Transcript It*s All Too Much - Bakersfield College

Single-Case
Experimental
Designs
Passer Chapter 12
Slides Prepared by Alison L. O’Malley
Single-Case Experimental Designs
•Researchers systematically examine how
an IV influences the behavior of an
individual case
•Remember Sperry and Gazzaniga’s
split-brain patients?
It’s only me!
Single-Case Experimental Designs:
Key Features
•Each participant serves as his/her own
control or comparison
• Treatment phase vs. baseline phase
Single-Case Experimental Designs:
Key Features
•DV measured across multiple trials
•Built-in replication
•Data examined separately for each
participant
•Reliance on visual analysis of data
Single-Case Experimental Designs:
Advantages
•Affords systematic understanding of low
base rate phenomena
•Flexible
•Individual differences are examined rather
than “averaged out”
What distinguishes single-case
experimental designs from case studies?
Single-Case Experimental Designs
•B.F. Skinner employed this research
approach in his operant conditioning work
• Advocate of experimental analysis of behavior
• Human applications include
as applied behavior analysis
and beyond
Single-Case Experimental Designs:
ABAB Designs
•Sequence of phases in which treatment is
either absent or present
•Typically (but not necessarily), treatment
absent during “A” phase and present
during “B” phase
•Often called withdrawal or reversal designs
as second A phase often entails withdrawal
of treatment
ABAB Experiment
Target
behavior
ABAB Single-Case Designs
•As before, we must consider how potential
confounds might threaten a study’s
internal validity
•What are some advantages of using an
ABAB design as opposed to an AB or ABA
design?
Drawbacks of ABAB Designs
• Ethical concerns surrounding withdrawal of
treatment that seems to be effective
• If target behavior does not revert to baseline
levels during a withdrawal phase, it is difficult
to draw conclusions
• Multiple-baseline designs are more useful if
we have reason to believe a target behavior
will not revert to baseline levels
Multiple Baseline Designs
•Behavior measured repeatedly during a
baseline period and then during a
treatment period
•Some sort of replication is conducted
• Timing of treatment is varied to rule out
alternative explanations
• But watch out for contamination!
Multiple Baseline Design Across
Subjects
•Two or more participants exposed to the
same treatment
•Introduction of treatment is staggered
across participants in order to create
baseline periods of varying lengths
Multiple Baseline Design Across
Behaviors
• Same treatment applied to two or more
behaviors of the same individual
• Introduction of treatment is staggered across
behaviors in order to create baseline periods of
varying lengths
• Describe how this approach could be
implemented in a study with three players of
your favorite sport or musical instrument
Multiple Baseline Design Across
Settings
•Same treatment applied to the same
behavior in two or more settings
•Introduction of treatment is staggered across
settings
•How (if at all) do multiple-baseline designs
resolve the ethical quandaries associated
with ABAB designs?
Changing-Criterion Designs
• Initial baseline phase followed by a treatment
phase
• When target behavior reaches a certain
threshold, a new performance criterion is set
Criterion 1
Criterion 2
…
Extending Single-Case Designs
• “Case” not limited to an individual
• Can be a larger social unit (e.g., business)
classroom)
Internal Validity of Single-Case Designs
•Unlike group within-subjects designs, singlecase designs are rarely counterbalanced
•What are the implications of this for
internal validity?
Data Analysis
•Primary reliance on visual data analysis
•No standard application of statistical tests
•Difficult to examine interactions
•What would an ideal set of data look like?
External Validity
•Why are concerns about external validity
often raised with respect to single-case
experiments?
•How can single-case experimenters defend
themselves against these claims of limited
external validity?
Group vs. Single-Case Designs
A classmate maintains that single-case
experimental designs are far superior to
between- and within-subjects designs
because group designs analyze “averaged”
responses.
Respond to this argument.