New automated methods for large-scale acoustic analysis bring expanded opportunities for investigating social factors influencing variation and change in vowel systems. This paper explores social factors in a 108-speaker subset of a 250-speaker conversational corpus from Raleigh, North Carolina, where the community has shifted in nearly uniform fashion from a Southern vowel system to an aregional standard system. Age, occupation, parents' occupation, sex, and neighborhood are evaluated using linear mixed-effects models, with Z2-Z1 for each of the 5 front vowels as a separate dependent variable. While there are some significant occupational effects, year of birth is the strongest and most consistent social factor, indicating considerable uniformity during the course of change. Adding more speakers from the corpus will facilitate the use of other socioeconomic variables such as education level as well as finer-grained occupation variables, which may provide insight as to the mechanisms by which professional speakers lead the shift.