A draft of Bailey's Design of Comparative Experiments may be found here; I am referring to Chapter 2, figure 2.1.

Consider $t$ treatments. We have $N$ plots labeled $1, \dots, N$. We apply treatment $1$ to plots $1, \dots, r_1$; $2$ to plots $r_1 + 1, \dots, r_1 + r_2$, and so forth.

Let $\Omega = \{1, \dots, N\}$ be the set of plots and let $\mathcal{T}$ be the set of treatments, for which $|\mathcal{T}| = t$. We define $T: \Omega \to \mathcal{T}$ as the treatment factor. The treatment subspace $V_T$ consists of the vectors in $\mathbb{R}^{\Omega}$ which are constant on each treatment.

Bailey's text refers to the following figure:

enter image description here

What is the relevance of $V_T$? Usually when running an experiment, I would have the responses $\mathbf{y}$, the treatments $T$, and the covariates (of which there are none here), so I don't quite understand what $V_T$ is. The $\mathbf{u}_i$ seem to be nothing more than vectors of treatment indicators for $A$, $B$, and $C$.

(I believe you can ignore what the "typical $\mathbf{v}$" row is without any issues.)

  • $\begingroup$ Chapter 3 gives some examples of where $V_T$ and its decomposition into subspaces is of interest $\endgroup$ Feb 2 at 1:30


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Browse other questions tagged or ask your own question.