Stuart Buck found an interesting quote in psychology professor Richard G. Medlin's article, "Home Schooling and the Question of Socialization," Peabody Journal of Education, Vol. 75 (2000): 107-23:

Shyers (1992a, 1992b), in the most thorough study of home-schooled children's social behavior to date, tested 70 children who had been entirely home-schooled and 70 children who had always attended traditional schools. The two groups were matched in age (all were 8-10 years old), race, gender, family size, socioeconomic status, and number and frequency of extracurricular activities. Shyers measured self-concept and assertiveness and found no significant differences between the two groups.

The most intriguing part of the study, however, involved observing the children as they played and worked together. Small groups of children who all had the same school background were videotaped while playing in a large room equipped with toys such as puzzles, puppets, and dolls. The children were then videotaped again in a structured activity: working in teams putting puzzles together for prizes.

Each child's behavior was rated by two observers who did not know whether the children they were rating were home-schooled or traditionally schooled. The observers used the Direct Observation Form of the Child Behavior Checklist . . . , a checklist of 97 problem behaviors such as argues, brags or boasts, doesn't pay attention long, cries, disturbs other children, isolates self from others, shy or timid, and shows off. The results were striking -- the mean problem behavior score for children attending conventional schools was more than eight times higher than that of the home-schooled group. Shyers (1992a) described the traditionally schooled children as "aggressive, loud, and competitive" (p. 6). In contrast, the home-schooled children acted in friendly, positive ways.