Factorial Validity and Reliability of a Survey to Assess the Teaching Effectiveness of Graduates of Teacher Education Programs

Abstract

This study examined the factorial validity of a newly developed 28-item survey called the Teacher Effectiveness Survey (TES) that was designed to assess the teaching effectiveness of graduates of teacher education programs. A sample of 390 principals in 33 states was surveyed. This sample was divided into two subsamples and respondents were randomly assigned to either a training sample (Sample 1, n = 195) or a confirmatory sample (Sample 2, n = 195). An exploratory factor analysis was performed on the data obtained from Sample I to identify an appropriate factor model and to refine the instrument. This approach yielded a three-factor model from which four items were dropped from subsequent analyses. Sample 2 data were subjected to a confirmatory factor analysis using LISREL6, which produced three factors that were identified as Instruction (10 items), Interpersonal/Professional (8 items), and Leadership (6 items). Corresponding estimates of intemal consistency reliability (Cronbach's alpha) of .96, .92, and .92 were obtained. Intercorrelations among the three factors ranged from .80 to .87, suggesting the possibility of a global measure of effectiveness rather than three separate constructs. Alternative explanations are also discussed.