The determinants of the persistence of involvement in what was discussed in a group are examined, focusing on the discussion process. An experimental procedure was developed to automatically measure several aspects of the process including interpersonal influence that was measured in terms of the changes of preference for alternatives that subjects showed following a short speech by another subject. The procedure used 27 four-person groups, 15 all-male and 12 all-female. The results showed that the amount of interpersonal influence a member gave the others, rapid move toward agreement, participation by speech, and discussion time elapsed contributed to determining the persistence of involvement over the month following discussion. Which of these variables was effective depended on gender and the extent of the persistence concerned. The method was considered instrumental in training to improve social skills for group problem solving, group decision making and other group work.