Argument

How did statistical representativeness first appear in the United States? To answer this question we turn to the archives of the Department of Agriculture. Since the 1860s, agricultural statisticians have been working to come up with representative groups of farmers able to answer questions about crop production. The archives describe the methods and tools used to devise representative samples. But the explicit justifications of why a given sample is particularly representative are nearly always missing. Where can we find explanations for why one sampling method is representative and not another? Where can we find the why when we have only the how? This paper argues that we can turn to the theorists of democracy for answers, because selecting a sample and selecting a group of representatives have a lot in common.