High school drinking mediates the relationship between parental monitoring and college drinking: A longitudinal analysis

Table 3 Results of linear regression predicting alcohol consumption in high school ^a (n = 1,100).

	Bivariate Models					Multivariate Model ^c
	b	SE	t ( df )	sr ²	p	b	SE	t ( df )	sr ²	p
Parental Monitoring Score	-.13	.01	-10.49 (1,194)	.08	<.0001	-.12	.01	-9.26 (1,092)	.07	<.0001
Sex [Reference = Female]	.97	.16	6.14 (1,249)	.03	<.0001	.69	.16	4.32 (1,092)	.01	<.0001
Race [Reference = Non-White]	1.37	.17	7.91 (1,246)	.05	<.0001	1.26	.18	7.02 (1,092)	.04	<.0001
Religiosity^b [Reference = Slightly/Not Important]	-.56	.16	-3.46 (1,243)	.01	.0006	-.10	.16	-.59 (1,092)	<.01	.56
R ²						.14
*F (df, df) p*						24.82 (7, 1,092) p < .0001

Effects were evaluated using the null hypothesis test of b = 0 (tested as: b/SE) which evaluates the unique contribution of a variable in a regression equation.
^a High school alcohol consumption was defined as the typical number of drinks per drinking day during the past year at the screener.
^b Religiosity was dichotomized into a binary variable (i.e., extremely/moderately vs. slightly/not).
^c As a proxy for socioeconomic status, the effect of mother's education was held constant in the multivariate model. Effect size (sr²) for each explanatory variable was as follows: parental monitoring score (.07), sex (.01), race (.04), religiosity (<.01), mother's education (<.01).

ISSN: 1747-597X