CHILDREN IN RESIDENTIAL CARE - HOW TO EVALUATE BEHAVIORAL-CHANGE

This article reports the results of an observational longitudinal stud y carried out on four children who lived in a residential community be cause of their multiproblem families. The aim of this study tvas to ap ply the evaluative methodology proposed by Le Poultier (1990) in order to assess I) the processes of change in the children over time and II ) the relations between educational practices and observed changes. Th e findings confirm I) the existence of processes of change in the chil dren; II) the presence of correlations between the displayed behaviour of the the children and adult practices. In addition, what emerges is the existence of a dimension (transgression versus co-operation and a utonomy) underlying the educational project designed by the adults. Th e data refer to observations of the behaviour of each staff member on a weekly basis for a six-month period and at the same time, to the sel f-observations of each adult regarding the behaviour adopted towards e ach child. The data were analysed using: I) correlational coefficients to check the relation between the children's behaviour and the social work, II) Correspondence Analysis to permit a simultaneous representa tion of the developmental changes observed in the four children with r espect to social work.