Reseaches on Several Problems about Test Equating

In this paper equating methods and equating designs about test equating were reseached. The improved linear equating was proposed, this equation is depended on correlation coefficient of two test and depended upon reliability of the test, linear equating which was used before is a particular example when reliabilities of two test are same. The confidence interval estimae of predicted value was given, at the same time, a new equating design was proposed: Single group design about the test was composed of half of a tes and half of another test.