The Washington PostDemocracy Dies in Darkness

For Hillary Clinton, demographics aren’t quite destiny

Social video shows Democratic presidential candidate Bernie Sanders entering and leaving from a meeting with Rev. Al Sharpton at a restaurant in Harlem (Video: Sarah Parnass/The Washington Post)

With the Iowa and New Hampshire contests in the books, political junkies can finally satisfy our appetite for “hard data” on who actual voters support for their party’s presidential nominations — and what is motivating their decisions.

So what questions, if any, have the first contests answered about the Democratic race for nomination so far and how the rest of it will play out?

First, Sanders’s near-victory in Iowa and 22-point blowout New Hampshire proved his younger and independent-leaning base is motivated to show up for both low- and high-turnout contests, confirming the palpable enthusiasm seen at his college-town rallies was no mirage. His supporters are pumped, and that bodes very well for his candidacy.

Second, exit and entrance polling found age — not gender or ideology — is the dominant dividing line among Democratic voters, with younger voters backing Sanders far more than Clinton. In Iowa, Sanders’s support was 58 percentage points higher among voters under age 30 than among those ages 65 and older, and the age gap was 38 points in New Hampshire. In both states, it far exceeded differences between “very liberal” and “moderate” voters, men and women, and even whites and non-whites.

The lesson from the first two contests is that either candidate can reasonably claim they have the advantage going forward. Sanders can argue his ability to match the longtime front-runner in Iowa and defeat her in a state she won eight years ago indicates his support is on the rise and will overwhelm her — like Barack Obama in 2008. Clinton can argue Iowa and New Hampshire’s overwhelmingly white and liberal electorates are tailor-made to Sanders’s message. Clinton, the thinking goes, will dominate in states with larger African American and, to a lesser extent, Hispanic populations, among whom she receives greater support.

Neither candidate should feel too comfortable in those rationales, though, since it’s unclear how much Democrats in Iowa and New Hampshire differ from those in the remaining 48 states and three territories left to vote. There’s little doubt that Democrats in Iowa and New Hampshire are politically unique — both are far more white, and New Hampshire is far less religious and more politically liberal. It’s also a state that borders Sanders’s home of Vermont, adding a potential favorite next-door-neighbor element into the mix.

But demographic generalizations oversimplify efforts to project the first two states into the future. Put more plainly: The demographic splits we see in national polls didn't match what we saw in the only two states that have actually voted. Some of the biggest geographical differences are within demographic groups, with Clinton’s and Sanders’s support ranging widely between the early states and those beyond.

Take middle-aged Democrats between 40 and 64 years old. Clinton led among Democrats in this group by 34 percentage points in a January Post-ABC national poll, but that shrunk to a 19-point edge in Iowa and flipped to a 9-point Sanders lead in New Hampshire.

The contrast is just as sharp along racial lines; Sanders was roughly even with Clinton among non-white voters in New Hampshire (50 percent to 49 percent), but lost them by 24 points in Iowa and trailed by 45 percent in national polling.

The racial breakdowns in the first two states are based on small sample sizes — more than 90 percent of voters in each state were white — but they mark a massive warning sign for the Clinton campaign which has banked on black and Hispanic Democrats as a bulwark against Sanders’s popularity with young white voters. An average of 31 percent of Democratic voters were black or Hispanic across all states where Democratic exit polling was conducted in 2008, including 55 percent in South Carolina’s primary, which will be held Feb. 27.

Looking at the difference in national, Iowa and New Hampshire support within seven key demographic and political groups, Sanders’s New Hampshire support was on average eight points higher than in Iowa and 23 points higher than it was nationally.

The results suggest Iowa and New Hampshire might indeed be pro-Sanders outliers, but it would be a mistake to assume the remaining states tether closely to earlier national polls. The sequential nature of primaries mean voters get to listen to one another — i.e. Nevadans and South Carolinians adjust their opinions based on what they hear out of Iowa and New Hampshire or what they see when the candidates focus more intently on them. Sanders’s strong early performances may lead many Democrats to consider switching their support from Clinton.

There’s already some indication this is happening. A national Quinnipiac University poll conducted between Iowa last week and New Hampshire on Tuesday showed a sharp shift in support toward Sanders; after leading by 31 percentage points in December, Clinton’s edge shrunk to a statistically insignificant two percentage points. Clinton held barely any advantage among non-white Democrats, garnering 45 percent to Sanders’s 42.

No other national telephone poll has matched the Quinnipiac shift, but polls in upcoming primary states will provide new clues on whether Iowa and New Hampshire have shifted the race strongly in Sanders’s direction or not.