August 17, 2015

How would you even study that?

From XKCD

every_seven_seconds

“How would you even study that?” is an excellent question to ask when you see a surprising statistic in the media. Often the answer is “they didn’t,” but sometimes you get to find out about some really clever research technique.

avatar

Thomas Lumley (@tslumley) is Professor of Biostatistics at the University of Auckland. His research interests include semiparametric models, survey sampling, statistical computing, foundations of statistics, and whatever methodological problems his medical collaborators come up with. He also blogs at Biased and Inefficient See all posts by Thomas Lumley »

Comments

  • avatar
    John Egan

    OK I’ll take up the challenge!

    Point me to the empirical evidence—physiological—that supports than the anal wall is less robust than the vaginal wall.

    As is claimed in HIV prevention materials around the world.

    I’ll keep checking back. :)

    Kind regards,

    A sociologist

    2 years ago

    • avatar
      Thomas Lumley

      John,

      You may be mistaking me for a physiologist or possibly for Google. I didn’t say *I* knew the basis for every scientific claim ever made, just that people making the claims should know.

      2 years ago

      • avatar
        Thomas Lumley

        Also, I didn’t read that cartoon as anti-sociologist, more as sociologists wondering how other people could be so credulous.

        2 years ago

        • avatar
          John Egan

          Sociologists are notoriously thin-skinned. ;)

          2 years ago

  • avatar
    Megan Pledger

    I know it wasn’t really meant to be answered but I was walking around town thinking about it so here goes…

    Give a random selection of the target population cells phones and then at random (perhaps stratified/balanced) times during the day text them and ask them “How long before receiving this text were you thinking about sex?” and then do some survival analysis* to get the mean (adjusting for time of day/age, if neccesary).

    One potential issue, after the experiment is over, is that it could become a Pavlov’s dog situation where respondents start thinking about sex when they get a text.

    *Probably parametric to get around the right and left censoring issues.

    2 years ago