Am I the only parent who yells at my kids?

The actual numbers

Most articles answer this question with reassurance and no evidence — don't worry, every parent loses it sometimes! — which is useless, because you can tell it is being said to make you feel better rather than because anyone checked. So here is what was actually measured.

The best prevalence data comes from a nationally representative survey of 991 US parents published in the Journal of Marriage and Family (Straus & Field, 2003). It measured psychological aggression using the Parent-Child Conflict Tactics Scale — a category whose most commonly endorsed item is shouted, yelled or screamed at, alongside items like swearing at the child, calling them names, and threatening to send them away.

The findings:

• ~90% of parents reported at least one such act in the previous twelve months when their child was two years old.
• 98% reported it by the time the child was five.
• From ages 6 to 17, the rates stayed in the 90% range.

The authors' own summary was that psychological aggression is a near-universal disciplinary behaviour among American parents. Not common. Not widespread. Near-universal.

One honest caveat, because this site has a rule about not stretching a source further than it goes: that 90% is for the category, not for yelling on its own. Yelling is the most frequently reported behaviour inside it, but the figure also captures the other items. The precise share who yelled specifically is lower than 90% and higher than anything that would make you unusual. If someone quotes you a clean "90% of parents yell," they are rounding off a caveat that the study itself does not.

Alongside it, Pew Research Center's 2023 survey of US parents found 41% say parenting is tiring all or most of the time and 29% say it is stressful all or most of the time. That is the baseline the yelling happens on top of — see parental burnout for what that does to your ceiling.

Why it stays invisible

Here is the thing worth understanding, because it explains the whole strange experience of being a parent who yells: a behaviour can be almost universal and still feel rare, if nobody reports it.

Parents do talk about losing their temper. But they talk about it in a specific, pre-softened register — I snapped this morning, I have zero patience today, bedtime was a nightmare. What is almost never said is the actual content: the volume it came out at, the sentence you used, your child's face, the specific thing you said that you would not want repeated. That part stays in.

So every parent in the conversation is contributing an edited version and receiving edited versions back. Each one privately holds an unedited version of themselves and compares it against everyone else's edit. The predictable result is that a large group of people, all doing roughly the same thing, each conclude they are the worst one in the room. Social psychologists call this shape pluralistic ignorance: a norm that almost everyone privately violates and almost everyone publicly assumes they alone violate.

This is not a small effect at the margins. It is the reason the question at the top of this page gets typed into a search box hundreds of times a night by people who are each certain they already know the answer, and are each wrong.

The comparison you're running is rigged

There is a second mechanism stacked on top of the silence, and it is about where you see other parents.

You see them at the school gate. At the birthday party. In the supermarket aisle. In daylight, in public, during the fifteen minutes of the day when every adult present is performing competence at exactly the same intensity you are. What you do not see is that parent's hallway at 7:40am, their car on the way home, or the twenty minutes after the party when the sugar lands.

And you see yourself with total access — including the internal monologue, the thing you nearly said, and the replay at 11pm. You are comparing other people's exteriors to your own interior. That comparison has one available outcome, and everyone running it gets the same one.

Social media makes it worse in a way that is worth naming plainly, because it is a selection effect rather than dishonesty. The parenting content that circulates is produced by the parents who are having a good enough day to film, edit and post. Nobody posts the hallway. The absence of the hallway from your feed is not evidence that other people don't have one; it is evidence that hallways don't get filmed.

Common is not the same as fine

This is the part where a lot of reassurance pieces quietly go wrong, so let's be explicit about it.

Prevalence is an argument about you, not about the behaviour. The fact that nearly all parents yell tells you something true and useful about whether you are a uniquely bad person. It tells you nothing at all about whether yelling is good for your child, and it would be dishonest to let one slide into the other.

What the evidence actually distinguishes is incident versus pattern. The longitudinal work on harsh verbal discipline (Wang & Kenny, Child Development) looked at a sustained diet of shouting, cursing, name-calling and insults, and found it predicted increases in adolescent conduct problems and depressive symptoms — and, notably, that parental warmth did not buffer it. That is a real finding about a real pattern.

It is not a finding about the Tuesday you shouted about shoes and apologised at breakfast. Those are different objects, and the difference between them is mostly what happens in the next thirty seconds. We go through this properly in does yelling at kids cause damage, including the places where the honest answer is less comforting than you want.

So: take the shame off. Leave the repair on. Those are separable, and the whole point of knowing the prevalence is that it makes the second one easier rather than optional.

What actually changes when you find out

It would be easy to treat "you're not alone" as a nice feeling with no practical content. It has practical content, and the mechanism is specific.

Guilt and shame are not intensities of the same thing. They are different emotions with different behavioural outputs, and the distinction is one of the more robust findings in the research on self-conscious emotion.

• Guilt is about the act. I did a bad thing. It points outward, at the person you did it to. Its natural output is confession, apology, reparation — going back in.
• Shame is about the self. I am a bad thing. It points inward, and it is much heavier. Its natural output is hiding, defensiveness, and withdrawal from the person who saw it.

This matters at 11pm, because a parent operating on shame is not a parent who is about to go and do a good repair in the morning. Shame's whole logic is that the problem is you, so exposure makes it worse and the safest move is to say nothing and hope it fades. Which means the child gets the yell and not the repair — and the repair is the part that carries most of the outcome.

Finding out that the behaviour is ordinary does one narrow, valuable thing: it strips the "I am uniquely defective" premise out from under the shame, and what is left is guilt, which is actionable. The standard doesn't move. The paralysis does.

There is a reason the sentence people report as the most useful thing anyone said to them is almost never advice. It is some version of oh god, me too.

How much is too much?

People want a number here. There isn't a validated one, and anyone offering you "more than twice a week" is making it up. But the question is fair, so here is a better frame than a count — four questions about shape:

1. Is it moving from volume into content? "I said that too loudly" is a different thing from "I told him what he is." Once it reaches name-calling, contempt, or characterising the child rather than the behaviour, you have crossed from a loud bad moment into the territory the harsh-verbal-discipline research is actually about.
2. Is it becoming the tool rather than the overflow? Yelling as something that happens to you when you're depleted is one pattern. Yelling as the routine method for getting shoes on is another, and it is the one that entrenches.
3. Is it going unrepaired? A yell that gets acknowledged is a rupture with a return. A yell that gets pretended away leaves the child holding an explanation, and the explanation children default to is that they caused it.
4. Is it frightening — to them, or to you? If your child has started flinching, or if you have scared yourself with how close you came to something physical, that is not a content-marketing question. Talk to a doctor or a therapist.

Three yeses mean the conditions need changing, not that you need to try harder — see why do I yell at my kids for what those conditions usually are. A yes to the fourth means get a person involved.

What to do with this tonight

If you arrived here at 11pm after a bad evening, the useful sequence is short.

1. Take the number seriously for thirty seconds. Not as an excuse — as a correction to a factual error you were carrying. You believed you were an outlier. You were not. That belief was doing damage and it was also just wrong.
2. Write one plain sentence about what happened. Not an analysis. One line. Rumination persists partly because your brain is keeping the event loaded on the theory that it is unresolved; writing it down gives it somewhere to sit.
3. Decide the words for the morning now. Deciding them tonight is what stops the guilt turning into a night of replay with no exit. Four lines is enough, and we have them written out by age.
4. Stop re-watching it. The moment already happened. Watching it again is not a repair, and it is not penance either. The 2am version of this goes into that properly.

And then the part that is genuinely hard: say it to one person. Not the softened version — the real one. Almost every parent who does this discovers the same thing within about four seconds, which is that the person they told has their own version, and has been keeping it in for exactly the same reason.