Table 2. Information on measured covariate balance and treatment effect estimates for the Average Treatment on the Treated Propensity Score Techniques.

Propensity Score Technique (estimation , application)	Number of covariates with standardized bias < 0.25	Maximum standardized bias	Name of covariate with maximum standardized bias	Treatment Effect Estimate	95% Confidence Interval
					Low	High
Mode l 1 , 1:1	2	0.50	alcohol problems	1.17	0.69	1.99
Model 2 , 1:1	2	0.55	alcohol problems	1.54	0.87	2.74
Model 3 , 1:1	2	0.53	alcohol problems	1.34	0.78	2.31
Model 1 , Full	3	0.32	free lunch	1.43	0.90	2.26
Model 2 , Full	10	0.50	parental monitoring	1.70	1.07	2.70
Model 3 , Full	2	0.39	free lunch	1.79	1.14	2.82
Model 1 , Odds	3	0.30	free lunch	1.14	0.66	1.95
Model 2 , Odds	7	0.32	free lunch	1.12	0.63	1.99
Model 3 , Odds	0	0.21	intervention in school	1.33	0.76	2.33

Estimation Methods = Model 1: Multivariable Logistic Regression, Model 2: Multivariable Logistic Regression with product-terms, Model 3: Generalized Boosted Modeling Average Treatment on the Treated (ATT) application methods: 1:1 = nearest neighbor one to one matching, Full = full matching, Odds = weighting by the odds