Questioning using food to train?
Are you worried that using food for training is just bribery? That
your pet won’t do what you want if you don’t have a treat? Are you
considering that they may get fat? Why isn’t their love for you enough
to make them do things? Read on and find the answers to all these
questions and more regarding the use of food in training.
Using food for training? Isn’t that just bribery?
The fact is, Yes, sometimes. It can become bribery. But this doesn’t always have
to be. When food, or any other reward is used properly, it’s not a
bribe and the animal will comply or behave in the long run, whether you offer food every time or not. This goes for any reward. (By the way, not hurting or scaring your dog if they obey is blackmail.)
How do you keep it from becoming a bribe then?
For teaching new behaviors:
At first, food is used every time in order to provide sufficient
motivation for the animal to offer to repeat the behavior. This has to happen enough times to become a learned behavior. Animals require a lot of feedback at first, and rewards provide this. This is called Operant Conditioning. Everytime I do a specific behavior, this is the outcome. This is called Positive Reinforcement.
After the behavior has been taught, the reward can be phased out to intermittent use to keep the animal playing the odds. Think of a slot machine. People don’t win every time, but they continue playing anyways because they’re hoping that this time will be the time that they win. This is called a variable ratio schedule of reinforcement. It’s so effective, that many people actually get addicted to gambling. This is a powerful training tool.
It’s not quite the same but still pretty similar to humans going to work for their paycheck or salary. Most people don’t bring home money every single day that they work. They usually get their money on a designated pay day.
Or you can use what’s called a fixed schedule of (varied) reinforcement. I personally prefer this because it means that for every compliant response to a behavior request, the animal gets a reward. But instead of food every time, you offer a variety of different rewards. This would occur after the initial learning phase and you’re maintaining or brushing up on the particular behavior. This schedule of reinforcement allows both to happen simultaneously. It also provides a high rate of feedback, which is a great communication practice since animals can’t read minds (that we know of). The variety prevents boredom from one particular reward being used every single time, and it allows you to never get caught without a reward. You can run out of food, or forget a toy at home, but you always have other rewards to use.
Even so, further measures to help prevent food from becoming a bribe are:
1. Not having the food visible until you’re actually giving it to the animal. I like to wear a treat pouch turned behind my back or on a nearby surface that’s within reach but out of the animal’s eyesight.
2. Having the reward food be different than the food used for luring.
Example: A peanut in your left hand to lure a bird’s head down and a sunflower seed give from your right hand for the actual reward given to the bird.
3. Switch up your use of rewards after the behavior has been learned. Food is just the tip of the iceberg. Try out the fixed schedule of (varied) reinforcement:
You can use a variety of different foods together, throwing a ball,
giving affection*, play, etc. Your pet should be the one to decide what they find reinforcing. A non-reinforcing “reward” isn’t going to motivate a dog to repeat a behavior in order to earn more. For example, praise, just as is, isn’t going to cut it for a lot of animals. Just like most people aren’t going to go into work every day for a kind word or pat on the back from their boss. People want and need their paycheck.
You can strengthen rewards or create new reward options by pairing them (praise, for example) with food and then slowly phasing out the food when they become rewarding all on their own. These are known as conditioned or secondary reinforcers. These usually have to be occasionally paired with food again at intervals to maintain their value. Some other rewards are inherently reinforcing, such as tugging a toy, getting to sniff the grass, chasing squirrels, and so on. You have an endless supply of reinforcement to choose from!
*To help determine if your dog actually enjoys being pet enough to accept it as a reward, here are two helpful resources from the great Eileen (I recommend reading her other blog posts too!):
Here is a helpful video on how to reward and not bribe! http://www.youtube.com/watch?v=joghYP-WE3c
For emotional behavior problem solving:
In the beginning, food is used often in order to transfer an emotional or physiological reflexive response (the latter is called a respondent behavior, examples are flinching, drooling, tensing, and so on) from one stimulus (such as food appearing, a stranger ringing the doorbell, a loud noise, getting zapped by static on the carpet) to another stimulus. Then once this response has been transferred, we back off and eventually no longer need the food pairing. This is called Classical Conditioning.
Russian psychologist, Ivan Pavlov, demonstrated this in an experiment with dogs. Pavlov noticed that dogs would drool when he fed them. So he started to ring a bell every time before he would feed a specific group of dogs. Since the bell was rung every time before they got food, the bell became established as a predictor of food. Hence, the bell eventually produced the same salivation effect that the food would, even when it was no longer followed by food.
Another example you may be more familiar with is that hand slapping game. Someone rests their down facing hands on your hands and then they try to slap your hands before you can react and move your hands away. As you’ll notice, sometimes you’ll jerk your hands away, even when your opponent hasn’t moved theirs. Jerking your hands away is becoming a respondent behavior to your opponent resting their hands on top of yours.
This phenomena works with emotions too. If Fido loves bacon, this will release pleasurable feeling hormones in his brain when he gets some.* If you pair bacon to a second stimulus often enough for the brain to view it as a predictor of bacon, the hormones will start to release even when the second stimulus is encountered independent of bacon. We often use this to change a fearful/anxious response to one of calm, anticipation. It’s very effective at removing fear and anxiety because what was once distressing now predicts fabulous outcomes. This is called counter conditioning. It’s most effective when the second stimulus is introduced in very gradual degrees. This is called desensitization. The term for this combined approach is Desensitization and Counter Conditioning.
More examples of conditioning. Fido REALLY loves bacon! It’s VERY high value to him.
Every time Fido lies on his mat, he gets bacon. Lying on his mat becomes a pleasurable behavior. Lying on his mat has value for him.
Every time Fido gets a nail clipped, he gets bacon. Nail clipping becomes a pleasurable activity.
Bacon is called the Unconditioned Stimulus because Fido innately enjoys it, you don’t need to condition him to like it. The mat or nail clipping is called the Conditioned Stimulus, because Fido doesn’t innately enjoy it, you have to condition the emotional or physiological response to it.
But be aware that this works in reverse too!:
Every time Fido gets bacon, he then gets his nails clipped. Bacon starts to become displeasurable if he dislikes like nail clipping strongly enough. This can also happen if Fido is aware of the bacon before he’s aware of the nail clipping. Bacon predicts nail clipping, rather than nail clipping predicting bacon. You see the difference?
Tell me more! Doesn’t this apply to Operant Conditioning too then?
Yes! Something to keep in mind about using food for a food motivated
animal is that they like it! Enough to work for it! When they get that
coveted food, it releases those pleasurable feeling hormones in their brain.
Since that has been paired with performing the behaviors that
you ask, their brain becomes conditioned to provide the same happy
feeling chemicals long after they get the food or not. Complying feels
good! Rate of compliance goes up!
In fact, some learned behaviors can become so enjoyable in and of themselves, that you can even use them to reward a separate behavior. This is called a tertiary reinforcer.
Fido loves to do his “Speak” trick. You can ask him to “Speak” as a reward for shaking paws.
All that food… Won’t I just make my pet fat?
Your pet has to eat. That’s a fact. But eating out of a bowl is
boring and quickly over. Adult wild and feral animals spend the vast majority of their time
working to get food. They forage, hunt, or scavenge. This exercises their minds and their bodies while getting the food they need. This is commonly replicated in zoos and aquariums, where
the animals would otherwise stand around with nothing to do all day. Poorly run zoos and other institutions don’t practice a program where animals work for food. The boredom leads to a lot of behavioral issues.
To prevent boredom, a part of their meals are often put into foraging toys (like puzzles) and
hidden all over the exhibit, and a portion is also usually reserved for
training purposes. It’s playing with their food! Play is fun. Figuring out how to get the food is pleasurable. It’s the Seeking System* at work. This system releases those same pleasurable feeling hormones mentioned earlier. It’s the same effect as when you beat a game or win at something.
When using food for training, you don’t want to use huge pieces of food. The goal here is to give bite sized pieces so you can get in many reps and have the morsel quickly consumed.
Isn’t it cruel to withhold food unless my pet does what I say?
As stated above, making mealtime an enrichment opportunity isn’t unnatural or necessarily cruel. The goal is not to deprive your pet of food! But on the flip side, if your pet is too hungry, this can lead to distraction and lack of focus during training time, causing frustration if they don’t happen to earn a reward every time. A recent study shows that animals learn better when not training on an empty stomach.* Frustration problems can easily be prevented by setting your pet up for success so that they don’t make a lot of mistakes, and feeding half (or slightly less, depending on what works for your individual pet) of your pet’s meal in a food toy a little while before conducting a training session. This should take the edge off your pet’s hunger, while still motivating them to earn the rest.
There are studies that show that animals will actually prefer to work for their food rather than just eat from a bowl. This is called Contrafreeloading.
But my pet doesn’t get excited enough about food. What now?
Most healthy, non-stressed animals will choose to work for food if given the opportunity. If the pet feels comfortable working for food, then use that
to your advantage! But if they don’t, then you can simply use another kind
of reward. Or it could help to try again later when the animal feels
more relaxed, as some animals won’t eat when stressed. This isn’t such
a loss. Not a lot of critical thinking takes place when your pet is too
stressed out to focus anyway. Ideally, when training, you want your pet to remain below threshold (the exposure point where a stimulus is at a strong enough intensity to cause a reaction*.)
It’s also possible that you haven’t tried the right food yet. A “treat
test” might be in order. (https://ladychaunceybarkington.wordpress.com/2013/03/04/food-testing-your-student/.)
But if you’re feeding a highly desirable diet out of a bowl all day long, your pet likely won’t value it. It’s kind of like the story of Midas, everything he touched turned to gold so gold quickly lost it’s value to him. So you need to at least partially limit your pet’s free access to their food.
But for pets who don’t mind working for food, this is just another
enrichment activity to liven their day. And it makes training
enjoyable. Who doesn’t like getting what they enjoy?
In fact, enjoy this video of Willy the Pug passing over food in a bowl in order to play with his food puzzle toys. https://www.facebook.com/video.php?v=895042050557853&fref=nf
Why shouldn’t my pet do what I say just because they love me and want
to please me?
I’m sure this is one of the reasons they do some of the things you
tell them. But if your husband, wife, parents, boyfriend, girlfriend,
or other person you love asked you to do things all the time, would you do
everything they asked? You’d likely eventually tell them to get the remote control themself or microwave their own cup of coffee. So why should animals be any different? We all work to get
the things we want in life. Most humans want and work for money (which is often then used to buy food). Most animals want and will work for food. I know you’re going to bring up volunteer work! But even volunteer work is reinforcing for humans. We get business contacts, friends, resume building, a sense of satisfaction, and so on.
So, the good news is that by using rewards, like food, your pets will
associate your training with fun, trust, and those feel-good chemicals
mentioned earlier! Training, itself, becomes pleasurable. This will positively effect your results.
Here’s a comprehensive article on the science behind training and why rewards, like food, work: