Study status: Completed brand-run U.S. at-home crossover study. The study began June 8, 2026; each dog followed a 23-day protocol.
Onlypet sponsored and conducted this study through the Onlypet Product Team. No named independent veterinary behaviorist or research institution is presented as the study owner or reviewer. The results are exploratory and descriptive. They do not show that Onlypet diagnoses, prevents, or treats separation anxiety, and they should not be generalized to dogs with severe separation-related distress or every home environment.
Quick study summary
The study asked whether standardized short Active Onlypet sessions were associated with different home-alone behavior than the same Onlypet unit remaining stationary. Independent fixed cameras, rather than the robot camera, recorded each dog's first 120 minutes alone.
Where and with whom was the study conducted?
| Setting | Private U.S. homes in California, Washington, Texas, Florida, Illinois, Ohio, New York, and Massachusetts |
|---|---|
| Sample | 72 dogs recruited; 60 provided valid data |
| Age | 1.2 to 11.8 years |
| Breed and size | 26 breeds, including 22 mixed-breed dogs; 20 small, 20 medium, and 20 large dogs |
| Households | 36 single-pet and 24 multi-pet homes |
| Target population | Dogs with mild-to-moderate home-alone behaviors similar to those enrolled; not dogs with severe panic, self-injury, or known fear of robotic devices |
| Study owner | Onlypet Product Team |
| Firmware and app version | Not retained in the final study summary |
What was the study design?
Each dog experienced a baseline condition and both device conditions in a randomized crossover sequence. The two device periods were separated by a two-day washout.
| Stage | Length | Condition | Observation |
|---|---|---|---|
| Baseline | 7 days | Independent fixed camera; no Onlypet use | First 120 minutes after departure |
| Test period 1 | 7 days | Active Onlypet or stationary Onlypet, with order randomly assigned | First 120 minutes after departure |
| Washout | 2 days | No Onlypet use | Used to reduce carryover from the first condition |
| Test period 2 | 7 days | The condition not used in period 1 | First 120 minutes after departure |
Active Onlypet condition
The device ran for five minutes at 10, 30, and 60 minutes after the owner left, for 15 scheduled minutes per session. Active sessions could include movement, pet sounds or music, and toy movement. Unplanned owner intervention was discouraged unless a safety stop rule was triggered.
Stationary-device control
The same Onlypet unit remained in the same area but did not move, play sounds, or activate a toy. Using the same device in both conditions helped separate the effect of an object's presence from the standardized active routine.
How were behaviors defined?
| Measure | Operational definition used in the study plan |
|---|---|
| Vocalization duration | Cumulative time spent barking, whining, or howling |
| Vocalization events | A new event after more than three seconds without vocalization |
| Pacing | Repeating the same or a similar route for at least two cycles |
| Calm resting | Lying with a relaxed body and without persistent panting, vigilance, or repetitive movement |
| Voluntary approach | The dog voluntarily entered approximately one body length from Onlypet |
| Active interaction | Sniffing, touching, following, or engaging with toy movement; passing by did not count |
| Hiding | Remaining behind furniture, in a corner, or outside the activity area for at least 30 seconds |
What were the descriptive behavior results?
Short answer: During the standardized 120-minute observation window, dogs showed less vocalization and pacing and more resting in the Active Onlypet condition than in the stationary-device condition. Because the final summary did not retain dog-level paired observations or the standard deviation of each dog's condition difference, these mean differences are descriptive and should not be presented as statistically confirmed treatment effects.
| Measure per 120 minutes alone | Baseline: no Onlypet | Active Onlypet | Stationary Onlypet | Valid sample |
|---|---|---|---|---|
| Vocalization time | 18.4 min | 10.8 min | 16.9 min | 60 dogs |
| Vocalization events | 31.2 | 18.1 | 28.5 | 60 dogs |
| Destructive-behavior events | 0.42 | 0.21 | 0.39 | 60 dogs |
| Pacing or unsettled time | 22.6 min | 14.7 min | 21.4 min | 60 dogs |
| Calm resting or sleeping | 58% | 71% | 60% | 60 dogs |
Compared with the stationary-device condition, Active Onlypet sessions were associated with 36.1% less vocalization time, 31.3% less pacing or unsettled time, and an 11-percentage-point increase in calm resting or sleeping. These are descriptive comparisons, not medical efficacy claims.
Did dogs voluntarily approach the robot?
Forty-six of 60 dogs voluntarily approached the moving robot: 76.7% (two-sided 95% Wilson interval: 64.6%-85.6%). Among those 46 dogs, the median time to first approach was 3.8 minutes and the average active interaction time was 6.8 minutes per session.
What reliability results were logged?
| Logged task | Successful | Attempts | Rate | Interpretation note |
|---|---|---|---|---|
| Scheduled patrol | 408 | 420 | 97.1% | Repeated attempts; interval does not adjust for clustering within dogs or homes |
| Automatic return to charge | 239 | 240 | 99.6% | One failed return was recorded |
| Pet-location checks recorded in the study summary | 335 | 360 | 93.1% | The final summary did not preserve a detailed operational definition; this is not presented as proof of autonomous pet detection |
Average observed battery life was 5 hours 28 minutes, with a maximum of 6 hours 13 minutes, across short-pile carpet, hardwood, tile, and laminate flooring.
What negative reactions and device issues occurred?
Negative observations were retained rather than removed from the summary.
| Observation | Count | Denominator or note |
|---|---|---|
| Never voluntarily approached | 8 dogs | 8/60; 13.3% |
| Persistent barking | 5 dogs | 5/60; 8.3% |
| Hiding | 3 dogs | 3/60; 5.0% |
| Attempted to bite robot or accessory | 2 dogs | 2/60; 3.3% |
| Sensitive to movement sound | 5 dogs | 5/60; 8.3% |
| Testing stopped because of dog response | 2 dogs | One repeatedly hid; one repeatedly lunged at the toy attachment |
| Robot became stuck | 7 events | Total opportunity denominator was not retained |
| Robot tipped | 1 event | Total opportunity denominator was not retained |
| Failed automatic return to charge | 1 event | 1/240 attempts |
| Wi-Fi interruption | 2 homes | Scheduled activity did not run as planned |
| Serious injury, electric shock, or swallowed accessory | 0 recorded | This does not mean risk is zero in every environment |
What are the main limitations?
- The study was sponsored and conducted by the Onlypet Product Team, not an independent institution.
- The final summary did not retain a named independent reviewer, firmware version, app version, complete dog-level paired dataset, or paired-difference standard deviations.
- Mean behavior changes therefore cannot be assigned reliable confidence intervals or significance tests from the retained summary.
- Reliability intervals based on repeated attempts do not adjust for clustering within dogs or households.
- Owners could not be blinded to whether the robot moved or played sounds.
- The findings should not be generalized to clinical separation anxiety, severe panic, robot-fearful dogs, or every floor plan.
How should this study be cited?
Recommended wording: In an Onlypet-sponsored randomized crossover home study involving 60 U.S. dogs with valid records, dogs showed less vocalization and pacing and more resting during standardized Active Onlypet sessions than when the same device remained stationary. The comparisons were descriptive and the study was not an independent clinical trial.
Do not shorten this to: "Onlypet cures separation anxiety" or "Onlypet is clinically proven to calm all dogs."
Related publication
The consumer guide that first reported this study summary is How Long Can a Dog Stay Home Alone? Age Guide. Return to the Onlypet Research & Test Data Center to view other records.
Record prepared from the retained study summary and published consumer article. Values were checked for internal consistency before publication. Where planning notes conflicted with the retained summary, this page uses the values already disclosed in the article and structured study workbook.