On Measurement of Machiavellianism

Abstract

Recently. Bloom ( 1 ) and Shea and Beatty ( 5 ) have debated the reliability of the Mach V scale ( 2 ) , a popular instrument used to measure interpersonal manipulation. Shea and Beatty's main criticism concerns the internal consistency of the instrument; they note that, while Christie reported coefficients in the .Go's, other researchers consistently obtain coefficients under .GO, sometimes substantially so. Bloom, meanwhile, staunchly defends the Mach V, noting that internal consistency is not as essential in a research measure as is temporal consistency. The authors oveslooked two other problems inherent in the Mach V. First, Rogers and Semin ( 4 ) noted that, while Christie may have removed effects of social desirability with a triadic forced-choice format, he reintroduced them through the scoring system. Second, the Mach V is an ipsative instrument, whereby selection of one item necessitates rejection of another. This requires the use of nonparametric statistical techniques, lessening the power of analyses. Shea and Beatty ( 5 ) concluded that a more reliable measure of Machiavellianism should be devised and suggested a Likert scale as one alternative. What both they and Bloom (1) overlooked is that such an instrument already exists in the Mach IV, a 20item, seven-step Likert scale which has been used extensively in research. Its internal consistency is in the neighborhood of .70 to .80 ( 2 , G ) , which is certainly better than the coefficients for the Mach V. A 6-wk. test-retest coefficient of .76 for the Mach IV indicates a personality characteristic that is also stable over a moderate length of time. Using a short form of the Marlowe-Crowne Scale ( 3 ) , Zook and Sipps ( 6 ) found an over-all coefficient of .I5 with social desirability. This is a lower correlation than found for the Mach V and is certainly an acceptable value for use in research. In summary, the Mach IV has numerous advantages over the Mach V. First, reliability is better. Second, correlations with social desirability are lower. Third, parametric statistics can be used in analyses. Since information on validity is similar for the two scales, the Mach IV is recommended over the Mach V for research. REFERENCES 1 . BLOOM, R. W . (1984 ) Comment on Machiavellianism with Mach V: a psychometric Journal of Personality Assessment, 48 , 26-37, 2. CHRISTIE, R. ( 1 9 7 0 ) Scale construction. In R. Christie & F. Geis (Eds.), Studies in MachiavellianUm. New York: Academic Press. Pp. 10-34. 3. REYNOLDS, W . M . ( 1 9 8 2 ) Development of reliable and valid short forms of the Marlowe-Crowne Social Desirability Scale. Journal of Clinicat Psychology, 38, 119-125. 4. ROGERS, R. S., & S ~ I N , G. R. ( 1 9 7 3 ) Mach V: an improved scoring system based on a triadic choice model. Journal of Personality and Social Psychology, 27, 3440. 5. SHEA, M . T., & B E A m , J. R. ( 1 9 8 3 ) Measuring Machiavellianism with Mach V: a psychometric investigation. Journal of Personality Assessment, 47, 509-513. 6. ZOOK, A., 11, & SIPPS, G. J. ( i n press) Reliability data and sex differences with a gender-free Mach IV. Journal of Social Psychology. Accepted October 21, 1985.