Self-instruction: An analysis of the differential effects of instruction and reinforcement.

Researchers taught elementary students a short self-instruction script for schoolwork. They split the kids into two groups. One group got praise and tokens for using the script. The other group got praise only for right answers, not for using the script.

All kids first learned the script in short practice sessions. Then they worked alone while staff watched. The study tracked how many answers each child got right.

Every child who learned the script jumped from low to high accuracy, no matter which group they were in. The group that got no rewards for using the script kept their high scores later.

A comparison group that never got the script stayed at low accuracy until teachers finally taught them the same steps.

Green et al. (1987) ran a similar test with adults who have intellectual disability in a vocational program. Self-instruction again beat external instruction, and the gains lasted. The pattern looks the same across ages and settings.

Yet Billings et al. (1985) warns that the power may sit in stating goals aloud, not in handing out self-given tokens. Their college students improved most when they simply told others their study targets. No extra self-reinforcement helped. Together the three studies say: teach the self-talk script, let learners say their goal, then you can fade the prizes.

Wan Yunus et al. (2021) extends the idea to children with autism. A broader self-regulation package cut challenging behavior and lifted school performance. The core message—student-driven rules work—holds up across diagnoses and decades.

You can run this script in any classroom. Teach the child to stop, read the question, ask “What do I do next?”, then answer. After a few days, drop the tokens and just mark right answers. The child keeps the skill and you save reinforcement budget. Use it for math sheets, reading tasks, or even vocational jobs with older learners.