In the days leading up to Democratic gubernatorial candidate Bill Ritter's landslide election victory, three news outlets -- the Fort Collins Coloradoan, the Associated Press, and KDVR Fox 31 -- each echoed Republican candidate Bob Beauprez's assertion that recent polling showed he was “closing the gap” on Ritter. As Colorado Media Matters noted (here, here, and here), the reports cited no polls to support their claims. These reports also appeared to ignore the most recent polls, some of which showed Ritter's lead expanding within those polls' margins of error. Moreover, the most recent version of all but one publicly released poll showed Ritter leading by double digits in the race's final days.

These media outlets' insistence on reporting -- despite substantial evidence to the contrary -- that Beauprez might be “gaining some ground,” “closing in,” or “narrow[ing]” Ritter's lead followed Beauprez's own statement that he was “closing the gap,” as an October 31 Rocky Mountain News article quoted him as saying. The News article noted that Beauprez was “urging supporters to ignore polls showing him trailing Bill Ritter by double digits” and said Beauprez “cit[ed] in part an online Wall Street Journal-Zogby poll showing him within 2 percentage points of Ritter.” However, as the News reported, "[s]ome pollsters ... doubt the accuracy of online polling, fearing it can be skewed by regular survey participants and a void of seniors, low-income voters and others who don't use the Internet." A subsequent online Zogby poll, conducted October 23-27, showed Beauprez leading by 0.2 percentage points and had a margin of error of 4 percentage points.

At the time, other polls showed Ritter leading by large margins. While two of these polls did show Beauprez gaining ground within the margin of error* in comparison to previous versions of those polls, a third poll showed Ritter's lead growing. All three of these polls showed Ritter leading by double digits.

• A Rasmussen Reports poll, conducted October 22, showed Ritter with a 12-percentage-point lead and had a 4.5-percentage-point margin of error. A previous Rasmussen Reports poll, conducted September 19, showed Ritter with a 16-percentage-point lead and also had a 4.5-percentage-point margin of error.
• A Denver Post/Mason-Dixon poll, conducted October 24-26, showed Ritter leading by 12 percentage points with a 4-percentage-point margin of error. A previous Denver Post/Mason-Dixon poll, conducted October 3-5, showed Ritter leading by 15 percentage points and had a 4-percentage-point margin of error.
• A Ciruli Associates poll, released October 27, showed Ritter with a 19-percentage-point lead and had a margin of error of 4.4 percentage points. A previous Ciruli poll, conducted September 26-October 2, showed Ritter with a 10-percentage-point lead and had a 4.4-percentage-point margin of error.

Subsequent polling showed Ritter with an even larger lead. On November 2, 9News and SurveyUSA released the results of a poll conducted October 30-November 1. The poll showed Ritter with a 22-percentage-point lead and had a 3.8-percentage-point margin of error. A previous 9News/SurveyUSA poll, conducted October 20-22, showed Ritter with an 18-percentage-point lead and had a 4.1-percentage-point margin of error.

But one day after the 9News/SurveyUSA poll showed Ritter leading by 22 percentage points, the Coloradoan reported in a November 3 article that “recent polls suggest” that Beauprez “could be” closing in on Ritter. The Coloradoan provided no evidence for its assertion that Beauprez “could be” closing in, and in fact noted that “most publicly available polls” showed “Beauprez behind” Ritter.

From the November 3 Coloradoan article:

[U.S. Rep. Marilyn] Musgrave [R], a two-term incumbent from Fort Morgan, is facing state Rep. Angie Paccione, a Fort Collins Democrat, and Eric Eidsness, a Fort Collins Reform Party candidate, in the 4th Congressional District. Paccione and Democratic gubernatorial candidate Bill Ritter, who stumped at Colorado State University Wednesday, spent Thursday campaigning together in Brush, Fort Morgan and Sterling.

With Election Day four days away, most publicly available polls have Musgrave leading her race and Beauprez behind in his, though recent polls suggest Paccione could be closing in on Musgrave and Beauprez could be doing the same to Ritter.

The same day that the Coloradoan published its article, the News published the results of a Public Opinion Strategies poll. Like the 9News/SurveyUSA poll released the day before, the Public Opinion Strategies poll showed Ritter leading by 22 percentage points. It was conducted October 28-30 and had a margin of error of 4.38 percentage points. A previous Rocky Mountain News/CBS4/Public Opinion Strategies poll, conducted September 10-12, showed Ritter with a 17-percentage-point lead and had a margin of error of 4.38 percentage points. Describing the polling results, the News reported November 3, “With time running out, Republican Bob Beauprez has failed to gain any ground on Democrat Bill Ritter in the Colorado governor's race.”

Nonetheless, only one day later, a November 4 Associated Press article by Steven K. Paulson reported that "[r]ecent polls have shown" Beauprez “gaining some ground against Democrat Bill Ritter.” The AP cited no polling results to support this claim and made no mention of the polls directly contradicting it.

From the November 4 AP article:

Tom Petersen, a 43-year-old computer consultant from Severance, said he's worried about Musgrave and Beauprez, but thinks they will still win on Tuesday. He said Bush's visit is a big boost at a time Republicans need it.

Recent polls have shown both races getting closer, with Democrat Angie Paccione closing in on Musgrave in the race for the 4th Congressional District and Beauprez gaining some ground against Democrat Bill Ritter in the race to replace term-limited GOP Gov. Bill Owens.

The final public poll on the gubernatorial race was conducted via the Internet October 27-November 5 by the Polimetrix polling organization and released on November 6. That poll showed Ritter leading Beauprez by 20 percentage points. The poll did not report a margin of error. Polimetrix did not previously poll this race, so it had no previous poll for comparison.

One day later, KDVR Fox 31 reporter Sean Tobin claimed on the November 7 Election Day broadcast of Good Day Colorado at 6 a.m. that while “most polls” showed Ritter in the lead, Beauprez “has been able to narrow that lead in most of those polls, so this could go right down to the wire.” Like the Coloradoan and the AP, Tobin cited no polls to support this claim and made no mention of the polls that directly contradicted it.

According to the News' election results web page, Ritter was leading Beauprez by more than 15 percentage points on November 9, with 98 percent of precincts having reported.

* According to former American Statistical Association president Fritz Scheuren, “a rule of thumb is to multiply the currently reported margin of error by 1.7 to obtain a more accurate estimate of the margin of error for the lead of one candidate over another. Thus, a reported 3 percent margin of error becomes about 5 percent and a reported 4 percent margin of error becomes about 7 percent when the size of the lead is being considered.”