While response rates on email-solicited Internet surveys are often low [10], a low response rate does not necessarily produced biased estimates [18]. To identify potential sources of bias, we conducted a nonresponse analysis [cf.6]. Utilizing data provided by the registrar, we ran a series of one-sample and group means comparison tests to identify how our undergraduate sample differed from the average undergraduate at UNC. We found significant overrepresentation by females (p=0.000), older students (p=0.000), and white students (p=0.002). We then utilized a median split to look at the effect of response time. Early responders tended to be younger than late responders (t1: 20.67, t2: 21.36, p=0.012) but we did not see a significant time difference for gender (p=0.838).