期刊名称：Drustvena istrazivanja. Journal for General Social Issues
印刷版ISSN：1330-0288
出版年度：2001
卷号：10
期号：3 (53)
页码：485-501
出版社：Institute of Social Sciences IVO PILAR
摘要：Based on results obtained from 710 teachers from 71 primary school in Zagreb and Zagreb County, reliability and validity of the Moos-Insel Work Environment Scale was analyzed for the measurement of school climate. All estimates were derived for three types of units of analysis: individual raters, observed schools and artificial "schools" (samples of 10 randomly selected raters per group). Empirical data show that the Work Environment Scale is a reliable measure of school climate, but cannot be used for accurate assessment of individual differences in perceived climate within particular schools. Factor structures of individual and ecological differences are not identical showing that school entities as units of analysis and the individual rates obtained within a single school should not be used interchangeably. Applicability of Cronbach-α and splitsample correlation coefficients for the reliability estimation of the instruments aimed at the assessment of environmental characteristics was proved by logical, theoretical, and empirical arguments. From the theoretical as well as empirical standpoint, the interchange of ecological entities and individual descriptions within an environmental unit proved to be generally unjustified. This can only be done taking into account empirical evidence on reliability and validity of individual and environmental differences.
