Abstract

This article was devoted to an attempt to elucidate the statistical approach to research problems in education and psychology. In so doing, the functions of statistics, the attitudes of behavioral scientists towards statistics, the role of statistics in behavioral science etc. were examined. The author contends that statistics is first of all a theory of error - not a theory of error in the concrete - but a theory of the abstract or structural characteristics of error. The theory uses such characteristics to identify error, to measure its magnitude, and to provide ways of taking it into account. Statistical tests are used to compare obtained results with chance expectations. Whenever a research study is done and statistical results are obtained, the results must be checked against the statistical results expected on the basis of chance. A large part of the activity of mathematical statisticians is devoted to determining what these chance expectations are. On the basis of the above discussion, the author analyzes the reasons : (a) why university teachers and students must develop some mastery of statistics. (b) why statistics are important in educational and psychological research, as tools of scientific thought or as devices for summarizing results, drawing conclusions, making predictions and analyzing causal factors of complex and otherwise bewildering events.