Poll: This is the scandal that could actually hurt Trump’s chances

There may finally be a scandal that can take down a candidate who otherwise seems impervious to bad news. Republican leadership is abandoning Donald Trump en massse after video surfaced of him saying how he tried to sleep with a married woman and felt comfortable groping women. Whatever you may think about politicians, they are demonstrably good at reading the tea leaves in public opinion and they suspect something is different this time around.

For the past few months, I’ve been running a series of polls to understand the psychology of Trump supporters and also to understand if there are any limits (at all) to conservative party loyalty. I’ve asked about real scandals, like Trump not paying taxes and crazy hypotheticals, like if Trump wanted to round up all American muslims into internment camps. For the most part, nothing hurts this candidate, which explains why he’s risen in the polls over the last year after each cycle of media outrage predicting his downfall.

But, thanks to the help of Google’s Survey team, I ran a flash national opinion poll last night to see how this scandal compared — it’s much worse than anything that came before. The summary is that this latest scandal could cost Trump upwards of around 3% of the popular vote (at the extreme end).

This may not seem like a lot, but in an election, it’s the margins that count. The vast majority of voters will not change their mind — short of something truly catastrophic. Indeed, one reason why Trump seems to have gotten so popular over the last year is that so-called “scandals” may actually help him with some voters. I ran a small poll after the revelations that Trump didn’t pay taxes and (at the time of the writing) found that an equal number of respondents said their would both switch support away from him and switch support for him. I suspect this is because many voters are truly looking for someone that the media hates and defies norms in every possible way. This is an anti-establishment election.

But, in regards to the raw numbers, 1 or 2 percentage points can make or break a presidential election. It can also seriously cost down ballot candidates, such as Congressman, which is why many are now calling for his resignation.

The only other issue I’ve polled that really hurt Trump was a hypothetical asking what would happen if he proposed serious gun control: ~30% of his supporters said they’d be less likely to vote for him.

So, while I think the media may generally overhype previous scandals, this one does seem different.

One nerdy methodological note about Google polls: I use them for a few reasons. One, on controversial issues, especially related to Donald Trump, online polls have done differently than phone polls, because it appears that some demographic groups feel uncomfortable stating their opinion to a live person. So, on sensitive issues, they may be more accurate.

Second, Google is is really good at placing the margin of error prominently on their graphs. Pollsters like to interpret subgroups (like young voters or women voters). But, the margin or errors on subgroups are usually really really big (often over 10 points either way) because the sample size is so small. I like being transparent about polling uncertainty.

Anyway, as you read reports of mass Republican abandonment, it’s helpful to know that this one may be based on good evidence of a self-defeating scandal unfolding.

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.