Abstract: Disparities in Spatial Access to Social Services in the U.S. : A GIS Analysis from 1990 to 2014 (Society for Social Work and Research 23rd Annual Conference - Ending Gender Based, Family and Community Violence)

457P Disparities in Spatial Access to Social Services in the U.S. : A GIS Analysis from 1990 to 2014

Schedule:
Saturday, January 19, 2019
Continental Parlors 1-3, Ballroom Level (Hilton San Francisco)
* noted as presenting author
Yoosun Park, PhD, Associate Professor, Smith College, Northampton, MA
James Quinn, MA, Geographer, Columbia University, New York, NY
Philip Hurvitz, PhD, Reserach Assistant Professor, University of Washington, Seattle, WA
Jana Hirsch, PhD, Assistant Research Professor, Drexel University, PA
Kathryn Neckerman, PhD, Senior Research Scientist, Columbia Population Research Center, New York, NY
Gina Lovasi, PhD, Associate Professor, Drexel University, PA
Andrew Rundle, DrPH, Associate Professor, Columbia University, New York, NY
Background and Purpose:  Little is known about the spatial distribution, and change in distribution, of social services across the U.S, and the extent to which changes in spatial access to social services mirrors changes in population social circumstances through time.  Here we introduce the National Establishment Time Series (NETS) dataset as a tool for studying the dynamics of spatial access to social services in the U.S.  The NETS provides annual establishment-level data beginning in 1990, for U.S. business, nonprofit, community, institutional and government establishments. 

Methods.  Standard Industrial Classification 8322 (Individual and Family Social Services) was used to identify social service establishments in the NETS in each year, across the U.S. from 1990 to 2014.  Establishment addresses were geocoded and linked to 2010 Census tract areas. Tract areas within Core-Based Statistical Areas, including both metropolitan and micropolitan areas, were selected for analyses (N=67,217 tracts; 55,776 social service establishments in 1990 and 184,798 in 2014).  The Longitudinal Tract Database was used to estimate population totals, % of population living in poverty (poverty rate), and inflation adjusted median household income in 1990 and 2010 for 2010 Census tract areas. Tract areas were categorized into tertiles of population density and quintiles of poverty rate and median household income. Negative binomial regression with tract area as an offset was used to test whether year and tract level characteristics predict social services/km2 in each tract area.

Results.  Density of social services in 1990 was associated with 1990 population density: mean services per km2 were 0.02, 0.29 and 1.42 respectively for tracts in the first, second and third tertiles of population density.  Focusing further analyses on tracts in the highest tertile of population density, where most social services were located, the density of social services was highest in tracts in the top quintile of poverty rate (2.22/km2 in 1990 and 5.25/km2 in 2014). For tracts in the highest quintile of poverty rate in 1990, tracts in the top quintile of 1990-2010 change in inflation adjusted median household income (gained $9,430 or more) had the highest density of social services in 1990 (4.10/km2) and saw the largest 1990 to 2014 gain in social services (+5.02/km2).  High poverty tracts in the bottom quintile of 1990-2010 change in inflation adjusted median household income (loss of $10,570 or more) had the lowest density of social services in 1990 (1.50/km2, p<0.01 compared to top quintile) and saw the smallest 1990 to 2014 increase in social services (2.49/km2, p=<0.01 compared to top quintile).         

Conclusions and Implications:  In urban centers and surrounding suburbs in the U.S., changes in spatial access to social services from 1990 to 2014 did not align with concurrent changes in population social circumstances. High-poverty tracts in which household income dropped the most from 1990 to 2010 had the lowest access to services in 1990 and saw the smallest increases in services from 1990 to 2014. The NETS, combined with Census data, provides a tool for understanding historical trends in spatial access to services and for planning to meet future needs.