Table 2.

Summary of results of generalized linear mixed models indicating effect of patch (female number + vegetation density) and fish id (random factor) on (a) proportion of visits (I_visit), (b) proportion of time spent (I_time), in E1 set.

E1 set: (a) (H + 8F, M + 4F, L + 2F, N + 1F)
Null model: I_visit ~ (1|FishID) AIC: − 64.69
Selected model: I_visit ~ Patch + (1|FishID) AIC: − 86.05
Variable	Estimate	z-score	df	p
Intercept	0.35	0.03	12.04	≪ 0.01
Patch H	− 0.05	0.04	− 1.22	0.22
Patch L	− 0.18	0.04	− 4.35	≪ 0.01
Patch N	− 0.19	0.04	− 4.63	≪ 0.01

E1 set: (b) (H + 8F, M + 4F, L + 2F, N + 1F)
Null model: I_time ~ (1|FishID) AIC: − 5.42
Selected model: I_time ~ Patch + (1|FishID) AIC: − 24.25
Variable	Estimate	z-score	df	p
Intercept	0.42	0.04	10.48	≪ 0.01
Patch H	− 0.18	0.05	− 3.18	≪ 0.01
Patch L	− 0.24	0.05	− 4.29	0.01
Patch N	− 0.28	0.05	− 4.86	≪ 0.01