(2)
Further evidence on this point,in dollars-and-cents terms,comes from an early study carried out on the selection of telephone operators for AT&T in the San Francisco area(Rusmore&Toorenaar,1956).A test battery containing numerical and clerical ability measures was used as a predictor.Calculation was made of the savings attributable to the fact that many operators who would ultimately have failed were not hired and there fore did not have to be trained.The following figures correct these training savings for the additional recruiting costs that the tests introduce.Scores on the test battery may range from 0 to 80,and the minimum score for hiring could,of course,be set at any point between,depending on the availability of applicants in the labor market and the degree of selectivity desired.This study was published in 1956;the savings would be many times greater today because of inflation.