What Is the Changing Criterion Design?
The changing criterion design is a single-subject experimental design that demonstrates a functional relationship by showing that behavior changes in a stepwise fashion to match progressively adjusted criteria. After a baseline phase, the researcher sets an initial performance criterion. When the participant consistently meets that criterion, it is adjusted (usually made more demanding), and the process repeats through multiple phases.
Experimental control is demonstrated when the participant's behavior closely matches each new criterion. If the behavior consistently tracks the criterion across multiple phase changes, it is unlikely that extraneous variables are responsible for the observed pattern.
Structure of the Changing Criterion Design
Baseline Phase
As with all single-subject designs, data collection begins with a baseline phase. The researcher measures the target behavior without any intervention to establish current levels of performance. The baseline data inform the setting of the first criterion, which is typically a modest improvement over baseline levels.
Treatment Phases with Successive Criteria
Following baseline, the intervention is introduced along with the first performance criterion. The criterion specifies the level of behavior needed to earn reinforcement (or avoid a consequence). For example, if a client currently completes an average of 5 math problems per session, the first criterion might be set at 8 problems.
Once the client consistently meets the criterion (usually defined as meeting it for a specified number of consecutive sessions), the criterion is adjusted to a more demanding level. This process continues through multiple phases, with each new criterion representing a step toward the ultimate goal.
Setting Criterion Levels
The size of each criterion change is an important design decision. Criteria that change too drastically may result in the participant failing to meet them, while criteria that change too modestly may make it difficult to detect whether behavior is actually tracking the criterion or simply drifting upward naturally.
A general guideline is to set each new criterion at approximately 20% to 30% above (or below) the previous criterion, though this varies based on the behavior and the individual. The changes should be large enough to be clinically meaningful but small enough to be achievable.
Bidirectional Changes: Strengthening the Design
One way to strengthen the demonstration of experimental control in a changing criterion design is to include bidirectional changes. This means occasionally making the criterion less demanding before increasing it again. If the behavior decreases when the criterion decreases and then increases again when the criterion is raised, this provides stronger evidence that the behavior is controlled by the criterion rather than by other factors such as practice, maturation, or external motivation.
For example, if a client has been increasing their exercise duration from 10 minutes to 20 minutes to 30 minutes, a bidirectional change might temporarily lower the criterion to 20 minutes. If the client's exercise duration decreases to match the lowered criterion, this strongly supports the conclusion that the contingency (not some other variable) is controlling the behavior.
The Shaping Analogy
The changing criterion design bears a strong resemblance to shaping (differential reinforcement of successive approximations). In both cases, the criterion for reinforcement is gradually adjusted as the individual's behavior improves. The difference is that shaping is a clinical procedure for teaching behavior, while the changing criterion design is a research methodology for demonstrating that the procedure is causing the behavior change.
This connection makes the changing criterion design a natural choice when the intervention involves shaping or any procedure that produces gradual, incremental behavior change.
Clinical Examples
Example 1: Increasing Physical Activity
A behavior analyst works with an adult client to increase daily steps. Baseline data show an average of 2,000 steps per day. The first criterion is set at 3,000 steps. After the client consistently meets this for five days, the criterion increases to 4,500, then 6,000, then 8,000, and finally 10,000. At one point, the criterion is temporarily reduced to 6,000 to demonstrate bidirectional control.
Example 2: Reducing Cigarette Consumption
A client smokes an average of 20 cigarettes per day during baseline. The changing criterion design sets successive criteria: 16, 12, 8, 4, and finally 0 cigarettes per day. Reinforcement is delivered when the client meets or stays below the criterion for each phase.
Example 3: Academic Fluency
A student reads an average of 40 words correct per minute during baseline. Successive criteria are set at 55, 70, 85, and 100 words per minute. The student receives preferred activities contingent on meeting each criterion during assessment probes.
Strengths of the Changing Criterion Design
- No withdrawal required: The intervention remains in place throughout the study, which is both ethical and practical.
- Natural fit for shaping: It aligns perfectly with interventions that produce gradual behavior change.
- Straightforward implementation: The design is intuitive for practitioners and easy to explain to stakeholders.
- Flexible: Criterion levels can be adjusted based on the individual's progress.
Limitations
- Limited to gradual change: The design is not appropriate for behaviors that are expected to change rapidly or in an all-or-nothing fashion.
- Requires close tracking: Behavior must closely match each criterion for the design to demonstrate control. If behavior overshoots the criterion substantially, the evidence is weakened.
- Fewer phase changes weaken control: A design with only two or three criterion changes provides a weaker demonstration than one with five or more changes.
- Cannot compare treatments: Unlike the ATD, the changing criterion design does not allow comparison of different interventions.
BCBA Exam Tips
On the exam, choose the changing criterion design when the scenario describes gradual, stepwise behavior change, shaping procedures, or behaviors measured on a continuum. Remember that bidirectional changes strengthen the design, and that the number of criterion changes affects the strength of the experimental demonstration.