Poll news is no news
“New Poll Alert: Mike Michaud for Governor Leads Paul LePage”
WHO SAID IT
Headline on Mainedems.org, web site for the Maine Democratic Party, Sept. 29, 2014
This headline is only one of many examples during the recent campaign for Maine governor that relied on polling results to promote a candidate. This use — or abuse — of polling extends to partisan web sites and blogs, the news media and pundits, both partisan and otherwise. The poll cited in this Democratic blog was conducted by the University of New Hampshire Survey Center for the Portland Press Herald. It showed Democrat Mike Michaud with 40 percent, Republican incumbent Paul LePage with 38 percent and independent Eliot Cutler with 12 percent. The blog quoted party chairman Ben Grant: “The latest Press Herald poll confirms what we’re hearing on the ground. Mike Michaud is the only candidate with the overwhelming support from Maine Democrats, Republicans and independents needed to win this race and beat Paul LePage. Eliot Cutler’s support remains anemic and we fully expect more Maine voters to coalesce around Michaud as we get closer to November.”
When a Pan Atlantic SMS Group poll came out in October showing LePage and Michaud effectively tied, LePage supporters were quick to put their spin on it: “This poll makes clear what we already know, the naysayers who claim Governor LePage’s support would never grow were simply wrong, and more and more Maine people are going to reward Governor LePage for his hard work to create jobs, fix Maine’s budget, and reform welfare with their votes in November,” Maine GOP Executive Director Jason Savage told The Maine Wire, a conservative web site.
While similar headlines and quotes can be found throughout the campaign, the two cited are representative of how polls are used to manipulate the electoral process.
First, they are presented as telling voters who is winning.
But that’s not primarily what pollsters ask about.
The key question respondents are asked is who they would vote for “if the election were held today.” But, of course, it is not held the day the poll is taken. In the case of this poll, the poll was taken Sept. 18-25, about six weeks before Election Day. Such polls create useful data for candidates to know how well their campaign is going, but it’s ersatz news because nothing has happened except a few hundred people were asked a hypothetical question.
The second problem is that the standard deviation of the polls — known popularly as the margin of error — is often downplayed. Take the case of the UNH poll. The Democrats interpret the two percentage point lead over LePage as showing a Michaud lead, but the margin of error in the poll was 4.4 percent, meaning LePage could have been leading, not Michaud. The use of polls by partisans to mislead the public and the overreliance on polls by journalists to push manufactured news was underscored in the governor’s race by the fact that only one of the many polls was close to right. When the real votes were counted, LePage won with 48.2 percent to Michaud’s 43.3 percent and Cutler’s 8.4 percent. So all of the emphasis over who led in the polls was another case of shallow and meaningless “news.”
“Even though our poll is based on nothing more than a ‘what if’ question, we’re going to flog it so hard you’ll all think it means more than it does.”