Davis, J. H., Stasson, M., Ono, K., & Zimmerman, S. (1988). Effects of straw polls on group decision making: Sequential voting pattern, timing, and local majorities. Journal of Personality and Social Psychology, 55, 918-926.
Six-person juries (evenly divided between members inclined toward guilty and not-guilty verdicts) responded to a straw poll either early or late, with subjects inclined toward guilty or not guilty voting first, or simultaneous voting. Critical (fourth) voters in straw polls were somewhat influenced by the preceding voter sequence, and were substantially influenced by the timing of the poll. We interpreted the asymmetry in opinion change observed for members initially inclined toward guilty and toward not guilty in terms of the defendant protection norm (leniency bias), the salience of which had apparently been emphasized by discussion and voting, even the prospect of public voting. We extrapolated member-level opinion changes to the group level, illustrating the verdict consequences of individual-level opinion changes. Observed group consensus verdicts, however, deviated from predictions, implying a different role for procedural factors and emphasizing the importance of group-level data in such research settings.
Juror voting sensitive to sequencing effects in evenly divided juries. Individuals voting 4th preceded by 3 votes opposite their predeliberation preference much more likely than chance to switch their vote, but only when poll taken before discussion. Polls taken after minimal discussion showed leniency but no sequence effect. Weak impact of polling on final verdicts.