Original scientific paper
Units of Analysis and the Psychometric Treatment of Environment Attributes Measures
Alija KULENOVIĆ
Vlatka DOMOVIĆ
Abstract
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.
Keywords
Hrčak ID:
20018
URI
Publication date:
30.6.2001.
Visits: 2.358 *