Relationships between central tendency error and estimates of reliability based on Cronbach's coefficient alpha were explored. Using a Monte Carlo simulation, no differences were found between the means of distributions of alphas derived from “central tendency biased data” and “random data.” Further it was found that alphas' constancy was the result of systematic reduction in the item standard deviations and increase in inter-item correlations associated with central tendency error. It was noted, however, that in this simulation “equal error” among the test items was assumed; further research with other assumptions is recommended.
Get full access to this article
View all access options for this article.
References
1.
CronbachL.Coefficient alpha and the internal structure of tests. Psycbometrika, 1951, 16, 297–334.
2.
CuretonE.The definition and estimation of test reliability. Educational and Psychological Measurement, 1958, 18, 715–738.
3.
GuilfordJ.Psychometric methods. New York: McGraw-Hill, 1954.
4.
NunnallyJ.Psychometric theory. New York: McGraw-Hill, 1967.
5.
Rand Corp. A million random digits with 100,000 normal deviates. Glencoe, Ill.: Free Press, 1955.
