Small-N and Single

Download Report

Transcript Small-N and Single

Small-N and Single-subject
Designs
Susan Varni
Fall 2002
Small-N and Single-subject Designs
Where are we going?




Why Small-N Designs?
History of Small-N in Psychology
Variations on a Theme
Psychophysics
Why Are You Studying
Psychology?






Take 2 minutes to write down why you are
studying psychology.
What about psychology interests you?
What topics?
What profession do you aspire to?
What impassions you about psychology?
What do you hope to do with the
knowledge that you gain here?
Small-N Designs


Definition
Studies that use “several” participants with
data that is individually described and may
or may not be statistically analyzed
Pros of Small-N







Avoiding problems with the group mean*
Can examine participants from hard to find populations*
 *Key reasons currently!
 Following are both current & historical reasons
Can deal explicitly with individual (not group) behavior
Results are easy to interpret (often no stats!)
Avoids small, unimportant effects
Flexibility
Can focus on helping one (few) participant
Cons of Small-N



Hard to demonstrate causality
No controls in most cases
Lack of statistics





(Major “significance test” is the IOT)
Can’t really look at interaction effects
Counterbalancing is a problem
DV usually limited to response rates
Problem of external validity
History of Psychology and Small-N
Designs

Psychology really began with small-n



Ebbinghaus-nonsense syllables
Darwin-child development
Watson and Raynor-Little Albert
History of Psychology and Small-N
Designs

Wilhelm Wundt (1832-1920)


Leipzig Laboratory
“E” was the “S” and others were
“replications”
History of Psychology and Small-N
Designs

Edward Thorndike (1874-1949)

Animal Learning
“I then suggested [to James] experiments with the instinctive
and intelligent behavior of chicken as a topic, and this was
accepted. I kept these animals and conducted the experiments in
my room until the landlady’s protests were imperative. James
tried to get the few square feet required for me in the
laboratory…He was refused and with his habitual kindness and
devotion to underdogs and eccentric aspects of science,
harbored my chickens in the cellar of his own home for the rest
of the year” (Thorndike, 1936 as quoted in Hothersall, 1995)

Cats and Puzzle Boxes (results for every cat)
History of Psychology and Small-N
Designs

B.F. Skinner


Operant
Conditioning
Why are we
considering this
among small-n
designs?
Designs

Three Essential Elements



Operational definitions
Baselines (A)
Treatment (and continued monitoring) (B)
Baseline Designs


Once baseline has been maintained, it is
unlikely that a confounding event will
affect results on the same trial as the
manipulation
Design (A = baseline, B = treatment)
A
B
{ O1 X/O2 }
Phantom Pain (A-B)





Jonsson & Fisher, 1996
Phantom pain is hard to treat (don’t really
understand it) and methods have been
inconsistent
Hard to find participants
Studied one woman who has had both legs
amputated
Treated pain control by attention diversion
Phantom Limb Results
25
20
15
Duration
10
5
0
1
2
Treatment
3
4
Withdrawal Designs


It is unlikely that a confound will cease to
affect the results on the same trial the
manipulation is removed
It is more unlikely that a confound will reaffect the results on the same trial the
manipulation is re-introduced
Withdrawal Designs

Design (A = baseline, B = treatment)
1.
A
B
A
2.
A
B
A
B
3.
A
B
A
C
Cocaine Abstinence (A-B-A)




Silverman, Wong et al., 1998
Methods of keeping people in a methodone
treatment program from using cocaine
Used escalating reinforcement for cocaine-free
urine samples ($2.50 for first, add $2.96 for each
additional) (could get up to $1950 over the course
of the study)
Baseline, reinforcement, then removed
reinforcement
Results
80
70
60
50
40
Desire for Cocaine
30
20
10
6
3
10
7
4
1
4
1
0
Condom Taking (A-B-A-B)




Kirby et al., 1998
Drug abusers are at high-risk for HIV
Would prompting cocaine users result in
their taking (and hopefully using) more
condoms?
Treatment: After counseling, clients were
informed of risk and encouraged to take
condoms
Results
4.5
4
3.5
3
2.5
# condoms
2
1.5
1
25
20
18
15
12
9
6
4
1
0.5
0
Conversion Disorder (A-B-(A)-C)




Donahue, Thevenin, & Runyon, 1997
Conversion disorder: physical symptoms (often
following stressor) without organic cause
12-year old girl complained of throat constriction
(following choking on food)
Treated with (B) training meals (positive thoughts
while eating), and then with (C) reinforcement
(call brother, eat w/TV)
Results
91
90
89
88
87
86
85
84
83
82
81
B
1
2
3
Weight
C
4
5
6
7
8
9 10 11 12 13 14
Multiple Baseline Designs


If treatment start time is staggered, it is unlikely a
single confound can explain all perceived
treatment effects
3 Types



Same behavior in 2 or more individuals
2 or more different behaviors within the same
individual
Same behavior within the same individual but in 2 or
more different settings
Other Designs Mentioned in the Text

Changing criterion design (“shaping”)

A - B - C - B, where C is non-contingent reinf.

A - A1 - B - A1 - B (A1 a placebo for B)
Alternating treatments

 A-B-A-C-A-D-A
 A-D-A-C-A-B-A,
etc.
And now for something
completely different…


Although, not really different
Exercise


Melody Lateralization
Chromatic and luminance modulation
Psychophysics





Fechner, October 22, 1850  Elements of Psychophysics
(1860)
Profound insight about relation of physical stimuli () to
their mental representations ()
Concepts:
 Absolute threshold (min. energy to detect)
 Difference threshold (min. energy to notice change)
Methods (limits, constant stimuli, adjustment) are still in
use
Newer -- signal detection theory (SDT) separates our
sensitivity to stimuli from individual criteria for saying
“yes” or “no”