Small N Designs in Psychology

Small N Designs

• Focuses on closely examining a single individual or a small group.
• Designs are often called single-subject designs because the behavior of each research subject is considered individually.
• Data may be combined statistically or described individually for replication.
• Includes case studies, which are in-depth analyses of individuals or events.
• Designs have a long history and were used frequently by early experimental psychologists.

Research in Psychology Began with Small N

• Statistical analysis was in its infancy when psychology emerged as a new science.
• Francis Galton began to conceptualize correlations and inferential techniques like ANOVA.
• Widespread use of large N designs and inferential statistics occurred only after Fisher's work on the analysis of variance appeared in the 1930s.
• Some pioneers studied their own behavior (e.g., Ebbinghaus) or that of a single individual (e.g., Watson and Rayner's Little Albert study).
• In Wundt's lab, small N designs were dominant and typically involved a small number of research participants, often including the investigator.
• Participants were called observers because they typically observed their own behavior and mental processes.
• Fletcher B. Dresslar's study of "facial vision" is an example, where he showed the skill was more about hearing than vision.
• Dresslar tested two additional subjects to replicate the initial finding demonstrating the replication strategy is a common feature of today's small N designs.
• Dresslar looked for a parsimonious explanation and found the "power to distinguish [the panels] was lost entirely" when ears were plugged.
• Facial vision was the ability to detect slight differences in reflected sound waves.
• Leta Hollingworth criticized large-scale questionary studies for rushing through data collection without perceiving the children as individuals.
• Edward L. Thorndike's puzzle box research is an early small N project and an impoprtant historical predecessor to B. F. Skinner's work on operant conditioning
• cats were studied individually in puzzle boxes, and Thorndike described his results.
• The cats learned to escape through "trial and error, with accidental success" and the "Law of Effect."
• Successful behaviors were repeated, while unsuccessful ones were eliminated.
• Thorndike challenged the prevailing ideas about the thinking abilities of animals and provided a more parsimonious explanation of problem-solving abilities.
• A model for learning eventually took the form of B. F. Skinner's experimental analysis of behavior.

Reasons for Small N Designs

• Studies using one or a few individuals continue to contribute to our knowledge of behavior.
• Small N studies cover the range from laboratory to field studies and from basic to applied research.

Occasional Misleading Results from Statistical Summaries of Grouped Data

• Summarizing data from large groups can fail to characterize individual behavior, lacking individual-subject validity.
• A lack of individual-subject validity can produce incorrect conclusions about behavior.
• Concept-learning experiments demonstrate the importance of examining individual data.
• Examine individual data to see if they mirror the grouped data for individual-subject validity.

Practical and Philosophical Problems with Large N Designs

• Necessary when potential subjects are rare or difficult to find, such as in clinical psychology or animal research.
• Ethical considerations and costs can limit sample size.
• Large N designs may not reflect individual behaviors, leading to philosophical preference for small N designs.
• BF Skinner advocated intensive study of individuals to derive general principles.
• "Study one rat for a thousand hours" rather than "a thousand rats for an hour each".
• Precise control over experimental conditions reduces random variability.

The Experimental Analysis of Behavior

• Skinner's system is the "experimental analysis of behavior," and provided the backbone for research designs used in applied behavior analysis.
• The behaviors result from a person's learning history, especially operant conditioning.
• A bit of behavior is modified by the consequences of the behavior.
• Positive consequences cause the behavior to recur.
• Rate of Response is the most important dependent variable to measure.
• Three things must be specified to predict and control behavior: the occasion, the response, and the reinforcing consequences.
• Skinner's contingencies of reinforcement means the behaviors are controlled by the environment.
• Operant conditioning uses an operant chamber, or Skinner box.
• A rat pressing a lever for food is an example of operant conditioning.
• Behavior can be brought under the environmental control of stimuli such as the light.
• A cumulative recorder continuously records and portrays the rate of bar-pressing behavior.

Applied Behavior Analysis

• Smith (1992) argued there is a distinction between two broad categories of scientists: contemplative and technological ideal.
• "Contemplative ideal types" focus on understanding the basic causes of events.
• "Technological ideal types" look for ways to use science to control and change the world.
• B. F. Skinner was in the the technological, and was interested in applying the results of experimental analysis of behavior to real-world problems.
• Skinner wanted to redesign society based on operant conditioning principles.
• Recommending the redesign of society made Skinner a controversial figure.
• Behaviorists describe their goals as the prediction and control of behavior.
• Controlling human behavior generates an ethical dilemma.

Small N Designs in Applied Behavior Analysis

• Watson and Rayner (1920) described ways to remove fear, but never tested them.
• Mary Cover Jones (1924) eliminated a boy's fear of rabbits, a pioneer of using behavioral methods to reduce fear.
• Behavioral approaches to therapy became popular in the 1960s and 1970s.
• Additional demonstrations of the effectiveness of learning-based procedures helped the approaches to therapy.

Elements of Single-Subject Designs

• The behaviour of a single individual must be shown to change as a result of the treatment being applied and not as a result of some other factor.
• The target behavior(s) must be operationally defined.
• A baseline level of responding must be established to asses treatment effects.
• The treatment should begin, and the behavior should continue to be monitored.
• The simplest single-subject design is the A-B design, but is weak.
• Internal Validity threats can be mitigated by using withdrawal designs.

Withdrawal Designs

• If the treatment is removed and the behaviour goes back to its baseline, then it is likely that the treatment is causing the effect.
• Researchers prefer an A-B-A-B design over the A-B-A design.
• The A-B-A-B design has the ethical advantage of finishing the study with the treatment in place.
• Flood, Wilder, Flood, and Masuda (2002) is an example, by training the schoolmates of children with ADHD to help alter off-task behaviors when doing math problems.

Multiple Baseline Designs

• They are used if a withdrawal design is not feasible.
• Several baseline measures are established then treatment is introduced at different times.
• If the behavior responds to the treatment, the behavior should change when the program is put into effect, and only then.

Changing Criterion Designs

• A target behavior is too difficult or complex for the person to reach all at once
• A procedure is inspired by the operant procedure of shaping,
• The procedure begins by establishing the usual baseline.
• The Deluca and Holborn (1992) study is an example of the effectiveness of excercise increase in obese and non-obese children by making bike riding more rewarding.

Alternating Treatments Designs

• It’s used to evaluate more than a single treatment approach within the same study.
• Different treatment strategies (usually two) are then alternated numerous times
• Lang and his research team (Lang et al., 2009) project on an 8 year old named Sue improving her repetitive actions by not paying attention to them.
• Functional play increased, stereotypy decreased, and the only increases in the screaming and falling behaviors occurred during non-AOC.

Evaluating Single-Subject Designs

• The designs have been helpful in assessing the effectiveness of operant approaches
• Derive from the behaviorist dictum that if conditions are precisely controlled, then orderly and predictable behavior will follow.
• They are foun effective in situations ranging from therapeutic behaviour change in individuals to community behaviour change in littering.
• These designs are not without critiscm

Case Study Designs

• A case study in abnormal psychology , for example, the case study approach is often used to understand the dynamics of specific disorders.
• A case study in can be usefull for experimental psychology, however, shedding light on basic psychological phenomena
• Case study approach is common in clinical work, in which the case of someone with a particular psychological disorder (e.g., depression) is used to illustrate the factors that lead to and influence the disorder and the methods of treatment for it.
• Case studies can provide inductive support for a theory
• Elizabeth Loftus(1979) show that memories for the events of our lives are often distorted by circumstances that intervene between the target event and the later recall of it.