/ Single-Case Design Graphs

Overview › Visual analysis › Lesson 1

Lesson 1 of 7About 15 minutes

The six features of visual analysis

Visual analysts look at six features of the data, first within each phase and then between phases. Together, they answer whether behavior changed and whether it changed when the conditions did.

You will be able to

  • Describe the level, trend, and variability of the data within each phase.
  • Judge immediacy, overlap, and consistency between phases, using the last three and first three points and the range of each phase.
  • Answer feature questions about new graphs, reaching 8 of your last 10 in the practice.

Where the six features come from. They are the features named in the What Works Clearinghouse single-case standards (Kratochwill et al., 2010), drawing on decades of behavior-analytic practice. The colored lines and bands on these figures are teaching aids. They don't belong on a publication graph.

Within a phase

1. Level

The typical value of the data in a phase, usually its mean or median. Compare levels between adjacent phases to see how much behavior changed.

AB graph of percentage of intervals on task with dashed mean lines at 30 percent in baseline and 76 percent in intervention.
A change in level. The mean rose from 30% in baseline to 76% in intervention.

2. Trend

The direction and steepness of the data path, often summarized by a best-fitting straight line. A baseline already trending in the direction you hope the intervention will produce makes any later change hard to credit to the intervention. To check, extend the baseline trend line into the next phase and see whether the new data simply follow it.

AB graph of percentage of intervals on task. Baseline rises steadily from 20 to 46 percent over 6 sessions, with a solid trend line fitted to it. The trend line continues as a dashed line through the intervention phase, where the data rise from 52 to 82 percent and stay within 1.5 points of that line.
A trend that simply continued. On-task behavior rose about 5 points per session in baseline, from 20% to 46%. Extended into the intervention phase (dashed), the baseline trend line predicts every intervention point, from 52% to 82%, within 1.5 points. The higher level was predicted by baseline, so it can't be credited to the intervention.

3. Variability

How much the data bounce around the phase's level, often described by the range. High variability makes the level hard to predict and needs more data before a phase change. Lesson 6 measures it from every point, with each point's distance from the level line, the variance, and the standard deviation.

AB graph of percentage of intervals on task with a narrow band around stable baseline data and a wide band around highly variable intervention data.
Stable, then variable. Baseline stays within a 4-point range. Intervention ranges over 45 points, so its level is uncertain.

Between phases

4. Immediacy of the effect

How quickly behavior changes after the phase change. Compare the last three points of one phase with the first three of the next. Immediate changes are more convincing, unless the intervention is expected to work gradually.

AB graph of percentage of intervals on task highlighting the last three baseline points and first three intervention points, which are similar; the change comes later.
A delayed change. The first three intervention points look just like the last three baseline points. Behavior changed three sessions later, which leaves room for other explanations.

5. Overlap

The proportion of data in one phase that falls within the range of the adjacent phase. Less overlap means more separation, but read it together with trend: if baseline was already moving toward the intervention level, intervention data can sit entirely outside the baseline range with no evidence of an effect. In the trend example above, no intervention point falls inside the baseline range, yet the graph gives no evidence that the intervention changed behavior.

AB graph of percentage of intervals on task with the baseline range shaded across both phases; three of seven intervention points fall inside it.
Partial overlap. Three of seven intervention points fall inside the baseline range (shaded). The intervention level is higher, but not cleanly separated.

6. Consistency across similar phases

Whether phases in the same condition look alike. Compare all the baseline phases, then all the intervention phases. Consistent patterns across repetitions are what make replication convincing.

Reversal graph of percentage of intervals on task with matching mean lines: both baselines near 30 percent and both interventions near 80 percent.
Consistent replications. Both baselines average 30%, and both intervention phases average about 80%.

Putting it together. No single feature decides the question. A strong effect usually shows a large change in level, little overlap, an immediate change, stable data within phases, no baseline trend toward the treatment level, and consistent patterns when phases repeat. Ask what the baseline predicted, and whether the next phase broke that prediction. For a full analysis of one graph, from each feature to a conclusion about control, see the worked example in Lesson 2.

Practice

Each graph asks about one feature. The answer is generated from the data, and the explanation shows the numbers.