The authors empirically demonstrate some of the trade-offs which can occur when content balancing is imposed in computerized adaptive testing (CAT) forms or conversely, when it is ignored. The authors contend that the content validity of a CAT form can actually change across a score scale when content balancing is ignored. However they caution that, efficiency and score precision can be severely reduced by over specifying content restrictions in a CAT form. The results from 2 simulation studies are presented as a means of highlighting some of the trade-offs that could occur between content and statistical considerations in CAT form assembly. (PsycINFO Database Record (c) 2003 APA, all rights reserved).