摘要：

The increasing diversity of methods of experimental control and measurement in studies of human learning indicates the need for an experimental approach to methodological problems. The reviewer emphasizes the value of a standardized methodology, based on the most valid and reliable methods of control and measurement, in increasing the validity and reliability of isolated experiments and in increasing the opportunity for a true systematization of the results of many experiments. Criteria for evaluating experimental methods are discussed under the headings validity, reliability, and conformity with the conditions of other experiments that have major systematic significance in the field. Work on the validity of experimental methods is reviewed in terms of the experimental error and the degree to which the experimental method represents the operations used to define the phenomenon indicated in the interpretation of results. The writer discusses the nature and application to experimental methods of learning of two criteria of reliability: (1) "that the experimental method is more adequate the smaller the variable error in the measurements obtained when it is used" and (2) "that a method is more adequate the greater the possibility of an accurate statistical analysis of the variable errors in the data obtained when that method is used." Bibliography of 146 titles. (PsycINFO Database Record (c) 2012 APA, all rights reserved)

展开