Measurement error significantly biases interaction effects and distorts researchers' inferences regarding interactive hypotheses. This article focuses on the single-indicator case and shows how to accurately estimate group slope differences by disattenuating interaction effects with errors-in-variables (EIV) regression. New analytic findings were…

Presents minimization and maximization models for parallel test construction under constraints. The minimization model constructs weakly and strongly parallel tests of minimum length, while the maximization model constructs weakly and strongly parallel tests with maximum test reliability. (Author/SLD)

The purpose of this paper is to highlight some psychometric cautions that should be observed when seeking to develop short form versions of tests. Several points are made: (1) score reliability is impacted directly by the characteristics of the sample and testing conditions; (2) sampling error has a direct influence on reliability and factor…

The purpose of this paper is to discuss Florida teachers' recommendations for improving the Florida Comprehensive Assessment Test (FCAT) and to compare their recommendations with those of Florida administrators. Although teachers' suggestions varied as to the types and extent of remedies needed to improve the FCAT, some common themes emerged. The…