Table 4.

Coefficients and model fit for the multilevel model in Experiment 6.

Fixed components (Intercept)	γ [95% CI]	SE γ	P
(Intercept)	2.57 [2.54, 2.60]	0.02	<0.001
Moral domain	0.97 [0.94, 1.01]	0.02	<0.001
Legality	−1.00 [−1.03, −0.98]	0.01	<0.001
Religious salience	0.03 [−0.03, 0.08]	0.03	0.408
Religiosity	−0.21 [−0.24, −0.19]	0.01	<0.001
Religious salience × Religiosity	−0.04 [−0.08, 0.01]	0.02	0.101
Moral domain × Religious salience	−0.09 [−0.15, −0.02]	0.03	0.009
Moral domain × Religiosity	−0.35 [−0.37, −0.32]	0.01	<0.001
Moral domain × Religious salience × Religiosity	−0.07 [−0.11, −0.02]	0.02	0.007
Legality × Religious salience	0.05 [0.003, 0.10]	0.03	0.037
Legality × Religiosity	0.17 [0.15, 0.19]	0.01	<0.001
Legality × Religious salience × Religiosity	0.05 [0.01, 0.09]	0.02	0.007
Random components	σ2	SD σ2	SE σ2
Subject intercepts	0.44	0.66	0.25
Moral domain	0.31	0.55	0.35
Legality	0.08	0.28	0.18
Residual	2.43	1.56	-
Intraclass correlations	ρ
Subject	0.15
Model R2	R2
Cox-Snelling Psudeo-R2	0.36

Note. σ [95% CI] = Unstandardized coefficients and 95% confidence intervals for fixed effects. SE γ = standard errors of fixed effects, p = p-values for fixed effects, σ² = variance estimates for the random components. SD σ² = standard deviation of the variances estimates. SE σ² standard errors of the variance components, ρ = intraclass correlation coefficient. R² = Cox-Snelling Psudeo-R² for global model fit. Fixed effects, random effects, and model R² were estimated using the lme4 package, and random effect standard errors were computed using the arm package (Gelman and Su, 2018). We omitted p-values for the variance components, because these estimates are often incorrect (Hox, Moerbeek, & Van de Schoot, 2010).