Abstract: | Interpretation of psychological measures is sometimes based on relationships between test characteristics (factor loadings, proportion of items keyed True, etc.). The direction of measurement adopted for a scale often determines how test characteristics are recorded. Failure to regulate direction of measurement, called here direction error, produces misleading results by affecting the distribution and intercorrelations of test characteristics. Examples are drawn from criterion analysis, the study of acquiescence, social desirability, and communality. A method for regulating direction of measurement is discussed. Its rationale is based on a distinction between directional and differential measurement. (34 ref.) (PsycINFO Database Record (c) 2010 APA, all rights reserved) |