| Abstrakt: |
Interest in educational technology has accelerated greatly during the last ten years with a consequent emphasis on evaluation and verification. Unfortunately, a good deal of the research work has yielded little in the way of unequivocal results. Some of the difficulties stem from the nature of educational measurement, as well as from the selection of inappropriate test‐statistics which inevitably make a number of assumptions about the data involved. This paper examines the nature of measurement, which is defined as “the assignment of numerals to objects or events according to rules,” the rules which must be followed, and the problems arising from the need to use indicants in order to measure learning. Four scales of measurement are recognised (nominal, ordinal, interval and ratio), and their characteristics and assumptions examined. The paper ends by considering how appropriate test‐statistics and power tests can be selected and employed in educational and psychological research. [ABSTRACT FROM PUBLISHER] |
