The Census Bureau and Its Accountability

• DOI: 10.1177/0275074006290057
• Author: Phung Nguyen
• Date: 01 June 2007
• Published date: 01 June 2007
• Subject Matter: Articles

The American Review
of Public Administration
Volume 37 Number 2
June 2007 226-243
The Census Bureau and
© 2007 Sage Publications
10.1177/0275074006290057
http://arp.sagepub.com
Its Accountability
hosted at
http://online.sagepub.com
Phung Nguyen
North Carolina Agricultural and
Technical State University, Greensboro
Since the early 1980s, the decennial census has been embroiled in unrelenting controversies
concerning the proposed use of statistical sampling to combat the undercount. These contro-
versies have produced numerous analyses by statisticians, legal scholars, political scientists,
and historians with different attributes and analytical frameworks. This article presents these
controversies from an administrative perspective, with emphasis on accountability and
accountability conflicts. As an administrative agency with its mission to produce accurate and
credible statistics—free from political manipulation—the Census Bureau is accountable to its
hierarchical authority in the Department of Commerce, Congress, the statistical community,
and the courts. Conflicts often take place among these accountability dimensions, breeding
controversies in census operations, hindering the bureau’s efficiency, and threatening the cred-
ibility of census statistics. This article analyzes the causes and consequences of these conflicts
and presents ideas for further exploration on the issue of accountability conflict and its impact
on the quality and credibility of census statistics.
Keywords:
census; accountability; undercount; sampling
Since the early 1980s, the decennial census has been embroiled in unremitting high-
profile controversies concerning the proposed use of statistical sampling to constrain
costs and combat the undercount problem. The proposal resulted in a grave disagreement
between the Department of Commerce’s hierarchical authority and the Census Bureau in
1990 and a full-scaled conflict between Congress and the bureau in 2000. These contro-
versies have been studied extensively by political scientists (Skerry, 2000; Wilson & Davis,
1997), statisticians (Bradburn, 1999; Brown et al., 1999), legal scholars (Taylor, 1996),
and historians (Anderson & Fienberg, 1999) with different attributes and analytical frame-
works. This article presents these controversies from a different perspective, an adminis-
trative perspective, with emphasis on accountability and accountability conflicts. As an
administrative agency with its mission to produce accurate and credible statistics—statistics
free from political manipulation—the Census Bureau has to be accountable to its hierar-
chical authority in the Department of Commerce, Congress, the statistical community,
and the courts. Conflicts often take place between and among these accountability dimen-
sions, breeding controversies in census operations, hindering the bureau’s efficiency, and
Author’s Note: I would like to thank three anonymous reviewers for their insightful comments on the original
version of this article.
Initial Submission: November 7, 2004
Accepted: April 13, 2006
226

---

Nguyen / Census Bureau Accountability
227
threatening the credibility of census statistics. Accountability conflicts have occurred in a
number of censuses and on various issues including counting Americans living abroad,
collecting race–ethnicity data, imputation, sampling, and the like. Sampling-related accountabil-
ity conflicts, however, have proven extraordinary, deserving attention from students of
public administration. The first part of the article presents the decennial census as a political
instrument for allocating political representation in the House of Representatives and the
Bureau of the Census with its with organizational characteristics aimed at protecting the
credibility of census statistics. The second part focuses on the four accountability dimensions
of the Census Bureau. Accountability conflicts and their consequences constitute the core of
the third and fourth parts. Ideas for further exploration aimed at reducing accountability
conflicts to enhance the quality and credibility of census statistics make up the main points of
the last part.
The Decennial Census, the Bureau of the Census,
and the Undercount Problem
In Article I, Section 2, Clause 3, the U.S. Constitution requires Congress to conduct an
“actual Enumeration” of the population every 10 years “in such Manner as they shall by
Law direct.” Relying on this grant of power, Congress has developed detailed legislation to
define the scope of the census, counting procedures, and formulas for allocating political
representation on the basis of census counts. In 1898, Congress established the Bureau of
the Census to supervise the decennial census.
The Census Bureau is a component of the federal statistical system, which consists of 72
statistical agencies within 12 cabinet departments and is loosely coordinated by the Office
of Management and Budget. The Census Bureau is located in the Department of Commerce
and is led by a director appointed by the president with confirmation by the Senate. The
director reports directly to the undersecretary in charge of statistics and administration, who
in turn reports to the secretary of commerce.
Among its many tasks, the Census Bureau compiles censuses of the U.S. population and
housing every 10 years. The goal of the bureau is to employ the headcount method to col-
lect and disseminate accurate, timely, relevant, and credible statistics for government pol-
icy making and policy implementation. The issue of credible statistics is paramount.
Because of its political purposes of allocating political representation, distributing federal
funds, and enforcing civil rights laws, the census scope and operations are shaped by polit-
ical forces. At the same time, to be credible, census statistics must be free from political
manipulation. In other words, the census is shaped by two competing forces: the political
control of the census taking, on one hand, and the credibility requirement of census
statistics to be independent from political manipulation, on the other. Census statistics are
of little or no value if they are actually, or are perceived to be, swayed by political manip-
ulation. Such statistics degrade the viability of the country’s population-based system of
representation and the quality of government policies; and if vital federal statistics are of
little or no credibility, serious questions must be raised about the vitality of American democ-
racy. Statistics must be understood not only as numerical quantities such as counts, aver-
ages, or ratios but also as a body of theories and methods for obtaining, analyzing,
presenting, and explaining credible data (in conformity with the fundamental values of

---

228
The American Review of Public Administration
society) for decision making. The credibility of census statistics is assured by the Privacy
Act of 1974, the Freedom of Information Act of 1966, and the Census Act, enacted in 1957,
amended in 1976, and codified in Title 13 of the U.S. Code.
The Census Act provides a fixed schedule for releasing census statistics. Population counts
for reapportionment must be delivered to the president within 9 months of the census date,
that is, by December 31 of each census year. The president then applies a mathematical for-
mula to these counts to calculate the numbers of representatives each state is entitled. And
within a week of the opening of the next session of Congress, the president reports the results
to the clerk of the House of Representatives. Within 15 days, the clerk of the house informs
each state governor of the number of representatives calculated by the president. In addition,
the act requires that when census population counts are presented to the president, they are
also released to the public and the president has to accept them as they are. The simultaneous
release assures that census counts are free from prior clearance from the White House or
Congress. This explains why the secretary’s action of transmitting data to the White House is
not promulgated in the Federal Register, as normally required by the Administrative
Procedure Act. Finally, although the Census Act gives the ultimate responsibility to conduct
the census to the secretary of the Commerce Department, the bureau’s director maintains full
authority in professional and technical matters. This grant of power enables the director (a) to
fully apply the theory and methods of the discipline of statistics to the compilation, analysis,
and dissemination of accurate, timely, and relevant data for various policy making purposes;
and (b) to acquire opportunities for achieving the perceived and real credibility of census sta-
tistics (i.e., to ensure the bureau’s independence from undue political interference).
Census statistics are used to apportion seats in the House of Representatives and for
other nonapportionment purposes.1 Every year, billions of dollars in federal grants are dis-
tributed to state and local governments using census statistics in the form of relative popu-
lation share, absolute population count, or its reciprocal (for computing income per capita).
This use might be altered if the American Community Survey, an annual sample survey of
250,000 households, is successfully incorporated into the 2010 census.2 State and local
governments rely on census data to draw legislative districts of equal population to comply
with the constitutional one person, one vote mandate and the statutory requirements of the
Voting Rights Act. Federal, state,...