Ch23 - Ch23 Ch23 Doing Research in Behavior Modification...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Ch23 Ch23 Doing Research in Behavior Modification Doing Research in Behavior Doing Research in Behavior Modification Dependent variable Independent variable The behavior being measured The manipulation (treatment) Baseline Levels of behavior occurring prior to treatment How long to run baseline Stable steady (predictable) state of responding trend observed is in opposite direction of predicted effect of treatment Doing Research in Behavior Doing Research in Behavior Modification The reversal­replication (ABAB) research design How many reversals do I run? No easy answer As many as necessary to demonstrate experimental control Doing Research in Behavior Doing Research in Behavior Modification The reversal­replication (ABAB) research design ABAB designs have limitations If the target behavior is dangerous (SIB, suicidal ideation etc.) it would be unethical to withdraw a successful treatment Behavior trapping If the behavior has come under control of natural reinforcers (socialization training) Learning effects If I teach you how to play pool, I can’t take away what you’ve learned Doing Research in Behavior Doing Research in Behavior Modification Multiple­baseline design (p303) Used to demonstrate effectiveness without reversing a treatment condition Multiple baseline across behaviors Step 1­Concurrent baseline across multiple behaviors Step 2­Introduce treatment sequentially across behaviors MBL assumes behaviors are independent Doing Research in Behavior Doing Research in Behavior Modification Multiple baseline across situations (settings) examines effects of treatment on a single behavior that occurs in several situations Step 1­Concurrent baseline across multiple settings Step 2­Introduce treatment sequentially across settings If behavior changes in the second and third setting before treatment is implemented in those settings, you can not claim the change was due to your treatment Doing Research in Behavior Doing Research in Behavior Modification Multiple baseline across people (participants) Studies effects of treatment on a single behavior occurring in different people Essentially replicating effects across participants Step 1­Concurrent baseline across multiple participants Step 2­Introduce treatment sequentially across participants Doing Research in Behavior Doing Research in Behavior Modification Changing criterion design (p305) Introduce successive changes in the behavioral criterion for application of the treatment Take a baseline Set a criterion for reinforcement (15% above prior phase) So if baseline performance on rate of pedaling was 90 RPMs, to get reinforced in phase 1 they would have to have an average RPM of approximately 103 RPM, and in phase 2 118 RPM etc. Doing Research in Behavior Doing Research in Behavior Modification Alternating­treatments (multi­element) design (p307) Comparing effects of different treatments on a specific behavior with the same individual Involves alternating 2 or more treatments considerably more rapidly than in a reversal Caveats of using an alternating treatment Interaction effects (carry over) Bx contrast Generalization between conditions Doing Research in Behavior Doing Research in Behavior Modification Data Analysis and Interpretation Seven guidelines used You have more confidence in your treatment effect 1. 2. 3. 4. 5. 6. 7. More replications, more confident The fewer overlapping data points between baseline and treatment The sooner an effect is observed after implementing treatment The larger the effect compared to baseline The more precisely the treatment procedures are specified The more reliable your response measure The more consistent the findings are with existing data and Bx theory Doing Research in Behavior Doing Research in Behavior Modification Data Analysis and Interpretation No control groups used No statistical analyses We are concerned about the behaviors of individuals, not group averages Statistics hide sources of variability and outliers and that’s what we are concerned with Social validity To what extent are the identified target Bxs important to the client (are they really a problem) Acceptability of the treatment to the client (and parents) The “consumers” are satisfied with the results ...
View Full Document

This note was uploaded on 09/13/2011 for the course PHY 205 taught by Professor Barlly during the Spring '11 term at University of Nevada, Las Vegas.

Ask a homework question - tutors are online