Table 1.

Parameter estimates from the best-fit, most parsimonious generalized linear model (binomial errors, log link) of annual deer persistence (number of months of deer presence, out of 36), modelled against landscape features.

Parameter	β estimate	std. error	t-value	p-value
(Intercept)	−0.16	0.10	−1.60	0.1090
Upland deciduous forest	0.03	0.00	7.07	<0.0001
Forest harvest blocks	0.07	0.01	7.76	<0.0001
Petroleum well sites	0.24	0.04	5.55	<0.0001
All trails (unpaved roads)	−0.88	0.17	−5.24	<0.0001
Industrial block features	0.41	0.06	7.37	<0.0001

*Null deviance = 465.4 on 60 df; residual deviance = 191.26 on 55 df; AIC = 397.85.