Page 157 - Zoo Animal Learning and Training

7.3 Two‐way Communication 129

VetBooks.ir the ante and offer more food. This only com- important aspects of shaping are the fluidity
of movement from one approximation to the
pounds the problem as some animals learn
waiting will result in higher value reinforcers
consequence sequence. The longer an animal
being offered inside. As with all reinforcers, next and the contiguity in the cue‐behaviour‐
food, especially favourite food items, should takes with a reinforcer, the farther away in
be presented after the animal has come inside time the animal is from the next behaviour.
the area and the door is closed. For instance, an otter can eat a 3‐g piece of
Trainers at free‐flight bird programmes fish in a couple seconds, but if the animal is
often experience problems associated with given a 20‐g piece of fish, it might take 10 sec-
prolonged use of the baiting strategy. To onds to eat the fish. Those extra eight sec-
encourage a hawk to fly to the glove, many onds can upset the flow resulting in the need
trainers will show the bird a particular food to move backward in the approximations to
item, such as a small piece of lean meat. From regain the momentum and get back on track.
the perch where the bird sits it can see the The shaping process is most successful
food and decide if that type or quantity or when the animal moves quickly across
reinforcer is worth the effort of flying to the approximations to the goal behaviour. The
glove. Often, after a short delay, and whilst size and type of food reinforcer should pro-
the trainer is ad‐libbing dialogue, the trainer mote quick consumption to keep the animal
reaches into the bait bag to add a piece of progressing smoothly through the approxi-
food to the offering. Too often if the bird mations and focused on criteria and conse-
delays longer the trainer will offer even more quences. The jackpot reinforcer not only
food, maybe even a whole, dead mouse, the disrupts the flow, it distracts the animal as it
bird’s favourite treat. I often wonder if these takes extra time to eat the increased quantity
birds are thinking, ‘these humans are so easy of food, and results in faster satiation. This
to train!’ small distraction may not ruin a training ses-
If baiting is used to help encourage behav- sion or even cause problems, but a jackpot in
iour, the trainer should provide an additional the middle of a shaping session may not help
backup reinforcer after the bird has landed the animal learn quicker or better either.
on the glove. If the trainer needs to prompt Jackpots and magnitude reinforcers may be
with a small piece of food in the early stages best used at the end of the training session to
of training, additional food items should be reinforce calm behaviour when the trainer
hidden in the glove. Occasionally, the bird leaves the training session.
should discover a whole mouse or other high
value reinforcer when it lands on the glove. 7.3.6 Fluency and Speed
As the consequence reinforcer hidden in the
glove encourages quick flights, the bait During shaping many trainers follow an 80%
should be faded out completely. rule, requiring 8 correct responses out of
every 10 trials (or 4 out of 5, etc.) before mov-
ing to the next approximation towards the
7.3.5 Jackpots
goal behaviour. Other trainers think of the
Trainers often talk about delivering ‘magni- 80% rule in a more subjective manner, such as
tude reinforcement’ or ‘jackpots’ for particu- a behaviour being 80% perfect before moving
larly high levels of performance of behaviour. to the next approximation. Either way, many
Their hope is to increase the likelihood the people follow a shaping strategy that involves
animal will recognise the large quantity of the animal performing multiple correct
food is delivered in response to a super‐ behaviours of a specific approximation before
criterion performance of behaviour, thus moving to the next approximation. In this
increasing performance in subsequent trials. manner a trainer puts a previously reinforced
However, that may not be the case. Two behaviour on extinction and selectively