Studies in the Design and Analysis of Dental Experiments

8. Application of Analysis of Covariance to Dental Research

¹ School of Public Health and Administrative Medicine, Faculty of Medicine, Columbia University, New York, N.Y.

When comparing various treatments a frequent design allocates the individuals, at random, to the several treatment groups. The measurements subsequently made are then compared by some application of the t-test or the analysis of variance. Random assignment of individuals equalizes the groups within chance with respect to initial factors, including concomitant variables to which the measurement in question may be related. An adjustment for inequalities in such a concomitant variable can be made, however, by an application of correlation methods referred to as the analysis of covariance.

The analysis of covariance is illustrated by comparing various analgesic agents with respect to pain threshold, as measured by the pulp-test reading, using pulse rate as the concomitant variable to be adjusted.