The results from the 3rd Class Test were the most disappointing I’ve ever seen. Yesterday was Class Test 4 and the results are only a shade better. Among those who took the test, exactly the same class average as Test 3, 74% (C). But that average comes from way more A’s and A-‘s (44, up from 26 last time), fewer B’s, C’s and D’s (good) but more dreadful fails. Oddly, we also had way more no-shows (53, up from 19 last time). So the best we can say is that the decline in test performance across the semester has been arrested. The average grade among those who passed was 80% (B-).
Two students got 100% on my ask-28-questions-grade-out-of-25 algorithm, but the highest score was 25/28. The rest: A, 14; A-, 30; B+, 39; B, 33; B- 24; C+, 20; C, 16; D 43; Fail, 53.
When I look at the performance on the individual questions, there were what I would term under “performance problems” with the questions involving stuff I’d gone over in class in the most recent weeks. Perhaps that’s a consequence of the low attendance:
- Just 40% of students in the class thought the strongest reason to think a Zombie apocalypse couldn’t happen is that it hasn’t so far; maybe that’s because only 49% of students were in class when I made that case.
- Just 52% of students were in class when I discussed the famous ‘experiment’ where Barry Marshall drank H. pylori laden broth, only to get sick not ulcers, thereby providing (anecdotal) evidence against his hypothesis that H pylori causes ulcers. Only 45% of the class knew that.
- Just 60% of the class was there when I showed the video that makes clear that to a reasonable approximation, the Drake equation shows that the main determinant of whether we will detect life in the Milky Way is how long civilizations transmit detectable signals; 53% of the class had taken that on board.
Most heart breaking:
Which of the following is currently an open question in science?
(a) the nature of dark energy
(b) the safety of childhood vaccines
(c) the cause of climate change
(d) the chemical composition of celestial bodies
(e) all of the above.
A staggering 55% of the class chose (e), meaning that I failed to get across the strong scientific consensus on (b) and (c), and many were not paying attention to Jason Wright’s strong guest performance when he showed we learned how to do (d) in 1859. I weep.
Just one question seems important for me to go over in class today. That’s one which makes clear I have not made clear enough that statistical significance per se tells you little about how big an effect is. Statistical significance does not necessarily mean biological significance.