Summary

In general, we’ve seen how the ANCOVA of designed experiments combines one categorical variable and one continuous explanatory variable. Thanks to the concept of general linear models, this mix of regression and ANOVA can be performed with the lm() function just like insimpler individual analyses.

The interpretation of the ANOVA and summary tables can get tricky. The equivalence of the F-test and t-test in these tables sometimes breaks down because the same comparisons aren’t being performed, despite what the row labels in the table sometimes imply.