This is the accessible text file for GAO report number GAO-08-1167T 
entitled '2010 Census: The Bureau's Plans for Reducing the Undercount 
Show Promise, but Key Uncertainties Remain' which was released on 
September 23, 2008.

This text file was formatted by the U.S. Government Accountability 
Office (GAO) to be accessible to users with visual impairments, as part 
of a longer term project to improve GAO products' accessibility. Every 
attempt has been made to maintain the structural and data integrity of 
the original printed product. Accessibility features, such as text 
descriptions of tables, consecutively numbered footnotes placed at the 
end of the file, and the text of agency comment letters, are provided 
but may not exactly duplicate the presentation or format of the printed 
version. The portable document format (PDF) file is an exact electronic 
replica of the printed version. We welcome your feedback. Please E-mail 
your comments regarding the contents or accessibility features of this 
document to Webmaster@gao.gov. 

This is a work of the U.S. government and is not subject to copyright 
protection in the United States. It may be reproduced and distributed 
in its entirety without further permission from GAO. Because this work 
may contain copyrighted images or other material, permission from the 
copyright holder may be necessary if you wish to reproduce this 
material separately. 

Testimony: 

Before the Subcommittee on Federal Financial Management, Government 
Information, Federal Services, and International Security, Committee on 
Homeland Security and Governmental Affairs, U.S. Senate: 

United States Government Accountability Office: 
GAO: 

For Release on Delivery: 
Expected at 10:30 a.m. EDT:
Tuesday, September 23, 2008: 

2010 Census: 

The Bureau's Plans for Reducing the Undercount Show Promise, but Key 
Uncertainties Remain: 

Statement of Robert Goldenkoff: 
Director, Strategic Issues: 

GAO-08-1167T:  

GAO Highlights: 

Highlights of GAO-08-1167T, a testimony before the Subcommittee on 
Federal Financial Management, Government Information, Federal Services, 
and International Security, Committee on Homeland Security and 
Governmental Affairs, U.S. Senate. 

Why GAO Did This Study: 

An accurate decennial census relies on finding and counting peopleóonly 
onceóin their usual place of residence, and collecting complete and 
correct information on them. This is a daunting task as the nationís 
population is growing steadily larger, more diverse, and according to 
the U.S. Census Bureau (Bureau), increasingly difficult to find and 
reluctant to participate in the census. Historically, undercounts have 
plagued the census and the differential impact on various 
subpopulations such as minorities and children is particularly 
problematic. 

GAO was asked to describe (1) key activities the Bureau plans to use to 
help reduce the differential undercount and improve participation, (2) 
the various challenges and opportunities that might affect the Bureauís 
ability to improve coverage in 2010, and (3) how different population 
estimates can impact the allocation of federal grant funds. This 
testimony is based primarily on GAOís issued work in which it evaluated 
the performance of various Census Bureau operations. 

What GAO Found: 

The Bureauís strategy for reducing the undercount and improving 
participation in the 2010 enumeration appears to be comprehensive, 
integrated, and shaped by the Bureauís experience in the 2000 Census. 
If implemented as planned, the various activities the Bureau is 
developing should position the agency to address the undercount. Key 
operations include building a complete and accurate address list, 
implementing an Integrated Communications Campaign to increase 
awareness and encourage participation, and fielding special enumeration 
programs targeted toward historically undercounted populations. For 
example, the Bureau develops its address list and maps over the course 
of a decade using a series of operations that sometimes overlap to 
ensure all housing units are included. Among other activities, 
temporary census workers go door to door across the country in an 
operation called address canvassing to verify addresses. To help find 
hidden housing units, the Bureauís workers look for clues such as two 
mailboxes or utility meters that could indicate additional households. 
Likewise, the Bureauís communications campaign includes paid media, 
public relations, and partnerships with national and grassroots 
organizations, among other efforts, some of which will be targeted 
toward hard-to-count groups. 

Despite the Bureauís ambitious plans, a number of challenges and 
uncertainties remain. For example, the performance of the handheld 
computers that is critical to address canvassing has technical 
shortcomings, while the communications campaign faces the historical 
challenge of converting awareness of the census to an actual response. 
Further, success will depend in large part on the extent to which the 
various operations (1) start and finish on schedule, (2) are 
implemented in the proper sequence, (3) are adequately tested and 
refined, and (4) receive appropriate staffing and funding. It will also 
be important for the Bureau to have a real-time monitoring capability 
to track the progress of the enumeration, target its resources to where 
they are most needed, and to quickly respond to various contingencies 
that could jeopardize the accuracy or cost of the count. 

Our past work indicates that the accuracy of state and local population 
estimates may have an effect, though modest, on the allocation of grant 
funds among the states. Many of the formulas used to allocate grant 
funds rely upon measures of the population, often in combination with 
other factors. For example, we analyzed the sensitivity of Social 
Services Block Grants (SSBG) to alternative population estimates, 
rather than the actual census. We selected SSBG for our analysis 
because the formula, which was based solely on population, and the 
resulting funding allocations were particularly sensitive to 
alternative population estimates. Based on our simulation of the 
funding formula, 27 states and the District of Columbia would have 
gained $4.2 million and 23 states would have lost $4.2 million of the 
$1.7 billion in 2004 SSBG funding. 

What GAO Recommends: 

At this time, GAO is not making any new recommendations. 

To view the full product, including the scope and methodology, click on 
[hyperlink, http://www.gao.gov/cgi-bin/getrpt?GAO-08-1167T]. For more 
information, contact Robert Goldenkoff at (202) 512-2757 or 
goldenkoffr@gao.gov. 

[End of section] 

Mr. Chairman, Senator Coburn, and Members of the Subcommittee: 

Thank you for the opportunity to be here today to discuss the 
challenges the U.S. Census Bureau (Bureau) faces in improving the 
accuracy and coverage of the 2010 decennial Census, and the strategies 
the Bureau plans to employ to reduce the undercount. An accurate 
decennial census relies on finding and counting people--only once--in 
their usual place of residence, and collecting complete and correct 
information from them. This is a daunting task as the nation's 
population is growing steadily larger, more diverse, and according to 
the Bureau, increasingly difficult to find and reluctant to participate 
in the census. Coverage improvement involves reduction in overcounting 
and undercounting. An undercount occurs when the census misses an 
individual who should have been enumerated; an overcount occurs when an 
individual is counted in error. 

What makes these errors particularly problematic is their differential 
impact on various subgroups. Minorities, renters, and children, for 
example, are more likely to be undercounted by the census while more 
affluent groups, such as people with vacation homes, are more likely to 
be enumerated more than once. As census data are used to apportion 
seats in Congress, redraw congressional districts, and allocate 
billions of dollars in federal assistance to state and local 
governments, improving coverage and reducing the differential 
undercount[Footnote 1] are critical. 

The Bureau has long recognized the importance of reducing the 
undercount and, in previous enumerations, has included operations and 
programs designed to improve coverage. As the Bureau moves toward 2010, 
however, besides such long-standing challenges to an accurate 
enumeration as the nation's linguistic diversity and privacy concerns, 
it also faces newly emerging issues such as local campaigns against 
illegal immigration and a post-September 11 environment that could 
heighten some groups' fears of government agencies. 

Today's hearing is particularly timely because, in the months that 
remain until Census Day, April 1, 2010, the Bureau will launch a series 
of operations aimed at reducing the differential undercount including 
building a complete and accurate address list, launching an Integrated 
Communications Campaign to increase awareness and encourage 
participation in the census, and implementing special enumeration 
programs targeted toward undercounted groups. Although not an 
exhaustive list, I am highlighting these actions in my statement to 
help illustrate the range of activities the Bureau employs to improve 
coverage at different phases of the census. 

As requested, my testimony will describe (1) how the Bureau plans to 
use these operations to help reduce the differential undercount and 
improve participation, (2) the various challenges and opportunities 
that might affect the Bureau's ability to improve coverage in 2010, and 
(3) how different population estimates can impact the allocation of 
federal grant funds. 

My remarks today are based primarily on reports we issued from 2000 
through 2008 on the planning and development of the 2010 Census, 
lessons learned from prior censuses, and the impact of population 
measures on federal funding allocations. Please see the final page of 
this testimony for related GAO products. Further, we reviewed recent 
documents on the Bureau's outreach and promotion plans as well as other 
efforts to reduce the undercount and interviewed Bureau officials about 
undercount challenges, plans to improve coverage among hard-to-count 
populations, and progress made towards addressing undercount issues 
from the 2000 Census. 

We conducted our review in accordance with generally accepted 
government auditing standards. Those standards require that we plan and 
perform the audit to obtain sufficient, appropriate evidence to provide 
a reasonable basis for our findings and conclusions based on our audit 
objectives. We believe that the evidence obtained provides a reasonable 
basis for our findings and conclusions based on our audit objectives. 

In summary, the Bureau has developed a wide variety of plans and 
programs to position it to address the differential undercount. 
Further, the Bureau's efforts are designed to reinforce one another, so 
that a household missed in one operation--say, address canvassing--can 
be picked up in a subsequent activity such as nonresponse follow-up. At 
the same time, the Bureau's plans reflect lessons learned from the 2000 
Census. 

Still, a number of hurdles and uncertainties remain, and success will 
depend in large part on the extent to which the various operations (1) 
start and finish on schedule, (2) are implemented in the proper 
sequence, (3) are adequately tested and refined, and (4) receive 
appropriate staffing and funding. It will also be important for the 
Bureau to closely track the progress of key census-taking activities, 
target its resources to where they are most needed, and ensure that it 
has the ability to quickly respond to various contingencies that could 
jeopardize the accuracy of the count. 

Background: 

The Bureau puts forth tremendous effort to conduct an accurate count of 
the nation's population. However, some degree of coverage error in the 
form of persons missed or counted more than once is inevitable. Two 
types of errors that can affect the accuracy of the enumeration are the 
omission of persons who should have been counted and erroneous 
enumerations of persons who should not have been counted. 

Historically, undercounts have plagued the census, although, according 
to the Bureau, they have generally diminished since 1940. For the 2000 
Census, for the first time in its history, the Bureau reported a slight 
net overcount of approximately 0.5 percent or about 1.3 million people. 
However, as shown in figure 1, coverage errors were not evenly 
distributed through the population. For example, there was an overcount 
of non-Hispanic Whites, and an undercount of non-Hispanic Blacks. 
Nevertheless, figure 1 also shows the strides the Bureau made in 
reducing the undercount in the 2000 Census compared to 1990. 

Figure 1: Comparison of Percent Net Undercounts, 1990 and 2000 
Censuses: 

[See PDF for image] 

This figure is a horizontal bar graph depicting the following data: 

Race/Hispanic Origin: American Indian/Alaska Native off reservations; 
1990 Census percent net overcount/undercount: [Empty]; 
2000 Census percent net overcount/undercount: +0.62. 

Race/Hispanic Origin: American Indian/Alaska Native on reservations; 
1990 Census percent net overcount/undercount: +12.22; 
2000 Census percent net overcount/undercount: -0.88. 

Race/Hispanic Origin: Native Hawaiian or other Pacific Islander; 
1990 Census percent net overcount/undercount: [Empty]; 
2000 Census percent net overcount/undercount: +2.12. 

Race/Hispanic Origin: Asian (Non-Hispanic); 
1990 Census percent net overcount/undercount: +2.36; 
2000 Census percent net overcount/undercount: -0.75. 

Race/Hispanic Origin: Hispanic origin; 
1990 Census percent net overcount/undercount: +4.99; 
2000 Census percent net overcount/undercount: +0.71. 

Race/Hispanic Origin: Black (Non-Hispanic); 
1990 Census percent net overcount/undercount: +4.57; 
2000 Census percent net overcount/undercount: +1.84. 

Race/Hispanic Origin: White (Non-Hispanic); 
1990 Census percent net overcount/undercount: +0.68; 
2000 Census percent net overcount/undercount: -1.13. 

Source: U.S. Census Bureau. 

Note: A negative number indicates an overcount. In 1990, Asian (Non- 
Hispanic) included Pacific Islanders. 

[End of figure] 

Importantly, the national net overcount of about 0.5 percent does not 
mean that 99.5 percent of the population was counted correctly in 2000. 
In fact, the number of persons who were counted twice in the census was 
partially offset by the number of persons who were missed by the 
census. We have long maintained that the sum of these numbers--known as 
gross error (rather than the difference between the two numbers or net 
error)--provides a more comprehensive measure of total error in the 
census. 

Participation in the census, as measured by the mail return rate, also 
affects the accuracy of census data. The Bureau calculates mail return 
rates as the percentage of questionnaires the Bureau receives from 
occupied housing units in the mail-back universe.[Footnote 2] Although 
individuals who fail to mail back their census forms might be counted 
by an enumerator during a subsequent operation called nonresponse 
follow-up, high mail return rates are critical to quality data. A 
Bureau evaluation of the 2000 Census found that responses from mail 
returns tend to be more accurate than those obtained during nonresponse 
follow-up. 

Historically, return rates have declined. According to the Bureau, in 
1970, for example, the overall mail return rate was 87.0 percent; in 
1980, 81.3 percent; and in 1990 and 2000, 74.1 percent. Importantly, as 
shown in figure 2, during the 2000 Census, differentials existed in the 
mail return rates of different demographic groups. For example, Whites 
had a higher mail return rate (77.5 percent) than the rate for all 
groups (74.1 percent), while nearly every other demographic group had 
lower return rates than the overall mail return rate. The lowest mail 
return rates were those of Pacific Islanders (54.6 percent) and those 
of two or more races (57.7 percent). Maintaining or increasing mail 
return rates, especially minority return rates, represents an important 
opportunity for the Bureau to improve the quality of census data. 

Figure 2: Return Rates by Race/Ethnic Groups during 2000 Census: 

[See PDF for image] 

This figure is a vertical bar graph depicting the following data: 

Return Rates by Race/Ethnic Groups during 2000 Census: 

Race/Ethnicity: White; 
Percent return rate: 77.5%. 

Race/Ethnicity: Black; 
Percent return rate: 59.7%. 

Race/Ethnicity: American Indian; 
Percent return rate: 64.5%. 

Race/Ethnicity: Asian; 
Percent return rate: 69.8. 

Race/Ethnicity: Pacific Islander; 
Percent return rate: 54.6%. 

Race/Ethnicity: Other race; 
Percent return rate: 58.7%. 

Race/Ethnicity: Two or more races; 
Percent return rate: 57.7%. 

Race/Ethnicity: Hispanic; 
Percent return rate: 64.5%. 

Race/Ethnicity: Non-Hispanic; 
Percent return rate: 75%. 

Return rate for all groups: 74.1%. 

Source: U.S. Census Bureau. 

[End of figure] 

A Number of Census Operations Are Aimed at Improving Coverage: 

In designing the 2010 Census, the Bureau recognized the importance of 
including a number of operations aimed at improving coverage and 
reducing the differential undercount. Three such efforts that I will 
highlight in my remarks today are (1) a complete and accurate address 
list, (2) an Integrated Communications Campaign to increase awareness 
and encourage participation, and (3) special enumeration programs 
targeted toward historically undercounted populations. These 
activities, along with a number of others planned for 2010, will 
position the Bureau to reduce the undercount. At the same time, each 
faces particular challenges and uncertainties that I will describe 
later in my statement. 

Building a Complete and Accurate Address List: 

The foundation of a successful census is a complete and accurate 
address list and the maps that go with it. The Bureau's Master Address 
File (MAF) is the inventory of the nation's roughly 133.7 million 
housing units. In so far as it is used to deliver questionnaires as 
well as to organize the collection and tabulation of the data, the MAF 
serves as the basic control for the census. 

The Bureau develops its address list and maps over the course of the 
decade using a series of operations that sometimes overlap to increase 
the accuracy of the list of all housing units are included. These 
operations include partnerships with the U.S. Postal Service and other 
federal agencies; state, local, and tribal governments; and local 
planning organizations. 

Three operations that can help include the hard-to-count are the 
Bureau's Local Update of Census Addresses (LUCA) program, address 
canvassing, and Group Quarters Validation. 

Local Update of Census Addresses: 

The LUCA program gives state, local, and tribal governments the 
opportunity to review and update the list of addresses and maps that 
the Bureau will use to deliver questionnaires within those communities. 
According to Bureau officials, LUCA helps identify hard-to-count 
populations and "hidden" housing units such as converted basements 
because local governments might know where such dwellings exist and 
have access to local data and records. In October 2008, the Bureau is 
scheduled to complete its reviews of participants' LUCA submissions and 
update the MAF and a related geographic database used for maps. 

Address Canvassing: 

In the address canvassing operation, thousands of temporary Bureau 
employees known as listers verify the addresses of all housing units-- 
including those addresses provided by localities in LUCA--by going door 
to door across the country. As part of this effort, listers add 
addresses that might not be in the Bureau's database. To help find 
hidden housing units it might otherwise miss, listers ask if there is 
more than one residence at a particular address, or to look for clues 
such as an outbuilding or two mailboxes or utility meters that could 
indicate additional households. Indeed, as shown in the picture on the 
left in figure 3, someone could be living in what appears to be a 
storage shed. Likewise, in the picture on the right, what appears to be 
a small, single-family house could contain another apartment as 
suggested by its two doorbells. 

Figure 3: Address Canvassing Can Help Locate Hidden Housing Units: 

[See PDF for image] 

This figure contains two photographs. The first photo includes the 
following caption: Housing unit or storage shed? 

The second photo includes the following caption: Two doorbells could 
indicate two housing units. 

Source: GAO. 

[End of figure] 

Group Quarters Validation: 

While the vast majority of U.S. residents live in residential housing 
units such as single-family houses, apartments and mobile homes, the 
2000 decennial enumerated over seven million people living in group 
situations such as college dormitories, nursing homes, migrant labor 
camps, prisons, and group homes, collectively known as "group 
quarters". Some group quarters, such as seasonal and migrant labor 
camps, can be difficult to locate because they are sometimes fenced-in 
or in remote locations away from main roads. The Bureau encountered a 
number of problems when enumerating group quarters during the 2000 
Census. For example, in 2000, communities reported instances where 
students in college dormitories were counted twice and prison inmates 
were counted in the wrong county. Additionally, group homes are 
sometimes difficult for census workers to spot because, as shown in 
figure 4, they can look the same as conventional housing units. 

Figure 4: Group Quarters or Single-Housing Unit? 

[See PDF for image] 

Photograph of a residential dwelling. 

Source: GAO, 2004. 

[End of figure] 

Since 1970, the Bureau has conducted a separate operation to enumerate 
the group quarters population. For 2010, the Bureau has plans to 
conduct Group Quarters Validation to validate the addresses found in 
the Address Canvassing operation and collect information about the type 
of group quarters. 

Encouraging Participation through the Integrated Communications 
Campaign: 

The Bureau's Integrated Communications Campaign is designed to increase 
the mail response rate, improve cooperation with enumerators, enhance 
the overall accuracy of the census, and reduce the differential 
undercount. The Bureau estimates it will spend $410 million on the 
Integrated Communications Campaign for the 2010 Census. 

In September 2007, the Bureau awarded its communications contract to 
DraftFCB, a communications firm hired to orchestrate a number of 
communications activities for the 2010 Census. DraftFCB's approach 
includes a specific focus on undercounted populations. As one example, 
the contractor worked with the Bureau to segment the nation's 
population into distinct "clusters" using socioeconomic, demographic, 
and other data from the 2000 Census that are correlated with a person's 
likelihood to participate in the census. Each cluster was given a hard- 
to-count score and the Bureau's communications efforts are to be 
targeted to those clusters with the highest scores. The four clusters 
with the highest hard-to-count scores made up 14 percent of the 
nation's occupied housing units based on data from the 2000 Census, and 
included the following demographic characteristics: renters, 
immigrants, non-English speakers, persons without higher education, 
persons receiving public assistance, and persons who are unemployed. 

Targeting the Bureau's communications campaign to hard-to-count 
populations will help the Bureau use its resources more effectively. 
This will be important because in constant 2010 dollars, the Bureau 
will be spending less on communications for the 2010 Census ($410 
million) compared to the 2000 Census ($480 million). 

The campaign strategy will be based on the theme "It's In Our Hands". 
According to the Bureau, this approach reflects a marketplace trend 
where communications are becoming more two-way or participatory, and 
can be seen, for example, in people creating their own content on the 
World Wide Web. The goal of the strategy is to encourage personal 
ownership and involvement that spreads the word about the census. The 
Bureau believes this approach will be more effective than if the 
message came from the government talking to the public. Further, the 
generic theme will be tailored to specific groups. For example, 
outreach targeted to families might carry the message, "The education 
of our children...It's in our hands," while the economically 
disadvantaged might receive "The power to matter...It's in our hands." 

The communications campaign consists of (1) paid media including 
national, local, outdoor, and online advertisements; (2) earned media 
and public relations such as news releases, media briefings, special 
events, podcasts, and blogs; (3) Census in Schools, a program designed 
to reach parents and guardians through their school age children, and 
(4) partnerships with key national and local grassroots organizations 
that have strong connections to their communities. 

Although the effects of the Bureau's communication efforts are 
difficult to measure, the Bureau reported some positive results from 
its 2000 Census marketing efforts with respect to raising awareness. 
For example, four population groups--non-Hispanic Blacks, non-Hispanic 
Whites, Asians, and Native Hawaiians indicated they were more likely to 
return the census form after the 2000 Census Partnership and Marketing 
Program than before its onset. However, the Bureau also reported that 
the 2000 Census Partnership and Marketing Program had mixed success in 
favorably impacting actual participation in the census. 

Of the various campaign components, the Census in Schools and 
partnership programs are specifically aimed at hard-to-count 
populations. The Census in Schools program provides curriculum and 
teaching materials that introduce students to the purpose and 
importance of the census as well as census activities and products. The 
program is also designed to engage students to encourage their parents 
to complete and return their census questionnaires. 

According to Bureau officials, although the Census in Schools program 
is not as extensive as the one conducted in the last decennial, they 
made a number of changes based on lessons learned from the 2000 Census. 
For example, the program will spend less on printing and base their 
2010 Census materials on materials used for the 2000 Census rather than 
create new materials from scratch. Moreover, similar to 2000, the 
Bureau is not reaching out to all schools but instead plans to target 
schools with large hard-to-count populations. Lower grades will be 
targeted as well, as Bureau officials believe their message has more 
traction with younger students. 

Under the partnership program, the Bureau plans to hire specialists to 
collaborate with local individuals and organizations, leveraging their 
knowledge and expertise to increase participation in the census within 
their communities. Partnership specialists are to be trained in, and 
help implement, various aspects of the census, as well as to reach out 
to key government and community leaders and gain commitments from 
community organizations to help the Bureau execute the enumeration. 

Reaching Out to Undercounted Populations Using Special Enumeration 
Activities: 

The Bureau operates a wide range of special enumeration programs--such 
as Be Counted, Questionnaire Assistance Centers (QAC) and Service-Based 
Enumeration--that target hard-to-count populations. Other activities, 
such as offering in-language questionnaires and replacement 
questionnaire mailings for nonresponding households, can help increase 
participation in non-English speaking populations as well as residents 
in areas with historically low responses rates. 

Be Counted program and Questionnaire Assistance Centers: 

The Bureau developed the Be Counted program to enumerate people who 
believe they did not receive a census questionnaire, or were otherwise 
not included in the census. The Be Counted form is a questionnaire 
intended to be placed in public locations such as stores, libraries, 
and other places where people congregate (see figure 5). QAC staff help 
people complete their Be Counted forms as well as other census forms. 
Census officials reported that approximately 560,000 people were 
enumerated through the Be Counted program in 2000 that might have 
otherwise been missed. Additionally, a Bureau evaluation found that Be 
Counted forms were more likely to include members of minority groups 
and children--two traditionally undercounted populations--when compared 
to the traditional mail forms. Plans for the 2010 decennial include 
30,000 QAC sites (a 25 percent increase over the 2000 Census) and 
40,000 Be Counted sites, which are oftentimes co-located with QACs but 
can be stand-alone sites. Partnership specialists are to help determine 
the location of the sites, which are to be operational for 4 weeks 
during the 2010 Census. The Be Counted forms are to be available in 
English, Chinese, Korean, Russian, Spanish, and Vietnamese. 

Figure 5: Be Counted Sites and Questionnaire Assistance Centers Can 
Help Improve Coverage: 

[See PDF for image] 

Photographs of Be Counted Sites and Questionnaire Assistance Centers. 

Source: GAO. 

[End of figure] 

Service-Based Enumeration program: 

The Bureau developed the Service-Based Enumeration program (SBE) for 
the 2000 Census to provide the homeless and others without conventional 
housing an opportunity to be included in the census. The program 
involves visiting selected service locations such as shelters, soup 
kitchens and regularly scheduled mobile food vans that serve people 
without conventional housing. The Bureau reported that during the 2000 
Census, the large percentages of historically undercounted populations 
were among the 171,000 people in emergency and transitional shelters 
enumerated through the program. For 2010, the Bureau plans to conduct 
address list updates of SBE locations by obtaining information about 
SBEs from the Internet and soliciting information from government 
agencies and advocacy organizations. 

In-Language Questionnaires and Other Efforts: 

The Bureau intends to notify respondents through the Integrated 
Communications Campaign that if a questionnaire in one of the 5 
languages other than English (Chinese, Korean, Russian, Spanish, or 
Vietnamese) is needed, the respondent should call the number provided 
on the questionnaire. The Bureau plans to provide language assistance 
guides for 59 languages, an increase from 49 languages in 2000. New in 
2010, the Bureau plans to send bilingual questionnaires to 
approximately 13 million households that are likely to need Spanish 
assistance, as determined by analyzing recent data from the American 
Community Survey (a related Bureau survey program). Moreover, for 2010, 
the Bureau plans a multi-part approach for replacement mailings that 
includes a blanket mailing of approximately 25-30 million replacement 
questionnaires to census tracts with low response rates several weeks 
after the initial questionnaire mailing. 

Challenges and Opportunities to Reduce Undercount: 

Although each of the operations I've described can position the Bureau 
to address the undercount, they also face challenges and uncertainties 
that, if not adequately resolved, could reduce the effectiveness of the 
Bureau's efforts. 

For example, with respect to address canvassing, the Bureau plans to 
provide listers with GPS-equipped handheld computers (HHC) to verify 
and correct addresses. Consequently, the performance of the HHCs is 
critical to the accurate, timely, and cost-effective completion of 
address canvassing. However, the Bureau's ability to collect and 
transmit address and mapping data using the HHC is not known. For 
example, the 2008 Dress Rehearsal--which was an opportunity for the 
Bureau to conduct development and testing of systems and prepare for 
the 2010 Census--revealed a number of technical problems with the HHC 
that included freeze-ups and data transmission issues. The problems 
with the HHC prompted the Bureau to make major design changes, and a 
limited field test is scheduled for December 2008 (GAO is making plans 
to observe this test). However, if after this test the HHC is found to 
be unreliable, the Bureau will have little time to make any 
refinements. 

Operations that were not tested during the Dress Rehearsal also 
introduce risks. These operations include the Be Counted program, 
Service-Based Enumeration, and Group Quarters Enumeration. Although the 
Bureau employed these operations during the 2000 Census, the Dress 
Rehearsal afforded the Bureau an opportunity to see how they might 
perform in concert with other activities planned for 2010, as well as 
identify improvements that could enhance their effectiveness. 

The Integrated Communications Campaign faces its own set of challenges, 
chief among which is the long-standing issue of converting awareness of 
the census to an actual response. As a rough illustration of this 
challenge, various polls conducted for the 2000 Census suggested that 
the public's awareness of the census was over 90 percent. Yet, as noted 
earlier, (1) the actual return rate was much lower--around 74 percent 
of the nation's households, and (2) the Bureau's evaluation of the 2000 
Census Partnership and Marketing Program found that it only had mixed 
success in encouraging actual participation. 

With respect to the partnership program, the Bureau plans to have 144 
partnership staff, including specialists, on-board nationwide by the 
end of September 2008, and ramp up to 680 partnership staff by 2010. 
According to Bureau officials, although this level of staffing is about 
the same as for the 2000 Census, the Bureau believes it is sufficient, 
and plans to deploy the partnership specialists more strategically by 
allocating more partnership specialists to regions with large hard-to- 
count populations. For example, the Atlanta region, (which includes 
Florida, Alabama, and Georgia), had 50 partnership specialists in 2000, 
but is to receive more than 70 partnership specialists in 2010. 
Although the strategic deployment is a reasonable approach, the impact 
of the reallocation on those regions that will receive fewer 
partnership specialists is unclear. 

Our evaluation of the 2000 Census Partnership Program found that there 
were mixed views regarding the adequacy of specialists staffing levels. 
[Footnote 3] Although partnership specialists we spoke to generally 
agreed that the Bureau hired enough specialists to carry out their 
activities, the managers of local census offices we interviewed noted 
that the partnership specialists' heavy workload may have limited the 
level of support they were able to provide. In 2010, to the extent that 
partnership specialists in regions with lower staffing levels wind up 
working with as many or more groups compared to 2000, or need to cover 
large geographic areas, they could find themselves thinly spread. 

Our observations during the 2000 Census highlighted some best practices 
that appeared to be key to successful partnership engagements, and 
might help the Bureau refine its partnering efforts in 2010.[Footnote 
4] For example, best practices for partners include (1) identifying 
'census champions' (i.e., people who will actively support the census 
and encourage others to do so), (2) integrating census-related efforts 
into partners' existing activities and events, and (3) leveraging 
resources by working with other partners and customizing census 
promotional materials to better resonate with local populations. For 
the Bureau, best practices include (1) providing adequate and timely 
information, guidance, and other resources to local partners on how 
they can support the census; (2) maintaining open communications with 
partners; and (3) encouraging the early involvement of partners in 
census activities. 

Another challenge lies in staying on schedule. In order to meet legally 
mandated data reporting requirements, census activities need to take 
place at specific times and in the proper sequence. For example, the 
Group Quarters Validation operation needs to be completed after the 
Address Canvassing operation; the Questionnaire Assistance Centers need 
to be properly staffed, equipped, and opened by a particular date; 
advertising needs to be synchronized with various phases of the 
enumeration; and the questionnaires and replacement mailings all need 
to be carried out at the right time. Given the tight deadlines, small 
glitches could cascade into significant problems with downstream 
operations. 

Another challenge will be to develop management information systems 
capable of tracking key operations to enable the Bureau to quickly 
address trouble spots. The Bureau did this successfully in 2000 with 
the system it used to track local census offices' progress in meeting 
their recruiting goals. At those offices where recruiting was found to 
be lagging, the Bureau was able to quickly raise pay rates and take 
other actions that enabled the Bureau to meet its goal. Less successful 
was the management information system used to track the Bureau's 
partnership efforts in 2000, which was found to be slow and not user- 
friendly, among other shortcomings, which limited its use as an 
effective management tool. For 2010, the Bureau intends to use a Web- 
based system that will enable it to manage the partnerships in real- 
time and determine, among other things, whether staff need to be 
redirected or reallocated. 

Potential Impact of Undercounts on Federal Funding: 

Our past work indicates that the accuracy of state and local population 
estimates may have an effect, though modest, on the allocation of grant 
funds among the states.[Footnote 5] Many of the formulas used to 
allocate grant funds rely upon measures of population, often in 
combination with other factors. In our June 2006 report, we analyzed 
the sensitivity of Social Services Block Grants (SSBG) to alternative 
population estimates, such as those derived by statistical methods that 
incorporate the number of people that were overcounted and undercounted 
in the census, rather than the actual census. 

To analyze the prospective impact of estimated population counts on the 
money allocated to the states through SSBG, we recalculated the state 
allocations using statistical estimates of the population that were 
developed for the 1990 and 2000 Censuses in lieu of the actual census 
numbers. We used the population estimates, which are based on the 2000 
Census counts, and then adjusted these population estimates by the 
difference between the 2000 official population counts and the 
statistical estimates of the population. 

We selected SSBG for our analysis because the formula for this block 
grant program, which was based solely on population, and the resulting 
funding allocations, were particularly sensitive to such alternative 
population estimates. In short, as shown in figure 6, in 2004, 27 
states and the District of Columbia would have gained $4.2 million and 
23 states would have lost $4.2 million of the $1.7 billion in SSBG 
funding. Based on our simulation of the funding formula for this block 
grant program, the largest percentage changes were for Washington, 
D.C., which would have gained 2.05 percent (or $67,000) in grant 
funding, and Minnesota, which would have lost 1.17 percent (or 
$344,000). While the shifting of these funding amounts may not seem 
significant in total, using an inaccurate count to allocate grant funds 
could adversely impact some states' ability to provide services to 
their residents. Reducing the undercount will alleviate this 
potentially adverse impact to states. 

Figure 6: Estimated Social Services Block Grant Percentage Change in 
Grant Funding Using Statistical Population Estimates for States: 

[See PDF for image] 

This figure is a horizontal bar graph depicting the following data: 

Estimated Social Services Block Grant Percentage Change in Grant 
Funding Using Statistical Population Estimates for States: 

District of Columbia: +2.05; 
Montana: +0.94; 
Nevada: +0.91; 
Alaska: +0.75; 
Virginia: +0.73; 
Georgia: +0.71; 
Maryland: +0.70; 
Hawaii: +0.67; 
California: +0.58; 
Texas: +0.51; 
New Mexico: +0.48; 
Colorado: +0.45; 
Louisiana: +0.38; 
Arkansas: +0.37; 
Utah: +0.37; 
North Carolina: +0.32; 
Oklahoma: +0.27; 
Washington: +0.25; 
New York: +0.21; 
Arizona: +0.16; 
Delaware: +0.16; 
Alabama: +0.13; 
Oregon: +0.12; 
South Carolina: +0.11; 
Idaho: +0.07; 
Wyoming: +0.07; 
Tennessee: +0.06; 
Mississippi: +0.05; 
Kentucky: -0.01; 
New Jersey: -0.04; 
Florida: -0.14; 
Connecticut: -0.27; 
West Virginia: -0.27; 
Nebraska: -0.33; 
Pennsylvania: -0.44; 
Michigan: -0.47; 
Massachusetts: -0.53; 
New Hampshire: -0.63; 
Vermont: -0.64; 
Rhode Island: -0.64; 
Maine: -0.70; 
Kansas: -0.79; 
Ohio: -0.79; 
South Dakota: -0.80; 
Missouri: -0.85; 
Illinois: -0.92; 
Iowa: -0.96; 
North Dakota: -0.97; 
Wisconsin: -1.00; 
Indiana: -1.15; 
Minnesota: -1.17. 

Source: GAO analysis of data from the Department of Commerce and the 
Department of Health and Human Services. 

[End of figure] 

This simulation was done for illustrative purposes only--to demonstrate 
the sensitivity of government programs to alternative population 
estimates that incorporate the number of people that were overcounted 
and undercounted in the census. Only the actual census numbers should 
be used for official purposes. This illustration further emphasizes the 
importance of an accurate decennial count. 

Concluding Observations: 

The Bureau's strategy for reducing the differential undercount appears 
to be comprehensive, integrated, and based on lessons learned from the 
2000 Census. If each of the various components is implemented as 
planned, they will likely position the Bureau to address the 
differential undercount. Still, the various programs we examined are 
generally in the planning or early implementation stages, and a number 
of uncertainties and challenges lie ahead as the activities become 
operational. Indeed, past experience has shown how the decennial census 
is an enormous and complex endeavor with numerous moving parts, and any 
shortcomings or missteps can have significant consequences for the 
ultimate cost or accuracy of the enumeration. 

With this in mind, the success of the Bureau's efforts aimed at the 
hard-to-count will depend in large part on the extent to which they (1) 
start and finish on schedule, (2) are implemented in the proper 
sequence, (3) are adequately tested, and (4) receive appropriate 
staffing and funding. It will also be important for the Bureau to have 
a real-time monitoring capability to track the progress of the 
enumeration, target the Bureau's resources to where they are most 
needed, and to quickly respond to various contingencies that could 
jeopardize the accuracy or cost of the count. In the months ahead, it 
will be important for Congress and the Bureau to continue to focus on 
these issues, as well as to be alert to newly emerging challenges. 

Chairman Carper, Senator Coburn, and Members of the Subcommittee, this 
concludes my prepared statement. I would be happy to respond to any 
questions you may have. 

If you have any questions on matters discussed in this testimony, 
please contact Robert Goldenkoff at (202) 512-2757 or by email at 
goldenkoffr@gao.gov. Other key contributors to this testimony include 
Ronald Fecso, Chief Statistician; Signora May, Assistant Director; 
Nicholas Alexander; Thomas Beall; Sarah Farkas; Richard Hung; Andrea 
Levine; Lisa Pearson; Sonya Phillips; Timothy Wexler; and Katherine 
Wulff. 

[End of section] 

Related GAO Products: 

2010 Census: Census Bureau's Decision to Continue with Handheld 
Computers for Address Canvassing Makes Planning and Testing Critical. 
[hyperlink, http://www.gao.gov/cgi-bin/getrpt?GAO-08-936]. Washington, 
D.C.: July 31, 2008. 

2010 Census: Plans for Decennial Census Operations and Technology Have 
Progressed, But Much Uncertainty Remains. [hyperlink, 
http://www.gao.gov/cgi-bin/getrpt?GAO-08-886T]. Washington, D.C.: June 
11, 2008. 

2010 Census: Bureau Needs to Specify How It Will Assess Coverage Follow-
up Techniques and When It Will Produce Coverage Measurement Results. 
[hyperlink, http://www.gao.gov/cgi-bin/getrpt?GAO-08-414]. Washington, 
D.C.: April 15, 2008. 

2010 Census: Population Measures Are Important for Federal Funding 
Allocations. [hyperlink, http://www.gao.gov/cgi-bin/getrpt?GAO-08-
230T]. Washington, D.C.: October 29, 2007. 

2010 Census: Diversity in Human Capital, Outreach Efforts Can Benefit 
the 2010 Census. [hyperlink, http://www.gao.gov/cgi-bin/getrpt?GAO-07-
1132T]. Washington, D.C.: July 26, 2007. 

2010 Census: Census Bureau Has Improved the Local Update of Census 
Addresses Program, but Challenges Remain. [hyperlink, 
http://www.gao.gov/cgi-bin/getrpt?GAO-07-736]. Washington, D.C.: June 
14, 2007. 

2010 Census: Redesigned Approach Holds Promise, but Census Bureau Needs 
to Annually Develop and Provide a Comprehensive Project Plan to Monitor 
Costs. [hyperlink, http://www.gao.gov/cgi-bin/getrpt?GAO-06-1009T]. 
Washington, D.C.: July 27, 2006. 

Federal Assistance: Illustrative Simulations of Using Statistical 
Population Estimates for Reallocating Certain Federal Funding. 
[hyperlink, http://www.gao.gov/cgi-bin/getrpt?GAO-06-567]. Washington, 
D.C.: June 22, 2006. 

2010 Census: Planning and Testing Activities Are Making Progress. 
[hyperlink, http://www.gao.gov/cgi-bin/getrpt?GAO-06-465T]. Washington, 
D.C.: March 1, 2006. 

2010 Census: Basic Design Has Potential, but Remaining Challenges Need 
Prompt Resolution. [hyperlink, http://www.gao.gov/cgi-bin/getrpt?GAO-05-
9]. Washington, D.C.: January 12, 2005. 

Decennial Census: Lessons Learned for Locating and Counting Migrant and 
Seasonal Farm Workers. [hyperlink, http://www.gao.gov/cgi-
bin/getrpt?GAO-03-605]. Washington, D.C.: July 3, 2003. 

Decennial Census: Methods for Collecting and Reporting Data on the 
Homeless and Others without Conventional Housing Need Refinement. 
[hyperlink, http://www.gao.gov/cgi-bin/getrpt?GAO-03-227]. Washington, 
D.C.: January 17, 2003. 

2000 Census: Review of Partnership Program Highlights Best Practices 
for Future Operations. [hyperlink, http://www.gao.gov/cgi-
bin/getrpt?GAO-01-579]. Washington, D.C.: August 20, 2001. 

[End of section] 

Footnotes: 

[1] Differential undercount describes subpopulations that are 
undercounted at a different rate than the total population. 

[2] The mail return rate differs from the mail response rate in that 
the mail response rate is calculated as a percentage of all the housing 
units in the mail-back universe, including those that are later 
discovered to be nonexistent or unoccupied. The Bureau uses this 
percentage as an indicator of its nonresponse follow-up workload. 

[3] GAO, 2000 Census: Review of Partnership Program Highlights Best 
Practices for Future Operations. [hyperlink, http://www.gao.gov/cgi-
bin/getrpt?GAO-01-579] (Washington, D.C.: Aug. 20, 2001). 

[4] [hyperlink, http://www.gao.gov/cgi-bin/getrpt?GAO-01-579]. 

[5] GAO, 2010 Census: Population Measures Are Important for Federal 
Funding Allocations. [hyperlink, http://www.gao.gov/cgi-bin/getrpt?GAO-
08-230T] (Washington, D.C.: Oct. 29, 2007) GAO, Federal Assistance: 
Illustrative Simulations of Using Statistical Population Estimates for 
Reallocating Certain Federal Funding. [hyperlink, 
http://www.gao.gov/cgi-bin/getrpt?GAO-06-567] (Washington, D.C.: Jun. 
22, 2006). 

[End of section] 

GAO's Mission: 

The Government Accountability Office, the audit, evaluation and 
investigative arm of Congress, exists to support Congress in meeting 
its constitutional responsibilities and to help improve the performance 
and accountability of the federal government for the American people. 
GAO examines the use of public funds; evaluates federal programs and 
policies; and provides analyses, recommendations, and other assistance 
to help Congress make informed oversight, policy, and funding 
decisions. GAO's commitment to good government is reflected in its core 
values of accountability, integrity, and reliability. 

Obtaining Copies of GAO Reports and Testimony: 

The fastest and easiest way to obtain copies of GAO documents at no 
cost is through GAO's Web site [hyperlink, http://www.gao.gov]. Each 
weekday, GAO posts newly released reports, testimony, and 
correspondence on its Web site. To have GAO e-mail you a list of newly 
posted products every afternoon, go to [hyperlink, http://www.gao.gov] 
and select "E-mail Updates." 

Order by Mail or Phone: 

The first copy of each printed report is free. Additional copies are $2 
each. A check or money order should be made out to the Superintendent 
of Documents. GAO also accepts VISA and Mastercard. Orders for 100 or 
more copies mailed to a single address are discounted 25 percent. 
Orders should be sent to: 

U.S. Government Accountability Office: 
441 G Street NW, Room LM: 
Washington, D.C. 20548: 

To order by Phone: 
Voice: (202) 512-6000: 
TDD: (202) 512-2537: 
Fax: (202) 512-6061: 

To Report Fraud, Waste, and Abuse in Federal Programs: 

Contact: 

Web site: [hyperlink, http://www.gao.gov/fraudnet/fraudnet.htm]: 
E-mail: fraudnet@gao.gov: 
Automated answering system: (800) 424-5454 or (202) 512-7470: 

Congressional Relations: 

Ralph Dawn, Managing Director, dawnr@gao.gov: 
(202) 512-4400: 
U.S. Government Accountability Office: 
441 G Street NW, Room 7125: 
Washington, D.C. 20548: 

Public Affairs: 

Chuck Young, Managing Director, youngc1@gao.gov: 
(202) 512-4800: 
U.S. Government Accountability Office: 
441 G Street NW, Room 7149: 
Washington, D.C. 20548: