Positive Reinforcement Pt 3 - Setting Yourself up for Success

See Parts 1 and 2 for a refresher.

Before I get to implementing basic training techniques, I'd like to talk about ways to set yourself up for success. This is a pretty big phrase in the training world, and it makes a lot of sense. If you don't set up a situation that is conducive to learning and succeeding (just like in the classroom), your success rates are going to drop dramatically. There is no sense in trying to force learning when it's not happening. Also, get this in your head now, IT IS THE TRAINER'S FAULT WHEN SOMETHING DOESN'T WORK... ESPECIALLY with animals who haven't chosen their situation and don't understand emotional or behavioral control or morals. They do what they do and that's that. If you go into this endeavor realizing that when something goes wrong, you approached it in a way that needs to change, I promise you it will work better. You cannot control someone or something else's behavior, but you can control your own. Even if your plan or implementation is genuinely a good one, it doesn't mean it works for that individual. And if that's the case, you chose incorrectly and need to reevaluate. Some cases are too challenging for some people or have extenuating circumstances, and that is completely ok. That's when you ask for outside help.

OK, now that I've gone into my little speech, here are a few tips for success:

Try to think of this endeavor as a partnership rather than you being "the boss." As we discussed before, animals/people who feel invested in the training rather than dominated by it are more likely to respond with positive results.
Make sure the environment is not hindering your results. If your subject is distressed by noise and activity, try to keep most of their life more calm and introduce these things step by step.
Trust is IMPERATIVE. Again, as we discussed before, you are making deposits in a trust account so that you have something to draw against when you have to do something negative. (Example: a medical treatment that you have to hold your subject for) If you have enough deposits into that account, a few withdrawals won't set you back too far. In that vein...
If you are new with your subject, have a trusted individual participate in negative interactions until you've built up some rapport with positive ones. And one more...
Don't "lie" to your subject unless you have no other recourse. For example, if you bribe an animal by showing it a fistful of treats, they are going to know there was more than one. They may not know a number, but they know there are many. Don't show them the many and then reward them with one. It ruins trust. Would you like someone who told you you'd be paid $500 for a job and then gave you $100 when it was completed? Furthermore, don't show something better and then switch it with a smaller treat when the task is completed. Basically, "Bait and switch" is bad to them too.
Bribery is tricky. Bribing is a wonderful too, but needs to be used wisely. First off, try not to "lure" your subject too much. For example, if you show your dog a treat and then walk with the treat to get them to go where you want, it's likely you will lose them at some point when they think it's not worth it. You're "lying" to them by saying they'll get the treat if they come this far and then moving instead. Did you like it when your siblings told you to jump for something and then yanked it higher every time you tried? Instead, try cutting the treats up into smaller bits. Have the animal come to a certain point and always give them the treat when they get there immediately. Then move farther back and ask again until you get to where you want. You can always phase out those steps slowly later.
Furthermore, phase out bribing as soon as possible with new behaviors. As I said, it's a great tool for new behaviors, but over time, the subject starts to have the ability to think, "hmmm.. is this behavior worth that treat? What if I ask for more?" Then you find yourself upping the ante. Try as best you can to move to rewarding after the fact so that this can't happen.
Don't ask for a difficult behavior all in one big jump. TAKE SMALL STEPS. It's called "approximations." You want to approximate toward your end goal. If I told you it was possible to have an animal willingly put its rump to a station to have an injection, you'd probably think I was crazy, but it's completely possible. If I told you that it took 2 weeks and fifteen different steps to get there, it makes a lot more sense.
Also, relax criteria when working toward new steps or changing the environment. For example, when E and I were visiting Jen, she relaxed her strict criteria for Elizabeth as far as watching movies, eating certain foods and behaving calmly at the table. She was less strict and allowed the behaviors to flex more than usual because she is a smart trainer. Elizabeth's routine was thrown off completely and expecting completely routine behavior in that instance would have spelled disaster for her discipline this weekend. Now that things are more "normal," Jen has reestablished her stricter criteria.
When looking at reward, try reward that your subject will work for. Spend time understanding that. If it's an animal, make sure they are not overweight if you're using food (usually the smart choice.) Overweight animals often have no need to work. If they are trickier about what they'll work for, try to find a special desired item that they ONLY get when training. NO EXCEPTIONS. This works for children well as children (and spouses. ;-) can often get whatever food whenever they want. Try to keep a special treat as a reward only for special behaviors. For example, keep M&Ms on a shelf only for potty training or sharing behaviors. Give them immediately after a success. There are also those rare subjects that work best for attention only. Give them extra special praise, hugs etc if success is attained.
For newer behaviors, REWARD EVERY TIME. Rewarding every time is a good habit to fall into in any case because even when we mean to, we often forget, so it ends up being a "variable reinforcement schedule" anyway. If we try to set up a "variable reinforcement schedule," we often screw it up. And for new behaviors, consistent reward is very important.
And finally (for now) consistency is more important than anything else. How many of you have had a bad boss but learned how to deal with him or her? Now, how many of you have had a group of managers who are all bad and have a horrible time trying to adjust to what each one wants? Which is worse? Usually the group of bad bosses because you can't get into any sort of pattern of behavior that consistently works which sets up the whole team for some icky failure.
Training is fluid. Nothing is ever fully trained forever in every circumstance. There will always be fall backs. When there are, take a few steps back in your plan and recover the behavior. Don't keep asking for the finished result with no response. You'll lose the behavior.
Don't invest in training your subject when you know you are feeling cranky or unfocused or when they are showing no interest. Either you will have a lovely breakdown in communication or you will just train your subject that it's ok to ignore your efforts. Come back later when you are ready. Let whatever it is go for the moment.
Please, please, please establish a "bridge." The communication will be so much clearer. We will get to that in the next installment.

There is SO much more, but that's what we'll go with for now. Questions?

Share On :

Pin It