• bjorney@lemmy.ca
    link
    fedilink
    arrow-up
    45
    arrow-down
    4
    ·
    11 months ago

    If male tears were the only control, then they run the risk of not finding any result. If you have 3 groups, you need a substantially larger sample size because you are running a less powerful statistical test.

    Easier to start with the test that’s most likely to work, and narrow it down from there if you succeed

    • queermunist she/her@lemmy.ml
      link
      fedilink
      arrow-up
      17
      arrow-down
      9
      ·
      11 months ago

      Having men sniff three different samples would still allow for saline as a control and wouldn’t really make the data set that much more complicated.

        • queermunist she/her@lemmy.ml
          link
          fedilink
          arrow-up
          15
          arrow-down
          9
          ·
          11 months ago

          Just college lab courses, but come on, it’s pretty basic. The experiment merely tests a single variable by changing it while keeping everything else the same. There could have been dozens of different samples that men sniffed and it wouldn’t really make the data complicated.

          It would increase the length of the test, though, so dozens of samples would have been cumbersome. But just two? Literally just “see how the test group responds to sample 1, sample 2, and the control sample”? That’s not complicated science. You probably did that in highschool lol

          • criitz@reddthat.com
            link
            fedilink
            arrow-up
            12
            arrow-down
            1
            ·
            11 months ago

            Testing multiple hypotheses this way still requires additional sample size because there is an increased error likelihood. From a statistical point of view, the most efficient test may be to stick to one variable like this.

          • DanglingFury@lemmy.world
            link
            fedilink
            arrow-up
            6
            ·
            edit-2
            11 months ago

            I’m guessing they had to stay within their funding/budget and didn’t want to reduce the sample size to increase the number of variables tested. MRIs are expensive

            • HubertManne@kbin.social
              link
              fedilink
              arrow-up
              0
              ·
              11 months ago

              they should just be getting time on the machine although maybe also tech time. either way doing multiple with a single individual is easier than more individuals.

              • Meowoem@sh.itjust.works
                link
                fedilink
                arrow-up
                2
                arrow-down
                1
                ·
                11 months ago

                But that makes it more complex because you have to start worrying about the order they’re done in because it might be different emotions playing your first or third game plus the effect might linger, take time to show, etc.

                Far better to answer one simple question and prove there is an effect then follow up tests can look at finding the bounds to that and starting to narrow in on identifying mechanics.

                • HubertManne@kbin.social
                  link
                  fedilink
                  arrow-up
                  1
                  ·
                  11 months ago

                  its been done with other things. its an mri so regardless of order the simuli should light up the brain regions. either it does it or it does not.