impressive but not informative

From the economist, David Hakes:

Think back to your first years in graduate school. The most mathematically complex papers required a great deal of time and effort to read. The papers were written as if to a private club, and we felt proud when we successfully entered the club. Although I copied the style of these overly complex and often poorly written papers in my first few research attempts, I grew out of it quite quickly. I didn’t do so on my own. I was lucky to be surrounded by mature confident researchers at my first academic appointment. They taught me that if you are confident in your research you will write to include, not exclude. You will write to inform, not impress. It is with apologies to my research and writing mentors that I report the following events.

The preference falsification in which I engaged was to intentionally take a simple clear research paper and make it so complex and obscure that it successfully impressed referees. That is, I wrote a paper to impress rather than inform—a violation of my most closely held beliefs regarding the proper intent of research. I often suspected that many papers I read were intentionally complex
and obscure, and now I am part of the conspiracy.

I suspect that adding overly-complex math to a paper is specific to economics, but I also suspect that every social science has issues like this, whether it’s writing 20 more pages of text than you need or adding extra controls to your model that don’t really improve the specification. Generally I think the review system works great. My papers have improved over the course of the review process.  But I can also see alternative ways of writing papers that might lead to clearer arguments, more straightforward results, etc. Altering the way we write articles though requires cooperation from reviewers who prefer to hold on to the conventions we’d like to see changed. And, in a way, even if they aren’t very efficient in distributing information or developing theory, these conventions may be effective in reducing conflict about the “right way” to do research.