single subject and group research design
TRANSCRIPT
Single Subject and GroupResearch Design
Terminology
Baseline
Data collected before the start of theintervention, change or experiment
Variability
Frequency and extent to which multiplesmeasures of behavior yield differentoutcomes
A high degree of variability within a givencondition usually indicates that the researcherhas achieved little control over the factorsinfluencing the behavior
Dependent Variable
Target behavior Observed to change in the presence of
the independent variables
Independent Variable
Used to change the targetbehavior/dependent variable
We aim to control the dependent variable byour use of the independent variable
E.g.: If I threw a $5 at you every time youraised your hand, you would probably raiseyour hand frequently. Raising your hand =dependent variable, use of money =independent variable
Changing one variable at atime… Allows us to accurately determine the
variable that is the cause of the change If we change more than 1 variable we
will not be able to say with certaintywhat caused our results
Reversal Design
Repeated measurements of a behavior in agiven setting that requires at least threeconsecutive phases– An initial baseline where the independent variable
is not present– An intervention phase where independent variable
is presented– A return to baseline conditions by removing the
independent variable– (money example)
Irreversible procedure
Some independent variables cannot beremoved
E.g.: Once we have taught a child tosay the word “Help”, we cannot removethe child’s ability to say the word “Help”
In this type of situation we modify ourresearch design
Multiple baseline design
Most widely used Allows researchers to analyze the effects of
the independent variable across multiplebehaviors, settings, subjects
Does not use withdrawal of treatment Used when target behavior is irreversible When withdrawal is undesirable, impractical
or unethical
Validity
Measurement has validity when it yieldsdata that are distinctly relevant to thephenomenon measured and to thereason for measuring it
Variability
Frequency and extent to which multiplesmeasures of behavior yield differentoutcomes
A high degree of variability within a givencondition usually indicates that the researcherhas achieved little control over the factorsinfluencing the behavior
Reliability
A consistent measurement
Clinical Significance
A result that is clinically significant isalso unlikely to have occurred bychance
But a result that is clinically significant isone where the change is of benefit tothe student
The Purpose of Research in BasicScience The discovery of functional relations
among dependent and independentvariables
This knowledge of functionalrelationships become principles thatadd to our knowledge of behavior
The Purpose of Research inApplied Science In ABA the focus is on improving the
lives of individuals through changingproblematic behaviors and increasingappropriate, beneficial behaviors
E.g. Teaching a child to request withPECS may eliminate problematicbehaviors such as self-injury ortantrums.
Example:
Rate of self-injury in student isdecreased from 200x an hour to 150xan hour– Statistically significant
Rate of self-injury in student isdecreased from 200x an hour to 5x anhour– Clinically significant
Group Research Design
Uses a pool of subjects from thepopulation - randomly assigned togroups: experimental group and controlgroup
Results are focused on generality ofresults to general population
Pretest of dependent variable
Single Subject Research Design
Also referred to as single case design Not always 1 subject (four to eight common) Repeated measure of each subject’s behavior are
obtained as she is exposed to each condition of thestudent
Subject often exposed to each conditions severaltimes over the course of an experiment
Measures of the subjects’ behavior during eachphase of the study provide the basis for comparingthe effects of experimental variables as they arepresented or withdrawn in subsequent conditions
Statistical tests may be used butfocus is on socially or clinicallysignificant results - observable,measurable changes; differenttests used
Statistical tests used to rule out“chance”
Data is collected throughoutexperiment
Comparison of pre and postmeasurements
All participants receive treatmentExperimental group receivestreatment, control group doesnot
Results of individual assessedData results are averaged fromall subjects
Measurement of dependent variableconstant throughout study
Initial pretest of dependentvariable done on all subjects
Not randomly assigned forethical reasons
Randomly assigned to group
Specific profiles to suit studySample from population
SingleGroup
So which one is best?
Not a matter of one being better than theother
Rather it is an issue of which is best suited tothe discipline
Behavior analysis is a science focused on theunderstanding and controlling of thebehaviors of the individual
Therefore we primarily use single subjectresearch, with a focus on the repetition ofresults to achieve socially significant behaviorchanges