In this study, 17 family child care providers took part in a seven-week, online training that used problem-based learning to address guidance and discipline practices. A website and an e-mail list provided the context in which two groups of participants discussed causes and solutions for three real cases of preschoolers' challenging behavior. Based on social constructivist theory, this research addressed two aspects of cognitive change and group problem-solving: the process of knowledge building within a group and individual appropriation of knowledge generated by the group.;Transcripts of group discussions were analyzed for evidence of knowledge building using the Degrees of Synthesis Model, a model developed by the author for the purpose of examining the level of cognitive interaction that participants' messages revealed in each discussion. Four levels of cognitive interaction were represented by the model: independent proposal, comparison, reactive modification, and integration. Results of the analysis indicate that the model was discriminant and moderately reliable. Results also indicated that participants showed variable levels of synthesis of one another's ideas and suggested that variability was due in part to characteristics of the cases being discussed. Further research is needed to refine the model and to provide greater understanding of variables that contribute to and inhibit group knowledge building.;Individual appropriation of knowledge generated by group discussion was measured by three different methods. Analysis of these data focused on the level of change in participants' thinking due to participation in group discussion. Results suggested that the degree of appropriation of new group knowledge is inversely related to the degree of affiliation with the larger child care professional community and inversely related to the degree of participation in the problem-solving discussion. Results suggested that more knowledgeable peers within the group discussion provided beneficial models of complex, theoretical thinking to less knowledgeable participants.;Results of self-report data indicated that problem-based learning delivered online was engaging for participants, conducive to their professional development, and feasible to implement. Further research is needed that will clarify the audience for whom it is most beneficial and specific variables that will optimize it as a training option.