Attention: State DOT and MPO staff evaluating CTPP 2000, Part 1:
I received a number of e-mails/phone calls since my last e-mail asking for staff to evaluate different levels of geography.
Some of the common issues are:
1. CTPP Part 1 data will not agree with SF 1 or Redistricting (Full count data). The CTPP data are based on the census long form, and are subject to weighting, just like other sample surveys. The SF1 and Redistricting data are full counts, and are more accurate. The best source to do spot-checks on the CTPP data is SF3.
2. The count of households WILL NOT EQUAL Occupied Housing Unit counts (in general).
This is because the all the household data in the census are derived by tabulating the "person weight" of the householder. Weights for occupied housing units are assigned based on housing unit counts.
3. Please be aware that all CTPP tables (except the mean, median, and standard deviation tables) are subject to the following rounding rules:
0 is kept as 0.
1-7 are rounded to 4
Anything above 7 is rounded to the nearest multiple of 5 (eg: 9 is rounded to 10, 17 is rounded to 15 etc)
Columns for Totals in each table may not match the sum of the categories because Totals are rounded independently of the cells. For example:
0 vehicle households = 6 is rounded to 4
1 vehicle households =14 is rounded to 15
2 vehicle households=8 is rounded to 10
3 vehicle households=8 is rounded to 10
4 vehicle households=3 is rounded to 4.
Total households=39 is rounded to 40 (and NOT 4+15+10+10+4 = 43 rounded to 45).
The difference between the totals and the sum of the parts is substantial for tables with more than 30 cells.
4. Missing geography: Sometimes, a few TAZs or BGs may not show up in your geography/map viewer or tables. This is because geography are shown only if at least ONE table in CTPP has some values. This means that those TAZs or BGs that don't show up have NO population or households in them (based on long form data).
5. Wrong Column headers: We noticed that the CTPP Access Tool wrongly labels column headers for one dimensional tables - this error will be fixed for the final version.
6. MSA/CMSA names are wrong: We think that the names for MSA/CMSA or PMSAs are incorrect, and should be fixed in the final version.
If you have issues DIFFERENT from the above 6, please call/e-mail me.
Thank you!
Nanda Srinivasan
202-366-5021
The Census Bureau just released its latest American Community
Survey-Alert (an ACS newsletter) Number 17. In this Alert there are
articles on
* ACS 2002 Data Are Now Available
* Redesigned ACS Web Site Launched
* ACS Content Fact Sheets
The Alerts and "lots of" information on the ACS can be found on the ASC
website at http://www.census.gov/acs/www/
--
Ed Christopher
Planning Specialist
Resource Center
Federal Highway Administration
19900 Governors Drive
Olympia Fields, Illinois 60461
708-283-3534 (V)
708-574-8131 (cell)
Hello from Utah.
You may be interested in a recent report:
"Commuting Patterns in Utah: County Trends for 1980, 1990, and 2000"
It is based on the CTPP 2000 county level data as well as the
county-to-county journey-to-work file.
It is available here:
http://www.business.utah.edu/bebr/uebr/uebrMayJun2003.pdf
Do you know of similar studies focusing on other states?
Best regards,
Pam
Pam Perlich
Senior Research Economist
Bureau of Economic and Business Research
University of Utah
1645 E. Campus Drive, Room 401
Salt Lake City, Utah 84112-9302
Voice: (801) 581-3358
Fax: (801) 581-3354
Email: Pam.Perlich(a)business.utah.edu
Greetings folks out there in CTPP-land -- I was wondering whether anyone
had yet tabulated or otherwise produced tables 2-003, 2-004, and 2-005 for
New York State's counties. (That's sex by occupation, industry, and class
of worker by place of work.) I don't have access to the advance release of
the CTPP myself but am trying to compile a report on the Upstate NY economy
based on place of work. If it's not kosher to release these tables to me,
don't -- but if you can I would appreciate it. Thanks. Please post directly
to me, not back to the list.
Rolf Pendall
-------------------------------------------------
Rolf Pendall, AICP, Ph.D.
Associate Professor, City & Regional Planning
Cornell University
212 W. Sibley Hall, Ithaca, NY 14853
607-255-5561 / 530-678-8103 (fax) / rjp17(a)cornell.eduhttp://www.people.cornell.edu/pages/rjp17/
Subject: Census News Brief #10
Date: Wed, 27 Aug 2003 19:13:58 EDT
National Research Council Panel Urges More Informed Planning for 2010
Census;
Census Funding Clears House Of Reps Intact;
House Subcommittee Hears Strong Support;
New Cost Estimate, for ACS; and more.
In its third report, issued last month, the National Research Councils
Panel on Research on Future Census Methods urged the Census Bureau to
complete evaluations of Census 2000 quickly in order to assess more
fully the costs and benefits of design options for 2010. The science
panel also expressed concern about the direction of plans to update and
improve the Master Address File (MAF), which represents the universe for
household enumeration.
The panel, chaired by Dr. Benjamin King, was established to review early
planning for the 2010 count. Its second interim report (the panel
previously published a letter report, as well) focuses primarily on
modernization of the Census Bureaus geographic databases and on
development of a technical infrastructure for the 2010 census.
The Census Bureau often describes its 2010 census planning strategy as a
three-legged stool comprised of (1) MAF/TIGER Enhancements Program;
(2) replacement of the census long form with an on-going American
Community Survey; and (3) early integrated planning for a short
form-only census. The Academy panel cited early integrated planning as
perhaps the one [component] most crucial to a successful 2010 census
and, in some respects, the hardest to accomplish. It noted that the
technical infrastructure for the census, needed to develop a logical
design process for 2010, was functional but developed at high cost and
high risk, without adequate time for development and testing for Census
2000. The panel said it was strongly impressed by early efforts to
reengineer the design process and encouraged the bureau to develop
comprehensive business process models for 2010 quickly, and to appoint a
system architect for the census. (Technical infrastructure refers
primarily to software programs and computer and telecommunications
systems that support census operations.)
The panel also evaluated plans to modernize the Census Bureaus
geographic database, called TIGER (Topologically Integrated Geographic
Encoding Reference System), and comprehensive housing list (MAF), for
the 2010 census. The report is generally supportive of the Bureaus
efforts to realign features in TIGER. The plausibility of the
realignment timetable would be bolstered considerably, the panel
suggested, by a detailed work plan that includes a schedule for updating
county information, and by specifying a process for detecting geographic
changes once the initial updating is completed.
The report is far more critical of proposals to improve the Master
Address File. The modernization program, the panel concluded, falls
seriously short with respect to the MAF, and is lacking a comprehensive
plan to add new addresses, identify duplicates, and generally ensure a
complete and accurate list. The Census Bureaus current strategy shows
signs of repeating costly errors from the 2000 experience, the panel
wrote, noting that despite using the Postal Services Delivery Sequence
File and instituting a pre-census local review of address lists, the
Bureau conducted a costly, 100-percent block canvass of addresses before
Census 2000. The panel recommended that the Census Bureau clarify how
it will improve the accuracy of the MAF, including how it will apply its
experiences from the 2000 census; appoint a coordinator to oversee MAF
development and maintenance; specify plans for local partnerships, with
an emphasis on improved communication and incentives for local sharing
of information; and strengthen the role of evaluations in developing
plans to update the MAF for 2010.
The report also contains initial feedback on the American Community
Survey (ACS). The panel encouraged Congress to fully fund the ACS over
the long term, while also noting statistical issues that required
further research, including the weighting and estimation scheme and the
relative accuracy of ACS and long form data.
Overall, the panel concluded that the strategy for the 2010 census
needs to be more fully informed by cost-benefit analyses of different
approaches. To date, the panel wrote, the plan for the 2010 census
has been presented to the panel with little supporting analysis.
The National Research Council (NRC) is the principal operating agency of
The National Academy of Sciences and The National Academy of
Engineering. The census panel was constituted by the NRCs Committee on
National Statistics. The report, Planning the 2010 Census: Second
Interim Report, is available on the National Academies Press website at
http://www.nap.edu. The panel will complete its work and issue a final
report at the end of the year.
Census funding update: By a vote of 400 21, the U.S. House of
Representatives passed a $41.2 billion spending measure that includes
funding for the Census Bureau, before adjourning for its August break.
The Fiscal Year 2004 (FY04) Commerce, Justice, and State, The Judiciary,
and Related Agencies Appropriations bill (H.R. 2799) allocates the full
$662 million for Census Bureau programs requested by President Bush.
The White House said in a Statement of Administration Policy that it
supported passage of the bill, and appreciates the full funding of
Census Bureau programs, but that it had several concerns about specific
provisions of H.R. 2799. The administration urged legislators to
restore $10 million cut from the Presidents request for the Commerce
Departments Economic and Statistics Administration, which houses the
Bureau of Economic Analysis (BEA). Only one provision of the mammoth
bill, which would block implementation of a controversial Federal
Communications Commission rule on media ownership, drew a veto threat.
The Senate delayed consideration of a Commerce spending bill until after
the August congressional recess. Congress resumes legislative work on
September 3. For details on Census Bureau and BEA funding in H.R. 2799,
see the July 10 and July 21 Census News Briefs.
House panel hears testimony in support of the ACS: The House
Subcommittee on Technology, Information Policy, Intergovernmental
Relations, and the Census held an oversight hearing in May to review
plans for the proposed American Community Survey. It was the panels
first hearing on Census Bureau programs since its creation as part of
the Government Reform Committee early in the 108th Congress. Rep.
William Lacy Clay (D-MO), the subcommittees top Democrat, is the only
panel member with previous service on a census oversight subcommittee.
In opening remarks, Chairman Adam Putnam (R-FL) noted the Census
Bureaus historical efforts to meet the continually changing needs of
our Nation for timely, quality data, and said Congress must now
consider another significant evolution: replacing the traditional
census long form with the ACS. To win support for ACS funding in
Congress, the chairman said, the agency must show that the ACS will
eliminate duplicative surveys at the Census Bureau, thereby saving
money. I simply would find it unbelievable that no surveys could be
eliminated when the ACS is fully implemented, the congressman said.
Chairman Putnam also pointed to privacy as another significant issue the
Census Bureau must address. The challenge, he said, is to obtain the
information that is needed to make informed decisions while at the same
time respecting the privacy rights of the public. While noting that
the Census Bureau benefits from the most protective privacy law on the
books, the congressman challenged the agency to explore new and
innovative ways to solicit voluntary cooperation from the public. The
bureau is evaluating a test conducted earlier this year at the request
of Congress, comparing the effects of mandatory versus voluntary
response on the cost and quality of ACS data.
Rep. Clay said he remains concerned about the quality of data the ACS
will produce and about adequate funding, as well as the long-term cost
of the ACS. He noted that the delayed start of the ACS would allow
stakeholders to evaluate additional years of data from test sites and
the national sample. The Census Bureau plans to release data tables
from the 2002 Supplementary Survey and test sites this fall.
The subcommittee heard testimony from the Under Secretary of Commerce
for Economic Affairs, the Census Bureau director, and a panel of five
witnesses representing a range of census data users. Under Secretary
Kathleen Cooper said the ACS will revolutionize the way we take the
census and meet the nations data needs. She described the survey as a
well-developed program that is at full implementation in 31 test
sites across the country and producing consistently high quality data
from these sites and a national Supplementary Survey of 700,000
households. Dr. Cooper reminded panel members that the content of the
ACS questionnaire is essentially the same as the census long form it is
designed to replace, because the Census Bureau must still meet federal
programmatic requirements for data. Director C. Louis Kincannon noted
that about $200 billion a year in federal program funds is distributed
on the basis of census data.
Referring to the revised plan to launch the ACS nationwide in the final
quarter of Fiscal Year 2004, Mr. Kincannon said tract level data from
the survey would be published in 2010, two years before it would be
available if the 2010 census included a traditional long form. The
first characteristics data, for places of 65,000 or more persons, would
be available starting in 2006, based on the first full year of data
collected in 2005. Mr. Kincannon acknowledged that tract level
estimates from the ACS would be subject to higher sampling error than
similar census long form data, because the five-year aggregates used to
produce ACS tract data involve a smaller overall sample (15 million
households over five years) than the long form (roughly 20 million
households). Sampling error, the director said, would be offset by
more complete responses [to the survey] that the permanent staff of
field representatives will collect during telephone and in-person
follow-up interviews to unresponsive households.
A re-engineered 2010 census that includes only a short form, with the
ACS replacing the long form, and a modernized address and mapping
system, will cost $11.2 billion, Mr. Kincannon told the subcommittee. A
traditional census plan would cost at least $12 billion, he said. In
response to a question from Chairman Putnam, the director said the
annual cost of a fully implemented ACS could be $165 million. Previous
estimates submitted to the census oversight subcommittee in April 2001
pegged the cost of the ACS at $131 million for the first year, which
under the original implementation plan would have included only nine
months of field follow-up visits, the most costly part of the survey.
Five witnesses presented the views of data users in state and local
government, business, and demographic research for a variety of
applications. All expressed support for the ACS and offered examples of
how annual estimates of social and economic characteristics would
improve decision-making for the data users they represented. Thomas
Reardon, Executive Director of the Fulton County Partnership,
McConnellsburg, Pennsylvania, discussed how annual data from the ACS
test site in his county helped the organization accurately estimate and
distribute flu vaccines for their small rural community. [T]he more
information we have about the people we serve, the more efficient we can
be with the resources we have to help them, Mr. Reardon testified.
Joseph Salvo, Director, Population Division, New York City Department of
City Planning, called the ACS a continuous barometer of the social and
economic condition of the nations communities. Noting that the
September 11, 2001 terrorist attacks diminished the usefulness of Census
2000 long form data for broad sections of the city, Dr. Salvo emphasized
the importance of current information for the cost-effective delivery of
government services. He also said the ACS would collect more accurate
information on population and neighborhood characteristics than the
census long form, citing successful follow-up interviews with
unresponsive households in the ACS Bronx test site, where a significant
number of households failed to complete long forms in Census 2000.
Joan Naymark, Director of Research and Planning for Target Corporation,
presented the views of business and industry data users as a
representative of the U.S. Chamber of Commerce. Ms. Naymark said the
ACS is vital to economic development and wise government and business
decision-making. She described Targets use of census data to locate
new stores and remodel existing ones, to offer merchandise and
advertising to match neighborhood characteristics, and to plan for work
force needs. Ms. Naymark said the ACS would provide more accurate data
than the census long form because of its timeliness, and she urged
Congress to fund full implementation of the ACS over the long term.
Ken Hodges, Director of Demography, Claritas, told subcommittee members
that, Support and even enthusiasm for the ACS are growing in the
private sector, for whom frequent data at the neighborhood level is a
top priority. Mr. Hodges acknowledged some concerns about the quality
of ACS estimates (including coverage of the population in group
quarters), but said many of the concerns also applied to the census,
and that the ACS offered greater opportunity to improve the reliability
of the data. Mr. Richard Ogburn, Principal Planner, South Florida
Regional Planning Council, testified that the ACS would support the
more effective allocation of scarce public resources through improved
assessment of need and better targeting of resources. He noted that
South Floridas population grew by 21 percent in each of the last two
decades. Without more timely data from the ACS, Mr. Ogburn said, local
governments and businesses must divert programmatic resources to collect
information on local characteristics. The Council analyzes area trends
and conditions to support land use and natural resource planning,
emergency preparedness, transportation, affordable housing, and other
important services, Mr. Ogburn told the panel.
House Democrats question no-adjustment decision for 2010: In a letter
to the current and former top Democrats on the House census oversight
subcommittee, Census Bureau Director C. Louis Kincannon reaffirmed his
agencys decision not to plan for a statistical adjustment of the 2010
census for reapportionment and redistricting purposes. Instead, Mr.
Kincannon wrote, the bureau would evaluate coverage improvement programs
used in Census 2000 that had a positive influence on the differential
undercount, to see if they can further reduce the gap in accuracy
between the count of Whites and of racial and ethnic minorities in
2010. Census 2000 programs that yielded a higher percentage of
enumerations of historically undercounted groups included the Be
Counted Program, Service-Based Enumeration (e.g. soup kitchens, homeless
shelters), Coverage Improvement Follow-up, and the Update/Enumerate and
Urban Update/Leave operations, the director said.
Rep. William Clay and Rep. Carolyn Maloney (D-NY) wrote to Mr. Kincannon
in May, saying they were surprised by his announcement at a National
Academy of Sciences census panel meeting that the Census Bureau would
not statistically correct the 2010 census numbers used for congressional
apportionment and redistricting. The decision, the legislators wrote,
is the only aspect of the 2010 design that has been finalized. Reps.
Clay and Maloney noted that the undercount of African Americans was
three percent higher than the undercount of non-Hispanic Whites in 2000
only a modest improvement, they said, from the 3.9 percent
differential undercount in the 1990 census. The legislators asked
about the Census Bureaus plans to measure the accuracy of the 2010
census and to correct for overcounting.
In his June 30th response, Director Kincannon said his comments at the
National Academy
confirms a widely held and often articulated conclusion based on an
exhaustive analysis of the post-census survey in Census 2000. The
Census Bureau learned that it could not both carry out an accuracy
measurement survey and analyze the results in time to meet the legal
deadline for delivering redistricting data to states, Mr. Kincannon
said. The bureau also has concluded that science is insufficiently
advanced to allow making statistical adjustment of a successful
decennial census in which the percentage of error is presumed to be so
small that adjustment would introduce as much or more error than it
fixes.
In addition to coverage improvement programs to improve accuracy, the
Census Bureau is looking at ways to clarify who should be included on
the census form, the director noted. Better presentation of residence
rules and an undercount detection question to identify households
where some residents may have tenuous attachments, are under
consideration. The bureau also is researching why the census
double-counts people and housing units. Improving and measuring
overall coverage and differential coverage in the census are primary
goals of the 2010 census, Mr. Kincannon wrote.
Upcoming news: The Census Bureau is expected to release the results of
a test of voluntary versus mandatory response to the American Community
Survey in the coming days. The test, conducted at the request of
Congress, took place over the course of several months earlier this
year. Congress asked the Bureau to assess the consequences of making
response to the survey voluntary. Response to the census (short and
long forms) is required by law. We will report on the results of the
evaluation when they become available.
Upcoming meetings: The next Decennial Census Advisory Committee meeting
will be October 9 10, 2003. The five Race and Ethnic Advisory
Committees will meet jointly October 1 - 3, 2003. The Census Bureau has
not yet published agendas for the meetings.
-----------------------
Census News Briefs are prepared by Terri Ann Lowenthal, an independent
consultant in Washington, DC. Please direct questions about the
information in this News Brief to Ms. Lowenthal at 202/484-3067 or by
e-mail at terriann2k(a)aol.com. Thank you to the Communications Consortium
Media Center for posting the News Briefs on the Census 2000 Initiative
web site, at www.census2000.org. Please feel free to circulate this
information to colleagues and other interested individuals.
--
Ed Christopher
Planning Specialist
Resource Center
Federal Highway Administration
19900 Governors Drive
Olympia Fields, Illinois 60461
708-283-3534 (V)
708-574-8131 (cell)
As of 8/27/03, 5-percent PUMS data for 20 states has been released.
Scheduled for release next Wednesday, 9/3/03, are: Arizona, California, Delaware,
Georgia, North & South Carolina, Tennessee and Warshington.
Census Bureau has not published a "look ahead" for data to be released either 9/17 or
9/24.
Also, the correspondence tables between residence area PUMAs (Public Use Microdata Areas)
and place-of-work PUMAS (POWPUMA5), is included in Appendix N, available on the web site
cited below.
*********
More information from the State Data Center listserv, on states scheduled for release on 9/10:
FYI...
The Public Use Microdata Sample files for Colorado, Connecticut,
Massachusetts, Minnesota, Montana, Utah, West Virginia, and Wisconsin will
be released to the public on September 10, 2003 at 12:01 AM. There is no
embargo period for these data.
The data can be accessed on 9/10/03 at:
http://www2.census.gov/census_2000/datasets/PUMS/FivePercent/ftp://ftp2.census.gov/census_2000/datasets/PUMS/FivePercent/
The PUMS files will be released on a weekly basis through September 2003.
********
C. Purvis, MTC
Attention: Everyone evaluating CTPP Part 1 data:
I have noticed that most of the staff evaluating the data are only looking at their LOWEST GEOGRAPHIC LEVEL (E.g.: block groups, and TAZs). We are noticing some errors for a few areas at other levels of geography (e.g.: MPO region, MSA/CMSA etc.)
While looking at the lowest geographic level is an excellent idea, please make sure you evaluate the data for other coarser levels (E.g.: Tract, MPO Region, and County). Please report any data issues/comments to me.
E-mail: Nanda.srinivasan(a)fhwa.dot.gov, Phone: 202-366-5021.
Thank you
Nanda Srinivasan
Attention: All MPOs and State DOTs
On August 21, 2003, Census Bureau (CB) completed mailing out Part 1 data for all the states. The CB is getting ready to finalize, and start mass-production of the Part 1 CDs. After the data are finalized, the CTPP Part 1 will be made available to the general public through the Bureau of Transportation Statistics webpage. We encourage you to examine the data for which you are familiar to determine if there are any systematic errors.
For most states, the deadline for agencies to send comments on their Part 1 dataset is September 15, 2003.
States where agencies can send their comments back by October 1, 2003 include:
Alabama
Arkansas
Connecticut
Hawaii
Kansas
Louisiana
Massachusetts
Minnesota
Mississippi
Missouri
New Hampshire
Oklahoma
For agencies in South Carolina, and Kentucky, the deadline is October 6, 2003.
To send comments, or ask questions, please contact Nanda Srinivasan (Phone: 202-366-5021, email: Nanda.srinivasan(a)fhwa.dot.gov).
This week, the CTPP Part 1 CDs for AL, AR, CT, HI, KS, LA, MN, MS, MO, NH, and OK were mailed by the Census Bureau. As a reminder, it has been taking between a week and 2 weeks for most people to receive the CDs in the mail, even if they are being sent first-class. There are 2 remaining states: KY and SC, and they should be mailed next week.
We are going to conduct a small test of different mailing operations to see if we can identify where exactly the delay is occuring, so that when it is time to mail the CTPP Parts 2 and 3 CDs, we will not have the same problem!
Elaine Murakami
206-220-4460 in Seattle
This message may be relevant to folks who use the CTPP data on households by the number of workers in the household (e.g., Table 1-65 households by workers in household by vehicles available).
I've had a conversation with Phil Salopek (Census Bureau) about the definition of "households by workers in household" and I've confirmed this information by examining the 1% PUMS for my region.
The CTPP Access Tool program defines "workers in household" as "see the definition for workers" and "see the definition for households" which is frankly not that useful. This could be improved.
The CTPP documentation could be improved by providing a precise definition of "workers in household." It may not be what you think it is!!! The CTPP documentation is based on standard census documentation such as that from SF3. For example, the term "workers in families" is defined as the number of civilian employed with a job LAST YEAR (Question #30a on the census long form.)
It is clear that the CTPP data on "workers in household" is based on "workers at work during census reference week" and is NOT BASED ON "workers at work, plus workers with a job but not at work"
So, the values for "zero worker households" in the CTPP will include "households with no workers" PLUS it will include "households with worker(s), but the worker(s) are NOT at work LAST WEEK"
The CTPP shows 539,755 "zero worker households" in the Bay Area, at 21.9 percent of total households. The 1-percent PUMS data shows 514,617 "zero worker households" in the Bay Area, at 20.9 percent of total households. 1-percent PUMS data, using the CTPP definition of workers in households, shows 538,741 zero-worker households. This means that we have 25,000 households that have workers, but all of the workers in the household were not at work "last week" (they were all having fun watching spring training baseball....)
(For those using PUMS, the CTPP definition of workers in households is based on where ESR=1 or ESR=4, where ESR = employment status recode variable. The alternative definition of workers in households is based on where ESR=1, 2, 4, or 5, which includes the weekly absentees (ESR=2 and 5).
I am assuming that the Census Bureau will not change the CTPP 2000 data since it is consistent with the 1990 CTPP data.
My plan, for validation of our workers-in-household & vehicles-in-household model (WHHAO), is to depend on the 5-percent PUMS data to provide PUMA-level control totals, and to adjust the CTPP zone-level data (Part 1, Table 65) to match with the PUMS data.
Hope this helps,
Chuck Purvis, MTC
**************************************************************
Charles L. Purvis, AICP
Principal Transportation Planner/Analyst
Metropolitan Transportation Commission
101 Eighth Street
Oakland, CA 94607-4700
(510) 464-7731 (office)
(510) 464-7848 (fax)
www: http://www.mtc.ca.gov/
Census WWW: http://www.bayareacensus.ca.gov/
**************************************************************
