View Single Post
  #6 (permalink)  
Old 12-08-2018, 11:23 AM
noodles123 noodles123 is offline
Senior Member
Parrots:
Umbrella Cockatoo- 11 years old
Join Date: Jul 2018
Thanks: 3,421
Thanked 5,816 Times in 2,343 Posts
noodles123 will become famous soon enough
Re: Need some advice on Kermitís training

Once he is doing the trick you want every time for 5-7 days or so (not a science, just my experience--depends on the bird and the skill), you can change (thin) the reinforcement schedule, from your current schedule (1:1 or continuous reinforcement) to a variable ratio or fixed ratio schedule of reinforcement. Thinning can help reduce burn-out (in addition to other benefits), but it isn't something that you do until a skill has been mastered.

A variable ratio (VR) schedule of reinforcement tends to produce the most enduring behaviors because the bird doesn't know when to expect the reward, so it keeps trying (like a slot machine). To prove its effectiveness, this same type of reinforcement is often inadvertently responsible for increasing/maintaining undesirable behaviors (like screaming for attention)---To illustrate my point with regard to attention screaming and variable reinforcement: An attention seeking screamer may scream, scream scream...people ignore, ignore, ignore and then give in during a moment of weakness. If this happens enough (or if one family member is more lax than the rest), the bird realizes that it just has to do the behavior enough times and it will eventually get a reaction (attention is the reinforcement in my example/theoretical bird who screams for attention).

With that in mind, the same type of variable reinforcement can be used to reinforce desirable behaviors and they tend to have greater sticking power long-term. The bird will be more likely to continue trying without giving up if it isn't sure when the reward will happen (but knows that it will happen eventually).

To start out thinning, after mastery of the skill, I would move from 1:1 reinforcement schedule (current) to a fixed ratio of reinforcement FR:2 (Meaning, for every 2 tricks, the bird gets one sunflower seed). After doing that for a period of time, you could thin further. If I were you, my goal would be to eventually achieve success using a variable ratio (VR) schedule. This is where, on AVERAGE, for ever X tricks, the bird gets a treat. For example, on a VR:3 schedule, this might mean it only takes 3 tricks one time (before treat is provided), and the 6 tricks another, and then 2 tricks the next time). It is meant to be inconsistent to keep the bird guessing, BUT you should only do this once you are certain that a skill has been mastered. After trying a VR:3, you would keep thinning until you are providing fewer and fewer unnatural "reinforcers".

You can step back if you try thinning and it seems to not be working, but again, stick with 1:1 (continuous) reinforcement until you are positive the bird gets it and is able to do the trick consistently. Also, if you do progress to thinning and you try a new schedule of reinforcement, don't step back to a more dense schedule unless you are certain that the new one isn't working. The bird may get frustrated or take a bit to catch on, so don't immediately assume that it has failed and regress.... Does that kind of make sense?

Here is a good link (it's about kids, but it's the exact same thing): Schedules of Reinforcement - Educate Autism


Also- you said he he is sometimes not motivated by food. If this is the case, consider trying different foods, or see if you can figure out a new reward. In ABA the idea is that all behaviors serve a function and in order to reinforce a behavior, reinforcement must match the function of a behavior. In this case, you want a behavior to increase, so if food works, use it, but if food stops working, then you need to find something that is more reinforcing.


The 4 main categories of reinforcement and behavioral functions in ABA are:

Attention- the presence of a special person, a vocal response from a specific person, eye contact, shared activity, proximity etc.

Sensory (THIS DOESN'T REALLY APPLY HERE....it relates more to things like scratching an itch or sleeping when tired)

Escape (getting out of something undesirable- task, person, location)

Tangibles (food, toys, a sound, a special location etc)---> you are using this one currently, and since the desired behavior has increased, you have proof that food is reinforcing for him at least some of the time. If it stops working, then you will need to come up with something else, or thin it out so that he isn't getting satiated and burnt-out on the food reward.

Last edited by noodles123; 12-08-2018 at 12:05 PM.
Reply With Quote