≡ Menu

The Plural of Anecdote is Not Data

My open letter to Lou Dobbs brought me lots of e-mail — alas, almost all of it hostile.  Here (with the writer’s name withheld, of course) is the full text of one of these e-mails; its message is typical:

If you lived in Michigan, where factory after factory has a
for lease or for sale sign in front of it, you might find Mr. Dobbs a bit more

While I understand the appeal of anecdotes — and while I’m not among those economists who dismiss anecdotes as altogether irrelevant — it is nevertheless important to distinguis anecdotes from data.  It pains my brain to read so many people who insist that their anecdotes establish a sound case against the empirical record built from data collected and interpreted according to well-established theories.

I don’t know the person who wrote the above e-mail; I don’t know any of the persons who wrote to assure me that my imbecility was on parade in my open letter to Mr. Dobbs.  But I wonder how many of these people would accept the following line of argument:

Researcher Jones, like most other researchers in the field, reports that a majority of Americans are fluent only in English.  But I dispute this claim, for my wife — who is American — is fluent also in French.  Therefore, I reject the finding that most Americans are mono-lingual.

I suspect that few, if any, of my correspondents would find the above rejection of the claim that most Americans are mono-lingual to be compelling.  Yet these same people proudly make similar arguments about the consequences of trade — and, because they e-mail these arguments to me, obviously believe that these arguments have weight and merit.


Next post:

Previous post: