31 May
2007
31 May
'07
4:02 p.m.
The first CTPP product from ACS we hope will be a 3-year summary (ACS
records from 2005 thru 2007). Clearly, people want a flow tabulation.
The 3-year tabulations are limited to geographic units of 20,000 persons
or more. We currently don't have a DIFFERENT population threshold for
FLOW tabulation based on 3-years of ACS records, but the CTPP Technical
Group has been assuming we would want:
County -to- County
Place -to-Place
Perhaps: County - to-Place
Place -to- County
Using the same 20,000 population threshold. AND THEN, AFTER WORKPLACE
ALLOCATION IS COMPLETE, TO HAVE PUMA-TO-PUMA where PUMA of work is not
limited to COUNTY, but is a "real" sub-county PUMA where available.
The geography for tabulation for this first 3-yr ACS would be restricted
to Census 2000 geography, so we can't include "new" PUMAs defined for
Census 2010 until probably 2011 or 2012.
http://www.bea.gov/scb/pdf/2004/11November/1104Econ-Areas.pdf by
Kenneth P. Johnson and John R. Kort
There are 179 BEA Economic areas which are county-based. There are 344
component economic areas, which are merged into the 179 Economic Areas.
I am wondering if we might use either BEAs or CEAs to calculate either
quartile or quintiles to tabulate income for CTPP. I like the CEAs
because, for example in the Seattle area, the CEAs separates out the
western counties that are fairly rural in nature from the main urban
core counties.
Therefore the Part 3 tables might look like this:
TABLES for FLOW:
Table 1. Total workers
Table 2. Workers in group quarters (no characteristics)
Table 3. Means of Transportation to Work (7) (all workers)
Table 4 Income quartiles (4) or Income quintiles (5) (workers in
households) (Then, we would have to include a CHART documenting the
quartile or quintile values for each of the Economic areas, since they
would not be the same # from one area to the next).
Therefore, no crosstab of means of transportation by income in the FLOW
tabulation. Nandu has been researching using IPF routines to synthesize
a crosstab of means of transportation by income for FLOW for "base"
TAZs.
Because the BEA Economic Areas are large enough, there should be
sufficient ACS samples from which to calculate quartiles or quintiles
for household income. If there are approx 2 million completed h.u. ACS
forms each year, there would be about 11,000 households per BEA economic
area (BEA), or 5,800 households per Component economic area (CEA). This
avoids the problem of "too many" income categories that result from
trying to address variability between places like San Francisco-Oakland
and Nashville. Therefore a 3-year accumulated ACS would give us
approx 33,000 households for BEA, and 17,400 households for CEA. So,
the first step would need to be to calculate the values for each BEA or
CEA, and then tabulate the ACS records based on those values.
One of the biggest problems with this approach is that it would be up to
the data user to KNOW which BEA or CEA area was used. And, we would
have to check to see if there are MPOs that cross BEA or CEA boundaries.
Let me know what you think about this idea.
Elaine Murakami
FHWA Office of Planning
206-220-4460
31 May
31 May
5:46 p.m.
New subject: Brainstorming about INCOME variable for the future CTPP(using ACS records)
Elaine, thanks for raising this. You are actually hitting on a much
broader topic that we need to begin to pin down. What will a three-year
CTPP data product look like? What will the Disclosure Review Board (DRB)
let us have? Unfortunately neither of these are simple to answer but
before we can begin any discussions with the DRB, we need to have the
data product in focus. On that note and more relevant to your post I
have a coupla thoughts.
1. I would think that states and those working on more longer distance
models would find some type of income flow data usable, especially at
the CEA level. I am not sure how usable it would be within a region.
Hopefully others will weigh in.
2. Regarding the list of 3-year flow tables I would suggest adding in
and inserting "Workers in Households (no characteristics)" as table 2
followed by GQ, etc. I know one could get workers in households by
subtracting one table from another (total from GQ) but that just begins
to add a layer of complexity.
3. I know we talked about this with the DRB but I have to keep reminding
myself that they have already said that we could get the same two Part 3
tables that we got with CTPP 2000 without any thresholds applied. Do
you think that same exemption would apply to just the flows of either
household workers or GQ workers without any other characteristics? Even
with the big geography (20,000 pop zones) we might have some suppression
impacts given the smaller sample associated with a three year product
especially because we are talking about unweighted records.
4. I like your idea of the mixed geography, e.g. county to place and
place to county. Maybe this could lead to getting people to think about
mixed TAZs in the future--a set of residence zones and a different set
of workplace zones.
"Murakami, Elaine" wrote:
The first CTPP product from ACS we hope will be a 3-year summary (ACS
records from 2005 thru 2007). Clearly, people want a flow
tabulation. The 3-year tabulations are limited to geographic units
of 20,000 persons or more. We currently dont have a DIFFERENT
population threshold for FLOW tabulation based on 3-years of ACS
records, but the CTPP Technical Group has been assuming we would want:
County to- County
Place to-Place
Perhaps: County to-Place
Place to- County
Using the same 20,000 population threshold. AND THEN, AFTER
WORKPLACE ALLOCATION IS COMPLETE, TO HAVE PUMA-TO-PUMA where PUMA of
work is not limited to COUNTY, but is a real sub-county PUMA where
available.
The geography for tabulation for this first 3-yr ACS would be
restricted to Census 2000 geography, so we cant include new PUMAs
defined for Census 2010 until probably 2011 or 2012.
http://www.bea.gov/scb/pdf/2004/11November/1104Econ-Areas.pdf by
Kenneth P. Johnson and John R. Kort
There are 179 BEA Economic areas which are county-based. There are
344 component economic areas, which are merged into the 179 Economic
Areas. I am wondering if we might use either BEAs or CEAs to
calculate either quartile or quintiles to tabulate income for CTPP. I
like the CEAs because, for example in the Seattle area, the CEAs
separates out the western counties that are fairly rural in nature
from the main urban core counties.
Therefore the Part 3 tables might look like this:
TABLES for FLOW:
Table 1. Total workers
Table 2. Workers in group quarters (no characteristics)
Table 3. Means of Transportation to Work (7) (all workers)
Table 4 Income quartiles (4) or Income quintiles (5) (workers in
households) (Then, we would have to include a CHART documenting the
quartile or quintile values for each of the Economic areas, since they
would not be the same # from one area to the next).
Therefore, no crosstab of means of transportation by income in the
FLOW tabulation. Nandu has been researching using IPF routines to
synthesize a crosstab of means of transportation by income for FLOW
for base TAZs.
Because the BEA Economic Areas are large enough, there should be
sufficient ACS samples from which to calculate quartiles or quintiles
for household income. If there are approx 2 million completed h.u.
ACS forms each year, there would be about 11,000 households per BEA
economic area (BEA), or 5,800 households per Component economic area
(CEA). This avoids the problem of too many income categories that
result from trying to address variability between places like San
Francisco-Oakland and Nashville. Therefore a 3-year accumulated ACS
would give us approx 33,000 households for BEA, and 17,400 households
for CEA. So, the first step would need to be to calculate the values
for each BEA or CEA, and then tabulate the ACS records based on those
values.
One of the biggest problems with this
approach is that it would be up
to the data user to KNOW which BEA or CEA area was used. And, we
would have to check to see if there are MPOs that cross BEA or CEA
boundaries.
Let me know what you think about this idea.
Elaine Murakami
FHWA Office of Planning
206-220-4460
---------------------------------------------------------------
_______________________________________________
ctpp-news mailing list
ctpp-news(a)chrispy.net
http://www.chrispy.net/mailman/listinfo/ctpp-news
--
Ed Christopher
708-283-3534 (V)
708-574-8131 (cell)
FHWA RC-TST-PLN
19900 Governors Dr
Olympia Fields, IL 60461
1 Jun
1 Jun
1:19 p.m.
New subject: Brainstorming about INCOME variable for the futureCTPP(using ACS records)
Here is the link that includes an .xls document that shows WHICH county
is in WHICH BEA EA or CEA.
http://www.bea.gov/regional/xls/ea/eastructure.xls
This might help people think about whether this geographic unit might be
reasonable to incorporate, but I think the larger issue about INCOME is
whether people like the idea about regionally-based quartile and
quintiles, rather than fixed income ranges (In CTPP2000 we used for the
4 ranges: Below $15K, $15-35k; $35 - 75K, and $75+K).
Ed Christopher has an excellent point that your IDEAS on all aspects of
a 3-year ACS summary for a CTPP product are needed. We hope you will
share your thoughts with this listserv!
On the topic of data suppression, Nanda Srinivasan has been working at
the Census Bureau with the confidential files (he got "special sworn
status") and is examining IPF approaches to SYNTHESIZING small area
(e.g. TAZ-to-TAZ)flows for a table of MEANS OF TRANSPORTATION BY INCOME,
based on "real" data for larger zones (e.g. what we envision as "large
TAZs"). Also, an NCHRP 08-36 "quick response" project on disclosure
avoidance should be under contract very soon, with additional research
on the same topic, but using a different statistical approach. We hope
that these projects will result in methods that can be implemented on
new CTPP products and avoid the data suppression that hindered the
usability of Table 3-07 in the CTPP2000.
Elaine
206-220-4460
6 Jun
6 Jun
3:33 p.m.
New subject: Brainstorming about INCOME variable for the future CTPP(using ACS records)
To CTPP-News Maillist and Other Census Groupies,
Comments about Elaine Murakami’s email and Ed Christopher’s email
both dated 5/31/07 with subject: [Brainstorming] about income
variable….
Elaine lists four presumed-wanted flows: county to county, place to
place, county to place, and place to county. These are insufficient for
many subregional analyses when the principal city is part of one or more
counties. In these cases, we need city to county remainder data and
county remainder to city data.
In 1980 and 1990, the City of Chicago was a migpuma This allowed us
to do city to county remainder and county remainder to city analyses.
For instance, using the PUMS file we analyzed inmigration and
outmigration by characteristic between the City of Chicago and its
suburbs.
In 2000, the City of Chicago was not a migpuma even though many smaller
areas were migpumas. The Census Bureau person Ms. Showalter told me that
no city of any size was a migpuma unless it coincided with one or more
counties such as the City of New York. As a consequence, we were not
able to determine in and outmigration between the City of Chicago and
the suburbs for the 1995-2000 period.
In the future ACS-PUMS are large cities like the City of Chicago
identified as migpumas and powpumas? I noticed that the 2004 ACS PUMS
has only MIGSTATE (allowing to determine the state of origin of movers.)
Can we please restore large cities as migpumas and powpumas and
identify the pumas, migpumas, and powpumas on the ACS-PUMS.
On their website, the Census Bureau claims that the ACS-PUMS file is
used for academic research. But my experience as a city demographer
shows that the PUMS files are very important tools in city management.
Please help if you are in a position to do so. Thank you,
Marie Bousfield, Demographer
City of Chicago
Department of Planning and Development
City Hall, Room 703
121 North LaSalle Street
Chicago, Illinois 60602
Tel. (312) 744-6536
Fax. (312) 744-0759
Email mbousfield(a)cityofchicago.org
http://www.cityofchicago.org
5:19 p.m.
New subject: Brainstorming about INCOME variable for the futureCTPP(using ACS records)
Hello Marie --
Well, you learn something new everyday. I have never heard of a migpuma
before!
I agree with you that PUMA geography and the PUMS are not just for
academic research. We are happy that the CB now includes residence PUMA
as a tabulation geography as part of the 2005 ACS standard tabulations.
This benefits large cities, like Chicago. Maybe Table B07204 or C07204
"residence 1 year ago: State/County/place level" would be useful to
you?
Right now, in the ACS PUMS dataset, POW-PUMA is not a "real" PUMA. It is
generally limited to County, as the "basic" workplace coding completed
at the CB is to County-Place. In order to have a "real" POW-PUMA, the
Census Bureau will need to complete extended workplace allocation
(imputation) process for responses with ungeocoded workplace addresses.
This is because to assign workers to a "real" PUMA, the coding needs to
be complete to the census block level. Ungeocoded records continue to
be between 20-25 % of workers, similar to the Census 2000. I hope that
the extended workplace allocation will be established and part of the
NEXT CTPP, so that we can have PUMA-to-PUMA flows where the POW-PUMA is
not a county.
There has been some informal discussion among the CTPP Technical Working
Group about whether a "special" ACS PUMS file with the POW-PUMAs after
extended allocation has been completed would be possible, as part of
CTPP. We would have to take this to the CB's Disclosure Review Board
(as ALL tabulation and products go through the same review) for their
consideration.
This link is the ACS list of the POW-PUMAs. For example, 04890 includes
nearly 100 places and the "remainder" of Los Angeles County; 08100
includes 47 places and the remainder of San Diego County. I don't know
how many "real" PUMAs there are in Los Angeles County, but it is
probably nearly 100. I didn't look at Chicago, but I'm sure it has a
similar scale of loss of geographic detail.
http://www.census.gov/acs/www/Products/PUMS/C2SS/CodeList/2005/POWPUMA20
05lookup.htm
here is the list for the ACS MIG-PUMAs.
http://www.census.gov/acs/www/Products/PUMS/C2SS/CodeList/2005/MIGPUMA20
05lookup.htm
We understand that the Census Bureau staff are generally not allowed to
post messages to listservs, but I'm sure they will read your message and
consider your ideas for ACS PUMS files.
Elaine Murakami
FHWA Office of Planning
-----Original Message-----
From: ctpp-news-bounces(a)chrispy.net
[mailto:ctpp-news-bounces@chrispy.net] On Behalf Of
mbousfield(a)cityofchicago.org
Sent: Wednesday, June 06, 2007 1:34 PM
To: Ed Christopher
Cc: ctpp-news maillist
Subject: Re: [CTPP] Brainstorming about INCOME variable for the
futureCTPP(using ACS records)
To CTPP-News Maillist and Other Census Groupies,
Comments about Elaine Murakami's email and Ed Christopher's email
both dated 5/31/07 with subject: [Brainstorming] about income
variable....
Elaine lists four presumed-wanted flows: county to county, place to
place, county to place, and place to county. These are insufficient for
many subregional analyses when the principal city is part of one or more
counties. In these cases, we need city to county remainder data and
county remainder to city data.
In 1980 and 1990, the City of Chicago was a migpuma This allowed us
to do city to county remainder and county remainder to city analyses.
For instance, using the PUMS file we analyzed inmigration and
outmigration by characteristic between the City of Chicago and its
suburbs.
In 2000, the City of Chicago was not a migpuma even though many smaller
areas were migpumas. The Census Bureau person Ms. Showalter told me that
no city of any size was a migpuma unless it coincided with one or more
counties such as the City of New York. As a consequence, we were not
able to determine in and outmigration between the City of Chicago and
the suburbs for the 1995-2000 period.
In the future ACS-PUMS are large cities like the City of Chicago
identified as migpumas and powpumas? I noticed that the 2004 ACS PUMS
has only MIGSTATE (allowing to determine the state of origin of movers.)
Can we please restore large cities as migpumas and powpumas and
identify the pumas, migpumas, and powpumas on the ACS-PUMS.
On their website, the Census Bureau claims that the ACS-PUMS file is
used for academic research. But my experience as a city demographer
shows that the PUMS files are very important tools in city management.
Please help if you are in a position to do so. Thank you,
Marie Bousfield, Demographer
City of Chicago
Department of Planning and Development
City Hall, Room 703
121 North LaSalle Street
Chicago, Illinois 60602
Tel. (312) 744-6536
Fax. (312) 744-0759
Email mbousfield(a)cityofchicago.org
http://www.cityofchicago.org
_______________________________________________
ctpp-news mailing list
ctpp-news(a)chrispy.net
http://www.chrispy.net/mailman/listinfo/ctpp-news
8 Jun
8 Jun
10:52 a.m.
New subject: Brainstorming about INCOME variable for the futureCTPP(using ACS records)
To CTPP-News Maillist and Other Census Groupies,
Much thanks to all who shared their migpuma concerns.
The attached file defines the migpumas and the powpumas in function of
the pumas (all 5%). I downloaded it from Blodgett’s Mable database.
For instance, for Illinois (17), the pumas 3501-3519 cover the City of
Chicago and pumas 3401-3414 cover the remainder of Cook County which we
call suburban Cook County. The file shows that powpuma 3500 coincides
with the city of Chicago and that migpuma 3490 coincides with Cook
County.
We at the City defined the pumas but the definitions of the powpumas
and migpumas were done by the CB. In 1990, Chicago was a migpuma but
in 2000 the CB decided it was no longer. According to Clara, no county
remainders were migpumas except Los Angeles city, why not Chicago city?
I planned to extract the population sizes off all the migpumas and show
that there are many migpumas that are smaller than the City of Chicago
thereby showing a lack of equal access to data. According to
“Principles and Practices for Federal Statistical Agencies” by
Martin, Straf, and Citro, federal statistical agencies have to ensure
equal access to data. Are large subcounty cities with concentrations of
minority persons getting equal access to data?
Thank you,
Marie Bousfield, Demographer
City of Chicago
Department of Planning and Development
City Hall, Room 1003
121 North LaSalle Street
Chicago, Illinois 60602
Tel. (312) 744-6536
Fax. (312) 744-0759
Email mbousfield(a)cityofchicago.org
http://www.cityofchicago.org
>> "Murakami, Elaine"
<Elaine.Murakami(a)fhwa.dot.gov> 6/6/2007 5:19 PM
>>
Hello Marie --
Well, you learn something new everyday. I have never heard of a
migpuma
before!
I agree with you that PUMA geography and the PUMS are not just for
academic research. We are happy that the CB now includes residence
PUMA
as a tabulation geography as part of the 2005 ACS standard
tabulations.
This benefits large cities, like Chicago. Maybe Table B07204 or
C07204
"residence 1 year ago: State/County/place level" would be useful to
you?
Right now, in the ACS PUMS dataset, POW-PUMA is not a "real" PUMA. It
is
generally limited to County, as the "basic" workplace coding completed
at the CB is to County-Place. In order to have a "real" POW-PUMA, the
Census Bureau will need to complete extended workplace allocation
(imputation) process for responses with ungeocoded workplace
addresses.
This is because to assign workers to a "real" PUMA, the coding needs
to
be complete to the census block level. Ungeocoded records continue to
be between 20-25 % of workers, similar to the Census 2000. I hope
that
the extended workplace allocation will be established and part of the
NEXT CTPP, so that we can have PUMA-to-PUMA flows where the POW-PUMA
is
not a county.
There has been some informal discussion among the CTPP Technical
Working
Group about whether a "special" ACS PUMS file with the POW-PUMAs after
extended allocation has been completed would be possible, as part of
CTPP. We would have to take this to the CB's Disclosure Review Board
(as ALL tabulation and products go through the same review) for their
consideration.
This link is the ACS list of the POW-PUMAs. For example, 04890
includes
nearly 100 places and the "remainder" of Los Angeles County; 08100
includes 47 places and the remainder of San Diego County. I don't
know
how many "real" PUMAs there are in Los Angeles County, but it is
probably nearly 100. I didn't look at Chicago, but I'm sure it has a
similar scale of loss of geographic detail.
http://www.census.gov/acs/www/Products/PUMS/C2SS/CodeList/2005/POWPUMA20
05lookup.htm
here is the list for the ACS MIG-PUMAs.
http://www.census.gov/acs/www/Products/PUMS/C2SS/CodeList/2005/MIGPUMA20
05lookup.htm
We understand that the Census Bureau staff are generally not allowed
to
post messages to listservs, but I'm sure they will read your message
and
consider your ideas for ACS PUMS files.
Elaine Murakami
FHWA Office of Planning
-----Original Message-----
From: ctpp-news-bounces(a)chrispy.net
[mailto:ctpp-news-bounces@chrispy.net] On Behalf Of
mbousfield(a)cityofchicago.org
Sent: Wednesday, June 06, 2007 1:34 PM
To: Ed Christopher
Cc: ctpp-news maillist
Subject: Re: [CTPP] Brainstorming about INCOME variable for the
futureCTPP(using ACS records)
To CTPP-News Maillist and Other Census Groupies,
Comments about Elaine Murakami's email and Ed Christopher's email
both dated 5/31/07 with subject: [Brainstorming] about income
variable....
Elaine lists four presumed-wanted flows: county to county, place to
place, county to place, and place to county. These are insufficient
for
many subregional analyses when the principal city is part of one or
more
counties. In these cases, we need city to county remainder data and
county remainder to city data.
In 1980 and 1990, the City of Chicago was a migpuma This allowed us
to do city to county remainder and county remainder to city analyses.
For instance, using the PUMS file we analyzed inmigration and
outmigration by characteristic between the City of Chicago and its
suburbs.
In 2000, the City of Chicago was not a migpuma even though many
smaller
areas were migpumas. The Census Bureau person Ms. Showalter told me
that
no city of any size was a migpuma unless it coincided with one or more
counties such as the City of New York. As a consequence, we were not
able to determine in and outmigration between the City of Chicago and
the suburbs for the 1995-2000 period.
In the future ACS-PUMS are large cities like the City of Chicago
identified as migpumas and powpumas? I noticed that the 2004 ACS PUMS
has only MIGSTATE (allowing to determine the state of origin of
movers.)
Can we please restore large cities as migpumas and powpumas and
identify the pumas, migpumas, and powpumas on the ACS-PUMS.
On their website, the Census Bureau claims that the ACS-PUMS file is
used for academic research. But my experience as a city demographer
shows that the PUMS files are very important tools in city management.
Please help if you are in a position to do so. Thank you,
Marie Bousfield, Demographer
City of Chicago
Department of Planning and Development
City Hall, Room 703
121 North LaSalle Street
Chicago, Illinois 60602
Tel. (312) 744-6536
Fax. (312) 744-0759
Email mbousfield(a)cityofchicago.org
http://www.cityofchicago.org
_______________________________________________
ctpp-news mailing list
ctpp-news(a)chrispy.net
http://www.chrispy.net/mailman/listinfo/ctpp-news
noon
New subject: Brainstorming about INCOME variable for thefutureCTPP(using ACS records)
Marie--I think you might be confusing some folks in the transportation
world when your talk about MIGPUMA and POWPUMA. We really do not look
at the world quite that way. You are absolutely right in what you say
about what the CB produces as part of their standard products which is
one of the reasons the transportation community has gone off and
purchased its own set of special tabulations.
One of the issues that frustrated the transportation with the 2000 data
was that you could not get data that was a standard product that had
PUMA of Work coded on the pums record. You could only get county of
work. That meant your flow would be PUMA to county of work. The
transportation community has asked all along for to see the PUMS data
coded to PUMA of work but the CB for whatever reason doesn't code all
the records down to that area of geography.
For our CTPP tabulation we had developed a process to extend the
geocoding down to the smallest areas (block faces or what ever they are
called today) for all the records that could not get coded. We called it
extended allocation and Elaine has early research work going on to help
get it incorporated into the standard process schema at the CB. The
extended allocation process was what was used for all the CTPP data
which is why we have the flows that we do at so many geographies and
with such high quality.
While the CB might not have the product you need unless you are LA, the
CTPP will produce what you are after for 2000, 1990 and even 1980 in the
Chicago region. CTPP did have place as summary geography. CTPP does
not have PUMA a summary geography for 2000 but in the future it will.
I am not sure if I helped clarify or only confused the matters. One of
the things that makes all of this even more confusing is that we all use
slightly different terms. I never thought of the data in terms of MIG
and POW PUMAs.
Marie--good to hear from you. What always drove me crazy we those few
residents who lived in Chicago but were actually were in Dupage County.
It was up around O'Hare and in 1990 kept causing my tables not to
balance, until I finally figured out. For 2000 in CTPP you do not have
to worry about all the tables balancing due the rounding and threshold
requirements on the data which is a whole other story....
mbousfield(a)cityofchicago.org wrote:
To CTPP-News Maillist and Other Census Groupies,
Much thanks to all who shared their migpuma concerns.
The attached file defines the migpumas and the powpumas in function of
the pumas (all 5%). I downloaded it from Blodgettâs Mable database.
For instance, for Illinois (17), the pumas 3501-3519 cover the City of
Chicago and pumas 3401-3414 cover the remainder of Cook County which we
call suburban Cook County. The file shows that powpuma 3500 coincides
with the city of Chicago and that migpuma 3490 coincides with Cook
County.
We at the City defined the pumas but the definitions of the powpumas
and migpumas were done by the CB. In 1990, Chicago was a migpuma but
in 2000 the CB decided it was no longer. According to Clara, no county
remainders were migpumas except Los Angeles city, why not Chicago city?
I planned to extract the population sizes off all the migpumas and show
that there are many migpumas that are smaller than the City of Chicago
thereby showing a lack of equal access to data. According to
âPrinciples and Practices for Federal Statistical Agenciesâ by
Martin, Straf, and Citro, federal statistical agencies have to ensure
equal access to data. Are large subcounty cities with concentrations of
minority persons getting equal access to data?
Thank you,
Marie Bousfield, Demographer
City of Chicago
Department of Planning and Development
City Hall, Room 1003
121 North LaSalle Street
Chicago, Illinois 60602
Tel. (312) 744-6536
Fax. (312) 744-0759
Email mbousfield(a)cityofchicago.org
http://www.cityofchicago.org
--
Ed Christopher
708-283-3534 (V)
708-574-8131 (cell)
FHWA RC-TST-PLN
19900 Governors Dr
Olympia Fields, IL 60461
>> "Murakami, Elaine"
<Elaine.Murakami(a)fhwa.dot.gov> 6/6/2007 5:19 PM
>>
Hello Marie --
Well, you learn something new everyday. I have never heard of a
migpuma
before!
I agree with you that PUMA geography and the PUMS are not just for
academic research. We are happy that the CB now includes residence
PUMA
as a tabulation geography as part of the 2005 ACS standard
tabulations.
This benefits large cities, like Chicago. Maybe Table B07204 or
C07204
"residence 1 year ago: State/County/place level" would be useful to
you?
Right now, in the ACS PUMS dataset, POW-PUMA is not a "real" PUMA. It
is
generally limited to County, as the "basic" workplace coding completed
at the CB is to County-Place. In order to have a "real" POW-PUMA, the
Census Bureau will need to complete extended workplace allocation
(imputation) process for responses with ungeocoded workplace
addresses.
This is because to assign workers to a "real" PUMA, the coding needs
to
be complete to the census block level. Ungeocoded records continue to
be between 20-25 % of workers, similar to the Census 2000. I hope
that
the extended workplace allocation will be established and part of the
NEXT CTPP, so that we can have PUMA-to-PUMA flows where the POW-PUMA
is
not a county.
There has been some informal discussion among the CTPP Technical
Working
Group about whether a "special" ACS PUMS file with the POW-PUMAs after
extended allocation has been completed would be possible, as part of
CTPP. We would have to take this to the CB's Disclosure Review Board
(as ALL tabulation and products go through the same review) for their
consideration.
This link is the ACS list of the POW-PUMAs. For example, 04890
includes
nearly 100 places and the "remainder" of Los Angeles County; 08100
includes 47 places and the remainder of San Diego County. I don't
know
how many "real" PUMAs there are in Los Angeles County, but it is
probably nearly 100. I didn't look at Chicago, but I'm sure it has a
similar scale of loss of geographic detail.
http://www.census.gov/acs/www/Products/PUMS/C2SS/CodeList/2005/POWPUMA20
05lookup.htm
here is the list for the ACS MIG-PUMAs.
http://www.census.gov/acs/www/Products/PUMS/C2SS/CodeList/2005/MIGPUMA20
05lookup.htm
We understand that the Census Bureau staff are generally not allowed
to
post messages to listservs, but I'm sure they will read your message
and
consider your ideas for ACS PUMS files.
Elaine Murakami
FHWA Office of Planning
-----Original Message-----
From: ctpp-news-bounces(a)chrispy.net
[mailto:ctpp-news-bounces@chrispy.net] On Behalf Of
mbousfield(a)cityofchicago.org
Sent: Wednesday, June 06, 2007 1:34 PM
To: Ed Christopher
Cc: ctpp-news maillist
Subject: Re: [CTPP] Brainstorming about INCOME variable for the
futureCTPP(using ACS records)
To CTPP-News Maillist and Other Census Groupies,
Comments about Elaine Murakami's email and Ed Christopher's email
both dated 5/31/07 with subject: [Brainstorming] about income
variable....
Elaine lists four presumed-wanted flows: county to county, place to
place, county to place, and place to county. These are insufficient
for
many subregional analyses when the principal city is part of one or
more
counties. In these cases, we need city to county remainder data and
county remainder to city data.
In 1980 and 1990, the City of Chicago was a migpuma This allowed us
to do city to county remainder and county remainder to city analyses.
For instance, using the PUMS file we analyzed inmigration and
outmigration by characteristic between the City of Chicago and its
suburbs.
In 2000, the City of Chicago was not a migpuma even though many
smaller
areas were migpumas. The Census Bureau person Ms. Showalter told me
that
no city of any size was a migpuma unless it coincided with one or more
counties such as the City of New York. As a consequence, we were not
able to determine in and outmigration between the City of Chicago and
the suburbs for the 1995-2000 period.
In the future ACS-PUMS are large cities like the City of Chicago
identified as migpumas and powpumas? I noticed that the 2004 ACS PUMS
has only MIGSTATE (allowing to determine the state of origin of
movers.)
Can we please restore large cities as migpumas and powpumas and
identify the pumas, migpumas, and powpumas on the ACS-PUMS.
On their website, the Census Bureau claims that the ACS-PUMS file is
used for academic research. But my experience as a city demographer
shows that the PUMS files are very important tools in city management.
Please help if you are in a position to do so. Thank you,
Marie Bousfield, Demographer
City of Chicago
Department of Planning and Development
City Hall, Room 703
121 North LaSalle Street
Chicago, Illinois 60602
Tel. (312) 744-6536
Fax. (312) 744-0759
Email mbousfield(a)cityofchicago.org
http://www.cityofchicago.org
_______________________________________________
ctpp-news mailing list
ctpp-news(a)chrispy.net
http://www.chrispy.net/mailman/listinfo/ctpp-news
------------------------------------------------------------------------
Name: puma5_xtract.xls
puma5_xtract.xls Type: Microsoft Excel Worksheet (application/vnd.ms-excel)
Encoding: base64
13 Jun
13 Jun
3:14 p.m.
New subject: Brainstorming about INCOME variable for thefutureCTPP(using ACS records)
At the recent ACS Peer Exchange, sponsored by the FHWA/FTA
Transportation Planning Capacity Building program, held May 10-11, 2007;
Nancy Torrieri shared a paper copy of some of the question changes for
the 2008 ACS content. I just noticed that Question 14 "Did this person
live in this house or apartment 1 year ago?" has a NEW line for STREET
ADDRESS, in addition to city/town; county; State; and zipcode.
If respondents complete this line, then, coding to a "real" PUMA, not
just to county should be possible.
Elaine Murakami
FHWA Office of Planning
-----Original Message-----
From: ctpp-news-bounces(a)chrispy.net
[mailto:ctpp-news-bounces@chrispy.net] On Behalf Of
mbousfield(a)cityofchicago.org
Sent: Friday, June 08, 2007 8:52 AM
To: Ed Christopher; Murakami, Elaine
Cc: ctpp-news maillist; douglas.w.hillmer(a)census.gov
Subject: RE: [CTPP] Brainstorming about INCOME variable for
thefutureCTPP(using ACS records)
To CTPP-News Maillist and Other Census Groupies,
Much thanks to all who shared their migpuma concerns.
The attached file defines the migpumas and the powpumas in function of
the pumas (all 5%). I downloaded it from Blodgett's Mable database.
For instance, for Illinois (17), the pumas 3501-3519 cover the City of
Chicago and pumas 3401-3414 cover the remainder of Cook County which we
call suburban Cook County. The file shows that powpuma 3500 coincides
with the city of Chicago and that migpuma 3490 coincides with Cook
County.
We at the City defined the pumas but the definitions of the powpumas
and migpumas were done by the CB. In 1990, Chicago was a migpuma but
in 2000 the CB decided it was no longer. According to Clara, no county
remainders were migpumas except Los Angeles city, why not Chicago city?
I planned to extract the population sizes off all the migpumas and show
that there are many migpumas that are smaller than the City of Chicago
thereby showing a lack of equal access to data. According to
"Principles and Practices for Federal Statistical Agencies" by
Martin, Straf, and Citro, federal statistical agencies have to ensure
equal access to data. Are large subcounty cities with concentrations of
minority persons getting equal access to data?
Thank you,
Marie Bousfield, Demographer
City of Chicago
Department of Planning and Development
City Hall, Room 1003
121 North LaSalle Street
Chicago, Illinois 60602
Tel. (312) 744-6536
Fax. (312) 744-0759
Email mbousfield(a)cityofchicago.org
http://www.cityofchicago.org
6166
days inactive
6179
days old
7 comments
3 participants
participants (3)
-
Ed Christopher -
mbousfield＠cityofchicago.org -
Murakami, Elaine