Spatial Regression Models for the Social Sciences. Jun Zhu
1970 to 2010 in 1,837 Wisconsin MCDs and its relevant factors based on a geographically referenced longitudinal data set. The second example relates migration from 1995 to 2000 to individual, household, and community characteristics in Wisconsin. The third example examines poverty in association with demographic characteristics and socioeconomic conditions from 2000 to 2010 at the county level in the contiguous United States. Readers can apply our research procedures to their research at various levels of units of analysis, such as countries, regions, counties, census tracts, metro/nonmetro areas, neighborhoods, communities, block groups, and others.
This book is composed of eight chapters divided into thirty sections. Chapter 1 has provided a brief summary of spatial social science theories and thinking as well as introductions to spatial effects and the primary data example used throughout this book. Chapter 2 addresses some important concepts and issues of spatial regression models and methods, including exploratory data analysis, neighborhood structure and spatial weight matrix, spatial dependence and heterogeneity, and exploratory spatial data analysis.
Chapter 3 introduces spatial regression models dealing with spatial dependence, including spatial lag models and spatial error models. Advanced spatial regression models dealing with spatial dependence, including spatial error models with spatially lagged responses, spatial cross-regressive models, and multilevel linear regression models, are introduced in Chapter 4. Chapter 5 introduces spatial regression models and methods dealing with spatial heterogeneity, including aspatial regression models, spatial regime models, and geographically weighted regression. Although both this chapter and Chapter 2 are recommended to be read in full, Chapters 3 to 5, which each introduce one method, are recommended to be read consecutively but do not have to be—readers can go to the method of interest directly.
Chapter 6 discusses extended spatial regime models and approaches for dealing with both spatial dependence and spatial heterogeneity in spatial regression analysis. Chapter 7 introduces some more advanced spatial regression models, including spatio-temporal regression models, spatial regression forecasting models, and geographically weighted regression for forecasting. A general procedure for studying social science phenomena with the spatial dimension in mind is suggested in Chapter 8 using the poverty data example and R code.
Study Questions
1 Which social science disciplines have a spatial aspect? How? To what extent?
2 What are spatial effect, spatial data analysis, geographic analysis, and spatial analysis?
3 What is the difference between spatial data analysis and geographic analysis?
4 What is the difference between spatial analysis and spatial data analysis?
5 What are the four types of spatial data analysis? What are they mainly used for?
6 How is the spatial dimension of population change addressed in related disciplines?
7 For your area of research, are spatial concepts and/or theories used? If so, what are they?
8 For your area of research, how do empirical studies typically address the spatial dimension?
Descriptions of Images and Figures
Back to Figure
The figure shows a detailed map of the state of Wisconsin in the United States.
The state is surrounded by Illinois in the south, the Mississippi River in the southwest with Iowa beyond that, Minnesota in the northwest, Lake Superior in the north, Michigan in the northeast, and Lake Michigan in the east. The state is divided into 72 counties.
The urban areas include the following counties:
Milwaukee County
Dane County
Waukesha County
Kenosha County
Brown County
Marathon County
Eau Claire County
Dane County
La Cross County
Rock County
Sheboygan County
Outagamie County
Winnebago County
In the map:
Urbanized areas are marked in dark gray
County boundaries are in bold outline
MCD boundaries are lighter
The scale on the map ranges from 0 to 80 miles, with the numbers 20 and 40 also marked out.
The compass rose indicates north is toward the top of the page.
The data is attributed to National Atlas of the United States and US Census Bureau on the map.
Back to Figure
The graph in the figure has a vertical axis denoting population numbers from 1 million to 6 million in increments of 1 million and a horizontal axis that represents period in years from 1840 to 2010.
An upward sloping line shows the growth in Wisconsin’s population since 1840. The curve is sharply upward sloping, reaching 2 million in 1900, 3 million in 1930, 4 million in 1960, and 5 million in 1990.
A note below the graph mentions that the source of the information are the Decennial censuses, U.S. Census Bureau.
Back to Figure
The figure is a map of Wisconsin that shows the classification of rural, suburban, and urban Minor Civil Division (MCDs) in Wisconsin. 18 counties are shown as suburban MCDs; around 15 clusters of urban MCDs exist, and the remaining are classified as rural.
The legend below the map states dark areas are urban, paler areas are suburban, and the palest are rural.
The compass rose indicates north is toward the top of the page.
The map scale ranges from 0 to 80 miles with 20 and 40 miles also shown and 0 to 120 kilometers with 30 and 60 kilometers also shown.
The source of the map is attributed to Chi (2012).
Back to Figure
The figure consists of 4 maps of Wisconsin that show population change at the MCD level in 1970–1980, 1980–1990, 1990–2000, and 2000–2010.
The first graph is for the period 1970–1980, when there was high population growth in several areas indicated by dark patches across the map.
The second graph is for the period 1980–1990, when most areas showed a large decline in population growth rates across the state.
The third graph is for the 1990–2000 period, when the population once again showed significant patches of growth in some areas.
The last graph depicts the 2000–2010 period, when there is growth in the population in certain areas only while others show signs of depopulation.
The legend below the graphs specify the population growth percentage;
Black shading means 128 percent to 15 percent
Dark