ABAB Design Ethical considerations Is it ethical to
ABAB Design • Ethical considerations – Is it ethical to remove a treatment that appears to be beneficial (i. e. , implement the second “A” baseline stage)? – Dilemma between goal of understanding and goal of creating change.
Multiple-Baselines Design • The multiple-baselines design solves this ethical problem. – – Also uses a comparison between baseline and treatment Does not withdraw treatment establish several baselines implement treatment in one baseline at a time • Multiple baselines can be implemented – across situations • One individual observed in several situations – across individuals • Several individuals are observed – across behaviors • Several behaviors of one individual are measured
Multiple-Baselines Design • Comparing baseline to treatment – Observing individual(s) behavior(s) across time to establish baselines – Implement treatment when there is a steady baseline – Only implement one treatment at a time for a particular baseline • To interpret the treatment – If treatment is effective, behavior changes when the treatment is implemented and does not change for the remaining baselines. – Behavior changes only when the treatment is implemented and directly following the introduction of treatment.
Multiple-Baselines Design • Multiple baseline across situations – Example of Leslie with selective mutism – Would speak at home but not in other situations – Target several situations; restaurant, meeting adults, playing with peers – Start collecting baseline data in these situations • Several baselines started around the same time – Apply treatment to one situation at a time – In this case treatment was role playing with reinforcement at home
Multiple-Baselines Design • Multiple baselines across individuals • Example: Allison and Ayllon (1980) – evaluating the effectiveness of a behavioral coaching on athletics – procedures included systematic verbal feedback, positive and negative reinforcement – Establish baselines for each of the individuals • Start baselines around the same time – When the behavior has stabilized – An intervention is introduced one individual at a time
Multiple-Baselines Design
Multiple-Baselines Design Multiple baseline across behaviors Example: A schoolteacher is using a behavioral intervention that will help manage the behavior of a problem child in the classroom. The child does not stay at her desk when asked to do so, does not remain quiet during “quiet times, ” and exhibits other behaviors that disrupt the teaching environment. Use a positive reinforcer such as tokens or small toys in a multiple-baseline across behaviors design to improve the child’s behavior.
Multiple-Baselines Design • Multiple-baselines across behaviors: – Behavior #1: Leaving desk. – Implement treatment (reward system) on Day 3 – Behavior improves (decrease in frequency)
Multiple-Baselines Design • Multiple-baselines across behaviors: – Behavior #2: Blurt out questions without raising hand – Implement treatment (reward system) on Day 6 – Behavior improves (decrease in frequency)
Multiple-Baselines Design • Multiple-baselines across behaviors: – Behavior #3: Not quiet during quiet times – Implement treatment (reward system) on Day 8 – Behavior improves (decrease in frequency)
Multiple-Baselines Design • How many baselines are needed? – A minimum of two; three or four are recommended. • FIGURE 9. 3 Frequency of spoon-banging responses across baseline, treatment, and follow-up phases of study • What if behavior changes before the intervention? – The reasons for behavior change become hard to interpret; researchers analyze the situation to see if some aspect of treatment generalized. • What if the treatment generalizes to other behaviors, situations, or individuals? – Researchers should try to anticipate when generalization may occur and modify their research design accordingly.
Problems with All Single-Subject Designs • Problems with Baseline Records – Unstable baselines • increasing, or decreasing trends in behavior, the effects of treatment are hard to interpret. – Extreme variability in baseline behavior • difficult to detect a clear discontinuity in behavior when treatment is implemented. – Solutions: – Look for factors that may contribute to variability – wait for baseline behavior to stabilize – average baseline data points across observations.
Problems with All Single-Subject Designs Problems with Baseline Records
Problems with Single-Subject Designs • Baseline Records, – Whether increasing or decreasing baseline trends are a problem depends on the desired direction of behavior change. – Suppose the goal is to increase the frequency of a behavior. • If the baseline shows an increasing frequency of behavior, determining whether behavior increases following treatment will be difficult. • However, if the baseline shows a decreasing trend and treatment reverses this trend, we can be confident about the effect of the treatment.
Problems with Single-Subject Designs • External Validity – Single-subject designs are frequently criticized for their limited external validity. – Will treatment effects observed for one individual generalize to other individuals? – Reasons why external validity may not be limited: • Treatments are usually powerful. • Multiple-baselines designs can be used to demonstrate generality of effects. • Group treatment can be used to demonstrate effectiveness of treatment.
Quasi-Experimental Designs Chapter Ten • Goal: To improve the conditions in which people live and work. • Natural settings: Messy, “real world” — hard to establish experimental control. • Quasi-experiments: Experimental procedures that approximate the conditions of highly controlled laboratory experiments. • Program Evaluation: Applied research used to learn whether real-world treatments work.
Experiments in Natural Settings. • Researchers do experiments in natural settings for several reasons – Test the external validity of a laboratory finding – Attempt to improve conditions under which people live and work – Determining if a treatment is effective • make decisions about continuing the treatment • make decisions about spending money • make decisions about investing more time and effort • Natural Settings: such as schools, work place, government, hospitals, playground, etc.
Characteristics of True Experiments • True experiments have a high degree of control – manipulate an independent variable – random assignment to conditions – hold conditions constant • Allows comparison of treated to untreated groups • A true experiment is one that leads to an unambiguous outcome regarding what caused a result on the dependent variable. • However, this much control usually can not be done in a natural setting so there is less internal validity. – A trade off between better external validity and less internal validity
Differences between experiments in the lab and in natural settings • Experiments that are conducted outside the laboratory are likely to differ in a number of significant ways from those done in the laboratory. • Control in experiments from – manipulation of I. V. – random assignment of participants – hold other factors constant • External validity – Better in natural settings then the laboratory • Goals – basic laboratory research to understanding how things work will have very good internal validity – Applied research in natural settings tries to improve the lives of people and has internal validity • Consequences – Applied research in a natural setting can impact large numbers of people • For example: Head Start program, Sesame Street program • Sometimes referred to as “social experiments”
FIGURE 10. 1 As a social experiment, Sesame Street was designed to improve the education of hundreds of thousands of children.
Obstacles to Conducting True Experiments in Natural Settings • Obtaining permission and gaining access to participants – Can not get a random sample of participants • Random assignment of participants – Random assignment is the best way to determine if a new treatment really is effective – May not be able to use random assignment because of intact groups such as classrooms of children – Also can be viewed as unfair because some people who may need treatment don’t receive it. – A waiting-list control group may be used so that people randomly assigned to the control group receive treatment after the study is completed.
- Slides: 22