We all know that someone is going to have to blink in the current BCTF strike, and if media reports are somewhere close to accurate, the sticking points seem to be class size and class composition.
There are very good reasons for this stalemate, though. Primary among those reasons is the fact that the BC government illegally tore up their contract with the BCTF in 2002 and stripped the ability of the BCTF to bargain on class size and composition. That was a stupid move. [EDIT September 5: Also stupid is the government’s inclusion of clause E80 in their new proposals, which would negate any court decision on their previous provocation of the BCTF.]
But more than that, the BCTF insists that class size and composition are vital determinants of student achievement, and this is obviously true. It doesn’t take much to realize that reducing class size will necessarily benefit students such that they will be better able to learn or that increasing support for students who exhibit challenging behaviours (and support for their teachers) will lead to a better learning environment for students. These things are obviously true…or maybe not.
The truth is that before anyone can make a bold proclamation like those above, they need data. By data I don’t mean ‘My kid was in a big class last year and so I know…’ or ‘I read a study once…’ (which usually means ‘I read a headline once…).
What we need to consider is what the preponderance of the data seems to indicate with respect to class size and composition. The way to get that view is to perform a meta analysis by looking at a wide variety of research studies on the issues, controlling for various factors in the data and determining a rating of how much of an effect a certain intervention has on student learning. This rating is called an effect size.
Fortunately, we have that. John Hattie has invested a huge amount of time and effort into analyzing the data on student learning. But he went a step further than performing a meta analysis. He performed a meta analysis on over 1000 meta analyses.
He found that the average effect size of schooling was +0.4. What that number means is relatively inconsequential for this post except for the fact that it gives us a number to which we can compare various interventions. So if the average student in an average class should see an improvement of +0.4, we can see that if an intervention has an effect lower than +0.4 (maybe +0.3), then that effect is weaker than simply putting a kid through a year of school. A negative effect size indicates that a student has regressed as a result of a particular intervention.
There are some structures in our current school system in BC that have measurable negative effects on student learning. One of those is summer vacation, which has an effect size of -0.02. This means that the preponderance of the data suggest that summer vacation has a negative effect on student achievement.
There are other teaching strategies that are extremely powerful such as student reported grades which Hattie calculated has an effect size of +1.44, providing formative evaluation (+0.9), and feedback (+0.75).
So, what did Hattie find in the data on class size and composition? Not what you might think.
With respect to class composition, Hattie found that
- providing comprehensive interventions for learning disabled students had an effect size of +0.77, almost double what should be expected in the average classroom;
- classroom behavioural interventions had an effect size of +0.68.
So clearly, these two data points show that classroom composition supports have a very significant positive effect on student learning, as predicted by our intuitive sense that helping struggling students with a little more support pays off.
With respect to class size, however, the data show that the effect of reducing the number of students from 25-30 students to 15-20 students in a class is positive, but it is a small effect at only +0.22. In addition, there were many interventions that Hattie identified that are entirely appropriate in large classes which have significant positive effects.
- classroom discussion … +0.82
- metacognitive strategies … +0.68
- cooperative vs individualistic learning … +0.59
- peer tutoring … +0.55
- cooperative vs competitive learning … +0.54
- small group learning … +0.49
So reducing class size has a small impact on student achievement and it costs an enormous amount of money whereas using any of the above teaching strategies has a very large impact for very little money.
So, if student achievement really is the central concern for both sides of the dispute between the BCTF and the BC government, then the compromise that we all know will have to happen is that the BCTF needs to drop class size from its list of demands and the government needs to drop class composition as a sticking point.
Class composition has a much stronger effect on student achievement than does class size and there are accessible (and cheap) strategies that all teachers can (and should anyways) employ in their daily teaching routines that not only help manage larger classes, but that have a much stronger effect on student achievement than would reducing class size.
There you have it. If that doesn’t work, then they should go with Rock/Paper/Scissors; best 4/7; winner takes all.
For more information on Hattie’s work, please see visible-learning.org.