Table 1.

Unstandardized estimates (SE in brackets) for the contextual effect (studies 1a–1e) and the contextual effect controlling for norms (studies 1a, 1d, and 1e)

	Study 1a†		Study 1b		Study 1c		Study 1d		Study 1e‡
	β (SE)	P	β (SE)	P	β (SE)	P	β (SE)	P	β (SE)	P
Within-level effect	−0.189 (0.009)	<0.001	−0.351 (0.023)	<0.001	−0.082 (0.039)	0.035	−0.555 (0.101)	<0.001	−0.21 (0.05)	<0.001
Between-level effect	−0.738 (0.099)	<0.001	−0.663 (0.062)	<0.001	−0.416 (0.162)	0.010	−1.465 (0.342)	<0.001	−0.47 (0.12)	<0.001
Contextual effect	−0.549 (0.101)	<0.001	−0.311 (0.073)	<0.001	−0.334 (0.177)	0.059	−0.910 (0.377)	0.016	−0.25 (0.14)	<0.001
Effect size of contextual effect	0.33		0.22		0.35		0.34		0.21
Contextual effect without controls	−0.495 (0.102)	<0.001	−0.282 (0.069)	<0.001	−0.270 (0.166)	0.104	−0.929 (0.379)	0.014	−0.226 (0.131)	0.085
Contextual effect controlling for norms	−0.135 (0.084)	0.106					0.33 (0.37)	0.37	0.42 (0.27)	0.542
Indirect effect of context§	−0.418 (0.097)	<0.001					−1.39 (0.44)	0.002	−0.71 (0.23)	0.010

^†All estimates are controlled for between-country differences in variables.

^‡The contextual effect for all groups is reported; due to sample sizes, further differentiation by ethnic group was not possible.

^§The indirect effect reflects the effect of contact on prejudice via norms (assessed as diversity beliefs) on the social context level, and indicates whether the contextual effect is significantly reduced after controlling for norms on the social context level.