Experimental Design Flashcards
Experimental Control
When a predictable change in behavior DV can be reliably produced by the systematic manipulation of aspect in environment
Behavior is
Individual
Continous
Determined-by functional relations
Extrinsic-variable
6 Components of Experiments in ABA
At least one SUBJECT At least one BEHAVIOR At least one SETTING At least one TREATMENT A measurement SYSTEM An EXPERIMENTAL DESIGN
Experimental Design
Brief but specific statement of what researchers want to learn from conducting experiments
At least one SUBJECT
ABA uses single-subject designs
Single subject acts as control
At least one BEHAVIOR
Some have more than one DV
Multiple DV can show data patterns serving as controls to evaluate and replicate effects on IV
Assess presents of IV’s effect on behaviors
At least one SETTING
Control 2 sets of environment variables to demonstrate experimental control; IV and extraneous variables
Hard to control environment in natural settings
At least one TREATMENT
Particular aspect of the environment which is manipulates to find affects on behavior
AKA for At least one SUBJECT
Single Case Design
Within Subject Design
Intra Subject Design
AKA for At least one BEHAVIOR
DV
AKA for At least one TREATMENT
IV
Intervention
Experimental Variable
A Measurement System and Ongoing Analysis of Data
Observation and recording procedures must be conducted standardized
Must detect changes in level, trend, and variability
An Experimental Design
The particular arrangement of conditions in a study so that a meaningful comparison of effects of presence, absence or different values of the IV can be made
2 Types of Experimental Design
Nonparametric Analysis
Parametric Analysis
Nonparametric Analysis
IV either present or absent during study
Parametric Analysis
Value of IV is manipulated. Seeks to discover for the differential effects of a range of values
Treatment Package
When multiple IV’s are bundled into one program such as a token economy plus praises plus time-out
Component Analysis
Looks for effect of each part of the treatment package
Steady State Responding
A pattern of responding that exhibits very little variation in its measured dimensional quantities over a period of time
Baseline Logic
Refers to experimental reasoning inherent in single-subject experimental designs
P- Prediction
V- Verification
R- Replication
Steady State Strategy
Repeated exposures of a given subject to a condition while trying to eliminate extraneous influences on behavior by obtaining a stable pattern of responding before introducing the next condition
Function of Baseline Data
Control Condition
Does not imply the absence of intervention
4 Patterns of Baseline Data
DAVS Descending Ascending Variable Stable
Descending Baseline
Shows behaviors is already changing
Generally should not implement IV when baseline is descending unless you want behavior to increase
Ascending Baseline
Shows behavior is changing
Should not implement IV when baselines is ascending unless you want the behavior to decrease
Variable Baseline
No clear trend
Wait it out because its due to change in environment
Stable Baseline
No evidence of ascend or descend trend
Best way to look at effects of IV on DV
3 Parts of Baseline Logic
PVR
Prediction
Verification
Replication
Prediction
Anticipate outcome of unknown measurement
Data should be collected until stability is clear
More points better
Verification
A previously predicted level of baseline responding by termination or withdrawal of the treatment variable
Replication
Is the essence of believability
Shows reliability
Replication accomplished by reintroducing the IV
5 Main Experimental Designs
MCRAW Multiple Baseline Changing Criterion Reversal Alternating Treatments Withdrawal
Multiple Baseline
Most widely used
Flexible
Staggered implementation of the intervention in a step-wise fashion across behaviors, settings, and subjects
Use when reversing is unethical
PVR in Multiple Baseline
A functional relation requires a change in behavior with the onset of the intervention
Apply IV to Bx1 when you can predict Bx will stay the same
If Bx 1 and 2 remain unchanged this verifies
If do change then its been replicated
Most commonly 3-5 tiers
Multiple Baseline Across Bx
Two or more Bx of same subject
After steady baseline IV is applied
When steady state of responding to IV is reached then IV is applied to the next Bx
Multiple Baseline Across Settings
Single Bx two or more settings
Steady baseline- IV applied to first setting
When steady response then IV is applied to next setting
Multiple Baseline Across Subjects
One target Bx for two or more subjects in the same setting
Steady baseline- IV applied to first subject
When steady response IV is applied to next subject
2 Variations of Multiple Baseline
Multiple Probe Design
Delayed Multiple Baseline
Inherently weaker
Multiple Probe Design
Analyze relations b/w IV and acquisition skill sequence
Instead of baselines, probes provide basis for determining behavior change has occurred before intervention
Delayed Multiple Design
Initial baseline and intervention begin and following baselines are added or delayed
Effective when reversal design is not possible, limited resources, new behavior, subject or setting
Shorter baselines do not show interdependence of DV’s
Guide for Multiple Baseline Design
Select independent similar baselines
Select concurrent and related multiple baselines
Do not apply IV to next behavior too soon
Vary significantly lengths of baselines
Intervene on the most stable baseline first
Advantages of Multiple Baseline Design
Successful intervention do not have to be removed
Evaluates generalization
Easy to implement
Disadvantages of Multiple Baseline Design
Functional relationship is not shown
Effectiveness of IV is demonstrated but information regarding function of Bx
IV may be delayed for certain Bx, settings, subjects
Takes resources
Changing Criterion Design
Experimental design where initial baseline is followed by;
Series of treatment phases consisting of gradually changing criteria for reinforcement or punishment
Only one Bx
Bx needs to be in subjects rep.
Evaluates step like treatment applied gradually
Technically variation of multiple baseline
PVR for Changing Criterion Design
Graphs in changing criterion should have lines separating a lot to show a functional relationship
Experimental control is evidenced by extent the level of responding changes to conform to each new criterion
If data do not fall around lines- little control
The greater vertical distance between lines the more control
Guidelines for Changing Criterion Design
Length of phases
Magnitude of criterion changes
Number of criterion changes
The more criterion changes the better proof of experimental control
Advantages for Changing Criterion Design
Does not require reversal
Enables experimental analysis within context
Disadvantages for Changing Criterion Design
Target Bx must already be in rep
Not appropriate for analyzing effects of a shaping program
No comparison design
Reversal Design
An experimental design in which the responding is reversed to level obtained in previous condition
IV is withdrawn or reversed
Alternation between baseline and particular intervention
Each reversal strengthens experimental control
3 Consecutive Phases of Reversal Design
Initial Baseline A
Intervention B
Return the Baseline A
PVR in Reversal Design
Involves PVR
IV is responsible for behavior change if repetition of baseline and treatment approximate the original phases
Solid points- actual measure
Open points- predicted data
Data in shade box- verifies prediction
Data in cross hatched box- data replicates
5 Variations of ABAB Design
Repeated Reversals BAB Reversals Multiple Treatment Design NCR Reversal Technique DRO/DRI/DRA Reversal Technique
Repeated Reversals
Simple extension of ABAB
More reversals the stronger evidence
BAB Reversals
3 Phases IV IV removed IV reintroduced No baseline Best design when client displays severe behaviors
Sequence Effects
Effects on a subjects behavior in a given condition that are the result of the subjects experience with prior condition
Multiple Treatment Reversal
A type of reversal design that compares two or more IV’s compared to baseline and/or to each other
Can cause sequence effects
NCR Reversal Technique
An experimental technique for showing the effects of reinforcement by using NCR as a control instead of baseline
Reinforcement presented of fixed or variable schedule
DRO/DRI/DRA Reversal Technique
Showing effects of reinforcement by DRO, DRA,DRI as a control
DRO- Following any behavior other than target
DRI- Any behavior incompatible
DRA- Any alternative behavior
Irreversibility
The level of behavior observed in an earlier phase that cannot reproduced even though experimental conditions are the same as they were.
Alternating Treatments Design
A design in which 2 or more conditions are presented in rapidly alternating succession independent of the level of responding and effects on Bx
Compares to IV’s to one another
Based on SD
AKA’s for Alternating Treatments Design
SCAMMM Simultaneous Treatment Design Concurrent Schedules Design Alternating Treatments Design Multi-Element Baseline Design Multi-Element Design Multiple Schedules Design
PVR in Alternating Treatments Design
Inspection of differences between data paths
Functional relation shown when one data point is higher than another and no overlapping
Degree of differential effects produced by 2 treatments determined by vertical distance between data paths
Each point plays all three roles of PVR
3 Variations of Alternating Treatments Design
Single Phase w/o Baseline
With Baseline
With Baseline and Final Best Treatment Phase
Problems Avoided by Alternating Treatments Design
Irreversibility Sequence Effects Unstable Data No treatment withdrawal Can begin immediately
Disadvantages of Alternating Treatments Design
Multiple treatment interference
Limited capacity
Treatments should be different
Withdrawal Design
Synonymous with reversal design
2 Types of Validity
Internal
External
Internal Validity
The extent to which an experiments show changes in behavior are a function of the IV and not uncontrolled variables
4 Confounding Threats to Internal Validity
Measurement Confounds
IV Confounds
Subject Confounds
Setting Confounds
Measurement Confounds
Number and intricacy of behaviors targeting May occur due to; Observer drift Reactivity Observer bias
IV Confounds
IV’s are complicated
Can reduce by placebo or double blind control
Subject Confounds
Maturation
Setting Confounds
Studies in natural settings are prone to confounding
Bootleg reinforcement can occur
Confounding Variables
Uncontrolled factor known or suspected to exert influence on the DV
Extraneous Variables
An aspect of ENVIRONMENT that must be held constant to prevent unplanned environmental variation
External Validity
Study’s results are generalizable to other subjects, settings, and or behaviors
Functional relation discovered should hold under different conditions
Replication established external validity
Treatment Integrity
IV is implemented and carried out as planned
2 Types of Errors Evaluating ABA Research
Type I- False Positive
Type II- False Negative
Visual analysis used in ABA tends to lead more to Type II errors
