A Memory Study Employing a Factorial Experimental Design and the Analysis of Variance

Abstract

According to recent articles by Baxter (1), Dunlap (3), and Garrett and Zubin (9), psychologists have only begun to utilize the experimental designs and analytical techniques developed by Fisher and other investigators in agriculture. It is the purpose of this report to contribute another example of the application of these methods to a psychological problem. The investigation (10) upon which this article is based follows a series of experimental studies of and retention by English and several associates. These studies have become best known through the four most recent experimental reports by English and Edwards (4, 5, 7, 8), and two theoretical articles in a controversial exchange with Buxton (2, 6). It was the purpose of the present investigation to critically examine the and studies and to design further experiments incorporating improvements of which these studies stood in need, but utilizing the basic experimental technique which they had developed. It was the underlying purpose of English and his associates to arrive at a more carefully controlled experimental comparison of logical and rote memory. The extent to which their experimental method yields information which may be directly interpreted in terms of the basic problem has been questioned, but that is of no concern here. For the purpose of this article, it is feasible to begin by uncritically accepting their technique and not questioning its relevance to the basic problem. The general procedure used in the present investigation, is the same as that used by English and Edwards. The experimenter reads to the experimental subjects carefully selected subject matter over which a true-false test consisting of verbatim and summary items has been prepared. A verbatim, or V, item is as nearly as possible in the words of the text, covers a relatively small unit of thought, and is of such a nature that a correct answer is largely, theoretically might be purely, a function of rote memory. A summary, or S, item departs from the wording of the text, is based on the significance of larger portions of text, and is of such a nature that a correct answer indicates understanding of the point covered. The test, containing both types of items, is given immediately after presentation of the text. The same items are later used as a retest. This enables a comparison of the retention of and items-a comparison which was interpreted by English and his associates in terms of rote and logical memory. In the earlier studies the verbatim-summary comparison was based on the averages of subjects' scores on the two parts of the test. In later studies English and Edwards gave up the group averages and based the comparison on a new type of item score-test-retest patterns of and wrong. On any item which can be simply scored as right or wrong, any one subject could be on the first test and wrong on the retest (the RW pattern), wrong on