A simple new method to triangulate self-reported risk behavior data--the bean method.

Background: Human immunodeficiency virus (HIV)/sexually transmitted disease (STD)-related interventions rely on self-reported risk behavioral data, which are biased toward social desirability. It is warranted to develop methods for data triangulation.

Methods: Some questions on HIV/STD-related behaviors that were asked in 5 face-to-face (FTF) surveys were repeated at the end of the surveys, using the newly developed bean method (BM), which requested participants to transfer a bean of a particular color from a large jar to a smaller one, depending on whether the answer was affirmative (yellow, white, green, or red) or not (black); both jars contained 500 mixed-up colored beans. At the end of each day, the number of beans for each color was counted and was compared with the results of the FTF interviews. Participants were ensured that the researchers would not trace their responses.

Results: Among female sex workers who were noninjecting drug users, the BM as compared with the FTF interviews, reported higher prevalence of inconsistent condom use in the last 6 months, unprotected sex with the last client, and STD in the last 6 months. Among men who have sex with men in Hong Kong, the BM as compared to the FTF interviews, reported higher prevalence of unprotected sex with another man and having >3 sex partners in the last 12 months. No statistically significant between-method differences were reported among injecting drug users.

Conclusions: Results of HIV/STD-related behaviors are dependent on mode of data collection. The simple BM can be used regularly to triangulate self-reported HIV-related behavioral data obtained from FTF interviews.