Written by Christopher Hein
Having been a school teacher as well as an academic investigating the economics of education, the class size debate isn’t just a big topic for me personally, but also one of the big topics of the field. This debate goes way beyond academia; reducing class size is a common item on the political agenda. It is debated not only because of the financial implications of needing more teachers or political ideology, but also because the academic world has no clear answer to whether it is a good idea or not. But why is this the case?
First of all, figuring out if pupils benefit from being in smaller classes is not easy. There are many studies out there investigating this question, but the most reliable evidence requires experiments. As you can imagine, running experiments with a society is more complicated than in a scientific laboratory. But when all financial and ethical considerations have been addressed they involve randomisation, in this case of pupils, to varying class sizes.
The best known social experiment, as these kinds of studies are called, investigating class size is the Tennessee STAR experiment from the 1980s. Here before entering primary school, pupils were allocated at random to differing class sizes. The team evaluating this experiment did find that children in smaller classes outperformed their peers in larger classes at the end of grade one, but surprisingly the gap didn’t grow over time.
Another way of exploring the effect of class size is to exploit the fact that many countries have policies stating the maximum class size and there are statistical methods, known as ‘quasi-experiments’ that are designed to investigate such cases. The most famous research using this ‘artificia’l experimental methodology comes from Israel, where Maimonides’ rule, as this policy is called there, doesn’t allow classes to exceed 40 pupils. As before, the research does not deliver conclusive results. Yes, pupils in smaller classes appear to outperform their peers in larger classes, but this is not consistently the case across all age-groups. This type of research has since been conducted in other countries such as France, Poland and Bangladesh. The findings from France and Poland find that being in a smaller class is beneficial, yet the effect is negligible. On the other hand the findings from Bangladesh show that pupils in larger classes appear to do better.
So what could be going on? In the economics of education we often use the analogy of a factory, where an output, for example a pupil’s achievement on a standardised test is the combined result of a number of inputs, such as
1. Aspects of the pupil him or herself, such as ‘natural ability or talent’, motivation, etc
2. His or her family support and the socialisation he or she receives at home
3. The school he or she attends which comprises the behaviour of the
a. School leaders
c. Pupil’s peers.
It quickly becomes clear that a change in class size is likely to trigger a whole load of other behaviour and it is very hard to say what their overall effect might be for pupil achievement.
A. The pupil’s family
If we assume that parents care about the quality of their children’s education, then it is likely that parents will react in some kind or another. For example, should the parents believe that the reduction in class size has benefitted their children, they might reduce the amount of extra tuition they provide for them and pay for music lessons instead. These examples are not just made up, evidence from Norway and the US suggests that parental support behaviour depends on how they perceive the quality of their children’s schooling.
Similarly parents might also decide to send their children to different schools that the parents think provide a better quality education.
B. The peer group
Another obvious example is the peer group. Research shows that strategically grouping pupils together can indeed be beneficial. This implies that if a policy change would force school leaders to adjust class sizes, how they redistribute pupils is crucial. Of course randomising pupils into classes creates random groupings of pupils in classes, but one needs to remember in the studies above that parents may have chosen the school first before their children were allocated to their class. Therefore a critical factor is the socio-cultural heterogeneity of the school – different schools will have a different intake of pupils.
C. Schools and school leaders
Similar to the above, schools also differ in their intake of teachers. Some schools might be more attractive due to their geographical location so that they can attract different teachers. School leaders play a crucial role, too, as the way they lead and manage their school will affect the teachers they hire and how they keep their teachers motivated, for example. School leaders also decide how to equip which classroom and which class and which teacher should be allocated to them. While schools in countries such as the UK and US might all be well-equipped, this is not necessarily the case all around the world. The effect of resources though is not to be underestimated as the available ‘infrastructure’ will determine the possible behaviour patterns, for example of teachers.
Imagine for a moment a class that suddenly doubles in size. How is the same teacher going to deal with this? The teacher may change they way he or she teaches, some might resort to ‘preaching from the front of the class’ whereas others might resort to group work methods. Similarly the teacher may need to adapt their classroom management methods to manage the potential for more noise or develop new strategies how to identify an individual pupil’s learning needs. The choices he or she makes to do this are likely to affect how well the pupils learn.
So as you can see, whether reducing class size really is beneficial for pupil achievement isn’t that straightforward. Even randomising pupils into classes of different size doesn’t automatically solve all the challenges I outlined above and more research is needed. But regardless of having conclusive evidence for this specific question, we should remain optimistic and use the wealth of evidence that already exists to focus on improving our children’s education in other ways.
This blog is based on the first chapter of my thesis, where I use data from Sub-Saharan Africa to explore some of the mechanisms described here. It starts on page 17.