This is a question that comes up a lot with people who are exploring the uses of positive reinforcement and clicker training: “When can I stop using food?”
And my response is: “When can your boss stop paying you for the work you do?”
I know, that’s a bit facetious, but I think it gets the point across.
In clicker training, the click marks the behavior that you’re reinforcing (so the dog understands what the reward is for), but the click (or “Yes!” or other marker signal) is also a promise to “pay” and should be followed by a reinforcer.
In most cases, that means a treat, simply because that’s quick and easy. But you can use lots of other things the dog finds rewarding: specifically any opportunity to do something the dog really likes or wants to do. That may be moving forward again when practicing loose-leash walking, stopping to sniff something interesting after practicing “attention” work with you, the opportunity to play, or even the opportunity to do whatever it is you just clicked the dog for not doing (chase a squirrel, for example).
In some cases, the treat’s just not good enough, or the dog actually wants something else more (like chase that squirrel when it’s not safe for her to do so). But it’s important that you at least offer something you think will be reinforcing to the dog.
If your dog doesn’t take it, you have some homework to do: figure out what other things you can use as reinforcers — whether that’s food your dog likes more than what you’ve offered, or her favorite toy, or a quick game of tug, or a “Yay, Yippee!” party where you and your dog dance around like fools, or whatever.
It’s also important that you take a look at why your dog didn’t find something reinforcing when you expected her to. A good example of this is a dog who won’t take any treat, no matter how yummy that treat would be. Generally this is because the dog is too stressed to eat — in which case you need to figure out how to set things up so the dog’s less stressed (typically keeping more distance between you and whatever’s stressing your dog, or keeping training sessions shorter so you can get some good reinforcement in before your dog’s overwhelmed).
In some cases you just need to look at the food treat you’re offering and ask “would I eat that?” A dry, grain-based biscuit is likely only a little more appealing to your dog than it is to you. “Milk-bones” just don’t cut it for most dogs, especially if you’re asking your dog to do difficult work for such mediocre treats.
Bacon, on the other hand…. or liver…. or deli turkey or chicken or beef…. or high-end treats such as “Real Meat” treats… or occasionally “junk food” like almost anything in the shiny packages in the “treat” aisle of the pet store — those are things your dog may find irresistable in situations where even “medium-value” treats such as Natural Balance Dog Food Log aren’t enticing enough.
So, does that all this mean you need to keep shoveling food into your dog for the rest of her life?
Well, yes, actually, she has to eat if you don’t want her to starve.
I know, that’s facetious, too, but hopefully it gets the point across as well.
Typically, there are three objections to “having” to use food:
1. “My dog will get fat!”
The answer to this is to stop feeding meals and start using the food you’d be feeding her anyway when training. That’s how many positively-trained service animals and police dogs get their meals — as payment for the tasks they do throughout the day.
Of course, you may need to look at what you’re feeding your dog at mealtime. Just as with dry grain-based “treats,” dry grain-based kibble may not be high-value enough to be used effectively in training throughout the day. (It may not be all that healthy for your dog, either, but that’s another topic entirely.) You need to find something that’s healthy and well balanced but still really appealing to your dog.
If kibble just isn’t going to cut it (especially with work in the “real world” with a reactive dog) you can try canned/wet food in a squeeze tube, or you may need to use “treats” instead. As long as you’re choosing nutritious treats, you can cut down on your dog’s meals at mealtime without jeopardizing your dog’s health.
2. “I don’t want to have to carry food around with me all the time!”
This one’s easy!
Make a list of everything your dog likes to do and/or have.
Eating and food will be on that list, but so should a bunch of other things!
Learn to use reinforcements already present in the environment.
And one of those should be you! If you can’t be reinforcing to your dog without food or toys, you have some basic relationship-building to do. The more fun your dog has with you, the more intrinsically rewarding just being with you will be. You can make yourself more rewarding by paying attention to your dog — really paying attention, the sort of rapt attention you want your dog to pay to you, but make sure you use polite doggie language with soft eyes that blink rather than intense eyes that stare. You can make yourself more rewarding by playing with your dog (a quick game of “chase me!” with you as “victim” will be rewarding and reinforce your dog for following you) or by making an utter fool of yourself with your dog (see the “Yay, Yippee!” party description above).
Basically anything your dog likes to do (including behaviors you’ve taught them that they’re fluent in and perform enthusiastically) can be used as a reinforcer.
3. “My dog should work for me, not for food!”
Well, yes and no.
Ideally, our dogs would comply with all our requests promptly and cheerfully.
Realistically, though, there has to be something in it for them.
Think of it this way:
You do what you’re asked and expected to do at work primarily because you get paid to do so — in money, of course, but also in “perks” such as health care, paid vacation time, maybe employee discounts on merchandise, bonuses (more money!), etc. With luck, you also enjoy your work — but you still expect to get paid.
If you volunteer your time for some cause, your “pay” is in the satisfaction of doing someting you believe in, and doing it probably makes you feel good. But if the “work” aspect of your volunteer work doesn’t pay off enough for you (in terms of social interaction with other volunteers or people or animals you may be helping through your volunteer work, or in terms of personal satisfaction at making a difference), you may begin to cut back on the amount of time you’re willing to volunteer, limit the tasks you’re willing to take on (walking the dogs at the shelter, sure, but no thank you to cleaning the kennels!), or stop volunteering your time and energy altogether.
You probably do “favors” for family members and for friends — and you probably even do them unasked. But you also get “paid” for those favors — in the form of return favors, of satisfying relationships and companionship, and in the form of satisfaction from doing something for someone you love. But if you don’t see enough of those return favors (and reciprocal companionship and love) you may start to resent doing what’s asked (”why can’t you do the laundry for once?!?”) or even resent the person who’s asking. (Since we’re talking about dogs that are trained using positive reinforcement and clicker training, I won’t go into the dynamics of living in an abusive relationship….)
Whether you’re working, volunteering, or doing favors, your enthusiasm and willingness and promptness in doing those things will depend on how much you enjoy doing the tasks involved, and how much you’re getting out of doing those things (money, friendship, satisfaction, etc.). The more difficult or unpleasant the task, the more worthwhile the “return” on your investment of time and effort you may need to even consider taking on the task.
Why should it be any different for our dogs?
Fortunately, you can get to the point where your dogs are complying with your requests promptly and cheerfully, without you having to give them food after every behavior.
The key to that is to make the behaviors you’re asking for intrinsically rewarding. The dog does them because he likes to do them.
And the key to making a behavior intrinsically rewarding is to build up a very long, rich reinforcement history for that behavior.
Every time you ask your dog to do something, and your dog complies, and you reinforce your dog for complying, you’re putting a buck in your “relationship bank.”
Every time you ask your dog to do something and your dog doesn’t comply, you’ve taken five bucks out of your “relationship bank.”
So part of the art of being a good trainer is to know when to ask your dog to do something (or ask her to refrain from doing something she wants to do and do something else). You’re essentially betting five bucks that she’ll do it, and if you guessed wrong and she doesn’t do as you asked, you’ve just lost five bucks.
If you lose the bet and your dog doesn’t comply, there are only 3 reasons that happened:
- 1. Your dog didn’t understand what you were asking. (We may think “he knows better!” but dogs don’t lie and don’t “blow people off” without good reason.)
- 2. What you’re asking is too hard for your dog in the given situation. This could be due to distractions, fear or anxiety, pain, tiredness, or having a “reactive” dog rather than a “thinking” dog (in which case your dog might literally not even be aware that you’ve requested something!).
- 3. The dog doesn’t see anything in it for her that makes it worth the energy expenditure (mental or physical).
So it’s your job as a trainer to to make sure your dog understands what you’re asking, is able to do it under those specific circumstances, and knows she’ll get something worthwhile out of it.
So how do you get from the “compliance pays off every time” stage to the “working for you rather than food” stage?
Practice, practice, practice! Put those bucks in that “relationship bank!” Don’t drill behaviors — asking for the same behavior over and over again can get pretty boring pretty quickly and even yummy treats might not overcome the “not again!” feeling your dog gets when you ask for the same thing over and over again. Instead, look for opportunities to quickly run your dog through several behaviors she knows really well. Keep “training sessions” (or “practice sessions”) short — 2-5 minutes — but do several throughout the day. You can get 10 or more “touches” (hand targeting) into a minute-long session pretty easily.
Other behaviors may take longer to perform (especially “stays” or behaviors with any duration), but you can still build up the balance in your relationship bank by practicing the behavior a couple times a day.
Once your dog is able to comply enthusiastically 8-9 times out of every 10 requests in a non-distracting environment, you can start “variably reinforcing” that behavior under those conditions. That means you only reinforce the behavior with food some the time. The behavior should always pay off with a “good dog!” or simply your body language saying “thank you! that’s right!” but the “primary” reinforcer (food or play) can now be intermittant. This will actually strengthen the behavior, as it strengthens anticipation: “Will this behavior pay off this time?” (Think of slot machines… they’re on a variable schedule of reinforcement!)
Keep up the high rate of reinforcement for the behavior in more challenging situations (with distractions, or in the case of our reactive dogs in the presence of their triggers), until you’re getting 8-9 successes out of every 10 requests under those conditions, then you can begin reinforcing that behavior intermittantly under those situations as well.
With a strong enough reinforcement history, the behaviors you have on cue will become intrinsically rewarding.
Then your dog will work for you rather than food.
But you should still reward your dog — with love, companionship, and attention.