How to pronounce“quartiles”in English

So we put it into the quartiles, high to low.

which we use to assign the units in group K to quartiles.

each observation has been assigned to one of the four quartiles.

We're going to put the observations into the four quartiles in the holdout fold.

And we can't tell the difference between quartiles 1 and 2.

So it looks like, the high quartiles are above the low quartiles.

And use that model to build a drive model mapping x each of four quartiles based on the tau hats from this, the CATE iteration,.

And then we'll finally estimate the average treatment effect in each of the four quartiles for fold k using the outcomes in fold k.

by definition, 37 and a half percent and so on if you just divide them into quartiles.

A related exercise we can do, is to basically break the data into four quartiles of estimated CATEs.

And then using those estimates, we define a mapping from covariates to the quartiles of the k distribution.

And then looking held out data to see for each quartile of CATE, what is the estimated treatment effect for each of those quartiles.

Well in 2017, we're able to pick out quartile 1, but we're not able to distinguish between the top three quartiles.

But it's also pretty hard to tell the difference between quartile three and quartile four.