Previous knowledge and memory.

Abstract

This paper is concerned: (a) with the relationship between previous knowledge and memory, (b), with differences between recall and recognition, and (c) with set for time of memory test while learning. Subjects listened to 144 short statements, half of which were true and half false. They had to rate each statement on a five‐point scale for certainty of their truth or falsehood. A test of memory was then carried out for half of the statements, and another test was given a week later for the other half. In each test, half the statements had to be completed (recall), and each of the remaining statements had to be discriminated from a paired statement which resembled it in a standard way (recognition). While rating the statements, half the subjects did not expect a test of memory at all, a quarter were led to expect an immediate test, and the remaining quarter were led to expect a test in a week. After the first test of memory, no subjects were led to expect a second test. A control group of subjects took a test of ‘memory’ without hearing the statements.It was found that a subject was more likely to be certain that a false statement was false, than that a true statement was true. For both true and false statements, memory was best for statements about whose truth the subject was certain, and worst for statements about whose truth he had no idea. The control subject was much less successful than the experimental subject, but tended to score on the same statements. Previous knowledge can thus assist both genuine memory and also guess‐work.With memory for a particular word as a criterion, there was no difference between recall and two‐choice recognition when the word was in a statement about whose truth the subject was certain and correct, and memory was tested almost immediately. But recognition was more successful than recall when the word was in a statement about whose truth the subject had no idea, or when memory was tested after a week, although differences were only significant in the latter case.Recognition was found to be a less sensitive measure of previous knowledge than recall. Success in recognition was determined to a considerable extent by factors intrinsic to the test of recognition. Part of the reduced sensitivity of recognition was attributed to this cause.Immediate memory was best when a test of memory was expected immediately, and worst when no test was expected, but the differences were not significant. If anything, the trend for memory after a week was in the reverse direction.