Effects of modeling versus instructions on sensitivity to reinforcement schedules.
A quick model of how to track reinforcement schedules keeps kids accurate when the rules later change.
01Research in Context
What this study did
The team worked with 8- to young learners, some with ADHD and some without.
They taught the kids to track two reinforcement schedules: one rich, one lean.
In one condition, the teacher simply told the kids how to respond.
In the other, the teacher showed a quick model of the tracking moves.
Later, the schedules flipped so the rich became lean and vice versa.
What they found
Both groups learned the first schedules fast.
When the schedules swapped, only the kids who saw the model kept up.
The instruction-only kids kept pressing at the old, now-wrong rate.
How this fits with other research
Bondy (1982) found that once you train a response with prompts and praise, kids stick with it even when a new model appears. Burack et al. (2004) flips the order: show the model first and it beats later instructions.
Wilson et al. (2020) also saw modeling beat prompting for teens with autism learning cooking skills. The same modeling edge shows up across ages and tasks.
Killeen (2023) theory says richer schedules build behavioral momentum. Burack et al. (2004) proves that momentum can be broken if you first teach with a strong model instead of weak instructions.
Why it matters
If you want learners to stay flexible when reinforcement rates change, show them what to do, don’t just tell them. A five-second demo now saves you from retraining later. Use modeling at the start of any new schedule game, token board, or point system. Your future self will thank you when the system shifts and the kids still track correctly.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Before starting a new token board, demo three correct token placements yourself and have the learner imitate once.
02At a glance
03Original abstract
This study examined the effects of modeling versus instructions on the choices of 3 typically developing children and 3 children with attention deficit hyperactivity disorder (ADHD) whose academic responding showed insensitivity to reinforcement schedules. During baseline, students chose between successively presented pairs of mathematics problems associated with different variable-interval schedules of reinforcement. After responding proved insensitive to the schedules, sessions were preceded by either instructions or modeling, counterbalanced across students in a multiple baseline design across subjects. During the instruction condition, students were told how to distribute responding to earn the most reinforcers. During the modeling condition, students observed the experimenter performing the task while describing her distribution of responding to obtain the most reinforcers. Once responding approximated obtained reinforcement under either condition, the schedules of reinforcement were changed, and neither instruction nor modeling was provided. Both instruction and modeling interventions quickly produced patterns of response allocation that approximated obtained rates of reinforcement, but responding established with modeling was more sensitive to subsequent changes in the reinforcement schedules than responding established with instructions. Results were similar for students with and without ADHD.
Journal of applied behavior analysis, 2004 · doi:10.1901/jaba.2004.37-267