The two Census 2000 datasets, SF3 and CTPP Part 1, are derived from the
same set of Census 2000 long form raw data, so they really should show
the same patterns (households by household income) at the region,
county, tract, block group and TAZ level. The SF3 (Summary File #3)
definitely does NOT have TAZ-level tabulations, but it sounds like you
have a process to use census block groups to aggregate/disaggregate into
your TAZ structure.
The SF3 table P-52 provides data by 17 income categories, ranging from
"less than 5,000" to "$200,000 or more".
The CTPP Part 1, Table 64, provides data by 25 income categories (+1
category for "total households) also ranging from < $5,000, to "$150,000
or more". The CTPP has better details in income categories less than
$50,000, in increments of $2,500. This compares to SF3 which increments
income categories by $5,000 ranges. So, if you really wanted to have
data on "households less than $27,500" then you should use the CTPP.
The other major benefit of the CTPP is that these detailed household
income categories are cross-classfiied by various other variables,
including person in household (Table -64), the number of workers in
household (Table 1-66), and by the number of vehicles in the household
(Table 1-67). These are valuable if you are examining very small area
geography information for possible cross-classification travel models,
say, shop/other trips (or tours) by household size by income level,
The major "cost" of using the CTPP is the independent rounding of cell
values inflicted on the part 1 tables. So, the sum of the households by
the 25 income categories may not precisely match the "total households"
also provided in the same CTPP table. This is an annoyance, but can be
overcome by normalizing (adjusting) the CTPP cell values to either match
the "CTPP total households" value, or the "SF3 total households"
are not subject to the rounding nonsense.
My recommendation would be to use the CTPP, and then normalize the CTPP
tables to the SF1 precise count of households for your TAZ. This means
you need a precise correspondence file of blocks-to-TAZ, so that you can
"roll-up" the SF1 counts of 100% short form data (total pop, household
pop, households, pop by race/ethnicity/age/sex) to your TAZes.
Note that neither the SF1 or SF3 files were subjected to the rounding
rules. This means that the numbers should always add up precisely to the
Hope this helps,
Chuck Purvis, MTC - SF Bay Area
>> "Kendra Watkins"
<kwatkins(a)mrcog-nm.gov> 06/01/06 3:02 PM >>>
I am compiling income
data for our region at the TAZ level and have 2
sources to choose from, the CTPP TAZ income data by place of residence
or the SF3 block group income data modified to fit our DASZ structure.
wonder if I could get some thoughts on which source might best
a 2000 snapshot by TAZ.
Senior Data Analyst
Mid-Region Council of Governments
809 Copper Ave. NW
Albuquerque, NM 87102