The Hawthorne Experiments

Abstract

The Hawthorne experiments performed fifty years ago concluded that variations in factor workers' productivity were to be accounted for in terms of human factors. Franke and Kaul (1978) retrieved the original data of these remarkable experiments and showed through application of stepwise multiple regression analysis that over 90% of the observed variance in worker output could be accounted for in terms of the measured variables of the First Relay Experiment, rather than human factors. The present effort reinterprets several of the coefficients reported by Franke and Kaul and suggests that their conclusions rest on a strong algebraic component of the analysis as well as on violated assumptions, confounded variables, index correlations, the meaning of measurement, and so forth. Reanalysis is here performed with SSA-1, an ordinal technique, leading to conclusions in support of those of the original researchers.