When I read this paper on Saturday, it genuinely made me stop and think. Nathalie Aerts, Thijs Bol and Eddie Brummelman started from a theory that has become increasingly influential over the past few years. According to the Social Class–Academic Context Mismatch theory, competitive school environments should disadvantage students from lower socioeconomic backgrounds. That was honestly my expectation as well, based on the studies I have read. The reasoning is fairly straightforward: competition aligns more closely with the values many higher-SES children grow up with, whereas students from lower-SES backgrounds are expected to feel less at home, develop less positive beliefs about their own abilities and ultimately achieve less. At first sight, that sounds perfectly plausible.

The researchers then did exactly what you hope researchers will do. They preregistered their hypotheses and analysis plan before analysing PISA 2018 data from more than 315,000 fifteen-year-olds across 69 countries. And that is where things became interesting.

Not only did they fail to find evidence that competitive school environments widen socioeconomic achievement gaps, but they actually found the opposite. Students from lower socioeconomic backgrounds reported slightly higher self-efficacy, a more positive view of their own reading ability and slightly better performance in both reading and mathematics when they perceived their schools as more competitive. The effects are small, something the authors themselves repeatedly acknowledge, but they are also remarkably consistent and point in the opposite direction from what the theory predicted.

Does that mean schools should become more competitive? I don’t think so, but not because I have any ideological objection to competition. Rather, while reading the paper, I realised that I was constantly picturing a different kind of competition from the one the researchers were actually measuring.

I immediately thought about league tables, public rankings, norm-referenced grading, selective admissions, or schools deliberately encouraging students to compete with one another. That is not what this study examined. Instead, PISA asked students how competitive they perceived their school. This included whether students competed with one another and whether competition was considered important. The researchers also calculated how competitive students in each school described themselves.

That is a rather different concept. So the question now is: what exactly are we measuring?

Perhaps students perceive a school as competitive because everyone is trying to outdo one another. But perhaps they also perceive a school as competitive because expectations are high, achievement is valued, teachers communicate clear goals, and effort is taken seriously. In other words, maybe perceived competitiveness is partly capturing an ambitious school culture. If this were the case, I think these findings suddenly look rather different.

A substantial body of research suggests that high expectations, clear instructional goals and well-structured learning environments particularly benefit students from disadvantaged backgrounds. Think of the literature on teacher expectations, collective teacher efficacy or effective instruction. These schools combine ambition with support. Students know what is expected of them, and they receive help to reach those expectations.

Perhaps that is what students are describing when they report that their school feels competitive. Interestingly, this interpretation is not entirely different from the explanation offered by the authors themselves. They suggest that competitive environments may provide clear goals, structure and rewards that give lower-SES students additional scaffolding.

The problem is that, imho, this study never actually measures those things. It does not measure expectations, classroom structure, instructional quality or teacher support. To be clear, I think that explanation remains plausible, but speculative,  mine too.

Another important limitation of this study is that it’s cross-sectional. Once again, we can’t determine cause and effect. Perhaps ambitious school cultures make students perceive their schools as more competitive. Perhaps successful students are simply more likely to experience their environment as competitive. Or perhaps both are explained by some third factor that was not measured. Yes, this is something I need to write again and again.

As I mentioned earlier, the effect sizes are small. That is not uncommon with a dataset of more than 300,000 students. Statistical significance does not automatically imply educational significance. The authors themselves emphasise this point several times. Even so, I do think this is an important paper. Not because it proves that competition is good. But because it subjects an influential theory to a particularly rigorous test and finds precisely the opposite of what had been predicted. Some people might see that as a problem for science. I see it as science working exactly as it should. Theories generate predictions. Good studies test those predictions. Sometimes reality turns out to be more complicated than our theories suggest.

So my takeaway is not that “competition works.” My takeaway is that we may not yet fully understand what students mean when they describe their school as competitive. And that, to me, is an excellent reason for more research.