Squirrels are very clever problem solvers. For example, give us a peanut butter smoothie in a glass too tall for us to dip into, and we’ll learn to drink through a straw. So I shouldn’t have been surprised when Nancy and Michelle, our twin five-month-olds, came into my office yesterday to ask if they could go play in the grass beneath Árbol Squirrel. They had already asked Mrs. Squirrel, and she said no as she was busy and could not go along to watch them. “Your mother already said no,” I replied.
“I know,” Nancy cheerfully chittered.
“So we’re asking you,” Michelle added.
I laughed and replied, “You sound like pollsters.”
It seems the Republican narratives about President Obama’s failure to stop rising gas prices were not making a dent in public polling. Most Americans seemed to know the president can’t do much to affect gas prices. With the economy growing enough jobs to absorb a quarter-million new and returning workers last month and Republicans paying a hefty political price for their War on Women, the GOP was fishing around for an economic argument. Gas prices seemed like a good cudgel. But voters weren’t buying that argument.
No problem. Just ask someone else!
Two new polls, one by the New York Times/CBS News and another by ABC News/Washington Post – show President Obama’s support plummeting in the past month. Both polls cite gas prices as the reason. Today the Washington Post offers the headline: Voters blame president for gas prices, experts say not so fast.
Why the sudden shift? Both polls changed their sampling. Deniac revealed the change in the ABC/WaPo crosstabs yesterday at The People’s View:
But what’s behind the headlines? What’s beyond the pundits and the pontifications? Why did President Obama’s approval numbers really go down? Or did it even actually go down? What really accounts for the reversal from last month? The real answer: they polled more Republicans this time than last. Ta-da! I decided to compare the party ID numbers attached to this poll vs. the ones attached to the last poll.
Sure enough, the ABC/WaPo pollster sampled 3% more Republicans and 4% fewer Democrats, for a net swing of 7% in favor of the GOP. Tweeter Jefferson Obama found the same shift in the CBS/NYT poll, which used the lowest sampling of Democrats in 20 years.
I disagree with Deaniac on one point:
It is important to note here that the analysis above is not meant to show that ABC News and Washington Post pollsters somehow “cooked” the numbers. These variations – 3 points here and 4 points there, especially given the poll’s margin of error of +/- 4 percentage points – are normal statistical anomalies.
To get a poll sample of 1000, pollsters don’t make 1000 calls, or even enough calls to get 1000 responses, and post crosstabs based on who they reached. Instead, pollsters set the crosstabs first – based on their predictions of the electorate – and make more than enough calls to for every demographic subgroup. The pollsters then randomly select responses to fill out their model. The people making calls did not get responses from fewer Democrats and more Republicans this month. The pollsters sampled responses from fewer Democrats and more Republicans this month.
Simply, they asked someone else.
The pollsters could argue the change was justified, as the Hotline‘s Josh Kraushaar did:
For comparison’s sake, Obama-friendly electorate in ’08 was D+7; 2010 midterm electorate was D+0. Expect ’12 to be somewhere in between.
But it’s dishonest to trumpet that February-to-March change in poll results, without mentioning the change in sampling model.
For the record, I knew Nancy and Michelle had already asked Mrs. Squirrel, and I gave them the same answer she had. They’ll have to come up with another clever trick. And being squirrels, I’m sure they will.
Good day and good nuts.