Nonreactive testing: Evaluating the effect of withholding feedback in predictive learning.

Area of Science:

• Cognitive Psychology
• Associative Learning
• Human Cognition

Background:

• Cue-outcome learning is typically assessed using no-feedback testing to prevent new learning.
• The assumption is that withholding feedback during testing does not influence participants' understanding of relationships.

Purpose of the Study:

• To investigate whether "no-feedback" testing in associative learning experiments can introduce new learning or alter existing causal judgments.
• To determine the conditions under which no-feedback testing might affect learning assessments.

Main Methods:

• Four experiments were conducted where participants learned cue-outcome (food-allergy) relationships.
• Testing involved presenting cues without outcome feedback, with variations in how missing information was presented (e.g., "?").
• A factorial design manipulated cover stories and feedback presentation to isolate effects.

Main Results:

• Withholding feedback had no impact on causal ratings in initial experiments.
• Presenting missing feedback as a "?" within the scenario led to a regression toward intermediate causal ratings.
• The "?" feedback, rather than the cover story, was the primary driver of this regression-to-baseline effect.

Conclusions:

• No-feedback testing is generally appropriate for assessing associative learning.
• The validity of no-feedback testing depends on how missing information is presented; avoiding uncertainty is crucial.
• Careful procedural design is necessary to ensure no-feedback testing accurately reflects learned associations without introducing bias.