Saturday, May 28, 2011

On the multiplication of weasel words

Economists and other social scientists love their weasel words (and phrases) such as "suggests", "may mean" and "seems to indicate". Sometimes these weasely words serve as a helpful shorthand for the fact that the estimates being discussed have standard errors, or that the a given empirical pattern may have multiple causes. Other times they just signal authorial timidity or a desire to avoid full responsibility for one's conclusions.

Reading a gradual student paper this morning, I was reminded that regardless of the reason for including weasel words, one of my rules of writing is that a given sentence should have at most one weasel word or phrase. Consider the sentence
This suggests that the longitudinal weights may underweight non-participants.
To accomplished the desired level of weaseliness, the writer (who is actually quite a clear and organized writer) does not need both "suggests" and "may". The sentence
This suggests that the longitudinal weights underweight non-participants.
does just fine at conveying the intended meaning, as does the sentence
My findings indicate that the longitudinal weights may underweight non-participants.
The bottom line: one weasel word or phrase per sentence.

As an aside, in the context of the paper from which it is drawn, the sentence actually does not require any weasel words at all, as the results shown in the paper clearly indicate underweighting of non-participants. But that, perhaps, is a subject for a different post.

No comments:

Post a Comment