This report was designed to respond to two major methodological shortcomings in the item bias literature: (1) misfitting test models; and (2) the use of significance tests. Specifically, the goals of the research were to describe a newly developed method known as the "plot method" for identifying potentially biased test items and to conduct several methodological investigations associated with applying the plot method. Following an introduction to the theory which forms the basis for the three-parameter logistic model and the model itself, the plot method first described by Shepard was presented. Advantages of the plot method were that (1) the problem of sample size was controlled for through the baseline plots; (2) the baseline plots provided a basis for interpreting the importance of particular independent variables on the invariance property of item difficulty parameter estimates; and (3) the concept of replication could replace the concept of statistical significance testing. The plot method resulted in reasonably stable determinations of potentially biased test items with small samples. Other methods with small sample sizes were not as successful. Methodological findings provided direction for future applications of the plot method. References and figures are appended. (Author/PN)