| Summary: | The present thesis introduces multilevel confirmatory factor analysis (MCFA) as an appropriate methodology to use when data are hierarchically structured and an identification of theoretical concepts of a given inventory is desired. The main purpose of this thesis is to introduce all possible methods and issues which are connected with attitude research. The MCFA methodology is demonstrated on the example of the attitudes toward physical education (ATPE). Data were collected from high school students (n = 1157) nested within PE classes (m = 87). A different factor structure of ATPE was determined at the student level (within-group) and at the PE class level (between-group). At the within-group level, Wear's four-factor theory of ATPE was confirmed. However, the existence of two wording factors (positive and negative) and the school PE-related factor was indicated. At the between-group level, the resulting three-factor model exhibited only a marginal fit. The structure of theoretical concepts and their indicators was indistinct, because the items were originally developed to measure ATPE at the within-group level. The analysis included four cases: (1) single-level continuous, (2) multilevel continuous, (3) single-level ordinal, and (4) multilevel ordinal. All four approaches were compared in terms...
|
|---|
