Chapter 2

Research Methods
0 Basic Terms and Definitions
1 Independent and Dependent Variables
1 Functional Relationships
1 Stimulus and Response
1 Overt and Covert Behavior
1 Appetitive and Aversive Stimuli
1 Establishing Operations:
1 Deprivation and Satiation
1 Contiguity and Contingency
0 Measurement of Behavior
1 Behavioral Definitions
41/2 Recording Methods (Rate of response, intensity, duration,
speed, latency, interval recording, time-sample recording,
topography, number of errors)
½ Assessing Reliability
0 Research Designs
2 Descriptive Research : naturalistic observation, case
studies,
9 (1/2each) Experimental Research : experimental research,
control group design, single-subject design – simple
comparison, reversal design, multiple baseline design,
changing criterion design
Based on an actual conversation that took place
between a “relationship expert” and a caller on
a radio call-in show:
“Hi Dr. Kramer. I need some advice. I’m
wondering if I should get married or break off
my engagement and finish university first.”
“How old are you?”
“Twenty-one.”
“Break off your engagement. Statistically, your
marriage has a much better chance of surviving
if you don’t get married until your late 20s.”
“Oh, okay.”

How to do Behavioural Research ?

• Variable : a characteristic of a person, place,
or thing that can change (vary) over time or
from one situation to another.

• Independent Variable : systematically varied

• Dependent Variable : outcome
• Functional relationship – cause-and-effect
relationship
• Stimulus - any event that can potentially
influence behavior
• Response - a particular instance of a
behavior
• The response of one organism can act as a
• Overt behavior - potential for being directly
observed
• Covert behavior - perceived or experienced by
the person performing the behavior - not
publicly observable.
• Appetitive stimulus - an event that an
organism will seek out
• Aversive stimulus - an event that an organism
will avoid.
• Appetitive and aversive stimuli also defined
as events usually described as pleasant or
unpleasant - often quite accurate but be
careful
• Establishing operation - procedure that
affects the appetitiveness or aversiveness of a
stimulus
• Deprivation - prolonged absence of an event
that tends to increase the appetitiveness of
that event.
• Contiguity means “closeness or nearness”.
• Temporal contiguity - extent to which events
occur close together in time, e.g., thunder
and lightening.
• Spatial contiguity - extent to which events
are situated close to each other in space e.g.,
door knock and door bell.
• Spatial contiguity is perhaps not as strong as
temporal contiguity in learning
• Contingency - predictive relationship
between two events, such that the occurrence
of one event predicts the probable
occurrence of another.
• Behavioral definition –
• Objective – some observable aspect e.g.,
yelling or striking rather than feeling of
anger
• Unambiguous – clearly defined –
• It is beneficial in the intervention or rearing
or educational practices as it clearly states
appropriateness or inappropriateness of a
behavior.
• Recording methods :
• Rate of response - frequency with which a
response occur in a certain period of time.
Most appropriate when the response is of
brief duration, with a well-defined start and
finish.
• Cumulative recorder - measures the total
number of responses over time and provides
a graphic depiction of the rate of behavior. A
slow rate of response yields a gradually rising
line, whereas fast rate of response yields a
steep line. It also indicates the time at which
reinforcer is delivered.
• Intensity - force or magnitude of the
behavior, e.g., amount of salivation or force
with which a rat presses a lever.
• Topography - physical form of the behavior -
neatness of writing, dressing appropriately,
brush properly etc.
• Duration - length of time an individual
repeatedly or continuously performs a
certain behavior - appropriate when
concerned with increasing or decreasing the
length of time the behavior occurs - study
time, or TV time.
• Speed - how quickly or slowly a behavior
occurs - length of time it takes to run
through a maze.
• Latency - length of time required for the
behavior to begin.
• Latency, speed, and duration are confused
because they involve time measurement.
Consider the behavior of an athlete who
• Interval recording - measurement of whether
or not a behavior occurs within a series of
continuous intervals - proportion of intervals
in which at least one incident occurred would
be our measure - not concerned with how
many times incidents occurred in each
interval - appropriate when responses occur
at a very high rate or where it may be
difficult to count separate responses or if it is
difficult to determine the point at which the
behavior starts and stops.
• Time-sample Recording - measures whether
or not a behavior occurs within a series of
discontinuous intervals (intervals that are
spaced apart) - behavior is intermittently
sampled - percentage of these sampled
intervals in which an incident occurred is the
measure of interest - this method of
recording is very time efficient for our
observer, who can spend most of the day
working on other tasks or making
• Number of Errors - responses which can be
categorized as right or wrong
• Assessing Reliability : Whenever possible,
researchers attempt to utilize two or more
independent observers for at least part of the
assessment period, and then calculate the
extent to which the observers agree versus
disagree on their observations.
• The measure of interobserver reliability in
this case is calculated as the number of
intervals during which the observers agree
divided by the total number of intervals,
• There are no hard and fast rules about what
constitutes adequate reliability, but 80% is
often regarded as the minimum acceptable
level and 90% as the preferred level.
• If reliability is considered to be inadequate,
then our behavioral definition may need to be
revised or our observers may need further
training in identifying specific incidents of
• Research Methods and Designs :
• Behavioral scientists have as many
preconceptions about behavior as the
average person does - perhaps even more,
because it is their job to study behavior - and
are therefore quite susceptible to viewing
behavior from a biased perspective.
• To avoid, or at least minimize, such biases,
researchers attempt to define their variables
objectively and unambiguously and make
their observations in a consistent and
uniform manner.
• Descriptive Methods
• Descriptive research involves simply
describing the behavior and the situation
within which it occurs.
• Descriptive methods do not involve the
manipulation of any variables.
• Two commonly used descriptive methods are
• Naturalistic observation involves the
systematic observation and recording of
behavior in its natural environment.
• The naturalistic approach is excellent for
gaining rich, detailed information about a
behavior and the circumstances in which it
typically occurs.
• A major problem with this approach is that it
often leaves us uncertain as to which
variables are most important in determining
the behavior.
• Thus, the naturalistic observation approach
is often insufficient for gaining a full
understanding of a behavior and the
variables that influence it.
• Case studies : it involves the intensive
examination of one or a few individuals.
• Case studies can be done in natural settings
(as a form of naturalistic observation), or
• Case Study involves the intensive
examination of one or a few individuals.
• Case studies can be done in natural settings
(as a form of naturalistic observation), or
they may involve detailed examination in a
more structured setting such as a clinic.
• Case studies are especially prevalent in
medical research.
• The case study approach is frequently
employed in some areas of clinical
psychology, especially with respect to
relatively rare disorders - for example, fugue
states (popularly known as split personality).
• As with naturalistic observations, it is
important to ensure that case studies are
based on systematic observation and that
researcher bias has been reduced to a
minimum.
• In addition to the problem of researcher bias,
• Because case studies often involve only one
person, we have no way of knowing if the
case being described is the norm or the
exception.
• The major limitation, however, for both case
studies and other descriptive approaches, is
that it is usually difficult to determine which
variables areresponsible for which behavior.
• Nevertheless, despite these limitations, the
case study method of research, just like the
naturalistic observation method, often
provides a valuable starting point for further
investigations.
• Experimental research :
• In general, in their quest to discover cause-
and-effect relationships (that is functional
relationships) between environmental events,
and behavior, behavioral researchers have a
strong preference for the exeprimental
approach to research.
• In an experiment, one or more independent
variables are systematically varied to
determine their effect on a dependent
variable (behavior you suspect will change as
a result of changes in the independent
variable). Any differences in behavior across
the different conditions of the experiment are
presumed to be caused by the differences in
the independent variable.
• Behavioral researches use two main types of
experimental designs : control group designs
and single-subject designs.
• Control Group Designs : In the simplest form
of this design, individuals are randomly
assigned to either an experimental (or
treatment) group or a control group;
individuals assigned to the experimental
group are exposed to a certain manipulation
or treatment, whereas those assigned to the
control group are not.
• Control group designs may be made for more
than one independent variable, e.g., factorial
designs. In factorial designs it is possible to
study interaction effects for finer
understanding of the variables.
• A particular type of control group design, is a
comparative design. In a comparative design
different species constitute one of the
independent variables. If it often used to test
an evolutionary hypothesis regarding the
differences in selective pressures for a
particular learning trait between species.
Comparative designs can be simple or
• Single-Subject Designs :
• Single-subject designs require one or a few
subjects to conduct an entire experiment.
• There are several types of single-subject
designs, four of which are :
• Simple-Comparison (AB) Design
• Reversal Design
• Multiple-Baseline Design
• Changing-Criterion Design
• In a simple-comparison design, behavior in a
baseline condition is compared to behavior in
a treatment condition.
No. of Cigarettes smoked
20

15 No. of Cigarettes
smoked
10

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14
• Reversal design : This design is also called an
ABA or ABAB design depending on the
number of reversals carried out. A reversal
design is a type of single-subject design that
involves repeated alternations between a
baseline period and a treatment period. If the
behavior systematically changes each time
the treatment is instituted and later
withdrawn, then a functional relationship has
been demonstrated between the treatment
and the behavior.
• Multiple-Baseline Design : In a multiple-
baseline design, a treatment is instituted at
successive points in time for two or more
• Persons – X, Y, Z
• Settings – at home, school and coffee shop
• Behaviors – smoking, swearing, nail-biting
• Appropriate when treatment is likely to
produce a permanent change in behavior, or
• Nevertheless, this design is limited because
we need to have more than one person,
setting, or behavior to which the treatment
can be applied.
• Another limitation is that the treatment
effect might generalize across the different
settings or behaviors before the treatment is
instituted within those settings or behaviors.
• Changing-Criterion Design : In some
circumstances, the treatment is not intended
to produce a large, immediate change in
behavior but rather a gradual change over
time.
• A useful design for measuring such changes
is a changing-criterion design.
• In this type of design, the effect of the
treatment is demonstrated by how closely the
behavior matches a criterion that is
systematically altered.
• The design can, however, be greatly
strengthened by including periods in which
the criteria suddenly change in the opposite
direction (for example, in the case of Cory,
the number of cigarettes allowed would
sometimes be raised). If the behavior
continues to track the criteria closely even
when they change direction, then we will
have obtained strong evidence for the
effectiveness of the treatment. In a sense, we
have created a changing-criterion design that
incorporates certain aspects of a reversal
design.
• The reversal design and multiple-baseline
design are the most basic single-subject
designs, with the changing-criterion design
being less often utilized.
• Other types of single-subject designs have
also been devised, each having its advantages
and disadvantages.
• Dear Dr. Dee,
• I am suspicious that my boyfriend is having
an affair with his old girlfriend. Whenever
she is in town, he phones me significantly
less often. For example, between May and
August, when I know for a fact that she was
in town, he phoned me an average of 5.8
times per week, while between September
and December, when she was out of town, he
phoned an average of 6.4 times per week. I
worked it out, and sure enough, this is a
statistically significant difference!
• But when I confronted him with this hard
evidence of his unfaithfulness, he denied it
and said that I’m being paranoid.
• Am I Being Paranoid?
• Dear Am I,
• Given the evidence that you have presented, I
would have to say yes, you are being
paranoid. Worse than that, you are being a
• Quite apart from his old girlfriend being in town,
your boyfriend may be calling less frequently
between May and August for several other
reasons, such as spending more time in outdoor
activities or visiting with relatives.
• Such other possibilities need to be assessed before
you can draw any conclusions about your
boyfriend’s unfaithfulness. You also need to
recognize that statistically significant differences
do not provide hard evidence of anything. What
they provide is supportive evidence for a certain
possibility.
• Thus, even with a highly significant difference
between two sets of scores, there is still a slight
possibility that the difference is actually due to
chance variation. As well, you need to consider
that a difference that is statistically significant
may not be meaningfully significant. In fact, the
difference you have described seems quite small. I
bet that if you chart the number of phone calls
week by week, as in a simple-comparison design,
you will have a hard time spotting much of a
difference between the May–August period and the
September–December period. And in this
• Use of animals in behavioral research :
• Animal research has greatly contributed to
• our understanding and treatment of
serious diseases and illnesses,
• our understanding of basic physiological
processes
• Our understanding of basic principles of
behavior
• Advantages of using animals in research are
• Ability to control their genetic makeup
• Ability to control their learning history
• Ability to control experimental
environment – stricter control, extraneous
variables, intervening period interactions
• Ethical consideration in potentially
harmful and aversive experimentation
• An animal model is a procedure that uses
animals to mimic a particular human
characteristic or symptom, such as drug
• Perhaps the most fundamental criticism of
animal research is that it is morally wrong
and that animals have rights similar to
humans.
• Animal rights activists oppose “inhumane”
research practices, such as confining animals
to cages, subjecting them to electric shock,
depriving them of food, and so on.
• Beginning in the 1800s, researchers have
reacted to such criticism by developing
guidelines that weigh the benefits of research
against injurious or aversive nature of the
procedures. The first guidelines were
formulated in 1876, with the introduction of
the British Cruelty to Animals Act.
• Today, researchers in most professional
organizations, including the American
Psychological Association are regulated by
ethical standards that provide strict
guidelines for the care and use of animals.
CONGRUEN INCONGRUE
T NT

• RED • RED
• GREEN • GREEN
• BLUE • BLUE