How do you determine if your intervention is working? Many of us analyze data on a visual display. As we learned in the last article, we can get powerful information about our data through level (average rate of responding) and celeration (quantified trend).
Both of these types of analytics are within condition analysis.
Within Condition Analysis
What can we discover by analyzing data within conditions — baseline, intervention A, intervention B, and so forth? “Data within a given condition are examined to determine (a) the number of data points, (b) the nature and extent of variability in the data, (c) the absolute and relative level of the behavioral measure, and (d) the direction and degree of any trend in the data” (Cooper, Heron & Heward, 2007, p. 150).
For each condition, at least 3-5 data points are needed. Without multiple data points, core conditions b-d (above) cannot be measured — and you can’t easily tell the effect of your intervention.
On the Standard Celeration Chart (SCC), within-condition change analytics aren’t just depicted visually, but numerically. In other words, statistical information — like level and celeration — is included on the standard chart.
Level is “The average of data within a condition” (Kennedy, 2005). Level can be calculated in 3 ways: by the arithmetical mean, median, or geometric mean. Median and geometric mean are best, as they are less affected by outliers — wild, far-off numbers — as Dr. Kubina shows.
Level is helpful because it shows, at a glance, the intensity of the behavior through the average rate of responding. It’s valuable to know the representative level for SIBs, Mands, letter sounds, and so forth. It’s valuable to know which of 10 students has a high rate of “squirms during circle time” or low rate of “hear-writes spelling word” so you can easily flag them.
Celeration is “the change in frequency over time” (Graf & Lindsley, 2002). Celeration includes the information that trend offers (up, down, flat), plus it quantifies the speed of change to provide a learning rate. For instance, x2.0 means the behavior has doubled. And ÷2.0 means the behavior has halved.
Further, behavior analysts can extend the celeration line to project when the learner will or will not make their goal. In other words, you can see if your client will learn letter sounds before the end of the month. You can see if your client will extinguish an escape behavior before the end of the quarter.
And… if a learner isn’t on track, you can instantly decide to make a change to help the learner.
So what can I do with level and celeration?
With the Standard Celeration Chart, you can
- Know the average rate of responding (level) within a condition
- Compare the average rate of responding in 2 or more conditions
- Know the speed of behavior change (celeration) within a condition
- Compare the speed of behavior change within 2 or more conditions
With these comparisons, you are empowered to make data-driven decisions — based not on vague line trends, but quantified level and celeration.
Already watched the webinar video? Take the BACB Type 2 CEU quiz here. Worth 1 credit.
Cooper, J.O., Heron, T. E. & Heward, W. L. (2007). Applied behavior analysis. Upper Saddle River, N.J.: Pearson/Merrill-Prentice Hall.
Graf, S., & Lindsley, O. (2002). Standard Celeration Charting 2002. Poland, OH: Graf Implements.
Kennedy, C. H. (2005). Single-case designs for educational research. Boston: Pearson/A & B.
*Paid content by Chartlytics.