What each approach actually does
Positive reinforcement adds something the dog wants (a treat, praise, play) immediately after a desired behaviour, increasing the likelihood it happens again. Aversive tools work through punishment or negative reinforcement, adding something unpleasant (a static correction, a spray, a leash jerk) to reduce an unwanted behaviour, or removing pressure only once the dog complies.
Both approaches can produce behaviour change. The meaningful differences are in speed, side effects, and what happens to the relationship between dog and owner along the way.
What the research generally shows
A growing body of studies comparing training methods has found that dogs trained with aversive methods show more stress-related behaviours, lip licking, yawning, lower body posture, cowering, both during training sessions and, in some studies, at other times in daily life, compared to dogs trained primarily with reward-based methods. Some research has also linked aversive-method use to a higher reported incidence of aggression toward the owner or unfamiliar people, particularly with tools like shock collars.
Reward-based training, by comparison, is generally associated with dogs that appear more relaxed during training and show fewer problem behaviours reported by owners overall. It's worth noting the evidence isn't unanimous or perfect, much of it relies on owner-reported outcomes and observational studies rather than tightly controlled trials, but the overall pattern consistently favours reward-based methods on welfare grounds.
Why aversive tools can feel like they "work" faster
Punishment can produce fast, visible suppression of a behaviour, a dog stops pulling the instant a prong collar bites, or stops barking the instant a spray hits. That immediate feedback loop is part of why aversive tools remain popular; the change is dramatic and instant, in a way training an alternative behaviour with rewards simply isn't.
The trade-off is that suppression doesn't teach the dog what to do instead, and it doesn't address why the behaviour was happening. A dog that stops pulling because of a prong correction hasn't learned loose-lead walking is rewarding, they've learned pulling hurts, which is a narrower and more fragile lesson, and one that can generalise into anxiety around walks generally if applied inconsistently.
Where the debate gets genuinely nuanced
Very few credentialled behaviourists argue that positive reinforcement alone solves every case instantly, especially with entrenched, high-stakes behaviours like serious aggression, where safety and speed of intervention matter and a purely rewards-only approach may be too slow on its own. What most current professional guidance converges on is: start with reward-based methods as the default, use management to prevent rehearsal of unwanted behaviour, and treat aversive tools as something to avoid rather than reach for casually, reserved, if used at all, for cases assessed individually by a qualified professional, not a default first option.
The practical takeaway
For everyday training, loose-lead walking, recall, greeting manners, reducing barking, positive reinforcement is the well-supported default, and it builds a foundation of trust that tends to make future training easier rather than harder. If you're dealing with serious reactivity or aggression, that's a case for a qualified trainer or veterinary behaviourist rather than any DIY tool, aversive or otherwise.