Saturday, August 15, 2009

Misinterpreting Obama's numbers

Charlie Cook declared in his column yesterday that Barack Obama would be 'unlikely' to win Virginia today.

False. This is where we lose perspective on what the President's falling approval numbers really mean. Ok so his national approval rating was in the 60s a couple months ago and now everyone (except Rasmussen) has it in the 50s. Guess what? If you have 13% more of the people with you than against you, as today's RCP average suggests, you get reelected in a landslide.

I've said it before but it bears repeating. Obama's numbers were artificially high for most of the year. Republicans who would probably never vote for him were giving him a chance and saying they approved of his job performance. Now they, and conservative leaning independents, are not and that's bringing Obama's numbers down. But is he actually bleeding the support of people who voted for him in November? Our numbers say no. Over the course of Virginia and North Carolina polls over the last two weeks we found a 95% correlation between how people voted for President last year and whether they now approve or disapprove of Obama. 3% didn't vote for Obama and now approved of him and 2% did vote for Obama but now disapproved of him. He still has pretty much all the support he had on election day!

Now Cook may be thinking Virginia is a special case, and we did put out a dreadful approval poll for him there last week. But that was with a 2009 electorate. Reweight it for the November 2008 electorate and you had Obama's approval at 50/44, basically identical to his 52-46 margin of victory in the state.

Barack Obama would definitely win Virginia if there was an election today. His numbers may be down from their lofty heights earlier in 2009 but unless you expected him to win 480 electoral votes in 2012, it's important to be careful about the extent to which his current poll numbers really suggests any trouble for his reelection prospects three years from now.


Anonymous said...

I would just add that once health care reform passes, I would expect his numbers to go back up to artifically high levels, because the people (and the media) all love a "winner".

Anonymous said...


You keep on assuming that your numbers last week represent only the "2009 electorate" when the 2008 electorate could have morphed into the 2009 electorate.

Tom, the percentage of Republicans in VA is basically unchanged. What is different is the percentage of dems and indies. The indies that became dems on NOvember 4th are now indies again due to dissatisfaction with Obama.

If your theory was so solid, why wasn't there a dramatic increase in the percentage of Republicans in your sample?

Anonymous said...

Tom, Tom,

You have discounted the possibility that more Obama voters are unwilling to admit that they voted for him.

We saw the same phenomenon when people were asked whether they voted for Kerry or Bush. There were always more Kerry voters than Bush voters in polls even when that clearly wasn't the case.

Face the music Tom. Judging by your preliminary numbers and the Vick question, it appears that you will join Scott Rasmussen and the Economist in showing Obama under 50% in terms of national approval.

Anonymous said...


Any thoughts on the Daily Kos/Research 2K weekly national poll? It seems to be a joke that they include non-voters in their sample.

How could Obama be at 60% favorables nationwide when his favorables in PA according to Daily Kos/Research 2K in its PA from this week was 55%? Its last CA poll, also conducted this week, had him at 63%.

PA is to 2-3 points to the left of the national average while CA is 8-10 points ot the left of the national average.

So how could Obama have a lower favorability number in PA than his national favorability according to Research 2K/Daily Kos?

Anonymous said...

Where were all these nay sayers about Obama, when George Bush JR.
ran our deficit up in the trillions?
The Rebublicans are having a hard time swollowing the fact that they lost.
The Poll in our Paper the "Baxter Bulletin" a pro-Republican paper, says the majority of Arkansans prefer Rush Limbaugh, to Obama, when in fact we think he is a hate monger, and a dangerous man.
On one of the back pages in the paper, they say Obama's economy policies are recovering.

Anonymous said...

You keep on assuming that your numbers last week represent only the "2009 electorate" when the 2008 electorate could have morphed into the 2009 electorate.

Ugh. Come on. You're displaying your ignorance of the basic fundamentals of off-year elections.

The "2009" electorate is about 60% of the size of the 2008 electorate. This is what we call an "off-year" election. There aren't as many voters in off-year elections as there are in Presidential elections. In fact, there are less voters in an off-year election (2009) than there will be in the midterm election (2010) with a contested Senate seat. Governor's races draw less attention than Senate races.

What's my point here? The point is that the "2009 electorate" looks bad because Republicans are energized about their gubernatorial candidate and Democrats are not. So that electorate contains more Republicans.

They're also energized by Obama's blackness. That seems to have fired up the GOP too!

Web Statistics