Interviewers' voices can influence phone interview cooperation, yet we know little about how their accents and regional speech variation (e.g., pace) impact response. Taken together, these carry many associations and inferences for respondents. For example, a shared accent may increase feelings of liking and affiliation between respondent and interviewer, producing higher cooperation. Conversely, a dissimilar accent may increase feelings of social distance and increased refusals. Despite little information about accent effects in survey interviews, many US phone surveys employ interviewers from outside the sampled region. Using digital audio recordings to classify interviewer accent, we evaluate the impact of accent and regional speech variability on cooperation and refusal in several phone surveys, comparing results by in-region and out-of-region calls. We hypothesize that interviewers' regional accent, speech rate, pitch variability, and fluency predict first contact outcome, controlling for their personal and professional characteristics. We discuss findings in the context of interviewer training and staffing best practices, and potential accent effects on survey error.