Natural consequences
A natural consequence is whatever reality produces on its own. Money saved for a bicycle, spent instead on something that breaks by Friday, and the bicycle does not arrive. A coat left at home on a cold morning. A PE kit forgotten. A tower built badly that falls over.
The adult's entire job here is to allow it, and the difficult part is saying nothing afterwards. A natural consequence delivered with an observation that this was predictable stops being a natural consequence — the lesson transfers from reality to the parent, and what the child takes away is that adults enjoy being right. Reality is a more impartial teacher and it does not need a spokesman.
It works in the favourable direction too, which is the half usually overlooked. A child who saves the money and gets the bicycle has experienced something they brought about themselves, and that is the strongest form of reinforcement available to anybody. It also costs nothing and requires no adult at all.
Where it does not apply is straightforward. Anything dangerous. Anything irreversible. Anything so delayed that the child cannot connect the two events. And anything where the consequence lands on somebody other than the child who caused it, which rules out a great deal of what happens between siblings.
Logical consequences
Where the natural consequence is unsafe, absent or too remote, an adult has to supply one — and the supplied version only works if the connection is visible to the child. Otherwise it is a penalty with a story attached.
Four tests are worth applying, and almost every consequence that goes wrong fails at least one:
- Related. It follows from what happened rather than being drawn from a general pool of things the child values. Paint left out and the carpet marked leads to the cleaning up. It does not lead to no television.
- Respectful. Delivered without humiliation, and not in front of an audience. A consequence that comes with contempt is remembered for the contempt.
- Reasonable. Proportionate in size and, above all, short. A bike left out overnight going into the shed for a week is mostly a punishment of whoever has to enforce it for a week, and the lesson was complete on day one.
- Revealed in advance wherever possible. The useful question is whether the child could have stated, beforehand, what would follow. If not, it is a surprise penalty, and surprise penalties teach vigilance rather than judgement.
Delivery matters as much as design. The consequence is the message, and it should arrive without a speech attached. Explaining at length while imposing something converts a clean piece of cause and effect into a row, and the child will remember the row.
There is a reason this works better than straightforward punishment, and the old-fashioned word for it is fairness. Most children past about five have a sharp sense of whether something is just, and a consequence that is visibly connected to what they did tends to be accepted — grudgingly, but accepted — because there is nothing to argue with. One that is arbitrary invites an argument about the adult instead of a reflection on the behaviour, and the argument is usually the more attractive option.
The actual difference from punishment
The distinction is not about harshness, and a logical consequence can be considerably less pleasant than a punishment.
The difference is in what each one is aiming at. A punishment aims to make a child feel bad enough not to repeat something. A consequence aims to let them see what follows from what. The first approach has a structural problem: when it stops working, the only available adjustment is to make it worse. The second does not need escalating, because its force comes from the connection rather than from the discomfort.
There is a reliable test for which one is actually being administered. If the adult is angry at the moment of delivery, it is a punishment, whatever it is being called. The useful response to noticing that is to say so and defer — a consequence decided in twenty minutes' time will be better than one decided now, and nothing is lost by the delay. Active listening covers the related problem of the first three seconds.
Matching it to the age
The connection has to be one the child can actually hold. A five-year-old cannot link effort today to an outcome in a year, so for young children a consequence needs to be close to immediate and small — within the hour rather than at the weekend.
Longer chains become available with age, and by adolescence a young person can follow a line of reasoning several steps out. But the field narrows at the same time: most of what matters by then is not enforceable, and influence has quietly replaced control. A consequence that cannot be imposed without a confrontation nobody can win is not a consequence, and recognising that earlier rather than later saves a good deal.
Worth keeping in view throughout: very few consequences at any age are irreversible. A bad term is not a ruined life. There are many good universities, many routes that do not run through one at all, and almost nothing decided at fifteen that cannot be revisited at twenty.
Physical punishment, and what the evidence shows
The legal position across the UK is now split. Striking a child is unlawful in Scotland and in Wales. It remains lawful in England and Northern Ireland, where a defence of reasonable punishment still exists — a proposed ban in England was dropped in 2025, and proposals in Northern Ireland have since been dropped as well.
The evidence is more settled than the law. Analysis from UCL using the Millennium Cohort Study, which has followed around 19,000 children born in 2000 to 2002, found that children in England physically punished at ages three, five and seven were 5.7 percentage points more likely to miss five GCSEs at grades 9 to 4 including English and maths. Children punished between three and seven were 33 per cent more likely to report risky behaviour towards other people at fourteen — including being 41 per cent more likely to bully a sibling and 35 per cent more likely to have hit, pushed or shoved someone.
The researchers are clear that residual confounding cannot be entirely ruled out, which is the honest caveat on any study of this kind. What is missing on the other side is any body of evidence showing a benefit. The finding is not that physical punishment works less well than the alternatives; it is that nothing has been identified that it does well.
The sibling figure deserves a second look, because it is the sharpest thing in the data. The behaviour physical punishment predicts is the behaviour it is generally administered to prevent. A child who is hit learns that hitting is what adults do when sufficiently annoyed, and they have no reason to think the rule does not apply to them.
Around one in five ten-year-olds in the UK had been physically punished in some way as recently as 2021, so this is not a marginal practice being described from a distance.
Frightening a child out of it
The other instinct, usually reserved for adolescents, is to confront a young person with where a course of action leads. It feels like the responsible use of a parent's superior knowledge, and it has been tested formally.
The test was Scared Straight and programmes like it, which took young people into prisons to be addressed by serving prisoners. The Campbell Collaboration reviewed the randomised evidence: nine studies across eight American states, 946 young people aged fourteen to twenty. Of the seven studies reporting reoffending, the programmes significantly increased the odds of offending. The review concluded that they cause more harm than doing nothing at all and should not be used.
The reason it generalises beyond prison visits is worth understanding. A dramatic worst case presented to an adolescent invites the conclusion that the warning is exaggerated — and once that conclusion is drawn, everything else from the same source is discounted along with it. The credibility spent on the dramatic warning is not recoverable for the ordinary one.
Which points the same way as the rest of this page. A consequence that is specific, near-term and plainly real does more than a remote catastrophe, and that is true at six and at sixteen.
Rewards, briefly
Reward charts are the mirror image and they fail for a related reason. They frequently work at first and then stop, and the child is left in a worse position than before, because whatever reason they originally had for doing the thing has been replaced by a reason that has just expired.
They have a narrow use: a specific, newly-learned behaviour, over a short run, with the scaffolding removed afterwards. They are a poor instrument for anything that needs to last. Praise for the effort rather than the result does the same job without the expiry date.
What it comes down to
Three things, and the rest is detail. The answer does not move under pressure. The consequence is connected to what happened and it is brief. And the adult is not angry when it is delivered.
Most households already know what the consequence ought to be. The difficulty is almost never in choosing one — it is in following through on a Tuesday evening when everybody is tired and it would be so much easier not to. That, rather than any question of technique, is what the whole subject actually turns on.
On where this sits in the wider picture of warmth and limits, parenting styles. For the age at which it all begins, toddler tantrums.