Looking at the tag's excerpt history, I see that the excerpt is very sparse and we haven't tried many versions to see what might work best. My first suggestion would be to see if editing the excerpt helps.

For instance, we could start the excerpt with:

There seems to be an alternative to the chisquared test, in place of $\sum (O-E)^2/E$, just do $\sum (O-E)^2$, this is asked about at stats.stackexchange.com/questions/605760/…. I once found a paper about this, but cannot find it again now. Anybody knows?

@kjetilbhalvorsen I am unconvinced of its value. It's not wrong so much as difficult to implement and extremely specialized. It's difficult because it's almost impossible to determine what the null distribution should be and it is specialized insofar as (viewing it as a weighted version of the usual chi-squared test) it discounts excursions in rare cells and favors the most populous ones.

I see a lot of people writing "dependant variable" and not "dependent variable", as I have learnt it ... Is that some variant of english?

@kjetilbhalvorsen No, it's just a common spelling error.

If I spell it "dependant" rather than "dependent" it is a spelling error.