Statistics in American psychology: the social construction of experimental and correlational psychology, 1900-1930

Abstract

Methodology makes visible to a scientific community the phenomena disclosed by research activity. The acceptability of methodology is inextricably tied to the acceptability of the data resulting from a particular method. Therefore it is clear that it is important not only to understand the methodology of a particular science, but of equal importance to understand the processes through which particular methodologies have become acceptable to a community of scientists. This thesis focuses on the processes through which statistical methods became acceptable to psychologists.I identify two competing research traditionsspecifically experimental and correlational psychology - and display their different interpretations and uses of statistics. I argue, however, that it would be wrong to credit the opposition between these research communities as owing to their conflicting ideas about the meaning of statistical methods. Rather, their conflict stems from differences over how science should be practiced, how labor within the research community should be organized, how knowledge should be generated, who should generate it, and who should apply it. In other words, I approach these conflicts over the interpretation (and uses) of statistics as reflecting differences in the intellectual, social and technological interests that operated within these research communities.