an assessment in which answers are scored in such a way as to establish differences in responses among groups already known to differ. For example, on a test measuring problem solving, the correct alternative among the response choices would be the one preferred by members of a criterion group who were administered the test previously. See also empirical-criterion keying.