Table 1:

Parameters used in the fitted $\left(\widehat{\alpha }\right)$ and true (α_t) models for the two cases.

Model	Fixed effects	Random intercept variance	Random time variance	Random treatment variance	Residual variance
Time-fitted random treatment
Fitted model	$\widehat{\Phi }$	$\widehat{{\tau }^2}$	$\widehat{{\gamma }^2}$	0	$\widehat{{\sigma }^2}$
True model	Φ t	${\tau }_t^2$	0	${\eta }_t^2$	${\sigma }_t^2$
Treatment-fitted random time
Fitted model	$\widehat{\Phi }$	$\widehat{{\tau }^2}$	0	$\widehat{{\eta }^2}$	$\widehat{{\sigma }^2}$
True model	Φ t	${\tau }_t^2$	${\gamma }_t^2$	0	${\sigma }_t^2$