This past term I again taught “The Physics of Life,” my biophysics-for-non-science-majors course. It went fairly well, but I was less happy with it than I’ve been in the past. Why? And what can I do about this in the future, especially this coming term, when I’m again teaching a (different) general-education class for non-science majors?
Graph no. 1 Here’s a histogram of the final grades for the course:
The most abundant grade was a “C.” There were several “A”s, but relatively few “B”s. (Ignore the “F”s, since most or all of these are from people who dropped the course. The “D”s are real.) Students either did quite well, or quite badly, with fewer in between than one might expect. Is this term particularly unusual?
Graph no. 2 Here are the grade histograms for all the times I’ve taught this course:
The “B”s are a bit more squished than before, and the “C” block is a bit fatter, but it has always been the case that there are fewer As and Bs than I’d like. Of course, I could just conclude that this is fine; I structure the course as best as I can, and students get the grade they get. Let’s dig a little deeper, though.
If you can read this, you get an “A”
There were several components that made up the overall course grade: a handful of quizzes, two exams, weekly homework assignments, writing assignments on popular science articles, short quizzes on assigned readings, and more. For each of these, I calculated the correlation between students’ scores on that component and their overall course grade. (Actually, I used the overall grade calculated without counting that component, to avoid a trivial correlation.) The piece with the strongest correlation with the overall grade was the reading quiz score:
Graph no. 3 Overall course score vs. reading quiz score.
What are these reading quizzes? Before about half of the classes, I would assign a few-page reading, perhaps a newspaper article, or something I wrote myself that gave some background information on class topics. The quiz questions were easy, not testing difficult concepts but rather key facts or points from the reading. Most people did well with these, but many students did quite badly, with answers that implied that they didn’t bother to read at all. (Or, I suppose, they might have abyssmal reading comprehension skills.)
Along similar lines, I asked students in the mid-course evaluation I handed out how much time they spend per week on homework, and on reviewing course material (notes, readings, etc.). The distribution of the total time looks like:
Graph no. 4 Study time.
The (self-reported) mean is just 4 hours! (N=28 respondents, about 1/2 the class.) University policy, by the way, prescribes about 8 hours per week [1]. Admittedly, this is a low-level general education course, and I would expect students to spend more time on courses that are central to their major and less time on this, but 4 hours, which includes doing homework, is too low if one hopes to have a decent understanding of the material in a university course.
Overall, my impression from interaction with students and from data like that shown above, is that there are too many students who spend too little time to do well in this, and similar, courses.
Carrots, sticks, or neither?
What should I do about this?
One option is, of course, nothing. Students seem to like the course …
Graph no. 5 Student evaluation scores for “Course Quality” and “Instructional Quality” for my Physics 171 course, as well as the Physics department average for all “100-level” (non-science-major) courses since 2011. The possible range is 1-5.
… and I don’t get any tangible rewards by trying to make the course better.
The other option is, of course, something, both because of my own intrinsic motivations and because I do actually believe that learning the material in these courses makes students better people and the world a better place. In last term’s course, I hope that students do gain an understanding of how they and the other living things around them function (or malfunction) due to physical laws and constraints, and in next term’s course, I hope that students will understand the principles behind energy use, applications such as ground transportation, and climate change well enough to have intelligent conversations about current events and policy. So: what should I do?
Here are a few things I’ll try this term.
- Again have reading quizzes, but have more of them, and give them more weight towards the overall grade, which may help them be taken more seriously. (Of course, this being effective assumes that the readings are actually useful, and aren’t just a signal of conscientiousness.)
- Explicitly tell students about my and the University’s expectations for time committments. (I’ve never actually done this, because I assumed, wrongly, that it was obvious.)
- Ask students to state, at the start of the term, how much time per week they expect or intend to devote to the course, and what grade they expect to get. Perhaps the act of declaring this explicitly will serve as a motivation.
- I do this already, but: stress that I expect everyone to be able to succeed in the course. As noted, many people do very well and, as intended, they represent a variety of non-science majors.
(Aside from developing more readings and reading quizzes, this thankfully doesn’t require much more work from me.)
We’ll see if this works! I don’t expect major impacts — I am constantly amazed by how many students are spending $10-30 thousand per year on tuition with remarkably little introspection about or engagement in the process. But even minor impacts would be nice!
Misc. item #1: clickers
If you’re curious about which component of course had the weakest correlation with the overall score: it’s the clicker-based in-class participation grade. Clickers are great for helping student engagement and getting real-time feedback on what students understand, but a participation score for this has almost no relationship to understanding anything.
The linear regression is silly, since it’s dominated by the handful of points at the left.
Misc. item #2: comments on homework
My favorite comment from the course evaluations was:
The homework assignments were very useful. I did poorly on almost all of them, which helped me know that there were concepts I wasn’t understanding.
I wish more students had this attitude!
Misc. item #3:
Yes, I know there are six graphs. The last one doesn’t count.
Today’s illustration…
…a cardinal, based on a photo in a guide to birds. My first painting of 2017!
Footnotes
[1] The expectation is 120 hours per term for a 4-credit class, which this is. There are 10 weeks per term (not counting final exam week), and class meets for 4 hours per week, leaving 8 hours per week outside of class as the expected time required.
The Eighteenth Elephant 
Hi Raghu,
Interesting post. Thanks. one thought I had is perhaps its not true that all students could do well in this course if only they tried. Recall Steve Hsu’s and Jim Schombert’s analysis of SAT correlation to physics ability. It doesn’t imply there is a born-in physics gene, but it does show that the SAT (unbeknown to its authors) tests primarily physics ability (there is no correlation of SAT to success in any other major).
Physics is different.
Your course is general science, so is supposed to be accessible to any student. But the truth is that the SAT correlation probably still resides for your course by its very nature – the ability to think logically and apply abstract concepts to real world problems.
I don’t don’t think you should dumb down the course so low-SAT students can pass it. (Actually it would be great if you could acquire SAT scores for your students, but that’s probably not possible.) But you might have to face the fact that nothing you can do will remove that residual tail of Ds. (Although you might be able to decrease the Fs, as physics education research shows)
Thanks for the comment. I strongly doubt this is the case for this course, both because of its content, and because I have talked quite a bit one-on-one with students (e.g. at office hours) who do struggle with math and logic, but who do quite well in the class if they put a reasonable amount of time into it, and who are (justifiably) proud of this. I don’t doubt that SAT scores are correlated w/ grades even in a 100-level class, as they are with everything, but it’s certainly not a sharp threshold that demarcates the passing and non-passing students. It would, I agree, be interesting to know how performance in my class correlates with other things, like SAT score, overall GPA, etc.
Are the reading quizzes a minor component of the grade? If they appear ancillary, and don’t count for much, then those that do well might be the people that care enough (about grades or learning) to spend time on that assignment.
Yes, that’s true — they’re only 5% of the total, and the correlation with overall grade probably points as much to conscientiousness of the students who do the reading regardless of points as to the value of the readings themselves. This term, I’m making reading quizzes 15% of the overall score, and having more of them. We’ll see what happens!
Hi Raghu,
Thanks for the insight. A thought on reading quizzes: Have you considered having students taking reading notes? A professor at my undergraduate school required this for physics courses, with the point being to ask meaningful questions about the text within the notes, as opposed to summarizing the text. While I struggled with these notes and initially did much of the latter, I found myself engaging with the text much more when I started asking questions the text did NOT answer. The added benefit of asking these questions was that the professor made a concerted effort to answer them either on the notes (we handed them in each week) or in class. Just a thought.
read: take* (It’s early)