Brainstorming about INCOME variable for the future CTPP (using ACS records)

List overview All Threads
Download

newer

older

RE: [CTPP] Brainstorming about...

RE: [CTPP] Census Data in High...

Murakami, Elaine

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

Attachments:

attachment.htm (text/html — 10.0 KB)

Show replies by date

Ed Christopher

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

Murakami, Elaine

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

mbousfield＠cityofchicago.org

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

Murakami, Elaine

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

mbousfield＠cityofchicago.org

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 >>

Ed Christopher

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

>> "Murakami, Elaine" <Elaine.Murakami(a)fhwa.dot.gov> 6/6/2007 5:19 PM >>

-- Ed Christopher 708-283-3534 (V) 708-574-8131 (cell) FHWA RC-TST-PLN 19900 Governors Dr Olympia Fields, IL 60461

Murakami, Elaine

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

6357

days inactive

6370

days old

ctpp-news@mailman.chrispy.net

Manage subscription

7 comments

3 participants

tags (0)

participants (3)

Ed Christopher
mbousfield＠cityofchicago.org
Murakami, Elaine