Sarah Palin’s misreading of polling data
“You’ve just abandoned the Reagan Democrats with this amnesty bill. It was the loss of working class voters in swing states that cost us the 2012 election, not the Hispanic vote.... You disrespect Hispanics with your assumption that they desire ignoring the rule of law.”
— Former Alaska governor Sarah Palin (R) in a Facebook post titled “Great Job, GOP Establishment, June 28, 2013
Given how many variables can affect a national election, it is sometimes foolhardy to suggest one group or another swung the election. For instance, President Obama and former Massachusetts governor Mitt Romney split 49-49 percent among heterosexuals, while Obama overwhelmingly (76-22 percent) won the votes of gays. But just focusing on that fact would ignore the efforts made by Obama to reach out to other groups to build his winning coalition.
Palin’s comment, posted in response to bipartisan passage of a comprehensive immigration bill in the Senate, struck us as interesting because it assumed Romney lost the so-called Reagan Democrats and that Hispanic voters do not favor a path to citizenship for illegal immigrants. What do the data show? (We were assisted by Scott Clement, analyst at The Washington Post’s Capital Insight, in finding some of these data.)
First of all, there have been a whole series of polls asking Hispanics about their views on overhauling immigration laws. Excluding advocacy polls with loaded questions, there is a consensus that Hispanics overwhelmingly support a path to citizenship. Here are two key samples:
Voter exit poll, 2012: 77 percent of Hispanics believe illegal immigrants should be offered a chance to apply for legal status.
WSJ-NBC poll, April 2013: 82 percent of Hispanics either strongly favor or somewhat favor a pathway to citizenship.
The 2012 exit poll of voters also found that Americans, by a margin of 65 to 29 percent, believe illegal immigrants should be offered a chance to apply for legal status. The WSJ-NBC poll found 64 percent of all Americans favor a pathway to citizenship.
Okay, so a core part of Palin’s assumption is just simply wrong. What about her notion that the loss of “working class voters” in swing states cost the GOP the election? That is a bit more fuzzy, in part because it depends on the definition of “working class” — Clement says there are a dozen different ways to define the term — but Palin’s reasoning is also debatable.
The most basic way to define a working-class voter is someone who lacks a college degree. In 2012, Obama won 51 percent of the national popular vote among non-college voters, 50 percent among college voters. The gap, in fact, has been fairly narrow over the last eight elections, with Democrats doing better among non-college voters than college graduates four times, Republicans four times.
But note that Palin referred to “Reagan Democrats.” This generally refers to white working-class Northern voters. Again, there are different ways to slice this. The National Journal has a handy interactive which allows you to examine the results of the elections between 1980 and 2008, slicing and dicing various categories.
You can see that Republicans already had an edge among white non-college voters — but not among white Democratic non-college voters. Even Reagan lost those voters by a margin of 68 to 27 percent, and the gap only increased in later elections. In fact, many of statistics showing how poorly Democrats do among white voters change substantially when Southern voters are removed; Democrats are competitive with Republicans in winning white voters, especially working-class voters, in New England, the Midwest and the West Coast.
Palin specifically mentioned the loss of Reagan Democrats in swing states. Let’s take a look at Ohio, the ultimate swing state, which Obama won 51 to 48 percent. Using an interactive that shows the difference between 2008 and 2012, you can see that Romney improved over John McCain’s share of white voters, by 5 percentage points, and among people with incomes of less than $30,000 a year; Obama slightly improved among voters with no college degree.
So why did Romney lose even after apparently improving his share of white working class voters? The reason is that white voters made up a smaller percentage of the overall vote in Ohio, falling four percentage points to 79 percent, while the African American vote increased four percentage points, to 15 percent.
A similar dynamic took place in Pennsylvania, where Obama experienced a six percentage point decline in the share of the white vote, but still won because Latinos and other voters of color increased their share of the vote — and went overwhelmingly for the president.
Another key swing state was Colorado, where again Obama lost ground among white voters but minority voters bolstered their share of the electorate and carried the president to victory. In Colorado, Obama’s share of the Hispanic vote was an astonishing 87 percent. (These figures come from an interesting analysis, “The Obama Coalition in the 2012 Election and Beyond,” by Ruy Teixeira and John Halpin.)
Overall, in fact, Obama lost ground among white working-class voters, garnering 36 percent in 2012, compared to 40 percent in 2008, according to a recent article in the New Republic by Teixeira and Andrew Levison. The article notes that even these results may have been distorted by unique events, giving Obama a greater share of such voters than might ordinarily be expected: “In 2008 the financial crisis and a desire to ‘clean house’ boosted white working class support for Obama. In 2012, the GOP fielded a uniquely aloof and unsympathetic Republican candidate.”
In a fascinating article titled “The Case of the Missing White Voters, Revisited,” for Real Clear Politics, Sean Trende concludes, based on population growth, that there were about 6 million white voters who failed to come to the polls. He identifies them as “H. Ross Perot voters” — largely downscale, Northern, rural whites. That may not be quite the same as “Reagan Democrats,” but close enough.
But Trende also concludes that these missing voters, by themselves, were not enough to cost Romney the election because he would have needed to win almost 90 percent of these votes; Trende estimated that Romney at best would have won 70 percent.
The Pinocchio Test
Even given the different ways one can slice poll data, we see little evidence for Palin’s assertions.
If anything, Obama did more poorly among white working-class voters in 2012 than in 2008, even in key swing states. Palin could be talking about the fact that not as many anticipated white voters showed up at the polls, but as Trende shows, even if those missing voters had cast ballots, it would not have made a difference in 2012 election result.
Meanwhile, her claim on Hispanic voters is not borne out by the data that show overwhelming support for policies that would provide a path to citizenship.
Check out our candidate Pinocchio Tracker