Are you really reinforcing your dog’s good behavior?

If you’re a positive reinforcement trainer, you should know the answer to this question!

Consequences are important

Thorndike’s Law summed up training with “consequences drive behavior.”  This is an important idea if you want to train your dog, or your cat, pet rat, or Alpaca.  The rule holds true for all animals.

Consequences in positive dog-training

When you make the commitment to train using positive reinforcement, you have decided to use things that your dog loves as consequences during training, and to forgo deliberately using positive punishers—in this context “positive” means adding a punisher, such as a collar pop or an electric shock, and is distinguished from a “negative” punisher, which is taking away something good).  This seems like a basic concept, but its application is more complex than simply providing treats.

For effective positive dog-training, the following is critical.

  1. First, to the extent possible, manage the environment so that it is less likely that your dog will practice unwanted behaviors, e.g., using a baby gate to keep your dog out of the family room when children are playing on the floor to prevent scratches and nipping, or crating an unattended puppy.
  2. Next, manage the environment so that your dog is not reinforced for unwanted behaviors, e.g., keeping food off counters while training a “counter surfer”.
  3. Now, shift your focus from stopping “bad” behaviors to catching your dog being good instead.
  4. Last, when your dog does something correctly, know how to reinforce that behavior effectively.

Management

Management is always a good place to start.  Management has two goals.  The first objective is to arrange the environment, as much as we can, to prevent our dogs from rehearsing an unwanted behavior over and over.  This is because a frequently rehearsed behavior will only get stronger.  The second objective is to eliminate any source of reinforcement for the undesired behavior.  This step is important because behavior that is not reinforced will move into extinction, and eventually disappear, whereas behaviors that are reinforced will happen more frequently. 

So why not simply stop with management? 

With some issues, you can.  But when we are working with an existing behavior problem, extinction can be an unnecessarily frustrating experience for your dog, and extinction does not tell your dog what they should do instead—and what they come up with on their own may be equally problematic, or worse!

Catch them being right!

Shifting your focus from finding fault to identifying what behaviors you want to see more of can take a lot of work because many of us learned through correction ourselves—and often suffered because of it.  From ticketing speeders to having naughty children stay after school, our culture is permeated with finding fault.  But you can, with practice, start noticing your dog (and the people around you) when they do something right, or even something that is simply okay—if you reinforce approximations toward desirable behavior, you will get there faster than you might expect.  The more you do this, the more positive behavior-changes you will see around you—which is reinforcing to you too!

So how does reinforcement work?

Like consequences, reinforcement is also a simple concept that is not always straightforward in practice.  The first thing to know about reinforcement is that your treat (or toy or praise) is only reinforcement if it changes your dog’s behavior

Consider, if your dog sits and you give him a Cheerio and he spits it out, you would probably conclude that you did not successfully reinforce that sit.  But if your dog sits and you give him a Cheerio and he eats the Cheerio, have you actually reinforced the sit?  Maybe.  If he immediately sits again, hoping for more Cheerios, it probably did function as reinforcement.  If he wanders off after eating the Cheerio, or jumps up, or barks, the Cheerio probably wasn’t reinforcing to your dog at that time. 

This is why many trainers encourage students to use “high-value treats” in class and in distracting environments.  A bored dog, hanging around the house waiting for something to do, might like Cheerios and find them reinforcing.  That same dog, at class or in a park with strange smells, other dogs, and lots going onaround him, may no longer be all that interested in Cheerios—maybe sniffing or meeting other dogs is more reinforcing in this environment at this time.  Steak, however, might do the trick—or it might not.  The individual animal (or person) being reinforced always decides what constitutes reinforcement.

What should you use?

In most contexts a high-quality training treat, if your dog likes it, will be the easiest way to reinforce your dog.  In more challenging environments, you might try cut-up block cheese, small bits of hot dog (microwaved on paper towels to make them less slimy), boiled chicken breasts cut into small pieces, or steak trimmings.

But don’t stop at food.  Tug games, fetch (ball or Frisbee), and personal play are all excellent reinforcers once your dog has learned to love these games.  Unlike food, which will loose value as your dog gets full, these forms of reinforcement will retain their value during most training sessions—although they too could lose potency during a tiring, physically draining session or on a hot day.

The environment is also a great source of natural reinforcement.  If your dog loves going outside, ask for a behavior and, when your dog performs it correctly, let them out to play.  Sniffing bushes can reinforce loose-leash walking. A belly rub can reinforce lying down quietly while people talk or watch T.V.

Getting started

Draw up three lists.  The first one will rank your training environments from easy to the most challenging.  Think about what sort of reward would likely function as reinforcement in each context—e.g., Cheerios for performing a sit in the kitchen, cheese for sitting in class, and a fun game of tug for sitting with you on the agility course when another dog is running in the ring beside you.

Next, list everything your dog loves.  To the extent you can, order this list from highest-value rewards to the lowest (or the reverse).  Think about how you could use each of these items in your training—compare this list with the first one you drew up.

Next, draw up an ordered list of distractions.  Consider whether any of these “distractions” could be converted into reinforcers, at least if you set up the training session well.  Distractions are by definition super interesting and often are desirable to your dog; therefore, many distractions are likely to function as reinforcement if accessed by your dog as a consequence of engaging in training with you.

Last, have fun reinforcing your dog’s behavior effectively!  Your dog will love you for it.

Happy training!

Making training rewarding

Hexi Playing - cropped

It feels like this should be an easy topic, no real training or thought needed, and on some level this is true. We all have a general idea how to dole out a reward. But it’s oh so easy to reward ineffectively! I think most of our issues stem from a failure to recognize (or remember) that rewards are ANYTHING that makes a behavior more likely to reoccur. So, a reward is something your dog actually likes and wants, offered in a manner that links it to a behavior so that your dog is likely to do behavior again—like when your dog almost pulls your shoulder out of it’s socket as he drags you to a bush, and then rewards himself by sniffing at it happily.

One way to reward ineffectively is to accidentally reward a behavior you don’t want!

Another way rewards can run off the tracks is when some of us confuse treats with rewards. Pet food marketers encourage this error after all, and it’s easy to feel you have rewarded your dog simply because you have shoved a treat at them—I think we are all guilty of this at times. But does your dog like the treat? Or maybe your dog is bored with a treat that was once super yummy—until 6 months passed by with absolutely no variation. You don’t have to hang out with dog trainers very long to hear someone indignantly say, “but he always liked these before”! Would you like potato chips if they were all you were ever allowed to snack on?

I might want to test this theory on myself, just to be sure! Potato chips are pretty darn good.

Maybe your treats are great, but your demeanor screams boredom or frustration or inattention when you offer your dog their treat. Another common thing that I see is a person leaning down over their space-sensitive dog as they offer the treat, counteracting the pleasure the dog might have felt over the treat. All these types of things can easily affect whether a tidbit is received as a reward by the dog.

Think how you feel when someone slaps food in front of you and then marches off. It’s not dinner. It’s an insult. Consider context when you’re evaluating your methods for rewarding your dog.

A related problem is the ease with which we forget that there is a whole world of rewards out there that we can draw upon. We are creatures of habit after all. But we needn’t limit ourselves to one brand (although my dogs do like Zuke’s®). Even hot dogs or cheese can get old, so although food is really helpful for precision, and treat placement, toys and games can be fun too and more motivating for some dogs or certain behaviors. Tugging (let your dog win) is a great reward for a recall or for holding two-on/two-off in agility. Frisbee or retrieving a ball can be highly motivating, and works-off calories from all the teats you’re using too.

Watch your dogs’ demeanor as you branch out. There are dogs who just don’t want you to toss the ball; they want to hold on to it. A squeaky toy that one dog loves may scare a sound-sensitive rescue.

So, now you’re convinced you should branch out, but don’t stop at toys. Why not look for other opportunities to reward good behavior. Let your dog out to play after a nice “sit” or “down.” Tell your dog to “go sniff” to reward his decision to not tug you all the way to the well-marked mailbox. Also, what things in the environment are rewarding behaviors you don’t like? You may need to cut off access to this environment-based reward if you want the annoying behavior to slowly extinguish itself.

Quantity can also be important, along with delivery (e.g., one treat after another to draw out the fun, or a quick jackpot or a 5 minute game of tug for a great job). A puritanical insistence on avoiding handouts (the nothing-in-life-is-free plan) can also undermine certain types of training, such as classical conditioning, as can the ever-popular fear of over-treating. Outside of calories (and cost), I’m not sure why this is of such concern for so many people. But over and over, I see people reward their young dog for finding a source after a long, hard search with, drum roll, a single, teeny, tiny treat—even after they were encouraged to go wild—and then they wonder why their dog is distracted by all the good scents occurring naturally in the environment. Hint—they are more interesting than the reward that you’ve conditioned your dog to associate with the scent.

I’m not suggesting that we treat anything and everything, but we forget at our peril that a better-than-all-previous-efforts effort deserves a bigger-than-most payout even if the behavior still needs work. Any effort may warrant a jackpot if it was in the presence of an overwhelming distraction, or took an exceptionally long time to complete. I cannot emphasize enough how an occasional piece (or 5) of steak, or a meatball, or a long tug session in exchange for a better-than-most effort may be all that is needed to help motivate a dog accustomed to working in a sea of teeny, tiny kibble.

So ask your dog to learn a lot and work hard, but remember that nothing in life is free. Pay up!

*Next time we’ll talk about rewards, I’ll explore how searching can function as a reward.

Slaying dragons is difficult when you ignore them…

…Or when you decide to start with the mother of all dragons!

Slaying DragonsI read an article this morning, a parable really, that addresses the common practice of talking about change, promoting it even, while running scared from the reality of it.  There’s a similar phenomenon seen in dog training where a handler says she wants her dog’s behavior to change.  This handler attends classes to promote change.  She fights traffic to get to this class after work.  She pays for this class.  While in class, she listens carefully to what her instructor says.  Then, while waiting her turn, or coming or going from the building, this handler reverts instantly to gripping the end of her leash, hanging on tight, and tolerating all of the problem behaviors she just spent an hour learning how to address.  And then she wonders why nothing ever changes.

Behavior changes slowly.  Habits are easily entrenched, but hard to fill in or reroute.  There is no magic formula.  To change behavior we all must make small changes.  Work with new “baby” behaviors patiently and consistently.  You have to nurture a future habit until it becomes fixed.  Then you’re ready to make the next change, and so on.

This is hard however.  First, you have to know where you want to go with your behavior issue, whether its your dog’s or your own.  After all, a fun trick that you once taught your dog may turn into a “zombie” problem that keeps rising from the extinguished behavior pile to haunt your obedience routine, or simply piss off guests, if you didn’t look ahead at what your long term expectations really were.  For example, I have a “rewind” trick that now rises up in finishes and other exercises from time to time–I find it funny, but it could really upset a trainer with more exacting standards.  I have another trick where my dog peeks through my legs.  I like it, but not all guests are amused and my agility instructor can’t stand it!

It is also hard because we don’t break things down and think small enough.  For example, teaching “stays.”  Each session I ask students to start teaching “stay” first by teaching the dog to resist following a treat when the handler moves their hand away from the dog.  Then, I’ll have them simply build some time into their dog’s sit: one one thousand, treat.  Then move to: one one thousand, two one thousand and treat.  Now build this up to a solid 5 seconds.  At this point, I’ll ask them to add the actual word, “stay”, but go back to counting one one thousand, the hold period that your puppy has already mastered (and not a 5 minute stay), mark and treat, then release.  Now build up to five seconds slowly before stepping away–and when you do, lower the time held.

Invariably, students ask their puppy, who just acquired the concept of sit, to stay as they walk out to the end of the leash.  Of course the puppy follows, bouncing along beside them happily–or she might wander off to sniff the puppy next to her.  The student then calls and scolds the puppy (breaking down the puppy’s nascent heel and recall) and drags him back to try again–hopefully I catch each student before this cycle gets too far!

Knowing this, I have to wonder why I still decide I’m going to make a change, like taking up meditation, or eating well, or starting this blog, and then alternate between attempting to create an award-winning site and completely ignoring my page, sometimes for years!  I should listen to myself and start small.  Baby steps.  Write an entry each week (I almost said day, but caught myself).  When this is habit, I will grow the behavior.  I’m getting better and better at expecting smaller changes from my dogs, maybe I can start expecting less of myself, and getting further in the end.