Table 2.
Estimates with SE from beta regressions for the percentage of matrices claimed as correct.
| Predictor | Model 1 | Model 2 | Model 3 | Model 4 | Model 5 |
|---|---|---|---|---|---|
| Intercept | 29.84 (3.61)*** | 30.32 (6.27)*** | 30.27 (6.19)*** | 31.18 (6.78)*** | 29.86 (6.49)*** |
| Mauritius | 11.32 (4.18)** | 11.10 (4.44)* | 9.28 (4.21) * | 7.93 (4.77)Ϯ | 9.27 (5.19)Ϯ |
| Czech Republic | −9.61 (3.34)** | −9.63 (3.44)** | −9.38 (3.39)* | −10.50 (3.48)** | −9.75 (4.20)* |
| Secular | 3.04 (3.93) | 2.45 (3.92) | 3.48 (3.95) | 2.74 (3.97) | 3.03 (3.93) |
| Control | 5.99 (4.07) | 5.40 (4.05) | 6.01 (4.06) | 6.19 (4.16) | 7.60 (4.38)Ϯ |
| Religiosity | −5.31 (2.40)* | −4.97 (2.48)* | −4.97 (2.43)* | ||
| Secular*Religiosity | 7.55 (3.37)* | 7.54 (3.45)* | 7.32 (3.37)* | ||
| Control*Religiosity | 6.60 (3.18)* | 6.49 (3.28)* | 6.26 (3.18)* | ||
| Ritual | −1.55 (1.47) | ||||
| Secular* Ritual | 5.40 (2.12)* | ||||
| Control* Ritual | 3.54 (2.06)Ϯ | ||||
| Females vs. Males | 7.90 (3.50)* | 8.47 (3.48)* | |||
| Age | 0.04 (0.20) | 0.04 (0.20) | |||
| Positivity | −2.16 (2.24) | ||||
| Negativity | −2.70 (2.12) | ||||
| Tempo | −1.61 (1.84) | ||||
| Impact | 1.99 (1.86) | ||||
| Cox-Snell R2 | 0.124 | 0.147 | 0.157 | 0.166 | 0.175 |
In all models, we control for the effects of site. The religious condition and the USA site were set as reference categories (intercept). The first model contains only the effects of condition (compared to the religious condition) while controlling for the effects of site. The second model includes a Condition*Religiosity interaction, describing the effects of religiosity on cheating in the religious condition. The two predictors specified as interactions (Secular*Religiosity and Control*Religiosity) are comparisons with this effect. Again, we control for site. The third model has an identical design to the second, only with a Condition*Ritual participation interaction. Since the effects of ritual participation on morality were not as strong as those of religiosity, we retained the latter factor for subsequent models. The fourth model contains site and condition effects, the significant interaction, and demographic covariates. The fifth model controls also for different characteristics of our musical stimuli.
p < 0.1;
p < 0.05;
p < 0.01;
p < 0.001.