Psy 9: Chp. 10-Single Subject Research

Single-Subject Research

Studies one person or a small number of participants intensively, with repeated measurement and experimental manipulation or control. "Single-subject" does not require only one participant.

Group Research

Usually studies many participants and summarizes results with group means, standard deviations, and inferential statistics.

Single-Subject Research vs. Case Study

A case study describes an individual in depth, while single-subject research experimentally manipulates conditions, collects highly structured quantitative data, and tests causal effects.

Behavioral Roots

Single-subject research is strongly associated with behaviorism, Skinner's experimental analysis of behavior, and applied behavior analysis, although it can be used with other theoretical perspectives.

Individual Differences

Group averages can hide meaningful effects for particular people. Single-subject research asks whether an intervention works clearly and consistently for the individual.

Strong, Consistent Effects

Researchers look for effects large and reliable enough to be visible in repeated behavior, rather than merely tiny average differences.

Social Validity

The target behavior should be socially important and the treatment should produce a substantial change that can be implemented reliably in the real-world context where it matters.

Why Is Social Validity Important?

Researchers need to determine not only whether behavior changed, but whether the change was meaningful and useful in real life.

Repeated Measurement

The dependent variable is measured many times over time rather than only once at pretest and once at posttest.

Phases or Conditions

Distinct phases are labeled A, B, C, and so on. A often represents baseline and B usually represents treatment.

Baseline

The control or comparison phase that shows behavior before the treatment is introduced.

Steady State Strategy

Researchers usually wait until behavior becomes relatively consistent within a phase before changing conditions, making a treatment-related change easier to detect.

Internal-Validity Logic

Confidence increases when the dependent variable changes systematically when the independent variable is introduced, removed, or introduced at different times across baselines.

ABA Reversal Design

A = baseline, B = treatment, A = return to baseline. If behavior changes when treatment begins and moves back toward baseline when treatment is removed, this supports a treatment effect.

ABAB Design

Adds a second treatment phase to an ABA design. Reproducing the treatment effect after the second introduction strengthens the causal conclusion.

Multiple-Treatment Reversal

Uses additional treatment phases, such as A-B-C-B-C, to compare more than one treatment, provided effects can be reversed and carryover is controlled.

Limitation of Reversal Designs

Reversal is inappropriate when treatment creates lasting change, withdrawing treatment would be unethical, or behavior cannot reasonably return to baseline.

Multiple-Baseline Design Across Participants

Establish baselines for several participants and introduce the same treatment at different times for each person.

Multiple-Baseline Design Across Behaviors

For one participant, establish baselines for several behaviors and introduce treatment to each behavior at different times.

Multiple-Baseline Design Across Settings

For one participant, measure the same behavior in different settings and introduce treatment at different times in each setting.

Why Are Multiple-Baseline Designs Persuasive?

If behavior changes only after treatment is introduced on each staggered baseline, alternative explanations such as a single outside event become less plausible.

Key Advantage of Multiple-Baseline Designs

They can demonstrate treatment effects without withdrawing an effective or irreversible treatment.

When Is a Multiple-Baseline Design Preferable to a Reversal Design?

When withdrawing an effective treatment would be unethical, the treatment produces lasting change, or behavior cannot reasonably return to baseline.

Visual Inspection

Plot each participant's repeated data and judge whether changes in the dependent variable closely follow changes in the treatment condition.

Level

A noticeable shift upward or downward in the typical value of the dependent variable from one phase to another.

Trend

A gradual increase or decrease across observations. A trend that changes direction after treatment can support an effect.

Latency

How quickly the dependent variable begins changing after the condition changes. Shorter latency generally strengthens the treatment interpretation.

PND

Percentage of non-overlapping data; the percentage of treatment observations more extreme than the most extreme relevant baseline observation.

What Does a Higher PND Mean?

A higher PND suggests a stronger effect. PND supplements rather than replaces visual inspection.

Data Analysis in Single-Subject Research

Emphasizes individual graphs and clear treatment-related changes. Formal statistics can supplement visual inspection.

External Validity in Single-Subject Research

Generalization is built through replication across people, behaviors, settings, and studies rather than assumed from one group average.

When Is Single-Subject Research Especially Useful?

When the research question concerns individual treatment response, rare cases, clinical or applied change, or strong effects that may differ across people.

Converging Evidence

Confidence is strongest when single-subject, group, correlational, and other methods with different strengths and weaknesses point toward the same conclusion.

Single-Subject vs. Group Research

Single-subject research focuses on repeated measurements and treatment-related changes within individuals, while group research typically compares many participants using group averages and inferential statistics.

How Does Single-Subject Research Establish Experimental Control?

By repeatedly measuring behavior and showing that changes in the dependent variable systematically follow changes in the independent variable.

What Should You Look for When Reading a Single-Subject Graph?

Identify baseline and treatment phases, look for clear and timely changes in level or trend, and determine whether the effect reverses or replicates.

How Do You Decide Whether a Treatment Caused a Change?

Look for a clear and timely change in the dependent variable when treatment is introduced, removed, or introduced at different times across baselines.

What Should Be Considered in Applied Single-Subject Research?

Whether the treatment caused a clear change and whether that change has social validity, meaning it is meaningful and useful in real life.