Introduction and objectives: Social climate is an understudied factor in residential youth care (RYC) institutions. Already in the 1950’s, the World Health Organization stated that “atmosphere” is an important factor in psychiatric treatment, but a very difficult element to measure. Assessing the features of social climate is further complicated by the fact that residents are both effected by and contribute to it. Only a handful of studies have investigated social climate in RYC institutions, including youth prisons. Research reveals that for adolescents a positive social climate is associated with better coping strategies, less running away behavior, less peer victimization and less social and behavioral problems, while a negative social climate can lead to passive coping strategies, more running away behavior, and more peer victimization. Previous studies have used different instruments for measuring social climate. One of them is The Community Oriented Programs Environment Scale (COPES), a 100-item dimensional instrument that assesses ten features of social climate in residential care facilities. The COPES has never been validated in a RYC setting, because the original target group has been elderly residential care and adult psychiatric institutions. Moreover, psychometric qualities have been found questionable in studies following the original scale development. In addition, previous research has provided criticism on the fact that the 100-item instrument is too long and shows redundancy in content. Therefore, validating the COPES is important to formulate further generalizations and associations between social climate and important outcome variables (e.g. adolescent mental health and quality of life) in RYC.
The first objective of this paper was to validate the COPES on adolescents and staff leaders in a country-wide sample of Norwegian RYC institutions. If the COPES cannot be validated, the second aim was to revise the instrument by improving the psychometric quality and conceptual validity of the instrument. The third objective was to shorten the COPES by using both classical test theory and modern test theory techniques. The goal was to reduce the number of items while retaining acceptable psychometric properties and sufficient information to differentiate respondents at different levels of social climate.
Method: All adolescents between the ages of 12-23 living in residential youth care institutions in Norway were invited to participate in the study. The final sample consisted of 400 adolescents and 140 staff leaders that were recruited from 88% of all Norwegian RYC institutions. The COPES was administered to both adolescents and staff leaders.
Given that there was on average less than 5% missing data, estimation maximization was used to impute missing values on the COPES. In the initial analysis, the data was fitted to the original COPES structure by means of confirmatory factor analysis. Following this analysis, a step-based approach in scale improvement and scale shortening was carried out. First, items with almost zero variance were removed from the questionnaire. Second, items with low or non-significant factor loadings were removed. Third, items that show a redundancy on content were removed from the questionnaire. Finally, item response theory was used to evaluate whether the remaining items provide sufficient differentiation on different levels of social climate.
Results and conclusions: Results from the confirmatory factor analysis indicated that the data did not fit the original structure. Furthermore, results revealed that a short-version of the COPES can improve the psychometric properties, as well as usability, and conceptual validity in RYC institutions. The revised short version of the COPES for measuring social climate in RYC institutions will be presented at the conference, illustrating improved predictive validity for measuring social climate in RYC institutions.
Residential child care , Program evaluation and quality in child welfare
