20 problems with punishment in animal training

20-effectsRecently there was a video post in my Facebook feed that caught my attention.

Typically, on Facebook, I’m a bit of a lurker. I’m not very active, and when I do watch videos I often don’t share, like or comment – even when perhaps I should.

This time, I watched, feeling my jaw gradually dropping in disbelief, and then I actually left a comment.

I wrote:

“I’m speechless”.

I know, that was lame.

But I didn’t have time for an essay, and then I was flooded by the rest of the FB flow, so the film slipped to the back of my mind – where it’s been festering.

A few weeks ago, I wrote that I was speechless. But in the time that’s gone by, I’ve realized that I should do the opposite.

I should speak up.

I figure, if people don’t realize that using punishment in animal training might be problematic, they need to be informed.

Jump straight to the 20 problems. 

I am by no means the first to voice my concerns, but this obviously needs repeating.

The video I watched proposed an automatic correction feature of an “on-collar training and tracking device” for dogs. Apparently, it’s a thing that you attach to a collar that can track your dog’s activity and also deliver electrical stimulation (a euphemism for shocks), noise or vibrations.

What was suggested in the film was that when the dog does anything that we don’t like, such as barking or trespassing in forbidden areas, it can be zapped, automatically, without us even being present. And so the dog will stop barking or trespassing.

I’ve seen this referred to as the New Standard of Dog Training.

For me, it’s Completely Outdated and Potentially Harmful.

The thing is, punishment works – unwanted behaviour is eliminated. At least sometimes. So people keep using punishment.

But, here’s another thing.

There is a price to be paid for using punishment. It may be a very small price, almost imperceptible. Or it may be very large.

It is the potential for serious fallout that worries me.

Using punishment in animal training is the equivalent to taking medication that only works sometimes and has humungous, not to mention common, side effects.

Frankly, I wouldn’t risk it unless there’s no other option.

At the very least, people should be aware that there are potentially serious side effects, rather than stumbling blindly into a punishment scenario.

Also, people should be aware of the alternatives, so that they can make an informed choice.

And yes, in a way I’m living under a rock. I’m not a dog owner, and such gadgets are not allowed in Sweden. I realize that in many other places they’re the norm, and nobody raises an eyebrow, let alone drops a jaw.

Perspective shift

Let’s take the animal’s perspective for a minute.

Imagine you’re a dog.

Come on now. I know you want to.

Medium-sized, brown, long hair. Goofy ears. I’m unfamiliar with the different breeds, but the dog that’s in my mind’s eye looks sort of like Lady in that old Disney film. You know the one.

Imagine Lady’s standing on the doorstep to the kitchen, and there’s a chicken on the counter, as depicted in the video that caused the jaw-dropping.

Setting Lady up to fail: a chicken on the counter.

The smell is delicious. And it’s within reach, too.

OK Karolina, but why are we doing this?

Bear with me. We’re imagining that we’re Lady so that we can get an understanding of her:

  • Conflicting motivations
  • Learning experience when zapped

Eeeehhhh…. will anyone read this?

At this point in crafting this blog post, I started listing and explaining all the potential problems with punishment.

In some detail.

Crafting illustrations.

Finding references.

It turned long. Very long. It was still messy, and at 5424 words, when it occurred to me: “what if nobody wants to read this?”

Who am I writing this for, anyway?

I figure, most people who use punishment probably don’t want to hear that they’re doing anything that might be considered suboptimal. I suspect that they may not read it, or zoom out by the time we get to reason 8.

And people who avoid punishment may not want to scroll through and read all of that – they may just skim the intro while nodding affirmatively.

So I’ve put the extended blog post on hold for the time being, and will let you decide.

Say what?

Keep reading.

Below you’ll find two sections:

  • An infographic listing 20 potential problems with punishment.
  • A few suggestions of how to get rid of counter surfing without resorting to punishment.

So, here’s what’s going to happen.

I’m going to leave it up to you: if you want me to expand on my list of 20 problems with punishment, let me know by pressing the button below.  If at least 100 people hit the button, I’ll wrap it up as an e-book and send it to your inbox. As a bonus, you’ll also get to download the infographic below – regardless of whether we hit the magic 100 or not. I’ll also keep you updated on free webinars, online courses and other blog posts.

If you just want the infographic and couldn’t care less about the e-book, hit the button below anyway. I won’t take offence if you never read the e-book. Honestly. I mean, 5400 words..!

*revision: it actually took just a couple of days to get to 100 sign-ups. By now, several hundred people have the e-book!*

The 20 problems with punishment.

If no other information is available to the animal, such as being fitted with a remote electronic collar, these are the potential problems that have been identified when using punishment to stop unwanted behaviour.

Describing the problem in one sentence is HARD and any suggestion for improvement is very welcome..!

This brings us to the question: if punishment is such a risky choice, what options are there?

Actually, there are several ways of addressing the problem of counter surfing.

Setting up for failure – or success?

What are Lady’s options?

Currently, the kitchen counter is fun, and the floor is boring.

This set-up is likely what motivates Lady to explore off-limits.

One hugely effective way of shifting that balance is to make the kitchen counter boring, and the floor fun.

Very likely, Lady will shift her behaviour too – you’re setting her up to be successful.

She’ll have no reason to go to the forbidden areas, because they would be boring.

Shifting the balance of reinforcement. Don’t take this literally.

Shifting the balance of reinforcement requires two (2!) things:

  • Thing 1: Remove or reduce the source of reinforcement in the forbidden areas. Don’t leave the chicken on the counter. In fact: don’t leave even a crumb of anything remotely resembling anything edible or interesting. Forbidden areas should be boring.
  • Thing 2: Make the allowed areas fun. While Lady probably wouldn’t mind, it’s actually not advisable to put three chickens on the floor. Rather, toys, puzzles, food mazes and whatnot – whatever she likes to engage in and that will keep her occupied.

Teaching impulse control

Also, you may want to do some impulse control training with Lady, so that she learns that by refraining from taking something she wants, something fabulous will happen.

Here are some key pointers on how I would teach a so-called leave-it cue:

  • Present inaccessible food (hidden in the hand) – as soon as Lady looks away or turns away from it, she gets a treat.
  • Gradually shift the criterion so that Lady gives eye contact rather than focusing on the food in your hand. An intermediate step is to first uncover the food when she looks away, then reward her for staying away from the uncovered food (which is more difficult than staying away from covered food).
  • Don’t use any signal to tell Lady when she’s doing the wrong thing such as mugging (many other trainers will say “wrong” or “eh-eh” or the like) – let her work out for herself that mugging gets her nowhere and eye contact gets her something fabulous.
  • Don’t pull your hand away when she tries to mug your hand – that could make it into a game, setting of her hunter’s instincts.
  • Change the set-up several times (direction, location, food type etc) so that she can learn to generalize.
  • Generalize the concept further by putting on a leash and throwing food out of reach on the floor – use a harness rather than collar (reduces strain on Lady’s neck if she lunges for the food). When she looks away from the food – give a treat.
  • Gradually shift the criterion so that Lady looks at you rather than the food on the floor.
  • Once she is reasonably reliably avoiding mugging, introduce the leave-it cue.
  • Work on duration.
  • Gradually make it even more difficult – responding to the cue even when you’re not in the room.

Emily Larham (aka kikopup) beautifully illustrates this training sequence in the video below, and explains it in more detail.

The New Standard of Animal Training

To me, a punishment-based remote system is not a New Standard of Animal Training. That approach has too many potentially severe side-effects – it would never be my first choice of addressing problem behaviour.

Rather, when dealing with counter surfing, I would arrange the environment so that the unwanted behaviour is unlikely to occur, and so that Lady has plenty of other, acceptable, options.

I would set her up for success – rather than failure.

I would teach Lady that it’s worthwhile to ignore temptation – reward her for doing something else, such as orienting towards me and giving eye contact, or a relaxed lie-down in a specific location such as on a conveniently placed mat.

Building relationships based on trust and a history of positive reinforcement rather than punishment will result in a more confident, happy Lady.

I know you’re burning to ask this question.

But what if the non-punitive approaches don’t work? Or the scenario is vastly different?

What about life-or-death situations?

Shouldn’t you ever, ever use punishment?

When to use punishment

Despite the list of downsides, there is a place for punishment.

If Lady was about to get run over by a car, I would do anything to prevent that from happening. That goes without saying. There are dangerous situations where you immediately need to stop behaviour, to save lives or avoid danger.

Also, getting an adrenaline rush is normal in such a situation, which may lead to a harsh tone of voice and forceful behaviour. This might be punishing to the animal, regardless of the intentions.

Still, if that happened, I would consider two things afterwards:

  • How badly did I just mess up? Which side effects can I expect to see? Can I reduce the impact?
  • I wasn’t prepared for this – what can I do to avoid this scenario in the future?

I think many animal owners run into these acute situations at some point, where you need to act quickly and resolutely in a way that’s aversive to the animal and therefore constitutes a punisher.

Animals recover well from the occasional punisher, that wouldn’t worry me too much. After all, that’s life. We stub our toe. Miss the bus. Get rained on.

“The occasional defensive punishment is likely to be treated less as a shock than as a signal that a reasonable boundary has been overstepped” Murray Sidman

I can relate to that.

But that’s another story. That’s the emergency situation.

Remote training with shock collars are not emergency situations.

Even if I had a dog, or shock collars were allowed here, I would be very hesitant to use them.  For one, I’d be afraid that I wouldn’t do it right. I’d try solving my training problems in other ways; after all, positive reinforcement is forgiving.

Punishment is not – one mistake might take forever to recover.

But sometimes you may run into a problem behaviour that can’t be solved through any other means, and that can’t just be ignored.

In such cases, punishment may be a solution.

For me and many others, punishment is a last resort, to be considered when nothing else works – it’s not the first thing to try, or even the fifth.

Actually, following the concept of the Humane Hierarchy (also known as LIMA or LIEBI), it’s about the eight or so, depending on whether we’re discussing negative punishment (removing something desired) or positive punishment (adding something aversive).

Nerd warning: the following includes terms that I would expect only experienced animal trainers would be familiar with. I mention it because I think beginners need to see this hierarchical thinking even if the details may be a bit complicated.

The humane list or LIEBI / LIMA (brought to the world of animal training by Susan Friedman). Solving behavioural problems is analogous to choosing which exit to take when driving. Try each exit before moving on to the next. The latter choices involve speed bumps, giving way, stopping altogether and consulting the map – meaning thinking it through and consulting others before going down those roads.
  • First you check whether any health or nutrition issues may be causing the problem
  • Then you see if you can address the problem by making changes to the environment triggering the response (nerd term: antecedent arrangements)
  • If that doesn’t work, try reinforcing some other behaviour – still doing nothing about the unwanted behaviour
  • If that’s useless, remove reinforcers for the unwanted behaviour while reinforcing another
  • If the problem persists, try extinction, negative punishment or negative reinforcement – these are all potentially aversive to the animal and some of the side effects listed above may result.
  • Finally, if all these fail and consulting with others have given no new insights, try positive punishment as a last resort. Try to mitigate the potential side effects – an empowered animal should recover quickly if this is a rare occurrence.

This approach assumes that you’ve mastered the different techniques on each stage, and that you perform a functional assessment to ensure that you’ve understood what is maintaining the unwanted behaviour. It’s beyond the scope of this blog post to go into details of how to do this, but I will be covering it in one of my upcoming online courses.

When choosing a training approach, remember: Positive reinforcement works, and punishment is dangerous.

It’s not to be undertaken lightly.

Actually, I wouldn’t use punishment alone, without giving any further information to the animal. I guess that’s one reason that I was indeed speechless when first seeing the remote punishment video: it really set the animal up to fail and didn’t in any way mitigate the potential adverse effects of punishment.

Interested in checking out the upcoming courses? Sign up be informed when I update this blog, give free online webinars or when my online courses about animal training, behaviour and welfare are available! Admission is typically only  open for a short while, so don’t miss the opportunity!

Reading list:

Koob (2013). Negative reinforcement in drug addiction: the darkness within. Current opinion in neurobiology.

Friedman, 2008. What’s wrong with this picture: effectiveness is not enough.

Schilder & van der Borg, 2004. Training dogs with help of the shock collar: short and long term behavioural effects.

Sidman, 1989. Coercion and its fallout.

9 Replies to “20 problems with punishment in animal training”

  1. Very helpful post. I trained my older beagle with mild negative reinforcement (tug on collar with release for wanted behavior), my younger beagle with clicker. The attitude difference is considerable. Clicker can appear to take longer in the beginning, until you get the hand of timing and setting the situation for success. I could never go back to negative reinforcement.

    PS – the reference link to Friedman (2008) does not work

Leave a Reply

Your email address will not be published.