While I admit the logic is somewhat tautologous, that's not quite the conclusion I'm drawing. I, first of all, look at the average rating of the writers involved in the show, and then look at how the make-up of the show's writers is substantially different in the early years vs the later years.

The only thing this experiment could show is that if the ratings changed, but the writers didn't, then the writers are not responsible. It can't show the reverse.

Everything in the article is just properties of arithmatic averages in sorted sequences.

I don't see how that's circular reasoning. The writers who worked on bad episodes presumably had something to do with making them bad.

Potentially it was bad writers, yes. But this data does not prove this; it just shows that some writers tenure correspond with worse episodes. But so does for example the year, or potentially many other influencing factors.

If we’re going down that rabbit hole, it doesn’t even prove there is such a thing as a bad episode. It could just be that people who don’t like the simpsons became more likely to use IMDB as time went on.