Google executives acknowledged this month they need to do a better job surfacing user-generated content after the recent Reddit blackouts.

  • goetzit@lemmy.world
    link
    fedilink
    English
    arrow-up
    3
    ·
    1 year ago

    Don’t you also need to worry about your sample population being biased? You’d only be sampling people who sought out a dislike plugin, these people might be much more likely to dislike a video. Is there any way to account for that?

    • tool@r.rosettast0ned.com
      link
      fedilink
      English
      arrow-up
      1
      ·
      1 year ago

      You’d have to have a separate cohort of non-plugin users & another with a sampling of both, I think. Run some regressions on those data and I think you’d be able to tease out any bias that exists.