When do detentions work?

In most schools, Mark resists initially in some way: complaining, or muttering perhaps. The sanction is imposed (he accepts a demerit or attends a detention), possibly through additional effort from me or a colleague. Then he’s back in class. And sooner or later, he does/doesn’t do the same thing again.

I’ve questioned the power of sanctions to change individuals’ behaviour. Recently, I tracked down a study answering this question directly; inevitably, the answer was more complicated than I expected. So when do sanctions work? What makes them work? What do we do if they don’t work?

Who do detentions work for? The evidence

The study looked at how 314 American 3rd – 8th graders (8-14 year olds) in one school responded to detentions, suspensions and other sanctions (Atkins et al., 2002). The researchers identified three groups:

• Group 1 didn’t receive a sanction all year (117 students)
• Group 2 received sanctions in Autumn but none in Spring (62 students)
• Group 3 received sanctions in Autumn and in Spring (75 students)*

This graph shows the average number of detentions for students in each group, each term:

Atkins et al., 2002, Figure 1b

For Group 2, detentions apparently worked. They got detentions in Autumn, fewer in Winter, and none in Spring. In the Autumn, Group 2 were indistinguishable from Group 3, in terms of the nature and severity of their misbehaviour. By Spring, their behaviour was indistinguishable from Group 1, who never got a detention.

For Group 3, on the other hand, the number of detentions (and suspensions – not shown on this chart) increased as the year progressed. Detentions led Group 2’s behaviour to improve, but they seemed to make Group 3’s behaviour worse. Not only were detentions and suspensions ineffective, they “may have served as rewards both for students and for teachers (Atkins et al., 2002, p.368).”

So sanctions work for some students, and really don’t help others. This post discusses what makes sanctions effective and what we can do if they don’t seem to be working.

1) What makes a sanction effective?

Potential/actual punishment is meant to change how attractive an action is: a student is meant to think ‘I could call out, but it’s not worth it.’ It may not be effective if it:

• Isn’t immediate: how people choose depends on the options available and when they’re choosing; we prefer immediate rewards and delayed costs. Being seen as funny among peers now may outweigh the cost of a detention on Friday.
• Is uncertain: a delayed sanction is uncertain – a student may hope their teacher forgets, lacks time to follow up or can be talked out of it; the student may be ill or miss school.
• Is a rote punishment: sitting in silence, or even reading, may look like a punishment, but it neither deters nor corrects. Some students like sitting in silence; if they’re not rethinking their behaviour, little is being achieved.
• Creates conflict: Students don’t want sanctions; teachers make mistakes and are often in a hurry. Yet unless we issue a sanction carefully, we may punish the misbehaviour but create fresh conflict

We have to address these challenges if our sanctions are to work. We can:

• Prioritise immediacy over intensity: a ten-minute discussion after the lesson is likely to be more powerful than an hour’s detention on Friday. (If a delay is unavoidable, we may make sanctions feel more immediate by encouraging students to imagine them now: “Think how frustrating it will feel to stay in over lunchtime… you need to focus on the task to avoid that.”)
• Ensure sanctions stick: students must be sure we mean what we say. Laura McInerney’s brilliant blog post on this notes that being willing to use and enforce detentions meant that she “needed to give them out a lot less.” (A schoolwide system may increase certainty).
• Make detentions productive: they needn’t be unpleasant – losing free time is a punishment in itself – but they shouldn’t be fun (reading is not a punishment!) And they should involve something relevant to and amending for their behaviour. Laura’s blog post describes a Book of Consequences: simply asking students why they’re there, why their actions are problematic and what they’re going to do next promises some kind of improvement; it can also reveal who needs a brief reminder and who needs intensive mentoring.
• Say why and what next: Doug Lemov’s Art of the Consequence (2015, p.406). covers this brilliantly. We need to mention the cause, say what the student should be doing, and offer ‘bounce back’ encouragement – then get on with the lesson: “You’re calling out again, we’ll discuss this after the lesson, make the last ten minutes a success.” (We may also put students in credit: “You’ve done so well this lesson, don’t mess it up now.”)

These measures make sanctions more effective. Yet for some students, they still won’t work.

2) Sanctions don’t work for some students

The best-designed sanction may not help some students: Group 3 who got more and more detentions as the year progressed. Atkins et al. note that their findings concur with “a wealth of evidence that a focus on punishment alone will not impact many of the most severe types of offenses or reeducate the highest rate offender,” and that, in many cases, they increase the rate of problematic behaviour (2002, p.369). For example, half of adults reoffend within a year of release from prison; for sentences of less than a year the figure is even higher (46.8%; MOJ 2020, p.11).

A rational explanation for misbehaviour assumes that students calculate:

(1) the benefit that one stands to gain from the crime; (2) the probability of getting caught; and (3) the expected punishment if one is caught. By comparing the first component (the gain) with the last two components (the costs) , the rational human being can determine whether a particular crime is worth it or not.”

(Ariely, 2013, p.14)

Sanctions won’t work if:

1. Students aren’t calculating. They may be behaving habitually, in which case their behaviour is more automatic than controlled. Atkins et al.’s Group 3 differed from their peers: among other things, they were much more impulsive.
2. The benefits (like looking good among peers) outweigh the costs (looking bad to teachers/parents). It’s worth recalling Graham Nuthall’s conclusion: “When there is a clash between the peer culture and the teacher’s management procedures, the peer culture wins every time (2007, p.37).”

If students aren’t calculating, or if our sanctions don’t weigh enough on their calculations, increasing punishments will achieve little – or, as Atkins et al. found, prove counter-productive. I’ve noted elsewhere that I used reward and punishment instinctively, but that this is a naive and incomplete model of human behaviour. I’ve suggested a framework for encouraging behaviour change: we may think about habit formation, social norms and how we frame choices, for example. If sanctions aren’t working, more/harsher sanctions are unlikely to help: we need to identify why the student is behaving as they are, and help them change.

3) Sanctions still matter

Yet misbehaviour still needs to be sanctioned, even if we don’t expect the sanction to change the individual’s behaviour. First, because we don’t know who sanctions will work for until long after the event – letting students off may rob them of an incentive to change.

More importantly, sanctions show the wider community – the class, the school, society – what behaviour is expected and accepted. Dan Ariely (2013) has tested what influences misbehaviour extensively, focusing on cheating. In one experiment, students completing a task for financial reward saw a peer apparently cheating. If the peer’s behaviour wasn’t sanctioned, cheating doubled among the rest of the group (p.198-204).** Some students would never dream of misbehaving. But Ariely’s description of locks in society may stand for sanctions in schools as well. Most locks won’t stop a determined thief. They deter “morally honest people who might be tempted to try your door if it had no lock…. It’s not that 98 percent of people are immoral or will cheat anytime the opportunity arises; it’s more likely that most of us need little reminders to keep ourselves on the right path. (p.38).” Group 1 in the study described above never received sanctions. Maybe some were saints. For the rest, seeing their peers misbehaving with impunity might have tempted them astray. Sanctions help the community, not just the individual.

Conclusions

1. Sanctions work for some students. They are more likely to work if they are immediate, enforced, meaningful and carefully-delivered.
2. Sanctions don’t work for some students. If three detentions haven’t made a difference, we need to do something else which will lead to actual change in behaviour and habits.
3. Sanctions show the whole community what’s acceptable: sanctioning misbehaviour remains important, even if we don’t expect the individual to benefit.

* Numbers don’t add to 314 because some students received detentions in winter as well.

** In another condition, the peer (a collaborator of the experimenter) asks the experimenter if the instructions make cheating possible, but does not cheat themselves. The cheating rate decreased. So it’s not rational calculation but social norms which makes a difference.