New Political History: Some Statistical Questions Answered

William G. Shade's ( 1981) "New Political History: Some Statistical Questions Raised" has two sometimes conflicting purposes: first, to remind historians to "think statistically" and to "give more self-conscious attention to the details and logic of research design," and second, to defend such ethnocultural historians as Ronald P. Formisano and Paul Kleppner against published criticisms. Too often confusing the former with the latter aim, Shade attains neither. His article is further compromised by distortions of other scholars' work and neglect of relevant literature published since 1974.1 With Shade's two major prescriptions - plan research carefully and use genuinely multivariate methods-we have no quarrel. Often breached in practice, these familiar commandments can never be repeated too many times. To his lack of conceptual rigor, to his employment of a series of either meaningless or misleading "tests," and to several of his methodological dicta, we do take exception-thus the necessity for this note