Monday, February 8, 2016

Is gender driving the Clinton vs. Sanders generation gap? Nope.

The pundits have been largely fixated on the roles that gender and race are playing in the Democratic primaries, but the Matt Bruenig Election Team has had our eye on one key demographic divide from the beginning: age. Clinton's catastrophic 14% showing among young voters in Iowa proves that the generation divide is by far her greatest challenge - and one that she has, despite her best efforts, been completely unable to overcome. Here for example is how that gap compared to Iowa's gender gap:

Still, when we make this sort of comparison, we often hear the same question: how much of the age gap reflects variable gender populations among different age groups? Since there many more old women than old men, it's entirely possible that the disproportionate number of olds favoring Clinton could actually reflect something about gender driving support for the candidates rather than age.

Fortunately, a new Reuters polls breaks down support for the candidates by age and gender, allowing us to drill into this a bit more. By averaging male and female support for teach candidate in each age group, we can factor out the effect of differing gender populations; this number gives us the odds that a person of a given age will support a given candidate, regardless of their gender.

A few significant trends stand out:
  • The obvious split is between voters under and over 30. That's when Sanders' 30.6 point lead on Clinton reverses into a 20.8 point deficit.
  • If you are not in the lead, your numbers hover within the margin of error of around 27%. Sanders' support deteriorates as voters get older, but the difference isn't statistically significant.
  • Similarly, Clinton's lead fluctuates within the margin of error of around 47%; it's hard to read too much into those changes.
  • Sanders is significantly more polarizing than Clinton among different age groups, primarily because of his extraordinary support among younger voters.
Of course, these numbers may still be overdetermined by other considerations, particularly economic factors - but this should be enough to put to rest the theory that gender is what's driving the generation gap.