Single Subject
Designs
Adv. Experimental
Methods & Statistics
PSYC 4310 / COGS 6310
Michael J. Kalsher
Department of
Cognitive Science
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2012, Michael Kalsher
1
Outline
• Introduction to Single-Subject
Methodology
• Applied Behavior Analysis
– Single-Subject Designs
• Reversal Design
• Multiple Baseline Design
• Optional outside assignment
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
2
Single-Subject Research Designs
Most often used in applied fields of psychology,
education, and human behavior in which the subject
serves as his/her own control, rather than using
another individual/group.
Sensitive to individual organism differences as
compared to group designs which are sensitive to
group averages.
Useful for evaluating the effects of a variety of
interventions in applied research.
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
3
Single-Subject Research Designs
Performance
Especially useful in situations where averaging together
the performance of individuals might produce a
misleading picture of the phenomenon under
investigation.
Time
Example: Averaging the data of a group of participants in a one-trial learning experiment might produce a
smooth learning curve like the one above. An examination of each participant’s data would instead show an
abrupt change from no learning to 100% performance, but at different points in time during the experiment.
The learning curve presented here thus provides a false picture of how learning is actually taking place.
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
4
The Logic of Single-Subject Experiments
Conventional experimental designs
• Are strategies for dealing with unwanted variation in performance.
• Attempt to eliminate all sources of systematic variation other than
that produced by the experimental manipulation and then use
design techniques (random assignment) and statistical techniques
to filter out sources of unsystematic variance.
Single-subject experimental designs
• Address the signal-noise problem by identifying the sources of
uncontrolled variance and then controlling these variables in the
experiment.
• Sometimes termed “steady-state methodologies” because they
compare the effects of an experimental manipulation to some
baseline steady state in the same individual.
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
5
Applied Behavior Analysis:
Putting Single-Subject Methodology to Work
Defined as the science of applying experimentally derived
principles of behavior, including single-subject methodologies,
to improve socially significant behavior. Behaviors are defined
in observable and measurable terms in order to assess change
over time. The behavior is analyzed within the environment to
determine what factors are influencing the behavior.
ABA contributes to solving important real-world problems in a
number of domains, including: developmental disabilities,
mental illness, education and special education, rehabilitation,
community psychology, clinical psychology, business and
industry, self-management, child management, sports
psychology, health psychology, and gerontology.
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
6
Behavioral Principles:
Functional definitions
Reinforcement: Procedures that result in an
increased probability of responding.
Punishment: Procedures that result in a decreased
probability of responding (usually temporary).
Apply vs. Remove; Appetitive vs. Aversive
Positive reinforcement ($ for weight loss)
Negative reinforcement (Nag until he picks up his socks)
Positive punishment (Speeding ticket, scoldings)
Negative punishment (take the keys to the family car)
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
7
A-B
One baseline and one treatment phase. Not a
true research design .
A-B-A-B Reversal
Two (or more) baseline phases for the same
Design
behavior of one subject.
Multiple-Baseline
Baseline and treatment phases for two or more
(Across Behaviors) different behaviors of one subject. Treatment is
staggered across behaviors .
Multiple-Baseline
Baseline and treatment phases for the same
(Across Subjects ) behavior of two or more subjects. Treatment is
staggered across subjects .
Multiple-Baseline
Baseline and treatment phases for the same
(Across Settings )
behavior of the same subject in two or more
settings. Treatment is staggered across
setting s .
Changing Criterion A baseline phase and treatment phase for one
Design
subject. In the treatment phase, there are
progressive performance criteria or increasing
goal levels of the behavior .
A Summary of Single-Subject
Research Designs
Source: Miltenberger, R.G. (2001). Behavior modification: Principles and procedures. Belmont,
CA: Wadsworth.
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
8
A-B Design:
Decreasing Smoking
Number of cigarettes smoked during baseline and following an intervention in
which a person agreed to smoke one fewer cigarette per day every second day.
Source: Miltenberger, R.G. (2001). Behavior modification: Principles and procedures, 2nd ed. Belmont, CA: Wadsworth.
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
9
A-B Design:
Increasing Time Spent Studying
Hours of studying during baseline and following an intervention in which a a behavioral
contract was used to increase the amount of time a student spent studying.
Source: Miltenberger, R.G. (2001). Behavior modification: Principles and procedures, 2nd ed. Belmont, CA: Wadsworth.
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
10
A-B-A-B Design
The effect of a teacher’s demands on a mentally retarded adolescent’s aggression. The observed
pattern of behavior changes makes it unlikely that any other variable caused the behavior to change.
Source: Miltenberger, R.G. (2001). Behavior modification: Principles and procedures, 2nd ed. Belmont, CA: Wadsworth.
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
11
Multiple-Baseline
Across Behaviors
The occurrence of four different
social behaviors exhibited by a shy
adolescent during baseline and
following a social skills training
program. Note that behavior
change occurs only after the
intervention is applied.
Source: Miltenberger, R.G. (2001). Behavior modification:
Principles and procedures, 2nd ed. Belmont, CA: Wadsworth.
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
12
Multiple-Baseline
Across Subjects
The percentage of time that four
ER nurses wear protective
gloves when examining patients.
The intervention, comprised of
feedback from their supervisor, is
staggered over time and results
in an increase in the glovewearing behavior for each of the
four nurses. Note that behavior
change occurs only after the
intervention is applied.
Source: Miltenberger, R.G. (2001). Behavior modification:
Principles and procedures, 2nd ed. Belmont, CA:
Wadsworth.
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
13
Multiple-Baseline
Across Settings
Percentage of intervals of disruptive behavior by a student during baseline and treatment (a
revised curriculum) in two settings: a morning class (a.m.) and an afternoon class (p.m.).
Note that behavior change occurs only after the intervention is applied.
Source: Miltenberger, R.G. (2001). Behavior modification: Principles and procedures, 2nd ed. Belmont, CA: Wadsworth.
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
14
Changing Criterion Design:
Decreasing caffeine consumption
Caffeine consumption decreased to a level below the criterion each time the criterion was lowered.
Source: Miltenberger, R.G. (2001). Behavior modification: Principles and procedures, 2nd ed. Belmont, CA: Wadsworth.
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
15
Applying Single-Subject Methods to LargeScale Systems: Increasing Safety-Belt Use
Media campaigns employing incentives or disincentives to promote
safety belt use were studied by applying an ABA reversal design at two
large naval bases with safety belt use mandates in effect. The base
police delivered each type of intervention by issuing warning tickets for
the disincentive program or by entering the license plate numbers of
vehicles with drivers buckled up in a prize drawing for the incentive
program.
At one navy base (daily traffic volume = 75,000 vehicles), use of safety
belts increased from a baseline of 51% (33,173 observations) to 61.%
(28,517 observations ) during the 4-week incentive phase. At the
other base (daily traffic volume = 46,000 vehicles), the baseline belt
use of 55% (17,221 observations), increased to 79% (18,305
observations) during the 3-week disincentive phase.
Kalsher et al. (1989). Safety belt promotion on a naval base: A comparison of
incentives vs. disincentives. Journal of Safety Research, 20, 103-113.
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
16
Promotional Materials:
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
Incentive Program
17
Promotional Materials:
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
Disincentive Program
18
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
19
Applying Single-Subject Methods to
Increase Recycling on a University Campus
Rapid declines in available landfill space have sparked interest in recycling
programs focused on increasing the quantity of materials recycled. The focus
on quantity, rather than quality, has produced a glut of paper which currently
comprises nearly 40% of all solid waste produced each year in the U.S.
Recycling facilities are selective with regard to the paper materials they accept
because contaminants reduce the recyclability of the collected materials.
A “low-tech” sort separation intervention was used to decrease contaminants in
recyclable paper collected from four campus buildings. A multiple baseline
design across settings was used to evaluate the intervention, that consisted of
an educational pamphlet, interactive group discussion, attention-getting posters,
and environmental arrangement of color-coded paper collection bins. Following
the intervention, the percentage of correctly sorted paper increased from 25.5%
during baseline to 83.5%. The results show that “low-tech” interventions can
be used successfully to improve paper recycling practices in office settings.
Kalsher et al. (1989). Promoting recycling behavior in office environments. Proceedings
of the Human Factors and Ergonomics Society, 37, 484-488.
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
20
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
21
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
22
Applying Single-Subject Methods
to Treat Chronic Pain
This study assessed the covariation of behaviors associated with chronic pain
within and across behavioral and drug approaches to treatment. Problems of
screaming and five other behaviors were measured across conditions of varying
behavioral contingencies and varying administration (time since medication and
dosage) of Parsidol during attempts to treat the muscle pain of a 24-year-old
male with dystonia musculorum deformans.
Results indicated that: (a) pain behaviors covaried during behavioral and drug
conditions even though the behavioral intervention only targeted screaming; (b)
effects were greater on non-targeted behaviors during periods that followed
rather than preceded drug administration; and (c) in contrast to behavioral
observation data, physiological measures of neuromuscular activity (EMG) did
not differ across conditions. These results suggest that functional responseresponse relationships exist in patients as the result of their illness experience.
Kalsher et al. (1985). Behavioral covariation in the treatment of chronic pain. Journal
of Behavior Therapy & Experimental Psychiatry, 16, 331-339..
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
23
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
24
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
25
Applying Single-Subject Methods
to Treat Self-Injurious Behavior
This research systematically investigated environmental correlates of selfinjurious behavior (SIB) which were later used as the basis for treatment.
Specifically, developmentally disabled subjects were exposed to a series of
conditions designed to identify factors that maintain SIB: attention contingent on
self-injurious behavior (positive reinforcement), escape from or avoidance of
demands contingent on SIB (negative reinforcement), along (automatic
reinforcement), and play (control).
Results showed that each subject’s SIB occurred more frequently in the
demand condition, suggesting that the behavior served an avoidance or escape
function. Treatment was comprised of extinction plus reinforcement for
tolerance of medical examination procedures that would accompany
ophthalmologic surgery to reattach retinal tissue destroyed by the subject’s selfinjury. Results showed that the treatment was successful in eliminating SIB and
that its effects transferred across eight new therapists and three physicians.
Iwata et al. (1990). Experimental analysis and extinction of self-injurious escape
behavior. Journal of Applied Behavior Analysis, 23, 11-27.
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
26
Functional Assessment:
Identifying the causes of behavior
Percentage of 10-sec intervals of self-injurious behavior
as a function of environmental conditions.
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
27
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
28
Applying Single-Subject Methods to Investigate
Environmental Determinants of Alcohol Consumption
Two field experiments were conducted to examine
situational determinants of alcohol consumption at
university fraternity parties. During these parties,
individual drinking rates of beer and mixed drinks, or
beer only, were monitored under varying environmental
conditions. When exiting the parties, students’ blood
alcohol concentrations (BACs) were obtained with a
breathalyzer. The purpose of the research was to
assess the impact of environmental features on
drinking behavior and subsequent intoxication.
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
29
Experiment 1:
Assessing the influence of
labeling practices on drinking
In the first experiment, participants
attended two different parties.
At the first party, three types of beer
(Budweiser, Bud Light, and LA)
were available in kegs labeled “A”,
“B”, and “C”.
At the second party, the three kegs
were labeled according to beer
content (i.e., Budweiser, Bud Light,
and LA).
Results showed significant
differences in drink choice across
parties, depending on how the beer
kegs were labeled.
Blind taste-test results
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
30
Students’ beverage choices and drinking rates differed
as a function of labeling condition.
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
31
Experiment 2:
Assessing the influence of drink
delivery format
In Experiment 2, the alcoholic beverages available to student of legaldrinking age at the party (mixed drinks and beer) were served either by
bartenders or served by themselves.
Male and female beer drinkers assigned to the “Self-Serve” condition
drank at a higher rate and consumed more of their preferred beverage
type than did those drinkers served by a bartender, or by those
consuming mixed drinks in the self-serve condition. This increase was
highest for male partiers.
Male and female mixed drink consumers assigned to the Self-Serve
condition drank at the lowest rate and consumed the least amount of
their preferred beverage type.
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
32
Experiment 2:
Assessing the influence of drink
delivery format
M = .119
M = .105
M = .088
M = .067
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher
33
Optional Assignment
Identify a behavior, or behaviors, that you would like to
change in yourself, or another person.
Design a measurement system and assess the baseline
rate of the behavior for at least one week, then implement
an intervention based on reinforcement, NOT punishment
principles.
Continue to collect data on the behavior to assess any
changes. Finally, write-up the “study” and submit it for
grading. It should contain an introduction to the problem,
a method section, a results and discussion section, with
relevant graphs, and at least three references from peerreviewed journals.
34
PSYC 4310/6310
Advanced Experimental Methods and Statistics
© 2011, Michael Kalsher