Identifying controlling variables for math computation fluency through experimental analysis: the interaction of stimulus control and reinforcing consequences.

Three children worked one-on-one with a teacher. Each session mixed easy math facts the child could answer in under two seconds with harder facts that took longer.

The teacher switched the type of praise the child earned. Sometimes the child got a point for every right answer. Other times the child got a point only for finishing a whole sheet. The team watched which praise kept the child working fastest.

When the problems were already easy and fluent, the children’s speed changed with the praise rule. The same child would speed up or slow down in clear steps.

When the problems were not yet fluent, the children’s speed stayed flat no matter how praise was delivered. Fluency, not praise, was the hidden switch that let the reinforcer show its power.

Lejuez et al. (2001) saw the same thing in a different way. They doubled the rate of reinforcers and watched computer tasks become harder to stop. Both studies show that how often and how you deliver a reinforcer decides how much punch it has.

Langthorne et al. (2007) push us to look one step earlier. They say to add a quick check for motivating operations—like hunger for attention—before we even start the analysis. Pair their MO scan with L et al.’s fluent-probe rule and you get a cleaner, faster picture of what is really driving the behavior.

Hoffmann et al. (2017) add a time twist: ten minutes of iPad beats thirty seconds of bubbles. Together these papers tell a simple story—match the item, the time window, and the child’s skill level or your data will stay flat.

Next time you run an experimental analysis, start with facts the learner can already answer in two seconds. If you see clear jumps across conditions, you know the reinforcer is real. If the line stays flat, do not blame the reinforcer—build fluency first, then test again. This one switch saves hours of guesswork.