In research paper we examine that teacher evaluation is conducted to improve the teaching effectiveness of the faculty. However, the evaluation itself has to be effective, which is possible when two issues are first addressed quite skillfully. First is the factors (or criteria) of successful teacher and teaching. And second is 'who will evaluate?' Unless these challenges are addressed, it is difficult to practice a result-oriented evaluation of teachers, particularly at the higher education level. Existing research tells that there is a universal set of factors (variables) with commonly used operationalization (measurements) to evaluate teachers. Adjustments are possible to add more factors and/or attributes (Questions) according to the situation. For example, besides common evaluations, public and private teachers will also be evaluated on different grounds. It is global practice that teachers are evaluated according to their 'Teaching Methods (TM)', 'Utilization of the Qualification & Experience (UQE)', 'Teachers Attitude & Behavior (TAB)', 'Classroom Management (CM)', 'Care of Individual Differences (CID)', Co-Curricular Activities (CCA)', 'Sense of Responsibility (SR)', and 'Evaluation Technique (ET)' using a set of questions (about the attributes) with either dichotomous (i.e., yes/no) or continuous (i.e., 5-Point or 7-Point) scales. Mostly students, heads and colleagues are the evaluators. It is ARGUED in this research that the above practices are based on several 'assumptions or preconceived expectations.' One of the assumptions is that all the respondents give similar importance (value) to every variable and attribute. However, it is obviously possible that 'CID' can be more important for a student while heads may be attaching higher value with 'SR'. This diversity can also be found within one group of students or heads or colleagues. For instance male students may give priority to the CCA while female students can put TAB as higher level factor. In this study, three samples of heads, colleagues and students were selected (using statistical procedures) who all filled the same questionnaire with 49 questions on 9 variables. Using SPSS (V. 12.0) data was analyzed both in a combined manner as well as separately for heads, colleagues and students. To find the 'Best-Fit' of factors predicting the behavior of a particular group, 'Stepwise-Regression' was applied. For computing the demographic impacts the researcher used 'Tests of Significance' including, 't-Test' and 'ANOVA' to compute the mean differences. The results support our hypothesis about the differences of groups from each other as well as variations in the selection of factors due to the demographic diversities within each group of evaluators. Eight predictors (TM, UQE, TAB, CM, CID, CCA, SR, & ET) were used to explain the dependent variable (Overall Score--OS). The findings of this study report that heads' attitude is determined by SR (p = 0.028) & ET (p = 0.015), Colleagues prefer CID (p = 0.008), CCA (p = 0.035) & SR (p = 0.000) while, students used TM (p = 0.039) & ET (p = 0.000) for evaluating the teachers in Gomal University. It is obvious that out of eight independent variables, very few are playing any role in the evaluation process. Most of the factors have been excluded in stepwise regression. Furthermore, the predictors (factors) selected by heads, colleagues and students are also more different than similar. Likewise, heads have no difference of opinion across all the demographic groupings. But colleagues and students are significantly different on 'Faculty' but similar on other demographics. Furthermore, teachers have some difference of opinion due to their grouping on 'Domicile'. To cut short, it has been hypothesized, tested and proved that all the evaluators do not give similar importance to the factors of evaluation. So their evaluation does not include only the placing of a teacher on the scale, it is also affected by whether or not the evaluator considers the predictor important.