Research objectives were to determine whether first ballot votes predicted final verdicts in actual felony juries, to examine the first vote in relation to other major juror tasks in courtroom trials, to assess the correspondence between first ballots and final verdicts, and to explore methodological issues in jury research. The study sample included jurors who had served on felony cases in Fayette County, Kentucky. These jurors participated in telephone interviews during which questions were asked about the first ballot vote. Of the 142 respondents, 83 were female (58.5 percent), the median age of respondents was 42 years at the time of jury duty, and the overwhelming majority of jurors were white. Interview data indicated that first ballots predicted jury verdicts at a high level. It was probable, however, that influence occurred prior to the first ballot, making it unlikely that the distribution of votes on the first ballot was equivalent to individual juror inclinations at the time they entered into deliberation. Results suggested that a factor in the magnitude of the relationship between first ballot votes and final verdicts was whether jurors discussed the case before the first ballot. Jury deliberation appeared to play a significant role in shaping juror verdicts. The only apparent support for a leniency bias arose when agreement among jurors was taken into consideration and deliberations were evidence-driven. Methodological issues in the study of real juries are discussed. 25 references and 3 tables