This paper presents an introduction to the use of latent trait models for the estimation of domain scores. It was shown that these models provided an advantage over classical test theory and binomial error models in that unbiased estimates of true domain scores could be obtained even when items were not randomly selected from a universe of items. This allowed the full use of item statistics during test development. Furthermore, it was argued that when all domain subtests measured the same underlying trait, then all subtest domain scales might be related to each other via the underlying theta scale. Finally, a method was presented whereby the unidimensionality assumption could be tentatively tested. This was to fit a LISREL congeneric measurement model to the observed subtest variance-covariance matrix. (Author/GDC)