Use four graphs to compare level, trend and variability across observations. Review the visible patterns and why explaining their cause requires broader evaluation evidence.
Read level, trend and variability separately. Describe the visible pattern and use the broader evaluation evidence to assess its cause.
Start with the axes: these graphs show occurrences per 10-minute observation, ordered from left to right. Then inspect the height of the points, their direction and how much they fluctuate. Each point represents one observation. Connecting lines help you follow the sequence of recorded values.
Key terms
Phase
A period of observations under a particular condition, separated when that condition changes.
What to look for
Distinction
Look for
How to decide
Level and trend
Typical height and direction over time
Separate where values sit from whether they rise or fall.
Variability
Spread and fluctuation between points
Look beyond the first and last values.
Causal limit
Other possible explanations and design
Describe association unless the evidence supports more.
Worked examples
Reading a pattern
Follow the points from left to right
An RBT counts the same defined response during six 10-minute observations. Read the graph to describe how the counts change over time. The response definition and collection method remain the same.
Six observations
Read across the observations, then follow each point to the count scale. Every observation lasts 10 minutes.
Relevant facts
The horizontal axis orders observations 1 through 6; the vertical axis counts occurrences.
Each point is two occurrences lower than the one before it: 12, 10, 8, 6, 4, 2.
The graph shows a decreasing trend.
12 → 10 → 8 → 6 → 4 → 2
The points consistently move downward as observations progress. There is little fluctuation around that downward path, even though the first and last values are far apart. This is a decreasing trend with low variability around the trend. Explaining the decline requires additional evaluation evidence.
Comparing patterns
Same mean, different variability
Graphs A and B show the same response measure in two six-observation series. Both use 10-minute observations and identical axis scales. Compare the typical level and the fluctuation across each complete series.
Graph A
Every plotted value is 5.
Graph B
Counts repeatedly rise and fall. The mean is still 5.
Relevant facts
Both series total 30 occurrences across six observations, giving each a mean of 5.
Graph A stays at 5; Graph B repeatedly moves between low and high counts.
Both have a mean of 5. Graph B has greater variability.
Mean for each: 30 ÷ 6 = 5. Range: A = 5 − 5 = 0; B = 9 − 1 = 8.
Graph A is flat. Graph B repeatedly rises and falls across the series. Their shared mean summarizes level, while the individual points show the difference in variability. Compare graphs on the same scale so visual height and spread are meaningful.
Interpreting a change
Describe the level change
A new procedure begins between observations 3 and 4. The RBT uses the same response definition and 10-minute observation length throughout. The graph records the timing of the change, but no other evaluation evidence is supplied.
Before and after a procedure change
Compare the two clusters of points. The dashed boundary marks the timing of the procedure change.
Relevant facts
Before the boundary, counts are 12, 11 and 12. After it, they are 5, 4 and 5.
The dashed line marks when the new procedure begins. A causal conclusion requires evaluation of the broader evidence.
Counts are at a lower level after the procedure begins. Establishing the reason for that decrease requires further evaluation evidence.
Before: 12, 11, 12 → after: 5, 4, 5.
Within each phase, values stay in a relatively narrow band. The main visible change is a drop in level between phases. Consider other changes and the evaluation design when assessing what caused the decrease.
What changes the answer?
Return to Graphs A and B. Keep every point and both scales unchanged and change the question.
Graph A: 5, 5, 5, 5, 5, 5.
Graph B: 1, 9, 2, 8, 1, 9.
Which series has the higher mean count?
Neither: both means are 5.
Each series totals 30 across six observations. The high points in B are balanced by its low points.
Which series fluctuates more between observations?
Graph B.
A stays at 5; B repeatedly moves between low and high values. The series share a mean and differ in variability.
Identify what the question asks before choosing a graph: typical level, direction over time, fluctuation, or a causal explanation require different evidence.
Remember
Read the axes and units first. Level is height; trend is direction; variability is fluctuation around the pattern.
Assess trend across the series. Compare the individual values to see how patterns differ around the same average.
A change after a procedure begins describes timing. Causation needs more evidence.
