Table 3.

Effect of predictor variables (settlement, reproductive status, cohabitation, Julian day, and family ID) estimated by the Generalised Additive Model (with Poisson error distribution) fitted to predict the length of inhabitation period by porcupine families in a settlement in Central Italy.

	Estimate	Std. error	z value	Pr( >|t|)
Parametric coefficients
(Intercept)	1.724	0.100	17.202	< 0.001
Settlement (main/secondary)	0.252	0.049	5.122	< 0.001
Reproductive status	0.896	0.056	15.879	< 0.001
Cohabitation	1.360	0.059	23.184	< 0.001

	Edf	Ref.df	F	p-value
Approximate significance of smooth terms
s (Julian day)	1.944	2.000	621.4	< 0.001
s (family)	3.585	4.000	60.9	< 0.001

Coefficients (Estimate) for the linear terms in the model (settlement, reproductive status, cohabitation), standard errors (Std. Error), test statistic and its significance (z value and p-value) are shown in the first part of the Table. Effective degrees of freedom (edf, representing the complexity of the smooth terms), together with test statistic (F), its significance (p-value) and degree of freedom (Ref.df) used to assess overall significance of the smooth terms, are shown in the second section.