (for example, those who have the most to gain). We calculate the observed difference between the treated and untreated groups and ask ourselves whether this suggests a causal association between the intervention and the outcome we observe.

Without providing a definitive answer, we proceed to another exercise. Here, each student randomly selects which of the friends they have listed receives the hypothetical intervention via coinflip. Then, we again proceed to calculate the difference between the two groups. Students begin to realize that the difference between these two groups is not the same as the one generated when selecting students into groups by rule. We also plot the histogram of the differences that all the students create to show that they are normally distributed. We vary the number of students in such an exercise and see that the shape of the distribution changes. Again, I pose questions that allow the students to analyze the stylized facts they are observing themselves. Why are the differences generated by each simulation different from each other? Why are they different from the protocol that assigned people to treatment by rule? Why do we see the differences of the simulations creating a bell curve distribution? And so on.

This is the sense in which students are “protagonists of their own learning” in the courses I design. I create the materials and learning environment where their own processes of investigation lead them to develop the necessary skills associated with a methodology or an understanding of a topic. For example, students who perform the above-mentioned exercise quickly begin to understand the challenge of determining causal relationships in data on their own. They realize that rule-based selection speaks more about the characteristics of the individuals who select into a treatment than the treatment itself. After the students have sufficiently analyzed and examined the problem on their own – without any need to absorb knowledge from me, their teacher – it becomes very straightforward to describe the potential outcomes framework with more formal language – by integrating additional readings, lecture materials, additional exercises, etc. 

The below quotes demonstrate how the course helped students think more critically about the use of various methodologies to advance research in the social sciences.

End of Course Student Reflections – Makerere University

Select quotes from end of course student reflections:

“Every moment of the session was stupendous since we learnt something new which changed our attitudes about research since our minds fueled new associations between different ideas and it was interactive in nature which made everything interesting.”

-          (Student #1)

“I recommend that this kind of research should be introduced at even the undergraduate level especially to the third year students since they also do research so that people can be able to appreciate research at an earlier level and build their cognitive skills earlier enough and also learn that research is not about data collection, entry and running models but to get the deeper understanding of what it is so that they  do not take research like something burdensome but as something they cannot live without.”

-          (Student #1)

“Research is not just about running regression models in a software, as usual be excited with having produced a bunch of output. Research is about curiosity of the world around us, gaining knowledge and understanding of a given phenomenon by critically and analytically thinking through it then you can identify suitable models to use considering the availability of data and the end user.”

-          (Student #2)

“I also discovered that we have the answers within ourselves just that sometimes we are not confident enough and at times we are lazy to use our brains. Research is real and should be fun, like something we do with passion not taking it as a punishment. It’s not about hurrying to finish but rather conceptualizing the concepts and coming out with best ideas and solutions that will help to practically solve the problem in the real world.”

-          (Student #3)

“The course has changed the way I look at science because it has made me become conscious about the world around me. I feel like everything around me is a research question whether big or small and can be inquired through these tools that are readily available through science.”

-          (Student #4)