This would not pass peer review for a journal as written.

Maybe the conclusion is correct, or maybe not, but as written the methodology is under specified, statistics are not supported, and there too many confounders not addressed. One should not take anything from this without a better write up. Just misunderstanding what n= means is a huge flag.

Since the author is here, I have to ask: Why a blog post and not an actual paper? Why spray this onto the internet without validating the work? Or, conversely, why not caveat the work as exploratory data science?