Things I Hate

Things I hate

  1. Constructions of the following form:

“Home (Foreign) produces and consumes two goods, with x being its natural import (export) good and y being its natural export (import) good.”

Why do I hate it? First of all because I find these sentences take me about twice as long to understand. But you may say why should my mental disability sway you away from this beguilingly compact way of expressing yourself? The second argument against it is that parentheses have a traditional use in written prose. They are to provide information that amplifies or qualifies the preceding text. The third point (which is closely related to the second point) is that a sentence should make sense when it is read out loud. Try reading the above sentence out loud. It emerges as nonsense.

So how should the sentence have been written? How about the following:

“Each country consumes two goods. The home country is the natural importer of x and the foreign country is the natural importer of y.”

  1. Precise (but wrong) and vague (but useless) reports of robustness.

Over and over I have heard presenters claim. “I did that and the results are exactly the same” or “identical”. Yet if we switch econometric method or the change the measurement of the RHS variables, the chances of literally identical results is very small. If the results are almost the same, you might say “I did that and all the reported coefficients are the same out to 2 decimals.”

I also dislike “qualitatively the same” since the reader does not know which qualitative features are maintained (signs? significance levels? of just the variables of interest or all included variables?). Our regression results are quantitative and trying to characterize two sets of results as qualitatively the same or similar just does not convey enough precision. Instead try to devise comparisons such as the following. “When estimated with 2SLS, the coefficients on X1 and X2 retain the same signs and stay within a standard error of the baseline results.”