How Common Is Social Anxiety in Teenagers?
Sources last read 2026-08-19.
The short answer
NIMH estimates that 9.1 percent of adolescents had social anxiety disorder, and that figure only means anything with its context attached. It is a lifetime estimate, so it says 9.1 percent met the criteria at some point by 18, not that 9.1 percent have it now. The criteria used were DSM-IV rather than DSM-5, and the survey behind it was fielded from February 2001 to January 2004, which makes the data roughly twenty-two years old. Printed in the same sentence by NIMH, and almost nowhere else, is the number that bears on a decision: an estimated 1.3 percent had severe impairment.
The figure everyone quotes, and the four things stripped off it
NIMH's statistics page carries one sentence that almost every article in this subject is built on. Read it with the four caveats set out directly underneath, because they are part of it: an estimated 9.1 percent of adolescents had social anxiety disorder, and an estimated 1.3 percent had severe impairment.
Four pieces of context travel with that first number, and they are usually removed:
- It is lifetime, not current. It means 9.1 percent met the criteria at some point by age 18. It does not mean 9.1 percent have social anxiety disorder right now.
- The criteria were DSM-IV, not DSM-5. The manual changed after this was measured.
- The survey was fielded from February 2001 to January 2004. That is the National Comorbidity Survey Adolescent Supplement.
- So the data are about twenty-two years old. Not wrong, but old, and old in a way that matters for a subject where the definition itself moved.
None of that makes the figure useless. It makes it a different figure from the one most pages print. Nine percent of teenagers walking around with a disorder right now is a frightening sentence. Nine percent meeting a set of older criteria at any point across the whole of childhood and adolescence is a much calmer one, and it is the one the source supports.
The 1.3 percent is the number to hold onto
Sitting in the same NIMH sentence is the severe impairment estimate: 1.3 percent.
Set the two side by side and roughly one in seven of the adolescents who met the lifetime criteria had severe impairment. The other six met the criteria without being severely held back by them.
That ratio is the most decision-relevant thing in this subject, and it is close to absent from the first page of search results. It is what stands between a parent and the assumption that meeting a threshold and being in trouble are the same event. They are not.
How the estimate splits
All of the numbers below come from the same lifetime estimate, under DSM-IV criteria, from that 2001 to 2004 fieldwork. Read them with the same four caveats attached.
By sex: 11.2 percent for females, 7.0 percent for males.
By age band: 7.7 percent at 13 to 14, 9.7 percent at 15 to 16, 10.1 percent at 17 to 18.
The age pattern is what you would expect from a lifetime estimate. Older teenagers have had more years in which to have met the criteria at some point, so the figure climbs. It is not evidence that seventeen year olds are having a harder time than fourteen year olds this week.
What a prevalence figure is actually for
It answers one question well: is this the kind of thing that happens to a lot of families. Yes, on any reading of the data.
It answers no other question at all. It cannot tell you whether what you are seeing in your house counts, whether it will last, or whether you should call anyone. Parents arrive at these numbers looking for a verdict, and the numbers do not contain one.
A worked example, for illustration
Brandon's father read the same figure on four different pages in one evening. On three of them it appeared bare, with nothing around it. He came away thinking it described how many teenagers have this right now.
The fourth page carried the severe impairment estimate next to it. That was the first thing he had read all night that felt like it was about a decision rather than about a category.
What he stopped doing after that was trying to place Brandon inside a percentage. What he started doing instead was writing down which situations Brandon had backed out of, and when. A number about a national sample was never going to answer a question about one seventeen year old.
What this page cannot tell you
Whether your teenager is in any of these percentages. Prevalence runs one direction only, from population to background, and never from background to a person.
If you came here to find out whether what you are seeing is ordinary, the more useful pages in this section are the one on where the line between shyness and the disorder sits and the one on what avoidance looks like from the outside.
Common questions
So about one teenager in eleven in my son's class has this?
No, and that step is where the figure usually goes wrong. It is a lifetime estimate from a national survey of adolescents, and it counts anyone who met the criteria at any point by 18. It is not a count of who meets them today, it was not measured in your district, and the interviews behind it were done between February 2001 and January 2004 under DSM-IV criteria. A classroom is not a sample and the arithmetic does not carry across.
Why does the severe impairment number matter more?
Because it is closer to the question you are actually asking. NIMH's estimate is that 1.3 percent of adolescents had severe impairment, against the 9.1 percent lifetime figure under DSM-IV criteria from that same 2001 to 2004 survey, now about twenty-two years old. It counts anyone who met the criteria at some point by 18, not who meets them today. That works out at roughly one in seven of those who met the criteria at some point. Meeting criteria and being badly held back by them are not the same thing, and most of the pages quoting the larger number never mention the smaller one.
Is it more common in girls?
In that survey it was. The lifetime estimate was 11.2 percent for females and 7.0 percent for males, on the same DSM-IV criteria from the same 2001 to 2004 fieldwork. That is a difference in a twenty-two year old national estimate, not a statement about your teenager.
Why is this site quoting data from more than twenty years ago?
Because it is the figure NIMH publishes on its own statistics page, and because the age of a number is part of the number. Any page that prints it without saying when the interviews happened is handing you less than it has. We would rather give you the figure and its birthday together.
If it is common, does that mean I should stop worrying?
A prevalence figure tells you about a population and nothing at all about one person. It can stop you feeling like the only house on the street with this going on, which is worth something. It cannot tell you whether your teenager needs anything, and no honest page will let it try.
Checked against
NIMH statistics, social anxiety disorder
nimh.nih.gov
NIMH science update, national survey on social phobia and shyness
nimh.nih.gov
NIMH, Social Anxiety Disorder: More Than Just Shyness
nimh.nih.gov