Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)



First Advisor

Abbas Tashakkori


Two studies were conducted as part of this research effort. The purpose of these studies was to determine the effects of assessor and assessee gender, ethnicity, and assessment role on performance observation ratings. Study I was Causal Comparative in nature and involved analyzing actual performance observation ratings received on the Louisiana Teacher Assessment Instruments (LTAl). Study II was an experimental study and involved analyzing experimentally manipulated, teacher-performance-observation ratings received on an instrument entitled "Survey of Effective Teaching Behavior." The data were collected in the spring of 1995 and expand upon the findings of Study I. There were essentially three different issues of bias to be addressed. Issue I addressed gender bias, Issue II addressed ethnicity bias, and Issue III addressed role- related bias within the assessment ratings, that is, Issue III examined the prevalence of bias attributable to the type of "role" assumed by the assessors within the assessment context. Study I results indicate significant main effects of assessee gender. Female assessees scored higher than male assessees on all components. The results also indicate that some differences in assessment ratings are attributable to assessee/assessee ethnicity. Caucasian assessees had consistently higher ratings than African-American assessees, regardless of assessor ethnicity. Results regarding Role Bias indicated that only in one of the eight components are the differences in assessment ratings attributable to assessor role. In addition, those assessors assuming the role of principal give higher ratings, yet master teachers have a slightly higher overall mean component rating. It is concluded that despite some statistically significant effects, magnitude of bias due to gender, ethnicity, or role was small. However, it is methodologically important that we examine the possibility of gender, ethnicity, or role biases that may devalue assessment results. As the nation moves toward teacher assessment systems that rely on observational rating performances, we must be prepared to extrapolate true assessment ratings from those that are confounded by bias. Differences in assessment results are tolerable but only if they are not the result of gender, ethnicity, or role biases rather than due to true differences in assessees' performance.