Laserfiche WebLink
RBH test has validity which is comparable to that of the MPRS test. <br />However the RBH test results in a difference between mean scores <br />of blacks and whites of .62 standard deviation. Dr. Sackett's <br />estimate of the difference in mean scores under an amended MPRS <br />testing process was .5 standard deviation for the CG predictor and <br />.3 standard deviation for the SR predictor. There is no <br />justification for the great expense of a transportability study of <br />a test which has no greater validity and is likely to result in <br />greater adverse impact. <br />Plaintiffs have also suggested that the MPRS be required to <br />use the AB forms of the Watson Glaser test. The greatest problem <br />with this suggestion is chat the AB forms are available in advance <br />CO the test takers. A test of this type cannot be relied on to <br />give valid results if test takers have prior access to the test. <br />Even if test security were not a problem, there is no evidence that <br />the AB forms of t.he test have equal or greater validity than the 42 <br />forms or that they would have less adverse impact. The only <br />argument for using the newer forms of the test relates to the <br />objectionable questions in the older forms. If these questions are <br />eliminated fro.m the older forms, there is no reason to compel the <br />use of Che newer ones. <br />There was some discussion at trial of further diluting the <br />impact of the cognitive Watson Glaser test by adding another <br />cognitive test, such as a content-based test, to the MPRS test <br />battery. The proble.m with this suggestion is that no concent-based <br />tests were identified which do not have any adverse impact. As <br />33