18 Mar
2003
18 Mar
'03
1:31 p.m.
1. Census 2000, summary file #3, Table P51 provides data on workers by class of worker,
including the self-employed and unpaid family. This shows that 10.1 percent of U.S.
workers are self-employed or unpaid family; 11.6 percent in California; and 10.6 percent
in the Bay Area. Interestingly our self-employed share ranges upwards to 23.2 percent of
the employed labor force in Marin County (home of George Lucas and other self-employed
clones). So, counties with high self-employed shares are probably either agrarian
economies or entrepeneurial economies.
2. The more I research, the less I find out I really know. There are several sources of
data on employment, including the BLS, BEA and Census. One very good source of information
is the book "Socioeconomic Data for Understanding Your Regional Economy" by
Cortright and Reamer. It is available on their EDA-supported web site at: www.econdata.net
Check it out.
3. Imputation (or allocation, using the Census Bureau's terminology) will definitely
be an error issue for place-of-work or other partial responses, though if the census
respondent said "no" to all of the employment-related questions, then it's
fairly certain that the Bureau couldn't change those answers (whether the answers are
correct or not.) I don't know what happens if a person reported their place-of-work
and means of transportation but didn't report other employment-related questions
(industry, occupation) (probably those values are imputed/allocated). Definitely we will
be seeing these "allocation flags" in PUMS as well as other census data
products.... And we shouldn't forget sampling error (decennial census long form is
still only a 1-in-6 survey); as well as non-standard errors (persons not telling the
truth! egads!)
Chuck
>> Patty Becker <pbecker(a)umich.edu>
03/18/03 11:03AM >>>
Why would the census be missing the folks who--I
agree--are missed in
ES202? The self-employed, public workers, etc. should still be covered
just fine in the census--as well as anyone else. Re non-response: you'd e
surprised what they can impute!
Patty Becker
At 09:18 AM 03/18/2003 -0500, you wrote:
The gap between the employment out there and what gets
reported in the
Census is a lot higher when you factor in jobs not covered by the ES202
system, both wage/salary and "self-employed" proprietors. (The 6% upward
adjustment ABAG reports to you seems low, BEA figures would indicate about
a 20% adjustment needed.) Non-response could be the biggest factor - I
imagine the Census Bureau would have a hard time "imputing" number of
workers if a household ignores all the work-related questions on the long
form.