@TomJewell

The introduction to Chapter 2 (reliability/precision) acknowledges that the sampling process can lead to significant differences in the determined coefficients,
that they therefore [2014, p. 37] "should be estimated separately for all relevant subgroups"
and that this is also important to assess potential consequences for #TestFairness.

And in Chapter 5 focuses especially in the discussion of #Norming of #TestScores on such aspects, including relating to our previous discussion...