[Excerpt] Imagine that you have set out to buy a used car. You examine eight cars before making your choice, test driving some of them and rejecting others at first glance (due for example to excessive rust). A researcher asks you to rate each of the eight cars in terms of overall quality.
The researcher proceeds to sharply criticize you for carrying out an unsystematic search process. Your failure to test-drive every car and to ask the same questions to the dealers about each car has caused you to do a poor job of rank-ordering the cars. You respond that, since you could only afford one car, you had no interest in rank-ordering or in assigning ratings to the entire set of cars. It seems unfair to be criticized for poor performance of a task which was unrelated to your original mission of buying the best used car available.
This paper explores the possibility that a similar misspecification of the goals of employee selection has caused researchers to criticize selectors for behavior which may not adversely affect the goal of hiring the best individual from among a group of candidates.