Toward an "Awareness" of the Relationship between Task Performance and Own Verbal Accounts of that Performance.

Hammonds (2006) ran lab tasks with college students. The goal was to see if people could learn even when they could not fully explain what they learned.

After each task, the students gave a verbal report. The team compared how well the students did with how complete their words were.

Students mastered the tasks, but their verbal accounts were often patchy. Learning happened even when the words did not match the actions.

The authors say "awareness" is just how much verbal behavior you can emit about the task, not a hidden mind state.

Lee et al. (2020) extends this idea. They showed that accurate feedback only helps when learners can see their results. If outcomes are hidden, even wrong feedback works, because the learner has no verbal access to judge accuracy.

Domjan (2026) gives historical backup. The review warns that Thorndike's law of effect is often misquoted in our teaching slides. Frank's move to treat "awareness" as verbal behavior fits the same call for conceptual clarity.

Alsop et al. (1995) shows a parallel effect: when reinforcement becomes intermittent, people follow verbal rules less. Together these studies say verbal control is fragile and does not have to line up with actual performance.

Do not trust a client's self-report as proof of learning. Probe performance directly, then ask for words second. If a learner can do the skill but cannot explain it, keep teaching the skill and add separate verbal training later. This split keeps your data clean and your program moving.