17.1 Why adjust for baseline?
When a continuous outcome is measured at baseline and follow-up, ANCOVA compares treatment groups at follow-up while adjusting for the baseline measurement. This can improve precision compared with analyzing change scores alone (Vickers and Altman 2001).
A basic ANCOVA model is
\[ Y_{12}=\beta_0+\beta_1Y_0+\beta_2I(A)+\beta_3I(B)+\varepsilon. \]
Here, \(\beta_2\) and \(\beta_3\) represent adjusted treatment differences relative to Control.
Before interpreting these as constant treatment effects, check whether the treatment effect changes with the baseline value by adding a baseline-by-treatment interaction.