Demographic change, like environmental change, occurs at different levels and is understandable only in the context of sufficient data. To address their research questions, demographers and other social scientists are increasingly turning to the Internet for electronic volumes of Census data, which until recently were available only on tape and before that in paper publications stacked away in libraries.
Economists Steven Levitt of the Harvard Society of Fellows and Julie Berry Cullen of the Massachusetts Institute of Technology tapped the Internet in their unique nationwide analysis of the relationship between crime and urban flight. "What others before us have done is compare, say, just the percent change in a city's population and what crime is or was a while ago," said Levitt. "What we wanted to do was get beyond just the aggregate numbers. We wanted to look at the relationship by household, and ask, for example, 'Is it the case that rich city residents are more likely to leave cities as crime rises?'"
To address these questions, Levitt and Berry Cullen obtained Census data extracts called Public Use Microdata Samples (PUMS) from NASA's Socioeconomic Data and Applications Center (SEDAC). PUMS data files, which the Census Bureau strips of identifying information before making available to researchers, contain detailed information on the characteristics of each housing unit and the people in it. The PUMS data are the best available about individual people, according to Levitt.
Using FBI's Uniform Crime Reports and a portion of PUMS data files representing five percent of all American households, Levitt and Berry Cullen established that city residents do move in response to crime rates. "For each serious crime—either a violent crime or a serious property crime—it looks like city populations fall by about one person," Levitt said. Drawing on more than 10 million demographic observations, Levitt and Berry Cullen also found that households fleeing crime most likely consist of families with children, have high income levels, and resettle in the suburbs.
At the state level, Tom Steahr is using Census data in a pilot study of the distribution of Connecticut's poverty populations. Steahr, a University of Connecticut demographer, is interested in the relationship between poverty severity and local labor markets.
"What I'm looking for is a relationship between the needs of local labor markets and the characteristics and abilities of the poverty populations," Steahr said. "Are they related to each other and what makes it difficult for some people to remove themselves from poverty roles? Is it because their skills and abilities don't match the needs of the local labor market, which may be different from the state or the regional economy?"
To answer these questions, Steahr is attempting to take the available data to a more meaningful level of detail. To avoid treating the entire state as a homogenous unit, he developed an algorithm to determine how Connecticut's towns relate to each other in terms of labor needs and transactions. Using the algorithm, Steahr assembled the state's 169 towns into 23 "local labor market areas," small groups of towns that tend to share workers.
"People shop within these local labor market areas and they visit within them," Steahr said, "and basically life goes on within these relatively small geographic groups of towns." Working with "blockgroups," which are the smallest units used to tabulate 1990 Census data (and which consist of contiguous city blocks that typically contain a few thousand people), Steahr is able to access the information he needs for people living at various levels of poverty in Connecticut.
Cullen, J. B. and S. D. Levitt. 1996. Crime, Urban Flight, and the Consequences for Cities. National Bureau of Economic Research Working Paper No. 5737
Steahr, T. 1996. Crime, Urban Flight, and the Consequences for Cities, National Bureau of Economic Research Working Paper No. 5737.
For more information
NASA Socioeconomic Data and Applications Center (SEDAC)
Public User Microdata Samples (PUMS)
|About the remote sensing data used|
|Data||Public User Microdata Samples (PUMS)|
|DAAC||NASA Socioeconomic Data and Applications Center (SEDAC)|