Homeowners and Renters |
Description of Data |
Data
sources The Census data used in the report come from two sources - Summary Tape File 2a for 1990 and Summary Files 1 and 2 for 2000. The form in which housing data are tabulated by the census requires that we use households as the unit of analysis. Households are assigned to racial/ethnic categories based on the race/ethnicity of the household head. We use split tracts in order to distinguish accurately between city and suburban locations; split tracts located entirely in the city or suburbs are aggregated to whole tracts. Group definition Data suppression For Asians, SF1 refers to "Asians alone," resulting in an understatement of the number of Asian householders in the tract compared to the SF2 definition. Due to suppression, SF2 counts only 88.9% of Asians in metropolitan areas in the United States. Replacement from SF1 tables allows us to bring this up to 97.8%. Use of SF1 to replace suppressed data for non-Hispanic blacks is more complex and also results in an understatement of the number of black householders. SF1 does not provide a table for non-Hispanic blacks; the number must be calculated from other tables. We begin with the number of blacks alone (blacks in combination with another race are lost). From this count we wish to remove Hispanic blacks. We approximate the number of Hispanic blacks by subtracting Hispanic whites (all "white alone" less "non-Hispanic white alone") from all Hispanics. More precisely, the number is calculated by this formula: NHB = B-(H-(W-NHW)), where NHB is non-Hispanic blacks, B is all blacks (alone), H is all Hispanics, W is all whites (alone), and NHW is non-Hispanic whites (alone). Where the calculation yields a negative number, it is set to zero. This number is used only in those tracts where the actual count is suppressed in SF2. Due to suppression,
SF2 counts only 96.6% of non-Hispanic blacks in metropolitan areas in
the United States. Replacement from SF1 tables allows us to bring this
up to 98.1%. |