What the research records
Effective control.
Lasting learning.
Different animals, tasks and protocols provide different kinds of evidence. Read the result with the conditions tested.
01 / Andelt · 1999
Experienced predators.
Attacks interrupted.
Five captive coyotes had each killed a lamb in preliminary testing. Across repeated tests over 22 weeks, collar stimulation averted all 13 subsequent attempted attacks. Later attempts became less likely and avoidance increased.
What it supports: strong predatory motivation need not make the intervention ineffective. Coyotes are not domestic dogs, and there was no untreated control. Thirteen attack episodes is not a stimulation count.
[3]
02 / Johnson & Wynne · 2024
A clear difference
in the first test.
All six analysed collar dogs passed the first post-training lure test; none of eleven dogs in the two food-based groups did. Most stimulation occurred in the first session and declined sharply by the second.
Two of eight assigned collar dogs were excluded after exceeding the stimulation limit. Counting both as failures gives six successes out of eight, versus zero of eleven. This small, short experiment compared specific protocols, not every training method or real-world situation.
[4]
03 / Christiansen and colleagues · 2001
Learned sheep avoidance.
One year later.
Among 114 hunting dogs reassessed about a year later, only one of the 17 dogs that had received e-collar training in the first year attacked sheep at follow-up.
Dogs were treated because they approached sheep, rather than randomly allocated. The follow-up cohort included untreated dogs. The finding supports retention in this selected group, not a universal success rate.
[5]
04 / Dale and colleagues · 2013
Retention without
wearing the device.
All 65 dogs showed avoidance after kiwi-stimulus training. All 15 dogs in a subset retested one month later retained avoidance without wearing an electronic collar.
The targets were training materials, not live birds. The result shows that useful learned avoidance need not disappear when the collar is removed; it does not measure prevention of wild-bird deaths.
[6]
05 / Tortora · 1983
A structured programme.
Specified aggression eliminated.
Safety training eliminated the specified avoidance-motivated aggression in all 36 dogs tested. It combined alternative responses, electrical stimulation, warning and safety signals, and transfer to owners. Dogs also met criteria without the collar.
Follow-up was primarily owner-reported: all 36 dogs through one year, 18 at two years and eight at three. The finding concerns this diagnosis and complete programme, not indiscriminate punishment or every form of aggression.
[7][8]
06 / Salgirli and colleagues · 2012
Control under
strong distraction.
In a crossover study of 42 police dogs, 39 showed the defined learning response in the electronic-collar condition: maintaining heel despite a provoking decoy.
These were same-day tests involving mostly experienced dogs. The study demonstrates control in that task, not a comprehensive comparison against reward-based training.
[9]
Tortora’s comparison and later appraisalA separate experiment randomly allocated 18 dogs to three groups. The complete programme outperformed play training alone and play plus direct punishment; comparison groups subsequently improved with the complete programme. A 2022 systematic review rated Tortora high quality and recognised its marked reduction in aggression. That appraisal concerns this study, not a general endorsement of collars. [7][8]
