A FACTOR ANALYSIS STUDY OF SEMANTIC DIFFERENTIAL SCALES ON VOCATIONAL AGRICULTURE CONCEPTS

The problem of this study was to determine the factorial structure of Osgood, Suci, and Tannenbaum's traditional semantic differential (SD) scales on vocational agriculture concepts. The study was designed to answer the following questions: (1) Are the evaluative, potency, and activity factors clearly evident? (2) Are those scales traditionally associated with one of the factors loading relatively high and pure on that factor? (3) Will scales maintain their traditional positive-negative posture on vocational agriculture concepts? (4) Does a multidimensional semantic space exist for vocational agriculture concepts when measured by traditional SD scales?;Data were collected from a stratified random sample of vocational agriculture teachers in Washington. The population was stratified by teaching assignment (general or specialized) and by years of teaching experience ((LESSTHEQ) 10 and (GREATERTHEQ) 11).;Principal component analyses with varimax and quartimax rotation were used in analysis of the data. Prior to principal component analysis the data were treated by a specially written computer program to better meet assumptions of interval level data. Scales were considered loading high and pure on a factor when the loading was (GREATERTHEQ) .40 on one factor and (LESSTHEQ) .23 on other factors. Factors with eigenvalues (GREATERTHEQ) 1.00 were retained for analysis.;The evaluative, potency, and activity factors were not evident on any concept. There was no evidence of scales loading high and pure on the same factor over all concepts. A multi-dimensional semantic space was deemed not to exist based on the lack of factor structure when all concepts were considered. All scales maintained traditional positive-negative postures on all concepts.;A SD instrument, consisting of 20 important vocational agriculture concepts, along with four selected bipolar adjective scales from each of the three traditional Osgood factors, was used to collect data.;It was concluded that SD scales should be factor analyzed when used with different sets of concepts. Assuming that scales will maintain their traditional factor structure can lead to erroneous results and conclusions.