EVALUATING STUDENT LEARNING:

Grading

Instructors may use a variety of grades during the quarter, from points, to letters, to percentages; all but a few schools and colleges use the University's four-point decimal system to assign final course grades.

During your department's orientation you will probably be informed of the average grade given in the department as a whole and in the particular course(s) where you have responsibilities. These norms are usually seen as guides rather than mandates, but many Teaching Assistants find them useful during the first year.

The most current scale with letter equivalents, as defined by the Faculty Senate, follows:

		4.0
		3.9     A
		3.8
		3.7
		3.6
		3.5     A-
		3.4 
		3.3 
		3.2     B+
		3.1 
		3.0 
		2.9     B
		2.8 
		2.7 
		2.6 
		2.5     B-
		2.4 
		2.3 
		2.2     C+
		2.1 
		2.0 
		1.9     C
		1.8 
		1.7 
		1.6 
		1.5     C-
		1.4 
		1.3 
		1.2     D+
		1.1 
		1.0 
		0.9     D
		0.8 
		0.7     D- (lowest possible passing grade)

		

All grades below 0.7 are recorded as 0.0 on a student's official record. Other features of the UW grading system are described in detail in the University Catalog, and you will want to familiarize yourself with them. http://www.washington.edu/students/gencat/front/grading-sys.html

Determining Grades

Grading is an extremely complex task. Grades do not exist in a vacuum, but are part of the instructional process and serve as a feedback loop between instructor and student. It follows, then, that grading policy should be consistent with the learning objectives for the course. If you have responsibility for planning the course, including devising exams, quizzes or assignments, you will want to keep the connection between your grading policy and your learning objectives firmly in mind. If, on the other hand, your role is less comprehensive, it is particularly important to discuss grading standards and policies with the professor, course coordinator, or TA coordinator before you begin your grading duties. It is also important to familiarize yourself with your department's grading norms and policies.

Whether or not you are ultimately responsible for assigning course grades, students will ask you about grading policies and criteria. In order to be prepared for such inquiries, ask yourself first, whether your grading strategy is based on an independent judgment of each student's achievement of course goals (mastery-based or criterion-referenced grading), or whether the grade will be based on the student's performance relative to other students in the class (norm-referenced grading).

Mastery-Based (Criterion-Referenced) Grading

If you concentrate solely the students' mastery of the material, you are using what is sometimes called mastery-based or criterion-referenced grading. Each student's grade is determined by comparing her/his accomplishments to a pre-set list of criteria. In this form of grading, the interpretation of a student's performance is in no way influenced by the performance of other students. Thus, you are not concerned about how widely the scores are distributed. In using the system, you may find yourself awarding all A's -- if all students get over 90 on an exam, for example, or if they all demonstrate understanding of a given concept or methodology in an essay, paper or research project. If, on the other hand, they all get scores below 80, and if you have beforehand said "80 and above is a B," then you will not be giving any A's or B's. The underlying assumption is that it is more important to assess how well each student achieves the specific course goals rather than how much each student achieves in relation to the other students in the class. The approach requires that you:

• set very clear, measurable goals/objectives against which to evaluate student performance;
• determine how best to measure individual achievement for each of the goals.

Norm-Referenced Grading

The norm-referenced system of grading determines students achievement of course goals in relation to other students in the course. An individual score is compared to that of a "normative" group. In a normal (bell-shaped) curve, the smallest groups occur on either end of the distribution, and are awarded A's and F's. The largest group is the middle group, and those persons are assigned C's, while D's and B's are assigned to the next-largest groups. If you concentrate solely on a normative distribution and grade on a curve, you are predetermining that a set percentage of students will receive a set grade, irrespective of the quality of their work (for example, deciding that one-third of the class will get C's on the midterm, even if they correctly answer over 90 percent of the questions).

Students think of "grading on a curve" in a variety of ways. Students will often say to you: "This grade doesn't seem fair, can't you grade on the curve?" A good strategy for answering this is to ask students what they mean by "curve." Typically, they don't mean a normal curve, since that would destine a preset percentage of them to D's and F's. Rather, they are saying that you seem to have given too few A's and too many C's, and you should award more B's and A's to "even out" the grades. In other words, they perceive UW grades as symmetrically distributed around a mean of 3.0. If this is indeed what the students are implying, explain that they really don't want you to curve the class but simply to raise grades.

In most cases, grading at the UW seems to move between some combination of criterion- and norm-referenced. Thus, if you assign grades based on how students perform relative to other students and also how well they achieve course goals, then you might look for natural breaks in your class's distribution as an easy way to make at least preliminary distinctions. If you are evaluating essays or research papers, one very helpful way to proceed is to rank the papers before assigning any grades. Natural clusters often occur in such a process, and you also can get a better feel for your specific criteria by saying "This one is better than this one because..."

If your grades are skewed at the high or low end, or are not in line with your colleagues, meet with your colleagues to discuss what your questions or assignments are designed to evaluate. If you are using a criteria-based grading system, see if a consensus can be reached among your colleagues on what constitutes effective student work. Comparing grades on a set of essays can be one of the most productive ways of arriving at a common language and standard for assessment, thus preventing students from shopping around for sections.

Forming and Discussing Criteria

Grading is easier -- and less likely to be contested -- if you have been making the evaluation criteria for individual assignments clear from the beginning. Handing out your standards at the beginning of the quarter will not necessarily be sufficient to clarify things for the students. Be prepared to repeat the information several times for reinforcement -- out loud, on the board; in handouts, and, most importantly, integrated into each discussion of assignments and results. Ideally, your grading criteria should be implicit in everything you say in class; the ways you define and analyze problems and present evidence should model the very processes you want to see in student work.

Additional ideas for discussing grades with students:

• If you are grading on the percentage of points a student earns ("95 percent and over is an A," etc.) then work out a system for translating those percentages into the decimal system. But when communicating this system to the students, indicate some broad guidelines about percentages in terms of the entire course, rather than on every exam or graded piece of work. Especially on early tests, consider leaving the raw score as a percentage only, rather than assigning it a decimal or letter grade. This can avoid repeated queries as to whether an 87 is a 3.3 or 3.4, or B or a B plus, etc.
• Be consistent and equitable.
• Make sure students know what types of questions will be asked, what types of evidence they will be expected to present, or what procedures they will be expected to follow. Whenever possible, hand out sample questions ahead of time.
• Make sure students understand why they are being tested on certain material -- what is being measured, how it is being measured, and what the test has to do with course objectives. Are students being asked to recall information, recognize patterns or analogies, draw inferences, make connections, originate a thesis, or what?
• When students ask to have a grade changed, or contest an answer, don't act hastily. Avoid spot judgments by scheduling a meeting for a few days later. Research the issue, prepare a response and a rationale, and, if necessary, talk with the course supervisor or TA coordinator about it. If students are not satisfied with your response, refer them to the course supervisor or TA coordinator for a determination. In those rare instances where you encounter a dissatisfied student, document your interactions as promptly as possible so you have accurate notes for later discussions.

Resources

The most complete resource on grading at the University of Washington is available at http://depts.washington.edu/grading/. CIDR consultants are also available to assist with issues related to establishing and carrying out a grading system.

For assistance or additional information on test construction, consult CIDR at 543-6588 or send a message to cidr@u.washington.edu.