Final Exam: Proficiency Needed for Guaranteed Pass

Question

Is it common to guarantee students a passing course grade if they demonstrate proficiency in some subset of topics/skills on a final exam?

Note that this is in contrast to setting an upper bound on a final exam score that would guarantee a failing grade for a student, as in this question.

If this is a process you have experience with, it would be helpful to hear about the following:

1. Is it common for students to disregard all other assignments, choosing to gamble the whole course on just the final exam?
2. What and when do you let students know about this required set of skills?
3. How were these qualifying skills chosen, and who wrote the final exam?
4. What set of conditions is required to earn a "B" or an "A" in the course?
5. How did students do in the following course?
6. What other issues (good or bad) come with this model?

And, of course, do you know of research on such a grading scheme?

In particular, I am imagining an undergraduate mathematics course in the United States, where a cumulative final exam is given at the end of a semester, typically figuring heavily into the overall course grade. Exams may be written by individual instructors for their courses, or created by committee and given to all sections of the same course. Here, "passing" refers to the minimum course grade required to enroll in the subsequent course (e.g. from Algebra 1 to Algebra 2).

Answer 1

In my undergraduate university, we had a similar system in the mathematics department, only instead of one final exam, we had our test grades calculated based on our top X - 1 tests, where X is the number of tests.

For example, if you had four exams, only the three highest scores were counted and the lowest score was dropped. 3 100s and a 0? You get an A. 0 on the first test, and 90s on the rest? Still an A.

In my case, I had gotten three solid As in calc. On the last day of classes before finals week, the professor checked my test grades, marked a zero for my final exam, and congratulated me on my A. He had no expectation that I would take the test because I had demonstrated proficiency three times already and he would probably not have found anything different about how well I knew calculus.

I'm sure some people will argue whether or not only actually demonstrating knowledge 75% of the course material is representative of 100% proficiency, but people have to put in at effort to reap the reward of no final. People who are driven to do well will do so regardless of any shortcuts you provide them.

The system you've described where your grade is only determined based on your performance on the final sounds like an even better deal, but I do not believe would be as effective as an X - 1 drop system, and would even harm student performance.

• Students do not get accurate gauging of their own performance through their semester.
• Students do not participate in class (you could end up teaching an empty room if nothing matters).
• Students can potentially fail before realizing they have a problem.
• If you're doing a half-semester Withdrawal setup, there's no grade reflection at the half-way point for them to consider dropping.

The only people such a system would benefit are those that have basically mastered the subject and need to retake it for some reason. Otherwise, I wholly think that it's better to have the student work their way through a course, rather than allow them to treat the final of the course as the only performance checkpoint.

Answer 2

I will add a student's perspective, realizing consequences down the road: I took a python (programming) class that treated the grade like this, though for us the entire grade was based upon the final project, not a single test. I was one of the students who skipped all of the assignments until the very end, because by the time I took the class I was already relatively accomplished in 3 other languages, and figured I could breeze through the final when I wanted to.

For me, statistically, it was a good gamble. I produced a perfect (satisfied all criteria) final project and got myself an A with almost no effort.

Other students in the course did this also, but with only mixed results. Many had a super hard time on the final project because they hadn't been following along, and they didn't understand the syntax well enough when it came time to produce something more complex.

The professor said that the reason for this method came down to the fact that measuring programming ability can only be done by determining whether a person can, with a set of guidelines, produce a program or not. If you satisfied all of the criteria of the project, the A was yours. If you were unable to do so, then the professor would look at any past assignments to determine whether you really knew your stuff or not.

The professor was also lazy. He wouldn't grade assignments unless you asked him to grade yours. He said that it saved him a ton of time by the end of the semester, because a lot of his students just ended up turning in perfect final projects, and then he wouldn't have to bother to grade their other assignments throughout the semester.

There wasn't a followup course for this, it was a standalone, so I can't really say how people did as far as doing well in the next course. What I can say though, is that I remember almost nothing from the course, and had to completely relearn python many years later from scratch. A lot of that is because the language changed pretty drastically in the 8 or so years between when I learned it and when I used it professionally, and more because I never encountered a situation when I said, "Man, I just really want to use python for this." so my skills started going stale almost immediately.

I knew, as a student, that I was taking a risk approaching the course the way I did. But I also knew that I had a pile of other coursework to do, and it benefited my other grades to focus less on a class that was primarily a repeat with different syntax. As a computer science student who transferred from my community college to a 4 year, I ended up with plenty of "I know this language, and already took this class, but don't have credit because my university won't give me 300 level credit for a CC course." Every time I had one of those, I wished that they would have graded like my Python professor.

Answer 3

I suppose that something like this happens but it isn't common. Thankfully. In fact, providing a simple path to a grade that doesn't involve actual mastery of the subject but just answering some subset of questions about it is likely to be counterproductive with at least a subset of students.

Students are busy and will want to take advantage of any short cuts offered.

I doubt that there is specific research that speaks to this, but I also think that the great bulk of research on learning would speak against the practice. For example, see The Art of Changing the Brain, by Zull.

At the end of the day would you as a professor prefer to take on a student who got a lot of good grades somehow or one who really learned a subject and could apply in in creative ways?

I do know of something like the opposite practice, actually. In fact I once benefited from it. I was taking an undergraduate physics course and crammed for the first exam, staying up all night. I failed it. The next exam I got a D, the next I got a C, then a B. I got an A on the final exam. The professor was viewed a something of a hard case, but he gave me an A for the course. I learned how to learn and I improved. But there was noting in the syllabus that suggested that I would be graded that way. And I doubt that any strategy other than steady improvement would have resulted in that outcome.

Similarly, I occasionally let students avoid the final exam altogether provided that they had proven throughout the course that the A grade was appropriate for them. Some other professors would grade the final so that you could only improve your grade there, which is a bit less radical.

If you make grading into a game, a fraction, possibly large, will treat it as a game to their detriment. Stress long term learning via reinforcement and feedback, not high risk exams. Yes, they will avoid other work and think they can "get away with it." Don't encourage such behavior.

However, there is a method of grading that might be a bit similar in structure to what you suggest: cumulative grading.

In cumulative grading every graded student activity has a number of points assigned and the points of all such activities add up to some number, say 1000. On each activity you earn a number of points toward the total. Publish in advance what number of points are required for each grade, say, 900 for an A. Then the student always knows where they are in the course and what grade they have earned. They also know what effort is required in the remaining part of the course to achieve the grade they feel they need.

Students can game this system, of course, working until they have earned, say, a B, and then putting in no further effort.

If you combine this with permission to repeat old work for some additional points then you will encourage people to get the reinforcement they may need. You also don't get complaints about your grading.