While major test publishers randomly select school districts for their national norming studies, a survey of "accepting" and "declining" districts supports the hypothesis that self-selection bias results in overrepresentation of districts which already use a specific publisher's tests or instructional materials. (Author/BW)