U.S. voter participation has not fallen steadily over the past few decades

We have an agreement with the Washington Monthly where they occasionally repost items from the Monkey Cage. So occasionally I go over to the site and take a look. Today I found this article by Samuel Knight, who writes:

It’s no secret that voter participation in the US has fallen steadily over the past few decades.

Perhaps this is true under some measures, but not if you look at vote for president:


Samuel Knight is listed as an intern for the Washington Monthly, and it’s pretty cool that they let their interns post on the blog. This should just be a lesson that even things that are no secret are not necessarily true. (I have a lot more sympathy when an intern makes a little mistake on a blog than when a bigger mistake is made in the ostensibly fact-checked New Yorker.)

No big deal, I just happened to notice this one and thought it would be good to clarify. I wouldn’t have bothered with it at all except that it’s in the same site that reposts some of our blog.

8 Responses to U.S. voter participation has not fallen steadily over the past few decades

  1. joni June 6, 2011 at 7:00 am #

    I’ve just gotten into your blog, and I like to pretend I understand it. (I’m a learning consultant/special ed teacher, and we use test results/statistics to analyze and advise on the best educational environment/program for developmentally disabled kids; Correlation and causation are big buzzwords in my field.)
    Anyway, when I clicked on the link to your prior article about the New Yorler article, I found an error: Carter spelled Cater, below the graphs. I’m sorry; I do not mean to do anything but make you aware.
    Keep on keeping on…

  2. Paul Gronke June 6, 2011 at 1:33 pm #


    It is amazing how long this meme has maintained. I generally attribute this to journalists who took political science back in the 80s–back when the discipline really WAS concerned about steadily declining voter participation. A related meme is “increasing number of independents” which is also constantly mentioned but is also no longer accurate.

    Perhaps Rosenstone and Hansen just don’t make the turnout point clearly enough. Perhaps many of us teaching political science were trained in the 80s and 90s and haven’t updated our notes.

    I don’t know. But this one has amazingly long legs.

  3. Dan June 6, 2011 at 4:59 pm #

    Think there’s any chance of targeting corrected?

  4. Joel June 6, 2011 at 8:55 pm #

    well, in fairness, this “meme” was still held to be true by most in the discipline until the uptick that started in 2000, yes?

    as a lay observer, i can’t help but wonder what it would look like if you removed the 1948 datum and then “fit” a line across the rest.

    in short, i’m not sure why you’re all so convinced that the meme is wrong. the trend that you’re dismissing was/is more sustained than the trend that you think changes the story. do you expect turnout to remain above 60% in 2012?

    • Andrew Gelman June 6, 2011 at 11:11 pm #


      It’s not a big deal, it’s just a false statement, that’s all. I prefer true statements. If you want to stay that voter turnout generally declined during the 1960-2000 period, then say that. If you want to say that you expect turnout to decline in 2012, then go for it. Just about all mistakes have some kernel of truth, and that’s fine. I just think it works better to separate the facts from the speculation.

  5. Kostas Gemenis June 12, 2011 at 6:02 pm #

    Why fit a line and not a loess curve?

    • Andrew Gelman June 13, 2011 at 9:36 am #

      I didn’t fit anything, I just connected the dots.

      • Kostas Gemenis June 13, 2011 at 2:33 pm #

        Yes, obviously. The comment was meant for the suggestion by Joel above.