A 2025 analysis of 47,444 dogs found that the year a dog entered the study was most strongly associated with its reported Trainability score: dogs enrolled from 2021 through 2023 scored lower on average than dogs enrolled in 2020. The study does not identify one cause, and it cannot explain why any particular dog behaves as it does.
The findings are a large population snapshot, not a diagnosis. They show how owner-reported behavior differed across groups, while leaving room for individual history, health, environment and learning. The primary paper is available in PLOS ONE.
As an Amazon Associate I earn from qualifying purchases.
What the study actually measured
Li, Sexton and colleagues analyzed complete Health and Life Experience Survey records from 47,444 dogs enrolled in the U.S. Dog Aging Project between 2020 and 2023. The analysis used the project’s 2023 curated data release, so the enrollment year refers to when each dog joined the project, not necessarily when a behavior was first observed.
Recommended Free Tools
Owners answered a shortened version of the Canine Behavioral Assessment and Research Questionnaire (C-BARQ). It contains approximately 40 behavior items, compared with about 100 in the full questionnaire. The authors report that the shortened form had been validated against the longer version at the domain level, but the answers still represent an owner’s observations and interpretation.
#1 Best Overall
Rather than treat every questionnaire item as a separate outcome, the researchers used principal component analysis to group correlated measures. Four composite behavior domains became the outcomes in multivariable regression models:
| Composite | What it represents in this analysis |
|---|---|
| Fear | Reported fear-related behavior patterns. |
| Attention/Excitability | Reported attention and arousal or excitability patterns. |
| Aggression | Reported aggressive behavior patterns. |
| Trainability | Reported ease of learning and responding to training-related cues. |
The four retained components explained 50% of the variability in the measured behavior scores. That is a substantial summary of the questionnaire, but it also means half of the variation was not explained by those four components. The study found no statistically significant or meaningful interactions that were retained in the final models.
The clearest result: later cohorts had lower Trainability scores
Enrollment year had the strongest influence on the Trainability composite. Compared with dogs enrolled in 2020, dogs entering in 2021, 2022 and 2023 had lower average reported Trainability scores. The difference between the 2020 and 2023 groups was the smallest of the year-to-year comparisons reported by the authors; the important point is that all three later cohorts were lower than the 2020 reference group.
This is a cohort association, not evidence that a calendar year changed a dog’s ability to learn. The analysis did not follow one group of dogs through an identical training program, and it did not measure every event that could have affected a household. A dog enrolled in 2022 was not necessarily the same age, living situation or training stage as a dog enrolled in 2020.
What did not move in the same way
The year pattern was not a general shift across every behavior domain. The authors found no significant change in the composite Fear score across the 2020–2023 enrollment years. Aggression was lower for dogs enrolled in 2023 than for those enrolled in 2020, but the results did not show a steady, year-by-year decline.
| Behavior domain | Enrollment-year finding | How to read it |
|---|---|---|
| Trainability | 2021, 2022 and 2023 cohorts had lower averages than 2020. | The strongest year association in the models; no single cause was established. |
| Fear | No significant 2020–2023 change. | The study does not support a broad claim that fear rose or fell by cohort. |
| Aggression | 2023 was lower than 2020, without a consistent annual trend. | A two-cohort difference is not the same as a progressive pattern. |
Why the authors discuss the pandemic—and why that is not a proven explanation
The researchers suggest that pandemic-era household changes could have made training more difficult for some dogs. People acquired dogs, work schedules and daily routines shifted, and opportunities for classes, socialization or other training may have changed. These are plausible explanations for a cohort difference, and the study’s authors raise them as possibilities.
However, the dataset cannot determine which of those factors mattered, how much they mattered or whether they explain the Trainability association at all. The analysis did not isolate pandemic exposure as an experimental variable. It also cannot separate the effect of a household routine from the dog’s age at enrollment, previous training, environment or other unmeasured circumstances.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Fear, stress, early trauma, household routines and environmental exposures can influence behavior in other settings, but this particular comparison did not establish any of them as the cause of the enrollment-year result. Treat them as context for interpreting the finding, not as a conclusion about your dog.
Which dog characteristics were associated with the scores?
Across one or more behavior domains, reported scores varied with breed type, life stage, sex, spay/neuter status, body size, U.S. region and primary activity. These are statistical differences between groups after adjustment in the study’s models. They are not rules that describe every member of a breed, sex or size category.
Size differences
Smaller dogs had higher average Fear and Aggression scores and lower average Trainability scores than larger dogs in this analysis. “Higher” and “lower” refer to group averages on the composites. They do not mean that a small dog is destined to be fearful or aggressive, or that a large dog is automatically easy to train.
Rank #3
Life stage, sex and reproductive status
The models included life stage, sex and whether a dog was spayed or neutered because these recorded characteristics were associated with one or more reported domains. An association cannot show that reproductive status or age caused a behavior score, because those characteristics travel with many other differences in care, health, history and living conditions.
Free tools Windows power users keep installed
One-click scans. No signup required.
Breed type, region and primary activity
Breed type, U.S. region and the household’s stated primary activity also appeared among the characteristics linked with reported behavior. Breed labels can combine genetically and behaviorally diverse dogs, and regional or activity categories can reflect differences in housing, routines, access to training and owner expectations. The study is therefore better for describing broad patterns than for predicting an individual dog’s response.
Who was represented in the 47,444-dog sample?
The project is a U.S. longitudinal cohort, not a census of all dogs. In the rounded breakdown reported by the paper, 55% of dogs were mature adults, 5% were puppies, 21% were young adults and 18% were seniors. Those displayed percentages total 99% because of rounding.
| Life-stage group | Share reported in the paper |
|---|---|
| Mature adults | 55% |
| Puppies | 5% |
| Young adults | 21% |
| Seniors | 18% |
Because participation required enrollment in the Dog Aging Project and completion of the relevant survey, the sample may differ from the wider U.S. dog population. A very large sample improves the precision and range of a description; it does not automatically make the participants representative of every household or remove all sources of bias.
How much confidence should you place in owner-reported behavior?
C-BARQ is useful for collecting comparable observations from many households, but it is not an objective sensor attached to the dog. The authors note inherent subjectivity in owner responses. People can differ in what they notice, what they consider “aggressive” or “fearful,” and how they interpret the same event.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Repair Windows errors before they cause bigger problemsFix Now →Rank #4
- Index supreme
- Interview by the author appendix
- #1 NEW YORK TIMES BEST SELLER.
The paper also discusses possible confounding and bias linked to an owner’s circumstances, experience with a particular dog and consistency in interpreting behavior. For example, an owner who has lived with several dogs may use a different internal baseline from a first-time owner. A household under unusual stress may also observe or report behavior differently.
These limits do not make the results useless. They define what the numbers can support: population-level associations in a very large owner-reported dataset. They do not support a diagnosis, a breed label or a causal explanation for an individual pet.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What this means if your dog behaves differently
Do not use a group average as a prediction
Your dog’s size, breed type, sex, age or enrollment-era equivalent cannot determine its score from this paper. Individual dogs vary widely within every category, and the composites summarize reported tendencies rather than measuring a fixed trait.
Look for a pattern rather than one isolated incident
Repeated observations can be more informative than a single frustrating walk or training session. Note the setting, trigger, distance, people or animals present, what happened immediately before the behavior and how quickly your dog recovered. Consistent notes help you and a qualified professional distinguish a recurring pattern from a one-off event.
Treat sudden change as an individual problem
A sudden increase in fear, aggression, confusion or difficulty responding is outside what this cohort comparison can explain. The study cannot rule in or rule out pain, illness, medication effects, sensory change, injury or a stressful event in a particular dog. Arrange a veterinary assessment for an abrupt or concerning change, especially when behavior is accompanied by physical symptoms or a safety risk.
Best Value
Use training history and environment as working clues, not assumptions
Changes in schedule, exercise, access to classes, social contact and household members may affect learning opportunities. They are reasonable subjects to review with a veterinarian or qualified behavior professional, but the paper does not tell you that any one of them caused your dog’s behavior. Avoid punishment-based responses when fear or aggression is involved; prioritize safety and seek individualized guidance.
What the study can and cannot explain
| Question | What the evidence supports | What it does not support |
|---|---|---|
| Did behavior differ by enrollment year? | Yes. Later cohorts had lower average Trainability than the 2020 cohort; Fear did not significantly change, and 2023 Aggression was lower than 2020 without a steady trend. | That a specific year caused a specific dog’s behavior. |
| Do dog characteristics matter? | Breed type, life stage, sex, reproductive status, size, region and primary activity were associated with one or more domains. | That a characteristic determines an individual dog’s temperament. |
| Can the survey measure behavior at scale? | Yes. A shortened C-BARQ, validated at the domain level, supplied comparable owner reports for 47,444 dogs. | That owner ratings are free from subjectivity or confounding. |
| Is there one explanation for the cohort difference? | No single explanation was established; pandemic-related routine and training changes are possibilities raised by the authors. | A claim that the pandemic, breed, size or any other factor caused the result. |
Why this baseline matters for future dog-aging research
The value of the project is not a new temperament label for your dog. It is the unusually large starting point for tracking how behavior and health relate over time. The study’s co-author Courtney Sexton described the goal as having “a starting point” for following behavior changes in tens of thousands of dogs as they age and studying how behavior and health are linked; that statement appears in the Virginia Tech announcement carried by EurekAlert!.
Future follow-up can test whether the same dogs’ behavior composites change with age, illness, lifestyle or other measured experiences. Longitudinal observations may also help separate a temporary household change from a persistent behavioral pattern. Until those analyses are available, the present results should be read as a baseline description of associations, not a finished theory of canine behavior.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
The practical answer to “what really explains the way yours behaves?”
This study points to several layers rather than one answer. Dogs in different enrollment cohorts reported different average Trainability scores, and characteristics such as size, life stage, sex, reproductive status, breed type, region and activity were associated with some domains. Yet the data cannot identify the specific driver of an individual dog’s behavior, and owner-reported measures carry unavoidable subjectivity.
For your dog, the useful next step is individual assessment: observe behavior repeatedly, record context, consider health and environmental changes, and obtain veterinary or qualified behavior help when the change is sudden, severe or unsafe. The 47,444-dog analysis can provide perspective; it cannot replace that case-by-case evaluation.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
Leave a Reply