I’ve been working through Box 8.5 of Nielsen & Chuang’s textbook (Quantum Computation and Quantum Information) related to single-qubit quantum process tomography.
Specifically, in equations (8.175) and (8.176), the output density matrices $\rho_2'$ and $\rho_3'$ are constructed from linear combinations of $\mathcal{E(|+\rangle\langle+|)}$ and $\mathcal{E(|-\rangle\langle-|)}$. However, when I compare the results of those expressions to the definitions given in earlier equations—namely:
$$\rho_2=\rho_1X, \quad \rho_3=X\rho_1$$
I notice that the results from (8.175) and (8.176) actually yield $\rho_3'$ and $\rho_2'$, respectively. In other words, the labels seem to be reversed.
So my question is: Is this a labelling error in the book (i.e., $\rho_2'$ and $\rho_3'$ should be swapped in (8.175)-(8.176))?
Or is there another subtlety in the definitions or assumptions that I’m missing?