Hacker News Reader

>really comes down to being statistically literate

A lot of people writing scientific papers aren't statistically literate.

It also seems that what was considered first rate methodology even a decade ago is now considered deeply unreliable.

A lot of it depends on who (as in, which group of scientists) is doing the considering.

Most scientists who don’t have grad level math education fail miserably in understanding statistics. Even the top ones. I was the statistics expert in my lab when doing my PhD and I had no business being one. All I was good at was calling out that one way ANOvA is just a t-test and when they should just stick to non parametric methods.

> one way ANOvA is just a t-test

If there are only two groups, yes. There are often more than two groups.

Also helps that Tüfekçi got her undergraduate degrees at good universities in Turkey which means she had to go through and score well on the grueling university entrance exams.

The die hard masking advocate is your example? Wow