We have long known that classroom management matters. It matters for learning, for behaviour, and for the well-being of both pupils and teachers. Yet what teachers actually do to keep lessons running smoothly often remains vague. It is usually described in broad terms such as “being strict”, “being warm”, or “supporting autonomy”. There is not much attention to the fine-grained reality of everyday classroom interaction.
A recent large-scale study by Mei Tan and Dorottya Demszky offers a different perspective. Instead of relying on surveys or global observation scales, the researchers focused on something that is always present in classrooms: language. They analysed transcripts from more than 1,600 primary school lessons and used natural language processing to identify teacher utterances that function as classroom management talk. This includes all forms of language aimed at maintaining order and guiding behaviour. Rather than teaching subject content, these utterances serve another purpose.
The results are striking. On average, almost one-quarter of what teachers say in class is directly related to classroom management. About 7% of all utterances consist of explicit behavioural interventions, such as reprimands or direct attempts to control conduct. Classroom management language ranges from barely noticeable cues like “sssh” or “eyes on me”, to simple instructions such as “sit down” or “stop talking”. It also includes warnings about consequences, threats of sanctions, or references to parents and removal from the classroom.
What is particularly revealing is where the variation lies. Differences among teachers are far greater than those between schools or lesson moments. In other words, classroom management language is mainly a teacher-level characteristic. It reflects relatively stable patterns in how individual teachers interact with pupils. Rather than being driven primarily by context or by particularly difficult classes, it is influenced by the teacher.
Experience also plays a role. Early-career teachers tend to use more behavioural language and more reprimands than their more experienced colleagues. This is not surprising. Beginning teachers are still developing routines and learning how to regulate classroom dynamics efficiently. With experience, teachers appear to shift towards language that is more instrumental and less emotionally charged. Even if the overall amount of management talk does not necessarily decline, the style changes.
The study also raises uncomfortable but important equity issues. Classrooms with higher proportions of pupils of colour showed higher levels of language associated with exclusionary discipline, even when controlling for other factors. This suggests that structural patterns of inequality may already be visible in everyday classroom talk. Such patterns appear long before formal disciplinary measures come into play.
Crucially, the findings should not be interpreted as a simple judgement that “more classroom management talk is bad”. Well-organised lessons inevitably contain many small corrections and cues. The key issue is not quantity alone, but the type and tone of language used and how it accumulates over time. Light-touch guidance is very different from repeated public reprimands or threats of punishment.
What makes this study particularly valuable is its methodological approach. By analysing what teachers actually say, rather than what they report doing or what observers summarise afterwards, it opens a new window on classroom practice. It allows us to see patterns that are normally invisible. We can talk about classroom management in much more concrete terms.
For teacher education and professional development, this offers an interesting direction. Instead of promoting abstract models of discipline or behaviour management, we can begin by helping teachers reflect on their own interaction patterns: how often they intervene, how they phrase corrections, and how their language shapes the classroom climate. Classroom management then becomes not only a matter of rules and routines, but also of everyday words.
In that sense, this research reminds us that classroom order is not created solely by policies, but by thousands of small utterances that together define what learning feels like for pupils.