See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/351462307

The ABC's of Animal Training and Behavior

Presentation · June 2018

DOI: 10.13140/RG.2.2.20553.47208/1

CITATIONS READS

0 629

1 author:

Eduardo J Fernandez
University of Adelaide
88 PUBLICATIONS 995 CITATIONS

SEE PROFILE

All content following this page was uploaded by Eduardo J Fernandez on 17 May 2021.

The user has requested enhancement of the downloaded file.

 THE ABC’S OF
ANIMAL
TRAINING
AND BEHAVIOR
Eduardo J. Fernandez, PhD
  Three Parts:

The 


1. What is behavior?
2. How do we define behavior?
Lecture:  3. How do we measure behavior?
  “I can train any animal to do anything!”
 Can you?
 “Behavior is behavior.”
Commonly  BUT…

held  BEHAVIOR ISN’T ARBITRARY!

The organism is important!
training


 Concepts such as “instinctive drift”

concepts (Breland & Breland, 1961) and niche-

related mechanisms (Timberlake, 2001;
2002) remain core to any understanding
of animal behavior.
  i.e., the arbitrary behavior
(general-process) approach.
 Is all behavior equally
trainable in (a) any
organism, and (b) with all
rewards?

The  Examples:
 How easy would it be to
General- train a cow to run for food?

Process
 To train a dog to growl for
food?

Approach  To teach a Tasmanian devil

to scream food?
 What WOULD make a
Tasmanian devil scream?
Training tiny,
adorable
monsters.
  Point being behavior isn’t just behavior.
 The organism is important.
 The organism comes to the table with a
repertoire:

The  Niche-related and other evolved

mechanisms (Ethology).

Pragmatic  Reinforcement histories

(Behavior Analysis).
Approach  Both animal behavior practitioners and
scientists need to know about the animal
species and the individuals.
The Pragmatic Approach
(cont.)
 Regardless:
 TRAINING IS PRAGMATIC!
 You have two things in front of you:
 ENVIRONMENT
 BEHAVIOR

 Everything else (including supposed reinforcement

histories) are just constructs.
 Simple point: Know your organism (individual and
species). But focus on what you can SEE and CHANGE.
  Behavior is everything an
organism does.
 E.g., running, eating,
breathing, thinking.
 But then, what isn’t behavior?
 Behavior is not NOT
something!
 The Dead Man’s Test
So what is  Can a dead man do it? If
so, it’s not behavior.
behavior?  Again, training is pragmatic.
If you want to change
behavior, train what you can
see.
 Similarly, focus on what you
want, not what you DON’T
want.
Defining
Behavior:

 The ABC’s
 Defining consequences
 Who is the trainer/learner?
The 3-Term Contingency

 Behavior does not occur in a vacuum.

 There are the stimuli/events that occur BEFORE and AFTER
behavior that affect all behavior.

A – Antecedent
 B – Behavior
 C – Consequence
 Also known as the 3-term
contingency (it’s a contingent
relationship).
Antecedents

 Stimuli/events that
occur BEFORE a
response.
 Obtain stimulus
control due to their
contingent relationship
with behavior.
  Stimuli that acquire proper
operant control (stimulus
control) over behavior are
called discriminative stimuli
(Sd).
Antecdents  Whether you identify a
proper cue, all operant
(cont.) behavior obtains a cue.
 It just isn’t always
something you identify, or
even want.
Stimulus Control
and Clever Hans

 A horse that could

count:
 "If the eighth day of
the month comes on
a Tuesday, what is
the date of the
following Friday?“
 BUT…the stimulus
control was in the
trainer, not the
math.
  Operant contingencies are
defined by the effect a
consequence has on the
future occurrence of some
behavior.
 Again, like antecedents,
Consequences these are defined and
described in observable,
functional terms, and as
part of a contingency.
  Any consequence that leads to an
increase in the future occurrence of
some behavior.
 The above is a FUNCTIONAL definition.
Consequences -  If it increases behavior, it’s
reinforcement. If it doesn’t, it’s not.
Reinforcement
 Perception/Enjoyment is not part of the
equation: (Walmart child example).
Consequences
- Punishment

 Any consequence that

leads to a decrease in
the future occurrence
of some behavior.
 Again, punishment is
defined functionally.
 By this functional
definition,
punishment works.
 …to DECREASE
behavior.
  All that “positive” and “negative” mean
with respect to reinforcement and
punishment is the addition and removal
of some stimulus.
 Again, these are FUNCTIONAL
Consequences definitions.
If I deliver a stimulus and that increases
(+ and -) 
a behavior, that’s R+. If I remove a
stimulus and it INCREASES behavior,
that’s R-.
Operant Conditioning:
The four
consequences/contingencies
  As the Clever Hans example provided,
what is being learned, or even who does
the learning, often comes into question.
 “I was once quite convinced that I had
trained two porpoises to jump on
command (a hand signal from me) until a
visiting scientist with a stopwatch
The informed me that they were jumping
Trainer/Learner every twenty-nine seconds.” (Pryor,
1984).
Dichotomy
 Learning involves a relationship.
 Also, some might say they did that to her
on porpoise…
The
Trainer/Learner
Relationship
 All organisms learn!
 As a trainer, you should be
learning while teaching.
 Your rewards are
effective behaviors.
 We learn just like our
animals:
Through effective
consequences.

DON’T FORGET TO
BE A LEARNER TOO!
Measuring
Behavior:

 Reinforcer Selection
 Shaping
 Measuring Everything!
So why  I. People believe weird things.
should a  Don’t believe your own bull”doodie”.

trainer  II. We’re bad at remembering things.

 Anecdote is a great starting point. But
care about data is fact.
 III. DATA ENDS ARGUMENTS!
data?
  How do you know you have the best
reward?
Example:  Or, the most optimal reward for your
current training protocol?
Reinforcer  In working with human populations, both
teachers and caregivers who used
Selection rewards regularly picked less optimal
rewards (Cote et al., 2007; Green et al.,
1988; Reid, Everson, & Green, 1999).
Systematic
Reinforcer
Assessment
 Paired-choice
preference
assessments are one of
the simplest
systematic methods
for assessing potential
rewards.
 First used with applied
animal behavior
(Fernandez, Dorey, &
Rosales-Ruiz, 2004).
 It’s simple, it’s fast,
and you just present
2 things at a time:
Animal: _________________________ Date: __________________

Trainer: ______________________ ___ Time: __________________

PAIRED-CHOICE PREFERENCE ASSESSMENT

ITEMS

1. ______________________________ = /6

2. ______________________________ = /6

3. ______________________________ = /6

4. ______________________________ = /6

Left = /12 Right = /12

1 - 2 2 - 3

1 - 3 2 - 4

1 - 4 3 - 2

2 - 1 4 - 2

3 - 1 3 - 4

4 - 1 4 - 3
 Shaping is often called an “art”.
 But, there’s science/function to ALL
behavior!
 Taking training/shaping data forces you
to focus on behavior.
Training/Shaping
 It also requires you generate a shaping Protocols
plan.
 In short, just taking training data makes
you more systematic.
 Ex: Fernandez & Dorey (2020):
https://doi.org/10.1080/10888705.2020.1753191
Animal: _________________________ Date: __________________

Trainer: ______________________ Time: __________________

Observer: _____________________ Session: ________________

Each animal will be delivered a stimulus for 5-sec. Each response will be either
checked as a + or – for each trial. Trainers will initiate a new trial following a 5-sec
black out period. Correct responses during training trials will be followed by a click
and a treat, and an immediate termination of the trial.

trial 1 trial 2 trial 3 trial 4 trial 5 trial 6 trial 7 trial 8 trial 9 trial 10
 trial trial trial trial trial trial trial trial trial trial
1 2 3 4 5 6 7 8 9 10

Target 6” + - + - + + - - - + = 5/10
Target 6” + + + - + + - + + + = 8/10
Target 1’ - - - - - - - - - - = 0/10
Target 6” + + + - + + - + + + = 8/10

Target 9” + + + - + + + + + + = 9/10
Target 1’ + + + - + + - + + + = 8/10

Target 1’ + + + + + + - + + + = 9/10
 Target Training

10
9
8
7
# of correct

6
5 Target Training
4
3
2
1
0
6" 6" 1' 6" 9" 1' 1'
 Ethologists/Animal Behaviorists rely
extensively on ethograms.
 These are useful for trainers as well. Measuring
 What is your dog doing in ‘x’ vs ‘y’
environments? Time of day? With other
Everything
people? (ethograms)
 Equally important, what is happening
before and after your dog does ‘x’?
(ABC!)
 DEVELOPING AN ETHOGRAM

There are approximately 5 steps to creating an ethogram. Here they are, in order:

1. What’s your question?

-You need to have an idea of what your basic question is. Are you interested in
how a group of animals interact together? Levels of abnormal behaviors, like
pacing? This will guide what you then look for in step #2.

2. Informal observations

-At this step, it’s largely field notes. You observe the animals of interest, writing
extensive notes about what the animals typically do. One main step here is trying
to discriminate between mutually exclusive behaviors (e.g., sitting vs. lying
down). You’ll need to use these notes to come up with your actual behavioral
definitions/operationalizations.

3. Deciding on behaviors

-Once you’ve finished your informal observations, you’ll need to come up with
all the behaviors you want to observe, and then specific definitions for each
behavior. Generally, you split behaviors into classes of behaviors. For instance,
“inactive” may be an entire class of behaviors, with “lying down”, “sitting”,
“standing”, etc., all being behaviors under that class.

4. Defining behaviors

-Once you’ve decided all the behaviors, you need to define them. In defining
behaviors, you need to be specific about the observable event that occurs. In
other words, don’t impute any feelings/emotions into how the behavior is defined
(e.g., animal is tired). They generally should be mutually exclusive (any one
event can only fit into one behavior) and exhaustive (every possible event can be
accounted for by a defined behavior).

5. Determining recording intervals

-After you’ve come up with your ethogram, you need to decide how you will
record them. Generally, I use instantaneous time sampling methods (pinpoint
sampling), where you record whatever behavior occurs every ‘x’ period of time.
For instance, you might record whatever an animal does every 30 seconds for 1
hour. Whatever occurs on the 30 second beep is what is recorded.

-There are other methods that could be used, depending on the type of question
you ask. If you think you want to use a different method of recording, consult an
ethological methods book, such as Philip N. Lehner’s Handbook of Ethological
Methods, Second Edition.
Animal: _______________________ Date: __________________

Feeder : _______________________ Time: __________________

Observer: _____________________ Phase/Session: ___________

Data will be taken by instantaneous time sampling every 15 s. Left side of box
records response, right side records area animal was in.

:15 :30 :45 :60 :15 :30 :45 :60

Minute
1-2
Minute
3-4
Minute
5-6
Minute
7-8
Minute
9-10
Minute
11-12
Minute
13-14
Minute
15-16
Minute
17-18
Minute
19-20
Minute
21-22
Minute
23-24
Minute
25-26
Minute
27-28
Minute
29-30
Comments:
  WPZ hippo pool was neither heated nor
filtered.

Example:  Pools were dumped approx. every 1-2

days.
The Hippo  Some keepers suggested that the hippos
used the pool less frequently during
Pool Use “dirty” conditions (i.e., further from
refill).
Dilemma  In addition, it was also suggested that
the hippos used the pools less during
colder conditions.
Subjects and Data

Gertie Lily Lupe

 Data: Jan – July (6 months).

 493 sessions, 246.5 hours.
 4,486 temperature data points.
Hippo
Ethogram
 8 behaviors measured every 30 seconds for .5 hours.

 Manipulate (Ma): Manipulating any NON-FOOD item, such as

enrichment.

 Locomotion (Lo): Walking, either forward or backward, non-

repetitively.

 Foraging (Fo): Eating any food, usually provided by the keeper.

 Interacting with Other Hippo (IOH): Any social interaction with

another hippo (i.e., oriented to another hippo and vocalizing;
contacting another hippo).

 Lying Down (LD): Lying down or sitting.

 Standing (St): Standing with no locomotion.

 Submerged (Su): Hippo is completely under the water.

 Other (Ot): Hippo engages in a behavior that does not meet the above
behaviors (e.g., “startled”).
Hippos – Pool Use vs. Water
Change
DAYS SINCE DUMP
70
% OF OCCURRENCE

55

40

25 Water
Land

10

0 1 2
(196) (242) (55)
Hippos - Behavior/Pool Use
vs. Temperature
60
TEMPERATURE CHANGES
Lying Down
45 Locomotion

30
% OF OCCURRENCE

15

75

60

45
WATER
LAND

30
46

51

56

61

66
-6

-6

-7
-5

-5
0

2
View publication stats

Questions?

THANK YOU!