Research conducted via the Internet has the potential to reach important clinical populations of participants who would not participate in traditional studies. Concerns exist, however, about the validity of samples recruited in this manner, especially when participants are anonymous and never have contact with study staff. This study evaluated two anonymous samples that were recruited over the Internet to test an online program designed to help problem drinkers. The two studies were conducted 3 years apart, and different recruitment strategies were utilized. Despite these differences, the two samples were highly similar in demographic and clinical features. Correlations that have been found between variables in traditional non-anonymous studies were also found in both online samples, supporting the validity of the data that was collected. Appropriate skepticism is required when critically evaluating Internet studies. Nevertheless, the results of this study indicate that it is possible to obtain stable, valid data from anonymous participants over the Internet, even when there are significant differences in the way the participants are obtained.