| Abstract: | ABSTRACT Previously it was reported that regional variations in well-being (poverty, per capita income, and family income) among Appalachian counties did not originate from regional variations in urbanization, but from regional differences in well-being among nonmetropolitan counties. It was argued that southern Appalachian counties had higher levels of well-being at the end of the 1980s because nonmetropolian counties in southern Appalachia experienced greater economic growth during the 1980s than did nonmetropolitan counties in other Appalachian regions. In this paper these data are reanalyzed to test to what extent the original findings are affected by the presence (and failure to control) spatial autocorrelation. Using a spatial lag model it is shown that correcting for spatial autocorrelation statistically altered the original results. However, substantively, the conclusions from the original analysis did not change: regional differences in county well-being in Appalachia are largely the product of regional differences among nonmetropolitan counties, even after correcting the model for spatial autocorrelation. |
