Table 4.

Multilevel regression model for male and female friend closeness predicting alcohol consumption 1 year later

Variable	M1	M2	M3
Level 2 (school-level) variables
School mean alcohol consumption	.65**	.65**	.64**
Level 1 (person-level) variables
Age	.02*	.02	.02*
Race	.02*	.02	.02*
Gender	−.03	−.03	−.05
Number of friends nominated	.02*	.02	.02
Percentage of opposite-sex friends	−.01	−.01	−.01
Romantic relationship status	−.04	−.04	−.04
Male friend closeness	.17**	.12	.17*
Female friend closeness	.04	.04	.07*
Alcohol consumption (AC) at Wave I	.45**	.45**	.46**
Average AC for male friends	.06×	.12*	.07×
Average AC for female friends	.03	.03	.01
Gender × male friend closeness	–	.13*	–
Male friend closeness × male     friend AC	–	−.15×	–
Gender × male friend AC	–	−.09**	–
Male friend closeness × male friend     AC × gender	–	.12	–
Gender × female friend closeness	–	–	−.04
Female friend closeness × female     friend AC	–	–	−.16*
Gender × female friend AC	–	–	.03
Female friend closeness × female     friend AC × gender	–	–	.18**

Gender coded as 0 = boys, 1 = girls. Romantic partner status coded as 0 = no romantic partner, 1 = yes romantic partner. Race coded as 1 = White/European American, 2 = Black/African American, 3 = Hispanic/Latino, 4 = Asian American, 5 = Native American, 6 = other/biracial)

^×p < .07;

^*p < .05;

^**p < .01