Table 5. Binary logistic regression equation employed to predict the outcome of stereotypic behaviour (0 = absence and 1 presence) and FCM levels (0 = low and 1 high) in tigers and leopards using the biological and environmental variables.

Species	Included	B (SE)	95% CI for odds ratio
			Lower	Odds ratio	Upper
Tiger Stereotype	Constant	3.777 (1.389)	-	43.67	-
	Pool presence	-2.594* (1.462)	0.004	0.075	1.311
	Stone presence	3.386* (1.769)	0.906	29.028	929.813
	Keeper attitude—positive		-	-	-
	Keeper attitude—neutral	-2.553* (0.891)	0.014	0.078	0.446
	Keeper attitude—negative	17.470* (1.797E4)	0.000	3.865E7	-
	Enclosure—large	-3.412* (1.214)	0.003	0.033	0.356
	Sociality—group	-1.854* (0.672)	0.042	0.157	0.585
	R2 = 0.43 (Hosmer and Lemeshow), 0.249 (Cox & Snell), 0.404 (Nagelkerke); Model χ2 (6) = 32.630, p<0.001; * significant at p < 0.05
Leopard Stereotype	Constant	-27.226 (2.556E3)	-	-	-
	Origin—wild	170.979* (1.534E4)	0.000	1.799E74	-
	Pool—absence	55.838* (5.112E3)	0.000	1.778E24	-
	Den—presence	-147.013* (1.739E4)	0.000	0.000	-
	Age	-14.306* (1.278E3)	0.000	0.000	-
	Tree—low %	48.257* (1.329E4)	0.000	9.073E20	-
	R2 = 0.81 (Hosmer and Lemeshow), 0.615 (Cox & Snell), 0.824 (Nagelkerke); Model χ2 (5) = 47.771, p<0.001; * significant at p < 0.05
Tiger FCM	Constant	20.180 (1.704E4)	-	-	-
	Health problem—nil	-		5.807E8	-
	Health problem—low	-40.653 * (2.112E4)	0.000	0.000	-
	Health problem—high	-62.755* (2.473E4)	0.000	0.000	-
	Enclosure—small	21.572* (1.248E4)	0.000	2.338E9	-
	R2 = 0.64 (Hosmer and Lemeshow), 0.649 (Cox & Snell), 0.909 (Nagelkerke); Model χ2 (3) = 23.023, p<0.001; * significant at p < 0.05
Leopard FCM	Constant	-21.203* (1.340E4)		0.000	-
	Stone—absence	20.797* (1.340E4)	0.000	1.077E9	-
	Keeper attitude—neutral	21.608* (2.842E4)	0.000	2.423E9	-
	R2 = 0.94 (Hosmer and Lemeshow), 0.675 (Cox & Snell), 0.95 (Nagelkerke); Model χ2 (3) = 17.995, p<0.001; * significant at p < 0.05