We reached a big milestone Friday as my students read their first primary research articles and we discussed them during our live session. Coming into the course, most of my students knew that correlation doesn’t imply causation. For example, just because cities with more police tend to have higher crime rates doesn’t imply that hiring more police induces people to commit more crime. In a sense, econometrics is just a set of statistical methods for going beyond correlation and estimating actual causal effects.
The medical literature is a great place to start because the methods tend to be more straight-forward, the papers are written for a less technical audience, and the papers are just a few pages long. The two articles we read Friday were:
- Balcells, Eva et al (2011) “Soft drinks consumption, diet quality and BMI in a Mediterranean population,” Public Health Nutrition
- Messerli, Franz H. (2012) “Chocolate Consumption, Cognitive Function and Nobel Laureates,” New England Journal of Medicine
When I’m teaching students how to read primary research, I walk through the articles in class in a very particular way:
- Identify the main research question. Why do we care?
- Identify the data. Is it representative of an interesting population?
- Identify the methods being used. Do we know all of them? We should know some, considering I assigned the paper! It’s also important to learn how to read around things you can’t yet understand.
- Show the descriptive statistics. Focus on the correlation between the dependent variable and the independent variable of interest.
- Discuss what might be driving this correlation. Are there potential confounding variables? Could there actually be a causal link in either or both directions? Could it be completely spurious?
- Look at the regression results. How should we interpret the coefficient estimates? Did they control for what we wanted them too?
- Wrap up. At the end of the day, what do we learn from this paper?
My students are no longer slaves to reporters at the New York Times–They can now read (some) of the primary research upon which newspaper articles are based and form their own opinions about it. Hopefully they will continue to practice and improve the skills they acquire here after the class is over, but even if they don’t, it’s remarkable how far we’ve come in just four weeks.