That’s what I thought too. For example if they “control” for factors like IQ or social economic status, then the correlations will be reduced.

This isn’t a surprise unless you think the delay of gratification is itself the cause of success (seems like a straw man so they can claim to “challenge” the original study)

There is more info from one of the authors here which includes the preregistration document: https://x.com/jess_sperber/status/1818100487964496119

Edit: Also, I think the associations of 0.17 prove the title is false