The Validity and Reliability of Value Measurements

ABSTRACT Four scaling techniques for measuring human values were compared, using the measurement properties of test-retest reliability, discriminatory power, convergent validity, and achievement of substantive results consistent with theoretical expectations. The standard rank-order version of the Rokeach Value Survey was superior to a 100-point rating, magnitude estimation, and the handgrip scaling procedure with respect to each of these fundamental measurement requirements. Because it exhibited less measurement error than the other methods, rank-order scaling is shown to be the best technique for measuring human values even though it achieves only an ordinal level of measurement.