BACKGROUND: The Internet has become an increasingly popular venue for men who have sex with men (MSM) to meet potential sex partners. Given this rapid increase in online sex-seeking among MSM, Internet-based interventions represent an important HIV (human immunodeficiency virus) prevention strategy. Unfortunately, black and Hispanic MSM, who are disproportionately impacted by the HIV epidemic in the United States, have been underrepresented in online research studies. OBJECTIVE: Our objective was to examine and quantify factors associated with underrecruitment and underretention of MSM of color in an online HIV behavioral risk research study of MSM recruited from an online social networking site. METHODS: Internet-using MSM were recruited through banner advertisements on MySpace.com targeted at men who reported in their MySpace profile their age as at least 18 and their sexual orientation as gay, bisexual, or unsure. Multivariable logistic regression models were used to estimate the odds stratified by race and ethnicity of the MySpace user clicking through the banner advertisement. To characterize survey retention, Kaplan-Meier survival curves and multivariable Cox proportional hazards models identified factors associated with survey dropout. RESULTS: Over 30,000 MySpace users clicked on the study banner advertisements (click-through rate of 0.37%, or 30,599 clicks from 8,257,271 impressions). Black (0.36% or 6474 clicks from 1,785,088 impressions) and Hispanic (0.35% or 8873 clicks from 2,510,434 impressions) MySpace users had a lower click-through rate compared with white (0.48% or 6995 clicks from 1,464,262 impressions) MySpace users. However, black men had increased odds of click-through for advertisements displaying a black model versus a white model (adjusted odds ratio [OR] = 1.83, 95% confidence interval [CI] 1.72 - 1.95), and Hispanic participants had increased odds of click-through when shown an advertisement displaying an Asian model versus a white model (adjusted OR = 1.70, 95% CI 1.62 - 1.79). Of the 9005 men who consented to participate, 6258 (69%) completed the entire survey. Among participants reporting only male sex partners, black non-Hispanic and Hispanic participants were significantly more likely to drop out of the survey relative to white non-Hispanic participants (hazard ratio [HR] = 1.6, 95% CI 1.4 - 1.8 and HR = 1.3, 95% CI 1.1 - 1.4, respectively). Men with a college-level of education were more likely to complete the survey than those with a high-school level of education (HR = 0.8, 95% CI 0.7 - 0.9), while men who self-identified as heterosexual were more likely to drop out of the survey compared with men who self-identified as gay (HR = 2.1, 95% CI 1.1 - 3.7). CONCLUSIONS: This analysis identified several factors associated with recruitment and retention of MSM in an online survey. Differential click-through rates and increased survey dropout by MSM of color indicate that methods to recruit and retain black and Hispanic MSM in Internet-based research studies are paramount. Although targeting banner advertisements to MSM of color by changing the racial/ethnic composition of the advertisements may increase click-through, decreasing attrition of these study participants once they are engaged in the survey remains a challenge.