November 25, 2022

For many Republicans, the final midterm election results were surprising— especially in the states of Arizona, Nevada, and Pennsylvania. Based on polling, it looked likely that Kari Lake, Adam Laxalt, and Mehmet Oz would win their contests. I believe that there are only three likely explanations for the unexpected results: late campaign surges, polling error, or aggressive ballot-harvesting.

‘); googletag.cmd.push(function () { googletag.display(‘div-gpt-ad-1609268089992-0’); }); }

A late surge?

Some Republicans are falling for an illogical narrative put forth by Democrats and the mainstream media. It goes like this: Although Republican candidates were higher in the polls, at the last minute people voted for Democrats because:

  • they feared the end of democracy
  • they hated Donald Trump
  • they were angered by the Dobbs abortion decision
  • they saw more anti-Republican campaign ads, due to financing differences
  • they were turned off by fighting during the primaries
  • they weren’t impressed by the poor quality GOP candidates

Certainly, these were real and very potent factors for many if not all Democrats, and for some independents. However, they were all “baked into the cake” long before the election and long before pre-election voter polls were administered. They were not last-minute issues.

‘); googletag.cmd.push(function () { googletag.display(‘div-gpt-ad-1609270365559-0’); }); }

I can think of a couple significant last-minute campaign developments, but they worked against the Democrats— not for them. Many pundits thought John Fetterman’s late October debate performance would put him into a political death spiral. Indeed, he was polling lower than Oz, just before the election. However, he ended up winning by over 4 points— somehow.

And Katie Hobbs was roundly criticized for her deer-in-the-headlights interviews near the end of the campaign. Just before the election, Hobbs was polling substantially below Lake, yet she won. Lake’s pre-election dominance is evident from the polling averages compiled by FiveThirtyEight:

Figure 1: Lake-Hobbs polling trends Oct 1 to Nov 8

When we look at the trend shown in Figure 1, and at the ending polling disparity (2.4%), it is hard to see how Hobbs managed to win.

Polling error?