Set an enthusiastic teacher in front of their class, the students will become enthusiastic too. And vice versa, enthusiastic students can also make their teacher enjoy teaching more. It does sounds logical, and in my own research I once found how contagious passion can be. This also fits nicely with the idea that what teachers and students do and feel influences each other. This is also an underlying idea behind the Rose of Leary or his Circumplex. We use this modeloften in education to talk about how the behaviour of one person can evoke a certain behaviour in another.
A new study by André Seidel, Mareike Trauernicht, Diana Raufelder and Rebecca Lazarides in the British Journal of Educational Psychology investigated this reciprocal influence. You probably guessed it already. It turns out to be a bit more nuanced than I described in my first paragraph.
In their work, the researchers looked specifically at the emotions of enjoyment and anxiety. Are these passed on between a teacher and their students and vice versa? But also, what is the role of the quality of their relationship in this? The researchers originally expected a fairly straightforward interaction, as I already described. Their hypotheses therefore go in both directions. Teachers’ emotions would predict students’ emotions. Students’ emotions would predict teachers’ emotions. And the quality of the relationship could possibly function as an intermediate link.
What I like about this study is that the researchers did not simply ask teachers and students at a single point in time how they felt, what would tell us not that much as it’s maybe just that one moment. Instead they did a (limited) longitudinal approach in which they followed 1,638 students (average age of almost fourteen) and 118 teachers. The students were spread across 86 German classes. In the study were three measurement points, each about 6 months apart. The researchers looked at two subjects separately: mathematics and German.
The three times, teachers indicated how much enjoyment and anxiety they experienced when teaching that class and how they assessed the quality of their relationship with it. The students reported how much enjoyment and anxiety they experienced during the lessons.
To analyse the data, the researchers used multilevel cross-lagged panel models. That sounds more complicated than it is. I will try to explain it using this study. Suppose that the teacher’s enjoyment at the first measurement point is related to students’ enjoyment six months later. The researchers then took into account how much enjoyment those students already had at the first measurement point.
By doing it this way, they could get a bit closer to answering whether changes at one measurement point are related to later changes at the other. This is more useful than looking only at whether both values correlate at the same moment. You noticed it; I wrote correlate. It remains, of course, an observational study and not an experiment. But in this case I think this is slightly less relevant.
First of all, the researchers found that during math classes, if the teacher had a better relationship with their group at the first measurement point, this was indeed linked to the teacher experiencing greater enjoyment at the second measurement point. Sounds both logical and nice. That same enjoyment was then again linked with a better relationship when they surveyed everybody the third time. The enjoyment students reported at the second time also predicted how positively the teacher assessed the relationship with the class six months later. This may look like “Mission accomplished”: something reciprocal is clearly happening.
But… this is not the simple chain reaction the researchers originally expected to find. The data didn’t show that more enjoyment for the teacher, through a better relationship, would lead to more enjoyment among students. For anxiety, the picture was somewhat similar. Student anxiety predicted a less positive later assessment of the relationship by the teacher.
In the case of German as a subject, the researchers came across another variation on this theme. There were fewer examples of this reciprocity through the relationship. There was, however, a direct connection between teacher and student. Increased teacher anxiety at the first measurement point predicted increased student anxiety at the second measurement point. The standardized association was β = .26. In contrast, for enjoyment in German lessons, there was only one significant cross-lagged connection.
So the emotions of teachers and students seem to be connected, with indications of reciprocal influence. But no, there does not appear to be a universal emotional cycle in which the teacher’s enthusiasm, through better relationships, sparks students’ enthusiasm, which in turn makes the teacher more enthusiastic again. The researchers themselves conclude that they did find transmission effects, but that the paths differ a lot more than they had expected beforehand.
What fascinates me personally the most is the difference between mathematics and German. There is an important warning here: the students, of course, had different teachers for both subjects. It’s therefore impossible to conclude that the subject itself is the cause of these differences. The researchers treat the school subject as a possible contextual factor and, among other things, point to differences in social dynamics across subjects. I am not going to make any jokes now about the personality differences between math teachers and German teachers.
This brings me back to Leary. To be clear, this study is not about his Rose or Circumplex. It cannot provide evidence pro or contra the model. The researchers do not, for example, measure the interpersonal behaviours Leary discusses. But I do think this study adds an important nuance to how we sometimes use similar models in education.
The general idea that teachers and students reciprocally influence each other’s behaviour, and emotions fits perfectly well with what this study describes. But the step from reciprocal influence to predictable reciprocal influence turns out to be a bit one, maybe even a step too far.
Based on this study, we can see that things can become risky when one translates a model into statements like: “If you, as a teacher, do X, then the student will respond with Y.” This study suggests that we should at least replace that “will” with “can.” The direction of the relationships can differ. Several expected relationships do not appear at all. Also, the school subject seems to matter.
This study, like all studies, has its limitations. I already mentioned the observational nature. But also the sample is both large and small at the same time. While there are many students, there are relatively few teachers/classes. This limits the analyses at that level. Also important to note: only the teachers rated the quality of the relationship. When I write about the ‘quality of the teacher-student relationship’, it may sound as if the study measured an objective characteristic of that relationship. In reality, we only know how the teacher experiences the relationship with the class. Students may see it very differently.