Maximization of reinforcement by two autistic students with accurate and inaccurate instructions.

Two autistic students picked between tasks. One task came with accurate instructions. The other came with instructions that were wrong.

The students earned a reinforcer every time they chose the task that paid off, no matter what the instructions said. Later the pay switched to only sometimes.

When the reinforcer came every time, both students quickly followed either the accurate or the inaccurate rule and got the most tokens.

When the reinforcer came only sometimes, one student kept maximizing. The other student stopped maximizing and began to pick at random.

Nottingham et al. (2020) later showed that giving extra information every other trial, not every trial, helped autistic kids learn faster. Their result lines up with the first half of Newman et al. (1991): intermittent can still work.

El-Boghdedy et al. (2023) went further. They faded the teacher out and thinned reinforcers to almost nothing. All three teens stayed on task. Their success shows you can keep intermittent schedules working if you fade slowly and add other supports.

Dugdale et al. (2000) and Galizio et al. (2020) show the flip side. They taught autistic learners to choose variety instead of repetition. Together these studies say: autistic learners can follow rules, break rules, or invent new rules depending on what earns the payoff.

You now know that accurate instructions are not sacred. Autistic learners can use wrong rules just fine when the payoff is steady. Before you thin reinforcement, check if the learner still gets enough wins. If not, add prompts, delay slowly, or mix in other supports so the rule keeps working.