Useful experimental designs and rank order statistics in educational research

Abstract

Experimental educational research is of great impact because it illuminates cause-and-effect relationships by accumulating empirical evidence. The present article does not propose new methods but brings three useful experimental designs as well as appropriate statistical procedures (rank order statistics) to the attention of the reader to conduct educational experiments, even with small samples. By means of their systematic use combined with the process-product paradigm of experimental educational research, the influence of essential variables (teacher, context, and process variables) in schools, universities, and other educational institutions can be investigated. The statistical procedures described in this article guarantee that small samples (e.g. a school class) can be successfully used, and that product variables (e.g. knowledge, comprehension, transfer) are only required to meet the criteria of an ordinal scale. The experimental designs and statistical procedures are exemplified by hypothetical data and detailed calculations.