Pragmatic Application of Risk Ratios in Behavior Analysis
Pick the risk-ratio formula that matches your single-case or group question—both are valid when used right.
01Research in Context
What this study did
Joslyn et al. (2024) wrote a position paper. They defend an alternative way to calculate risk ratios for single-case data. They do not run new experiments. They explain why the formula still helps you decide if a treatment works.
The paper answers critics who say the formula is wrong. It shows when the old and new formulas give the same answer and when they do not.
What they found
The authors found both formulas are right, but they answer different questions. The traditional formula works best when you compare groups. The alternative formula works best when you track one client across phases.
You can pick the formula that matches your question. The paper gives clear rules for choosing.
How this fits with other research
Hinson (1988) said comparison studies ask the wrong question. Joslyn et al. (2024) agree, but they add a tool for when you must compare. The new paper does not reject the old warning—it gives you a safer way to compare when needed.
Ferron et al. (2017) showed masked visual analysis spots strong effects in single-case designs. Joslyn et al. (2024) offer a second tool: the risk ratio. You can use both. Masked visual analysis tells you if an effect is real. The risk ratio tells you how big it is.
Herrnstein et al. (1979) built a math model for VI schedules. Joslyn et al. (2024) build a math tool for single-case decisions. Both papers aim to make behavior analysis more exact.
Why it matters
You no longer need to fear risk ratios. If you run a single-case design, use the alternative formula to show parents and payers how strong the change is. If you later pool cases, switch to the traditional formula. One paper gives you both maps.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Next time you graph a reversal design, plug the data into the alternative risk-ratio formula and add the number to your summary.
02At a glance
03Original abstract
Joslyn and Morris (2024) recently published a tutorial on adapting and applying risk ratios to within-subject behavioral data. Newland (2024; this issue) offers a commentary on Joslyn and Morris that expands the discussion of risk ratios and how they are typically applied in group comparison research. He is also critical of the approach described by Joslyn and Morris, and advocates against its use primarily because it differs from the traditional calculation. Although we agree with many of Newland’s assertions about the benefits of the traditional approach, we disagree that it is the only method of evaluating relative risk that may be useful to behavior analysts. In this response to Newland, we summarize and respond to his concerns, discuss our own concerns with his perspective, consider variables that may affect the relative utility of the two approaches, and provide concluding remarks.
Perspectives on Behavior Science, 2024 · doi:10.1007/s40614-024-00425-1