Abstract: Could the 1936 Literary Digest Poll fiasco have been avoided? "Scrupulous bookkeeping" versus statistical adjustment of data
In 1972, George H. Gallup, the "father" of modern polling, wrote that the Literary Digest had "the evidence in hand to make a correct forecast in the 1936 presidential election." He was not the first, nor the last, to make this observation. All the Digest had to do, he went on, was to perform "a simple statistical correction". Why did the Literary Digest not adjust the data it published for all to see, thus avoiding embarrassment and, ultimately (in 1938), doom? Gallup saw this as "one of the ironies of polling history". This paper will explore the clash between the norms (e.g. "objectivity", "impartiality") of a long established profession (news reporting) and a relatively recent technology (statistical data analysis), which had not yet gained legitimacy among news reporters.• Back to CV.
