“…Questions about alcohol and illicit drug use are likely to be sensitive, because they are associated with a strong social desirability response bias (Duffy & Waterton, 1984). Previously, many different survey methodshave been used to measure sensitive behaviors,such as self-administeredpaperquestionnaires (Wright, Aquilino, & Supple, 1998), face-to-face interviews (Konings, Bantebya, Caraël, Bagenda, & Mertens, 1995), computerassisted personal interviews (Tourangeau & Smith, 1996), computer-assisted telephone interviews (CATIs; Aquilino & Lo Sciuto, 1990), computer-assisted self-interviews (Bonevski, Sanson-Fisher, Campbell, & Ireland, 1997;Millstein & Irwin, 1983), and audio computer-assisted self-interviews (Tourangeau & Smith, 1996;Turner et al, 1998). Those methods that involve self-administration and greater anonymity have been found to produce higher reporting rates on sensitive issues (Kobak et al, 1997;Tourangeau & Smith, 1996).…”