Bob Petrulis

         What do grades mean? What purposes do they serve? How should they be assigned, and what roles should instructors, students, and institutions play? Evaluating and grading student learning is a contentious issue. When imposed on a collaborative educational setting, the limitations, paradoxes, and difficulties of the traditional grading system are magnified.

         This article draws on the insights of a number of colleagues whom I interviewed to gain a view of the practical issues that attend the grading process in collaborative classrooms. The issues are complex and my colleague informants have each developed unique ways to resolve or mitigate them, as called for by the individual circumstances that confront them. Therefore, my purpose here is not to provide a cookbook for grading, but to lay out the issues and illustrate a range of possible approaches.

         Collaborative classrooms have many distinctive features. Most instructors recognize that when students are engaged with one another and with the course content-when they "own" their learning process-they learn and retain more. To enhance learning, instructors have developed pedagogical and curricular strategies, such as group projects and student-led discussion groups, which increase students' opportunities to work collaboratively and to actively construct the knowledge they take away from their courses.

         Although courses based on learning community curricular models have gained a foothold at many colleges and universities, the vast majority of classes at most institutions are presented in the traditional single instructor format, and within this structure, many instructors effectively employ collaborative techniques. While gathering information for this article, I talked with teachers of both traditional "discrete" courses and learning community programs. Many issues are common to both, but learning community programs add the complexities of interdisciplinary study, team-teaching, and many forms of coordination among several instructors. Issues unique to team-taught, coordinated studies programs are discussed in the final section of this article.

Grading in the Collaborative Classroom

         A cornerstone of collaborative education is that the student takes an ever-increasing role in directing her own learning. At term s end, however, most instructors must conform to the expectation that students be given grades which report their academic performance.

         In the collaborative classroom, the tension between the goal of nurturing student self-direction and the requirement for an external (to the student) evaluation of performance intensifies the dilemmas and problems of grading students. When the instructor uses her own criteria to judge students' work, the chance that students will perceive the value of their own evaluations is reduced. When the grade becomes a central motivator, students become mercenaries, instructors become frustrated, and in the end everybody becomes cynical.

         It is possible, however, to lessen the effect of this conflict, and even to channel it in positive ways which promote learning. Students can be brought into the process as evaluators of their own and someaspects of each others' work, as evaluators of the course itself, and even to help develop the criteria by which the evaluations will occur.

         Jim Harnish's collaborative preparation for exams (described elsewhere in this Handbook) provides one strategy. Susan Wyche-Smith s "Inkshedding," (also in this Handbook) in which students can make known their own judgments about course content or process, provides another.

         Even when instructors are comfortable with a traditional, authority-based role in the grading process, they express concern about the meaning of the grades they give. Is a B+ given to one student equivalent to the B+ given to another? And what, precisely, does any grade say about the students'academic performance abilities? Concerns about inconsistency in grading policies and grade inflation are among the perennial issues that all teachers must confront. Few institutions have maintained the sustained, campus-wide conversations needed to develop consistent responses to these questions.

         Recent calls for increased rigor in grading have highlighted some basic assumptions about how grades should be determined and awarded. For example, one current proposal, that transcripts should show the average grade for each course, is based on the supposition that a B earned in a course whose mean grade is C is "worth" more than one in which the mean grade is B, and that classes in which relatively few high grades are given are more rigorous and challenging than those with higher grade means. This goes to the heart of one of the basic questions that any instructor must answer when designing a grading scheme: should grades be given on a norm- or criterion-referenced basis?

Norm-Referenced vs. Criterion-Referenced Grading

         Because the differences between these approaches have far-reaching implications for pedagogy, curriculum, and course design, a short review and comparison will be helpful here. Norm-referenced grading, also known as grading on the curve, means that students are compared to one another and graded relative to the class s performance as a whole. Using this approach, student performance on assignments and tests is distributed along a normal (bell-shaped) curve.

         Norm-referenced grading schemes typically use the mean score for the class as the mid-point (given a grade of C). A pre-determined number of students receive each grade, A through F, as illustrated in Figure 1.

         Few of those instructors who use norm-referenced grading apply it with complete consistency. Confronted with a particularly bright class, most instructors balk at penalizing good students for having the bad luck to enroll in a class with a cohort of other very capable students. On the other hand, few instructors would reduce the mean grade for a class with a larger than usual share of lower-ability students. A serious problem with norm-referenced grading is that, no matter what the class s level of knowledge and ability, and no matter how much they learn, a predictable proportion of students will receive each grade. Since its essential purpose is to sort students into categories based on relative performance, norm-referenced grading and evaluation is often used to winnow students for limited places in selective educational programs.

         Norm-referenced grading promotes competition among students, since they must vie for a scarce resource (grades above the mean). This may be the most significant reason that few, if any, instructors use this approach in collaborative classrooms. While individual students may be encouraged to study and perform to the best of their abilities, they are also discouraged from helping their fellows, since any improvement in others performance raises the class mean and thus lowers one s own grade. Conversely, the lower everyone else s scores, the higher the individual s grade. When norm-referenced grading is used, students are given every incentive not to collaborate with one another.

         A more subtle problem with norm-referenced grading is that a strict correspondence between the evaluation methods used and the course s instructional goals is not necessary to yield the required grade distribution. The specific learning objectives of norm-referenced classes are often kept hidden, in part out of concern that instruction not "give away" the test or the teacher s priorities, since this might tend to skew the curve.

         (Of course, as a practical matter, the vast majority of instructors strive to evaluate their students' performance fairly regardless of the grading method used, but this often reduces to treating each student the same, rather than carefully relating the goals of the course to students' demonstration of abilities.)

         So, if norm-referenced grading is so fraught with problems, what is the alternative?

         Although the instructors I interviewed have developed a wide array of grading policies and procedures, all of them are based on the criterion-referenced approach. As the name implies, this approach establishes clear evaluation criteria up front, and compares student performance to them. Theoretically, if all students in the class meet the stated performance criteria, all could receive a grade of A.

         In fact, if the ideal grade distribution in the norm-referenced class is the bell-shaped curve with a mean grade of C, the ideal in the criterion-referenced scheme is that all students earn A s (Figure 2). I should quickly point out, however, that in practical terms, criterion-referenced grade distributions tend to look like those of other courses. An entire class of students seldom has the ability, motivation, and circumstances to earn straight A s. In addition, remember that criterion-referenced grading is not exclusive to courses which incorporate collaborative teaching methods. It is in wide use throughout the educational system.

         Instructors in criterion-referenced classes need not worry about skewing the curve or giving students "unfair advantage" because their cards are on the table from the beginning, and because the students are not competing with each other for high grades. The instructor can consider individual or subgroup needs and devise strategies which maximize the learning opportunities presented to each individual student. Thus it is possible for the instructor s role to shift from that of aloof, powerful authority to one of coach, facilitator, and collaborator.

         When students are evaluated based on predefined criteria, they are freed to collaborate with one another and with the instructor. With criterion-referenced grading, a rich learning environment is to everyone s advantage, so students are rewarded for finding ways to help each other, and for contributing to class and small group discussions.

         What prevents instructors who use criterion-referenced grading from setting the performance criteria so low that everyone can pass with ease? Instructors certainly experience some pressure from students and perhaps administrators to do so. However, when criteria are public, as they must be in collaborative classrooms, they may be scrutinized (and criticized) by any member of the educational community. Further, when the goals are clear, students often want to be held to high standards. Thus, collaborative approaches require that instructors not only know what the criteria are, but also articulate them so that all members of the class understand them well.

         I hope it is clear by now that the choice of a basic grading approach affects the day-to-day conduct of the classroom in fundamentally significant ways, and that criterion-referenced grading opens a larger universe of possibilities to the instructor because the students can engage with each other to enhance their learning.

Evaluating the Work of Groups

         Few tools in the collaborative learning kit are as potentially powerful as small groups. Employing small groups and group projects creates possibilities for interaction and learning that could not be accomplished easily in other ways, as many of the articles in this Handbook illustrate.

         However, grading students group work has controversial aspects. There is an ethic prevalent in academe and in American society at large which holds that students should be evaluated on the basis of their own individual work, and not that of others. This sensibility is challenged when students are evaluated on the basis of their interaction in a group, or the quality of a jointly-created product. In addition, students are understandably hesitant to place a portion-however small-of their educational fates in the hands of their peers until bonds of trust have formed.

         When students are asked to work collaboratively and to submit group products for a grade, instructors must confront inevitable differences among students and the conflicts that often erupt. Highly motivated students often resent being grouped with less committed peers. Shy students sometimes feel overwhelmed by more assertive group members. Language, cultural, and social differences, and the resulting status issues can wreak havoc with group dynamics.

         Instructors in collaborative classrooms use many methods for forming groups. These fall into three basic categories: student-selected, random, and instructor-selected. A more complete treatment of forming and using groups is beyond the scope of this article, but a short discussion of the implications for grading follows. Self-selected groups offer the greatest amount of autonomy to students but may tend to isolate subgroups of students from one another. Random assignment to groups can overcome this problem, although the laws of chance argue that some awkward combinations will result. One coordinated studies course used this method to create heterogeneous small groups until a quarter in which most of the African-American students were assigned to the same group.

         Instructor-selected grouping can ensure that each group contains what the instructor expects will be a productive mix of student characteristics. The criteria used to form such groups vary. One instructor uses the Myers-Briggs personality type indicator to sort students. He tries to assign students so there is a mix of personality profiles, and other characteristics in each group. Others mix academic ability, age, learning styles, or other characteristics. Another strategy is to group students with certain characteristics together. For example, one colleague regularly puts talkative and quiet students into separate groups.

         Grading group performance requires that the instructor be clear about the ways instructional groups relate to course objectives. Will each individual group member produce a product to be evaluated or will there be only a collective product? Does the instructor grade the group s product as a whole? How much do individual contributions to the group effort count? If learning to work in groups is a course objective, should the instructor attempt to monitor and evaluate the dynamics of the group? What is an appropriate mix of individual and group grades, and will student evaluations of themselves and their groups be part of the process? A variety of factors come into play as instructors design the evaluation process for group work. Figure 3 shows a representation of three continua on which some of the most significant design considerations can be located.
0: Representation of three continua on which some of the most significant design considerations can be located.

         The first dimension (a) represents the amount of control the instructor retains or gives to the students for determining grades. Some instructors have developed innovative ways to share responsibility with their students for evaluation of group projects. Several have developed checklists or evaluation schemes for students to use. A few develop the evaluation criteria jointly with their students, either applying the criteria to assign grades or having students use the criteria to grade themselves and each other. Such evaluation schemes often include criteria related to group process, individual contributions to the group, and the product itself.

         Dimension (b) of Figure 3 indicates whether the grades assigned to group members are based on criteria related to individual or group performance. Some instructors elect to give one grade to all participants based on the group s output, regardless of individual effort. Others grade only individual products or effort. Others find a mixture of individual and group criteria to be most effective.

         The third dimension (c) of Figure 3 shows the weight given to process versus product outcomes. That is, will the students be graded on the quality of their work in the groups, for attendance, contribution, and so forth? How much will this influence the grade as compared to a tangible product of the group s effort? Some groups are formed mainly for the purpose of providing an arena for students to discuss difficult educational topics which might not be fully aired in open class, and the product might be an ungraded verbal report to the class. In such cases, to paraphrase Woody Allen, 90% of the grade is for showing up. Other groups are given more formally structured assignments where the product is the focus. In that case, group dynamics might not be directly measured at all.

Portfolios

         Many instructors have incorporated portfolios into their courses. A portfolio consists of a collection of students' work compiled over the course of the term. Depending on the instructor s purposes, it might include the student s best work, or a record of the student s development, such as successive drafts of a paper. In addition, students may be asked to include a self-evaluation or other reflections on their experiences as learners. Some instructors base their grade exclusively on the portfolios their students submit.

Grading Attendance

         The collaborative classroom depends on promoting student involvement with the subject matter, with other students, and with the instructor. Since students are actively engaged in the process of creating the knowledge they will gain from the course, they must be present in class for discussions and group work. Whether in small groups or in the class as a whole, the entire enterprise depends on interaction.
         Several of the instructors I interviewed figure attendance into the grade to some extent. Typically, attendance might contribute from 5% to 15% to the final grade. Other instructors cite the logistical difficulty of keeping track of attendance, or the fact that their students are adults and should be treated as such, and say they do not factor attendance into the grade directly. However, contribution to discussions or participation in small groups is usually a component of the grading scheme, and this, of course, requires that students attend class.

Student Self-Evaluation

         If a central purpose of collaborative learning is to help students become independent, self-directed learners, then a significant instructional goal is to help them develop a capacity for self-evaluation. Several of my informant-colleagues have incorporated some form of self-evaluation into their courses and learning community programs.

         The evaluations are not always factored into the course grade directly. Asking students to evaluate their own performance can help them develop insights into their learning. Instructors may use the students' self-assessments as one guide among several as they evaluate student work.

         One instructor allows his students to choose the amount that self-evaluations and peer-evaluations will influence their grade. Others specify a proportion of the final grade which will be determined by self-evaluation.

Grading in Learning Community Programs

         Learning communities which involve more than one instructor in a teaching team add another dimension of complexity to the design of evaluation and grading schemes. Instructional teams must resolve questions about their collaboratively designed curriculum and pedagogy. They must coordinate schedules, assignments, policies, and so forth. When grading assignments, they must be in agreement about the grading criteria to be used so that all instructors give consistent grades.

         The teaching teams I interviewed, all of which had conducted several learning community programs, reported that they had been able to achieve consensus about grading and other pedagogical issues. However, conflicts within teaching teams can arise from a complex of differences of personality, style, approach, and purposes. Some of these may be lessened by making underlying assumptions and intentions explicit. Teams sometimes begin this process by taking the Teaching Goals Inventory. This framework allows them to explore their individual purposes and goals, and agree on the goals for the learning community. It reveals wide variety in instructional purposes and values even among instructors teaching sections of the same course. The Teaching Goals Inventory can be useful because it helps instructors identify a wide variety of types of goals. Less formal processes can also be productive, such as simply discussing the kinds of student performance the team wants and ways it might be evoked.

One Course Grade, or a Grade For Each Component?

         As part of the program design process, instructors in team-taught learning communities must decide whether to award a single grade for all course components or different grades for each. Both approaches have staunch advocates who provide convincing arguments in favor of their positions.

         Those who advocate a single grade for all components of the learning community say that an interdisciplinary program should be seen as a unit, even though students are given college credit as though they have taken several individual courses, and that it is fruitless-as well as wrong-headed-to try to tease the individual disciplinary strands apart. A single grade for all the student s work in a coordinated studies program honors the holistic character of the learning community approach and allows instructors to collaborate much more closely as they develop expectations for student performance, grading criteria, and ways they will evaluate it.

         Many learning community programs are broken into seminar sections of 15 to 25 students each which meet regularly with one member of the teaching team. The seminar instructor takes primary responsibility to grade her students, although students may appeal their grades to other members of the teaching team.

         On the other hand, partisans of the separate-grade-for-each-course approach argue that students should not be unfairly rewarded or penalized for performance in one component of the program on the basis of their performance in others. For example, if one component of the program is English composition, students who are good writers but who find it difficult to speak up during discussions should not be given a grade in the composition component which is heavily influenced by their oral communication skills. To do so, these instructors believe, is unfair to the student and presents a misleading picture on the student s transcript. It also fails to take account of the expertise of individual instructional team members. How, they ask, can an historian grade an assignment related to literature, or a literature instructor grade a history assignment?

         In some cases, assignments are given more than one grade-one grade for compositional aspects and another from a discipline-specific perspective, for example. However, at least some students report being confused and even upset by this practice. They find it difficult to resolve the conflicts they experience when writing to satisfy multiple sets of evaluation criteria, and for some, this may impede learning. Whether this is a problem from the instructors' perspective depends on their course objectives.

         The individual grade approach is seldom strictly applied. Instructors usually consult with one another about borderline students even though they may retain the ultimate prerogative to assign the student s grade in "their" component of the program. And it is common practice to base some portion of the final grade on a jointly evaluated assignment. So, course component grades might be based, for example, on individual work for 75% of the grade, and on a group project which crosses disciplinary lines for the remaining 25% of the grades in all components.

         Whether a learning community team takes one approach or the other depends on the underlying assumptions used by course designers as they make curricular decisions. In the case of the teams I interviewed, when coverage of standardized content or development of specific skills loomed large in instructors' concerns, the latter approach was used; teams used the single-grade approach when encouraging students to bring many threads of thought together and to embark on an intensely personal, creative educational journey.

Can Students Withdraw from One Course Component?

         A related issue is that of course withdrawals. In some cases, students are given an all-or-nothing choice: either take all components of the course or withdraw from the course completely. They are not allowed to withdraw from one component while continuing in the remainder of the course. The reasoning behind this policy is that the course components are so inextricably intertwined that the course s integrity is lost if one component is missing. However, of concern when this policy is used is that it raises a barrier to enrollment in the course in the case where a student s financial aid is dependent upon maintaining a specified credit load. Students may perceive taking the course as too risky.

         Other instructors not only allow students to withdraw from a component of the course in which they are performing poorly, some actively encourage students to do so when the circumstances warrant. This can provide an out to students who are overwhelmed by the course work, but who could do well if they concentrate on areas of relative strength.

Conclusions

         Although there are serious limitations and problems with grading as the currency of learning, instructors in collaborative settings have developed productive ways to approach grading in which students become owners of the evaluation criteria to be used. When instructors take the role of coach or facilitator, rather than dispassionate judge, the grading scheme must change in kind. In the best cases, as illustrated throughout this Handbook, assessment and grading itself becomes part of the teaching/learning process, and is harnessed to reinforce learning objectives.

back to the Table of Contents