I found this story about historiography from Rodney Stark’s book Cities of God: The Real Story of How Christianity Became an Urban Movement and Conquered Rome (pp. 209-211) fascinating, and wanted to share it with you. Here is the quotation: 

In 1962, Arthur Schlesinger Jr.–on leave from the Harvard history department to serve as a White House intellectual for John F. Kennedy–told an assembled audience of American scholars that “almost all important [historical] questions are important precisely because they are not susceptible to quantitative answers.” Such arrogance thrilled many of his listeners, as clever nonsense so often does. For others it prompted reflections on how someone so poorly trained had risen so high in the profession of history. In truth, many of the really significant historical questions demand quantitative answers. They do so because they involve statement of proportion: they turn on words such as none, few, some, many, most, all, along with never, rarely, seldom, often, usually, always, and so on.

… let us turn back, to Arthur Schlesinger and the book that made his reputation: The Age of Jackson. For decades before Schlesinger wrote, the central question posed about “Jacksonian democracy” was how Old Hickory had managed to motivate millions more Americans to vote in his presidential elections than ever had done so before. All the stars of American historiography had addressed the matter, including Charles and Mary Beard, Richard Hofstadter, and John Back McMasters. A remarkable jumble of explanations had been offered, but everyone agreed with the Beards that Jackson was swept into office by “the roaring flood of the new democracy.” Thus, Schlesinger launched his career by attempting to explain the “immense popular vote” received by Jackson in 1828 when he was elected by a “mighty democratic uprising.” Notice that Schlesinger was not content simply to assert that Jackson was elected. He stressed the proportion of the victory–it was “immense” and “mighty.” His book was awarded the Pulitzer Prize in 1946 because reviewers found Schlesinger’s explanation of Jackson’s huge appeal to the ‘common man’ so convincing.

But trouble soon arose. In 1960, two years before Schlesinger’s expressions of contempt for quantification, Richard P. McCormick bothered to count the votes. He demonstrated conclusively that what was notable about the Jacksonian elections was low voter turnout! There was no “mighty” or “immense” outpouring. More votes had been cast in many previous elections. What seems to have so misled historians for so long was that these were apparently the first presidential elections in which attention was paid to the total popular vote, as opposed to merely reporting the results of the electoral college. Confronted with large numbers of votes, no one bothered to calculate whether these were many more or fewer than usual. As a result, generations of historical analysis was patent nonsense, having been devoted to explaining something that hadn’t happened!

… although McCormick’s expose was published in the American Historical Review, the most distinguished journal in the field, it was generally ignored, and many textbooks continued for several more decades to discuss Jackson’s immense appeal. To the best of my knowledge, Schlesinger never recanted.