Book (2022)/Absolute grading

Absolute grading (criterion-referenced) measures the student's success in relation to the criteria required to achieve that grade level. Usually, the criterion is the number of points or a percentage of the total number of points that the student must achieve in order to receive the appropriate grade. The simplest way to grade is to determine what percentage of the total number of points is needed to achieve a certain grade. For example, an A grade requires 90% and up, a B grade 80-90%, and so on. The problem with this approach lies in the arbitrary setting of the boundaries between individual grades. If we set similar point boundaries in advance, the author of the test must “hit” within them. If the test or one of its versions is more difficult or, conversely, easier than the creator of the grading scale assumed, the resulting assessment will also be perceived as disproportionately strict or, on the contrary, benevolent. Therefore, for more important tests, it is advisable to set the borderlines using the estimates of a larger number of experts, for example according to the Angoff or Ebel method.

In absolute grading, unlike relative grading, a student's grading is not influenced by the performance of others and is not based on mutual comparison with students in a group. If we were to test a group of significantly above-average students, they might all get good grades, and conversely, if a group of weak students happened to come together, no one might get good grades. Students are not competing with each other and are therefore more likely to work together. This can also be beneficial for their active involvement in learning, which is often based on cooperation. The grading of an individual student is not affected by the overall result of the class.

Absolute and relative grading are actually somewhat intertwined. Most teachers set criteria based on their experience with typical student performance. This brings relative elements into the absolute grading. Similarly, teachers sometimes retain some flexibility in absolute grading by telling students in advance that the criteria on the first run of the test may be relaxed if too few students score well. For example, the 90% threshold for obtaining an A grade may be reduced to 85%. If the test would be more difficult for the students than the teacher imagined, he can reduce the assessment criteria in this way. The opposite procedure, where the teacher would tighten the criteria because too many students achieved a good grade, is not recommended.

Another way to grade students according to criteria is to set course objectives and assign grades based on how well the student has achieved them (e.g. A = student has achieved all major and minor course objectives, B = student has achieved all major and several minor objectives, etc. .).

A more sophisticated form of absolute classification distinguishes between different types or levels of knowledge and skills that a student demonstrates on different tasks. Greater emphasis is placed on those that reflect higher levels of mastery of the material. This approach reflects both the amount of material and its level of cognitive complexity. For example, we can divide the learning objectives of our course into two groups: basic and advanced. The basic objectives refer to the minimum necessary knowledge and skills that students must acquire. Advanced objectives, on the other hand, represent higher levels of skills, such as using critical thinking, solving complex problems, and the like.

It may be easier, at least initially, to use two completely separate tests to determine how well the basic and advanced learning objectives have been achieved. This will make it easier to evaluate the exam and keep records of it. By separating the tests, it is also easier to focus on the individual learning objectives and prepare test questions for them. It tends to be relatively easy to assess the basic learning objectives. Assessing the extent to which the advanced learning objectives have been achieved is usually more difficult, as it is more difficult to devise test items covering the ability to apply the acquired knowledge.

Different requirements for student performance can be set for passing both types of test, as indicated in Table 5.7.2.

Table 5.7.2 Example of a possible setting of absolute standardization for grading a basic and an advanced test on a five-point grading scale.

Grade level	Basic test	Advanced test
A	90 % or above	85 % or above
B	90 % or above	75–84 %
C	80 % or above	60–74 %
D	80 % or above	50–59 %
F	less than 80 %	less than 50 %

In the given example, we require students to demonstrate mastery of at least 80% of the basic learning objectives and 50% of the advanced objectives. If we require the setting of success thresholds to be more objective, we can use one of the expert estimation methods described above.

From a higher education perspective, criterion assessment is the most desirable. Although it is more demanding for teachers, it requires thinking about the expected learning outcomes, but it is transparent for students and the derived grades should be defensible from a reasonably objective point of view – students should be able to trace their grades according to specific performances in solving set items. Criterion evaluation, with its transparency, creates an important framework for the involvement of students in the learning process.

With absolute evaluation, it is also appropriate at the same time to monitor the distribution of grades in the study group – in other words, to monitor the results of the criterion grading model from the perspective of the relative evaluation model. If we find that too many students are getting poor grade, or, on the contrary, good grades, or the distribution is skewed in some way, then this may indicate that something is wrong and that the grading process needs to be reviewed. For example, this may be a problem with the overall difficulty of the assessment items (for example, unchallenging exam questions or too few questions, or items that do not differentiate between students of different knowledge and skill levels). Best practices for grading in higher education are based on predominantly criterion-referenced assessment, slightly modified by relative correction and feedback.