ABA graph maker: session data to a clean line graph

An ABA graph is a picture of one person’s behavior over time, under different conditions. The most common type of graph in ABA is the equal-interval line graph: each session is a data point, the points are joined by lines within a phase, and sessions sit at equal distances along the x-axis. Phase change lines mark where conditions change.

It is the graph behind almost every single-case design, and the one your supervisor expects you to draw and read. Graphing is part of both the RBT and the BCBA Test Content Outlines: on the RBT exam it sits in the Data Collection and Graphing domain, which you can practice on its own in the free RBT practice test.

In an ABA line graph, the x-axis (horizontal) represents time: sessions, days or dates. The y-axis (vertical) represents the measure of behavior, such as a count, a rate per minute, a duration or the percentage of correct trials. The y-axis starts at 0, and its label should say exactly what was measured and in which unit, so a reader never has to guess.

• Phase change lines: a dashed vertical line between the last session of one condition and the first session of the next. It tells the reader that something about the environment changed at that point.
• Condition labels: a short name centered above each phase, such as Baseline, FCT or Maintenance.
• Data paths are not connected across a phase change line. Each condition is its own data set, and a line drawn across the change would suggest a continuity that is not there. The path also breaks at a missed session, so a gap stays visible.
• Black and white, axes only: no gridlines, no box, no 3-D, distinct markers instead of colors. That is how JABA and Cooper, Heron and Heward draw them, and it still reads when photocopied.

Visual analysis of an ABA graph comes down to three questions, asked within each phase and then compared across phases. Level: where do the data sit on the y-axis, on average? Trend: are they going up, down or staying flat, and how steeply? Variability: how much do they bounce from one session to the next?

Then look at the change between phases. How quickly did the data change after the phase change line (immediacy)? How much do the data points of the two phases overlap? A clear, immediate change in level or trend with little overlap is the pattern that suggests the intervention made a difference. A baseline that was already trending in the direction you wanted makes the same intervention data much harder to interpret.

In supervision, a useful habit is to describe the graph out loud before deciding anything: “Baseline is high and variable, eight to twelve per session, no clear trend. After FCT started, aggression dropped within three sessions and kept trending down, while mands went up.” That sentence is the clinical reasoning, and the graph is the evidence for it.

Good visual analysis needs good data. The session data usually come from a paper sheet, such as a frequency or DTT sheet from the free ABA data sheet generator, and the targets often from a skills assessment like the VB-MAPP. Check the data with the IOA calculator before you draw conclusions from the graph.

The equal-interval line graph is the most preferred graph in ABA for session-by-session data, because it shows level, trend and variability at a glance and is what this ABA graph maker draws. Cumulative records, bar graphs and the semi-logarithmic Standard Celeration Chart each answer a narrower question, listed below.

• Cumulative record: each session adds to a running total, so the line can only go up or stay flat. The slope shows the rate of responding. Useful for skills that build up over time, such as the number of targets mastered.
• Bar graph: good for comparing totals or averages between conditions or people, such as mean percentage correct per phase. It hides session-by-session variability, so it does not replace the line graph for decisions.
• Semi-logarithmic chart, including the Standard Celeration Chart from Precision Teaching: the y-axis is logarithmic, so equal proportional changes look equal. It shows how fast a rate is speeding up or slowing down (celeration) and is the tool of choice for frequency-building work.

One ABA graph panel, one set of axes with up to three data series and up to eight phases, covers the single-case designs most BCBAs and students graph: A-B, reversal, changing criterion and alternating treatments. A multiple-baseline design needs one panel per tier, which you make here one at a time and stack. Here is how each design fits:

• A-B design: baseline, then intervention. The simplest picture of change, but on its own it cannot rule out other explanations.
• Reversal (A-B-A-B): the intervention is withdrawn and introduced again. Reuse the same phases in the grid and the graph labels each condition and draws the three phase change lines.
• Changing criterion: one intervention phase per criterion level. Use the criterion line for the current goal, or name each phase after its criterion.
• Alternating treatments: conditions alternate session by session, so they are drawn as separate series (one per treatment) without phase change lines, rather than as phases.
• Multiple baseline: the staggered panels of a multiple-baseline design need one graph per tier (per behavior, setting or learner). Make one graph per tier here and stack them in your report; the tool does not draw the connecting step line between panels.