Doctoral Dissertation Defense: Sadeq Sohrabie

Abstract

The Effect of Faculty Leadership Style on the Results of Student Evaluation of Teachers

Mohammed Sadeq Sohrabie
University of the Incarnate Word, 2019

Research Focus. The purpose of this study was to study the effect of college instructors’ classroom leadership style and the differences between the perceptions of students and faculty as indicated in Student Evaluation of Teaching (SET). This effect was examined using leadership style factors data collected through the Multifactor Leadership Questionnaire (MLQ), SET scores collected through the evaluation questionnaire, and demographic variables. The few researchers who have examined the relationship between SET and leadership style, have used simple correlation analysis methods in limited settings and they haven’t offered a construct. My effort was to examine a construct of these relationships in U.S. universities and use a more complex research method for this purpose. There are three research questions under examinations: 1) Are the MLQ transactional and transformational leadership factors conceptually and empirically independent and valid? 2) Do the MLQ transactional and transformational leadership factors predict SET scores? 3) Do the MLQ transactional and transformational leadership perception difference, between faculty and students predict SET scores?

Research Methods. This research study was based on a post-positivist paradigm using a quantitative approach to test the conceptual framework reflecting my personal philosophy. Phillips and Burbules (2000) called post-positivism a pluralistic paradigm in research that helps in conducting scientific socio-educational research that reaches partial conclusions and points the way to further research. The sample for this study included undergraduate students in regular four-year degree programs from all class standings. Two universities were involved in the data collection process, one public and one private university in Southwest Texas. 400 students were recruited from classes taught by 10 faculty from both universities. Using a complete case approach, data were analyzed with 361 observations. Data analysis was done in 3 steps: 1) establishing the validity of the factors describing transactional and transformational leadership styles using confirmatory factor analysis utilizing the data from MLQ questionnaire; 2) testing the first conceptual model of relationship between leadership factors and SET using structural equation modeling; and 3) testing the final model of relationship between leadership concept differences between students and faculty and its effect on SET using GSEM.

Research Results/Findings. Confirmatory factor analysis demonstrated that the 7 transformational and transactional factor model in their original constructs as recommended by MLQ developers was not valid. The fitted valid factor construct that I have examined represented a very complex relationship construct among factors and items. Using this fitted model as the basis for the SEM analysis, it was demonstrated that leadership factors from student perspectives alone cannot predict SET scores, while student age, gender difference between students and faculty and expected grade can predict SET scores. Testing a third model, it was demonstrated that difference between perception of students and faculty from faculty leadership style can actually predict the SET scores, almost as strongly as the expected grade can predict SET scores.

Conclusions From Research. The MLQ transformational/transactional leadership factors in their original structure were not valid. There are many cross loadings across items and factors. The majority of these cross-loading were negligible, but they showed that the nature of the data was not as simple as it might look in its original structure These factors even in their fitted structural model cannot predict the small variance in the SET scores. However, when looking at the differences between perception of leadership between student and faculty, they can predict SET scores. Student age, gender difference between student and faculty and expected grade at the end of semester all can also predict the SET scores. All these variables have positive meaningful correlations with SET scores.