<<RE: Re[2]: Employer File Update for Decennial Census >>
Several of you have responded to my question on what is to be done with
journey-to-work coding, including two responses from the Census. I've
attached a response I sent directly to Phil Salopek of the Census Bureau
because I think it touches on something common to many of your agencies -
starting "fresh" with a new database to review when many of you have put in
a lot of effort during the 90's to put together a geocodable listing on your
own.
Many of your comments made references to presentations and discussions at
the annual TRB meeting held earlier this month. I did not go to TRB this
year and no doubt many other MPO staffs were not there or made it to these
discussions. I don't know if others have already asked the kind of
questions I'm asking or not, but let me know if you see any problems with
the issues I'm raising - or if you think they're not serious problems at
all.
Please place Earnest R. Lloyd (e-mail: elloyd(a)nctcog.dst.tx.us) and Rocky
Gardiner (e-mail: rgardiner(a)nctcog.dst.tx.us) on the CTPP mail list.
Thank you,
Earnest R. Lloyd
Research Associate
Research and Information Services Department
North Central Texas Council of Governments
P.O. BOX 5888
Arlington, TX 76005-5888
Phone: (817) 695 - 9161
FAX: (817) 640 - 4428
E-mail: elloyd(a)nctcog.dst.tx.us
I chatted with Bob and Phil after the A1D08 committee meeting and asked the ? 2 them. The answer was reasonable (so I think they should be able to get an answer to us soon), but since I can't remember exactly what they said, I don't want to paraphrase what I thought I heard. Hopefully they can deliver the product we need, but if you have an employer that has multiple locations (here in the Seattle area, Boeing has always been our big problem with Census products) it is best to work directly with that employer.
I have a question for the CTPP group. About this time 10 years ago the
Census requested of MPO's an up-to-date list of employer addresses for use
in the journey-to-work coding. If you're like us, you've made lots of
improvements locally in the geocoding of these addresses. But I have not
heard a word from the Census on supplying them with this list for the 2000
geocoding, and I've sent two e-mails to the Census inquiring about this that
to date have gone unanswered. Does anyone know what, if anything, is going
on in this area? Is it being handled through a different avenue?
-----Original Message-----
From: ed c [SMTP:berwyned(a)mcs.com]
Sent: Monday, January 25, 1999 4:24 PM
To: ctpp maillist
Subject: [CTPP] BREAKING NEWS ALERT
January 25, 1999
From: Census2000 <Census2000(a)ccmc.org>
5 - 4 Court Decision Bars Sampling For Purposes of
Congressional Apportionment;
Political Debate Shifts to Use of Sampling for Other
Purposes
Justices Express Widely Differing Views on Meaning of Census
Act
In a close decision issued yesterday, Justice O'Connor led a
slim majority of five Justices in reaching the conclusion
that section 195 of the Census Act (title 13, United States
Code) "prohibits the proposed uses of statistical sampling
in calculating the population for purposes of
apportionment." The Justices expressed a wide range of
views on whether the Act allows the use of sampling to
derive the state population totals used for congressional
apportionment. Justice Scalia authored a separate opinion
concurring in O'Connor's conclusion, in which Justice Thomas
joined in full and Justice Kennedy and Chief Justice
Rehnquist joined in substantial part.
Justice Stevens, author of the dissenting opinion, was
joined by Justices Souter, Ginsburg, and Breyer in reaching
the conclusion that "[t]he Census Act unambiguously
authorizes the Secretary of Commerce to use sampling
procedures when taking the decennial census." Justice Breyer
set forth somewhat different reasons for reaching the same
conclusion in a concurring opinion.
The Justices also split several ways on whether the parties
challenging the census plan had legal standing to do so.
The five Justices in the majority concluded that the
plaintiffs in Glavin v. Clinton had standing because at
least one litigant, a resident of Indiana, successfully
demonstrated (based on analysis done by University of
Wisconsin government professor, Dr. Ronald Weber) that his
state would lose a congressional seat if statistical methods
were used in the 2000 census. The Justices also noted that
residents of some counties would be injured by the use of
sampling because their votes would be diluted relative to
counties with higher undercount rates. Because it reached a
decision on the merits of the Glavin case, the Court
dismissed the similar case brought by the U.S. House of
Representatives at the direction of former Speaker Newt
Gingrich.
In his dissenting opinion, Justice Stevens said he would
have granted standing to the House plaintiffs, as well, and
then overturned the lower court decision in that case.
Justice Breyer, however, agreed with the majority's
conclusion on standing. Justices Ginsburg and Souter took
exception to the majority's consideration of intrastate
redistricting in deciding the standing question.
The Court's opinion: Justice O'Connor placed great weight on
what she described as a 200-year congressional prohibition
on the use of sampling to calculate the apportionment
population. Through various amendments to the Census Act,
O'Connor said, Congress authorized an increased level of
sampling for collecting information other than the census
counts used for congressional apportionment. When Congress
amended section 195 in 1976, O'Connor wrote, it "changed a
provision that permitted the use of sampling for purposes
other than apportionment into one that required that
sampling be used for such purposes if 'feasible'. They also
added to the existing delegation of authority to the
Secretary to carry out the decennial census a statement
indicating that despite the move to mandatory use of
sampling in collecting non-apportionment information, the
Secretary retained substantial authority to determine the
manner in which the decennial census is conducted."
(emphasis in italics in original)
Justice O'Connor also noted that when Congress debated the
1976 amendments to section 195, it did not discuss the
possible effect of the change on the apportionment
calculation. This silence, O'Connor wrote, demonstrates
that "Members of Congress voting on the bill read the text
of the statute, as do we, to prohibit the use of sampling in
determining the population for apportionment purposes."
In his concurring opinion, Justice Scalia ridiculed the
dissenters' efforts to reconcile the apparent prohibition on
sampling for apportionment purposes in section 195 with the
Secretary's authority to use sampling in the census in
section 141(a), which directs the Secretary to conduct the
census. He was also swayed, Scalia wrote, by his strong
doubt that the Constitution permits sampling in the census.
The dissent: The dissenting opinion differed sharply with
the reasoning of the majority, calling the Court's
conclusion "an unusual tour de force." The Census Act,
Justice Stevens wrote, contains two provisions on sampling:
"an unlimited authorization" in section 141(a), giving the
Secretary of Commerce "unqualified authority" to use
sampling in taking the census, and "a limited mandate" in
section 195, directing the Secretary to use sampling.
However, under the second provision, the Secretary may
decide not to use sampling in calculating the apportionment
population or when he considers it infeasible, Stevens
noted. Moreover, he said, the two provisions must be read
consistently, with the provision relating specifically to
the decennial census (section 141) prevailing over the
other.
Justice Stevens said that the 1976 amendments "unambiguously
endorsed the use of sampling." "[T]he Court's
interpretation of the 1976 amendment to section 141 drains
it of any meaning," Stevens wrote.
While the Court's opinion did not reach the Constitutional
issue, the dissenters also concluded that the requirement
for an "actual enumeration" does not proscribe methods
Congress may direct for the census. Quoting an earlier
Supreme Court opinion (relating to the inclusion of overseas
military personnel in the census) that the census must serve
"the constitutional goal of equal representation," Stevens
said that using the most complete and accurate methods would
best serve that purpose. In his opinion concurring with the
four dissenters, Justice Breyer put forth his primary reason
for concluding that the law does not bar the uses of
sampling proposed by the Census Bureau in its plan for 2000.
Breyer said he believes that Congress only intended to
prevent the use of sampling as a substitute for traditional
counting methods, not their use as a supplement to a direct
counting effort. He pointed to the collection of
demographic and housing information on the census long form
as evidence of this distinction, saying that section 195
requires the Secretary of Commerce to collect data for
purposes other than apportionment strictly on a sample
basis. Therefore, Breyer said, the apportionment
"exception" to the use of sampling referred only to the
collection of information based entirely on a sample. He
also noted that the Bureau has used statistical estimation
procedures such as imputation in past censuses to account
for people from whom it could not obtain responses directly.
The Administration's response: At a press conference
yesterday, Commerce Secretary William Daley called the
Court's decision "disappoint[ing]," but noted that the
decision only dealt with the use of sampling for
congressional apportionment and "affirmed the legality of
sampling for other purposes." The Census Bureau, Daley
said, is reviewing both the opinion and the dress rehearsal
results, to ensure that "the 2000 census is designed to
produce the most accurate accounting of the American
people." In response to reporters' questions, he alluded to
the possibility that the Bureau would produce two sets of
census numbers, but did not indicate that such a decision
had been made.
White House Press Secretary Joe Lockhart also called the
Court's opinion "a limited decision under the existing
statute" that did not address whether sampling could be used
for redistricting or the allocation of federal funds.
Questions about the information contained in this News Alert
may be directed to TerriAnn Lowenthal at (202) 484-2270 or,
by e-mail at <terriann2k(a)aol.com>. Please direct all
requests to receive News Alerts, and all changes in
address/phone/fax/e-mail, to Census 2000 at
<Census2000(a)ccmc.org> or 202/326-8700. Please feel free to
circulate this information to colleagues and other
interested individuals.
********** C E N S U S 2 0 0 0 B U L L E T I N
**********
Vol. 3 - No. 1 Jan. 25,
1999
Following the U.S. Supreme Court's 5-to-4 ruling today
against the use, for purposes of congressional
apportionment, of statistical sampling in Census 2000,
Commerce Department Secretary William M. Daley issued the
following statement:
"To begin with, let me say that we are very grateful that
the Supreme Court issued its ruling quickly, although we
obviously are disappointed with the decision.
"Our focus will continue to be conducting the most accurate
census possible. "As the President said in his State of the
Union Address, 'since every person in America counts, every
American ought to be counted.'
"As everyone knows, the 1990 census was the first in fifty
years to be less accurate than its predecessor.
"It contained 12 million mistakes, undercounting millions of
Americans, especially children and members of racial and
ethnic minority groups. "The Census Bureau proposed the use
of sampling to correct these errors, an approach that was
strongly supported by the National Academy of Sciences and
the statistical community.
"It is very important to understand that the only issue
before the Supreme Court was the use of sampling for
apportioning the seats in the House of Representatives among
the States.
"The court itself summarized its ruling by stating 'we
conclude that the Census Act prohibits the proposed uses of
statistical sampling in calculating the population for
purposes of apportionment.'
"In reaching that conclusion, the Court actually affirmed
the legality of sampling for other purposes.
"The Census Bureau is reviewing the Court's opinion and the
results of the dress rehearsal in order to ensure that the
2000 census is designed to produce the most accurate
accounting of the American people and, of course, to comply
fully with the law.
"Thank you very much."
Earlier in the day, Census Bureau Director Kenneth Prewitt
said, "As I have said many times, the Census Bureau is
committed to only one principle -- to provide the most
accurate, scientifically sound census possible in accordance
with the law.
"The Census Bureau career professionals will conduct the
census for 2000 that provides the nation apportionment
numbers that do not rely on statistical sampling." For
further information about Census 2000 Bulletins, contact J.
Paul Wyatt on 301-457-3052 (fax: 301-457-3670; e-mail:
pio(a)census.gov).
January 25, 1999
SUPREME COURT DECISION ON CENSUS IS ISSUED TODAY
The Supreme Court today ruled 5-4 against the use of
statistical sampling for narrow purposes of reapportionment,
the distribution of seats in the U.S. House of
Representatives among the states. But the Court noted that
the Census Act as amended "now requires the Secretary to use
statistical sampling in assembling the myriad demographic
data that are collected in connection with the decennial
census." (p. 21)
The full text of the Court's ruling in Department of
Commerce v. House of Representatives can be found on the
World Wide Web at
http://supct.law.cornell.edu/supct/html/98-404.ZS.html
Further information and reactions will be released as they
become available with a full News Alert later today.
Census Bureau Releases Dress Rehearsal Results
Plan for Census without Sampling Sent to Congress
The Census Bureau released the first set of population
counts from its Census Dress Rehearsal conducted last year
in Sacramento, CA; Columbia, SC, and surrounding areas; and
the Menominee American Indian Reservation, WI. Following
the schedule required by law for the decennial census, the
Bureau published totals for each of the sites within nine
months of the dress rehearsal Census Day in late April.
More detailed population numbers and related socio-economic
information will be available later this year.
The population of Sacramento, where the Bureau executed its
original Census 2000 plan that includes sampling and
statistical methods, was 403,313. The Bureau also reported
that without the use of sampling methods, the count would
have been 349,197 (a difference of 54,116). The quality
check survey (called a post enumeration survey, or PES)
measured a net undercount of 6.3 percent, or 25,572 people,
which was corrected with scientific techniques to reach the
final count. Other uses of sampling and statistical
estimation added 28,544 (7.1 percent of the citys total
population) to the count, as well. The Bureaus 1998
funding bill included a provision sought by sampling
opponents that requires the agency to publish at least two
sets of census counts, produced with and without the use of
statistical methods.
The Bureau also implemented its original census plan in
Menominee County, WI, home to an American Indian reservation
of the same name. The site-level count was 4,738, 3.0
percent (or 143 people) more than the 4,595 people counted
without using statistical methods. The Bureau plans to
visit all unresponsive households on Indian reservations,
instead of conducting follow-up visits on a sample basis,
but it will conduct the quality check survey in these areas
in order to adjust for undercounts. According to the PES
conducted in 1990, that census missed more than 12 percent
of American Indians living on reservations.
In the Columbia, SC, area, the Bureau conducted the dress
rehearsal using only traditional census methods and reported
a population total of 662,140. The Bureau administered a
post-census survey to measure accuracy but has not yet
reported the results of that quality check.
More information about the dress rehearsal is available on
the Bureaus web site, at
http://www.census.gov/Press-Release/www/date.html.
"Traditional" census plan released: The Census Bureau has
sent to Congress a plan for the 2000 census that does not
include sampling and statistical methods to finish the count
of unresponsive households or to correct undercount and
overcounts measured by a post-census survey. The plan is
based on four broad strategies, three of which mirror the
strategies developed for the original census design that
includes some sampling: building partnerships, simplifying
the census process, and using state-of-the-art technology to
improve efficiency and accuracy. The fourth strategy
using "special techniques to improve coverage" replaces
"us[ing] statistical methods" in the original plan.
The Bureau has been preparing to implement two census plans
since the fall of 1998, when Congress tacked such a
requirement onto the agencys annual funding bill. In its
"traditional" census plan, the Bureau says that improvements
to the 1990 census design will "enhance the quality of
Census 2000". However, it also concludes that, "consistent
with the judgements of most other experts in this area, [a]
census using statistical methods would be more accurate
generally, succeed in reducing the differential undercount
of minorities and children, and be more cost-effective than
the alternative."
The traditional plan calls for expanded operations
associated with the direct counting effort, such as opening
more local census offices (520 v. 476 under the original
plan) and hiring more enumerators (not quantified in new
plan). More enumerators will be needed to follow-up with
all households, not just a sample, which do not respond by
mail. Enumerators also will visit all housing units
identified as vacant or nonexistent by the Postal Service,
instead of the sample follow-up originally planned. And
instead of a quality check survey to provide the basis for
correcting undercounts and overcounts using statistical
methods, the Bureau will employ "coverage improvement"
operations to try and increase accuracy. After the
follow-up phase is completed, census workers will
double-check housing units that were listed as vacant by
enumerators on a previous visit, but which did not have an
undeliverable postal address. They also will visit all
households for which a questionnaire returned by mail was
not filled out or was subsequently lost. The alternative
plan includes a post enumeration survey to measure the
accuracy of the counts, but those results will not be used
to correct the official figures.
The Bureau also has decided to accept census responses via
the Internet, regardless of the final census design. This
option will be available only for households receiving the
short form (respondents must provide an identification
number printed on the form they receive at home) and only in
English.
The Bureau noted that its "Census 2000 Operational Plan
Using Traditional Census-Taking Methods" may change as a
result of congressional funding decisions or evaluations of
last years dress rehearsal. The original Census 2000
plan, unveiled in February 1996, has also been modified over
time based on recommendations from the Bureaus advisory
committees and outside evaluators such as the Commerce
Departments Inspector General and National Academy of
Sciences panels.
Congressional field hearings continue: The House Committee
on Government Reform will hold a field hearing in Phoenix,
AZ, on January 29, to examine "community based approaches
for a better enumeration." A primary focus will be efforts
to improve the count of American Indians living on
reservations. According to the post-enumeration survey
conducted as part of the 1990 census, the undercount for
this population was over 12 percent. The hearing will be
held in the Phoenix City Council Chambers, 200 West
Jefferson St., starting at 2:00 p.m. The panels census
subcommittee held its first field hearing in Miami last
December. (The committee has not formally organized its
subcommittees for the 106th Congress.)
Congressional committee assignments: The House and Senate
continue to fill seats on committees and subcommittees that
oversee and fund the census. Because of space
considerations, we will provide that information in a
separate News Alert later this week.
Questions about the information contained in this News Alert
may be directed to TerriAnn Lowenthal at (202) 484-2270 or,
by e-mail at <terriann2k(a)aol.com>. Please direct all
requests to receive News Alerts, and all changes in
address/phone/fax/e-mail, to Census 2000 at
<Census2000(a)ccmc.org> or 202/326-8700. Please feel free to
circulate this information to colleagues and other
interested individuals.
Date: Sat, 9 Jan 1999
From: Census2000 <Census2000(a)ccmc.org>
Changes in House Leadership May Affect Census Oversight;
Census Subcommittee Given Reprieve
Also: What's In Store for 1999?
The recent changes in the Republican leadership of the U.S.
House of Representatives could affect Congressional
monitoring of final preparations for the 2000 census. With
the ascension of Rep. J. Dennis Hastert (R-IL) to the
Speaker's position, plans to shift oversight of the census
to an internal watchdog committee have been scrapped. (See
December14, 1998 News Alert for background.) Instead, the
Subcommittee on the Census will continue its work as part of
the renamed Committee on Government Reform (previously the
Committee on Government Reform and Oversight). Rep. Dan
Miller (R-FL) and Rep. Carolyn Maloney (D-NY) are expected
to continue as the subcommittee's chairman and ranking
minority member, respectively, but Republicans have proposed
reducing the panel's membership to four Republicans and two
Democrats, down from a 5 - 3 split last year. Final
decisions on the subcommittee's structure and membership
will be made when the full committee holds its
organizational meeting.
Hastert, a member of the census subcommittee when it was
created a year ago, was sworn-in as Speaker when the 106th
Congress convened yesterday. The new Speaker also chaired
the House panel overseeing the census in 1997. He has been
a strong critic of the Census Bureau's plan to use
statistical sampling methods in the 2000 census and has
raised concerns about the effect of the census long form on
response rates.
Looking ahead in 1999: As the Census Bureau makes final
preparations for the 2000 census, here are some key
activities and decisions that stakeholders can expect in the
coming months:
* Census budget: The President will submit
his budget request for Fiscal Year 2000 to Congress in early
February. The Census Bureau's original plan calls for
about $2.1 billion for census operations next year.
Congress and the President also must agree on how the Census
Bureau can spend the remainder of this year's $1.027
allocation after June 15, 1999, the deadline set for a final
decision on the use of sampling.
* Congressional oversight: The House and
Senate committees responsible for overseeing census
activities must still reorganize for the new Congress. The
House Subcommittee on the Census has announced tentative
plans for a field hearing in Phoenix, AZ, in late January.
* Legal challenges to sampling: The Supreme
Court is expected to issue a decision in the spring in two
cases challenging the use of sampling in the census. The
plaintiffs in U.S. House of Representatives v. U.S.
Department of Commerce and Glavin v. Clinton contend that
the Census Act and the Constitution prohibit sampling to
determine the State population totals used to apportion
seats in Congress. The Supreme Court heard oral arguments
in both cases on November 30.
* Advisory committees: The 2000 Census
Advisory Committee to the Secretary of Commerce will issue
its final report in mid-February. The report is expected to
include recommendations on the outreach and promotion
campaign, hiring efforts, and the use of sampling to
supplement traditional counting methods.
* Census Monitoring Board: The Board has not
yet announced plans for future meetings or hearings. By
law, it must submit reports to Congress by February 1 and
April 1, 1999, and semi-annually thereafter until its sunset
in September 2001.
* National Academy of Sciences panels: The
Panel to Evaluate Alternative Census Methodologies, chaired
by Dr. Keith Rust of Westat, Inc., is expected to issue its
final report in a month or so. The new Panel to Review the
2000 Census, chaired by former Bureau of Labor Statistics
Commissioner Janet Norwood, will continue to monitor and
evaluate final design decisions and implementation.
* Race and ethnic data: The Office of
Management and Budget (OMB) is expected by March to issue
draft guidelines for tabulating multiple responses to the
question about race in the census and other federal
surveys. OMB revised the official standards for collecting
racial data in October 1997, allowing respondents to report
more than one race but rejecting a proposal to add a new
"multiracial" category.
Congressional hearings (continued): In the final News Alert
of 1998, we provided an initial report on the House census
subcommittee's first field hearing, held in Miami on
December 10. Local officials and community-based
organizations also made the following recommendations to
improve the census process in their testimony before the
panel.
Rep. Carrie P. Meek (D-FL) urged passage of legislation she
introduced in the last Congress to allow recipients of
Federal benefits or pensions to work in temporary census
positions without losing their benefits as a result of the
income they earn in those jobs. Other witnesses
representing historically hard-to-count populations endorsed
Rep. Meek's proposal as a way to promote the hiring of
enumerators who live in the neighborhoods they will
canvass. Many of the witnesses noted that immigrants, who
make up a majority of the Miami area's population, are more
likely to fear giving information about themselves to the
government. They urged the Census Bureau to stress the
confidentiality of census responses in its promotional
campaign.
The Alliance for Aging, Inc. noted that non-traditional
housing arrangements may result in an undercount of
low-income elderly people. A leader in South Florida's large
Haitian-American community cited fear of government
authorities, overcrowded housing conditions, and lack of
English proficiency as significant factors that contributed
to an undercount of Haitians in 1990. This witness
emphasized the need for educational materials in Creole and
Spanish, the importance of hiring enumerators who speak
Creole and are trusted in the community, and the allocation
of funds to begin census outreach efforts in 1999.
Nearly all of the witnesses said that despite their
recommendations for improved outreach and operational
enhancements, they support the use of scientific sampling to
reduce the disproportionate undercount of minorities and the
poor. Rep. Meek called a census design that doesn't include
sampling "a rickety, out-dated, baroquely complex approach
to counting people door-to-door." However, Dario Moreno,
Associate Professor of Political Science at Florida
International University, said that statistical methods
could invite "politicization" of the results by ignoring
local demographic differences in designing the sample to
measure the undercount. He urged the Census Bureau to work
closely with local organizations that serve immigrants and
undertake a "massive foreign-language public education
program" as a way to reduce the undercount.
Media watch: On January 11, National Public Radio will
broadcast a live interview with Census Bureau Director
Kenneth Prewitt on the Diane Rehm Show. The program is
broadcast weekdays at 10:00 a.m. EST from WAMU/88.5 FM in
Washington, D.C., and is carried on many NPR stations across
the country (see the WAMU web site for a list of stations).
Dr. Prewitt will discuss current issues related to the 2000
census. Listeners may call 1-800-433-8850 to ask questions
or comment during the program.
Questions about the information contained in this News Alert
may be directed to TerriAnn Lowenthal at (202) 484-2270 or,
by e-mail at <terriann2k(a)aol.com>. Please direct all
requests to receive News Alerts, and all changes in
address/phone/fax/e-mail, to Census 2000 at
<Census2000(a)ccmc.org> or 202/326-8700. Please feel free to
circulate this information to colleagues and other
interested individuals.
