UC Berkeley

Teachers, parents, and school administrators have become interested in implementing programs to teach children social interaction skills for developing friendships. These social skills training programs typically teach rote skills without addressing a fundamental aspect of social behavior, children's beliefs about their potential for increasing social skillfulness. Beliefs about one's ability to increase competence have been shown to have significant effects on performance and yet these beliefs are rarely considered or incorporated into social skills programs. This dissertation includes two scale development studies conducted to bring the field closer to effective measurement of children's beliefs in this domain. Such a measure will allow researchers to study how interventions shape children's beliefs, ultimately leading to interventions that better remediate social problem-solving weaknesses.

Results of exploratory factor analyses supported a one-factor structure with two composites instead of a five-factor structure. Item-level analyses--including frequencies, p values, and point-biserial correlations--were indicative of some potential weaknesses in the items, including poor item discrimination and some non-significant item-total correlations. Qualitative item-level analyses suggested that the items were not able to elicit the range of responses that was expected; overall students tended to endorse statements consistent with the idea that all children have the potential to improve friendship skills through effort and practice. Test-retest and split-half reliability measures indicated that the items were eliciting fairly reliable responses. Validity evidence supported the assertion that items were eliciting information about the purported constructs.