Table 4.

Comparison of Standardized Biases After Matching: Male Participants

Variable	All male participants	Method 1	Method 2	Method 3	Method 4	Method 5
Propensity score	0.44	0.08		0.06	−0.01	0.08
Maternal substance use history	−0.12	0.08		0.12	−0.01	−0.02
Maternal years of schooling	−0.05	−0.01		0.16	0.04	0.06
Family income	−0.18	−0.11		0.20	−0.12	−0.04
Below poverty threshold	0.14	0.16		0.14	0.15	0.05
Teacher rating of underachievement	−0.03	−0.06		0.14	−0.16	−0.10
Teacher rating of aggression	0.24	0.00		0.13	−0.04	−0.02
Teacher rating of shyness	−0.09	0.01		0.08	−0.03	−0.02
Teacher rating of immaturity	0.07	0.05		0.10	−0.01	0.03
Teacher rating of inattention	0.13	0.00		0.13	−0.06	−0.04
Precision relative to a 1:1 match		1.00		0.95	1.15	0.95
N	216	156		216	216	216

Note. Standardized bias is defined as the weighted difference in means divided by the standard deviation in the full group of participants who were not heavy marijuana users. Numbers in bold indicate standardized biases that decreased in comparison with the original sample. Methods are as follows: 1 = 1:1 optimal nearest neighbor matching; 2 = 2:1 optimal nearest neighbor matching (could not be done for male participants because there were not twice as many male participants who were not heavy marijuana users as there were male participants who were heavy users); 3 = six subclasses; 4 = full matching; 5 = constrained full matching.