Ahimsa Dog Training, Seattle
Cart

Are You Sabotaging Your Training?

In my first training class, with Spoon (that’s her, to the left) I was told to lure my dog into a sit and simultaneously say, “Sit.”  Then I was to give her the treat and say, “Good sit!”

I doubt the trainer specifically wanted to teach my dog to only work if food was present, but I was sabotaging my training, and we were on the fast track to teaching her to ignore me if I didn’t have a treat!

The reason I can say that with confidence is because I now know about two scientific concepts, blocking and overshadowing, that prevent animals (including humans) from learning a cue.

Overshadowing: if two cues are presented at the same time (like the word sit and the lifting treat), the more obvious one will be paired with the outcome and the other will be harder to learn (Pavlov, 1927). Which do you think is more obvious to a dog? The hand with food in it or the word?

Blocking: if two cues are presented at the same time (like the word sit and the lifting treat) and it is learned that one of them works by itself (the lifting treat), then it is significantly harder to learn the one presented only in tandem (Kamin, 1969). This is called learned inattention.

The cues I gave as examples are the lifting treat hand and the word sit, but the presence of the food is also a cue, as is your body language.

Furthermore, I was also supposed to say “Good sit!” which specifically paired he the word sit sometimes with NO CONSEQUENCE at all, making it doubly difficult for the dog to learn. Oh dear. (I wrote more on using the cue in your praise in another article).

So how do you avoid sabotaging your training?

  • If you are using luring, teach the hand signal first, then transfer the cue to a verbal cue.  Do this by luring just a few times and treating when the dog is successful (to luring a sit,  put a treat to the dog’s nose, and moving it up so the rear goes down). Then quickly shift into using a pretend treat and luring with an empty hand. Click any attempts, and treat.  Gradually only click full sits.  Once the hand signal works to cue the dog to sit, then transfer the cue by putting two full seconds between the word sit and the hand signal. So it would be “Sit…….hand signal” or if the dog starts to sit without needing the signal, click and treat!
  • Better yet, use shaping or capturing (click & treat when the dog sits on his own or when he starts to sit). Once the dog is offering a sit, begin to insert the cue before he does it, as if you are cueing the sit.  Then only pay for sits that you have asked for, or else the food becomes the more obvious cue and the word sit will be blocked.
  • Clean up your body language so that only the word sit and the hand signal predict rewards when the dog sits.
  • A natural consequence of using food or rewards in training is that the cue you want to teach the dog may be overshadowed by the presence of food.  One way to help overcome that is to simply teach multiple behaviors, like sit, down, and stand, so that the dog has to pay attention to your words.
  • I also recommend using a counterexample to show that the cues are important on their own, even without food. Counterexamples have been proven to help teach previously overshadowed cues in humans (Heckler, Kaminsky, & Sloutsky, 2008).  Ask for sit, down, etc. when you have no treats on you. When your dog does them, say your marker word, like “Yes!” and run to the treat jar or refrigerator to get the dog a reward. Or use play: pull a hidden toy out of your pocket or instigate a game of chase-the-human.

Related blog post: To Treat or Not to Treat: Training Reliable Behaviors

Resources

Heckler, A. F., Kaminski, J. A., & Sloutsky, V. M. (2008). Learning associations that run counter to biases in learning: Overcoming overshadowing and learned inattention. In Proceedings of the XXX annual conference of the cognitive science society. (pp. 511-6). Austin, TX: Cognitive Science Society

Kamin L. J. (1969). Predictability, surprise, attention, and conditioning. In B.A Campbell & R. M. Church (Eds.), Punishment. New York: Appleton-Century-Crofts.

Pavlov, I. P. (1927). Conditioned Reflexes. London: Oxford Univ. Press.

Written by Grisha Stewart, Ahimsa Dog Training, Seattle [Slashdot] [Digg] [Reddit] [del.icio.us] [Facebook] [Technorati] [Google] [StumbleUpon] Tweet This Post!
Training tips and news by email:

rss RSS feed

2 Responses to “Are You Sabotaging Your Training?”

  1. Rachel R. Says:

    I have been using reward-based clicker training for about 1.5 months to train our two-year-old Husky-Border Collie mix (adopted from a shelter in Philadelphia, PA). I would point out that rewards can be anything – not just food. I want my dog to sit in even the most challenging situations (particularly challenging – waiting at the door before and after a walk). So, I tell him to sit, click him when sits, and then his reward is opening the front door (clearly a valuable reward, as he is actively trying to go through the closed door). I am definitely using food rewards for training de novo and particularly challenging behaviors (i.e. – “stay” while I am preparing your dinner), but once I am in the reinforcement phase, I am trying to be more creative and motivate the dog with whatever it is he clearly wants. I am new to clicker and positive-reinforcement based training, but I have been seeing results since almost immediately after bringing our dog home. I am interested to hear the thoughts of the professional trainers on non-food rewards.

  2. Hugh Says:

    I wanted to find out about ‘overshadowing’ and Google brought me here. Many thanks for providing such a clear and precise and detailed explanation – all I needed to know. I will print your rules on avoiding these pitfalls, pin them to my fridge and stay away from all trainers who make these common mistakes. Obedience was always our downfall; it should now be so much easier.

Comments? Thoughts?      (Please use our contact form instead to email a dog trainer.)

 

 

 

ahimsa dog training seattle [Home]   [Puppy School]   [Dog Classes]   [Contact Us]   [Dog Training Philosophy]   [Testimonials]   [FAQ]   [Free Dog Training Advice]   [Dog Trainer Bios]   [Puppy Help]   [Links]   [K9 Friends]   [Ahimsa Dog Training Gear]

© 2003-2013 Ahimsa Dog Training, LLC. Seattle, Washington State, USA.  
925 NW 49th Street, Suite C, Seattle, Washington, 98107.   206.364.4072    Email info@doggiezen.com