The effects of migration, place, and occupation on adult mortality in the American north, 1740-1880.

D emographic indices are typically constructed out of aggregated statistical data derived from different subpopulations, and that is true not only on the national level, but also on the local level. Such indices are therefore potentially misleading means, because they obscure differences between component subpopulations. Moreover, even when attempts are made to disaggregate national statistics, the units used have many of the same faults. First, they mask much internal variation, and, second, it is often assumed that the units themselves are self-contained. Frequently, the units are administrative subdivisions of the larger nation and are conceived of as places, but migration in and out of these so-called places is rarely considered a factor influencing mortality within them. The challenge of demography, then, depends on knowing the changing composition and different vital rates of subgroups within mobile populations and on uncovering the various strands of the population and weighing their differences and interdependencies. But given the general lack of detailed data for even small subpopulations, such a task is often very difficult. Some might argue that the smaller the group and the smaller the variation within it, the more meaningful is the mean. But even micro populations are not fixed and uniform; within single families, there may be considerable variation (Adams and Kasakoff 1999). Others might argue that on the level of the large nation-state, small differences become inconsequential, especially as modern communications and standards of public health may tend to create a uniform population. But such homogeneity does not exist at at any level of analysis that we have examined. The underlying social reality is so complex that the usual demographic rates have to be taken as no more than summaries of dubious convenience. In this article, we use individual-level data to present three examples of the ways in which migration affected adult life expectancy in the American North from 1740 to 1880. We found that to understand mortality changes, particularly among adults, we had to consider the accumulated lifetime experiences of the population, experiences that for many, if not the majority, took place in more than one locality and occupation. People moving spatially and socially brought their demographic histories with them. Depending on the particular disease environment and the source of the migrants, new arrivals might have either raised or lowered mean life expectancy in their new location. Such mixtures are not only difficult to sort out statistically, but they also raise questions about the demographic stability of places. There is, therefore, a fundamental tension between looking at the lives of individuals who are parts of migration streams linking spatial and occupational units and the units themselves. Are the rates in a particular spatial unit the result of the migrants who entered into it or the people born there? (We will address this question later by means of the example of cities.) The following examples help us develop the argument that migration is important in the study of mortality, especially as it affects those analyses in which it cannot be studied directly.