Decomposing the Effects of Spatial and Infrastructural Connectivity on the Development of the Northern Plains

• In Event: 442 - Section on Methodology Paper Session. Open Topic II

Abstract

In this paper we examine population growth on the Northern Great Plains between 1880 and 1930. We are interested in the idea that, all other things being equal, neighboring counties tended to exhibit similar patterns of growth. In practice, however, the notion of proximity was not simply a matter of geographic distance. Counties bound together through infrastructural ties were arguably “closer” to one another than counties that simply happened to be adjacent to one another. Drawing on a unique spatial dataset that depicts the geographic location of rail lines at ten-year intervals across the period in question, we distinguish between the effects of spatial and infrastructural connectivity by decomposing traditional measures of autocorrelation. Infrastructural effects are further decomposed so as to allow for variation across rail lines, reflecting the fact that the railroad network was actually comprised of a set of overlapping networks, each of which was tied to a specific corporation. To the extent that these organizations were sufficiently different from one other, we expect to find considerable heterogeneity in the corresponding set of network effects. We find that not only were patterns of population growth in any given county associated with patterns of growth in neighboring counties, but that the degree of association varied depending on which rail line linked the counties in question. We find that the magnitude of the observed infrastructural effects peaked just after the turn of the century at which point they began to decline.

Authors

• Adam Slez, University of Virginia

• Katherine J. Curtis, University of Wisconsin-Madison