There are four class tests during the semester. I take the average of the best two as the final test grade. This allows students to improve. Improvement is of course the aim of education. It also eases complaints and fear. For many of the students, their grade on the first class test was their first C or D ever.
The breakdown for the overall test grades for 2014: A, 23 (including two with 100%); A-, 55; B+, 21; B, 29; B-, 27; C+, 13; C, 8; D, 8; Fails, 2.
Interestingly, that means 42% of the class got an A of some sort. The corresponding figure for the blogs is 16%. So despite the endless complaints that my tests are too hard, people actually do better on the tests than on the blogs. I hadn’t thought of that before, but looking at it, I see that was also true in previous years (e.g. 2013 blogs, 2013 class tests). So there is more scope for improvement with the blogs. That’s especially interesting since they can be done any time, on anything, under no time pressure, no exam freak-out scene, with endless help freely available from the TAs and with lots of personalized feedback from the graders….
Memo to self: make sure you mention this to the class next year.
This went well. A class average of 80% among those who showed.
Altogether, 18 students got an A, including 8 with 100% (on my ask-28-questions-grade-out-of-25 algorithm – again, no one got everything right). The remaining students: A-, 37; B+, 24; B, 16; B-, 17; C+, 12; C, 7; D, 33; Fails, 10; no shows (imagine not even trying!), 12.
The only disappointing question was one revealing that half the class were asleep (or absent) when I talked about cancer rates in elephants and blue whales. They have less cancer than humans even though they have more cells, more cell replication and so more mutation potential. It stands to reason, then, that elephants and whales must have methods for controlling cancer we don’t. If we could work out what they are, we might be able to use them to control cancer in humans.
I also slightly worry that I went on about the dangers of generalizing from an anecdote so much during the semester that the entire class completely dismisses anecdotes as evidence of anything. That’s an over reaction. Anecdotes can mean something: the first person to get cancer from smoking was an anecdote. The second was another anecdote, the third another…. You just have to be super cautious about drawing conclusions from them. But don’t dismiss them altogether. If someone keeps a cell phone in their bra and gets breast cancer, that could mean something…
Both those questions were suggested by students (I give extra credit for suggesting exam questions I use). There was also a question in a previous test suggested by a student which most of the class got wrong. I wonder if the students are better at testing each other’s deep understanding than I am. They might be. Perhaps students learn to get on my wave length rather than to deeply understand the material. Multiple wave length’s might stretch them better. Interesting thought.
Class Test 3 ran yesterday. The class average among those who took the test was 83%, up 5% (half a grade) on the second test, and almost an entire grade on the first test. But as always, the average hardly captures the picture. Most gratifying was the number of great performances. No one got everything right, but five students got 26 out of 28 questions, and 16 got 100% on my ask-28-questions-grade-out-of-25 algorithm. Altogether, 33 students got an A, and another 33 got an A-. So 66 students got some sort of A, a third of the class. So much better than any of the previous tests — and maybe a record for any of the 19 class tests I have run so far (must check).
The rest of the grades broke out thus: 16 B+, 13 B, 25 B-, 18 C+, 24 C, 20 D, with 5 fails. Seven students failed to take the test despite reminder emails.
So this is all going in the right direction. As always, what to do about the students who are not doing well? Unless anyone suggests something else (and I am always, always open to better ideas), the best I can do is repeat what I said here and beg and implore students unhappy with their scores to come to the revision sessions I run. For those who came to a revision session last time and still did poorly, the recipe hasn’t changed. Keep at it. Revise, question, ask. Learning requires action. If it’s not yet working, keep going. I can help. But only if asked. If I could figure out how students could learn without effort, I would be very rich.
Class test 2 ran yesterday. The class average among those who took the test was 77%, slightly up on the first test. Some students did outstandingly well. No one got everything right, but two got 100% on my ask-28-questions-grade-out-of-25 algorithm. There were 5 A‘s, 5 A-‘s, 34 B+‘s, 23 B‘s, 38 B-‘s, 21 C+‘s, 16 C‘s, 40 D‘s and 7 fails. Eight students did not take the test despite reminder emails.
What to make of these results?
The glass half full view: The test was a bit harder than last time, so it is good that the average is up a bit and in general more students are getting B’s and fewer getting a C+ or less. So things are going in the right direction and as we go over and over things for the rest of the semester, that trend will continue.
The glass half empty view: About a quarter of the class are not getting it. These are the students I really care about. Making a difference to them will generate the real impact of the course. These students are not on my wave-length. They see the world differently. If I can help empower them to see the world critically, big things will happen to them. They’ll have an impact on me too.
How to do I empower them? Well, I’ll run revision sessions on this test. I will beg, grovel and implore poorly-performing students to come to those. Experience tells me that if anyone turns up, those sessions will be dominated by students who did really well (and they are very welcome too). I will beg, grovel and implore all the students to go over their tests and pop quizzes and identify their problems, the first step to solving them. Experience tells me that the students doing well are good at that. As always, the challenge now is to figure out how to motivate the poorly-performing students to help themselves to help me help them improve. Everyone can do this. Not everyone will. My job is to make as many as possible rise to the challenge. I’m not going to lower the bar.
Many students did poorly on the first class test and with the second class test looming, they’ve been emailing asking for office hours (American-speak for wanting a 1:1 meeting). By-and-large, I’ve been refusing to see them. Not because I want don’t want to talk or help, but because I somehow have to help students to try to first help themselves. Teaching them that is perhaps even more important than the content of the course (?).
My view is that students who are doing poorly have to go over the pop quizzes and the class tests and figure out what they can’t understand. I try to force them to do this with the class tests – I tell them what they got wrong, but not what the right answer is. I want them to try to figure it out. If they can’t, and then I force them to write to me explaining what they can’t understand and why they can’t understand it, well hell, two things happen. First, I’ve forced them to identify what they can’t figure out and think hard about it. Often times (most times?) when they think hard about it, they understand it. And second, once we have put our joint finger on what the problem is, I can help. I can’t help with undefined problems. So far this semester only two students have written with specific problems (Class Test X, question Y, I do not see why it could not be (a) and (c) because… or, I do not understand what a confounding variable is because…). I can work with those students. Together, I think we’ve made progress.
Industry, employers, scientists, business-folk, politicians – everyone in challenging positions – agree that identifying the problem (defining it well) is the first step to solving it (and maybe the hardest part). And we have it in us all to work at that. We don’t need a professor to do that for us. Professors who just provide the answers to questions no one understands cheat everyone.
Improvement is rewarded on this course. I set the grading algorithm up so that poor performance early on has no impact on the final grade.
That honeymoon period stops Monday with Class Test 3. I take the top two marks from the four class tests; this means for many students, Monday’s score will be one they are stuck with. Similarly, we are now in the 3rd Blog Period; I take the best score from three periods, so this blog period will determine the final blog grade for many students.
Some students have already done well on their blogs and class tests. In principle they need do nothing more than the final exam. But that is a bad strategy: the more practice people get on my class tests, pop quizzes and class material, the better they will do in the final exam.
A couple of years ago, a student who failed the course wrote to say he did not realize blogging counted 40% of the grade. I have just written to the 22 students in this years class who have yet to start blogging. You can’t even be in the running to pass this course if you don’t blog. It’s as simple as that.
Here’s how things currently stand for overall grades. Fail, 26; D, 11; C, 22; C+, 33; B-, 24; B, 33; B+, 11; A-, 9; A, 1. Right now, no one is stuck with those grades. From Monday, things start to congeal.
Class Test 1: 60 students with unacceptably low scores (D or worse)
Class Test 2: 43 students
with unacceptably low scores.
I did the first test revision tutorial last Thursday at 5:30pm. At 5:30, this is what the classroom looked like:
Ten minutes later, it looked like this:
Four more turned up near 6pm. I think all six learned a lot. One said as she left that she felt a lot more confident now. Fabulous.
Tonight I did a rerun. Nine students appeared. They all seemed more confident as they left.
I like the banter of these sessions, and I like looking into the whites of students’ eyes. Its the only way to figure out what works and what does not.
Several of the students who showed had good scores. So dozens of poorly performing students aren’t taking advantage of these sessions. All something of a mystery to me.
A, 43; A-, 58; B+, 28; B, 14; B-, 7; C+, 12; C, 3; D, 2; Fails, 5.
Two of the fails look like people who dropped out without dropping the course. Tragically, the other three fails failed because they did no or insufficient blogging.
At the other end of the spectrum, five students got >100% with all the extra credits etc.
It always amazes me how much better students do on the final exam than they do on the four class tests I run during the semester. Obviously I like to think it is that because my teaching eventually makes an impact. More likely, it is the fact that 20% of the final grade rests on the exam….
The average exam score was 87%, 10% up on the class tests. Again, no students got every question correct, but ten got all but one of the 28 correct, and fully 35 got 100% in my grading algorithm
. Pleasingly, for all questions, the majority answer was the right one.
Grade distribution: A, 74; A-,18; B+, 20; B, 19; B-, 16; C+, 8; C, 1; D, 11; Fails, 5, including 3 no shows.
The exam had the same structure and questions style as the class tests. The first half of the exam is on things discussed in class. I was able to populate that part of the test entirely with questions suggested by the students. I give 2.5% extra credit for any suggested question that I use; about ten students made suggestions. Any students who felt the first half of the test was tough can blame those class mates.