| Abstract: | Examined 13 research studies dealing with differences between negroes and caucasians in the validity of employment and training selection procedures. It was found that 100 of the 160 validity coefficients computed were not significant for either group, indicating the need for validation prior to operational use. Significant differences in the coefficients between groups were found only in 7 instances. Statistical significance of single group validity occurred in 33 instances. The combination of ethnic groups yielded lower validity than for either group separately on only 3 out of 120 instances. It is concluded that there is very little evidence of differential validity. Single-group validity appears linked to the use of ratings (rather than more objective criterion measures), and small samples. (25 ref.) (PsycINFO Database Record (c) 2010 APA, all rights reserved) |
