Shelter’s Edge

Recently I have had a few conversations with shelters struggling with how to work with dogs with significant behavior issues. The discussions have largely been around what type of training would be acceptable to use. The question asked has been, “Should I use only positive reinforcement, and what do I do if it does not work?”

Here is a shocker – I know of no shelter and virtually no trainers that use JUST positive reinforcement when training dogs. Use a leash and collar? If so, chances are you are not using just positive reinforcement.

Stay with me here – the fact is, learning theory terms are really confusing:
- We use the word positive not to mean good, but to mean “to give.”
- Reinforcement simply means increasing the likelihood a behavior will occur again.
- Negative does not mean “bad,” it means to “take away.”
- Punishment means to “decrease the likelihood a behavior will occur again.”

So you can have positive AND negative reinforcement – they both increase the likelihood a behavior will occur again, but one increases the likelihood by giving something and one by taking something away. If a dog pulls on a leash, I may train him to walk on a loose leash by giving him a treat or something else he wants when he does not pull, so that over time he keeps a loose leash all the time for the chance for a treat; that is positive reinforcement. On the other hand, I may choose to stop my forward movement when the dog pulls on the leash and stay still until he puts slack in the leash. That is negative reinforcement – I remove the lack of forward motion when the pulling stops.

Most who say they use just positive reinforcement use “time outs” – placing the dog in a quiet place or clipping the leash to a wall and ignoring the dog when he does a particular behavior. Depending upon how this is done, it is often punishment (or could be negative reinforcement if the time out ceases when the dog settles), as we are putting the dog in time out when he does a particular behavior in hopes that we will decrease the likelihood the behavior will occur again.

When I worked in zoos, I trained animals to do some pretty amazing things with positive reinforcement. I trained orangutans to volunteer to get injections, Komodo dragons to enter crates on cue so we could pull their blood, bears and jaguars to open their mouths so we could touch and inspect their teeth, and much more. Much of the training was true, pure positive reinforcement-based training – the animal was given something appetitive (something that they wanted) when they did some level of close approximation to the behavior I was shaping. It was hard work that took great timing, consistency and a fair amount of creative planning. Teaching a spider monkey to not reject her offspring often started with simply training her to reach out her hand – and resulted in a complex behavior of picking up her offspring and holding the animal to her chest.

When the vets would schedule a surgery that depended upon the animal volunteering for an injection, they came with a plan B just in case the animal did not volunteer for injection. Positive reinforcement is powerful – and we can do so much with it – but as it is completely dependent on us always having a strong enough motivator (and/or supremely talented trainer), we cannot depend upon it always producing the desired response. Or to put it another way, if what I “give” is not as motivating as something else, it is possible the behavior I desire will not occur. With an incredibly skilled trainer, the likelihood is low, but the fact is most of us are not that skilled, and certainly our adopters have other things to do besides take training seminars…

What we do should not cause harm, what we do should not cause significant pain – but if I can save a dog’s life by causing temporary discomfort with negative reinforcement or positive punishment tools – I am all for it. How can we possibly, because of a misunderstood science, kill a dog because we choose not to use a proven training method that can modify or manage a behavior? The fact is, you already use tools that cause the dog discomfort (either negative reinforcement tools or positive punishment tools, depending upon how they are applied). Anyone using a Gentle Leader, Sense-ation harness or Weiss Walkie? They work because when pressure is applied it is uncomfortable (not painful – uncomfortable) and the dog quickly learns how to turn off the discomfort. We can take a dog who was at risk because he pulled uncontrollably and apply a management tool to manage the behavior immediately.

I am a huge proponent of positive reinforcement – and use it every chance I get. But there are times where what motivates the dog to do the behavior I am looking to modify is so strong, that only an exceptionally skilled trainer can modify the behavior – and then retaining that behavior becomes a challenge. A dog who is motivated to chase to kill small prey items may require a mock prey item as a positive reinforcer, and an adopter who is willing to carry said mock prey item in her pocketbook. If I can add other tools to my tool box, I can provide a more realistic management and retention program to my potential adopter – and open doors to dogs who had previously had only sanctuary or the euthanasia door open to them.

Related links:
“Time to Change Our Perspective Around Food Aggression”
About SAFER
“There Is Chimp Poo Hanging On My Office Wall”
“A Blog Story: Why Tortoises Rock”

Tags: Behavior & Enrichment, Euthanasia, Saving Lives, training

You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.