Table 2.

Proportion of animals reported as ‘responding’ to CMS in relation to estimates of the overall reliability of the model.

	n	Proportion responding
		25-50%	50-75%	>75%
Effects sometimes absent, or present but nonsignificant	9	2	6	1
Sucrose intake less reliable than other measures	6	0	4	2
CMS produces a reliable depressive phenotype	47	7	13	27

This table omits some labs that did not supply data on the proportion of responders, which includes the three labs where CMS was found to be ‘unreliable’.