A recent article by Sheingold calls for reviving the voting research agenda left unfinished by the Columbia School, including its emphasis on social networks as determinants of individual party choice. Two hypotheses about the relationship of social networks to party choice are tested on data from an Israeli election survey: (1) measures of social network ties to parties are more highly correlated with party choice than are standard background attributes such as class or ethnicity; (2) more important, the impact of social networks is substantial even after controlling for individuals' background attributes. The second hypothesis is based on the work of V. O. Key, and has not been tested on individual data previously. Both hypotheses are supported by the data.