The initial steps necessary for constructing an analogies test are described. Basic mathematics being the context, the aim of the test is to predict student performance in research statistics courses characteristic of graduate education and social science programs. Items involving analogies in a quantitative context were built, sorted into two forms of the test (Q-Analogies), administered to beginning statistics classes, and analyzed on the BICAL program. Two revised forms of the Q-Analogies Test were developed and administered to intermediate classes. A basic algebra test was also administered to the classes. Results suggest the following for the Q-Analogies: (1) it is superior to an algebra test in predicting success in quantitative courses; (2) it measures with high reliability; (3) scores on both forms are quite comparable and simply equated; (4) it discriminates between groups defined on the basis of mathematical expertise; and (5) it appears promising as a valid, reliable measure of quantitative ability and warrants further trial. (Author/GK)