Whereas G Studies determine the G Coefficients of an actual test, D Studies help with 'what if' scenarios, i.e. in finding most cost effective designs for actual tests, based on preliminary variance components from pilot studies.

Once we know the variance components, the G Coefficients depend only on the sample sizes of the facets of generalization by simple arithmetic calculations.

So let's look at the result of the G Study:

Now let's look how we do a D Study: