CHAPTER 12: Single-Case Experimental Designs1

Chapter 12: Single-Case Experimental Designs

A.LEARNING OUTCOMES. After studying this chapter students should be able to:

  • Describe the key features of single-case experimentation and its advantages.
  • Discuss the use of single-case designs in basic and applied research.
  • Describe the key features of ABAB, multiple-baseline, and changing-criterion designs.
  • Explain how each design attempts to rule out alternative causal explanations.
  • Identify strengths and limitations of each design.
  • Explain several ways in which single-case designs are applied to social units.
  • Discuss ethical issues that arise when using single-case designs.
  • Describe concerns regarding the internal and external validity of single-case designs.
  • Identify key issues pertaining to the data analysis and examination of interactions in single-case designs.

B.KEYWORDS

ABAB design / Multiple-baseline design across settings
Applied behavior analysis / Multiple-baseline design across subjects
Baseline phase / Psychophysics
Changing-criterion design / Single-case experimental designs
Experimental analysis of behavior / Target behavior
Multiple-baseline design / Treatment phase
Multiple-baseline design across behaviors / Withdrawal design

C.BRIEF CHAPTER OUTLINE

I. A Single-Case Approach to Experimentation

  1. Key Features
  2. Advantages
  3. Use in Basic and Applied Research

II.Types of Single-Case Designs

  1. ABAB Designs
  2. Multiple-Baseline Designs
  3. Changing-Criterion Designs
  4. Combined Designs

III.Extending Single-Case Designs to Social Units

  1. Buckle Up! A Withdrawal Design
  2. Slow Down! A Multiple-Baseline Design

IV.Issues Concerning Single-Case Designs

  1. Ethical Issues
  2. Internal Validity
  3. Data Analysis
  4. Examining Interactions
  5. External Validity

D. EXTENDED CHAPTER OUTLINE

*Much of this summary is taken verbatim from the text.

Introduction

This chapter is about single-case experimental designs, whichallow researchers to systematically examine how an independent variable influences the behavior of an individual case.

Part I: A Single-Case Approach to Experimentation

  1. Key features. Single-case approaches typically involve only one participant, but some involve more than one. What makes a single-case approach unique is that the researcher examines the results individually for each case and draws conclusions about how an independent variable influenced it. According to Barlow, Nock,Hersen (2009), most single-case designs have the following features:
  2. Each participant serves as his or her own control. This is achieved by measuring participants during baseline(control)and treatment(experimental) phases.
  3. The dependent variable is measured across multiple trials or sessions within each condition of the experiment.
  4. The research design includes built-in replicationsthat enable the scientist to assess whether the independent variable produces a reliable change in behavior.
  5. The data are examined separately for each participant because the researcher wants to know how the independent variable affects that individual individually.
  6. A visual analysis of the data may be the sole basis for judging whether the independent variable produced an effect.This is achieved by plotting the participant’s performance on a figure to illustrate the pattern of results over time under each phase of the experiment.
  7. Advantages. One of the biggest advantages of the single-case design is that they enable a researcher to conduct an experiment on a single participant. In addition, because most single-case studies employ only one individual it affords the researcher flexibility to alter the treatment as needed. Lastly, the single-case design eliminates the possibility of averaged results from a large group of subjects to mask an effect of the experimental treatment.
  8. Use in basic and applied research. The earliest psychological experiments were single-case designs. The development of statistical techniques ultimately led to studying how independent variables affected groups of participants.
  9. The experimental analysis of behavior is the intensive study, under controlled situations, of how environmental stimuli and consequences regulate an individual organism’s behavior. Skinner relied upon this type of research method to elucidate the principles of operant learning using nonhuman animal models (e.g., pigeons & rats).
  10. Applied behavior analysis is the field of clinical research in which operant principles are used to enhance human welfare. In 1968 the Journal of Applied Behavioral Analysis was created to publish works to this end.
  11. Other fields. Single-case designs are used in a number of fields, including cognitive neuroscience, clinical psychology, psychopharmacology, pediatric psychology, and rehabilitation medicine.

Part II: Types of Single-Case Designs

  1. ABAB designs involve a sequence of phases. In the A phase the treatment is absent and in the B phase it is present. The temporal sequence of the phases is indicated by reading from left to right. The ABAB design is the most basic of all single-case designs; it is also referred to as a withdrawal design because after the initial A phase, the subsequent A phase involves a withdrawal of the treatment. In the ABAB design, the target behavior is the behavior that the treatment is designed to change. Essentially it is a dependent variable affected by the absence or presence of the independent variable (i.e., treatment).
  2. Drawing causal inferences. The reason why single-case designs repeat the AB sequence twice is because if it was done only once history could be a potential alternative explanation for the change in behavior during the B phase. The addition of the second AB sequence further increases the study’s internal validity. Ending on the B phases is advantageous because the study ends while the treatment is in place.
  3. Special concerns. Reverting back to the A phase after a successful B phase is important in order to rule out other potential explanations for a change in behavior, but doing so does raise serious ethical questions. In applied research, scientists must weigh the potential risks and benefits of completing the ABAB design. Another drawback of the ABAB design is that on occasion during the second A phase, behavior does not go back to baseline levels. It is recommended that single-case designs not be used if their results are expected to exhibit this pattern of behavior.
  4. Multiple-baseline designs are those in which two or more participants are exposed to the same treatment and the switch from baseline to treatment periods is staggered across participants. In addition, the time of a participant’s exposure to baseline and treatment measures may change across participants. Variants of the multiple-baseline design can also be used to measure a single subject. In a multiple-baseline design across behaviors, the same treatment is applied to two or more distinct behaviors of the same individual and the switch from baseline to treatment periods is staggered across behaviors. In a multiple-baseline design across settings, the same treatment is applied to the same target behavior in two or more settings, and the switch from baseline to treatment is staggered across settings.
  1. Special concerns. Even though multiple baseline designs do not withdraw treatment, as do ABAB designs, there are still ethical concerns. In order to draw causal conclusions the implementation of treatment must be staggered across participants. That said, if a treatment shows promise to one participant is it ethical to withhold it from others until a later time? A second item to consider is that for multiple-baseline designs to draw causal conclusions, at least two participants (some argue more) must be examined. A final concern is contamination. It is possible that, as participants observe one another, they may influence each other’s behavior.
  1. In changing-criterion designs, an initial baseline phase is followed by a treatment phase that lasts until the target behavior reaches a criterion level and becomes stable. Treatment then continues in a series of additional phases, with a new performance criterion set in each phase. In a study of smoking cessation described in the text there were 20 criterion changes over 52 weeks! The concern with this design is when the change in behavior does not exhibit the same pattern of change in criterion.
  2. In addition to the designs described above, many researchers combine designsto create a unique design tailored for a particular participant.

Part III: Extending Single-Case Designs to Social Units

The “case” in single-case studies need not be an individual person, but can be a larger social unit such as a classroom of students. In some experiments the sample that constitutes each case remains the same and in others it does not.

  1. Buckle Up! A withdrawal design. Austin et al (1998) found that the proportion of restaurant patrons who fastened their seat belts when they left a restaurant increased when the hostess reminded them to do so than when she simply said “goodbye.” In this study the group of restaurant patrons was the unit under study.
  2. Slow Down! Amultiple-baseline Design. Van Houten and Malenfant (2004) examined whether the intensive enforcement of the speed limit would affect the proportion of drivers who yielded to pedestrians in two different locations. In the study drivers in one location were considered one unit, and drivers in the other location another unit.

Part IV: Issues Concerning Single-Case Designs

  1. Ethical issues. In single-case designs the participant eventually receives the treatment. This is considered to be an ethical advantage over other types of research designs. Still, the fact that treatment may be delayed for some participants constitutes a need for consideration. Researchers should always evaluate the ethical issues involved in a research study and adjust their methodology to ensure participant safety.
  1. Internal validity. Single-case studies reflect a within-subjects approach to research. In a group within-subjects design, counterbalancing is an important element to reduce the possibility that order effects could explain the change in behavior. Very few single-case studies use counterbalancing techniques. Because randomizing the order of treatment is not inherent to most single-case designs some researchers believe these designs are quasi-experimental design, at best. However, single-case studies certainly have more control than quasi-experiments do over the independent and dependent variables. To reduce the possibility that order effects contribute to behavior, researchers typically wait to implement future treatments until behavior in the current condition exhibits stability.
  2. Data analysis is very different for single-case designs. In general, alternative explanations for a change in behavior are ruled out by visually inspecting a temporal pattern of behavior. Most staunch single-case experimenters see no need for statistical methods to analyze their results, and those who do disagree about which method is appropriate to use.
  3. Examining the interactions. Single-case designs are not very effective in their ability to examine potential interactions between variables. First, the design needed to determine whether an interaction exists is very complicated. Second, it is difficult to determine whether the effect of two independent variables produces an interactive effect (AB) or an additive effect (A, then B) on behavior. Lastly, adding more than two treatments leads to possible carryover effects.
  4. External validity. Critics of the single-case design state that the use of a single subject leads to poor external validity. Proponents of this design believe the concerns about external validity are the same as designs that use groups of participants.

C.LECTURE AND CLASSROOM ENHANCEMENTS

PART I: A Single-Case Approach to Experimentation

A.Lecture/Discussion Topics

  • Single-case designs are not the same as case studies. Many undergraduate psychology students do not adequately understand the difference between single-case designs and case studies. Although each research method examines a single participant, or unit, a single-case design is an experimental design while the case study is purely descriptive. And since they are descriptive they are usually employed at the beginning stages of research or they are unique cases we want to know more about. In contrast, single-case designs enable a researcher to manipulate an independent variable to identify whether it has a causal effect on behavior. Emphasize to your students the difference between each design and how they are methodologically very different ways to study human behavior.
  • Applied behavior analysis. Applied behavior analysis (ABA) is a huge subfield in psychology. The Association for Applied Behavior Analysis was founded in 1974 and now has over 20,000 members internationally across 81 chapters. The Behavior Analyst Certification Board (BACB) was developed in 1998. It awards BS, MS, and doctoral-level certificates in ABA to students in BACB-approved ABA programs in order to ensure that those practicing ABA are adequately trained. (Note: Applied Behavior Analysis is often referred to as behavior modification.)

B.Classroom Exercise

  • Using single-case designs in real life. The single-case design makes it possible for a researcher to conduct a true experiment using only one participant. Chapter 12 begins with a description of a single-case design that was used to change an undesirable behavior. Most everyone has a pet (or a child!) whose behavior can be improved. Ask your students to develop a single-case design that they could use in their lives. For example, if a student is a new dog owner, he or she may think of an experiment that would test whether giving doggie treats each time the dog potties outside reduces the number of accidents in the house. This exercise can be used not only to demonstrate single-case research, but also how research can be applied in our daily lives.

C.Web Resources

  • This website has a variety of resources about single-case designs.

D.Additional References

Gast, D. L., & Ledford, J. (Eds.). (2010). Single-Subject Research in Behavioral Sciences.
Routledge.

Hammond, D., & Gast, D. L. (2010). Descriptive analysis of single subject research designs: 1983–-2007. Education and Training in Autism and Developmental Disabilities, 45(2), 187–202.

Moore, J., & Shook, G. L. (2001). Certification, accreditation, and quality control in behavior analysis. The Behavior Analyst, 24(1), 45.

PART II: Types of Single-Case Designs

A.Lecture/Discussion Topics

  • Why an ABAdesign just won’t do. The text emphasizes that single-case designs repeat the AB sequence in order to enhance internal validity; if a treatment does have an effect on behavior then we should see a change in behavior when the treatment is administered and the behavior should revert to baseline when the treatment is removed. So why not just add a second A phase and skip the second B phase altogether? To explain why this is not a good idea, consider the following example. Suppose a drug manufacturer wants to determine whether a new drug is effective at reducing anxiety. A participant who suffers from anxiety is recruited and is “measured” for several days under control conditions. The drug is then administered to the participant and it is quickly obvious from observation that her behavior is markedly improved. To make sure it is the drug and not something else that is reducing her anxiety, the researchers take away the drug. They find that once it is removed the participant’s anxiety returns to baseline levels. The problem here is that although the methodology is sufficient to support internal validity, ending a single-case study with the baseline phase when a treatment was beneficial to the subject is unethical.
  • Changing criterion design to take steps toward a goal. A changing criterion design is an excellent way for one to eventually achieve a goal, especially when that goal is very difficult to achieve. For example, let’s say a parent wants her child’s behavior to improve at school. Each day the child brings home a smiley face for which he or she will get some small reward (get a popsicle, watch a favorite TV program, etc.). Then, after the child has received three to five smiley faces for good behavior at school each week for two consecutive weeks the criterion is changed so that three smiley faces in a row becomes the criterion for a prize (the prize should be a “better” one than previously). This process of changing the criterion “upward” continues successively so that eventually the child goes for long stretches without being reinforced. It may take a while, but if this process is done consistently it should produce the desired outcome!

B.Classroom Exercise

  • Using single-case designs in real life: Part II. If students created a single-case design (see the suggested classroom exercise in Part I), ask them to modify its methodology so that it becomes a multiple-baseline or a changing-criterion design. Ask them to identify how altering the methodology would benefit, or weaken, the study based upon the research question asked.

C.Web Resource

  • Common misconceptions about applied behavior analysis.
  • Changing criterion designs andmultiple-baseline designs (YouTube).

D.Additional References

  • Review paper about multiple-baseline designs.

Barger-Anderson, R., Domaracki, J. W., Kearney-Vakulick, N., & Kubina Jr, R. M. (2004). Multiple baseline designs: The use of a single-case experimental design in literacy research. Reading Improvement, 41(4), 217.

PART III: Extending Single-Case Designs to Social Units

A.Lecture/Discussion Topics

  • Single-case designs in education. Aclassroom makes for an excellent environment in which to conduct single-case experiments using social groups.In fact, education is one of the fields that use single-case designs most frequently. In these experiments, the class, which may consist of many people, is a single social unit. For example, a teacher may record the class’s behavior for a period of time during the control condition (A phase), and then after a treatment has been implemented (B phase). The percentage of students who exhibit the behavior being measured (e,g,, raising their hand before speaking) is then plotted across the multiple measures obtained during baseline and treatment phases to determine whether treatment produced a change in behavior.

B.Classroom Exercises

  • Identifying social units. Single-case studies are used to examine an individual or a single group of individuals. Ask students to think of groups that could be studied with a single-case experimental design. Have them think about whether the results from a single group would provide better evidence about the independent variable–dependent dependent variable relationship than the results from a single-case study using only one participant. Have them provide a rationale for their answer.

C.Additional References