Table 2.

Analysis of 20 strains x ethanol treatment effects on 8 tests ^a

	Strain effect (df =19)			Ethanol effect (df =1)			Strain x Ethanol interaction (df =19)
Measure	F	P	est. ω2	t	P	d	F	P
Slips on balance beam 1.25 g/kgb	2.95	.0001	0.11	NM	b	1.14	NM	b
FSRR 1.25 g/kg fall latency (sec)	3.05	.00003	0.12	8.42	<.00001	0.96	0.7	NS
FSRR 1.75 g/kg fall latency (sec)	2.19	.003	0.07	8.54	<.00001	0.98	1.39	NS
ARR 1.25 g/kg fall latency (sec)	6.36	<.00001	0.26	<1	NS	NM	<1	NS
ARR 1.75 g/kg fall latency (sec)	5.97	<.00001	0.25	6.00	<.00001	0.69	2.00	.009
Open field distance (m) 1.75 g/kg	20.26	<.00001	0.56	6.38	<.00001	0.73	4.25	<.00001
Grip force (g) 1.75 g/kg	4.09	<.00001	0.17	18.0	<.00001	2.04	<1	NS
Observer rated ataxia 3.0 g/kgb	10.27	<.00001	0.38	NM	b	3.07	NM	b
Loss of rectal temp (°C) 3.0 g/kgb	7.94	<.00001	0.31	NM	b	2.11	NM	b

^aNM indicates computation was not meaningful. NS indicates effect was not significant at P < .05. Estimated effect size ω² is partial ω² computed from the F ratio for the strain effect. Effect size d for Ethanol versus Saline expresses standard deviations by which means differ, where standard deviation (SD) is based on MS_within from the ANOVA or within-strain SD for three variables denoted with b. Degrees of freedom within groups ranged from 265 to 271 for the various measures.

^bFor the balance beam, slip rates were so low under saline that within-group variances were close to zero and the strain x ethanol ANOVA was not meaningful. Likewise for observer rated ataxia and hypothermia, where all mice received ethanol and effects were compared with baseline measures, a two-way ANOVA was not possible. For these three measures, a one-way ANOVA to assess the strain effect was essentially a test of strain by ethanol interaction, where there was no doubt at all about the statistical significance of the ethanol effect.