On 7/19/2011 3:22 AM, M. Horton wrote:
The idea behind this is that it’s not fair to give an “F” student who didn’t learn anything up to 50 points in the grade book. But the student who didn’t do the assignment at all (and also didn’t learn anything) gets 50 points less than that…a zero. The logic is that on a 4, 3, 2, 1, 0 scale, students get the same grade for failing the assignment and not doing the assignment and the penalty for not doing an assignment isn’t career-ending with a giant cliff between F and zero. Doug Reeves talks about this quite a bit as well as the weaknesses of using average scores.
Let’s pretend for a moment that we have a set of assessments that are a good measure of the student’s level of knowledge and understanding. Let’s also suppose that the score that indicates the minimum acceptable level of proficiency–a D on the traditional A/B/C/D/F
scale–corresponds with a score of 60% on this set of assessments. A score of 59% and a score of 3% are both failing grades on one of these assessments, but they say very different things about a student’s level of mastery of the topic. I think it’s important to preserve this
distinction. A student who scores 59% is most of the way to a passing grade. A student who scores 30% is only half-way there, and a student who scores 3% has learned practically nothing.
However, I think the question that some people are trying to ask but haven’t quite done so is, “How much should performance on one topic affect the grade for the entire course?” For example, most high schools grade on a quarterly system. Suppose that during one quarter, I teach the topics of solutions, chemical reactions, moles, and stoichiometry. Suppose a student passes the solutions, chemical reactions, and moles tests with 70% on each, but fails the stoichiometry test abysmally with a 10%. This student was found to be proficient in three out of the four topics assessed. The student has an average of 55% (failing) on the
percentage system, but an average of 1.5 (passing) on the 4/3/2/1/0 scale (assuming 4 = 90-100%, 3 = 80-89%, 2 = 70-79%, 1 = 60-69% and 0 = less than 60%.)
Suppose another student scores 65% each on the solutions, chemical reactions, and moles tests, and 55% on the stoichiometry test. This student was also found to be (minimally) proficient in three out of the four topics assessed. However, this student has an overall average of 63% (passing) on the percentage system, but an average of 0.75 (failing) on the 4/3/2/1/0 system.
Suppose a third student scores 95% on the solutions test, but fails the chemical reactions, moles, and stoichiometry tests with scores of 55% on each. This student has failed three of the four topics, but has an average of 65% (passing) on the percentage system, and 1.0 (passing) on the 4/3/2/1/0 scale.
In both cases, two of the fictitious students passed three topics out of four and failed the course under either the percentage or the 4/3/2/1/0 system. In both cases, a third student passed only one topic and failed three, but passed the course under both systems.
Every grading system is arbitrary. Every grading system, especially when applied rigidly, results in some students failing who should have passed, and other students passing who should have failed. The “I look at your performance and subjectively assign you a grade” system is probably a much truer measure of student performance than any of the statistical systems described above, but it requires a lot more trust in the teacher than most students, parents, and administrators are willing to bestow.
Originally posted to the Chemistry Education Discussion Group.
