Accuracy versus speed in the generalized effort of learning-disabled children.
Reinforce accuracy first and the skill will pop up in new places; reinforce speed first and you only get faster trained sheets.
01Research in Context
What this study did
The team worked with late-elementary kids who had learning disabilities.
They used a token economy. Kids earned points for reading fast or for reading correctly.
The study switched the payoff rule every few days so each child served under both rules.
What they found
When points came only for correct answers, accuracy rose and the gain spread to new sheets the kids had never seen.
When points came only for quick answers, speed rose but the gain stayed on the trained sheets.
The accuracy bonus kept helping even after the points stopped.
How this fits with other research
Cordeira et al. (2022), Schneider et al. (2022), and Wong et al. (2022) all show the same pattern in kids with autism: looser or target-level mastery rules cut teaching time without hurting long-term recall.
Sprague et al. (1984) adds that the dimension you pick—accuracy or speed—decides what generalizes.
Together the four papers say: set the mastery gate on accuracy, not speed, and you get both faster acquisition and wider carry-over.
Why it matters
If you run reading fluency drills, tie the reward to correct answers first. Once accuracy is solid, you can add speed goals without losing the broad transfer. This one switch can shorten the number of lessons you need before the skill shows up in untaught material.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Put the token or point on a correct answer only; add a speed bonus only after the child hits 90% accuracy for two sessions.
02At a glance
03Original abstract
Reinforcement of effortful performance in a given academic task has been found to increase the subsequent performance of other academic tasks. The learned-effort hypothesis assumes that individuals learn which dimensions of task performance are correlated with reinforcement of high effort, and generalize across tasks. Therefore, reinforcement of increased effort in a given dimension of one task should result in greater generalized effort in the same dimension of transfer performance than in another dimension. In accord with this view, preadolescent learning-disabled students who received points for high reading accuracy subsequently produced more accurate drawings and stories than did students whose points had been based upon high reading speed or upon mere completion of the reading task. Students who received points for high reading speed subsequently constructed stories more quickly than did children whose points had been based upon high reading accuracy or upon reading-task completion. Consistent with the more explicit and frequent feedback for accuracy than for speed in most academic tasks, generalized accuracy was much more durable than generalized speed.
Journal of the experimental analysis of behavior, 1984 · doi:10.1901/jeab.1984.42-19