Academic research has long suffered from a sampling problem, particularly when researching animal behavior.
Particularly that of humans.
The effect is perhaps most widely known from the wrong idea that wolves have "alphas." The whole concept was drawn from observing wolves in captivity, who were not behaving the way they do in the wild.
When observing humans, researchers have often been limited by who they could research. Often, that has meant "college undergraduates," which do not represent the population as a whole.
Sometimes the limitation has been by how we asked the question — such as the highly skewed diagnostic criteria for autism and masking behaviors that resulted in a pattern of gender-based underdiagnosis for female-presenting persons.
Progress has been made on this front in the biological medical sciences — for example, in 2023, the issue of diversity in medical trials began to get serious attention.
But this same attention must be paid to behavioral research as well. The fact remains that a significant amount of our research about human behavior {1} suffers both from the sampling problem and problems of how the question was framed.
For example the "marshmallow study" doesn’t say what we think it does about willpower, particularly when controlled for socioeconomic variables.
And then there’s the Milgram "obedience" experiment — whether the original or the virtual reprise in the early aughts. Whether you interpret the results as how willing people are to obey authority — or instead, how far people will go when they think it’s for the "greater good" — I think it’s fair to hypothesize that the results would be very different depending on the neurotype of the participant.
What were the neurotypes of the participants in either study? We simply do not know; the question simply was not asked. Not only do we not know, but even if they were assessed at the time, until really recently those assessments were themselves flawed.
We literally cannot know the extent of the problem without redoing a lot of experiments, some of which you literally cannot legally (or ethically) do today. {2}
All prior psychological research about human behavior needs to have a big old warning label.
Featured Image by felixioncool from Pixabay
{1} Sociology has this issue as well, though perhaps to a lesser extent than experimental psychology.
{2} Hence why the Milgram experiment was repeated virtually.