A Troubling Pattern at PPH

Portland Press Herald's coverage of the Allen-Collins race fits a troubling pattern: Poll numbers that show Collins leading by wide margins--even when those results are out of date or from partisan sources--are played up; and timely results that suggest a closer race are ignored.

The facts:

There have been eight public polls on the Allen-Collins race this year.

Three came from Republican firms; four are from Rasmussen, a national polling outfit; and one was conducted by Pan Atlantic, a local company.

Results have ranged from a 25 point advantage for Collins all the way down to a seven point spread (twice).

Here's a chronology of PPH's coverage of the various polls.

--April 2: The anti-union Coalition for a Democratic Workplace releases a poll conducted several weeks earlier by a Republican pollster showing a 23 point lead for Sen. Collins. At PPH's political blog Political Extra, reporter Jonathan Kaplan publishes a post detailing the results.

--April 8: Rasmussen releases a poll taken a week earlier showing a safe 16 point advantage for Collins. Kaplan posts a summary of the results at Political Extra.

--May 16: Rasmussen releases a poll conducted two days earlier that shows a relatively narrow ten point spread between Sen. Collins and Rep. Allen. This is the narrowest advantage yet for Collins. PPH makes no mention of the poll in its newspaper or on its website.

--June 18: Rasmussen releases a poll conducted two days prior showing a narrow seven point margin in Collins' favor. PPH makes no mention of the poll in its newspaper or on its website.

--June 24: On its website, PPH publishes the results of a six week old poll sponsored by the GOP-aligned US Chamber of Commerce showing a 14 point lead for Collins.

The poll, released six days earlier, was conducted by a Republican polling firm. It has a larger margin of error and a smaller sample size than the Rasmussen poll released on the same day.

--July 8: Pan Atlantic releases a poll showing Collins ahead by a 25 point margin. This is the largest margin of any poll all year, including polls by Republican pollsters. PPH publishes results on its website as a "news update" reported by Dieter Bradbury.

--July 9: Despite the fact that its data is a full month old, PPH features a story on the Pan Atlantic poll on page A1 in print editions. The 727 word piece runs under the headline, "Poll finds solid lead for Collins in Maine."

It isn't until the seventh paragraph that the piece acknowledges that the poll is a month old. It isn't until the twelfth paragraph that the article contrasts the 25 point Pan Atlantic spread with the seven point margin in the Rasmussen poll from the identical time frame.

--July 18: Rasmussen releases a two day old poll which duplicates the seven point lead for Collins from its June survey. PPH doesn't run a "news update" on its website; it doesn't mention the poll in the next day's paper.

So what exactly is going on here? Why has PPH repeatedly featured partisan surveys and stale data while completely disregarding current, non-partisan polling?

The Rasmussen result from April showing a 16 point spread clearly rose to the level of "news." So why did the company's results mysteriously slip off of PPH's radar when the margin shrunk in later polls?

And how does PPH justify putting an outlier of a poll--with month-old data--on its front page when for nearly two months it failed to even mention equally salient, far more timely results?

As of this writing, PPH Managing Editor Bob Crider had not responded to an early Saturday e-mail requesting comment.