Some of my most popular posts are about common training errors. It seems that I have an infinite supply, and I’m willing to use myself as a naughty example. New errors keep popping into my consciousness (and my training) all the time.
In this post I’m going to focus on two main categories of errors: problems with criteria, and problems with food handling. Can you identify with any of these?
Criteria Training Errors
Raising criteria too quickly. I’ve talked about lumping a fair amount in this blog. Lumping can happen when you are trying to teach a dog a skill gradually, usually in successive approximations. When you lump, instead of building the skill gradually, you skip ahead and leave the dog back in the dust somewhere. You take too big a leap. It is super easy to do. Remember, we have the steps outlined in our heads, but the dogs don’t. But there’s another way to lump as well. Perhaps you have split the task up into appropriate slices, and the dog guesses them correctly, but you simply go too fast for the dog to learn the appropriate physical skills. One example of this is when your dog is learning something involves strength or dexterity. We often say that we are not really teaching behaviors; the dogs already know how to do them. Most dogs can already sit and lie down; when we talk about “training” those it means, with a few exceptions, that we are teaching them to do them when we ask them to. However, sometimes we really do teach skills. For instance, some of my service dog trainer friends teach their dogs a special retrieve using their front teeth to pick up tiny, delicate items like a credit card on a tile floor or an earring that has sunk into the carpet. Picking up a tiny item is not just a behavior, it’s a skill. There is risk involved for the dog and the item and a ton of dexterity required. Such things need many repetitions. Even if the next step in the process is easy to make obvious to the dog, the dog needs experience at the current level before continuing. As a more common example, any behaviors that involve strength training need this kind of care as well.
- Raising criteria too slowly. I hate to tell you this, but raising criteria too slowly can be as much of a problem as raising them too quickly. You would think you could solve the problems of lumping by splitting all tasks into tiny sections and taking your time and doing lots and lots of reps. Wouldn’t that be great? Unfortunately, the Matching Law can bite you in the butt if you do that. If you build up a huge reinforcement history for approximations or incomplete behaviors they will stick around. Check out my post on what it took to rehab Summer’s target behavior after I had reinforced all sorts of half-hearted versions. And from Dr. Jennifer Cattet’s excellent article on the Matching Law:
“While shaping, the longer we stay on intermediate behaviors, the more we strengthen those behaviors over the target behavior. If we try to get each step perfectly before moving to the next one, instead of moving on quickly, we make all those steps stronger and therefore more likely to be repeated. We often believe that shaping a behavior ultimately makes it stronger. If we apply matching law however, this simply doesn’t hold true. In shaping, it can take 20-40 clicks (or more) to get the target behavior. During the training session, we’re likely to have clicked the dog more often for intermediate behaviors than for the target behavior.”
I think it stinks that even being slow and careful and taking our time can have a down side, but there we have it.
- Not holding to criteria. It really isn’t fair that there should be so many problems with criteria. But here’s the final one for today. Let’s say you’ve trained up a nice behavior. Your dog has the physical skills and the understanding. But life intervenes and you let your criteria for the behavior loosen up. Since many of the things we ask dogs to do are counter to their natural behaviors (that’s why we are training them in the first place), the dogs will be more than happy to lapse back. Every time you are in a hurry and let your dogs skip the sit at the back door before going out; every time you don’t wait for your dog to be seated politely before putting down his dinner; every time you let your dog pull you to the car, even though you practice loose leash walking in your front yard all the time–guess what you are doing? Shooting yourself in the foot, that’s what. You are not holding to criteria, and you are letting your dog get bootleg reinforcement for the very behavior that you have worked and worked to modify.
Food Delivery Training Errors
Here are just a few of the common mistakes we can make with food.
- Starting your food delivery before marking the behavior; i.e., treating before clicking. I write a fair amount about cues, and try to pay attention to what the real cues for my dogs’ behaviors are. I have written about that in several posts including this one: 16 Behavioral Cues I Didn’t Train (But Are Still for Real). But guess what? Sometimes our markers are not what we think they are either. If you use a clicker, a spoken word, or a mouth click as a marker, you can have very good timing but still introduce an error of mechanics if you start to move your food delivery hand before or at the same time you click. Do that enough times, and your marker becomes your hand moving toward the pocket or treat bag rather than the sound you intended. Just think: it is utterly consistent. It’s got a 1 to 1 correspondence with food appearing. So watch out! This is an example of overshadowing. (Yvette Van Veen recently wrote a great article (Part 1 of 2, I believe) about that.) Now, a clicker in particular is a unique enough sound that it would probably be difficult to completely overshadow. It’s still understandable as a marker. So what’s the harm? Clicker, hand movement–both predict food, right? One problem is timing. If your timing with the clicker is perfect but the dog is paying attention to the visual of your hand movement, he is not getting the clear signal that you think you are sending. What if sometimes you move your hand after the click, sometimes at the same time, sometimes even before? Who knows what you are actually marking with your hand movement? Even worse, if hand movement towards the food becomes the marker/bridging stimulus–well, you’d better not do it casually, right? We take great pains not to click without treating. We want the click to be a trustworthy predictor of a reinforcer. But if the real marker is the hand movement, we are probably marking things all the time without even knowing it. I know that sometimes when working without a marker (or so I intend) I reach for my food and change my mind. That can be frustrating and confusing to the dog.
- Always keeping food in the same spot on your person. Dogs know when we have food and when we don’t. If you load up your left front pocket like I do when you get ready for a training session or strap on your treat pouch, the dog says, “Yay! We are going to train and I can win some tasty food.” So what about the rest of the time when you need your dog to do something? Have you set things up so that the presence of food in that pocket or pouch is part of the antecedent for the behaviors you ask for? If so, then having no food present means the cue for the behavior is not complete. It’s not that the dog is lazy or being tricky. You have built a pattern that predicts reinforcement, and then you have broken the pattern. Luckily there is a straightforward way to fix it. Instead of delivering from your pocket or treat pouch, deliver from other random places. Deliver from a covered container that is sitting on your bookshelf. Or another by your back door. Or something that you have secreted in the back yard. Now, don’t go around filling up the containers just before your training session. The point is that they are there all the time and fade into the background. Your dog learns that you can give her something great even when you have not a speck of food on your person. Sue Ailsby builds the process of gradually teaching the dog that the food can come from anywhere into her Training Levels if you want to see it spelled out. Most pro trainers can coach you on making the transition as well. Just don’t go cold turkey! It needs to be gradual, since we are often adding a delay between the behavior and the reinforcer. That has to be done carefully or the relationship between the two will break.
Reinforcing the dog for looking at the food. I have created one very intent food-staring dog, and a couple of others with a fairly strong habit. If you use a marker, what happens immediately after you mark? Usually the dog orients toward the food. So think about it: that food-looking behavior is getting ultra reinforced by immediately preceding the primary reinforcer. So isn’t it natural that this behavior would bleed into other parts of the training session? In the past I have worked on a process to decouple Summer’s eyes from my pocket. Here’s my movie about that: Default Eye Contact Before Cues. Unfortunately, I have not been consistent. But my efforts have paid off somewhat. My friend Marge Rogers solves this problem by building eye contact from the dog into every possible behavior. It works beautifully if you have the self-discipline to do it. And that brings us back to criteria, doesn’t it?
A Philosophy of Errors
I write about my mistakes because I’m in a unique position to do that, being a behavior nerd and training aficionado but not a pro trainer. I get lots of feedback from people saying that they learn a lot from those posts and enjoy them. I just want to add here, though, that while it’s human to be inconsistent, it is not great from the dog’s point of view. We should be honest, even forgiving about our own errors, but we don’t need to get too comfortable with them. Even if we don’t punish the dog for errors that we bring about ourselves, being inconsistent about what behaviors are being reinforced does bring behavioral extinction into the picture, which is known to be frustrating.
So you can look for some future posts from me where I will perform some training problem solving, and one criterion of the challenge will be how to maintain my own behavior change. Because to change our dogs’ behavior, we have to change our own. That’s how it works, and it’s only fair.
Which of these errors, if any, are familiar to you?
Copyright 2016 Eileen Anderson
Photo of Rounder the buff ridgeback courtesy of Marge Rogers of Rewarded Behavior Continues. Love and miss you, sweet boy.