Table 2.

A logistic regression modelling whether the demonstrators chose their demonstrator optimum. It includes a dummy for the final period of the blocks and % as optimal as predictors. Model 1 displays the data for the game against nature (asocial skills) and Model 2 displays the data for the coordination game (social norms). Robust standard errors in parentheses are clustered on the demonstrator. We also include the lower and upper limit 95% confidence interval for each estimate below this standard error.

Parameter	Estimate (game against nature)	Estimate (coordination game)
Intercept	0.321 * (0.161) 95% CI [0.01, 0.64]	−0.130 (0.154) 95% CI [−0.43, 0.17]
Final period dummy	0.378 *** (0.091) 95% CI [0.20, 0.56]	0.325 *** (0.092) 95% CI [0.15, 0.51]
% as optimal	0.178 (0.290) 95% CI [−0.39, 0.75]	1.146 *** (0.260) 95% CI [0.64, 1.66]

Asterisks denote the level of significance of our p-values, with the following key: *** p < 0.001; ** p < 0.01; * p < 0.05; trend, p = 0.05–0.10 significance.