In ANCOVA, the distinction between randomized and non-equivalent group designs is important. When participants are randomly assigned to groups, groups should not vary systematically on the covariate, and interpretation is straightforward. However, when intact groups that differ on the covariate are assigned to different treatments, the treatment is confounded with characteristics of the group, and the results of an adjustment may be not be interpretable. Any adjustment based on the covariate will result in adjustments of any correlated characteristics as well, and will result in the kinds of difficulties of interpretation that occur in correlational research.
Huitema, B. E. (1980). The analysis of covariance and alternatives.
New York: Wiley. Myers, J. L., & Well, A. D. (2002). Research design and statistical analysis (2nd ed.). Mahwah, NJ: Erlbaum.
Arnold D. Well University of Massachusetts
See also: Analysis of Variance
Was this article helpful?