Race, Wealth, and Solid Waste Facilities in North Carolina

Jennifer M. Norton; Steve Wing; Hester J. Lipscomb; Jay S. Kaufman; Stephen W. Marshall; Altha J. Cravey


Environ Health Perspect. 2007;115(9):1344-1350. 

In This Article

Materials and Methods

We used inhabited North Carolina census block groups as the unit of analysis to define communities in order to obtain racial and economic characteristics (n = 5,261).Census block groups are designed to contain between 600 and 3,000 people, with an optimum size of 1,500 people (U.S. Census Bureau 2001b). The choice of block group boundaries is intended, where possible, to reflect neighborhoods with distinct sociodemographic characteristics.

Solid Waste Facilities

The North Carolina Division of Waste Management (NCDWM) is responsible for issuing permits to solid waste facilities in the state. The NCDWM's electronic files of permitted solid waste facilities lacked sufficient information, including dates of operation and specific facility locations, to address the study aims. Thus, we obtained additional information through a systematic review of solid waste facility paper records maintained by the NCDWM. For the purpose of this study, solid waste facilities were defined as MSWLs, construction and demolition debris landfills (CDLFs), industrial solid waste landfills (INDUSLFs), tire landfills (TIRELFs), and waste transfer stations (TRANSFERs). Solid waste facilities that were issued a permit to operate (or equivalent) by 31 December 2003 were included in the study.

We used geographic coordinates to locate solid waste facilities within block groups. The North Carolina Center for Geographic Information Analysis (NCCGIA) provided geographic coordinates for some solid waste facilities. Complete addresses were not available for most solid waste facilities; therefore, address matching could not be used to locate facilities. We used maps obtained from NCDWM records, tax parcel maps, and the internet program TerraFly (Florida International University 2003) to determine latitudes and longitudes and to verify coordinates received from the NCDWM and the NCCGIA. The TerraFly interface was used to virtually fly over aerial images to locate solid waste facilities. When possible, latitudes and longitudes at the approximate centers of waste disposal areas or TRANSFER buildings were recorded. The coordinates of two solid waste facilities that could not be visually located were assigned to the centroid coordinates of the census block group that contained the road listed as the facility address. Census 2000 geographic boundary files were obtained from ESRI (2001). Geographic coordinates of solid waste facilities were spatially joined to census block groups using ArcGIS version 9.1 (ESRI, Redlands, CA).

Block Group Characteristics

We obtained block group population and wealth data from the U.S. Census Bureau decennial Census 2000 Summary File 3 (U.S. Census Bureau 2002a) and the 1990 Census Summary File 3 (GeoLytics Inc. 2003). Census 2000 data were used for analyses of existing facilities in 2003. For analyses of newly permitted facilities after 1990, we used a standard geographic area to compute changes over time so that changes in the population are not a reflection of changing geographic boundaries. The longitudinal analyses utilized data from GeoLytics, which estimates 1990 U.S. Census data in 2000 Census block group boundaries (GeoLytics Inc. 2003). Intercensal estimates were created by linear interpolation between 1990 and 2000. Linear extrapolation beyond the year 2000 yielded implausible values from some areas; therefore, we used year 2000 data to estimate the population for 2001-2003 for each block group.

The primary racial and ethnic groups in North Carolina are white non-Hispanic (70%), African-American non-Hispanic (21%), Hispanic (5%), Asian non-Hispanic (1%), and American Indian non-Hispanic (1%) (U.S. Census Bureau 2002a). Racial and ethnic groups other than white non-Hispanic share experiences of discrimination, low political power, and low socioeconomic status (Smelser et al. 2001). Hispanic identity is an ethnic, not a racial, classification. North Carolina's Hispanic population is largely made up of recent immigrants from Mexico and Central America (U.S. Census Bureau 2001a). Because of these shared characteristics and small population size of racial and ethnic groups other than white non-Hispanic and African American, race was categorized as either white non-Hispanic or other, and the percentage of persons of race and ethnicity other than white non-Hispanic was calculated for each block group.

Socioeconomic status may be measured by education, annual income, or wealth. Wealth varies less over time than annual income and is a primary dimension of social class and political power in the United States (Krieger et al. 1997). We used median house value for owner-occupied housing units as a measure of community wealth. For the cross-sectional analyses of facilities present in 2003, all owner-occupied housing units were used. For the longitudinal analyses of facilities permitted between 1990 and 2003, data were available only from the 1990 Census for specified owner-occupied housing units. Specified housing units do not include mobile homes, homes with a business on the property, or homes on > 10 acres of land (U.S. Census Bureau 2002b). Median house values in 1990 were adjusted to year 2000 dollars to account for inflation (U.S. Census Bureau 2004).

Several other factors were considered as alternative explanations of solid waste facilities. Landfills require land for waste disposal, which is more plentiful in rural areas. Many rural areas of North Carolina have high poverty levels, and eastern North Carolina is part of the Black Belt, the former slave plantation region of the South that is still home to many rural African Americans. Population density, expressed as persons per square mile, was used to measure rurality. We also considered region of the state as a determinant of locations of solid waste facilities. North Carolina has four regions that differ physically, racially, and economically: Mountain, Piedmont, Coastal Plain, and Tidewater (Gade et al. 1986). Because access to truck routes connecting points of waste generation to points of waste disposal is considered in locating solid waste sites, we also considered block group distance to the nearest major road and distance to the nearest city [defined by the U.S. Census Bureau (2001b) as "urbanized area/urban cluster"] measured from block group centroids as a determinant of the location of solid waste facilities.

Statistical Analysis

We used inhabited census block groups as the unit of analysis. For cross-sectional analyses, the outcome was whether or not the block group had ≥1 permitted solid waste facility on 31 December 2003. Crude and adjusted prevalence odds ratios (PORs) and 95% confidence intervals (CIs) were calculated using logistic regression with generalized estimating equations to account for the nesting of block groups within counties. Models for all facility types and for specific facility types were fit and indicator variables were created to compare block groups with higher percentages of people of color to those with lower percentages of people of color (< 10%), and to compare block groups with lower house values block groups to those with the highest house values (≥$100,000). The combined effects of race and house value were assessed by creating indicator variables for their cross-classification. Population density was modeled as a continuous variable. Region was considered using indicator variables with the Piedmont as the referent category. We considered distance ≥3 miles from a major population center in North Carolina, using an indicator variable with block groups <3 miles from a city as the referent category. Indicator variables were also used to evaluate distances from major roads using the following categories: block groups ≥1 mile from a U.S. highway and ≥10 miles from an interstate; block groups <1 mile from a U.S. highway or <10 miles from an interstate; and block groups <1 mile from a U.S. highway and < 10 miles from an interstate (referent).

Longitudinal analyses followed block groups through time between 1990 and 2003. Because the presence of an existing solid waste facility strongly affects permitting of another facility at the same location, facility-free block groups were analyzed separately from block groups with an existing facility. We used a block group-time (analogous to person-time) approach so that block groups contributed follow-up time in each category of race and wealth as they changed over time; block groups remained at risk of having a first facility permitted (for facility-free block groups) or for having an additional facility permitted (for block groups with existing facilities). Facility-free block groups entered follow-up in the cohort with existing facilities on the date of their first permitted facility, at which time they were at risk of receiving a second permitted facility. Crude and adjusted hazard ratios (HRs) were calculated using extended Cox proportional hazards regression to compare block groups with higher percentages of people of color and lower house values to block groups with lower percentages of people of color and higher house values. Because of the nesting of block groups within counties, SEs were computed using robust variance estimation.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.