ANCOVA for Randomized‐Block Designs

The typical way of characterizing the comparison of analysis of covariance (ANCOVA) with “blocking” confounds two issues that should be kept separate: design and analysis. The actual comparison that should be articulated is the analysis of the randomized-group design using ANCOVA versus the analysis of the randomized-block design using randomized-block (RB) ANOVA. The failure to distinguish the design from the analysis has led most researchers believe that ANCOVA does not apply to the RB design. This chapter demonstrates that it is desirable to analyze the typical RB design using either (1) conventional ANCOVA or (2) a more complex analysis that includes a covariate. An approach, described as a combined ANCOVA and block analysis is recommended for the case where one variable is used for block formation and a second variable is used as a covariate. Controlled Vocabulary Terms analysis of covariance; randomized block design