Seroadaptive behaviors are traditionally defined by self-reported sexual behavior history, regardless of whether they reflect purposely-adopted risk-mitigation strategies. Among MSM attending an STD clinic in Seattle, Washington 2013-2015 (N = 3751 visits), we used two seroadaptive behavior measures: (1) sexual behavior history reported via clinical computer-assisted self-interview (CASI) (behavioral definition); (2) purposely-adopted risk-reduction behaviors reported via research CASI (purposely-adopted definition). Pure serosorting (i.e. only HIV-concordant partners) was the most common behavior, reported (behavioral and purposely-adopted definition) by HIV-negative respondents at 43% and 60% of visits, respectively (kappa = 0.24; fair agreement) and by HIV-positive MSM at 30 and 34% (kappa = 0.25; fair agreement). Agreement of the two definitions was highest for consistent condom use [HIV-negative men (kappa = 0.72), HIV-positive men (kappa = 0.57)]. Overall HIV test positivity was 1.4 but 0.9% for pure serosorters. The two methods of operationalizing behaviors result in different estimates, thus the choice of which to employ should depend on the motivation for ascertaining behavioral information.