Unfortunately the p-values that are significant are only nominally significant. If you do a null experiment enough times it will sometimes look significant - ev

Unfortunately the p-values that are significant are only nominally significant. If you do a null experiment enough times it will sometimes look significant - even a broken clock is wrong sometimes. This doesn’t mean the paper isn’t detecting a real effect, its just that the noise in these experiments is large and the techniques or sample size insufficient to confidently overcome that noise. It’s a great contribution to it’s field, but I won’t be changing my habits based off it.