. Author manuscript; available in PMC: 2015 Aug 6.

Published in final edited form as: Addiction. 2014 Sep 24;110(1):59–68. doi: 10.1111/add.12706

Table 2.

Interaction effects^a between policy domain efficacies and implementation trend, 1999-2011.

Models	Predictors	Beta (95% CIs)	P-value
Model I (efficacy on binge drinking in general population)	Efficacy on binge drinking in general population	0.012 (–0.003, 0.028)	0.110
	Implementation trend	0.013 (0.010, 0.015)	<0.001
	Efficacy × trend interaction	–0.003 (–0.004, –0.002)	<0.001
Model II (efficacy on binge drinking in youth)	Efficacy on binge drinking in youth population	0.044 (0.019, 0.068)	0.001
	Implementation trend	0.019 (0.015, 0.023)	<0.001
	Efficacy × trend interaction	–0.005 (–0.006, –0.004)	<0.001
Model III (efficacy on impaired driving in general population)	Efficacy on alcohol-impaired driving in general population	0.021 (0.007, 0.035)	0.003
	Implementation trend	0.006 (0.003, 0.009)	<0.001
	Efficacy × trend interaction	–0.001 (–0.001, 0.001)	0.781
Model IV (efficacy on impaired driving in youth)	Efficacy on alcohol-impaired driving in general population	0.063 (0.038, 0.088)	<0.001
	Implementation trend	0.009 (0.005, 0.012)	<0.001
	Efficacy × trend interaction	–0.001 (–0.002, –0.001)	0.048

Regression models predicting the outcome of implementation ratings were based on individual state-policy-year observations (n = 51 states and DC × 29 policies × 13 years = 19 227). CI = confidence interval.