Participants maintained a social interaction diary, a variant of the Rochester Interaction Record, to describe their close personal relationships. Relationships were defined using relative frequency of contact with different individuals and via participants' descriptions of relational partners. Individuals with whom participants had more contact were more likely to be described as close friends than were individuals with whom participants had less contact. The results suggest that the characteristics of same-sex relationships depend less on the specific partners composing the relationship than do the characteristics of opposite-sex relationships. As hypothesized, the characteristics of close same-sex relationships did not differ from the characteristics of other same-sex relationships, although interactions with best friends were more intimate than interactions with other same-sex partners. In contrast, the characteristics of close opposite-sex relationships differed from the characteristics of other opposite-sex relationships. Men and women who were romantically involved had more contact with their romantic partners than with other opposite-sex persons, and they had more contact with their romantic partners than men and women who were not romantically involved had with their most frequent opposite-sex interaction partner. For women, interactions with most frequent opposite-sex interaction partners were more affectively rewarding and instrumentally positive if these men were romantic partners, whereas for men, the affective quality and instrumentality of interactions with most frequent opposite-sex interaction partners did not differ as function of whether these women were romantic partners.

Twice over two years, participants described their day-to-day social interactions for two weeks, and they described their social skills. Within each phase of the study, self-perceived social skills and the quality of social interaction were positively related. Moreover, the results of cross-lagged panel analyses suggested that changes in the quality of day-to-day interaction led to changes in perceived social skills, whereas changes in social skills did not lead to changes in the quality of interaction. In contrast, quantity of interaction and social skills were not related either within or across time. Consistent with some aspects of the sociometer hypothesis, improvements in the quality of people's social interactions led to increases in self-perceived social skills, and declines in the quality of social interactions led to decreases in self-perceived social skills.

This study investigated the relationship between the academic performance and social behavior of a sample of college students. In support of the primary hypothesis, for males, academic performance was negatively correlated with both the quality and quantity of their social interactions. For females, academic performance and social interaction were not correlated. This was true when both same and opposite sex social interactions were examined.