Wednesday, May 11, 2022

How Useful is "Generic Ballot" Polling ...


... or, rather,  is it useful and if so when and why?

"Generic ballot" polling is simple: Ask a bunch of voters whether they're likely to vote for 1) this party or 2) that party in the next election.

Ceteris paribus, it's useful in the sense of being broadly predictive.

But there are ceterisi that can go non-paribus in ways that make it less useful.

One is the local/incumbent factor: "For all those other races, I prefer Party X. But my district's Party Y congresscritter, well, he took care of my gammy's Social Security problem, and besides, Party X's candidate in my district looks like of shifty (read: The incumbent had a lot more money to spend making the challenger seem shifty than vice versa)."

Another is enthusiasm/turnout. If 51% of "generic ballot" respondents prefer Party X to Party Y, but Party Y gets out 90% of its supporters and Party X only manages to get out 50% of its supporters, Party X is going to lose.

For an obvious example, let's look at non-generic polling in the 2016 presidential election. More Americans polled for Hillary Clinton than for Donald Trump. But a lot of Clinton supporters weren't enthusiastic enough to act on that preference, while most Trump supporters were.

To quote a previous post of mine:

In 2012, Mitt Romney received about 60,000 votes in Erie County, Pennsylvania.

In 2012, Donald Trump also received about 60,000 votes in that county.

But in 2012 Barack Obama received about 91,000 votes in Erie County, while in 2016 Hillary Clinton received about 58,000.

As you can probably already tell, this factor is what's driving my skepticism when it comes to a "Red Wave" in November.

I think the Democrats are in good position to hold on to their 50 Senate seats and maybe even pick up one or two.

Not because most Americans prefer Democrats to Republicans, but because Democrats have a couple of tools at their disposal to get out their voters in higher percentages than Republicans.

One, likely coming in August, is large-scale student debt "forgiveness." That will tip some voters off the fence on the Democratic side, and some of them enthusiastically so.

Another, likely coming Real Soon Now, is the Supreme Court overturning Roe v. Wade.

And a third is the assistance of Donald Trump in making the midterms -- to the extent that they can be turned into a "national election" -- a referendum on Donald Trump rather than on Joe Biden. We've had two such referendums on Trump (2018 and 2020 -- 2016 was more a referendum on Hillary Clinton), and in both he got his ass whipped.

The Democrats don't have it in the bag, or even close, but it's looking a lot better for them than  "generic ballot" polling (which, last I noticed, had the GOP up by 7%) reflects.


blog comments powered by Disqus
Three Column Modification courtesy of The Blogger Guide
Some graphics and styles ported from a previous theme by Jenny Giannopoulou