Rich "The People's Pundit" Baris

We often ignore the years when exit polls correctly called the winner, though were still showing serious issues due to at least non-response bias. Veteran exit pollster Murray Edelman has spoken a lot about this. Here are some notable, egregious misses.

In 2004, the Kerry Campaign was elated. The Bush Campaign was in shock re Florida. The latter was confident they over-performed exit polls. By night's end, a Kerry exit poll win turned into a historically comfortable Bush win in a razor thin state he barely won 4 years before.

But that still left the Bush Campaign short in Ohio and Iowa, both of which exit polls suggested they would lose. The loss for Kerry and his supporters was so painful that they continued to argue Bush stole Ohio years after he won a second term. Why?

Exit polls showed it was mathematically impossible for Bush to win Ohio. That is why Russert and others were openly celebrating before the polls closed. Hyped expectations leave people entertaining conspiracies rather than attributing misses to known problems with exit polls.

Even in recent elections in which the winner was correctly predicted, the exit polls have overstated Democratic support. That's because Edelman was right re: non-response bias specifically re: older voters and other more conservative voting groups, issues pollsters now fail with.

But the final straw was in 2016, when the exit polls showed Hillary Clinton would win everywhere. It was a total defeat for Donald Trump, let alone failed to predict rather comfortable wins in Florida, North Carolina, Iowa, Ohio and elsewhere. Thus, the use of exit polls...

Thus, the use of exit polls conducted by Edison Research were outright abandoned and replaced by AP and their partners such as Fox News. The "voter analysis" has been touted as a far more accurate system due to ONLY the call for control of the U.S. House in 2018. Like polls...

Like pre-Election Day polls and exit polls, the success of voter analysis has been largely limited to that call for the U.S. House. State-level races in Florida and the Midwest, were not accurately gauged by voter analysis. The 2020 Democratic primary performance was even worse.

In Maine, it was mathematically impossible for Joe Biden to defeat Bernie Sanders, according to AP/Fox voter analysis. Bernie led with every age group, educational group AND both genders in the post-closure wave. Folks, Joe Biden won Maine. They're later revised to = reality.

In Minnesota, it was a very similar story. Bernie was going to defeat Biden despite the last-minute endorsement by the home state senator. Folks, Biden beat Bernie in Minnesota. Truth is the AP/Fox voter analysis struggled and struggles with the same groups Edelman discussed.

All that said, imagine how difficult it will be this year for either exit polls or voter analysis to report accurate results. Even if we could gauge preference in either VBM or in-person, which clearly is wrong, non-response & social desirability, will be even worse. In sum...