A Comparison of Analysis of Covariance and Blocking Designs in a Gender and Science Context.

Data from a distance education project, Integrated Science 7 (IS7), were used to compare block designs with analysis of covariance (ANCOVA) for their ability to increase statistical power. The IS7 program enables students in grades six through eight to study science via satellite. A sample of 1,802 students from a pilot of the program yielded data for the study. The treatment by blocks designed were formed using the posttest aptitude scores as the dependent variable, the pretest scores as the concomitant or blocking variable, and the gender of the subjects of the independent variable. Two-block, 5-block, and 10-block designs were compared to 2 ANCOVA analyses. With the data used, the 10-block design appeared preferable to ANCOVA, but overall results suggest that there is no one optimal method. The use of ANCOVA versus blocking is dependent on several conditions, as discussed. Eleven tables present data from the analyses. Contains 49 references. (SLD)