High expectations matter. By now, there is quite a bit of research suggesting that what teachers expect from their students can play a role in what happens in the classroom. Research also examines so-called wise feedback: critical feedback in which teachers make clear they have high standards while also indicating they believe the student can meet them. But what happens if you simply ask teachers to do this?

A new study by Yasemin Copur-Gencturk, Ian Thacker and Joseph Cimpian examined precisely this question. The results are interesting, but perhaps not entirely as you might expect.

For their study, the authors recruited 963 US teachers. The researchers asked the teachers to provide written feedback on students’ mathematics work. The scientists assigned them randomly to two groups. One group was simply asked to provide specific feedback as they normally would. The other group received an additional instruction: provide feedback that emphasises your high standards and makes clear that the student has the potential to meet those standards.

Importantly, the students’ work always remained the same, but… the researchers did vary the photographs shown alongside the work. This allowed them to examine whether the feedback differed depending on the gender and race of the student in the photograph. Did this simple instruction make a difference? Well, yes, but…

Teachers who were asked to communicate high expectations were more likely to mention both what was good and what could be improved in the student’s work. This happened in 38.6% of their feedback, compared with 26.4% in the control group. They were also about half as likely to give feedback that mentioned neither strengths nor weaknesses. So far, the good news.

But the teachers also started talking more about the students themselves. Positive comments about their mathematical ability increased from 2.1% to 8.1%. But negative comments about ability also increased, from 0.5% to 2.7%. Negative comments about students’ effort also increased, from 0.5% to 2.8%. This wasn’t what the researchers asked teachers to do. They had only been asked to communicate their high expectations and indicate that they believed the student could meet them.

The effects also varied by student. For both White and Black girls, the high-expectations instruction led to more negative comments about their mathematical ability. White boys, meanwhile, were the only group for whom negative feedback about effort did not significantly increase.

Mind you, these percentages are relatively small. But the pattern is interesting because the teachers evaluated the same student work. Only the photographs shown alongside it differed.

The intervention did not affect everything. In particular, it did not help the teachers become more skilful at analysing whether the solution was mathematically right or wrong. Additionally, the intervention did not provide more tailored suggestions for students.

This study is sobering for anyone who, like me, translates research into practice. As I mentioned, previous experiments with wise feedback have shown that communicating high standards, along with the message that a student can meet them, can have positive effects. But there is a difference between researchers who carefully design such feedback on the one hand and telling teachers to “communicate high expectations”. This study shows that the latter leaves room for interpretation.

Looking at the results, we see that teachers translated the instruction in different ways. Some of the changes seem potentially positive. For example, participants were more likely to mention both strengths and weaknesses. At the same time, the instruction also led the teachers to talk more about students’ ability and effort. Sometimes this happened negatively, and it also varied depending on the characteristics.

This study can not be used as an argument against the importance of high expectations. Nor does it show that the changed feedback harmed students. The researchers studied what teachers wrote, not how students responded or what they learned afterwards. The teachers also had no relationship with these fictitious students.

But while research can tell us that a particular approach works, it is important to realise that translating such a finding into a short recommendation for practice is itself another intervention. That intervention may not produce exactly what the original researchers had in mind.