Skip to content

“Would Republicans pay a price if they vote to impeach the president? Here’s what we know from 1974.”

I better post this one now because it might not be so relevant in 6 months . . .

Bob Erikson answers the question, “Would Republicans pay a price if they vote to impeach the president? Here’s what we know from 1974.” The conclusion: “Nixon loyalists paid the price—not Republicans who voted to impeach.”

This is consistent with some of my research with Jonathan Katz from awhile ago. See section 2.3 of this unfinished paper.


  1. D Kane says:

    > what we know

    Are you trolling us? That summary is almost a textbook case of the garden of forking paths! Leaving out 2 of 17 cases for reasons that are, at best, ad hoc.

    • Andrew says:


      As I wrote, Erikson and Wright’s conclusion accords what Katz and I found when we looked at the data earlier. So I guess I was focusing on their general message rather than the details. Ultimately there’s only so much that can be learned from such historical data, as Erikson and Wright say in their post.

  2. Terry says:

    So we have one observation (one impeachment vote) where public opinion turned against the president after the impeachment vote, and Republicans who voted against impeachment suffered electorally. (According to the WaPo article, the vote was in July 1974, and Nixon’s “guilt was clearly exposed in August”.

    This proves what exactly?

    I’m just taking a wild guess here, but it sounds like the electoral effects might depend on how public opinion evolves after the impeachment vote. If so, the WaPo article “proves” that an impeachment vote may hurt or help Republicans and it may help or hurt Democrats depending on developments after the vote.

    The WaPo article mentions this only in the very last paragraph. A more serious article might have mentioned this earlier. But then, of course, the anti-Trump upshot of the article would have been attenuated. Indeed, a more serious article might have given some thought to how public opinion might evolve as events shift from the Schiff impeachment inquiry to a Republican-controlled Senate trial. But, of course, that would have severely attenuated the anti-Trump tenor of the article. Indeed, it might have suggested that an impeachment vote in 2019 might actually harm Democrats and help Republicans.

    • Andrew says:


      I agree that this is n=1, or 2, or some small number. Just to be clear, you put “proves” in quotation marks, but Erikson and Wright never used that word in their article, nor do I think did they imply that they proved anything.

      Given Trump’s unpopularity, I doubt that an impeachment vote would harm Democrats and help Republicans. But I agree with you that we don’t know. Public opinion was far from static in 1974. What with political polarization, I’d expect opinion to be more stable now than it was in 1974, but things could shift. We don’t know, and we should look at historical data with their limitations in mind.

      • Terry says:


        I wrote “proves” as scare-quotes, not to indicate that it was a quote from Erikson and Wright. I can see the confusion.

      • Terry says:

        I didn’t notice that n=2 because the authors are so quick to dismiss the Clinton impeachment and because I didn’t understand their logic for dismissing it. On closer examination, their logic is rather odd. They write:

        The more recent impeachment of President Bill Clinton in 1998, on the other hand, offers no useful lessons for today, since congressional votes on impeachment were closely tied to members’ ideological leanings. Parties had already begun to polarize into what we see today.

        Of course, the context in 1974 was different. Among other differences, most Americans then got their news from the same sources; today, they consume news through partisan outlets and social media feeds that frame what they hear quite differently.

        Still, let’s take a closer look.

        They seem to be saying that things were very different in 1974 and that 1998 was more similar to today: (1) in 1998, “parties had already begun to polarize into what we see today“, and (2) most Americans in 1974 got their news from the same sources, which is very different from today. They also kind of admit this by saying “Still, lets take a closer look [at 1974]”.

        So they have two observations, and they ignore the one more similar to today in favor of the one that is “of course” different and that produces an anti-Trump result. I wonder why.

        • Andrew says:


          I’ve thought about the Clinton impeachment. I don’t recall whether it made its way into my paper with Katz, but we were planning to include it at some point.

          To me, the key difference between the Clinton impeachment and the Nixon and Trump impeachments is not the passage of time, but rather that Clinton was popular, and Nixon and Trump are not. Also, there was always doubt about an impeachment based on a sexual affair.

          So, I do think that Erickson and Wright could’ve written that aspect of their post more clearly, but I understand why they’d want to consider the Clinton impeachment separately, as there’s a big difference between a popular and an unpopular impeachment.

  3. Alberto says:

    At the moment, the Democrats look more likely to be hurt. Impeachment is unpopular in battleground states

    If opinion is more stable than in 1974 then state polls aren’t unlikely to turn, either.

  4. Dzhaughn says:

    What about the Clinton impeachment?

  5. jim says:

    General comment on Trump, polling, and distinctions that might not be obvious:

    I suppose it’s been done, but I wonder how effectively: an effort to separate Trump the man from Trump the administration and policy.

    The other day a columnist here wrote about how much “Trump’s trade war” is costing Washingtonians, but no one seems to care. Perhaps however much it’s supposedly costing Washingtonians, you can’t tell on the street: the economy is holding up, so the line that it’s costing people so much isn’t consistent with what people see.

    There are similarities between Clinton and Trump: in Clinton’s case, people cared about how he was running the country, not about who was under his desk. But in Trump’s case it’s not that much different: most people don’t care about Ukraine either.

  6. TMC says:

    There’s a big difference between this and previous impeachments. All the others were a situation where both sides agreed there was a law broken. In this situation the Republicans definitely do not, and a lot of Democrats also seem not to believe. Schiff’s hearings just had every single person who had direct knowledge of the events, Rep leaning or Dem leaning, testify that they thought Trump did not break the law.

Leave a Reply