Student Evaluations of College Instructors:
An Overview

Author: Patricia A. Gordon
Source: Valdosta State University, As partial requirements for PSY 702: Conditions of Learning, Dr. William G. Huitt, Instructor

Return to: | Readings in Educational Psychology | Educational Psychology Interactive |

Abstract

The paper provides an overview of student evaluations of college instructors. Reasons for and appropriate uses of student ratings are examined. Sources of validity as well as bias are reviewed. Student self-insight into implicit and public theories is related to the ability to make valid, thoughtful evaluations. Evaluations are described as multidimensional, stable, and a function of the instructor. Improvements to satisfy faculty, students, and to meet the needs of the information age conclude the paper.

Student Evaluations of College Instructors: An Overview

For nearly a decade, teaching in higher education has risen in status and is viewed by many to be of greater importance than the more traditional focus on research (Boyer & Rice, 1990; McKeachie, 1990; Paulsen & Feldman, 1995). In designing a supportive teaching environment, specific characteristics have been identified that facilitate teaching and drive the process of instructional improvement. These characteristics include: unambiguous commitment to and support of teaching and its improvement from administrators; shared values between faculty and administrators about the importance of teaching; involvement of faculty in planning and implementing programs to improve teaching, thereby creating a sense of faculty ownership; the presence of effective department chairs; frequent interaction and collaboration among faculty; a faculty development program; a broad, expanded view of scholarship; and rigorous evaluation of teaching (Paulsen & Feldman).

It is this last characteristic, a rigorous evaluation of teaching, that is the focus of this paper.

Student Involvement in Evaluation

Students need not be silent partners in the improvement of teaching in higher education. In actively seeking students' responses to instruction, colleges and universities are sending the important message that they honor and support the teaching-learning process. According to the North Carolina State University Handbook for Advising and Teaching (1994), student evaluation of teacher effectiveness serves two purposes. Most importantly, student evaluations provide instructors with important feedback from the consumer's point-of-view. Students can and do make important contributions to the teaching-learning process, and teachers must be receptive to their ideas.

Additionally, student evaluations are of value to administrators and department chairs in assessing perceived effectiveness of instruction (NC State University, 1994). This information is often used for both formative and summative purposes (Marsh, 1984) and occasionally to reward exemplary teaching (NC State University).

In summary, since the 1970's, there has been a consensus on the purpose of student evaluations at colleges and universities. Rifkin (1995) confirmed that the primary purpose is formative; that is, facilitating faculty growth, development, and self-improvement. Secondly, student evaluations are used for summative purposes and often play a vital part in tenure, promotion, reappointment, and salary decisions.

Regardless of purpose of student evaluations, formative or summative, their use implies belief in the following principles (Huitt, 1995; Stockham & Amann, 1994):

1.    Learning is an active process and student involvement is an integral part of that process.

2. Student characteristics and behaviors impact perception of and interaction with the teacher.

3. Teachers view their teaching with regard to the paradigms of their students in order to facilitate change and build for growth.

4. Teachers recognize that students can make important contributions to the teaching-learning process.

5. The teaching-learning process is dynamic and should change over time and with context.

Students are in a unique position to assess a variety of aspects concerning effective instruction. Scriven (1995) identified several sources of validity for student ratings of instruction. They include:

1.    students' ratings of their own increased knowledge and comprehension;
2.    perceived changes in motivation toward a) the subject taught, (b) a career associated with the subject, (c) further learning in that subject area;
3.   observed teacher behavior relevant to competent teaching, such as punctuality;
4.   identification of teaching style indicators, such as enthusiasm;
5.   test evaluation;
6.   student consumerism; that is, information not relevant to competent teaching, but important to other students, such as textbook cost, attendance policy, or homework.

Finally, although not systematically examined, student ratings should incorporate research on teaching (Marsh, 1984). Huitt (1995) proposed that educational research should help answer the question: "Why do some students learn more than others?" When one considers the review of models of the teaching-learning process as described by McIlwrath and Huitt (1995), it becomes apparent that student evaluations can be an important measure of teacher behaviors and teaching methods when they are related to a product or output measure of student learning and growth. It might be suggested that student evaluations are additionally relevant to presage or teacher characteristic variables (e.g., subject knowledge, teaching skills, teacher efficacy) for the next group of learners. Such a powerful triad serves to facilitate research in the teaching-learning environment in an attempt to answer major educational research questions.

Negative Implications

Although hundreds of papers have been written regarding the effectiveness of student evaluations on faculty performance, such papers cannot be easily summarized. Researcher opinions run the gamut from valid, reliable, and useful to invalid, unreliable, and useless (Aleamoni, 1981). However, there is one criteria that most researchers agree on: student evaluations should be non-normative (Rifkin, 1995). A criterion-referenced system that appraises faculty members according to a set of standards that encourages professional development is preferable to a system that rank-orders faculty on a particular set of items.

Both Scriven (1995) and Simpson (1995) suggested that student evaluations often contain the wrong global questions that tend to rank or compare teachers or serve to recommend a specific course or instructor to other students. An inappropriate question might be: "Is this one of the best courses you have ever taken?" An appropriate global question might involve an evaluation of overall teaching performance (Hativa, 1996).

Cashin and Downey (1992) and Abrami (1989) have deemed global questions appropriate for summative evaluations. Just as a single final summative grade is used to indicate the amount of student learning related to unquestionably multidimensional objectives, so too can effective teaching be summarized via global questions. Cashin and Downey acknowledged that formative questions allow for differences among teachers, courses, and environments, and serve as diagnostic tools for improving the teaching-learning process. Yet they, along with Abrami, favored the use of several global questions for summative purposes to indicate the lack of or existence of problems or difficulties.

Eley and Stecher (1997) favored behavioral observations over Likert agree/disagree type questions since they found that Likert type questions elicited global, impressionistic responses. Behavioral observations seemed to produce more objective responses in distinguishing favorable teaching qualities. In contrast, Cook (1989) found no difference between behaviorally anchored rating scales (BARS) formatted questions and Likert-type questionnaires. Students preferred Likert-type scales over BARS questionnaires, simply because it was less work. A related point, according to Scriven (1995), is to keep evaluations short because students tend to lose interest if the evaluation process becomes too wordy or lengthy.

In an effort to reduce instructor hostility toward student evaluations and to improve overall validity, Davis (1995) suggested faculty-developed evaluation instruments suited to individual departmental curriculum and purpose. Most universities currently use standardized forms that do not recognize individual course contours and content (Divoky, 1995). North Carolina State University (1994) has mandated that each department develop or adopt student evaluation instruments appropriate to the types of courses offered in each department.

Finally, the validity of student evaluations is dependent on the context of how and when they are administered. Two basic principles governing distribution, collection, and handling of evaluation instruments are anonymity and confidentiality (NC State University, 1994). Preferably, an instructor would not be present while students are completing questionnaires in an attempt to control pleas for sympathy or indulgence by the teacher (Scriven, 1995). Questionnaires should be distributed and collected by a non-teaching school employee or designated student representative, who delivers the completed questionnaires to the appropriate official, where they remain until grades are administered (NC State University, 1994).

While individual situations and personalities may be able to lend some credence to the description of student evaluations as unreliable, invalid, and useless, the literature does not support these claims. For over thirty years, exhaustive research has been compiled on student evaluations, and as will be discussed later, when they are appropriately developed and administered, they remain useful tools in impacting the teaching-learning process on the higher education front.

Students' Ability to Judge

As previously mentioned, students have a front row seat to observe teacher behaviors and classroom processes, and are the best judge of what they have learned (Scriven, 1995). The use of standardized evaluations implies, though, that students have the ability to make fair, overall judgments about teaching effectiveness based on legitimate factors that actually relate to effective teaching and learning. Harrison, Ryan, and Moore (1996) have suggested that college students, in general, possess self-insight into how they make judgments concerning their instructors since they have an implicit awareness of the relative importance of the factors they are considering.

Self-insight is a form of metacognition that reflects peoples' awareness of the value and processes involved in their understanding (Harrison, Ryan, & Moore, 1996). Are students aware of the processes they use to integrate information in order to make decisions concerning teacher effectiveness? "Yes," said Harrison, Ryan, and Moore. While students seemed to possess implicit theories associated with the occurrence or nonoccurrence of specific behaviors of instructors, students also possessed self-insight into how they make overall evaluations while incorporating those implicit theories.

Such implicit theories fall into the realm of student characteristics in Huitt's (1995) teaching-learning process model. Students tended to organize and evaluate their experiences with one instructor with respect to previous experiences and observed behaviors (Whitely & Doyle, 1976). These implicit mental theories are then brought into each particular learning situation. For example, if a male student had a negative experience with a female instructor, he might bring an implicit theory regarding gender and teacher effectiveness to other learning situations. Additionally, for an auditory learner, teacher characteristics such as clarity of speech, and speed of material presented might be implicitly ranked as more important to that student than to another more visual leaner. Students' opinions concerning the relative importance of specific teacher behaviors and characteristics including such items as enthusiasm, organization, rapport, and fairness, then, appeared to be brought with them to the learning situation (Harrison, Ryan, & Moore, 1996).

The construct of public theories involves a specific group of people, who because of common experiences, share implicit theories (Whitely & Doyle, 1976). These theories need not be communicated publicly, orally or written. Public theory indicates only common experiences (Wright & Rip, 1981). Regardless of the guiding theories, implicit or public, Harrison, Ryan, and Moore (1996) supported the idea that students possess the self-insight necessary to understand their personal or group weighting system and thereby make valid evaluations of teacher effectiveness.

Greenwald (1997) substantiated Harrison, Ryan, and Moore's (1996) contention that students do make valid evaluations of teachers. During the past 25 years, much research has been focused on said validity. Greenwald discovered that while validity studies peaked around 1980, historically the studies supporting student evaluations as valid remained constant. From 1991 to 1995, validity research was at an all time low, inferring that the 20 years of previous research had resolved most major issues concerning this topic (see Table 1).

[Place Table 1 about here]

Student Evaluations Are . . .

According to Marsh (1984, 1993), there are several reasons why evaluations of teaching should be considered as multifaceted. First, if effective teaching is multifaceted (e.g., Cruckshank, 1985; Gage and Berliner, 1992; Huitt, 1995), then instruments for student evaluations should reflect this multi-dimensionality. Second, there is no single criterion of effective teaching (e.g., McKeachie, 1990). Therefore, a construct approach to validate student ratings is necessary whereby the student evaluation ratings are shown to be related to additional indicators of effective teaching. Finally, different factors of student evaluations will correlate more highly with different indicators of effective teaching (e.g., McKeachie, 1973). Therefore, student ratings should not be summarized by one response to a single item or an unweighted average response to many items.

Throughout the 1980's and 1990's, Marsh (1983, 1984, 1993) studied the multi-dimensionality of student evaluations of teaching effectiveness by the administration of one million questionnaires representing 50,000 courses. Marsh's questionnaire, the Students' Evaluations of Educational Quality (SEEQ), contains thirty-three evaluation items, divided into nine subgroups (see Table 2).

[Place Table 2 About Here]

Over the years, Marsh (1983, 1984, 1993) consistently found that the single most important factor effecting student evaluations was amount learned and the least important factor was course difficulty. This finding was more recently supported by Ryan and Harrison (1995),

Cohen (1981) proposed that if a student knew their final grade in a course, and that grade was high, then the student's perception of amount learned would be high and he or she would directly attribute that grade to effective teaching, and the student rating would be correspondingly high. However, higher grades in a course may actually reflect grading leniency, rather than amount learned. Without comparison to some outcome measure, there is no way to tell.

In the statistical evaluation of the SEEQ, Marsh (1983) considered a host of what he called background characteristics. These background characteristics included prior subject interest, workload difficulty, expected grade, reason for taking course, class level, overall graded point average, year in school, course enrollment, percent majors (that is, is the student majoring in the same division or school in which the course is offered), and teacher rank (teaching assistant, instructor, assistant professor, associate professor, professor). Through statistical analysis of these background characteristics, Marsh concluded that four characteristics had the greatest correlation to the SEEQ's: 1) prior subject interest, 2) general interest, 3) expected grade, and 4) workload difficulty. The importance of prior subject knowledge should be of little surprise; Gage and Berliner (1992) proposed that the schemata we bring to an instructional situation are as important as the instruction itself.

In a simple analysis of background characteristics and SEEQ factors, Marsh (1983) considered the correlation substantial if the r-squared was greater than .05 (with a resulting r > .23). In a statistical analysis of 768 coefficients, 16 met the criteria and 14 of the 16 correlations involved prior knowledge. The largest single correlation (r = .56) was between prior subject interest and ratings of Learning/Value on the SEEQ (Marsh, 1983). The other three background characteristics were expected grade (positively correlated with Learning/Value and Group Interaction), course enrollment or class size (negatively correlated with Group Interaction and Individual Rapport) and average year in school, freshman or sophomore (negatively correlated with Group Interaction.) No other background variables were substantially correlated with any other SEEQ factors.

After determining the bivariate relationships previously mentioned, Marsh (1983) entered the background characteristics into a stepwise multiple regression to predict each of the SEEQ factors. Marsh (1983) used the number of times a background factor appeared in a regression equation as an indicator of importance. The most important variables, as determined by 48 regressions, were expected grades (42), workload/difficulty (30), prior subject interest (28), and general interest/reason (18). Student ratings tended to be more favorable when expected grades were higher, prior subject interest was higher, levels of workload/difficulty were higher, and the percentage of students taking the course for general interest was higher. None of the other background characteristics appeared more than 12 times in the 48 regressions equations.

Additional relationships between students' evaluations of teaching effectiveness and background characteristics were discussed by Marsh (1983, 1984. 1992) and are summarized in Table 3.

[Place Table 3 About Here]

Cranton and Smith (1990) suggested that a different factor sturcture emerges when different levels of analysis are considered. When looking at individual's opinions, two factors emerged: Interest/Atmosphere, which includes such items as inspiring interest and developing an atmosphere conducive to learning, and organization, which includes such items as clarifying material and making clear distinctions among topics. These two factors accounted for 78.8% and 9.8% of the variance, respectively.

When one considers class means as the unit of analysis (which focuses on differences among instructors), another factor structure emerged. The organizational factor included several additional items such as clear purpose of class and use of class time, and accounted for 83.8% of the variance. A second factor, labeled value, accounted for 8% of the variance and included such items as how much learned and overall value.

Marsh and Bailey (1993) analyzed the rating profiles of 123 instructors during a 13-year period. Ratings for each instructor included at least two undergraduate and two graduate courses, and the departments included in the research spanned all disciplines. Since the shape of a teacher profile graphically describes teacher characteristics and behaviors, and the level describes ratings from high to low, Marsh and Bailey questioned whether teachers' profile shape would remain constant over time and across courses. Results from their analysis indicated high consistency/ stability of the profile shape of the same instructor over time, regardless of course taught.

Hativa (1996) supported Marsh and Bailey's (1993) research concerning stability of profile shape, but found during the second year of her study that there was a significant change in profile level. If teachers are expected to modify behaviors and teaching strategies based on student evaluations, one would then expect Hativa's results: ratings levels increase as a result of student evaluations.

Marsh (1981) examined the importance of the effect of the instructor on student evaluations compared with the effect of a particular class being taught. Marsh compared two different courses taught by the same instructor and the same course taught by two different instructors. The 1,364 courses were grouped in sets of four: same instructor-same course; same instructor-different course; different instructor-same course; different instructor-different course. The mean correlation coefficients are shown in Table 4.

[Insert Table 4 About Here]

Marsh (1981) argued that the large differences in correlation coefficients between same instructor-same course (.7) and same instructor-different course (.52), especially when compared to the much lower correlation between different instructor-same course (.14), demonstrates that the effect of the teacher on student ratings was much larger than the effect of the course being taught. Therefore, student evaluations reflected the effects of the instructor, not the course.

Marsh and Hocevar (1984) examined the consistency of student evaluations for university instructors who taught the same course at least four times in four years. Over 31,000 students evaluated 314 instructors and the pattern of ratings for one instructor were similar throughout the four years.

In the analysis of Marsh's (1983, 1984, 1993) SEEQ's, the majority of the items we have discussed are either teacher characteristics or behaviors (as compared to contextual or outcome factors.) More recent studies have explored a additional teacher characteristics and behaviors and the impact they have on the teaching-learning process.

Murray, Rushton, and Paunonen (1990) discovered that different personality traits contributing to effective teaching differed markedly for different psychology courses. Students in introductory and general undergraduate courses rated extroversion (extroverted, sociable, attention-seeking) and level of neurosis (neurotic, anxious, and harm avoiding) as important instructor qualities. Of low importance was negative affect, as well as factors such as aggressive, impulsive, and subjective. Similarly, honor and senior honor students in optional courses rated extroversion and liberalism high and negative affect low. In comparison, honor and graduate students in required courses rated instructor achievement (intelligent, intellectually curious, amibitious) high, gave average ratings to liberalism and negative affect, and considered neuroticism as relatively unimportant. However, no measures of student achievement were collected.

Renaud and Murray (1996) investigated the effects of aging on (1) teacher personalities and (2) student perceptions of teaching effectiveness. Teaching effectiveness was inversely related to age due to the change in specific personality traits associated with effective teaching. One limitation to this study was the fact that the judgments regarding personality were made by a peer group and then correlated to archival student evaluations. Indeed, personalities do change with age, and the change may impact teaching behaviors, but this research was not substantiated by current student evaluations of the instructors with regard to their personalities.

While these and other characteristics such as gender (;Boggs, 1995; Freeman, 1994), fashion (Morris, Gorham, Cohen, & Huffman, 1996), and race (Hendrix, 1995) have been investigated as important teacher characteristics related to effective teaching, it is important to consider the findings of 20 plus years of Marsh's research that have been substantiated by others more recently (Nasser & Fadia, 1997; Ryan & Harrison, 1995): the single most important factor in determining effective teaching is output, that is, measures of achievement or amount learned.

Additionally, one should note that if students have preconceived ideas as to what constitutes good teaching, teachers, as well, have similar beliefs. Feldman (1988) compared the opinions of college faculty with those of college students as to what teacher characteristics resulted in what one might call good teaching. Faculty and students agreed on nine points. They include:

1. Knowledge of the subject/discipline;
2. Course preparation and organization;
3. Clarity and understandability;
4. Enthusiasm for subject/teaching;
5. Sensitivity to and concern with students' level and learning progress;
6. Availability and helpfulness;
7. Quality of examinations;
8. Impartiality in evaluating students; and
9. Overall fairness to students.

Potential Bias

Many faculty members insist that student evaluations are biased by factors that they believe are unrelated to effective teaching (Marsh & Overall, 1979). The faculty employed at a major university was asked to pick from a list of 17 characteristics that they believed would cause bias to student evaluations. Eight characteristics were chosen by over 50% of the faculty: Course difficulty--72%, grading leniency--68%, instructor popularity--63%, prior student interest--62%, course workload--60%, class size--60%, reason for taking course--55%, and grade point average--53%. These same faculty members indicated a need for quality feedback regarding effective teaching, but regarded student evaluations with skepticism, and were even more critical of other tactics such as self-evaluation and classroom visitation.

It is well established that student evaluative ratings of instruction correlate positively with expected course grades (Greenwald & Gillmore, 1997). Statistical correction to remove the unwanted inflation of ratings produced by lenient grades may be warranted. Student instructional ratings provide data for three distinct functions: personnel decisions, instructional improvement, and information to students. It is to achieve fairness in personnel decisions that adjustments for grades, class size, and perhaps other variables, are potentially most useful and justifiable. Gigliotti and Buchtel (1990) stated that while potential bias exists, actual bias of student ratings is low. Whether student evaluations are biased or not is not the issue of the Marsh and Overall (1979) study. The issue is the perception of bias by teachers. If such methods of evaluation are perceived as biased, instructors will hesitate to use them as important sources of information to facilitate modifications in teaching strategies and teacher behaviors in an effort to make teaching more effective.

Marsh (1983, 1993) suggested that there are methodological weakness common to much of the research on potential biases of student evaluations. They include:

1. Using correlation to argue causation.

2. Neglecting the difference between practical and statistical differences. Conclusions should be based on an index of effect size as well as tests of statistical significance.

3. Failure to consider the multivariate nature of student ratings and potential biases.

4. Selection of inappropriate unit analysis. Applications of student evaluations are most often based on class average responses, so this would be the appropriate unit of analysis.

5. Failure to research the replicability of findings in a similar situation and to generalize to different situations. This is generally a problem with small sample sizes or classes from a single department or college.

6. The lack of an explicit definition of bias. If a characteristic does impact teaching effectiveness, and it is reflected in student ratings, then it is not a bias.

McKeachie (1973) stated that student evaluations could be better understood if researchers did not concentrate on explaining background characteristics in terms of biases, but rather examined the relationships between the two. For example, Marsh's (1983, 1984, 1993) work on class size shows this variable is correlated to group interaction and individual rapport, but not with any other SEEQ characteristics. The larger the class size, the less likely that interaction takes place and the less likely a student is to establish personal rapport with the instructor. Will these factors bias the evaluation?

Mateo and Fernandez (1996) noted that the literature indicated instructors were rated more favorably in small groups and that one would expect a tendency for ratings of teaching effectiveness to fall as class size increases. In very small classes (i.e.,< 10 students) both teacher effectiveness and motivational and interactional skills were rated higher, but as class size increased to very large the researchers found no effects were worthy of mention. Mateo and Fernandez agreed with Marsh that evaluation of the teaching world is complex and results of evaluations studies are difficult to define with simple models, but rather require consideration of the complex relationships among all factors involved.

Naftulin, Ware, and Donnelly (1973) hired a professional actor to lecture enthusiastically and expressively to a group of graduate students. The lecture was exciting, but lacking in content. Despite the lack of content, the actor received favorable ratings. The actor's expressiveness correlated to high ratings of instructor enthusiasm, which was the factor most logically related to the manipulation. Expressiveness also resulted in a higher overall general rating and a small effect on achievement. While some researchers merely labeled the study problematic (e.g., Frey, 1979), others conducted a review and re-analysis of the Dr. Fox study (e.g., Abrami, Leventhal, & Perry, 1982; Marsh & Ware, 1982; Ware & Williams, 1979). For example, Abrami et al. found that content manipulations had a greater effect on achievement and a smaller effect on student ratings.

Improving the Evaluation Process

As stated previously, despite the general use of student evaluations in university systems during the past thirty years, many teachers remain hostile (Davis, 1995) and often do not use the feedback they receive (L'Hommedieu, Menges & Brinko, 1997). Davis directed faculty members to explain the importance of the evaluation process to students, in an effort to elicit a fair and thoughtful response by students. Cook (1989) advocated training student raters in an effort to reduce the halo effect on Likert scale evaluation instruments and to help produce quality ratings. According to Cook, students trained on the importance of ratings and given information and opportunity to provide quality ratings, do so with regularity.

Cook (1989) hypothesized that ratings following rater training showed less leniency and halo error without rater training. Five weeks before the end of a regular semester course, student raters listened to a 15-minute presentation on leniency and halo effect and were given a copy of the rating instrument to review. In addition, the presentation stressed the importance and benefit of student evaluations. The students retained the evaluation instrument until the end of the course and then completed their evaluations. A control group of raters received no training and no evaluation instruments prior to the day of evaluation. Leniency was lower on all six evaluation scales for students who received training than for students who did not. In addition, there was less halo error in the trained ratings. Cook stated that (1) the training session was effective in reducing leniency and halo error; and (2) rater training could be used to reduce psychometric error in student evaluations of teacher performance.

Ideally, educators agree, there should be multiple sources of feedback to facilitate effective teaching. Rifkin (1995) suggested a combination of self-evaluation, peer review, evidence of student achievement, administrator and alumni evaluations, and finally, student evaluations. A holistic approach to teacher evaluations, that is multiple sources and types of data, should yield a more objective and complete picture of the teaching-learning environment and the teacher in question. Seldon and Angelo (1994) suggested the collection of specific data from four different evaluative sources: students, faculty, faculty peers, and administrators (see Table 5).

[Place Table 5 about here]

Student evaluations are included in another assessment alternative: the teaching portfolio. Many colleges and universities see portfolios as a more credible means of assessment (Defina, 1996; Rifkin, 1995). A teacher portfolio would include a narrative section that identifies the teacher's goals and philosophy of education. In collaboration with peers and department chairs, teachers describe, in writing, responsibilities, performance standards, course syllabi, instructional techniques, professional development and self-improvement activities. In addition, evidence of teaching effectiveness and student achievement should be included. Portfolios, updated annually, are reflective of change, growth, and general improvements taking place from year to year.

Another issue is when to collect evaluation data. Evaluations most often occur at the end of a course. Students do not see the benefits of their evaluations, and often instructors must wait to implement any recommended changes (Davis, 1995; Stockham & Amann, 1994; Timpson, 1997). Davis suggested periodic evaluations administered during the fourth, eighth. and twelfth weeks of a course. Evaluation instruments might focus on such topics as understanding course objectives and their relevance, perception of the organization and presentation of material, attitudes toward grading and testing, and teacher perceptions. This strategy gives not only periodic feedback, but an opportunity for students to witness the effect of their evaluations on the teaching-learning process.

A relatively new method of formative student evaluation, labeled facilitated student feedback, provides for student assessment of instructors at midcourse via a collaborative effort between student and teacher (Stockham & Amann, 1994). This feedback process is divided into four steps: 1) facilitator selection and preparation; 2) class brainstorming/feedback session; 3) facilitators report; and 4) teacher's response. During facilitated student feedback, student brainstorming helps to generate ideas that are then written before providing verbal responses to the facilitator.

In addition to providing midcourse evaluations, facilitated student feedback increases teacher rapport with students, allows students to see that others have similar problems, and presents the teacher as a partner in the teaching/learning process, rather than as an adversary (Stockham & Amann, 1994).

Abbott, Wulff, Nyquist, Ropp, and Hess (1990) employed Small Group Instructional Diagnosis (SGID) to determine student opinion regarding the effectiveness of student evaluations on the teaching-learning process. SGID also employs the facilitator approach. A facilitator met first with small groups of students at the midterm of a college course to identify relevant instructional issues. These student concerns were reported by the facilitator to the instructor and at the following class meeting, the instructor responded to the students' concerns and comments.

Since many colleges are committed to collecting students opinions of teaching and learning with standardized evaluation forms at the end of a course, Abbott et. al. (1990) compared student satisfaction of a midterm, facilitated interview evaluation with the widely accepted traditional evaluation method. Students preferred midterm evaluations, in any form, over end-of-term evaluations. Students were most satisfied with midterm student interviews followed by an extended teacher response and least satisfied with standardized questionnaires administered at the end of the term.

As the shift into the information age progresses, it is logical to expect a variety of computer-based student evaluation instruments to appear on the scene. Divoky (1995) has developed a program that employs microcomputers to collect and evaluate the data for student evaluations. Divoky described this data collection method as efficient and practical. Thinking back to Scriven's (1995) warning about long, boring evaluations, the concept of computer-based evaluations fits the parameters of quick, easy, and familiar. This technology also allows for the generation of individualized evaluation instruments (Davis, 1995).

Divoky's (1995) interactive teaching evaluation is comprised of two phases. In phase I, each student develops and uses an individualized evaluation instrument. One hundred fifty evaluation statements, derived from 20 years of research in the field of student evaluations, were grouped according to teaching and classroom dimensions. Thirteen categories were defined.

The evaluation process begins with students reviewing the thirteen categories and selecting those that they wanted to use in the evaluation of a specific class and instructor. Students were also asked to provide a performance and an importance measure for each item. These measures are based on a ten point Likert-type scale. The selected items were stored and tabulated. This completed Phase I of the process.

During Phase II, students reviewed the newly formatted database. Item listed were presented and students were asked to vote on which items should be retained. Voting continued until a majority was reached. Performance and importance factors were also selected and then approved by a majority.

Students responded in a positive manner to this evaluation tool. They liked seeing what other students deemed important concerning effective teaching and often asked if they could submit their own items to be included in the data base. Although the voting process was time consuming, it helped to produce a unique instrument for a unique setting, thereby increasing the value of the diagnostic material it provided (Cook, 1989).

Regardless of student training, or the timing or format of the student evaluations, the question of what type of evaluation questions, formative or summative, still remains. d'Apollonia and Abrami (1997) stated that student ratings of instruction measure general instructional skill, which is in fact a composite of subskills "of what the rater believes is all the relevant information necessary for making accurate ratings" (Nathan & Tippins, 1990, p. 291). In contrast to Marsh (1993), these researchers believe that global ratings or a simple score representing General Instructional Skill should be used for summative evaluations. Moreover, specific ratings of instructors' effectiveness add little to the explained variance beyond that provided by global ratings (Cashin & Downey, 1992; Hativa & Raviv, 1993).

Cashin and Downey (1992) proposed the utilization of two global items for summative purposes: one to evaluate the instructor and the second evaluating the course. They found that such items accounted for more than 50% of the variance of weighted-composite criterion. They suggest that students respond to two global questions for summative purposes (using a 5-point Likert scale of 1 = definitely false and 5 = definitely true):

1. Overall, I rate this INSTRUCTOR an excellent teacher;
2. Overall, I rate this an excellent COURSE.

Abrami (1989) compared Marsh's multidimensional factor scores with his idea of overall or total ratings. Abrami viewed both ratings as multidimensional, but stated that they should relate to effectiveness criteria differently. Global ratings correlate more highly with student learning than do factor scores, with the exception of teaching skill, which is a fairly global rating factor itself. Bias appeared to be no more severe in global ratings than with factor scores, although the rationale for the bias is more difficult to discern.

In summary, both instructors and students found formative evaluations more suited to an ongoing interactive feedback mechanism that facilitated modifications to teaching and course content. Both Abrami (1989) and Marsh (1993) agree that global items are not valuable for formative evaluations. Additionally, they agree that weighted factor scores are superior to unweighted factor scores or individual item ratings.

While a host of concerns address the improvement of student evaluations of teachers, the primary concern remains how to use specific evaluative techniques. If formative evaluation is the goal, then multidimensional analysis containing weighted factor scores would be the method of evaluation. On the other hand, global questions would prove suitable and even preferable for summative evaluations. Finally, as noted by several researchers (e.g., L'Hommedieu, Menges & Brinko, 1990; Murray, Rushton, and Paunonen, 1990), one must always determine the reliability, validity, and utility of the evaluative instrument at the local institution at which it is used. L'Hommedieu et al. (1990) see little use to widening the scope of evaluation instruments. Rather, replication in this area is important for the purpose of establishing the usefulness of student ratings for instructional improvement. However, this is directly opposite the approach utilized at North Carolina State University (1994). At that institution individual department faculties both determined the design of the instrument as well as its use, either formative or summative.

Conclusions

As discussed in this paper, student opinion continues to be a major factor in the evaluation of teacher effectiveness at institutions of higher education. Regardless of the weight of such evaluations, or the form they take, we would be wise to consider the warnings and suggestions offered by Seldin and Angelo (1997) concerning the development and administration of student evaluations. They warn that the entire evaluation system, including any rating forms, should be designed to meet a specific set of specific purposes and needs and should have a clear connection to the academic rewards system. Faculty should be involved in the development of the entire process as they are likely to be skeptical and are more likely to utilize data if they assist in the development process (see Table 6).

[Place Table 6 about here]

Student ratings add a valuable component to the range of feedback for the evaluation of effective teaching. Based on the work with the SEEQ, Marsh and Roche (1997) summarized student evaluations as:

1. multidimensional;
2. reliable and stable
3. primarily a function of the instructor who teaches the course rather than the course;
4. relatively valid against a variety of indicators of effective teaching;
5. relatively unaffected by a variety of variables hypothesized as potential biases (e.g., grading leniency, class size, workload, prior subject interest); and
6. useful in improving teaching effectiveness when coupled with appropriate consultation.

In a review of the literature, McKeachie (1997) determined Marsh's work on multidimensionality is correct when the goal is formative evaluation for instructional improvement. However, he agrees with d'Apollonia and Abrami (1997) that a single global rating is more appropriate for summative evaluations made by personnel committees. Regardless of the type of evaluation questions used, Marsh and Roche (1997) and Murray et al. (1990) found support for the need to consult on specific teaching behaviors if improvement is to be made. Corrections and improvement can be facilitated by providing evaluations early in the learning process to aid instructors in making corrections.

Finally, according to McKeachie (1997), the burden of appropriate administration and use of evaluative tools falls on the shoulders of the personnel committee. The personnel committee should never be concerned with comparing teachers, but rather should use student evaluations to measure teaching effectiveness.

Two issues that can be addressed by the institution itself are grading inflation and class size. McKeachie (1997) suggested that grading inflation bias may be a function of the type of institution and should be a concern of the personnel committee. Additionally, if effective teaching is more likely to take place in smaller classes, then the personnel committee can again address this issue, if indeed effective teaching is their ultimate goal.

To reiterate what was presented earlier, the generation of evaluative scores may not be a real problem facing personnel committees. Rather, it is the use of these ratings by said committees that is, in fact, the larger problem that must be addressed. When and how student evaluations are used in the decision process by personnel committees is the major concern; not what necessarily composes the evaluation.

Student evaluation of teachers is just one component of an important process. Most higher education faculty believe that the teaching-learning process is an active, ever-changing interaction between the student, teacher, and environment (input and process within context) that should be focused on a particular outcome. It is therefore reasonable to expect students to make important contributions to this system and it is imperative that teachers be receptive to student feedback. However, there is no one correct method of teaching (Joyce & Weil, 1996). In fact, teaching and learning are contextual by nature, with different methods showing superiority for different outcomes and different students. This diversity of student characteristics, faculty objectives and preferred teaching methods, and institutional contexts must be considered when developing the evaluation process.

References

• Abbott, R. D., Wulff, D. H., Nyquist, J. D., Ropp, V. A., & Hess, C. W. (1990). Satisfaction with processes of collecting student opinions about instruction: The student perspective. Journal of Educational Psychology, 82(2), 201-206.
• Abrami, P. C. (1989). How should we use student ratings to evaluate teaching? Research in Higher Education. 80(2), 221-227.
• Abrami, P. C., d'Apollonia, S., & Cohen, P. (1990). Validity of student ratings of instruction: What we know and what we do not. Journal of Educational Psychology, 82(2), 219-231.
• Abrami, P. C., Leventhal, L., & Perry, R. P. (1982). Educational seduction. Review of Educational Research, 52, 446-464.
• Aleamoni, L. M. (1981). Student ratings of instruction. In J. Millman (Ed.). Handbook of teacher evaluation (pp. 110-145). Beverly Hills, CA: Sage.
• Boggs, C. (1995). Gender bias in teaching evaluations: Insights offered by communication research. (ERIC Document Reproduction Service No. ED387848)
• Boyer, E.,& Rice, E. (1990). The new American scholar. Princeton, NJ: Carnegie Foundation for the Advancement of Teaching.
• Cashin, W. E.,& Downey, R. G. (1992). Using global student ratings items for summative evaluations. Journal of educational Psychology, 84(4), 563-572.
• Cohen, P. A. (1981). Student ratings of instruction and student achievement: A meta-analysis of multisection validity studies. Research in Higher Education, 51, 281-309.
• Cook, S. (1989). Improving the quality of student ratings of instruction: A look at two strategies. Research in Higher Education, 30 (1), 31-45.
• Cranton, P., Smith, R. A. (1990). Reconsidering the unit of analysis: A model of student ratings of instruction. Journal of Educational Psychology, 82(2), 207-212.
• d'Apollonia, S., & Abrami, P. C. (1997). Navigating student ratings of instruction. American Psychologist, 52(11), 1198-1208.
• Davis, M. (1995, March). Staging a pre-emptive strike: Turning student evaluations of faculty from threat to asset. Paper presented at the annual meeting of the Conference on College Composition and Communication. Washington, DC.
• Defina, A. (1996). An effective alternative to faculty evaluation: The use of the teaching portfolio. (ERIC Document Reproduction Service No. ED394561)
• Divoky, J. (1995). Eliciting teaching evaluation information interactively. Journal of Education for Business, 70, (6), 317-32.
• Eley, M.,& Stecher, E. (1997). A comparison of two response scale formats used in teaching evaluation questionnaires. Assessment and Evaluation in Higher Education, 22, (1), 65-70.
• Feldman, K. A. (1988). Effective college teaching from the students' and faculty's view: Matched or mismatched priorities? Research in Higher Education, 28(4), 291-344.
• Feldman, K. A. (1989). Instructional effectiveness of college teachers as judged by teachers themselves, current and former students, colleagues, administrators, and external (neutral) observers. Research in Higher Education, 30(2), 137-169.
• Freeman, H. (1994). Student evaluations of college instructors: Effects of type of course, gender and gender role, and student gender. Journal of Educational Psychology, 86, (4), 627-630.
• Frey, P. W. (1979). The Dr. Fox effect and its implications. Instructional Evaluation, 3, 1-5.
• Gage, N. L.,& Berliner, D. C. (1992). Educational psychology, (5th ed.). Boston, MA: Houghton, Mifflin Company.
• Gigliotti, R. J., & Buchtel, F. S. (1990). Attributional bias and course evaluations. Journal of Educational Psychology, 82(2), 341-351.
• Greenwald, A. G. (1997). Validity concerns and usefulness of student ratings of instruction. American Psychologist, 52(11), 1182-1186.
• Greenwald, A. G., & Gilmore, G. (1997). Grading leniency is a removable contaminant of student ratings. American Psychologist, 52(11), 1209-1217.
• Harrison, P. D., Ryan, J. M., & Moore, P. (1996), College students' self-insight and common implicit theories of rating of teaching effectiveness. Journal of Educational Psychology, 88, (4), 775-782.
• Hativa, N. (1996). University instructors' ratings profiles: Stability over time, and disciplinary differences. Research in Higher Education, 37, (3), 341-365.
• Hativa, N., & Raviv, A. (1993). Using a single score for summative evaluation by students. Research in Higher Education, 34, 625-646.
• Hendrix, K. G. (1995, November). Student perceptions of the influence of race on professor credibility. Paper presented at the annual meeting of Speech Communication Association. San Antonio, Texas.
• Huitt, W. (1995). A systems model of the teaching/learning process. Educational Psychology Interactive. Valdosta, GA: College of Education, Valdosta State University.
• Joyce, B., & Weil, M. (1996). Models of teaching (5th ed.). Boston: Allyn & Bacon.
• L'Hommedieu, R., Menges, R., & Brinko, K. Methodological explanations for the modest effects of feedback from student ratings. Journal of Educational Psychology, 82(2), 232-240.
• Marsh, H. W. (1981). The use of path analysis to estimate teacher and course effects on student ratings of instrument effectiveness. Applied Psychological Measurement, 6. 47-60.
• Marsh, H. W. (1983). Multidimensional ratings of teaching effectiveness by students from different academic settings and their relation to student/course/instructor characteristics. Journal of Educational Psychology, 75, (1), 150-166.
• Marsh, H. W. (1984). Students' evaluations of university teaching: Dimensionality, reliability, validity, potential biases, and utility. Journal of Educational Psychology, 76, (5), 707-754.
• Marsh. H. W. (1993). Multidimensional students' evaluations of teaching effectiveness. Journal of Higher Education, 64, (1), 1-18.
• Marsh, H. W., & Bailey, M. (1993). Multidimensionality of students' evaluation of teaching effectiveness: A profile analysis. Journal of Higher Education, 64, (1), 1-15.
• Marsh, H. W., & Hocevar, D. (1984), The factorial invariance of students' evaluations of college teachers. American Educational Research Journal, 21, 341-366.
• Marsh, H. W., & Overall, J. U. (1979). Long-term stability of students' evaluations. Research in Higher Education, 10, 139-147.
• Marsh, H. W., & Roche, L. A. (1997). Making students' evaluations of teaching effectiveness effective: The critical issues of validity, bias, and utility. American Psychologist, 52(11), 1187-1197.
• Marsh, H. W., & Ware, J. E. (1982). Effects of expressiveness, content coverage, and incentive on multidimensional student rating scales. Journal of Educational Psychology, 74, 126-134.
• Mateo, M. A., & Fernandez, J. (1996). Incidence of class size on the evaluation of university teaching quality. Educational and Psychological Measurement, 56, (5), 771-778.
• McIlrath, D., & Huitt, W. (1995, December). The teaching-learning process: A discussion of models. Educational Psychology Interactive. Valdosta, GA: Valdosta State University.
• McKeachie, W. J. (1973). Correlates of students' ratings. In A. L. Sockloff (Ed.). Proceedings: The first invitational conference on faculty effectiveness evaluated by students (pp. 213-218). Temple University.
• McKeachie, W. J. (1990). Research on college teaching: The historical background. Journal of Educational Psychology, 82(2), 189-200.
• McKeachie, W. J. (1997). Student ratings: The validity of use. American Psychologist, 52(11), 1218-1225.
• Morris, T. L., Gorham, J., Stanley, H. C., & Huffman, D. (1996). Fashion in the classroom: Effects of attire on student perceptions of instructors in college classes. Communication Education, 45, (2), 135-147.
• Murray, H. G., Rushton, J. P., & Paunonen, S. V. (1990). Teacher personality traits and student instructional ratings in six types of university courses. Journal of Educational Psychology, 82(2), 250-261.
• Naftulin, D. H., Ware, J. E., & Donnelly, F. A. (1973). The Doctor Fox lecture: A paradigm of educational seduction. Journal of Medical Education, 48, 630-635.
• Nasser, F., & Glassman, D. (1997). Student evaluation of university teaching: Structure and relationship with student characteristics. (ERIC Document Reproduction Service No. ED407390)
• North Carolina State University handbook for advising and teaching. (1994). Available online: [http://www2.ncsu.edu/ncsu/provost/info/hat/current/ch10/0105.html]
• Paulsen, M. B., & Feldman, K. A. (1995). Taking teaching seriously: Meeting the challenge of instructional improvement. (ERIC Document Reproduction Service No. ED396615)
• Renaud, R. D., & Murray, H. G. (1996). Aging, personality, and teaching effectiveness in academic psychologists. Research in Higher Education, 37, (3), 323-339.
• Rifkin, T. (1995). The status and scope of faculty evaluation. (ERIC Reproduction Service No. ED385315)
• Ryan, J. M., & Harrison, P. D. (1995). The relationship between individual characteristics and overall assessment of teaching effectiveness across different instructional contexts. Research in Higher Education, 36, (5), 577-594.
• Scriven, M. (1995). Student ratings offer useful input to teacher evaluations. (ERIC Reproduction Service No. ED39824)
• Seldin, P., & Angelo, T. A. (1997). Assessing and evaluating faculty: When will we ever learn? (To use what we know). Proceedings of the AAHE 1997 Conference on Assessment and Quality Assessing Impact: Evidence and Action.
• Simpson, R. D. (1995). Uses and misuses of student evaluations of teaching effectiveness. Innovative Higher Education, 20, (1), 3-5.
• Stockham, S. L., & Amann, J. F. (1994). Facilitated student feedback to improve teaching and learning. Journal of Veterinary Medicine, 21, (2), 51-55.
• Timpson, W. W. (1997). Rethinking student evaluations and the improvement of teaching: Instruments for change at the University of Queensland. Journal of Higher Education, 22, (1), 55-65.
• Ware, J. E., & Williams, R. G. (1979). Seeing through the Dr. Fox effect: A response to Frey. Instructional Evaluation, 3, 6-10.
• Whitely, S. E., & Doyle, K. O. (1976). Implicit theories in student ratings. American Educational Research Journal, 13, 241-253.
• Wright, P., & Rip, P. D. (1981). Retrospective reports on the causes of decisions. Journal of Personality and Social Psychology, 40.

Table 1. Percentage of studies appraising student ratings of instruction as valid, biased, or neither

Table 2. Evaluation items included in the Students' Evaluations of Educational Quality (SEEQ)

Table 2. Evaluation items included in the SEEQ (continued)

Table 3. Relationships between students' evaluations of teaching effectiveness and background characteristics

Table 3. Relationships between students' evaluations of teaching effectiveness and background characteristics (continued)

Table 4. Correlations of instructor profiles based on student evaluations comparing instructors and courses taught

Table 5. Potential sources of data for faculty evaluation

Table 5. Potential sources of data for faculty evaluation (continued)

Table 6. Suggestions for appropriately evaluating faculty