Percentages in assessment

Ross Woods

"What percentage constitutes a passing grade?" For example, some people might say 60%, others might say 70% is higher, and others suggest that 80% represents "higher quality" education.

The fact is that the actual percentage figure is completely arbitrary. An assessor could easily dumb down all questions in order to require 100% correct. Similarly, an assessor could make them all so difficult that 20% correct is a good result.

In norm-based assessment, what really matters is the percentile needed to be deemed a passing grade. In norm-referenced assessment, assessors use a scaling adjustment so that, for example, the 25% of students with the lowest percentages are failed, the middle 50% do satisfactorily, and the top 25% get distinctions.

In a competency-based system, however, it is the other way around. First, write competencies to define a passing level of performance. Then set the testing system so that a pass fits the competency. In competency-based assessment, all knowledge is required so we can legitimately ask: "If it doesn't matter whether students get a particular question wrong, then nobody needs to get it right to pass. So why are you asking it?" In other words, it is quite legitimate to ask for 100% in these cases.

Some systems have it both ways; they use a rubric for assessors to classify answers as "unsatisfactory," "satisfactory," and one or more levels of exceptional. Other have several stages:

  1. Early questions only settle students into the testing environment.
  2. The next batch of questions tests the basics, and represent the competency standard. Students must get them all correct to pass.
  3. The following batch of questions are more difficult; their purpose is to enable students to get a grade above the minimum pass.