Do organizations that voluntarily participate in a program differ from non-participating organizations?

The goal of this study was to assess potential participation bias in a program that required voluntary participation of organizations. Evaluations of intervention studies are often limited to information provided by program participants, with little or nothing known about non-participants. Alcohol Risk Management (ARM) Express was a training program designed to encourage alcohol businesses to implement policies to prevent illegal alcohol sales. To assess whether establishment characteristics and practices predicted participation in the program, we evaluated recruitment and establishment characteristics and estimates of illegal alcohol sales rates. Type of establishment was associated with participation in the program, with gas stations, grocery and convenience stores participating more often than other businesses. Other establishment characteristics and rates of illegal alcohol sales were not associated with participation in the program, showing that high-risk alcohol establishments were just as likely as other alcohol establishments to voluntarily participate in an alcohol policy training program.