Archive for August 8th, 2005

The Matthew Effect

Monday, August 8th, 2005

A little while back ("Let's Play Internet!") I wrote about trying to track down an authoritative attribution for this quotation:

The plural of "anecdote" is not "data".

It was fun but inconclusive, barely penetrating the surface of the murky waters of misattribution. It's a game I enjoy playing, although I only get to play in little installment (unless someone is willing to pay me to do it!).

At any rate, because of that piece, I now have a steady stream of visitors to my blog who google for the phrase and end up here, which is a nice thing, but it's made me think that I should push on and see whether we can penetrate to the next level in uncovering the correct attribution for the phrase.

A casual start with Google seemed to indicate two conflicting attributions, of which these were typical:

Glancing through the results, it looked like Brinner was leading Kotsonis about 3 to 1, so I asked the Google for these keywords and got the indicated number of hits:

  • "anecdote plural kotsonis": 61
  • "anecdote plural brinner": 311

Overall, more people appeared to attribute the phrase to Mr. Brinner, although more careful checking showed that not all of those 311 results were independent of each other.

But here's a surprise during that last search: Google asked (in it's irritating fashion): "Did you mean: anecdote plural brenner?" So, what if I did? Those keywords resulted in 266 hits. However, that path ("anecdote plural brenner") quickly seemed to degenerate. Many of the hits simply happened to have those three words without any reference to the phrase in question. It also turned up more attributions however, further muddying the waters:

The last one is interesting because we'll see the name "George Stigler" again; otherwise, this route didn't seem to be giving results that were going to converge on any useful information.

I thought maybe I should compare with what the MSN-search oracle might give. It served up about 250 results for "the plural of anecdote is not data" (as a phrase, i.e., with quotation marks in the search string — by comparison, Google turned up 4,900 hits for the phrase, but the first 100 didn't add much to what we'd already seen.1) Most just added their weight to the assertion that it was first uttered by Brinner, with a few supporting the Kotsonis assertion.

There were also some interesting outliers:

These are all evidently spurious, one-off attributions; it might be interesting to know how they came to be, but I don't want to get sidetracked.2 In addition to "some wag", I've seen "some sociologists", "a learned professor", and "anonymous" for attributions. I find these untidy and lazy.

One blog entry, written by yet another person (Sam, at liberaldestert) trying to track down the origins of the phrase, offered this very authoritative sounding bit of information:

Lee Bolin (Tempe, AZ) emailed his comments about the origin of the phrase I asked about in a prior posting (Blogger link may not work; it's the posting for 11/18) that "the plural of 'anecdote' isn't 'facts'":

"I believe that the original aphorism is "The plural of anecdote is data." The negation, that the plural is NOT data, seems to be a recent reaction by academics to the original phrase. I do recall that I read the original quote from Senator Moynihan in Time, Newsweek, or U.S. News over a decade ago, but I do not now know the specific citation. I also do not if it was Senator Moynihan's original thought, or if he had borrowed it from someone else.

A variety of other people, notably Ben Wattenberg, have used that phrase over the years. Senator Moynihan's own use of it appears from time to time in the Congressional Record. The negation, "The plural of anecdote is not data", seems to have arisen fairly recently and is popular with persnickety social scientists.

The date on this entry is Tuesday, November 26, 2002, important in the context of the next quoted piece. From the evidence I've been through today, I'm ready to disagree with e-mail author Lee Bolin's two assertions: that the original phrase did not have "not" in it (I think it did), and that the phrase has anything at all to do with Senator Moynihan (it appears to owe nothing to him) except that he may have repeated it on some occasion. I fear that these may be fabrications, albeit well-intentioned, on Mr. Bolin's part.

Going back to Google, then, and asking about "anecdote plural stigler" gave 161 results, not as many as Mr. Brinner had, but number of hits can be a dangerous metric. The results for Stigler seemed to have a good variety of sources, which is suggestive that they were independent indications, and with dates that went back at least as early as 1995:

Most fundamentally, economists are mostly unmoved by industrial policy claims because, as George Stigler has quipped, “the plural of anecdote isn’t data.”
[Richard Beason and David Weinstein, "The MITI Myth", The American Enterprise Online, July/August 1995.]

Note that the "not" was part of the "quip" at this early date.

That search also turned up this very provocative nugget:

From: Difficult library reference questions list
Sent: Sunday, June 13, 2004 2:08 PM
To: STUMPERS-L@LISTSERV.DOM.EDU
Subject: ? Quotation. "Plural of anecdote is data."

We have been asked who first said or wrote: "The plural of anecdote is (not) data." The quotation is found both with and without the "not." We have searched standard quotation books (Bartlett, Oxford, Penguin, etc.) Lots of examples on the Web, the earliest being one from the economist George Stigler in 1991, but nothing to indicate he originated the phrase. Also one site which attributes it to Daniel P. Moynihan, but with no evidence. Any assistance in pinning down, if possible,the origin of the phrase will be appreciated. Sincerely

Stan Shiebert
Librarian
Arts, Recreation & Literature Department
Seattle Public Library

Don't you just love librarians! (And isn't it nice to know that there is a mailing-list for reference-librarian "stumpers"!) I'll have to write dozens of additional posts about libraries and my love for them and how I generally think of librarians as gods and goddesses, but anyway….

One wonders whether this reference to "one site which attributes it to Daniel P. Moynihan" is the site we'd just visited above, with the e-mail from Mr. Bolin.

This looks like a good place to pause in tracing the origins of the phrase, since we have an authoritative, trustworthy voice taking us back to 1991 with George Stigler as the originator. But one further (later) example might be amusing:

To paraphrase Nobel Prize-winning economist George Stigler, data are the plural of anecdote.
[Nathan S. Balke andMine K. Yucel, "Evaluating the Eleventh District's Beige Book - Brief Article", Economic & Financial Review, Oct, 2000.]

This one brings out two important points:

  • George Stigler is a Nobel Laureate, so he tends to win in the attribution game because of the "Matthew Effect"3
  • As I pointed out in my first piece on this subject, maintaining the quotation marks (presuming that they were original, which I suspect they were since they make the aphorism much, much wittier — if they weren't they should have been!) is vital for the meaning; our authors should have written: "data" is the plural of "anecdote" (which would have been funnier, although economists are not generally known for their sense of humor).

The unanswered question that comes to mind is: why did they feel the need to paraphrase in the first place?
———-
1Although there were several entries that seemed to be undertaking a very serious, deconstructionist sort of analysis of the phrase, which strikes me as rather silly, since it was a witticism to start with, not a dissertation in poetic form.

2The most likely explanation from the snippets that I read was that the person writing had received a communication from another person who used the phrase without attribution, so the writer atributed it to the correspondant.

3In quite a different context, I came across a short piece by physicist N. David Mermin, "Could Feynman Have Said This?" (Physics Today 57, no. 5, 2004 — subscription may be required), in which he was trying to track down first-hand evidence that Richard Feynman had actually said something frequently attributed to him.

Anyway, Mermin described the "Matthew effect":

The Matthew effect was enunciated by the great sociologist of science, Robert Merton [R. K. Merton, Science 159, 56 (1968)]. Merton worked in those innocent days when sociologists were interested only in the behavior of scientists and not in the content of their science. (To be fair to contemporary sociologists of science, I should modify that last phrase to "and not in the manifestations of that behavior in the content of their science.") I first learned of the Matthew effect more than 20 years ago, on the occasion of my first and, perhaps until now, only, victimization at the hands of the New York Times.

I learned the name for what the Times had done to me when I received a very nice note from P. W. Anderson in which he expressed his regret that the newspaper had given him exclusive credit for a nomenclatural advance that was entirely due to me. "A depressingly typical example of the Matthew effect" was how he characterized the misattribution. (I reported the entire history of this contretemps in these pages back in those dark ages [April 1981, page 46] before there were Reference Frame columns.) When I wrote back asking him what the Matthew effect was, he referred me to Merton.

It was Merton who identified and named the tendency always to assign exclusive scientific credit to the most eminent among all the plausible candidates. At least I hope it was he, though I'm sure Merton, who invented many wonderful jokes himself, would have been delighted if the credit for it turned out to be misattributed to him. Merton named the effect after the Gospel According to Matthew, because there it is written,

For unto every one that hath shall be given, and he shall have abundance: but from him that hath not shall be taken away even that which he hath.
—Mattthew 25:29.

Posted in All, Such Language! | 6 Comments »