J. Cohen's Review of Tukey’s Exploratory Data Analysis [pdf]

My friend, Joe Marasco <rchimed@comcast.net> recently circulated this endorsement to a closed list. I thought it might be of interest to the HN community. Joe writes:

I heartily endorse the attached article. He inveighs against the routine, mechanical, and mostly mindless misapplication of what he calls NHST, null hypothesis statistical testing. It does require that you slog through it at times, but the effort is worthwhile. It’s sort of like having to march a few miles through thigh-deep fresh snow; when you get to the cabin with the roaring fire in the hearth, it all makes sense.

I was pleased to see Cohen's laudatory review of Tukey’s Exploratory Data Analysis, one of those few books that survived my massive library downsizing of 2015. I hope to get around to reading more of it in the future.

The nonsense Cohen assails here is confined to what he calls the soft psychological (pseudo)sciences. I am not so confident that the disease has not metastasized to other less vulnerable targets, medical research in particular. Of course the consequences of sloppy reporting of results there are ironically much more severe, given Big Pharma’s thirst for profits and the FDA’s tenuous grip on the situation.