Two methodologies have been used in this work. Both begin with the SF3 report on family income and income to non-family households and use related data to estimate which blocks these families live in.
The first is an adaptation of the methodology used with the 1990 census. Two important data series were critical to that analysis: (1) housing values and rent levels at block level (from STF1) and (2) income by family size (from STP-19). Neither of those two data series are available at block level for the 2000 census. Our first methodology has been to take the data available from the 2000 census (SF3 tables P9, P76 and P79 available at block group level) and to use the block level relationships in the 1990 data to distribute families and non-family households to blocks in 2000. This methodology is appropriate for blocks that existed in 1990 and that have not changed substantially.
The second methodology has been to use regression analysis of county-wide block group data to develop an algorithm for estimating block level income. Median family income and median non-family household income are the dependent variables and 358 variables or combinations of variables were tested as the independent variables. The best predictor was found to be "Race of householder by tenure" (SF1 table H14 and SF3 table H11). SF3 block group data was used to develop the algorithm and then SF1 block level data was used to estimate income at block level. Families and non-family households were then distributed along the income curve using 1990 relationships for blocks which existed in 1990 and were substantially unchanged and using block group level data for new or changed blocks. This second methodology is especially important for blocks that have been newly defined since 1990 or which have changed substantially in terms of the numbers or nature of people living there.
All of the data gathered for this analysis is presented in these pages. Source data is organized by census and table and presented in these pages county-wide along with examples of block group and block level data. The most important of this data and our estimates are presented in detail for all blocks.