Saturday, July 14, 2007

A lesson in using food rewards

First a few cute faces I saw this week. Here is Baxter. He is a labradoodle. (Lab/poodle cross) He has gotten so much bigger than when I saw him last. (I forgot my camera that day so I had them send me a photo to share.) It is hard to see in the pic but his coat has gotten very wavy.

This sweetie is Pepper. He came from the local shelter and is doing fabulous being spoiled in his new home. His new owner just needed to learn a few things about dog behavior to make their walks more enjoyable.

Here is Orli enjoying a bone at her favorite spot in the window of her mom's store. (The Loyal Biscuit Company) I couldn't resist the photo opportunity.

Here is Surely after our weekly walk. It is hard to believe she once had a problem jumping into her mom's car.

This poor girl doesn't have a name just yet. She helped me at the prison yesterday by being a doggy distraction for Hope. Yes you can fool a dog with a fake dog sometimes. Hope gets very excited when she sees other dogs so this is a great way to work her without stressing a real dog out. The other prisoners got a real kick out of watching me give this dog commands. (Faking works best if you treat the dog just like a real one.)

Speaking of photos, check out this website: She does FABULOUS pet portraits. She is also set up at the Rockland Farmers Markets (non-rainy days) so you can walk over and get your pets photo done there as well. (Thursdays 9am to 1pm til October 25) She can also do some great work with photos you already have. Check it out.

And here is a photo of some of the prep work I do. Cutting up string cheese and hotdogs to use as food rewards during training sessions. I also use the shown store bought jerky treats called "Real Meat". It comes in many different meats like chicken, venison, fish, beef, etc. It is a soft jerky that is easily torn into smaller pieces.

Using food rewards is all about finding what the dog will work for. Think of it as a paycheck for the dog. I also use food as a lure to teach body positions without having to physically force the dog into them.

Now some people say they don't want to use food because it is bribery. Done incorrectly yes, it is bribery. There is a specific way you need to use food as a reward, and a specific way to phase food out of training. Showing your dog a treat before he will respond is a bribe. Rewarding your dog for doing something after they respond with a hidden food treat is a paycheck.

My goal is to teach the dog he has a reason to do what I ask. A dog sitting because we ask them to isn't natural to the dog after all. We need to teach them that we expect them to do as we ask when we ask. Initially that reward is food paired with the words I am teaching them. Eventually the reward may simply be verbal praise.

The alternative in training is called compulsion training which teaches dogs to respond in order to avoid an uncomfortable collar correction. Yes those methods work, yes they have been used for many years. There is a right and wrong way to do that training as well. After learning and using both, I prefer to use the more positive methods now whenever possible. I do believe it builds a nice solid bond between the dog and owner without fear. If I have clients that do want to use this method I teach them how to use it as fairly as possible to the dog. And some clients don't have a food option. (Drug dogs originally trained on prong collars for instance.) Though even those dogs get trained with lots of other rewards and praise too.

There are (at least) 5 ways we can praise our dogs. They are: Food, Verbal Praise, Touching, Play, and by Smiling. (Because dogs communicate with each other through body language and facial expressions they do learn to read ours.) You can see food is only one of many ways we can praise. Most people forget to reward their dogs for all good behavior, not just good behavior they want at that moment. (Say your dog is laying down calmly when people come in "What a good dog!")

Consider that dogs are "What's in it for me?" creatures. It is the Lassie myth that they live to please us. Granted smart dogs learn quickly that pleasing us gets them some great things, but it isn't the same as being motivated purely out of the love of making us happy. Dogs like to make themselves happy. They like to have fun, whatever that may mean to them. Chewing shoes and digging holes in gardens is great fun to many dogs. Our job is to teach them what we prefer them to do. Keep in mind that unsupervised dogs will make their own choices so management is frequently necessary as well as training.

Phasing out food needs to be done on an unpredictable schedule for the dog. In other words you go to a variable food reward schedule. They learn they don't know if they will get a treat, but they might. (Kind of like a slot machine. We know they don't always pay out but sometimes they do, so we keep playing.) This doesn't mean we don't praise them at all, we simply use our other methods of praise instead. Variable schedule example: the dog gets a food treat the first time, but not the second. They get one for the 3rd, but not the 5th and 6th. Then they get a jackpot. (meaning more than one treat) Jackpots need to be given one piece at a time. If you were to give 4 pieces of treat as a jackpot on one hand that only equals one treat to the dog. Hence giving it one piece at a time. This also gives you more time praising your dog.

I also use a reward marker word that equals a treat is forthcoming. I use the word yes! (happy tone) It is used AS the dog is doing the thing I have asked them to do. It is basicly clicker training without the clicker. The word yes! becomes the click. This helps us be more clear for our dogs. Many training issues I see are due to a communication problem between the owner and the dog. Sometimes the gap is simply a timing issue. Timing is important because dogs live in the moment. We humans aren't always very consistent either which is very confusing to our dogs as well.

Dogs do not have "This is good behavior, this is bad behavior" rules in their heads. It is all simply behavior to them. They will do any behavior that works to get them what they want. If we reinforce the good behavior and either ignore the bad (Example: jumping up on people-turn away and ignore them. If you push them off you are giving them attention by touching them, which is their intention and will not extinguish the behavior.) or address it by teaching them an alternative. (Example: stealing shoes to chew- teach leave it and replace shoe with kong toy or appropriate dog chewtoy.)

There are other ways we use food as well like during behavior modification. We can counter condition or use classical conditioning to teach a dog to expect good things when something they may fear appears. (Example: other dogs during a walk, baby strollers, loud skateboarders passing by.) Got a fearfully barking dog when people arrive? Pair the appearance with yummy food EVERY time people show up and you can change the emotion of that event for the dog.

I hope this helps explain the proper use of food in training. All training takes time, so one of our goals is to make that time with your dog fun for both of you. For most dogs food equals fun.

Have a dog that isn't food motivated? First try MANY different food tidbits. You may not have found the right one yet. Your other option is to find a toy your dog loves above all others and use that as his/her reward. And step up the other methods of praise. Find what your dog will work for and finds valuable. Remember, it is about THEIR perception of value that matters for rewards. Not ours.

(The lure reward method works great with deaf dogs. We simply use hand signals as cues instead of verbal.)

For more in depth information on some of these techniques check out the books: How to teach a new dog old tricks by Ian Dunbar and Dog-Friendly Dog Training by Andrea Arden.

Happy training!

No comments: