Classes where average grade % is failing - is this common?

Parents Forum
161

You misrepresent my position. My experience in STEM, and I think it continues in the STEM workd today, is that the problem solving aspect is not well taught in the classroom. The regurgutatable facts and methods are taught, but students get little training and practice in putting them together – yet that is the primary expectation on exams, and that is then used to weed out the students. Also, I believe a lot of the students who do well in the weeder classes are repeating the material – either to confirm their basis for moving forward or for an easy A. Then the profs congratulate themselves into thinking their teaching is fine because some students got it, and don’t ever examine whether their teaching or exam format leaves something to be desired.

I also did not say they should not solve completely new or unseen problems - but I do think those problems should make up a smaller percentage of exams than they often do. There is no earthly reason to give students an exam where most of the class can’t do more than half the problems. They don’t learn anything from that except discouragement – you don’t learn during or from your exams, at least I didn’t – in a well taught class, you learn prior to the exam, then show what you can do. If students can’t do over 50% of what is being tested, then the right skills are not being emphasized or the exam is poorly written.

I know this is just impossible for some profs to wrap their heads around, and most who have always done it this way don’t want to. But this is a really broken piece of the STEM world.

162

I also expect kids to not just learned but also internalize the concepts so they can use them to different situations heretofore unseened. Starting in high school and especially in college. And even more so for engineering students. That’s what they will be doing when they graduate. The problems they meet in their jobs will be new and hopefully they will be able to look into their bag of tools and knowledge and know which works best - it may not be the traditional solution.

Being about to spit back what the professor has taught or even given as homework isn’t very challenging. To me, that is C work, and I got my share of Cs in college. I can’t imagine a professor (or a team of them) composing an exam where most of the problems are unchallenging - especially in competitive, highly selective universities where the brightest attend. I suspect the students wouldn’t find much satisfaction in that either. I know when I managed to defeat a difficult math problem (and it felt like war solving it), I always felt like I just accomplished something great. Homework was also challenging in upper-level classes. Often, there was a problem or two that stretched our minds but with study groups and a few hours, we could get through them. Now, I’ll admit to not being a great math student - I never considered taking the Putnam - but I was adequate.

As for Putnam, the challenge is to score anything, not answer all the questions correctly. It’s to score anything. One sibling took it and scored maybe 1, 2 and was elated to get a non zero score. Test takers know this.

163

@intparent - Respectfully, there seems to be a lot of gender-based baggage that is coloring your view of this problem. I fully admit that there are major issues for women in many STEM fields that I have no idea how to solve. But I know that mistraining students isn't the answer.

Again, this isn’t about punishing students. If I recall correctly, one of the last classes I taught was an advanced senior course with 10 students. The exam averages may have been around a 50%, but as best I can remember I gave something like 8 A+/A/A- grades, 1 B+ grade, and 1 B grade. I also wrote up 100 pages of lecture notes because I thought the 2nd half of the textbook was unclear. I hardly think I was some sort of “Warrior” who was uninterested in my students’ success or who delighted in weeding them out.

There is another possibility buried in the excluded middle. It could be that you don’t understand the issues. Many people who have actual university teaching experience, both men and women, are telling you what the goals of a STEM class are and how they think it’s best assessed. It is possible that you are right and thousands of “old white male” professors (of all genders and races, apparently) are wrong. But I don’t think so.

Now this critique has a lot of merit.

First, we spend the first 18 years getting students up to the basics of calculus, then in the next 4 we try to take them well along the way to the research frontier. There simply isn’t any time to help students who can’t keep up if we want to graduate them in four years.

Second, in lecture we have to spend the bulk of the time simply teaching the material. There isn’t too much time left for teaching students how to apply it. Some of that’s done in section or at office hours, but for the most part the average student is just trying to do their problem set to turn in. I’m sure we could do better here. Maybe online education can help.

I don’t think there is a silver bullet. One problem is that freshman students often haven’t learned any real problem solving skills in high school. The high school setting is where it needs to start because they can practice on material that is easier and where their intuitions are stronger. Unfortunately, a lot of students get to college not even realizing that there is a deeper level of understanding that they need to get to, and get to now. Then they’re stuck developing problem solving skills while simultaneously struggling through their first ever encounter with hard material.

Rather than change college testing to be more like high school, we need to do the exact opposite of what you propose. We need to change high school testing to be more like college.

This is how it’s done in many other countries that seem to do a better job than we do teaching STEM subjects at the high school level (though the US university system is clearly the world’s best).

Regardless, there is no substitute for a student engaging with material. Deeper understanding is a journey. I can show them the path, light the path, help them if they stumble, but they need to walk the path themselves.

164

If the students can understand the steps well enough to complete the problems correctly on the exam, they aren’t leaving the course with zero knowledge. They have a foundation they can build upon if someone would take the time and trouble to teach them. If they know what you taught them, how are they faking their way through?

I don’t understand why students aren’t graded against the material instead of each other. You either know the material or you don’t. If several students end up getting A’s, then several students earned A’s. How does creating an artificial stratification so you can say one A is better than another help the students? You may be weeding out creative thinkers who approach problems differently than you do. You don’t even know if you’re comparing apples to apples. How do you know whether or not the students who do well on the exams have been exposed to the material before? You’re assuming they have a natural ability, but for all you know they may have studied the material before entering your class. Apparently, many high school students do that. Why wouldn’t they continue to do so in college?

I also don’t get why professors aren’t comparing what the students know to what they’re supposed to know when they leave the course. Testing them on things you haven’t taught them because you don’t have time to teach them everything and you want to weed out all but one type of learner is lazy and a waste of human resources. These exams don’t reflect the way the real world works. If you’re presented with a type of problem you haven’t seen before, you research and figure out a solution.

How do you know that a test that you designed is good enough to determine who is worthy enough to be an engineer? I can see that it may be helpful in classifying different types of learners, but you’re not scientists. How can you determine that what you think you’re proving is what you’re actually proving?

165

When my daughter was in hs, she was enrolled in Physics B. For her, it was really easy (especially considering she had completed the required honors physics prior to that). She decided to self-study Physics C (not offered at our school). And she had a lot more trouble with it than I would have expected from a student who was coasting through Physics B with no effort and who had already mastered more than the required calculus. She said that the problems in Physics B were just plug and chug repetition of the examples and the problems in Physics C required actual thought and real problem-solving.

College problems sets should be where the kids learn the problem solving skills. The kids are often allowed and encouraged to work in study groups, to attend office hours and other organized supports. But students have to be careful with this. I think it’s possible for students to kind of coast along with these supports and seek help too readily rather than working harder at developing those problem solving skills. There is a big difference between nodding along as someone else shows how to solve something and actually coming up with it yourself. I think there is also a difference in students’ level of hs preparation for this kind of hard problem solving, both in the classroom and in the EC’s they have done. Math competitions are a good opportunity to practice these kinds of problem-solving skills.

166

If all they did was memorize a homework problem, they are not getting any foundation. I could memorize and regurgitate a sentence in Greek and have no idea what it means or how to produce any other sentences.

I’m not making any assumptions about whether their ability is “natural” or the result of intense effort or of having seen the material before. And frankly, that is irrelevant. All I care about is whether they can do what we have determined they should be able to do in order to succeed in the next level. And creative thinkers who do it wrong are still doing it wrong. If they creatively (and validly) get to the right answer, I’m smart enough to see that.

Now we’re back to giving them the exact homework problems which they could memorize and get along without learning anything.

Are you doing away with the exams altogether then? I’m not designing tests to determine who is worthy. I’m just trying to see if they have a clue.

167

Is it possible that the determination of what is needed at the next level is skewed – and there really isn’t enough time to teach what should be taught in a given class? I look at a subject like Bio – our knowledge has exploded in the past 25 years – but the intro classes still have the same title. There is a reason, I think, that most students I know hated AP Bio – way, way too much material for one class.

No prof wants to give up a scrap of their precious material – I know when my kid’s college made an effort to reduce the core requirements (fewer semesters of some subjects for all students, or quarter long classes instead of semesters that were supposed to have less material), many of the profs just stuffed the same material in the compressed schedule because they couldn’t stand to see their area cut back. Hence core is now probably more stressful than before, when the changes were intended to have the opposite impact.

You don’t have to give them the same problem. But conceptually the same permutation with different numbers/elements/etc is fine – if they knew why it worked that way in the homework, they should be able to do the same on the exam. But I don’t think that is what is happening most of the time – I think profs often do mashups that really aren’t remotely similar to the homework or what was covered in class, then congratulate themselves when almost all the students can’t do it.

168

“You don’t have to give them the same problem. But conceptually the same permutation with different numbers/elements/etc is fine – if they knew why it worked that way in the homework, they should be able to do the same on the exam.”

Wrong. That is not engineering.

Engineering is about solving problems that you know how to solve - it is about figuring out how to attack problems you haven’t seen before with techniques that you have seen. Engineering tests are the same way. The problems will you the physical laws you’ve seen before, but applied to a different situation.

Being able to solve a problem you’ve seen before should be trivial. It is the bare minimum passing standard for most engineering classes.

169

@intparent, I do believe there are very significant differences in preparation that students are getting at the hs level. If I remember your posting correctly, your daughter came out of a hs where few APs were offered. She jumped into physics, arguably one of the hardest college majors–at one of the toughest, most selective programs in the world. And there has been some struggle. This doesn’t surprise me. She probably has many classmates who were far better prepared in hs and I am sure even they are challenged. I know my daughter feels that many, if not the majority, of her classmates were better prepared for her difficult physics class. There were students from super-elite hs’s who literally spent half the time she did on the problem sets, yet did better on the exams. I’m guessing they didn’t learn Physics C from a test prep book.

170

From what I have heard, this is pretty common for engineering at the Ivy League schools and other comparable schools, especially at Cornell. While these courses are very hard, I do not understand how having a test average of 30-45% helps anyone.

171

It’s the academic version of fraternity hazing.

172

How does it hurt them if it is still passing?

173

@mathyone, these comments are also based on my own experience with STEM weeder classes at a large research university in my freshman year of college. I was discouraged by the very issues we are discussing today, and dropped out of the STEM major I was interested in. In retrospect, I think I could have been pretty kicka** in my area of interest – but was “weeded out”. One of many, I think.

It is one reason my D has been able to stick it out, IMHO – I see this system for what it is, and help her keep her chin up. She is quite good at some areas of her major, has a harder time in others, but has passed all classes and has over a 3.0. Some of her peers aren’t doing as well – she seems to be academically at about the midpoint of her class. But she has that female tendency to assume it is just her being a failure when she hits one of these profs who tests material and methods they aren’t actually teaching. Interestingly, her advisor has warned her about a couple other profs in her major, telling her that they don’t teach what you need to know to pass their exams, and he has offered to give her some extra tutoring in a couple of the classes to supplement because he doesn’t think the profs teach the class fairly and fully.

It isn’t like every STEM prof really sees teaching as their highest calling, too. Of course some do, but I think everyone knows some profs who care more about research than teaching, and really only want to interact with the star students.

174

Might as well summarize my thoughts about what is and isn’t reasonable on an exam.

What is fair to put on an exam:

  1. A modified 100-point scale that assigns letter grades according to a system that differs from 90=A, 80=B, etc. Depth of understanding may need to be measured in a way that does not fit well within that system, especially for technically deep classes where a small increase in ability may make you a lot more productive (e.g. for quantum mechanics, your ability to solve problems will increase greatly if you understand the mathematical side of it, so an A-level understanding might get you many points more than a B-level understanding of the material).
  2. Questions on the exam that test concepts pretty much the same as they were taught in class, with small differences. This is rudimentary understanding that shows that you are solid enough to move on (a “C” level of understanding).
  3. More in-depth questions that require a deeper understanding of how the principles learned within the class can be applied to somewhat similar problems. These won’t be obvious but they should be well within what would come to mind for someone who really has a solid grasp of the material. This is what generally qualifies as “A” level understanding, or “B” level understanding if they seem to see how the principles could be applied but make some amateur errors along the way.
  4. Making the modified scale a bit more lenient if it is decided that in retrospect, the exam was harder than it should have been. I’ve done this before when I’ve written exams and frankly, you can’t always know ahead of time that your exam won’t be more than you bargained for. This shouldn’t happen frequently but it’s fair if it happens once in a while.

What isn’t fair to put on an exam:

  1. Any form of normal distribution ranking that, rather than measuring who passes a predetermined bar of understanding for the given grade levels, just decides that X% of the class deserves a Y grade. This tells you nothing and it’s a stupid ego trip.
  2. Problems that involve more depth of understanding than students are taught. Obvious but I’ve seen it happen.
  3. Problems that can be solved using the tools learned in class but that require a depth of understanding that is generally gained by a graduate level study of the material. This isn’t the kind of thing that comes from intuition and brilliance, it comes from having seen similar problems before and developing understanding over time. One of my relatives who was a physics professor at a top school told me about how their department sometimes came together so that the professors would solve the problems that they were planning to give to undergraduates because they weren’t confident that any one of them could do the problem without making any mistakes. That’s not “setting the bar high,” that’s ego.
  4. Problems that involve stupid or irrelevant “clever” insights that are generally meaningless. The worst offender I’ve ever seen of this is an engineering exam that asked “What if we repeated X in space?” when the effects of that were neither relevant nor studied but rather involved some simplification of the material that the professor designed himself that may or may not be right (mechanics in space aren’t exactly intuitive or well understood). There’s nothing to be learned or evaluated from this kind of exam.
  5. Any form of exam design that favors non-merit criteria over innate ability and a decent but not excessive amount of study. This includes exams that are structured in such a way, by difficulty or extreme specificity or other criteria, such that learning the material the normal way is less helpful than borrowing old tests from former students, coming to office hours frequently to receive tips on the material, or having experience unrelated to the class.

I think that about sums it up.

175

My final comment - one last observation, based on AP exam results from a few years ago.

For the AP Chemistry exam, the average score is about a 40%. Getting a ‘5’ requires a 67%; a ‘4’ is a 54%; a ‘3’ is a 41%.

For the AP Calculus BC exam, the average score is about a 50%. Getting a ‘5’ requires a 64%; a ‘4’ is a 54%; a ‘3’ is a 41%…

For the AP Physics C exam (Mechanics), the average score is about a 37%. Getting a ‘5’ requires a 56%, a ‘4’ is a 43%, a ‘3’ is a 34%.

The curves for many of the other AP exams (STEM and non-STEM) are similar.

Not too different from what’s been discussed.

176

^PhD pass on Part 2 of my Qualifying exam was 60%.

177

Cuz CollegeBoard is what we all want to model after… you aren’t addressing the issues raised, you are just saying '“this is the way it works”. But I hear no good explanation for it NEEDING to work that way. I have had classes where it didn’t work that way (my Calc class in college comes to mind, actually) – there is no real reason for it except ego, laziness, and inertia.

178

No, people have explained why it needs to work that way. The ability to apply concepts and equations to solve new situations is important in the STEM fields. It is learned by struggling with the problem sets, not by looking at how the problem was done in the book and substituting in the new numbers.

179

Isn’t that what intparent is saying? It’s learned by struggling with the PROBLEM SETS, not the EXAMS.

180

No, intparent was saying that it’s not taught. The problem sets are designed to teach it. Kids who don’t struggle enough with the problems to develop those skills before seeking help tend not to do well. Kids who struggle alone and fail to take appropriate advantage of the resources available (study groups, office hours) also tend not to do well. I think you need to have put a certain amount of intellectual effort into something before seeing that extra step will click.

How many of you would like to see a freshly minted MD? After all, they’ve memorized everything they need to know, right? Why bother with all those years of clinical training? Oh, wait, maybe they might come across cases that are complicated, or cases that don’t fit the exact clinical description they memorized. Well, you be seen by them. I’ll stick with the ones who had some years of experience using the basics they learned in class to apply to the real world, which is inevitably more complicated.