Hierarchy and spatial contagion: population in American cities between 1990 and 2010
使用空间分层回归模型，分离空间传染和城市等级对人口规模的影响，发现大城市以集聚和空间互补为特征，小城市则更依赖空间互补和等级位置。
We use a spatial hierarchical regression model to isolate the effects of spatial contagion and urban hierarchy on population levels and apply this model to a unique dataset of American urban areas from 1990 to 2010 that includes all but the smallest rural communities. By disentangling the effects of urban hierarchy and spatial contagion on population levels, we find that large cities are characterized by urban agglomeration and spatial complementarity, while small cities are characterized by a greater degree of spatial complementarity and position in the urban hierarchy.