Pundit Boxscore for Thursday 22 January 2004

THE O'NEILL AFFAIR AS A SCANDAL CASE STUDY: The Paul O'Neill controversy provides an interesting case study because both the accuser and the accused are Republicans. This means that whichever side pundits take, they will find it difficult to avoid simultaneous praise and criticism of Republicans. Reader Dan Schaeffer questions whether the score of such columns are actually indicative of partisanship:

Linda Chavez's column gets a blue 16 rating due primarily to several (13) negative references to Paul O'Neill. O'Neill is a Republican, so I suppose those negative references are literally R- references... but given O'Neill's recent apostasy, it seems to me that making negative references to him -- essentially, attacking him for falling away from the party line -- would be more indicative of Republican partisanship, not less.

Mike C. asks a similar question and offers an example of a maverick on the Democratic side:

I'd just like to point out, that in your rankings of the pundits commenting on Paul O'Neill, there is a bias towards being counted as "pro-Democratic" when that is not the intent. Since O'Neill-bashing by the White House apologists is ranked as "R-" it results in them having a much lower "partisan" score. (Imagine if you will a column trashing Sen Zell Miller for breaking with the Democratic Party ---would that REALLY be an "anti-Democratic" critical column?

First of all, I think it's certainly true that individual references and even entire columns will sometimes have partisanship scores which do not correctly reflect the intent of the columnist. In 2002 Claudia Rosett of the WSJ OpinionJournal wrote an extremely negative column about Hillary Clinton which "beat the system" only because of its unusual form. As long as these kinds of departures occur infrequently, they shouldn't have a strong effect on a pundit's score over dozens or hundreds of columns. In other words, partisanship scores should become more representative as the sample size becomes larger. I've always said that at least one year of columns (usually 50 to 100) is needed before trying to draw any serious conclusions.

Second, notice that most of the O'Neill columns shown in yesterday's post have low partisanship scores whether the columnist was defending the Bush administration (Daniel Henninger, Linda Chavez, Cal Thomas, and most interestingly, Michael Kinsley) or attacking it (David Broder). And yet some pundits found a way to achieve a higher score when writing about the same issue. I don't think it's coincidental that the columns with the highest scores were written by two of last year's titans of partisanship, Paul Krugman and Molly Ivins.

Their two columns illustrate how Mr. Krugman and Ms. Ivins stay ahead of the Lying in Ponds pack -- by writing about nearly every issue in a more one-sided way than their colleagues. The other columnists grappled with the O'Neill story directly, making lots of references to O'Neill himself (Broder 13, Kinsley 18, Thomas 13, Chavez 13, Henninger 16). But there are many fewer references by Mr. Krugman (4) and Ms. Ivins (3), because they spend very little time praising Paul O'Neill despite having complete faith in the accuracy of his revelations. Instead, they merely use the revelations as a springboard for a familiar litany of attacks on the Bush administration. Extremely partisan columnists are able to use nearly any set of facts as a starting point to arrive at seemingly unrelated favorite arguments such as unemployment (Krugman) or criticism of Richard Perle (Ivins). Although none of our current pundits have done so, a partisan Republican could easily have written a column defending the administration from Mr. O'Neill's claims without taking much time to attack Mr. O'Neill himself, and their partisanship score could have been just as high.