In lieu of an abstract, here is a brief excerpt of the content:

Journal of Interdisciplinary History 30.4 (2000) 667-669



[Access article in PDF]

Book Review

Human Demography and Disease


Human Demography and Disease. By Susan Scott and Christopher J. Duncan (New York, Cambridge University Press, 1998) 354 pp. $74.95

The approach followed in this book is the systematic application of time-series analysis to British yearly statistical series of the past. The case study that illustrates much of the discussion focuses on Penrith, a poor parish of Cumberland, and the authors claim they have realized [End Page 667] "the first quantitative, integrated study of the population dynamics in a single human community" (114). They also look at urban communities where wool prices are related with mortality. They envision, on the basis of many such studies, the treatment of England as a meta-population, that is, a population of populations each with their own economic, epidemiologic, and meteorological characteristics. They analyze a number of time series of disease incidence, from Bills of Mortality to nineteenth-century registration statistics. With this simple instrument, they tackle many fascinating health issues of the past. The breadth of their effort is impressive, even when their inferences are not persuasive. Their use of demographic terminology is unorthodox (see the confusion of cohorts and periods, as when they tabulate temperature by cohort (98)), and their methodologies are often insufficiently laid out. How did they calculate female life tables from family reconstitution (68)?

The basic approach consists of extracting from the available series, with the help of computer-based statistical techniques, a succession of cleaned sinusoidal curves with measurable wavelength (expressed in years) and amplitude. Filters are used to eliminate random fluctuations, revealing cycles that would not have been visible to the naked eye. Several time series are compared, and their cross-correlation tested for significance. The authors warn repeatedly that correlation does not imply causality, but in their analysis of Penrith, they come dangerously close to inferring causality, anyway. They rely almost exclusively on the high significance of cross-correlation coefficients, and pay no attention to the size of the effects. Since the statistical techniques reveal cycles that are not apparent through casual observation, the issue of the strength and importance of the connections arises.

Parish statistics provide the basic demographic series of baptisms and burials (distinguishing between infant and adult burials.) The other time series are national wheat prices and records of weather conditions for Britain. British historical demographers have often commented on the inadequacy of national wheat prices to describe local food availability. The series reflect, to a large extent, the price at which institutions were purchasing grain, and much of the population did not rely on market prices for their food. The authors justify their use of these series for Penrith by the observation that "the population cycles correlate well with national grain price indices" (114). This explanation, however, makes the argument circular.

The book suggests that oscillations of wheat prices and winter temperatures were driving the different demographic cycles in Penrith. Since adult mortality was correlated with grain prices, expensive bread must have spelled malnutrition for pregnant or breast-feeding women, leading to deprivation of vital nutritional elements in the womb during early life and high infant mortality. The argument takes too much for granted, and it derives some of its intellectual background from a controversial view of historical populations constantly teetering on the edge of famine and severely affected by malnutrition. This approach can [End Page 668] clearly contribute to historical studies when used cautiously, but it often claims too much on the basis of too little evidence.

Etienne van de Walle
University of Pennsylvania

...

pdf

Share