I'm not saying it isn't true, just that it's not statistically supported by the evidence presented in the paper.
You can invent post hoc justifications for almost any positive conclusion (although this one is weak since the authors admit the effect cannot be detected in individuals, only in aggregate).
You can invent post hoc justifications for almost any positive conclusion (although this one is weak since the authors admit the effect cannot be detected in individuals, only in aggregate).
[edit: you can even p-hack and retroactively justify your way to entire subfields; see famously https://slatestarcodex.com/2019/05/07/5-httlpr-a-pointed-rev...]