Condom use as a dependent variable: measurement issues relevant to HIV prevention programs.

The value of condoms in efforts to slow the spread of HIV infection has been well established in the literature. Behavioral science faces the challenge of promoting condom use through intervention programs. As these programs are evaluated, multiple issues should be considered in relation to measuring participant use of condoms for the purposes of preventing HIV infection. Lack of attention to these issues is likely to create a large number of Type I and Type II errors. Ten common sources of error are described and corresponding recommendations for eliminating these errors are offered. A review of published studies shows that there is little consistency relevant to controlling for these sources of error. Incorporation of standardized methodology will allow for more accurate program evaluation and benefit researchers by facilitating comparisons across studies.