People tend to rate themselves as being better than average, especially when they are being considered by potential partners. We present large-scale validation for social comparative biases using data from OkCupid, one of the largest online dating websites in the US. We find strong evidence that self-enhancing biases in self-judgments for some traits, such as self-confidence and maturity, vary by age and geographic region. Furthermore, we find that biases for other traits, such as height and appreciativeness, remain constant. Our results also suggest that susceptibility to self-enhancement bias interacts with varying levels of online dating success and behaviors as measured by message interactions between users. Our work replicates previous social comparison findings on a massive scale, examines susceptibility to bias across a diverse demographic sample, and shows the effects of biases on behavioral outcome variables.