My wife was at a conference in Finland last week and sent me this link, along with the slide below, saying, “this might be interesting for your blog.” At first sight, it is about a large-scale experiment in which children in Finland start something resembling kindergarten a year earlier, at five instead of six. Primary school still starts at seven, by the way. This earlier start did not make much difference compared with the existing system. The researchers only found a positive effect for children who would otherwise not have attended early childhood education and care.

But the slide my wife sent me tells a much more interesting story. Before the results were known, the researchers asked around 600 people to predict how large the effects of the new approach would be. They included researchers, teachers, people working in early childhood education and care, and civil servants. So they all made their predictions in advance. The question is: who got it right?

What did they find? First of all, expectations for this innovation were pretty high. On average, the respondents overestimated the immediate effect on academic skills. They also expected those gains still to be visible when children started primary school, which turned out not to be the case. And they predicted positive effects on socioemotional skills that the researchers did not find either.

But those were the averages, because… not everyone was equally optimistic. People working in early childhood education and care generally expected the largest effects. Researchers and teachers, on the other hand, were considerably more cautious. According to the researchers, these differences were statistically significant for almost all outcomes and measurement points.

You can see this, for example, in the predictions for academic skills at age six. On average, people predicted an effect of around 0.16 standard deviations for the new approach. People working in early childhood education and care were clearly above that, while researchers expected a smaller effect. A year later, the same pattern was visible: once again, people working in early childhood education and care expected more than the researchers did.

And yet the more cautious forecasters did not really get it right either. Compared with children who were already attending early childhood education and care under the existing system, there was hardly any difference in the development of academic and socioemotional skills. The study was also so large that the researchers could rule out even fairly small effects on academic skills.

So nobody really got it right, but some groups were clearly more optimistic than others. I find this interesting because it tells us something about expectations and expectation management. My wife told me that during the presentation, it was mentioned that the policymakers surveyed in this study did even worse because they had the highest positive expectations. I have tried to find the actual study using search engines and AI, but nothing seems to have been published yet. So I am sharing all of this with the necessary caution.

The fact that the researchers were the ‘least wrong’ is somehow reassuring to me, and I can understand why. Researchers know how difficult it is to find large effects. Or at least they should know. There is also a certain logic to the idea that the people actually doing it, in this case, the early childhood educators, may have higher expectations than the people who take over the children afterward, the teachers.

But in any case, perhaps it is important to temper our expectations when introducing innovations, if only to avoid disappointment.

So by all means expect something from educational innovation, but perhaps don’t expect miracles.

Those only happen in classrooms.