a method that prevents the production of incorrect answers during the learning period. Specifically, learning occurs across several sessions but memory is not tested until the last session. It is thought to be more efficient than standard trial-and-error learning because it eliminates interference.