The big idea
Behavior analysts judge single-case data mainly by visual analysis: inspecting graphed data to decide whether behavior changed meaningfully and whether the design shows that the independent variable caused the change. Visual analysis is systematic, not a glance. It looks at specific features within each phase and between adjacent phases, then asks whether the pattern across the whole design rules out other explanations.
Critiquing data goes a step further. Even a dramatic graph can fail to support a conclusion if the baseline was too short, two variables changed at once or the measurement was unreliable. This task pulls together the logic of D.4, the threats in D.3 and the graph-reading skills of C.11.
Breaking it down
Within a phase: level, trend, variability
- Level is the value on the vertical axis around which the data cluster, often summarized by the mean or median. A change in level at a phase change is the most common sign of an effect.
- Trend is the overall direction of the data path (increasing, decreasing or flat) and how steep it is. A baseline trending in the direction the intervention is expected to produce weakens any conclusion.
- Variability is how much the data points differ from one another or from the trend. The more variable the data, the more points are needed before a decision.
Between phases: immediacy, overlap, consistency
- Immediacy of effect is how quickly the data change after a phase change. Compare the last three points of one phase with the first three of the next. An immediate change is easier to attribute to the IV than a delayed one.
- Overlap is how much of the data in one phase falls within the range of the adjacent phase. Less overlap means a clearer difference.
- Consistency across similar phases asks whether phases with the same condition look alike: both baselines similar, both intervention phases similar.
These six features are the ones used in widely cited single-case standards (Kratochwill et al., 2013).
The whole design
After reviewing each phase change, ask whether the design produced enough demonstrations of the effect at different points in time (usually at least three), and whether each change happened when, and only when, the IV changed.
Questions for a critique
- Was there enough baseline data, and was it steady or trending away from the expected effect?
- Did only one variable change at each phase change?
- Were interobserver agreement and treatment integrity measured and reported (see C.8 and C.12)?
- Is the vertical axis scaled honestly? A truncated or stretched axis can make a small change look large.
- Are data paths broken at phase changes, and are conditions clearly labeled?
- Is the change large enough to matter to the client (the effective dimension, A.5)?
Aids to visual analysis
Visual analysts sometimes disagree, especially when effects are small or data are variable. Structured aids help. The dual-criterion method projects the baseline mean line and trend line into the intervention phase and counts how many intervention points fall beyond both (Fisher et al., 2003). Nonoverlap indices and other effect sizes can supplement visual analysis but do not replace it.
Seeing it in practice
A supervisor reviews three graphs from a team’s quarterly report.
| Graph | What the supervisor sees | Interpretation |
|---|---|---|
| A-B-A-B for Elena’s requests | Clear level changes at every phase change, immediate, almost no overlap, similar baselines and similar intervention phases | Strong evidence of a functional relation |
| A-B for Omar’s aggression | Baseline was already decreasing, and intervention data continue the same slope | The change cannot be credited to the intervention; the trend came first |
| Multiple baseline across three settings for Grace’s on-task behavior | The first two tiers change when treated, but the third rose while still in baseline, right after the first tier was treated | Possible covariation or diffusion of treatment; control is weaker than it looks |
Common mix-ups
- Level vs. trend. Level is where the data sit; trend is where they are heading. Data can shift level with no trend, or trend with little change in average level.
- High variability is not failure. It signals that more data are needed or that the sources of variability should be examined.
- A big change is not proof. Without verification and replication, even a large change supports only prediction.
- A delayed effect is not automatically no effect, but it is harder to separate from history and maturation. Look for the same delay across replications.
- Overlap is judged against the adjacent phase, not against the whole study.
Exam tip: When asked whether a graph shows a functional relation, check two things in order. Did the data change in level, trend or variability at each phase change? Does the design give at least three demonstrations of that change at different times? A strong change shown only once is not enough.
Quick recap
- Within phases: level, trend, variability. Between phases: immediacy, overlap, consistency across similar phases.
- A baseline trend toward improvement undermines conclusions; a trend away from it does not.
- Critique the design too: baseline length, one change at a time, agreement and integrity data, honest axes.
- Visual analysis comes first; structured aids such as the dual-criterion method support it.
Further reading
- Cooper, J. O., Heron, T. E., & Heward, W. L. (2020). Applied behavior analysis (3rd ed.), chapters 6 and 10.
- Fisher, W. W., Kelley, M. E., & Lomas, J. E. (2003). Visual aids and structured criteria for improving visual inspection and interpretation of single-case designs. Journal of Applied Behavior Analysis, 36(3), 387-406.
- Kratochwill, T. R., Hitchcock, J. H., Horner, R. H., Levin, J. R., Odom, S. L., Rindskopf, D. M., & Shadish, W. R. (2013). Single-case intervention research design standards. Remedial and Special Education, 34(1), 26-38.
Key terms
- Visual analysis (Visual inspection)
- The systematic inspection of graphed data to judge whether behavior changed meaningfully and whether the design shows that the independent variable caused the change. It considers level, trend and variability within phases, and immediacy, overlap and consistency across phases.
- Level
- The value on the vertical axis of a graph around which a series of data points clusters, often summarized by the mean or median. A change in level between phases is a common sign of an intervention effect.
- Trend
- The overall direction of a data path (increasing, decreasing or flat) along with its steepness. A baseline trend toward the improvement the intervention is expected to produce makes any effect hard to interpret.
- Variability
- How much the data points in a phase differ from one another or from the trend line. The more variable the data, the more measurements are needed before a confident decision can be made.
- Immediacy of effect
- How quickly behavior changes after a change in conditions, usually judged by comparing the last few data points of one phase with the first few of the next. The more immediate the change, the easier it is to attribute to the independent variable.
- Overlap
- In visual analysis, the extent to which data points in one phase fall within the range of data in the adjacent phase. Less overlap indicates a clearer difference between conditions.
- Steady-state responding (Steady state)
- A pattern of responding that shows relatively little variation in level and trend over a series of measurements. Changing conditions at steady state makes it easier to see whether a change in behavior is due to the new condition.
Check yourself
5 questions on D.6. Choose an answer to see whether you're right and why.
Show the answer
Answer: D. Immediacy is how quickly behavior changes once the condition changes, judged by comparing the last few points of one phase with the first few of the next. Consistency across phases would compare this intervention phase with the later one, not describe the speed of change.
Show the answer
Answer: B. The behavior was already improving before intervention began, and the intervention data simply continue the same trend. There is no change in level or trend at the phase change, so nothing points to the intervention. A baseline trend toward improvement is a classic reason a graph cannot support a conclusion.
Show the answer
Answer: B. High variability makes the baseline a poor predictor, so any effect would be hard to see. More data, or identifying what drives the variation (such as sleep or schedule changes), gives a better basis for comparison. Removing data points would misrepresent the behavior.
Show the answer
Answer: B. Overlap compares the range of data in adjacent phases; the less overlap, the clearer the difference. How quickly behavior changes is immediacy, and similarity across same-condition phases is consistency.
Show the answer
Answer: B. Truncating the vertical axis exaggerates differences: a change from 44 to 41 can fill the graph. Critiquing data includes checking that the scale represents the size of the change honestly. Rate is an appropriate measure for self-injury.
Adds a step to your study record. Saved in this browser only.