A unified theory of all-or-none and incremental learning processes via a new application of study-test-rest...

The all-or-none (AON) versus incremental learning debate was newly examined using heart rate (HR) and galvanic skin response (GSR) under varied study-test-rest (STR) presentation programs. Fifty university students learned a list of 20 consonant-vowel-consonant-2-digit pairs in a simple learning situation. Three item types were analyzed: items learned, never learned, and in-between. Learning performance moved from the unlearned to the learned state in one jump in the AON fashion, despite nonsignificant incremental trends during some precriterion trials. New HR and GSR phenomena were unveiled: Persistent relaxation occurred abruptly a few trials before actual learning. The relaxation attained unconsciously, after intense conscious effort, might facilitate mastery. Unlearned items lacked such jumps for all 3 response measures. Of all conditions examined, STTTTTTT produced the most efficient learning with the greatest attention (HR), most relaxation (GSR), and least test anxiety (interprogram comparisons). By using Estes' mathematical derivations, a unified theory is advocated: AON learning can be regarded as a special case of incremental learning. The former is associated with very simple learning situations, whereas the latter involves complex learning situations. The 2 positions are not irreconcilable; instead, they represent 2 extremes of such experiments, the simple and complex learning situations.