1 TaylorRussell Tables Incremental validity the provide an

1 Taylor-Russell Tables • Incremental validity = the provide an estimate of improvement you get by the incremental validity of using a selection test, a test over selecting without a test

2 Taylor-Russell Tables • What you need: – Base rate – Definition of success • Must be clear (e. g. , “Increase sales by 10%”) – Selection ratio • What proportion of applicants will be hired – Validity coefficient for test • correlation of test score with some criterion (e. g. , a measure of work quality)

3 Taylor-Russell Tables • Taylor & Russell (1939) showed that a selection test is most useful when validity is high and selection ratio is low • If selection ratio is high, test doesn’t matter much (you hire people anyway) • If validity is low, test won’t help you (even if you are very picky)

Test is most useful when validity is high and selection ratio is low

5 WARNING: The Taylor-Russell tables are of value primarily in drawing attention to the importance of utilizing the selection-ratio concept rather than in furnishing dependable specific probabilities of expected improvement in effectiveness of selection. The following cautions should be noted to avoid unwarranted reliance upon low validity coefficients or misuse of the tables: (1) the specific probabilities indicated in the Taylor-Russell tables are quite inapplicable to triangular scattergrams; (2) even when an approximately normal, elliptical scattergram exists, the tables often yield an inaccurate estimate of the prospective gain in effectiveness. This is because the tables assume that the applicant group and the present employee group are similarly constituted. This is equivalent to assuming that at present 100% of applicants are being hired and retained or that our current selection procedures have zero validity or that both conditions prevail. (Psyc. INFO Database Record (c) 2006 APA, all rights reserved)