Small sample sizes don’t invalidate studies. They reduce the statistical certainty, but can still be accurate and there’s formulas for gauging how accurate a given sample size is based off the standard deviation.
The student only part does mean a sampling biase, but that doesn’t invalidate things either. Mostly just limits it such that we can only make the claim for students and not anyone else (but it’s still a meaningful result and provides justification for larger studies).
br3d@lemmy.world 1 year ago
This is such a common sneer, but it’s naive and misguided. In the hope a few people see this, consider a thought experiment: how many students would you have to push off a cliff to get solid evidence that pushing people off a cliff is harmful? The point being that when a phenomenon is powerful, it’s quite possible to study it even with small and selective samples.
exi@feddit.de 1 year ago
I get what you mean but please read the study: link.springer.com/article/…/s41347-023-00304-7
They themselves note that the very small sample size would be an issue. They say they would need 78 people for even remotely confident results, then they initially targeted 74 people, of which 20 dropped out.
Let me be clear that I too want to believe that social networks are bad for people, but studies like this one do very very little to provide any meaningful data to base my opinion on.
br3d@lemmy.world 1 year ago
It’s good to look at the original study, but what that tells me is that for the effect to emerge as significant despite the lack of participants, it was quite possibly even larger than the researchers estimated in their power analysis