• BodyBySisyphus [he/him]@hexbear.net
    link
    fedilink
    English
    arrow-up
    37
    ·
    edit-2
    3 days ago

    Oh boy, time for another round of Economists Try To Do Math:
    - A definitely real regression specification

    - Error bars? We don’t need to stinkin’ error bars. We use the Excel bar chart defaults like God intended. And axis labels are for people who want to look like they understand what they’re doing, which we certainly do not.

    R Squared of <2 fucking percent. The whole model explains 2% of the variation in income. 2%; the traits alone account for only 0.3%. Literally only one of the “Big five” traits was even statistically significant! This is a nothingburger! I would tell an undergrad nice effort but maybe ask a less stupid question next time.

    Who the fuck was dumb enough to publish this?

    Guenzel, Marius and Kogan, Shimon and Niessner, Marina and Shue, Kelly, AI Personality Extraction from Faces: Labor Market Implications (January 09, 2025). Available at SSRN: https://ssrn.com/abstract=5089827 or http://dx.doi.org/10.2139/ssrn.5089827

    Oh, so no one? The academic equivalent of self-publishing? The economist sure is going for the hard-hitting research. Anyone who was stupid enough to put their name on this thing deserves to be tarred and feathered.

    • elpaso [he/him]@hexbear.net
      link
      fedilink
      English
      arrow-up
      15
      ·
      3 days ago

      Those coauthors are from “well regarded” business schools too.

      Some competitive MBA programs want you to send in a photo. It’s literally “Hot or Not.” So this kind of bullshit from a business school does not surprise me.