University of Minnesota
Office of Human Resources
http://www.umn.edu/ohr
612-625-2016

Grading Systems

Jump to:


The two most common types of grading systems used at the University of Minnesota are norm-referenced and criterion-referenced. Many professors combine elements of each of these systems for determining student grades by using a system of anchoring or by presetting grading criteria which are later adjusted based on actual student performance.

Norm-Referenced Systems

Definition

In norm-referenced systems, students are evaluated in relationship to one another (e.g., the top 10% of students receive an A, the next 30% a B). This grading system rests on the assumption that the level of student performance will not vary much from class to class. In this system, the instructor usually determines the percentage of students assigned each grade, although it may be determined (or at least influenced) by departmental policy.

Advantages

Norm-referenced systems are very easy for instructors to use. They work well in situations requiring rigid differentiation among students, where, for example, due to program size restrictions, only a certain percentage of the students can advance to higher level courses. They are generally appropriate in large courses which do not encourage cooperation among students.

Disadvantages

One objection to norm-referenced systems is that an individual's grade is determined not only by his/her achievements, but also by the achievements of others. In a large, non-selective lecture class, you can be fairly confident that the class is representative of the student population; however,in small classes (under 40) the group may not be a representative sample. One student may get an A in a low-achieving section while a fellow student with the same score in a higher-achieving section recieves a B.

A second objection to norm-referenced grading is that it promotes competition rather than cooperation. When students are pitted against each other for the few As to be given out, they're less likely to be helpful to each other.

Possible Modification

When using a norm-referenced system in a small class, you need to modify the allocation of grades based on the caliber of students in the class. One method of modifying a norm-referenced system is anchoring. Jacobs and Chase in Developing and Using Tests Effectively: A Guide for Faculty (1992), describe the following ways to use an anchor:

"If instructors have taught a class several times and have used the same or an equivalent exam, then the distribution of test scores accumulated over many classes can serve as the anchor. The present class is compared with this cumulative distribution to judge the ability level of the group and the appropriate allocation of grades. Anchoring also works well in multi-section courses where the same text, same syllabus, and same examinations are used. The common examination can be used to reveal whether and how the class groups differ in achievement, and the grade in the individual sections can be adjusted accordingly.... If an instructor is teaching a class for the first time and has no other scores for comparison, a relevant and well-constructed teacher-made pretest may be used as an anchor."

Modifying the norm-referenced system by anchoring also helps mitigate feelings of competition among students since they may feel they are not directly in competition with each other.

Before deciding on a norm-referenced system, consider:

  • What is your expected class size? If your class size is smaller than 40, do not use a norm-referenced system unless you use anchoring to modify it.
  • Is it important for students to work cooperatively in this class (e.g., form study groups or work on team projects)? If the answer is yes, a norm-referenced system is not appropriate for your class.

Criterion-Referenced Systems

Definition

Norm-referenced tests measure students relative to each other. Criterion-referenced tests measure how well individual students do relative to pre-determined performance levels. Teachers use criterion-referenced tests when they want to determine how well each student has learned specific knowledge or skills.

In criterion-referenced systems, students are evaluated against an absolute scale, normally a set number of points or a percentage of the total (e.g., 95-100 = A, 88-94 = B). Since the standard in this grading system is absolute, it is possible that all students could get As or all students could get Ds.

Advantages

Students are not competing with each other and are thus more likely to actively help each other learn. A student's grade is not influenced by the caliber of the class.

Disadvantages

It is difficult to set a reasonable standard for students without a fair amount of teaching experience. Most experienced faculty set criteria based on their knowledge of how students usually perform; thus, criterion-referenced systems often become fairly similar to norm-referenced systems.

Possible modifications

Instructors sometimes choose to maintain some flexibility in their grading system by telling the class in advance that the threshold for grades may be lowered if it seems appropriate. Thus, if a first exam was more difficult for students than the instructor imagined, she/he can lower the grading criteria rather than trying to compensate for the difficulty of the first exam with an easier second exam. Raising the criteria because too many students achieved As, however, is never advisable.

Another way of doing criterion-referenced grading is by listing class objectives and assigning grades based on the extent to which the student achieved them. For example, A = Student has achieved all major and minor objectives of the course; B = Student has achieved all major objectives and several minor objectives; etc..

Before deciding on a criterion-referenced system, consider:

  • How will you determine reasonable criteria for students? If you are teaching the class for the first time, it is advisable to maintain some flexibility.

Other Systems

Some alternate systems of grading include contract grading, peer grading, and self-evaluation by students.

In contract grading, instructors list activities students can participate in or objectives they can achieve, usually attaching a specified number of points for each activity (e.g., book report = 30 points, term paper = 60 points). Students select the activities and/or objectives which will give them the grade they want and a contract is signed. It is advisable to have qualitative criteria stated in the contract in addition to listing the activities.

In some classes, a portion of a student's grade is determined by peers' evaluation of his/her performance. If students are told what to look for and how to grade, they generally can do a good job. The agreement between peer and instructor rating is about 80%. Peer grading is often used in composition classes and speech classes. If used, it should always be done anonymously.

Students can also be asked to assess their own work in the class and their assessment can be a portion of the final grade. This method has educational value since learning to assess one's own progress contributes to the university's goal of preparing students to be lifelong learners.

A research analysis found that the percentages of self-assessors whose grades agree with those of faculty graders vary from 33% to 99%. Experienced students tend to rate themselves similarly to faculty while less experienced students generally give themselves higher grades than a faculty grader. Students in science classes also produced self-assessments which closely matched faculty assessment. If self-assessment is used, the instructor and student should meet to discuss the student's achievement before the self-evaluation is made.

Characteristics of a Good Grading System

1. Grades should be relevant to major course objectives.

Although this may seem obvious, students often complain that there is no connection between the stated course objectives and the way they are evaluated. For example, one frequent lament goes something like this: "Professor X said the most important thing he wanted us to get out of this class is to be able to think critically about the material, but our entire grade was based on two multiple choice exams which tested our memory of names, dates, and definitions!"

When preparing your grading system for a course, begin with a list of your objectives for the course. Assign relative weights to the objectives in terms of their importance. Be sure the items you are including as part of the grade (e.g. exams, papers, projects) reflect the objectives and are weighted to reflect the importance of the objectives they are measuring.

2. Grades should have recognized meaning among potential users.
Because the purpose of grades is to communicate the extent to which students have learned the course materials, grades should be based primarily on the students' performance on exams, quizzes, papers, and other measures of learning specified at the beginning of the course. Items such as effort, attendance, or frequency of participation, although contributing factors to student learning, do not actually reflect the extent to which students have learned the course materials.
3. The grading process should be impartial and compare each student to the same criteria.
If you are willing to offer extra credit or opportunities to retake exams or rewrite assignments, the offer should be made to the whole class rather than only to individuals who request these opportunities.
4. Grades should be based on sufficient data to permit you to make valid evaluations of student achievement.
It is rarely justifiable to base students' grades solely on their performance on one or two exams. Unless the exams are extremely comprehensive, one or two exams would provide an inadequate sampling of course content and objectives. There is also a likelihood that an off-day could lower a student's grade considerably and be an inaccurate reflection of how much he/she has learned. Generally speaking, the greater the number and variety of items used to determine grades, the more valid and reliable the grades will be.
5. The basis for the grading should be statistically sound.
If you say that an exam is worth 15% of the total grade, use a procedure for combining scores that ensures that this will be the case. The University of Minnesota Office of Measurement Services can help you find an appropriate procedure.