The main purpose of this dissertation was to relate a study within the framework of Peer Student Group Research to the basic Process-Product Paradigm for Research on Teaching. Information about previous research within this field was given in the chapter on background. A total of 287 peer students at the Department of Education, Umeå University and at the teacher training colleges at Umeå and Luleå participated in this study. Thirty-one students worked individually and constituted a control group. The remaining 256 students were divided into 64 groups. Each group consisted of four students and was videoed during the problem-solving phase (60 minutes). The task dealt with a discipline case in the Swedish compulsory school. This case was discussed on the basis of four different psychological theories. A 2(Content, Group vs. Individual Content) X 2(Cohesiveness, High vs. Low Cohesiveness) X 2(Credit, Group vs. Individual Credit) design was used. A mediating model, i.e. Contribution(Analytical vs. Structural Contribution) X Persistence(Task Persistent vs. Task Non-Persistent Contribution) was used to observe the processes. Two of the videoed groups were analysed naturalistically. The results were measured psychometrically by means of an ANOVA. The ANOVA analysed main effects and interactions between the factors for the four process variables and for the product variable. The results of this study revealed the importance of studying Educational Productivity, i.e. both the process and the product. Furthermore, the study demonstrated that interactions could be discovered with psychometric methods but not with naturalistic methods. On the other hand, the naturalistic method revealed more sensitively how the group members worked and how they attempted to solve their task. The traditional assumption of group work vs. individual work was falsified. Individuals did better than groups on this type of problem. The results were discussed in both psychometric and naturalistic terms. / <p>Diss. Umeå : Umeå universitet, 1996</p> / digitalisering@umu

Many children in foster care have emotional or behavioral problems or are at risk for these problems. It is important to identify parents willing to foster children with these problems in order to ensure placement, care, stability and well-being of such children. This study presents a new 40-item self-report measure of the willingness of parents to foster children with emotional and behavioral problems, and two 20-item parallel forms of this measure. In addition, this study presents evidence of reliability and validity of scores derived from these measures with a national sample of 297 foster mothers. Coefficient alpha for these measures was 92 or greater, indicating excellent internal consistency reliability. Scores from these measures were unrelated to demographic characteristics, providing evidence of discriminant validity. In addition, scores from these measures were higher for foster mothers licensed to provide treatment foster care than for mothers only licensed to provide regular foster care, providing support for known groups validity. Finally, support for construct validity is provided by the fact that foster mothers with higher scores on these measures had fostered longer, were fostering more children at the time of this study, and had fewer children removed from their home at their request.