# How Common Is Separation Anxiety in Dogs? 5 Studies Compared

By Lopaw · Published: 2026-10-03

Separation anxiety estimates in dogs run from 2% to 86%. We compared five studies by sample, questionnaire and cut-off to show what each number counts.

Search for how common separation anxiety is in dogs and you will get figures like 20 to 40% and 85.9%, usually on the same results page and mostly without a primary source. Most of them answer different questions. We went back to five studies, read how each one counted, and put their methods side by side. For what separation anxiety is and how it is treated, [start with our complete guide to dog separation anxiety](/blog/dog-separation-anxiety).

> **Key Takeaways**
> - With a strict frequency bar, separation-related behavior affects about **2 to 6% of dogs** (Salonen 2020, Serpell & Powell 2026). At the looser "sometimes" bar, the US figure is 9.3%.
> - Studies that count **any single sign** land far higher: 22.1% of 1,807 UK dogs before lockdown (Harvey 2022) and 46.9% of a 145-puppy case/control subsample at six months (Dale 2024).
> - The widely quoted **85.9%** and the **9.3%** come from the same 43,517 US dogs. The difference comes from merging separation items with attachment and attention-seeking, which most dogs show.
> - Clinic figures describe caseloads: at Cornell's behavior clinic, separation anxiety made up 49.2% of anxiety, fear and phobia cases (Anderson 2022).
> - For your own dog, the useful question is whether the signs are frequent and start as you leave.

---

## The short answer: 2 to 6% at a strict bar, much higher if any sign counts

In general-population studies with a strict frequency bar, about 2 to 6% of dogs show separation-related behavior. That is 5 to 6% in Finland, where Salonen counted signs shown "often" or more, and 2.3% in the US at "usually" or "always". At the looser "sometimes" bar, the US figure rises to 9.3%. When a single occasional sign is enough, estimates reach 22% and more. The highest widely quoted figure, 85.9%, dropped to 9.3% when the same dataset was re-analyzed.

There is no single number because there is no single definition. Two terms get mixed up. **Separation anxiety** is a clinical diagnosis, made by a vet or behaviorist who has ruled out other causes. **Separation-related behavior** (SRB) is what the studies below measure: signs reported by owners in a questionnaire, such as barking, destruction or house soiling when the dog is alone. Every population figure in this post is SRB, and none comes from diagnosed dogs.

## The five studies side by side

Three things decide the number almost entirely: who was asked, which questionnaire they filled in, and what level of severity counted as a case.

| Study | Dogs and recruitment | Instrument | What counted | Figure |
|---|---|---|---|---|
| **Salonen 2020**, Finland | 13,715 dogs, 10 weeks to 17 years, self-selected online owner survey | Custom questionnaire, 0 to 4 frequency scale | Signs "often" or "very often" | **5 to 6%** |
| **Dog Aging Project**, US (Beaver 2024, Serpell & Powell 2026) | 43,517 dogs, owners enrolled in a national research cohort | Mini C-BARQ | Score of 2+ of 4 ("sometimes" or more). The original merged separation with attachment items | **85.9%** original, **9.3%** re-analyzed |
| **Harvey 2022**, UK | 1,807 dogs, online cohort recruited during the first lockdown | Owner survey on SRB signs | At least one sign | **22.1%** in February 2020 |
| **Dale 2024**, UK/IE (Generation Pup) | 145 puppies analyzed from a cohort of 1,987 | Owner survey at six months | At least one sign, case/control design | **46.9%** (68 of 145) |
| **Anderson 2022**, Cornell clinic | 1,800 dogs with a primary behavior problem, 1997 to 2017 | Clinical records | Diagnosis by the clinic | **49.2%** of anxiety cases |

**Salonen 2020.** A large Finnish dataset with a strict cut-off: only signs reported often or very often. The results section gives 5%, the discussion 6%, so both figures are quoted in good faith ([Salonen et al., 2020](https://pmc.ncbi.nlm.nih.gov/articles/PMC7058607/)).

**Dog Aging Project.** A national volunteer cohort, analyzed twice with different groupings of questions (see the next section).

**Harvey 2022.** A cohort followed through the pandemic. The pre-lockdown figure was recalled by owners in a survey run between May and July 2020, which makes it softer than the October figure ([Harvey et al., 2022](https://pmc.ncbi.nlm.nih.gov/articles/PMC8868415/)).

**Dale 2024.** A puppy cohort. The authors selected 145 of its dogs for a case/control comparison, so the 46.9% describes that subsample ([Dale et al., 2024](https://pmc.ncbi.nlm.nih.gov/articles/PMC11655275/)).

**Anderson 2022.** Twenty years of cases at a university behavior clinic. It tells you what brings dogs to a specialist ([Anderson et al., 2022](https://pmc.ncbi.nlm.nih.gov/articles/PMC8909650/)).

## Same 43,517 dogs, 85.9% or 9.3%

Both figures were calculated on the same Dog Aging Project data: 43,517 US dogs whose owners completed the mini C-BARQ. Both used the same threshold. The main difference is which questions were counted together.

The 85.9% comes from Beaver (2024) in the *Journal of Veterinary Behavior*. It was publicized through a Texas A&M press release titled "More than 99% of US dogs have a behavior problem", which listed "separation and attachment behaviors (85.9%)" among the top categories ([Texas A&M, 2025](https://stories.tamu.edu/news/2025/04/01/more-than-99-of-us-dogs-have-a-behavior-problem-texas-am-researcher-finds/)).

The mini C-BARQ scores separation-related behavior and attachment/attention-seeking as two separate factors, and the original analysis merged them. Attachment behavior, such as following you from room to room or asking for attention, is something most dogs show: on its own, 88.9% of the sample scored in the "moderate to serious" range (score 2 or more).

Separation-related behavior scored on its own factor, at the same threshold, came to 9.3%. In Serpell and Powell's words, grouping both sets of items together "has the effect of greatly inflating the prevalence of so-called separation anxiety" ([Serpell & Powell, 2026](https://doi.org/10.1016/j.jveb.2025.11.005)).

The threshold is the other lever. Both figures count a score of 2 or more on a 0 to 4 scale, which for separation items means the behavior happens "sometimes". Raise the bar to "usually" or "always" and the share of dogs with separation-related behavior falls to 2.3%. Serpell and Powell add one more caveat: the Dog Aging Project recruits volunteers, who may be more interested in dog behavior than average owners, so even these figures could run high. So the most quoted number on the subject comes from how the items were grouped.

## Why "any sign" studies land at 20 to 50%

If one episode of barking at the door counts as separation-related behavior, the share of dogs with it goes up several times over. These studies answer a different question: how many dogs show anything at all.

Harvey and colleagues found that 22.1% (400) of the 1,807 dogs in their UK cohort showed at least one sign in February 2020, according to their owners ([Harvey et al., 2022](https://pmc.ncbi.nlm.nih.gov/articles/PMC8868415/)). One sign might be a few minutes of whining, a chewed doorframe once, or an accident on the mat.

Dale and colleagues found that 46.9% of puppies (68 of 145) showed separation-related behavior at six months, again on an any-sign definition ([Dale et al., 2024](https://pmc.ncbi.nlm.nih.gov/articles/PMC11655275/)). These 145 puppies were selected from a cohort of 1,987 for a case/control analysis. So the figure describes that subsample, and "nearly half of all puppies" overstates it.

For owners, the lesson is the same: one sign on its own proves little, as our guide to [why a dog barks when you leave](/blog/why-does-my-dog-bark-when-i-leave) explains. Frequency, timing and whether the dog can eat tell you far more than the presence of one behavior.

## Clinic numbers are not population numbers

The share of separation anxiety among patients of a behavior clinic describes what people bring to a specialist. It does not tell you how many dogs have the problem. Dogs without problems never enter the sample.

Anderson and colleagues reviewed Cornell behavior clinic cases from 1997 to 2017. Separation anxiety accounted for 178 cases, 49.2% of all anxiety, fear and phobia diagnoses ([Anderson et al., 2022](https://pmc.ncbi.nlm.nih.gov/articles/PMC8909650/)). By our calculation, that is about one in ten of the 1,800 dogs with a primary behavior problem.

Neither number applies to dogs in general. The first is a share of anxious patients, the second a share of all patients, and every one of those dogs was referred because something was wrong. The "20 to 40% of dogs" figure that circulates without a source belongs here too: Harvey and colleagues attribute it to dogs "presented to behaviour clinics" ([Harvey et al., 2022](https://pmc.ncbi.nlm.nih.gov/articles/PMC8868415/)). No general-population study with a frequency threshold supports it. Only any-sign counts, like Harvey's 22.1%, reach that range.

## Did lockdown change it?

In Harvey's UK cohort, 117 of the 1,187 dogs that had no signs in February and were being left alone in October 2020 (9.9%) had developed new signs of separation-related behavior since lockdown began. Risk was uneven. The authors report "dogs whose leaving hours decreased most during lockdown restrictions being at increased risk of developing new signs of SRB" ([Harvey et al., 2022](https://pmc.ncbi.nlm.nih.gov/articles/PMC8868415/)). Living alone or with another dog made no difference.

The same study gives two headline shares that look like a trend: 22.1% in February 2020 (400 of 1,807) and 17.2% in October 2020 (260 of 1,510). The denominators differ, and the February answers were recalled months later, so the two cannot be compared directly.

The data supports one narrower point: dogs whose time alone dropped sharply were the ones most likely to develop new signs. If your routine has changed the same way, [building time alone back up gradually](/blog/how-to-train-a-dog-to-be-alone) works better than going straight back to full days. If the change is a return to work, follow [a six-week plan for going back to the office](/blog/dog-return-to-office).

## Which dogs are more likely to have it

Risk-factor evidence is weaker than the evidence on prevalence. Each effect's direction is reasonably clear, but its size is uncertain, so treat these as tendencies.

**Sex, age and breed.** In the Finnish data, separation-related behavior was slightly more common in males. Destruction and urination when alone were more common in younger dogs. Mixed breeds and Wheaten Terriers had the highest rates among the breeds studied ([Salonen et al., 2020](https://pmc.ncbi.nlm.nih.gov/articles/PMC7058607/)). The same study found that anxiety traits tend to cluster, so a noise-sensitive or fearful dog is more likely to struggle alone too.

**Age, in the other direction.** In Harvey's cohort, where the oldest dog was just over nine, each extra year of age raised the odds of developing new signs by October 2020. The link did not hold once other factors were accounted for ([Harvey et al., 2022](https://pmc.ncbi.nlm.nih.gov/articles/PMC8868415/)). Young dogs show more destruction. Older dogs may cope less well with a change in routine, and a sudden change in an older dog is a reason to see a vet first.

**Early life.** In the Generation Pup data, puppies that were restricted to a crate or room overnight by 16 weeks, and puppies that slept at least nine hours a night at that age, had lower risk at six months. Higher risk went with owners fussing over the dog on return in response to "bad" behavior and with aversive or time-out training methods. Poor house-training and owners aged 16 to 34 were also linked to higher risk ([Dale et al., 2024](https://pmc.ncbi.nlm.nih.gov/articles/PMC11655275/)). These are associations, and no single habit has been shown to cause the problem. They are still a good argument for getting [a rescue dog's first week alone](/blog/rescue-dog-first-week-alone) right.

## What the number means for your dog

A percentage cannot diagnose anything. Either your dog shows frequent signs that start right after you leave, or it does not, and one evening with a camera will usually tell you which.

[Film one absence and read the footage](/blog/dog-separation-anxiety-camera): when the first stress signal appears, whether the food gets eaten, and where any damage is. If the noise starts, peaks and fades within the first ten minutes, read [why ten minutes of howling is usually not anxiety](/blog/dog-howls-when-left-alone-then-stops) before assuming the worst. If the signs start within a minute or two and keep going, you have a threshold to work from. That number matters more than any study's.

Lopaw is an app for running that training: it records each session and works out the next step from your dog's threshold.

## FAQ

**What percentage of dogs have separation anxiety?**
About 2 to 6% at a strict frequency bar (Salonen 2020, Serpell & Powell 2026). At the looser "sometimes" bar, the US figure is 9.3%, and definitions that count any single sign give 22% or more.

**Why do separation anxiety statistics vary so much?**
Three reasons: who was surveyed, which questionnaire was used, and what severity counted as a case. The clearest example is the Dog Aging Project, where the same 43,517 dogs produced 85.9% when separation and attachment questions were merged and 9.3% when separation-related behavior was scored on its own.

**Is separation anxiety more common in puppies?**
Signs are common in puppies: 46.9% of a 145-puppy case/control subsample showed at least one at six months (Dale 2024), though that figure describes a selected subsample. In the Finnish data, destruction and urination when alone were more common in younger dogs. For [whether puppies grow out of it](/blog/can-dogs-grow-out-of-separation-anxiety), see the follow-up studies, and for [how to tell normal puppy crying from a real problem](/blog/puppy-separation-anxiety), see our puppy guide.

**Did separation anxiety increase after lockdown?**
In one UK cohort, 9.9% of dogs that had no signs in February and were being left alone in October 2020 (117 of 1,187) had developed new signs since lockdown began (Harvey 2022). Risk was highest where the dog's time alone had dropped the most.

## Sources

- Anderson K.H. et al., [Case Distribution, Sources, and Breeds of Dogs Presenting to a Veterinary Behavior Clinic in the United States from 1997 to 2017](https://pmc.ncbi.nlm.nih.gov/articles/PMC8909650/), *Animals* 12(5):576, 2022
- Beaver B.V., [The prevalence of behavior problems in dogs in the United States](https://doi.org/10.1016/j.jveb.2024.11.001), *Journal of Veterinary Behavior* 76:34-39, 2024
- Dale F.C. et al., [Canine separation-related behaviour at six months of age: Dog, owner and early-life risk factors identified using the 'Generation Pup' longitudinal study](https://pmc.ncbi.nlm.nih.gov/articles/PMC11655275/), *Animal Welfare* 33:e60, 2024
- Harvey N.D. et al., [Impact of Changes in Time Left Alone on Separation-Related Behaviour in UK Pet Dogs](https://pmc.ncbi.nlm.nih.gov/articles/PMC8868415/), *Animals* 12(4):482, 2022
- Salonen M. et al., [Prevalence, comorbidity, and breed differences in canine anxiety in 13,700 Finnish pet dogs](https://pmc.ncbi.nlm.nih.gov/articles/PMC7058607/), *Scientific Reports* 10:2962, 2020
- Serpell J.A., Powell L.R., [Prevalence and severity of behavior problems in dogs in the United States: A re-assessment](https://doi.org/10.1016/j.jveb.2025.11.005), *Journal of Veterinary Behavior* 83:8-13, 2026
- Texas A&M University, [More than 99% of US dogs have a behavior problem, Texas A&M researcher finds](https://stories.tamu.edu/news/2025/04/01/more-than-99-of-us-dogs-have-a-behavior-problem-texas-am-researcher-finds/), 2025

[Download Lopaw on the App Store](https://apps.apple.com/app/apple-store/id6793935429?pt=128805461&ct=blog_md&mt=8) · [Get Lopaw on Google Play](https://play.google.com/store/apps/details?id=com.nowaffl.lopaw&referrer=utm_source%3Dblog_md%26utm_medium%3Dlanding)