As industry's most popular selection device, the interview is too often misused to measure or predict numerous skills and abilities that can be measured or predicted better by other methods. The authors encountered such an abuse in a recent allegation of sex discrimination. This paper reports a study conducted for this litigation. Couched in a general lens model paradigm, the results demonstrate an inappropriate use of a selection technique to predict an ability that can be better measured by a test. This simple interview analog demonstrates the problems of replacing tests with inappropriate selection devices.