The second question addressed by this investigation was whether partner
type moderated the relationship between intervention condition (treatment/
no treatment) and condom use. Regression analyses were used with
percent condom use serving as the criterion variable. Given that preexisting
demographic differences were controlled for, condition and partner type
were entered in block 1, and the intersection of condition and partner type
was entered in the second block. Table 3 presents the regression analysis,
which lacked significant results.