Let’s continue our look into the mistakes people make when choosing to use punishment. Last week I wrote about two common mistakes people make: Correcting Unknown Behaviors and The Punisher Not being Unpleasant but those are only two of the issues I frequently see with people trying to use punishment.
Timing, it isn’t just for clickers
I am constantly talking about the need for proper timing when doing positive reinforcement training–either with a clicker or a verbal marker and it is no different if you are using punishment. The precision needed for the clicker to make a connection between behavior and reward is the same needed to deliver a punishment that the dog is able to connect to the unwanted behavior. More often than not, unfortunately, I see handlers issuing punishments that are so far after the behavior that the dog is not making the connection between the punishment and the behavior and is just getting yanked around or smacked on the nose.
One of the biggest culprits I see with this is barking. Dogs BARK-BARK-BARK-BARK and the handlers are trying to correct it but they don’t manage to correct the dog until it has stopped barking. I see it frequently with dogs new to my classes. The humans ignore the 5 minutes of quiet their dog just gave them and even ignore the 30 seconds of barking but once the dog stops barking to breath, they leash pop and tell him to be quiet.
Perhaps the worst of all has to do with house training or garbage digging/counter surfing. Family leaves Fido, a 5 month old puppy, alone uncrated for 6 hours and comes back to a wet spot on the rug and poop by the back door… instead of smacking THEMSELVES on the nose for leaving a puppy loose for so long, they then try to punish the dog for soiling the house. For all they know the dog could have peed 10 minutes after they left, the dog has absolutely no idea for what he is being punished.
Could you imagine getting a traffic violation ticket in the mail a week after the date listed and with the ticket not explaining what you did wrong, or where the ticket occurred, or when specifically you did the incorrect thing? Would that help you at all know what to change about your driving? This is what it’s like when dogs are punished well after the undesirable behavior finished–they have no idea what they did wrong and why their handlers are so upset.
When using corrections, it has to be as precisely timed as a click for the dog to make the connection to the unwanted behavior.
You’ve got to be consistent
Again, sound much like what I say when using a clicker? To teach a new behavior you need to reward the behavior every time it happens correctly. The dogs NEED multiple trials to really “figure out” what is getting that click and what is getting rewarded. If your dog is learning to sit and you reward it the first two times she sits but don’t the third–your dog will likely think that the sit was NOT the behavior you were clicking. Maybe it was the her head turn, her leg bends… she now second guesses when you ask for a sit. You’ve got to heavily and regularly reward the behavior you want so they learn exactly what is expected.
The same goes for punishment. If sometimes you punish a dog for a behavior and other times you either don’t (or worse yet reward it), your dog will never learn what behavior you do not want to see them doing. I see this ALL THE TIME with dogs who jump on people. Their people yank on the leash and yell at their dogs for jumping up one moment and then give them kisses for jumping up the next.
If you are using punishment to teach a dog what behavior you don’t want, you have to be consistent. If sometimes you punish the behavior and sometimes you don’t, the will not learn not to do that behavior. In order to teach them not to do that behavior, they have to be punished each time it happens so they learn it is not a “safe” behavior. If the punishment isn’t consistent, they can learn specific times it’s not safe (like when you are home) but will quickly learn times that it is safe (like when you are not home).
I think of it like a speeding ticket. These are punishments that happen SOMETIMES when you are speeding, but certainly not always and rarely do they even happen regularly. I would argue that most people who’ve gotten the sporadic speeding ticket still speed. Why? Because the punishment isn’t consistent and they’ve decided the risk of getting a rare speeding ticket is lower than the reward of speeding (or rather the reward to getting to their destination potentially quicker). The punishment is so inconsistent the behavior doesn’t change–although I would also argue that some of these people would probably slow down in the areas where they have a history of getting speeding tickets or they would become more vigilant looking for the source of potential punishment. This really really inconsistent punishment is totally ineffective (people still speed).
If, however, there was a device installed in every vehicle that would send an instant notification to the police EVERY time the user went over the speed limit and the driver would instantly get a notification that they would be receiving a fine in the mail… you can bet people would SLOW DOWN (at least until someone figured out how to by pass the system). The consistant nature and appropriate timing of this punishment would modify the behavior of a large portion of drivers (there are some who would decide that the punishment of the $200 speeding ticket wasn’t more punishing than the feeling of speeding–they probably wouldn’t speed EVERY time but they’d still make the choice to speed occasionally).
As, I’m sure you can see, there are some real similarities to the mechanics behind effective positive reinforcement training and effective use of punishment. We’ll check out a bit more about punishment tomorrow.