I just randomly clicked on your middle link "2020 (large-scale replication study): https://www.researchgate.net/profile/Emil-O-W-Kirkegaard/pub..." which, to be clear, did not replicate the results of the study.

Instead, it correlated pupil size with race. Given all of the socioeconomic associations with race, sounds kinda like a confounding variable could very well be at play.