Ch. 3 Flashcards
experimental designs
turn IV ON and OFF, identify functional variables to help positively influence behavior
2 kinds of exp. designs
1) group
-evaluate if behavior of treatment group is statically significant from that of control group
–difference is attributed to IV
2) single-subject
-expose individual to baseline and experimental phases to determine if IV systematically and reliably changes behavior
4 group design weaknesses
1) when IV is a therapeutic intervention, no one wants to be assigned to control group
2) not studying behavior of individual
3) behavior of treatment and control groups will differ simply because the people assigned to 2 groups are different
4) reliance on inferential statistics to evaluate if IV changed behavior
internal validity
experiment gives evidence that a functional relation exists between the IV and behavior change
confounds
variables that influence behavior within an experiment, but not controlled by researcher
4 single-subject designs
1) comparison (A-B)
2) reversal (A-B-A-B)
3) alternating treatments
4) multiple-baseline
comparison (A-B) design
aka quasi-experimental design, quasi because a well-designed experiment rules out confounds
-arranges baseline (A) phase and experimental (B) phase
-can’t rule out confounds, if confound is responsible, turning OFF the intervention will not have influence
stable
over repeated observations, there is little “bounce” (variability in behavior from one session to the next) and no systematic trend in behavior
reversal (A-B-A) design
-behavior is evaluated in repeatedly alternating baseline (A) and experimental (B) phases
-turning on and off and back ON the intervention allows us to replicate the intervention effect with the one person
-data not averaged
alternating treatments design
independent variable(s) is turned ON and OFF rapidly to evaluate if this systematically and repeatedly changes behavior
-2+ phases rapidly alternated in random order
-commonly used when trying to understand why problem behavior occurs
-establishes internal validity by repeatedly turning ON and OFF the IV and evaluating if this systematically and repeatedly changes behavior
reversal and alternating treatments designs establish internal validity by?
a) show that behavior is systematically and repeatedly changed when the IV is turned ON and OFF
-impossible if IV produces lasting effect that can’t be reversed when intervention is turned OFF
b) rule out confounds that are correlated w/ IV
-if a) and b) do not apply, use multiple-baseline design
standard deviation (SD)
average diff. between each data point and average of all data points in a condition
multiple-baseline design
evaluates functional relation between IV and behavior by a series of time-staggered A-B comparisons either across behaviors, situations, or people
used when intervention is expected to produce an irreversible effect, or it would be unethical to remove an effective intervention for an extended period of time
-series of A-B comparisons conducted at time lags across various behaviors, people, or settings
multiple-baseline across-behaviors design
time-staggered A-B replications are conducted across behaviors
multiple-baseline across-situations design
time-staggered A-B replications are demonstrated across situations
multiple-baseline across-participants design
time-staggered A-B replications are demonstrated across participants
4 features of single-subject designs
1) focus on behavior of individual
2) individual receives intervention
-make single-subjects useful for clinical behavior analysts who work with 1 patient at a time
3) behavior is repeatedly measured in each phase until we predict what’ll happen if nothing changes
4) evaluate internal validity by investigating if IV produces systematic and replicable effects on behavior
3 kinds of replications built into single-subject exp. designs
1) within-individual
-behavior of person is repeatedly observed after IV is turned ON
2) across-individual replication
-evaluate if effects of an IV can systematically and reliably influence behavior of 1+ individual
3) replication across labs or clinics
inferential statistics
used to decide if behavior changed when the IV was manipulated
-require big # of participants, not designed for time-series data like those graphed
approaches to address replication crisis
1) transparency to precent scientists from using “objective decision rules” that gives them the result they want
-ask researchers to publish in advance how they’ll conduct experiment and how they’ll analyze their data
2) visual analysis of behavior
-looking at graph of time-series single-subject behavior to evaluate if convincing change occurred when IV was introduced or removed
2 advantages of visual analysis over inferential statistics
1) requires that behavior analyst develop methods for producing big changes in behavior
2) recognizes subjectivity of scientific enterprise
-doesn’t attempt to hide behind “objective decision-rules” that are used by those prone to errors, biases, and prejudices
question to answer when conducting visual analysis
did behavior change when IV changed?
2 ways in which behavior can change from no-intervention baseline
1) trend
-systematic change in behavior over time
2) level
-prevalence of behavior during stable portion of phase or condition
least-squares linear regression
predicts what’ll happen to behavior in the future if nothing is changed
-extends arrow to right, beyond baseline, into future sessions
3 steps to guide visual analysis
1) draw trend arrow through baseline data to predict what’ll happen if IV isn’t turned on
2) evaluate if behavior in baseline is too variably (bouncy) to have confidence in prediction of trend arrow
3) draw trend or level lines through intervention data
-evaluate if there is a convincing change in trend or level (whichever change is of interest)
stimulation modeling analysis
made for single-subject time-series data and can be conducted w/ freely available software
-tests for changes in trend and level as supplement for visual analysis
