Does Anger Management Actually Work?

Search this question and you are sent to nursing journals, to index pages for open-access databases, and to two treatment centres whose headline is the word evidence-based. None of them answers the question a parent is asking, which is whether the thing works, by how much, and for what.

Here is the answer in plain English, and here is the part of it that is genuinely not known.

What has actually been measured?

Two bodies of research, and it matters which one anybody is quoting at you.

The first is cognitive behavioural treatment for anger in children and adolescents. The central study is a 2004 meta-analysis by Sukhodolsky, Kassinove and Gorman, which pooled 40 studies, 51 treatment-versus-control comparisons and 173 effect sizes, covering 1,953 children and teenagers. Eligible ages ran from 6 to 18, and the average age in the treatment groups was 12.5. It is clinical research, done in clinical conditions.

The second is anger programmes run in schools. Gansle’s 2005 meta-analysis pooled 20 peer reviewed studies of those.

There is no third body of research covering the commercial market: the online courses, the weekend workshops and the local classes sold to parents. That gap is not a technicality. It is the reason this page cannot tell you whether the course you have a browser tab open on does anything at all.

How big is the benefit, in words rather than decimals?

For the clinical treatment, the overall effect was d = 0.67, which sits in what researchers call the medium range.

Translated: it is real, it shows up reliably across studies, and it leaves you with a recognisable version of the same teenager. Fewer incidents, shorter ones, more of the arguments landing somewhere short of the wall. It is not a personality transplant and no honest reading of that number promises one.

The authors said as much themselves: “CBT is an effective treatment for anger-related problems in youth. Its effects are consistent with the effects of psychotherapy with children in general.” Read the second sentence twice. It is a ceiling statement, not a sales line. Anger treatment works about as well as child therapy works, which is well enough to be worth doing and not well enough to be a cure.

The result also held whether or not a study had been published. Published and unpublished studies gave close to the same figure, which is a good sign in a literature where the reverse is common.

Which parts of it do the work?

This is the most useful finding on the page for a parent choosing between options.

Effects varied a great deal by what the treatment actually consisted of, from 0.36 at the bottom to 0.79 at the top. The top of that range was skills training. Close behind it were multimodal programmes, which combine several approaches. Problem-solving work sat between the two. At the bottom, at 0.36, was affective education: talking about feelings, naming them, learning the vocabulary, without a practice component.

The same meta-analysis looked at which teaching ingredients went with the larger effects. Three were statistically significant: feedback, modelling and homework. In ordinary words, the programmes that did better were the ones where the teenager was shown the skill, made to practise it, and told how they did.

So if you are looking at a programme, the question is not how caring it sounds. It is whether anyone is going to practise anything.

Why do school programmes come out smaller?

Because they are shorter, lighter and delivered in a building with other priorities.

Gansle’s pooled figure across all outcomes was d = 0.31, about half the clinical figure. The largest single effect, .54, was for anger and externalising behaviour, which is the outcome most parents care about. The remaining domains came out lower, and the published abstract does not attach its numbers cleanly enough for this site to say which figure belongs to which, so we will not guess.

One line from that abstract deserves more attention than it gets. Treatment integrity means checking that a programme was actually delivered the way it was designed. Around 60 percent of the twenty articles discussed how important that is. Two of them measured it. The follow-up results, the ones that would tell you whether anything lasted, came from too few studies to lean on, and the author says so.

That is not an argument against the school group. It is an argument for keeping your expectations proportionate to it, and for asking who is delivering it and from what.

What does it not fix?

Three honest limits.

Who you ask changes the answer. In the clinical meta-analysis the effect varied by informant from minus 0.07 up to 0.69. Teenagers, parents, teachers and observers do not agree about how much better things got, and at the low end of that range the average measured benefit was nil or very slightly the wrong way. A programme that reports only self-report is reporting the easy end.

Treatment research is not life. These effects were observed in studies, with clinicians, with supervision. They do not transfer automatically to a course bought online, and they do not transfer to a parent reading a website, including this one.

Nothing here is a claim about your child. These are averages across nearly two thousand young people. They describe a distribution, not a prediction.

And one thing anger treatment is not for: it does not settle what happens with the school, the damage or the consequences at home. Those run on their own track.

What about the class the court or the school named?

The one federally published curriculum we could verify is SAMHSA’s anger management manual, a twelve week cognitive behavioural group model. It is free, it is real, and it was written for adults in substance use and mental health treatment, not for teenagers in a school programme.

Anger management is not a protected term. It is not a licensed profession. A manualised twelve week group run by a clinician and a two hour online course sold for forty dollars are both marketed with the same two words. Because nobody has studied the market as a category, this site will not tell you that anger management classes work, and it will not tell you they do not. It will tell you to ask the three questions in the last item below and to read what anger management classes actually are before paying anyone.

What this looks like in an ordinary house

Bettina is fourteen and has been in two shouting incidents at school this term. This is an illustration rather than a real family.

Her father finds three options. A forty dollar online course promising results in a weekend. A school-run group that meets six times at lunchtime. A clinician who runs a twelve week group and has a waiting list until March.

He does not need a decision tree. He needs three answers. The online course cannot say who wrote its curriculum. The school group turns out to be run by a counsellor working from a published programme, with practice tasks between sessions. The clinician’s group is manualised, involves role play and homework, and the waiting list is real.

He books the clinician’s group for March and takes the school group now, on the grounds that a smaller effect starting on Tuesday is worth having. He does not buy the weekend course.

By summer Bettina has had one incident instead of two, and the second-to-last argument at home ended with her leaving the room instead of the kitchen door. That is roughly what a medium effect looks like when it arrives in a house. It is worth the Tuesdays. It is not a new child.

If what you actually want to know is whether the amount of anger justifies any of this, is this amount of anger normal takes that question first, and the brain is not finished at 25 deals with the claim people will offer you as a reason to wait it out instead.