It’s not unusual for any news outlet to get one thing right and another thing wrong on a given day. But, in their Wednesday coverage of the tightening race in New Hampshire between Howard Dean and Wes Clark, the Associated Press managed to get the same thing right and wrong on the same day.

Mike Glover and Ron Fournier lead their January 14 AP story with the statement, “Howard Dean is airing a new anti-war ad in Iowa while plotting to stem Wesley Clark’s surge in New Hampshire, where the race has closed to within 10 percentage points and forced Dean to shift strategy.” They continue, “Locked in a close four-way fight in Iowa, site of the kickoff caucuses Jan. 19, Dean spent the day in Vermont with an eye toward New Hampshire’s Jan. 27 primary. Private polling by two campaigns in New Hampshire showed that Dean’s lead has shrunk to single digits from a first-of-the-year high of about 25 percentage points, officials said on condition of anonymity.” The casual reader would be left with the impression that the race had closed to within single digits.

In another January 14 story, however, AP reporter Tom Raum writes that “Private polling by two campaigns in New Hampshire showed that Dean’s lead has shrunk to single digits from a first-of-the-year high of about 25 percentage points, according to officials familiar with the polls. However, an independent poll, by the American Research Group, Inc., showed Dean with 32 percent support and Clark with 22 percent in the three-day period that ended Monday, far ahead of the rest of the field.”

Both the Glover/Fournier piece and the Raum piece cite private polling from two campaigns that disclosed a “single digit” gap between Dean and Clark. However, unlike Glover and Fournier, Raum accurately balances out his story with publicly available current polling data from the American Research Group.

However, this is not even the full story. Based on the margin of error in this poll (4%) the true percentages could actually be Dean 28%, Clark 26%, or Dean 36%, Clark 18%, since the margin of error applies to each candidate’s percentage. This poses the question - Why aren’t we seeing any headlines that read “Dean’s numbers doubles Clark?”

To be fair, the race between Dean and Clark is tightening. However, the gap began to close over a week ago, not just in the last few days. According to the American Research Group tracking poll, between the New Year and January 8th, Clark’s number moved up from 12 to 20 and Dean’s dropped from 39 to 32. Since then, the changes in the polling numbers for Dean and Clark have all been within the margin of error, and thus, not a statistically significant change at all.

—Thomas Lang

If you'd like to help CJR and win a chance at one of 10 free print subscriptions, take a brief survey for us here.

Thomas Lang was a writer at CJR Daily.