As promised, I will briefly discuss the strange paradox I saw in a slide show presented at a recent teachers’ conference I attended. The paradox is that in the OECD countries participating in PISA testing, the correlation between interest in science topics and performance on the PISA science test is negative. The point of the presenter at the conference was that we need to find ways to preserve or improve interest in science while keeping achievement high or raising it.
Much as I enjoyed the presentation, I wondered whether there aren’t other kinds of problem embedded in this paradox. Francesco Avvisati, the OECD analyst reporting the paradox in a larger presentation, thinks there may be. He notes, for example, that “self-enhancement,” which includes saying good things about oneself (including how keen one is on science), is viewed differently in different cultures. Not everyone in this world who is Peter Pan’s age goes around singing, “I’ve Got to Crow!” The tendency to crow or not appears to be bound to the culture of one’s upbringing, and may affect answers. Given that the same regions of the world that produce good test results also have low “self-enhancement,” as it’s called, one wonders whether some cultures talk the talk while others walk the walk. Such differences would complicate taking remedies from one part of the world and trying to apply them in another.
It is also interesting to find that teachers’ concentrating on discussing and explaining the applications and uses of science in the modern world is most likely to raise students’ interest in science, while their use of hands-on methods of instruction such as labs is most likely to raise science test scores. Readers of these postings will know that I favor genuine over “virtual” classrooms. PISA’s findings seem to support such a preference because they focus on things that teachers can do but that software cannot. (And “autonomous student inquiries” favored in constructivist classrooms are the worst thing for improvement of interest or test scores, suggesting that teachers’ guidance really matters, which of course all teachers, but not all educationists or software vendors, already know.)
Finally, Avvisati notes that the interest/test score paradox is undercut by a fascinating finding that within any particular school, the correlation between individual students’ interest and their test scores tends to be positive, but that this correlation vanishes when numbers are mixed between schools and average correlations are sought. Avvisati suggests that these results imply “diverging effects on student interest and test scores” in different school cultures.
If he is right, firing the teachers at schools that have lousy value-added averages may do little to improve scores if the schools’ cultures work against good teaching and learning. Maybe some “further study is needed” in the real causes of trouble before NCLB and RAT bring in the headsmen.