Table 4.
Estimates with SE from beta regressions for the percentage of matrices claimed as correct across our three sites.
| Mauritius | Czech Republic | USA | ||||
|---|---|---|---|---|---|---|
| Model A | Model B | Model A | Model B | Model A | Model B | |
| Intercept | 33.89 (5.57)** | 35.49 (15.24)** | 20.20 (3.57)*** | 21.70 (3.50) *** | 33.82 (5.09)** | 32.46 (9.78)** |
| Secular | 16.66 (8.10)* | 17.13 (11.33) | −2.54 (4.43) | −4.29 (4.46) | −3.17 (6.74) | 0.49 (6.77) |
| Control | 16.34 (7.95)* | 13.83 (8.27)Ϯ | 0.93 (4.88) | 0.23 (4.83) | 1.01 (7.03) | 4.48 (6.97) |
| Religiosity | −4.72 (5.23) | −4.40 (3.22) | −5.04 (3.60) | |||
| Secular*Religiosity | 2.41 (9.87) | 1.58 (4.20) | 9.91 (5.36)Ϯ | |||
| Control*Religiosity | 7.58 (7.52) | −0.57 (3.81) | 9.89 (5.01)Ϯ | |||
| Cox-Snell R2 | 0.071 | 0.085 | 0.008 | 0.077 | 0.005 | 0.068 |
Models A describe condition effects for the three sites: Mauritius, the Czech Republic, and the USA. Models B display a Condition*Religiosity interaction for each site. In all models, the religious condition was set as a reference category.
Ϯ
p < 0.1;
*
p < 0.05;
**
p < 0.01;
***
p < 0.001.