Levels of Replication
Plot every response as it happens—flat line means control, wiggles mean you need to adjust.
01Research in Context
What this study did
Iversen (2025) wrote a how-to paper, not an experiment.
He says every trial is a tiny replication. If the learner’s answers jump around inside one session, control is weak.
He tells us to graph minute-by-minute data so we can see the jumps.
What they found
The paper itself has no new numbers.
The big idea: replication is not just running the whole study again. It also happens every time the stimulus repeats inside one sitting.
When the line on your graph stays flat, stimulus control is replicating. When it wiggles, it is not.
How this fits with other research
Vladescu et al. (2021) already did this. They swapped set sizes trial-by-trial and watched learning speed change within the same hour. Their minute-level data showed the same pattern Iversen predicts.
Jessel et al. (2020) ran 10-minute IISCAs. Each quick test was a stack of mini-replications. Control showed up inside one session, just like Iversen says it should.
Sullivan et al. (2020) tracked resurgence second-by-second. The jumps in problem behavior were tiny replications of extinction bursts. Their fine-grain plot is a live picture of Iversen’s idea.
Why it matters
You can check your program before the day ends. Drop the data into a cumulative graph while the child is still at the table. If the curve bends the wrong way, fix the prompt or the reinforcer right then. No more waiting for next week’s probe.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Start a within-session cumulative graph for one skill; stop the session if the line plateaus or drops.
02At a glance
03Original abstract
In behavior analysis, replication is one of the most fundamental strategies used to establish generality of results. However, replication is not restricted to just repeating an experiment, whether directly or systematically. Replication is also a defining component of many procedures used in individual experiments in behavior analysis. For example, some methods, such as single-stimulus discrimination procedures, exhibit direct control over behavior with a series of mini-AB designs (trial and intertrial periods) repeated multiple times within a single session. Once stimulus control is acquired, replication is demonstrated each time stimulus presentation is followed by the appropriate response. Conditional discrimination methods have the same structure with more trial types or stimuli that control response selections. So, replication is built in not only across experiments but also in within-session experimental designs. This will be illustrated by examples showing fine-grained data analysis. The illustrations will confirm Pennypacker’s emphasis that moment-to-moment analyses of behavior are essential to successful replication.
Perspectives on Behavior Science, 2025 · doi:10.1007/s40614-024-00431-3