Table 4.

Generalized linear model for the probability of cooperation depending on the previous action of the focal player and the cooperation level in the neighbourhood in both treatments (shown in figure 6) $(\text{lm}(\text{formula}=\text{action}\sim \text{info}+N\mathrm{\_}\mathrm{coop}+\text{prev}\mathrm{\_}\mathrm{action}))$. (The variables are all significant: (i) ‘info’ is once again confirming the difference of behaviour between the two treatments with (TWI) and without (TWO) information, (ii) ‘N_coop’ is the number of cooperating neighbours of the focal player in the previous round, and (iii) ‘prev_action’ is the action of the focal player in the previous round.)

variable	estimate	s.e.	z-value	$\text{Pr}(|z|)$
(intercept)	0.215331	0.012826	16.789	<2 × 10−16***
info	−0.037874	0.010861	−3.487	0.000491***
N_coop	0.027424	0.005606	4.892	1.02 × 10−06***
prev_ action	0.352350	0.011089	31.775	<2 × 10−16***