The purpose of this study was to investigate the stability of factor structure across three levels of analysis: student scores, class means, and teacher's group means, as units of analysis. / It was hypothesized that there would be no similarity among extracted factors when analyzing the fifty items from the Survey of Study Habits and Attitudes (SSHA) developed by Brown and Haltzman (1966) across three levels of analysis. / The population was defined as high school students who attended school in the U.S.A. during the 1978-79 school year. Approximately 4,300 students comprised the sample collected by the Individualized Science Instructional System Project (ISIS). These 4,300 students were aggregated to 181 classes and 62 teacher's groups. / Factor analyses were conducted for the three levels of analysis. The stability of each factor was examined by varimax rotation. Procrustes solution was utilized to match all three initial factor structures to determine the stability of factor structure across different levels of analysis. / The methods of judgment were used in this study to determine the degree of similarity or dissimiliarity among the factor structures: (a) subjective judgment, based on observation of factor loadings and comparisons of the labeled factors across the three levels of analysis; (b) application of the Procrustes solution, which provided the correlation between factors and the root mean square which are two viable criteria to judge the similarity or dissimilarity of the factors extracted. A comparison of the results from labelling and factor matching showed a good match among the fellow factors across the three levels of analysis. / Results of the study led to the conclusion that as the level of aggregation increased, the correlation coefficient between variables increased constantly. Varimax rotation showed that Factor 1, which is a strong factor, is very stable across all three levels of analysis. Factors 2 and 3 are weaker and less stable. / Factor matching and factor labelling both revealed that all three factors across the three levels of analysis are stable; however, this stability is very high for Factor 1 and led the investigator to conclude that the strong factors would remain the same regardless of the level of analysis. Therefore, the null hypothesis of dissimilarity among extracted factors across different levels of analysis was rejected. / Source: Dissertation Abstracts International, Volume: 41-09, Section: A, page: 3785. / Thesis (Ph.D.)--The Florida State University, 1980.
Identifer | oai:union.ndltd.org:fsu.edu/oai:fsu.digital.flvc.org:fsu_74285 |
Contributors | GUDARZI, ABDOLREZA., Florida State University |
Source Sets | Florida State University |
Detected Language | English |
Type | Text |
Format | 116 p. |
Rights | On campus use only. |
Relation | Dissertation Abstracts International |
Page generated in 0.0016 seconds