Tuesday, October 09, 2012

The Polling Games!


Have intensified.  I came back from a political break to the news that all is lost and we better flee, because the Pew poll for October shows Romney running away with the presidency.  And what's even more horrendous, it's the women's opinions which did that.  So.

What's the next stage in the divine self-defense against angst-and-emotion overflow?

To get hold of that poll and to look at it, naturally.  But first I took a short detour to Nate Silver's blog.  Here's what he says:

But were it not for the Pew poll, our forecast would have been unchanged from Monday, with Mr. Romney’s chances holding at 21.6 percent.
The Pew poll, however, may well be the single best polling result that Mr. Romney has seen all year. It comes from a strong polling firm, and had a reasonably large sample size. Just as important is the trendline. Pew’s polls have been Democratic-leaning relative to the consensus this year; its last poll, for instance, had Mr. Obama 8 points ahead among likely voters. So this represents a very sharp reversal.
...
There are two smarter questions to ask about the Pew poll. First, is it really likely that Mr. Romney leads the race by 4 points right now? The consensus of the evidence, particularly the national tracking polls, would suggest otherwise. Instead, the forecast model’s conclusion is that the whole of the data is still consistent with a very narrow lead for Mr. Obama, albeit one that is considerably diminished since Denver.
It might be granted that the situation is more ambiguous than usual right now. But our forecast model looks at literally all of the polls; it estimates Mr. Romney’s post-debate bounce as being 2.5 percentage points, not quite enough to erase Mr. Obama’s pre-debate advantage.
The other valid line of inquiry concerns the timing of the poll. The Pew poll was conducted from Thursday through Sunday, although more of the interviews were conducted in the earlier part of that period. There’s nothing in the poll that really refutes the story that Mr. Romney initially received a very large bounce after the debate (perhaps somewhere on the order of 4 or 5 points, if not quite as large as Pew shows it), which has since faded some between the news cycle turning over and the favorable jobs report on Friday.
The evidence that Mr. Romney’s bounce is receding some is only modestly strong — as opposed to the evidence that he got a significant bounce in the first place, which is very strong. Still, the order in which polls are published does not exactly match the order in which they were actually conducted — and at turning points in the race, these details can matter.

I also noticed that the Daily Kos/SEIU State of the Nation Poll found a somewhat similar shift towards Romney after the first presidential debate:

The candidates for President are Democrat Barack Obama and Republican Mitt Romney. If the election was today, who would you vote for? Obama 47 (49)
Romney
49 (45)
That's a pretty disastrous six-point net swing in just a week, and the first time we've ever had Romney in the lead. It is inline with all other national polling showing Romney making gains in the wake of his debate performance last week. Both the Gallup and Rasmussen trackers saw their Romney bounce evaporate on Sunday. In this poll, 75 percent of the sample was gathered on Thursday and Friday, at the height of Romney's bounce. This is because PPP does call-backs: It identifies a random range of numbers and begins calling them on Thursday. If they get no answer, they keep trying the same numbers on subsequent days until they get the required number of responses (we ask for at least 1,000). This avoids the old tropes about young liberals being out partying on Friday nights, while conservatives are at church on Sunday mornings, etc.
So this week, 47 percent of responses were on Thursday, 28 percent on Friday, 17 percent on Saturday, and just 8 percent on Sunday. Romney won Thursday 49-48 and Friday 49-44 before losing steam over the weekend. While Romney won Thursday and Friday by a combined 2.5 points, he won Saturday and Sunday by just 0.5 percent.
So where did Romney gain? Among women, Obama went from a 15-point lead to a slimmer 51-45 edge. Meanwhile, Romney went from winning independents 44-41 to winning them 48-42. And just like the Ipsos poll showed last week, Romney further consolidated his base. They went from supporting him 85-13 last week, to 87-11 this week while Obama lost some Democrats, going from 88-9 last week, to 87-11 this week.

In both cases the shift seems to be due to women's opinions, and in both cases the results appear to capture the immediate post-debate reactions.

The Pew poll results are available here (pdf).  I don't see anything major wrong with it, though I wonder how it got such a low number of Hispanic registered or likely voters and young likely voters.

The crucial question asks whom the respondent would vote for if the election was held on the date the poll questions were asked.  That date was sometime last week, from Thursday to Saturday.  The answers to that question:

Forty-four percent of registered male voters stated that they would vote for Obama, 47% for Romney.  Forty-nine percent of registered female voters stated that they would vote for Obama, 45% for Romney.  The percentages for likely voters benefited Romney more:  Among likely male voters 51% chose Romney and 43% Obama, while among likely female voters the percentages for Romney and Obama were both 47%, or a tie.

This means that women are still more likely to vote for Obama than men are.  On the other hand, among the white non-Hispanic (what happened to the Hispanic voters, again?)  respondents in this poll the usual gender gap has essentially disappeared.  Both men and women  would have picked Romney over Obama at roughly the same percentages.

So what's going on here?  I'm not on expert in polling, and probably should not have written about this at all.  But I'd wait and see what happens in the next week or so in all the different polls.  This is because the two I mentioned here were both taken right after the first presidential debate and could be measuring a debate bounce.  The only way to be sure is to see whether Romney's apparent advantage in these two polls turns up in other, later polls.

It's true, naturally, that  the first presidential debate completely protected Romney when it comes to his views on reproductive freedom and whether women should be treated fairly in the labor force and so on.  Obama didn't mention women, either, and not a single question was about this giant wedge issue between the parties (sorta like "are women partly human or not at all human").  Neither did anyone refer to Romney's comments about the 47% being leeches and parasites.

People not following politics that eagerly might have taken the message from the debate that nothing else distinguishes the two candidates except their policies on taxes, jobs and the deficit.  The Pew poll respondents, on average,  thought that Romney did better on all those three questions.  That he also lied a lot more might not matter to the respondents.  Or they may not even know about those lies.

What's the morale in this story?  I'm not sure but Obama should certainly start practicing his debate skills and perhaps even remember that his re-election is crucially dependent on women's votes.

Finally, the Pew results really hinge on the percentage of respondents who self-identified as Republicans: The September Pew poll had 39% Democrats, 29% Republicans and 30% Independents.  While the percentage of Independents stayed the same in the October poll, the percentage of Democrats dropped to 31% and the percentage of Republicans rose to 36%.

This could be a real change in how voters self-identify.  On the other hand, it's always possible that any particular sample fails to be representative, simply because of bad sampling luck.  That jump looks pretty huge for just one single month, though, but who knows.