This week, CBS and The New York Times teamed up with Quinnipiac University to produce a poll of three key swing states: Florida, Ohio and Pennsylvania. The media is loudly trumpeting that the polls show Obama with sizable leads in each state, as well as showing him, for the first time, crossing the critical 50% threshold in the head-to-head match-up against Romney. They are deluding themselves. While the polls do show Obama with solid leads, he only gets there with a serious skewing of the poll sample. Another week, another flawed media poll, I guess.
The partisan sample of the polls were: FL D+9, OH D+8 and PA D+6. In 2008, the partisan split in Florida was D+3. In 2004, the split was R+4. Does anyone really believe that the Democrat vote in Florida this year is going to be 6 points larger than in 2008? In OH, today’s poll’s partisan split matches the 2008 result. In 2010, the state had an R+1 and in 2004 the state had an R+5 partisan split in the vote. The PA sample is largely in line with the vote in 2008, when the state had a D+7 split. Again, does anyone really expect Obama to match his 2008 turnout numbers?
Gallup has consistently shown a big drop-off in enthusiasm among Democrats this election, especially compared to 2008. Yet, the media continue to base their polling samples on a model that would mimic the 2008 election. Let me repeat…this isn’t going to happen.
There are a couple of other funky things in the poll that should raise questions about the headline results. After knowing the party affiliation of the respondent, they ask whether they consider themselves a “strong Democrat” or “Strong Republican.” Around 60% of GOP respondents consider themselves “strong Republicans”, but around 70% of Democrat respondents consider themselves “strong Democrats.” So, not only the partisan sample skewed toward Democrats, its skewed towards very active, base Democrat voters.
Also, when asked how they voted in 2008, respondents picked Obama by the following margins; OH +15, FL +13, and PA +14. Obama’s actual margin of victory in these states was; OH +5, FL +3 and PA +11.
Also, the poll was conducted over a 7 day period, more than double the normal window. That doesn’t necessarily skew the partisan sample, but it does potentially skew the overall results. Polls generally conduct their samples over a 2-3 day window, to ensure that respondents are generally reacting to the same information. 7 days is too long of a period to get an accurate “snapshot” of the electorate.
All this and Obama is still just at 50%?
I totally get that pollsters have to weight their samples to reflect their estimation of the electorate. But why is it that virtually every media poll this election weights their samples in a way that is beneficial for Democrats? Where is the poll that over-samples GOP voters? That every media poll errs in the same direction is what, in poker, we call a “tell”.
This poll, and many others like it, are simply propaganda. They are meant to reinforce the idea that Obama’s and the media’s attacks on Mitt Romney are working. They are meant to reverse the growing belief that Obama’s reelection is in serious jeopardy. But, the days of the media controlling the campaign narrative are over.
If the media is resorting to these tricks in July, imagine the lengths they will have to go in October? You’ve been warned.