Let’s Talk about Using Aversives in Training. I’ll Go First!

A speech balloon with the words, "This method is OK because..." in it.Today’s post is about being  thoughtful about the use of aversives. I write about that frequently, so how about if I describe what I mean by “thoughtful”?

  • I am going to present a description of an aversive method I used to use.
  • I am going to list many common justifications that could be offered as reasons why that method could be OK.
  • I’m going to describe the possible fallout from the method for the dog and for the handler.
  • I’m going to discuss the difference between justification and weighing the pros and cons,  taking into consideration possible side effects of the method.

Aversives

Paul Chance, Learning and Behavior, 7th edition, defined aversives as:

Stimuli the animal would avoid, given the option.

An aversive can be a thing or an event. The same thing can be sometimes aversive and sometimes not. As I’ve mentioned, aversives can range from something a tiny bit annoying, like a fly buzzing around or hair blowing in your face, to being chased by a predator and at risk of death. As I delineated in a blog about assessing stress in training, we all expose our dogs to aversives. knowingly or by accident. Aversive events of different magnitudes are happening all around us all the time.

Also, for clarity’s sake, let me repeat something I say with great regularity. I do not claim to protect my dogs from all aversive events.  The thing I pay keen attention to is using aversives in training. Many of us have been culturally programmed to assume that applying a pretty high level of aversiveness in training is OK. I consciously fight that programming, and know many others who do.

I grew up as a trainer not perceiving dog body language. Like many of us, I did not notice the effects on my dogs of training methods, or I misunderstood them. I’m still working on changing my habits. I try to notice everything that is aversive to my dogs through observation. I mean everything. I am their caregiver and they can’t tell me. If “everything” sounds extreme, keep in mind that I am consciously trying to overcome a strong tendency in the other direction.

But I pay an even higher level of attention to my training methods. These are things I can control. If I use an aversive,  I need to have considerations beyond, is this working?

Example

When teaching a duration sit/stay, what do you do if the dog pops up into a stand before being released?

One approach based on positive reinforcement would be to lower the criteria for the dog. Pause a moment, get the dog to sit again (with a verbal cue if you are at that point), and increase the rate of reinforcement. Return to a generally shorter duration, smaller distance, or lower level of distraction, whatever you have been working on, and employ whatever method you have been using for graduating those. Work back up again.

Leaning in as a training method

Leaning in as a training method

Another common approach is to lean over the dog until she sits back down. If the trainer were at a distance from the dog, she would walk towards her and into her space first, then lean.

This is a negative reinforcement approach, with a probable element of positive punishment.

In negative  reinforcement, something is removed after a behavior, which results in the behavior happening more often.

In positive punishment, something is added after a behavior, which results in the behavior happening less often.

It is possible for both of these processes to occur as the result of the application and removal of the same aversive, in this case, body pressure/proximity. The dog’s behavior of standing up during a sit/stay is punished, and holding the sit/stay sitting back down again is negatively reinforced. (See below for addendum.)

When the handler leans over the dog, the behavior that “turns off” the pressure is the dog sitting back down. The pressure relief generally includes a well timed move backwards by the trainer.

Although two quadrants are involved, this method is generally characterized as one of negative reinforcement since a specific behavior has increased: duration sitting. If we solely punished the dog for standing, standing could decrease, but there could still be “flaws” in the duration sitting, such as the dog lying down or rolling over. Another common characteristic of negative reinforcement is duration of the aversive, and that is present here. The pressure continues until the dog sits; it’s not just a “zap” when the dog stands up.

In the range of possible aversives, body pressure is fairly mild. It doesn’t hurt or generally startle the dog. There are no “tools” involved. No touching, yelling, or nagging. I was surprised when first told that this method was aversive. We don’t always think of it that way, and that’s part of my point.

Justification

OK, it’s “Be careful time,” as my stepdad used to say.

Here are the first two definitions of “justification” from Dictionary.com:

1. to show (an act, claim, statement, etc.) to be just or right
2. to defend or uphold as warranted or well-grounded — Dictionary.com

Justification is not a dirty word. But if our goal at the outset is to justify a training method, I would say we are on the wrong track. Justify is often used as a euphemism for “excuse.” (That is listed as one of the synonyms at the same site.) Wouldn’t it be better to assess the method, analyze its pros and cons, and observe its effects?

What follows are a bunch of reasons one could present to justify the use of body pressure. Many of them are true, some are not. However, just because a statement is true does not mean that it can stand alone as a good reason to do something, or is even relevant.

Possible Justifications

  • Body pressure is natural. Dogs and humans both understand and respond to body pressure.
  • Dogs do it to each other.
  • It works.
  • It can work fast.
  • It’s not painful or startling to the dog.
  • Negative reinforcement happens all the time in life  (i.e. it’s “natural” too).
  • Using body pressure is not as bad as using [fill in the blank] method.
  • It’s really positive reinforcement because you are adding distance when you relieve the pressure.
  • Other people use aversives too, even if they think they don’t.
  • It clarifies the situation for the dog, making the wrong response instantly clear. Now they know what to do, and what not to do.
  • It is empowering. You give the dog a problem to solve, that the human is too close. The dog can figure out the solution and take action to rectify it.
  • If you use body pressure a number of times, you have to do less and less to get the dog to sit back down. You can get it down to a miniscule movement on your part or even just a look from you.
  • Quadrants analysis is just not applicable here.
  • It’s OK to use an aversive if you try everything else first.

These are all very, very common reasons offered for the use of aversives. The problem with all of them–true or false–is that they don’t address the possible down sides of the method.

Clara guards the sprinkler

Clara demonstrating the dogs use body pressure

What “Assessing, Analyzing, and Observing” Looks Like

What if, instead of leaping to a justification, we instead examined the ramifications of the method? Aversives are known to have fallout. Careful consideration means that you consider the possible negative outcomes of any method. If you are a trainer, you disclose these to your client.

The possible negative effects of using aversive body pressure fall into two categories: The effects on the dog, and the effects on the trainer.

Possible Fallout for the Dog

  • Using pressure from your own body is a form of (mild) coercion. If you want your dog to have completely pleasant and happy associations with your presence, you are risking putting a dent in those associations.
  • The sit/stay is no longer a purely joyful way for the dog to earn a goodie. You have added an “or else.” This is known as poisoning the cue.
  • The dog has learned avoidance pays off. Avoidance behaviors, once reinforced, are very persistent.

Possible Fallout for the Trainer

  • Methods such as this (when they work) are immediately reinforcing for the trainer. You get what you want instantly: the dog sitting back down. This could be positively or negatively reinforcing for the trainer or both.
  • Because they work so well, the human behaviors of applying and relieving pressure can become habitual and unconscious.
  • Because you don’t likely reinforce with food for a while after applying the pressure, so as not to create a behavior chain for the dog of popping up, sitting down, and getting reinforced, you are thinning out your positive reinforcement schedule right after finding out that it was probably too thin to begin with.
  • It creates a shortcut, a disincentive for doing the training patiently with positive reinforcement.

Careful consideration means weighing these possible negative outcomes against the perceived benefits of the method. And it means looking with an eagle eye for any adverse effects on the dog if we do choose such a method.

The Effects Vary

This example is straight from my experience. I did formerly use this method, and can honestly state that it was an incredibly hard habit for me to break. I probably have not broken the habit completely. Since dogs are so much more perceptive of small movements than we are, I imagine I am still doing it at times and don’t even know it.

How aversive this method is to any individual dog varies. Plenty of well known positive reinforcement trainers use the body pressure method for proofing stays (also for teaching dogs to back up). It’s not a terrible thing for most dogs, although you can definitely find some videos of it where the dogs are not happy.

Thinking of my own dogs, I would be ill advised to use this method on Zani, who is already incredibly pressure sensitive.  I would likely undo the work I’ve been performing to classically condition a positive emotional response to proximity and handling if I used my body to coerce behavior from her.

On the other hand, my dog Clara is so thrilled to be close to me in most circumstances that the method might not even work on her. It might not even be aversive. If it were a little aversive, and it did work, the long-term effects on her would probably be negligible. That’s my gut assessment, based on her personality and strong bounce-back ability. However, the effects on me would be just the same. If it worked, I would be reinforced for using it. How soon would it be before I used it on the wrong dog?

Assessing Methods

This way of thinking about one’s own training can be applied to methods that we learn about or that are presented by other trainers.

If someone is promoting a method that includes use of aversives, to me it is a danger signal if they enumerate justifications similar to the ones I listed and do not discuss the risks for the animal and trainer.

But What if My Dog Tries to Dash Out the Door into Danger?

Step in front of her. Grab her if need be. And see this post.

Regarding Comments

I would love to hear from others who have assessed aversive methods they used to use and have decided to do otherwise. Examples of the opposite type of story–“I tried everything but ended up using punishment”–are a dime a dozen.  So let’s tell the stories about consciously and successfully leaving the aversives behind. Ready, set, go!

Addendum

An astute reader pointed out that if the pressure is released after the dog sits, as I explicitly described, the stay itself is not being maintained by negative reinforcement. (That could be possibly be achieved by the human looming obnoxiously over the dog, but that’s not what I described.) If the dog does stay after the trainer releases the pressure, it could be because the popping up was positively punished, or from a generalized reduction in activity from the use of an aversive (passivity can result in or look like a stay). The positive reinforcement previously used could also play a part, but I think it is more likely that immediately after the pressure move, the dog is more affected by the aversive use. Thank you for the correction, and any errors here remain my own!

Coming Up:

  • The Girl with the Paper Hat Part 2: The Matching Law
  • Punishment is not a Feeling
  • Why Counterconditioning Didn’t “Work”
  • What if Respondent Learning Didn’t Work?

Eileenanddogs on YouTube

 

Share Button

About eileenanddogs

Passionate amateur dog trainer, writer, and learning theory geek.

Eileen Anderson on Google+

This entry was posted in Behavior analysis, Escape/Avoidance, Negative Reinforcement, Punishment, Terminology and tagged , , , . Bookmark the permalink.

20 Responses to Let’s Talk about Using Aversives in Training. I’ll Go First!

  1. Hi Eileen,

    Wonderful, thoughtful article. Interestingly, according to the American Heritage Dictionary, “Justify” means:
    1. To demonstrate or prove to be just, right, or valid: justified each budgetary expense as necessary; anger that is justified by the circumstances.
    2. To free (a human) of the guilt and penalty attached to grievous sin.

    The proof of being “right” is not only the ego, irrelevant and relative but also immaterial and moot when it comes to aversives, unpleasant and/or stressful feelings and animal welfare. Being conscious and aware or our and the dogs’ actions and body language at all times is so important. Not only when dog training and modifying behaviors, but in day-to-day life. Conscious awareness, is a growing process and lifelong pursuit much like training, with practice, we get better. 🙂

    • Beautiful points, Russell. Thanks for the definition and the commentary. Absolutely agree about conscious awareness.

      Wow, and how about definition #2! Could have a field day with that one!

  2. This is absolutely invaluable as a resource, Eileen! Thank you so much!

  3. Lisa Schmidt says:

    OK – I am TRYING to just use positive methods with my 5 year old rescue but am at a loss in some situations and can’t think of positive alternatives. For example, he likes to bite flying insects like bumble-bees and wasps. I am concerned that he will be stung by a wasp in his mouth and would like him to stop. He has been stung or bitten a couple of times – I think only by a bumble bee, and while it hurt him, it is nothing compared to a wasp sting. Yelling a sharp NO! can stop him if he is chasing something and I always reward him for coming. I also reward him for lying still with buzzing insects around, The problem is that I can’t control the insects when we are out in the garden and I don’t want him to have to spend his summer inside the house. The urge to chase these insects are just very strong for him and my only way to stop him is to yell….! What else can I do that is a realistic alternative?

    • Lisa, I’ll respond a little later.

    • What if, instead of yelling “no” you had a phrase that your dog already understood, and would respond to fairly automatically? That’s what can happen if you pre-train a behavior, first in non-distracting settings (no bugs present), and later in increasingly distracting settings. The dog, once he has a really sufficient “reinforcement history” is conditioned to respond, almost before he has a chance to think about it. So, learning to train “leave it” incrementally, and using very good rewards (things the dog really really loves, not things you think he should love), is a good goal for you. You can also train him to play with something aside from the bugs, say a flirt pole. Dogs can be taught to chase on cue, and stop chasing on cue.
      Good resource for dogs that like to use their hunting software – “Harnessing the Hunter” DVD by Emily Larlham.

    • Lisa, I like what pawsforpraise described. The most important thing to me is that if you haven’t trained a “leave it” before, you can’t start with the flying insects! You have to start with something easier that you can completely control. If you work up to insects and are very consistent about it, the insects themselves can be a cue for your dog to turn away and come to you.

      Also it’s a great idea to do something else to harness that prey drive. Good luck! One of my dogs (Summer) chases flying insects and I know how intense that drive can be.

      Oh, and just in case this needs to be mentioned: don’t make any discrimination between the types of insects. If your dog wants to chase moths in the house, you need to apply the same techniques and rules there. You need to prevent the “chasing insects” behavior from occurring at all, even when the insects won’t hurt your dog.

  4. r0sey says:

    Also, re the “I tried everything but ended up using punishment” are a dime a dozen – I’m surrounded by trainers that support punishment and have tried it in various occasions on their advice and it has NEVER worked for me. As soon as I started introducing positive alternatives I saw a huge change in my dog. So I guess I’m the opposite – I tried everything and ended up using rewards!

  5. calkinsbetsy says:

    I recently convinced a great client with a jumpy, nippy adolescent golden retriever to teach the pup to jump on cue. This took a huge leap of faith from my client, bless her heart! And watching Donna Hill’s video on teaching opposites helped a lot. He’s already much better after three days. Who says that positive reinforcement takes longer? Not me!

    • That’s so cool, Betsy! Indeed, that’s a leap of faith. Susan Friedman always says to try to work out a way that the animal can do what it actually wants. Teaching a controlled jump is a great example of that. I bet the pup loves the solution, too. Thanks for the comment!

  6. brionylilo says:

    Hi,
    My retired guide dog Lilo used to have trouble with overhead obsticles.
    I used to give her very hard corrections as soon as I bumped my head.
    This resulted in her still knocking me in to things but quickly sitting to avoid the punishment or to make it easier to handle. Don’t know which.
    After realizing there had to be a better way I started stopping at every overhead I knew on a specific route and rewarding her with a click and a treat.
    A while after that we worked them without me stopping first. If she did it we had a party. If she didn’t I just ignored the bump and we’d retrace our steps and do it again.
    It took my dog 2 years before she stopped cringing if she made a mistake and I got a knock.

  7. Reblogged this on Take Paws and commented:
    Great post about Aversive training methods

  8. indiwind3 says:

    Wow. My first thought was “I don’t do that, because it strikes me as being intimidating.” Then I really thought about it and realized I HAVE done it, taken a step toward Indi (as I’m giving a no-reward marker) when she’s broken a stay, without even thinking about it! I don’t have a problem with using body-blocks in certain situations, such as keeping Indi back from the door or keeping the shelter dogs I walk from jumping on me, but I’d definitely like to be aware of when I’m doing it. I’m not a trainer, just work with my own little dog. Very informative and thought-provoking, thank you.

    • eileenanddogs says:

      indiwind, it is really, really easy to do without knowing it. Kudos to you for thinking about it. Now you can make a choice about it.

  9. Jessica says:

    Body pressure is one that I dither about a lot. Our veterinary behaviorist actually taught me to do it as an emergency maneuver, where I do find it both really useful and much less aversive than alternatives that involve physical contact. In the heat of the moment of, say, freaking out that the neighbor’s cat is within sight, body blocking him away from the window is much preferable to something like a collar grab. (There is no treat in the universe better than the cat, and that window is impossible to cover.) I would never use it for just regular training, though, because poor anxious Silas does not need more pressure of any kind. Still, it makes me very anxious to justify something that, I agree, is somewhere along the aversive continuum, even “for emergencies.”

    • eileenanddogs says:

      Jessica FWIW I think that’s a situation where you pick something that works quickly and is the most humane you can muster. Sounds like you are doing a good job. I think for my dogs, most body blocking is _more_ aversive than a collar grab. But that’s because I have conditioned the heck out of collar grabs.

  10. me says:

    I sincerely believe that aversives are a crutch for better teaching. I see it a lot in both child rearing and animal training. Care takers often jump to an aversive application, which usually requires little to no thought. And aversives are often what people quickly reach for when faced with a conundrum with positive training, instead of sticking it out and thinking it through. An easy way out that doesn’t require as much thought or understanding of the issue as a whole.

Comments are closed.