Page 90 - Zoo Animal Learning and Training
P. 90
62 4 Choosing the Right Method: Reinforcement vs Punishment
VetBooks.ir cases they are influenced by other outside public perception. These are not usually
scientifically based decisions, but they are
factors:
have a huge impact on what choices are
Speed of behavioural acquisition: in some important to most organisations and can
●
training environments getting results fast available to those who train animals.
may lead to job promotions and be per- Generally, we as humans are a compassionate
ceived as a better trainer. The choices species. We train animals in an attempt to
about which methods were used to get the give them better care and help them live in
desired behaviour/results may not matter our world safely. The methods that we choose
to those organisations. Or, whilst it may to use must be governed not only by efficacy
matter, they choose not to ask how the but also by ethics. Just because we can train
behaviour was acquired.
Behavioural reliability: in many training something doesn’t mean that we should. One
● of the biggest factors that guides us in our
circles the key to success is whether or not decision about which training methods to
the behaviour is just as reliable weeks, use is our personal ethics. There are ethical
months, or even years after its initial acqui- guidelines laid down by our employers, by
sition. The animal trained to give a blood our profession, and by our peers, but as indi-
sample to monitor its diabetic condition is viduals we are also bound by our own ethical
not very helpful if the behaviour breaks considerations and beliefs. Many wise train-
down after only one or two samples are ers and scientists have used ethics and writ-
drawn. The animal trained to cooperate in ten about the importance of having an ethical
a research study will fail to be useful if they framework to guide animal training in their
are not reliable for repeated trials. An ani- decision making. Three of the most signifi-
mal trained for educational programming cant include:
is not very useful if it cannot be counted on
week after week for those programmes. In ● Least intrusive and minimally aversive
the domestic animal world the importance (LIMA) principle (Lindsay 2005): Stephen
of reliability can be the difference between Lindsay describes what he refers to as a
life and death. The animal that is used to cynopraxic (dog friendly) approach to
detect bombs in a stadium needs to be as training that is reinforcement based, but
reliable in its ability to detect explosives recognises that the need for aversive tools
two years after completing training as it may at times be necessary. The LIMA prin-
was the week it completed training. The ciple advocates for a ‘least intrusive and
guide dog that assists its blind handler to minimally aversive’ approach to training.
navigate the world needs to be as skilled He emphasises that any ethical approach
three years after training as it was one must be competency based, because it
month after training. Behavioural reliabil- requires skill and experience to know when
ity is often a key indicator of success in to use a more aversive approach. He also
many organisations. describes the dangerous effects of unnec-
Professional organisations: many people essary escalation of utilising aversive tools.
●
who train are bound by rules and guidelines ● Hierarchy of effective procedures
put forth by professional associations and (Friedman 2009): Susan Friedman asks the
groups designed to manage a species or a question ‘Is effectiveness enough?’ Just
breed of animal. These guidelines are often because a tool gets the job done is that suf-
derived through compromises and discus- ficient reason to use it? She concludes that
sions by many professionals with a wide it is not, nor should it be the only criteria in
range of skills, knowledge, and agendas. determining which methods to use. She
Public relations: sometimes organisations proposes a hierarchy in which the use of
●
make choices based on appearances, and punishers are a last resort, used only when