2010 Census: Design Shows Progress, but Managing Technology	 
Acquisitions, Temporary Field Staff, and Gulf Region Enumeration 
Require Attention (24-APR-07, GAO-07-779T).			 
                                                                 
The decennial census is a Constitutionally-mandated activity that
produces data used to apportion congressional seats, redraw	 
congressional districts, and allocate billions of dollars in	 
federal assistance. The Census Bureau (Bureau) estimates the 2010
Census will cost $11.3 billion, making it the most expensive in  
the nation's history. This testimony discusses the Bureau's	 
progress in preparing for the 2010 Census to (1) implement	 
operations to increase the response rate and control costs; (2)  
use technology to increase productivity; (3) hire and train	 
temporary staff; and (4) plan an accurate census in areas	 
affected by hurricanes Katrina and Rita. The testimony is based  
on previously issued GAO reports and work nearing completion in  
which GAO observed recruiting, hiring, and training practices in 
the 2006 test, and visited localities that participated in the	 
Local Update of Addresses Dress Rehearsal as well in the Gulf	 
Coast region.							 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-07-779T					        
    ACCNO:   A68706						        
  TITLE:     2010 Census: Design Shows Progress, but Managing	      
Technology Acquisitions, Temporary Field Staff, and Gulf Region  
Enumeration Require Attention					 
     DATE:   04/24/2007 
  SUBJECT:   Census						 
	     Cost control					 
	     Data collection					 
	     Employee training					 
	     Employees						 
	     Hiring policies					 
	     Hurricane Katrina					 
	     Hurricane Rita					 
	     Operational testing				 
	     Personnel recruiting				 
	     Risk management					 
	     Staff utilization					 
	     Technology 					 
	     Temporary employment				 
	     2010 Decennial Census				 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Product.                                                 **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO-07-779T

   

     * [1]Background
     * [2]The Bureau Has Taken Steps to Increase Response Rates
     * [3]The Bureau's Plans for Greater Use of Automation and Technol
     * [4]The Bureau Can Improve Its Recruiting, Hiring, and Training

          * [5]The Bureau Can Refine Its Approach to Recruiting and Hiring
          * [6]Opportunities Exist for the Bureau to Improve Training for F

     * [7]Bureau Is Designing Decennial Activities in the Geographic A
     * [8]Contact and Acknowledgments
     * [9]Appendix I: Related GAO Products

          * [10]Order by Mail or Phone

Testimony

Before the Subcommittee on Information Policy, Census, and National
Archives, Committee on Oversight and Government Reform, House of
Representatives

United States Government Accountability Office

GAO

For Release on Delivery
Expected at 2:00 p.m. EDT
Tuesday, April 24, 2007

2010 CENSUS

Design Shows Progress, but Managing Technology Acquisitions, Temporary
Field Staff, and Gulf Region Enumeration Require Attention

Statement of Mathew J. Scire
Director, Strategic Issues

GAO-07-779T

Mr. Chairman, Mr. Turner, Members of the Subcommittee:

Thank you for the opportunity to be here today to discuss the status of
the Census Bureau's (Bureau) progress in preparing for the 2010 Census.
Based on issued and ongoing work, my testimony today addresses the
Bureau's efforts to prepare for the next decennial by (1) implementing
operations designed to improve the completeness and accuracy of the census
as well as to increase response rate and hence control costs, (2) using
automation and technology to increase productivity, (3) recruiting,
hiring, and training peak temporary staff of about 600,000 in 2010 in a
challenging environment, and (4) planning how to ensure an accurate
population count in areas affected by hurricanes Katrina and Rita.

As you know, Mr. Chairman, the census is a critical national effort
mandated by the Constitution. Census data are used to apportion seats in
the Congress, redraw congressional districts, allocate billions of dollars
in federal assistance to state and local governments, and for numerous
other public and private sector purposes. In addition, the census is a
complicated undertaking and substantial investment, requiring careful
planning, risk management, and oversight to ensure its ultimate success.
The Bureau estimates the 2010 Census will cost $11.3 billion over its
life-cycle, making it the most expensive census in our country's history,
even after adjusting for inflation. Since the 2000 Census, we have been
examining how the Bureau is preparing for the 2010 Census, including
incorporating lessons learned from the 2000 Census into its planning for
the 2010 decennial. Given the importance of a successful enumeration, for
the last 3 decennials, we have supported an approach to oversight that is
timely, rigorous, constructive, and holds the Bureau accountable for
results.

Today's hearing is particularly timely. The Bureau is now conducting the
2008 Dress Rehearsal's Local Update of Census Addresses (LUCA) program,
where local, state, and tribal governments are given the opportunity to
review and suggest changes to Census Bureau address lists and maps.
Beginning in early May 2007, the Bureau will deploy the hand-held mobile
computing devices (MCDs)--a keystone to the reengineered Census--to verify
address data as part of the address canvassing portion of the Dress
Rehearsal. We plan to be on-hand to observe the functionality and
usability of the MCDs at the dress rehearsal sites in North Carolina and
California where the Bureau will conduct a dry run of the full enumeration
planned for 2010. "Census Day" for this exercise is April 1, 2008.

As we have testified in the past, the Bureau's ongoing reengineering of
the decennial census--including changes in the survey design and greater
use of technology--could have important benefits in improved efficiencies
and cost-containment as well as the accuracy, quality, and consistency of
data collected. But these changes, as well as intervening events, raise
new risks that bear watching. Also, given the sheer size of census
operations, refinements to recruiting, hiring, and training can have
substantial results.

In summary, our recent work on the reengineering of and preparations for
the 2010 decennial have yielded a number of observations about actions the
Bureau has taken to promote an accurate and cost-effective census.

           o The Bureau has taken steps to increase response rates through
           such measures as moving to a short form, and utilizing a second
           mailing.
           o The MCD version being deployed during the upcoming dress
           rehearsal will be used for the first time in the field--this is
           the prototype under contract for the 2010 Census--and if it does
           not function as expected or needed, little time will be left for
           the Bureau to take corrective action. Overall, the Bureau's
           greater reliance on contractor-developed automation and technology
           for the 2010 Census call for greater focus on sound acquisition
           and management of these key investments.
           o The Bureau's efforts to recruit, hire, and train a sufficient
           workforce to enumerate an increasingly hard-to-find and reluctant
           population in a more technology-dependent census presents a unique
           challenge for the Bureau to refine its recruiting practices and
           enhance its training.
           o The intervention of hurricanes Katrina and Rita have placed
           additional demands on the Bureau to prepare for enumerating a
           large population displaced by these devastating storms, in an
           environment in which local governments' capacities are constrained
           and physical infrastructure and services have not yet returned to
           normal.

           Given these complexities, our message remains that the risks
           associated with the decennial must be closely monitored,
           evaluated, and managed, with mitigation plans in place where
           appropriate, to help ensure that accurate results are delivered on
           time and within projected costs.

           My remarks today are based primarily on reports that we have
           issued from 2002 through July 2006 on the planning and development
           of the 2010 Census, as well as the results of work nearing
           completion. (Please see app. I for a list of relevant reports.)
           For the 2004 field test, we visited Queens, New York, and several
           counties in rural south-central Georgia. We visited the Texas and
           South Dakota test sites during the 2006 field test. During these
           visits we observed such operations as the address canvassing
           operation--where workers go door to door verifying addresses and
           updating maps as part of the Bureau's effort to build a complete
           and accurate address list, and we observed the non-response
           follow-up operation (NRFU)--where enumerators collect information
           from those households that do not return their initial
           questionnaire. We also observed key recruiting, hiring, and
           training activities during the 2006 test completed last summer.
           During the autumn of 2006, we observed preparations for and the
           conduct of the Local Update of Census Addresses (LUCA) phase of
           the 2008 Dress Rehearsal in sites located in North Carolina and
           California, and in January 2007 we visited areas in Louisiana,
           Mississippi, and Texas affected by hurricanes Katrina and Rita. We
           conducted our work in accordance with generally accepted
           government auditing standards.
		   
		   Background

           The decennial census is the nation's largest, most complex survey.
           In April 2009, address canvassing--a field operation for verifying
           and correcting addresses for all households and street features
           contained on decennial maps--will begin. One year later, the
           Bureau will mail census questionnaires to the majority of the
           population in anticipation of Census Day, April 1, 2010. Those
           households who do not respond will be contacted by field staff
           through the NRFU operation to determine the number of people
           living in the house on Census Day, among other information. In
           addition to address canvassing and NRFU, the Bureau conducts other
           operations, for example, to gather data from residents from group
           quarters, such as prisons or college dormitories. The Bureau also
           employs different enumeration methods in certain settings, such as
           remote Alaska enumeration, in which people living in inaccessible
           communities must be contacted in January 2010 in anticipation of
           the spring thaw which makes travel difficult, or update/enumerate,
           a data collection method involving personal interviews, used in
           communities where many housing units may not have typical house
           number-street name mailing addresses. Further, the efforts of
           state and local government are enlisted to obtain a more complete
           address file through the LUCA program.

           The census is also conducted against a backdrop of immutable
           deadlines, and the census's elaborate chain of interrelated pre-
           and post-Census Day activities is predicated upon those dates. The
           Secretary of Commerce is legally required to (1) conduct the
           census on April 1 of the decennial year, (2) report the state
           population counts to the President for purposes of congressional
           apportionment by December 31 of the decennial year, and (3) send
           population tabulations to the states for purposes of redistricting
           no later than 1 year after the April 1 census date. To meet these
           mandated reporting requirements, census activities must occur at
           specific times and in the proper sequence. The table below shows
           some dates for selected, key decennial activities.

           Table 1: Timeline of Selected Key Decennial Events

           Source: GAO summary of Census Bureau data.

           The Bureau estimates that the 2010 Census will cost $11.3 billion
           over its life-cycle, making it the most expensive in the nation's
           history. While some cost growth is expected, partly because the
           number of housing units has increased, the estimated cost
           escalation has far exceeded the housing unit increase. The Bureau
           estimates that the number of housing units for the 2010 Census
           will increase by 10 percent over 2000 Census levels, but the
           average 2010 cost to enumerate a housing unit is expected to
           increase by about 29 percent from 2000 levels (from $56 to $72)
           (see fig. 1). As the Bureau plans for 2010, maintaining cost
           effectiveness will be one of the single greatest challenges
           confronting the agency.

           Figure 1: Decennial Census Average Cost per Housing Unit (in
           Constant Fiscal Year 2000 Dollars)

           According to the Bureau, the increasing cost of the census is
           caused by various societal trends--such as increasing privacy
           concerns, more non-English speakers, and people residing in
           makeshift and other nontraditional living arrangements--making it
           harder to find people and get them to participate in the census.
		   
		   The Bureau Has Taken Steps to Increase Response Rates

           The Bureau has reengineered the decennial census, including
           implementing new initiatives aimed at increasing the response
           rate. Furthermore, the Bureau also plans to begin to implement its
           outreach and communications campaign, an effort used in the 2000
           Census that was designed to increase awareness and encourage
           individuals to respond to the census questionnaire. Increasing the
           decennial's response rate can result in significant savings
           because the Bureau can reduce the staffing and costs related to
           NRFU, as well as yield more complete and accurate data. According
           to the Bureau, for every one-percentage point increase in the
           response rate, the Bureau will be able to save $75 million.

           The Bureau plans to increase response rate by several means,
           including conducting a short-form-only census. The Bureau is able
           to do this because in 1996 the Bureau began efforts to replace the
           decennial long form with the American Community Survey. Since
           1970, the overall mail response rate to the decennial census has
           been declining steadily, in part, because of the burden of
           responding to the long form, which was sent to a sample of
           respondents. In the 1980 Census, the overall mail response rate
           was 75 percent, 3 percentage points lower than it was in the 1970
           Census. In the 1990 census, the mail response rate dropped to 65
           percent but in 2000 appeared to be leveling off at about 64
           percent. In the 2000 Census when comparing the short form to the
           long form the Bureau found the short form response rate of 66.4
           percent was 12.5 percentage points higher that the long form
           response rate of 53.9 percent. While the difference between the
           long and short form response rates are significant, the Bureau in
           its initial assumption for the 2010 Census predicted that
           conducting a short-form-only census will yield only a 1-percent
           increase in the overall mail response rate.

           A targeted second mailing to households that fail to respond to
           the initial census questionnaire can increase the ultimate
           response rate. According to Bureau studies, sending a second
           questionnaire could yield a gain in overall response of 7 to 10
           percentage points from non-responding households, thus potentially
           saving the Bureau between $525 million to $700 million dollars
           (given that every 1 percentage point increase in response may save
           $75 million). In reports, we have highlighted that a targeted
           second mailing could boost the mail response rate, which in turn
           would result in considerable savings by reducing the number of
           costly personal visits enumerators would need to make to
           nonresponding households.^1 The Bureau has never before included
           this operation as part of a decennial census and over the decade
           has been testing its feasibility. A targeted second mailing was a
           part of 2006 test and boosted the response rate by 8.8 percent at
           the Austin, Texas test site. According to Bureau officials
           targeted second mailing will be a part of the 2010 Census design.
		   
^1GAO, 2010 Census: Basic Design Has Potential, but Remaining Challenges
Need Prompt Resolution, [11]GAO-05-9 (Washington, D.C.: Jan. 12, 2005) and
GAO, 2000 Census: Contingency Planning Needed to Address Risk That Pose a
Threat to a Successful Census, [12]GAO/GGD-00-6 (Washington, D.C.: Dec.
14, 1999).

           For the 2010 Census the Bureau also intends to increase response
           rates by undertaking a public awareness campaign as it did in the
           previous census. In the 2000 Census that effort was comprised of
           two major activities:
		   
		   o conducting the first-ever paid advertising campaign aimed at
           increasing the mail response rate, including the historically
           undercounted populations, and
           o leveraging the value of local knowledge by building 140,000
           partnerships at every level including state, local , and tribal
           governments; community-based organizations; and the media and
           private-sector organizations to elicit public participation in the
           census.

           In 2001 we reported that for the 2000 Census, it appeared that
           encouraging people to respond to the census questionnaire was
           successful, in part due to the Bureau's partnership efforts.^2 For
           example, according to the Bureau, it achieved an initial mail
           response rate of about 64 percent, 3 percentage points higher that
           it had anticipated when planning for NRFU. This was a noteworthy
           accomplishment and, as a result, the Bureau had over 3 million
           fewer housing units to follow-up with than it had initially
           planned.

           The Bureau will soon begin its outreach and communication effort
           for 2010. The Bureau plans to award the communications contract in
           August 2007 and will begin hiring partnership specialists at
           headquarters starting in fiscal year 2008.
		   
		   The Bureau's Plans for Greater Use of Automation and Technology
		   Demand Greater Risk Management

           The MCD is a keystone to the reengineered census. It allows the
           Bureau to automate operations and eliminate the need to print
           millions of paper questionnaires and maps used by census workers
           to conduct address canvassing and NRFU, as well as assisting to
           manage field staff's payroll. The benefits of using the MCD were
           tested in the 2004 and 2006 tests. According to the Bureau, during
           the 2004 Census Test, the MCD allowed the Bureau to successfully
           remove over 7,000 late mail returns from enumerators' assignments,
           reducing the total NRFU workload by nearly 6 percent. The ability
           to remove late mail returns from the Bureau's NRFU workload
           reduces costs, because census field workers no longer need to make
           expensive follow-up visits to households that return their
           questionnaire after the mail-back deadline. If the Bureau had
           possessed this capability during the 2000 Census, it could have
           eliminated the need to visit nearly 773,000 late-responding
           households and saved an estimated $22 million (based on our
           estimate that a 1-percentage-point increase in workload could add
           at least $34 million in direct salary, benefits, and travel costs
           to the price tag of NRFU^3).
		   
^2GAO, 2000 Census: Review of Partnership Program Highlights Best
Practices for Future Operations, [13]GAO-01-579 (Washington, D.C.: Aug.
20, 2001).

^3 [14]GAO/GGD-00-06.

           However, the Bureau's ability to collect and transmit data using
           the MCD is not fully tested and, at this point, constitutes a risk
           to the cost-effective implementation of the 2010 Census. During
           the 2004 test of NRFU and the 2006 test of address canvassing, the
           MCDs experienced significant reliability problems. For example,
           during the 2004 Census Test, the MCDs experienced transmission
           problems, memory overloads, and difficulties with a mapping
           feature--all of which added inefficiencies to the NRFU
           operation.^4 Moreover, during the 2006 test, the MCD's global
           positioning system (GPS) receiver, a satellite-based navigational
           system to help workers locate street addresses and collect
           coordinates for each structure in their assignment area, was also
           unreliable.

           Bureau officials believe the MCD's performance problems will be
           addressed through a contract awarded on March 30, 2006, to develop
           a new MCD. A prototype of the MCD has been developed and delivered
           by the contractor for use in the 2008 Dress Rehearsal. However,
           operational testing of the MCD will not occur until May 2007, when
           address canvassing for the 2008 Dress Rehearsal occurs, and if
           problems do emerge, little time will be left to develop, test, and
           incorporate refinements. In our May 2006 report, we highlighted
           the tight time frames to develop the MCD and recommended that
           systems being developed or provided by contractors for the 2010
           Census--including the MCD--be fully functional and ready to be
           assessed as part of the 2008 Dress Rehearsal. We are currently
           reviewing the cost, schedule and performance status of the
           contract for the MCDs.

           We plan to visit the dress rehearsal sites to determine the
           functionality of the devices to collect and transmit data. If
           after the 2008 Dress Rehearsal the MCD is found not to be
           reliable, the Bureau could be faced with the daunting possibility
           of having to revert to the costly, paper-based census used in
           2000.

           Although the greater use of automation offers the prospect of
           greater efficiency and effectiveness, these actions also introduce
           new risks. The automation of key census processes involves an
           extensive reliance on contractors. Consequently, contract
           oversight and management becomes a key challenge to a successful
           census. As part of the Bureau's plans to increase the use of
           automation and technology for the 2010 Census, the Bureau
           estimates that it will spend about $ 3 billion on information
           technology (IT) investments. The Bureau will be undertaking
           several major acquisitions, including the Decennial Response
           Integration System (DRIS)--a system for integrating paper and
           telephone responses; the Field Data Collection Automation (FDCA)
           program--the systems and support equipment for field office data
           collection activities including the MCDs to be used by
           enumerators; the Data Access and Dissemination System (DADS II)--a
           system for tabulating and disseminating data from the decennial
           census and other Bureau surveys to the public; and the
           modernization of the Master Address File/Topologically Integrated
           Geographic Encoding and Referencing (MAF/TIGER) system, which
           provides the address list, maps, and other geographic support
           services for the decennial and other Bureau surveys, known as the
           MAF/TIGER Accuracy Improvement Project (MTAIP). Together these and
           other systems are to support collection, processing, and
           dissemination of census data.

^4 [15]GAO-05-9 .

           In March 2006, we testified on the Bureau's acquisition and
           management of two key information technology system acquisitions
           for the 2010 Census--FDCA and the DRIS.^5 We reported on the
           Bureau's progress in implementing acquisitions and management
           capabilities for these initiatives. To effectively manage major IT
           programs, organizations should use sound acquisition and
           management processes to minimize risk and thereby maximize chances
           for success. Such processes include project and acquisition
           planning, solicitation, requirement development and management,
           and risk management. We reported that while the project offices
           responsible for these two contracts have carried out initial
           acquisition management activities, neither office had the full set
           of capabilities they needed to effectively manage the
           acquisitions, including implementing a full risk management
           process. We also made recommendations for the Bureau to implement
           key activities needed to effectively manage acquisitions. For
           example, we recommended that the Bureau's project office for DRIS
           complete a project plan and obtain stakeholder concurrence before
           initiating additional development work and obtain validation,
           management, and customer approval of DRIS requirements. In
           response to our recommendation, the Bureau has finalized the
           project plan for DRIS and has obtained stakeholders' commitment.
           As a result, the DRIS project office will have the direction that
           it needs to successfully avoid unanticipated changes.
		   
^5GAO, Census Bureau: Important Activities for Improving Management of Key
2010 Decennial Acquisitions Remain to be Done, [16]GAO-06-444T
(Washington, D.C.: Mar. 1, 2006).

           We are reviewing the cost, schedule, and performance status for
           DRIS, FDCA, MTAIP, and DADS II to determine whether the Bureau is
           adequately managing risks associated with these key systems.
           Effective risk management includes identifying and analyzing
           risks, assigning resources, developing risk mitigation plans and
           milestones for key mitigation deliverables, briefing senior-level
           managers on high-priority risks, and tracking risks to closure and
           is an important project management discipline to ensure that key
           technologies are delivered on time, within budget, and with the
           promised functionality. This discipline is essentially important
           given the immovable decennial census deadline. We are scheduled to
           complete our work on that engagement by September 2007.
		   
		   The Bureau Can Improve Its Recruiting, Hiring, and Training Efforts

           Prior to Census Day, Bureau field staff perform the address
           canvassing operation, during which they verify the addresses of
           all housing units. The Bureau estimates spending $350 million to
           hire about 74,000 field workers for the address canvassing
           operation. About 1 year later, the Bureau mails out questionnaires
           to about 130 million households nationwide. However, the Bureau
           expects that about 40 million households will not return the
           questionnaire. To collect information from those households, the
           Bureau hires temporary field staff--based out of local census
           offices--to visit each nonresponding household in its NRFU
           operation. The Bureau expects to spend over $2 billion to employ
           about 525,000 temporary field staff for that activity. As shown in
           fig. 2, in total the Bureau will recruit and test 3.8 million
           applicants for addressing canvassing and NRFU, hiring some 600,000
           people for the 2010 Census.
	
Figure 2: The Bureau's Recruiting and Hiring Timeline for Temporary Field
Staff During the 2010 Census

The Bureau Can Refine Its Approach to Recruiting and Hiring Temporary Field
Staff

For the 2010 Census, the Bureau plans to use a similar approach to recruit
and hire workers as it used during Census 2000. These strategies made the
Bureau a more attractive employer to prospective candidates and helped
provide a steady stream of applicants during Census 2000. Despite a tight
labor market, the Bureau attracted about 3.7 million qualified applicants
and hired about half a million enumerators at peak. Some of the broad
approaches from 2000 that the Bureau plans on implementing for the 2010
census include

           o recruiting five times more applicants than the needed number of
           field workers to ensure a considerable depth in the applicant pool
           from which to hire;
           o "frontloading" or hiring twice the number of people needed to do
           the work in anticipation of high levels of turnover;
           o exercising the flexibility to raise pay rates for local census
           offices that were encountering recruiting difficulties; and
           o launching a recruitment advertising campaign, which totaled over
           $2.3 million for Census 2000.

           As in 2000, the Bureau faces the daunting tasks of meeting its
           recruiting and hiring goals. However, it also faces additional
           challenges, such as demographic shifts whereby the population is
           increasingly diverse and difficult to locate, and newer
           challenges, like the Bureau's use of handheld computers for data
           collection in the field. It does plan some improvements to how it
           recruits and hires its temporary workforce to carry out the 2010
           Census. For example, the Bureau has conducted and incorporated
           information collected from employee debriefings that could improve
           its recruiting and hiring processes. Bureau officials believe this
           feedback would be helpful in evaluating and refining its hiring
           and recruiting processes and intend to incorporate some of that
           information for the 2008 Dress Rehearsal. However, it can do more
           to target its recruitment of field staff.

           The Bureau casts a wide net to recruit its temporary workforce to
           ensure it has a large enough applicant pool from which to hire. In
           commenting on a draft of this work, Commerce noted that the
           Bureau's priority is to reach out as broadly a possible to the
           diverse communities in the county to attract several million
           applicants. We recognize that when recruiting and hiring for
           hundreds of thousands of positions, the Bureau faces a challenge
           in assessing applicants' potential success or willingness to work.
           However, opportunities exist for the Bureau to hone its recruiting
           efforts to identify individuals who would be more likely to be
           effective at census work and willing to continue working
           throughout an operation. Along those same lines, the Bureau could
           also evaluate the factors associated with an applicant's success,
           willingness to work in an operation, and likelihood of attrition
           to refine its hiring. Despite Commerce's reservations about
           refining its current recruiting and hiring strategies, we believe
           that the Bureau could do more to understand what makes for a
           successful recruit and, by hiring such applicants, reduce
           operating costs.

           Another recruiting and hiring issue we identified in our completed
           work is related to how the crew leaders are selected. We found
           that the Bureau's tools for hiring crew leaders could better
           distinguish the skills needed for those positions. Crew leaders
           fill an important role in the Bureau's field activities because
           they supervise the work of crews of field workers; train field
           workers; and will be counted on to troubleshoot the MCDs. We found
           that despite the different skill requirements of crew leaders and
           other field staff--for example, while it was important for field
           staff working in the NRFU operation to have arithmetic and visual
           identification skills, crew leaders need those skills as well as
           additional skills, such as management, leadership, and creative
           thinking--the Bureau used the same set of hiring tools to hire
           individuals for crew leaders and other field positions during the
           2006 Census Test. In its review of the 2004 Census Test, the
           Department of Commerce Office of Inspector General OIG also
           reported that Bureau officials said that the applicants' the
           multiple-choice test does not capture the technical or supervisory
           skills needed by crew leaders.^6

           The Bureau hired a contractor to assess whether the tools used
           during the 2006 Census Test selected individuals with the skills
           necessary to conduct field work using MCDs;^7 however, the Bureau
           has no current plans to make changes to its hiring process that
           would include differentiated hiring tools for crew leaders and
           other positions. Without hiring tools that distinguish between
           skills needed for the crew leader and other positions, the Bureau
           does not have assurances that it is selecting crew leaders that
           can best perform important duties like providing training,
           managing other field staff, and troubleshooting handheld
           computers. In commenting on our draft, Commerce indicated that the
           Bureau needs to evaluate its hiring tools. It is also working to
           identify and test what the appropriate skills are for the crew
           leader position.

           Finally, we found that the Bureau does not collect performance
           data needed to rehire former workers from prior or ongoing
           operations to whom it may give hiring priority. Officials say they
           try to exclude those terminated for cause (such terminations can
           result when workers have performance or conduct problems such as
           selling drugs or striking another worker). Bureau officials point
           to its internal systems, which, they say preclude the rehiring of
           employees who were terminated for cause. However, the OIG and
           field officials told us that poor performers may not always be
           terminated. Without better information on employee performance,
           the Bureau cannot ensure that the weakest performers are not
           rehired. Over the course of the 2006 Census Test, almost 15
           percent of all field staff were rehired. If this percentage were
           to be rehired during the 2010 Census, the Bureau would not have
           performance data to meaningfully evaluate whether to rehire
           approximately 90,000 individuals. The Bureau believes that the
           pace of the decennial, particularly NRFU, is such that local
           census officials would not have enough time to consider past
           performance when making hiring decisions. However, we believe that
           the Bureau has enough time. For example, performance data could be
           collected during address canvassing to be used to assess workers
           for NRFU, nearly one-year later.
		   
^6Department of Commerce, Office of Inspector General, Improving Our
Measure of America: What the 2004 Census Test Can Teach Us in Planning for
the 2010 Decennial Census, OIG-16949 (Washington, D.C.: September 2004).

^7Bureau officials told us that final results of this study are not yet
available.

           Opportunities Exist for the Bureau to Improve Training for Field
		   Staff

           The Bureau has employed essentially the same approach to training
           since the 1970 Census. To conduct training, the Bureau solicits
           free or low-cost training spaces from local organizations, such as
           churches or libraries. Training classes typically include 15 to 20
           students. Crew leaders usually train their crews, with the
           assistance of at least one crew leader assistant, using a verbatim
           training approach, whereby crew leaders read training scripts
           word-for-word over the course of several days. Similarly, the crew
           leaders were themselves trained by their supervisors in a
           "train-the-trainers" approach. The length of training varies by
           operation; for NRFU, training took almost 42 hours over the course
           of 6 days during the 2006 test.

           The Bureau and others, including us, have reported that the Bureau
           should consider alternate approaches to training delivery. Our
           review of the 2004 Census Test found that, as a result of the
           demographic and technological changes that have taken place since
           1970, the Bureau might want to explore alternatives to its
           verbatim approach to training.^8 Moreover, in 2004, the OIG
           suggested the Bureau explore the use of interactive training
           methods, as the Bureau does for other non-decennial surveys.^9 For
           example, while many field staff we contacted during the 2006 test
           said their overall impression of training was generally positive,
           many added that videos or visuals would or might improve training.
           In addition, while the Bureau is providing some computer-based
           training on using the handheld computers in key operations,
           overall the Bureau has made limited changes to the approach it
           uses to deliver training and has not evaluated alternative
           approaches to providing training. It is notable that observations
           during the 2004 and 2006 tests showed that field staff may have
           missed important parts of training. Contractor employees saw
           students playing games on their MCDs during training for the 2006
           test,^10 and in 2004 the OIG saw students not paying attention and
           falling asleep in class, concluding that some may not have learned
           how to conduct census operations.^11
		   
^8 [17]GAO-05-9.

^9Department of Commerce OIG, OIG-16949.

^10The MCDs developed by Harris will not include software that will allow
field staff to play games during training.

^11Department of Commerce OIG, OIG-16949.

           The content of the Bureau's training for field staff also has not
           changed substantially since Census 2000, despite the fact that,
           according to the Bureau itself, collecting data from the nation's
           population has become increasingly difficult. Field workers we
           spoke to during the 2006 test noted two related issues on which
           they had not received sufficient training--dealing with reluctant
           respondents and handling location-specific challenges.

           According to the Department of Commerce OIG, in 2004 field staff
           complained that they felt unprepared to deal with reluctant
           respondents; the OIG report recommended the Bureau consider adding
           content to enhance training on this topic.^12 Moreover, our review
           of the Bureau's summaries of debriefings it conducted after the
           2006 test indicated that field staff found respondent reluctance
           to be a challenge.^13 Crew leaders noted that this was the most
           difficult task enumerators faced. In our field visits, we observed
           that without adequate preparation in dealing with reluctant
           responders, field staff developed their own strategies when
           confronted with these situations, resulting in inconsistent and
           sometimes inappropriate data collection methods. For example, when
           unable to contact respondents, one Texas enumerator looked up
           respondent information online, tried to find a phone number for
           another respondent from a neighborhood cat's collar, and illegally
           went through residents' mail.

           Field staff may also need more training in overcoming
           location-specific challenges, such as rural conditions on the
           Cheyenne River Indian Reservation in South Dakota; and counting
           the transient student population in Austin, Texas. For example, in
           Austin, one crew leader explained that training spent a lot of
           time on mobile homes--which did not exist in his area--but very
           little time on apartment buildings, which are common there. Based
           on our observations of the 2004 test, we suggested that the Bureau
           supplement the existing training with modules geared toward
           addressing the particular enumeration challenges that field staff
           are likely to encounter in specific locales.^14 During this
           review, the Bureau told us that it works with regional offices to
           develop 10-minute training modules for specific locations. For
           example, in 2000, Bureau officials said enumerators in Los Angeles
           were trained to look for small, hidden housing units, such as
           apartments in converted garages. Bureau officials said they
           provide guidance on the length of the modules and when they should
           be presented. However, they said they were not sure how often this
           kind of specialized training took place, nor had they allocated
           time during training to present specialized information.
		   
^12Department of Commerce OIG, OIG-16949.

^13As previously discussed, these sessions aimed to obtain information
that will improve Bureau procedures, including training. We reviewed
summaries of debriefings conducted for three operations--NRFU,
update/enumerate, and address canvassing.		   

           We believe the Bureau could do more to assist local offices
           provide training that recognizes local conditions. Specifically,
           based on work we will be reporting shortly, we will recommend that
           the Bureau centrally develop training modules covering enumeration
           strategies in a variety of situations, such as mobile homes, large
           apartment buildings, and migrant worker dwellings, which local
           officials can selectively insert into their training if there is a
           need to train their field staff on that topic. Such modules would
           enhance the effectiveness of training by giving greater attention
           to the challenges field staff are likely to face. In commenting on
           this recommendation, Commerce noted that the Bureau works with
           managers in each regional census center to customize a
           location-specific training module for local census offices.
           Nonetheless, developing modules for different types of locations
           centrally would allow the Bureau to control the consistency and
           quality of training throughout the nation.
		   
		   Bureau Is Designing Decennial Activities in the Geographic Area
		   Affected by Hurricanes Katrina and Rita

           As part of our evaluation of the Bureau's LUCA dress rehearsal, we
           visited the localities along the Gulf Coast to assess the effect
           the devastation caused by Hurricanes Katrina and Rita might have
           on LUCA and possibly other operations. The effects of Hurricanes
           Katrina and Rita are still visible throughout the Gulf Coast
           region. Hurricane Katrina alone destroyed or made uninhabitable an
           estimated 300,000 homes; in New Orleans, local officials reported
           that Hurricane Katrina damaged an estimated 123,000 housing units.
           Such changes in housing unit stock continue to present challenges
           to the implementation of the 2010 LUCA Program in the Gulf Coast
           region and possibly other operations. Many officials of local
           governments we visited in hurricane-affected areas said they have
           identified numerous housing units that have been or will be
           demolished as a result of hurricanes Katrina and Rita and
           subsequent deterioration. Conversely, many local governments
           estimate that there is new development of housing units in their
           respective jurisdictions. The localities we interviewed in the
           Gulf Coast region indicated that such changes in the housing stock
           of their jurisdictions are unlikely to subside before local
           governments begin reviewing and updating materials for the
           Bureau's 2010 LUCA Program--in August 2007.^15 Local government
           officials told us that changes in housing unit stock are often
           caused by difficulties families have in deciding whether to return
           to hurricane-affected areas. Local officials informed us that a
           family's decision to return is affected by various factors, such
           as the availability of insurance; timing of funding from
           Louisiana's "Road Home" program;^16 lack of availability of
           contractors; school systems that are closed; and lack of amenities
           such as grocery stores. As a result of the still changing housing
           unit stock, local governments in hurricane-affected areas may be
           unable to fully capture reliable information about their address
           lists before the beginning of LUCA this year or address canvassing
           in April 2009. Furthermore, operation of local governments
           themselves has been affected by the hurricanes (see fig. 3). These
           local governments are focused on reconstruction and at least two
           localities we spoke to questioned their ability to participate in
           LUCA.
		   
^14 [18]GAO-05-9 .

^15The period for local review and update of addresses and maps for the
2010 LUCA Program is August 2007-March 2008.

^16The "Road Home" Program was implemented by the State of Louisiana to
provide compensation of up to $150,000 for eligible homeowners affected by
hurricanes Katrina and Rita.

Figure 3: City Halls in Mississippi and Louisiana Have Been Destroyed and
City Officials Now Operate Out of Trailers

Note: Pictures are from January 2007.

The mixed condition of the housing stock in the Gulf Coast will increase
the Bureau's address canvassing workload. During our field work, we found
that hurricane-affected areas have many neighborhoods with abandoned and
vacant properties mixed in with occupied housing units. Bureau staff
conducting address canvassing in these areas may have an increased
workload due to the additional time necessary to distinguish between
abandoned, vacant and occupied housing units. Another potential issue is
that due to continuing changes in the condition in the housing stock,
housing units that are deemed vacant or abandoned during address
canvassing may be occupied on Census Day (Apr. 1, 2010). Bureau officials
said that they recognize there are issues with uninhabitable structures in
hurricane-affected zones. They noted that addresses marked as vacant or
uninhabitable during address canvassing in the Gulf Coast region will not
be deleted from the MAF, and said that they may adjust training for Bureau
staff in hurricane-affected areas.

Workforce shortages may also pose significant problems for the Bureau's
hiring efforts for address canvassing. The effects of hurricanes Katrina
and Rita caused a major shift in population away from the
hurricane-affected areas, especially in Louisiana. This migration
displaced many low-wage workers. Should this continue, it could affect the
availability of such workers for address canvassing and other decennial
census operations. Bureau officials recognize the potential difficulty of
attracting these workers, and have recommended that the Bureau be prepared
to meet hourly wage rates for future decennial staff that are considerably
higher than usual. It has noted that its Dallas regional office, which has
jurisdiction over hurricane-affected areas in Texas, Louisiana, and
Mississippi, will examine local unemployment rates to adjust pay rates in
the region, and use "every single entity" available to advertise for
workers in the New Orleans area.

Early in 2006, we recommended that the Bureau develop plans (prior to the
start of the 2010 LUCA Program in August 2007) to assess whether new
procedures, additional resources, or local partnerships, may be required
to update the MAF/TIGER database along the Gulf Coast--in the areas
affected by hurricanes Katrina and Rita.^17 The Bureau responded to our
recommendations by chartering a team to assess the effect of the storm
damage on the Bureau's address list and maps for areas along the Gulf
Coast and develop strategies with the potential to mitigate these effects.
The chartered team recommended that the Bureau consult with state and
regional officials (from the Gulf Coast) on how to make LUCA as successful
as possible, and hold special LUCA workshops for geographic areas
identified by the Bureau as needing additional assistance. While the
Bureau (through its chartered team, headquarters staff and Dallas regional
office) has proposed several changes to the 2010 LUCA Program for the Gulf
Coast region, there are no specific plans for implementing the proposed
changes.

In summary, Mr. Chairman, we recognize the Bureau faces formidable
challenges in successfully implementing a redesigned decennial census. It
must also overcome significant challenges of a demographic and
socioeconomic nature due to the nation's increasing diversity in language,
ethnicity, households, and housing type, as well as an increase in the
reluctance of the population to participate in the census. The need to
enumerate in the areas devastated by hurricanes Katrina and Rita is one
more significant difficulty the Bureau faces. We applaud the moves the
Bureau has undertaken to redesign the census; we have stated in the past,
and believe still, that the reengineering, if successful, can help control
costs and improve cost effectiveness and efficiency. Yet, there is more
that the Bureau can do in examining and refining its recruiting, hiring,
and training practices and in preparing to enumerate in the
hurricane-affected areas. Also, the functionality and usability of the
MCD--a key piece of hardware in the reengineered census---bears watching
as does the oversight and management of information technology
investments. All told, these areas continue to call for risk mitigation
plans by the Bureau and careful monitoring and oversight by the Commerce
Department, the Office of Management and Budget, the Congress, GAO, and
other key stakeholders. As in the past, we look forward to supporting this
subcommittee's oversight efforts to promote a timely, complete, accurate,
and cost-effective census.

^17GAO, 2010 Census: Census Bureau Needs to Take Prompt Actions to Resolve
Long-standing and Emerging Address and Mapping Challenges, [19]GAO-06-272
(Wash. D.C.: June 15, 2006), and GAO, 2010 Census: Costs and Risks Must be
Closely Monitored and Evaluated with Mitigation Plans in Place,
[20]GAO-06-822T (Wash. D.C.: June 6, 2006).

Contact and Acknowledgments

For further information regarding this testimony, please contact Mathew J.
Scire on (202) 512-6806, or by email at [email protected].

Individuals making contributions to this testimony included Betty Clark,
Carlos Hazera, Shirley Hwang, Andrea Levine, Lisa Pearson, Mark Ryan, Niti
Tandon, and Timothy Wexler.

GAO Products Appendix I: Related GAO Products

2010 Census: Redesigned Approach Holds Promise, but Census Bureau Needs to
Annually Develop and Provide a Comprehensive Project Plan to Monitor
Costs. [21]GAO-06-1009T . Washington, D.C.: July 27, 2006.

2010 Census: Census Bureau Needs to Take Prompt Actions to Resolve
Long-standing and Emerging Address and Mapping Challenges. [22]GAO-06-272
. Washington, D.C.: June 15, 2006.

2010 Census: Costs and Risks Must be Closely Monitored and Evaluated with
Mitigation Plans in Place. [23]GAO-06-822T . Washington, D.C.: June 6,
2006.

2010 Census: Census Bureau Generally Follows Selected Leading Acquisition
Planning Practices, but Continued Management Attention Is Needed to Help
Ensure Success. [24]GAO-06-277 . Washington, D.C.: May 18, 2006.

Census Bureau: Important Activities for Improving Management of Key 2010
Decennial Acquisitions Remain to be Done. [25]GAO-06-444T . Washington,
D.C.: March 1, 2006.

2010 Census: Planning and Testing Activities Are Making Progress.
[26]GAO-06-465T . Washington D.C.: March 1, 2006.

Information Technology Management: Census Bureau Has Implemented Many Key
Practices, but Additional Actions Are Needed. [27]GAO-05-661 . Washington,
D.C.: June 16, 2005.

2010 Census: Basic Design Has Potential, but Remaining Challenges Need
Prompt Resolution. [28]GAO-05-09 . Washington, D.C.: January 12, 2005.

Data Quality: Census Bureau Needs to Accelerate Efforts to Develop and
Implement Data Quality Review Standards. [29]GAO-05-86 . Washington, D.C.:
November 17, 2004.

Census 2000: Design Choices Contributed to Inaccuracies in Coverage
Evaluation Estimates. [30]GAO-05-71 . Washington, D.C.: November 12, 2004.

American Community Survey: Key Unresolved Issues. [31]GAO-05-82 .
Washington, D.C.: October 8, 2004.

2010 Census: Counting Americans Overseas as Part of the Decennial Census
Would Not Be Cost-Effective. [32]GAO-04-898 . Washington, D.C.: August 19,
2004.

2010 Census: Overseas Enumeration Test Raises Need for Clear Policy
Direction. [33]GAO-04-470 . Washington, D.C.: May 21, 2004.

2010 Census: Cost and Design Issues Need to Be Addressed Soon.
[34]GAO-04-37 . Washington, D.C.: January 15, 2004.

Decennial Census: Lessons Learned for Locating and Counting Migrant and
Seasonal Farm Workers. [35]GAO-03-605 . Washington, D.C.: July 3, 2003.

Decennial Census: Methods for Collecting and Reporting Hispanic Subgroup
Data Need Refinement. [36]GAO-03-228 . Washington, D.C.: January 17, 2003.

Decennial Census: Methods for Collecting and Reporting Data on the
Homeless and Others Without Conventional Housing Need Refinement.
[37]GAO-03-227 . Washington, D.C.: January 17, 2003.

2000 Census: Lessons Learned for Planning a More Cost-Effective 2010
Census. [38]GAO-03-40 . Washington, D.C.: October 31, 2002.

The American Community Survey: Accuracy and Timeliness Issues.
[39]GAO-02-956R . Washington, D.C.: September 30, 2002.

(450589)

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. However, 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.

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 ( 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
www.gao.gov and select "Subscribe to 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: www.gao.gov/fraudnet/fraudnet.htm
E-mail: [email protected]
Automated answering system: (800) 424-5454 or (202) 512-7470

Congressional Relations

Gloria Jarmon, Managing Director, [email protected] (202) 512-4400 U.S.
Government Accountability Office, 441 G Street NW, Room 7125 Washington,
D.C. 20548

Public Affairs

Paul Anderson, Managing Director, [email protected] (202) 512-4800
U.S. Government Accountability Office, 441 G Street NW, Room 7149
Washington, D.C. 20548

www.gao.gov/cgi-bin/getrpt?GAO-07-779T .

To view the full product, including the scope
and methodology, click on the link above.

For more information, contact Mathew J. Scire at (202) 512-6806 or
[email protected].

Highlights of [47]GAO-07-779T , testimony before the Subcommittee on
Information Policy, Census, and National Archives, Committee on Oversight
and Government Reform, House of Representatives

April 24, 2007

2010 CENSUS

Design Shows Progress, but Managing Technology Acquisitions, Temporary
Field Staff, and Gulf Region Enumeration Require Attention

The decennial census is a Constitutionally-mandated activity that produces
data used to apportion congressional seats, redraw congressional
districts, and allocate billions of dollars in federal assistance. The
Census Bureau (Bureau) estimates the 2010 Census will cost $11.3 billion,
making it the most expensive in the nation's history. This testimony
discusses the Bureau's progress in preparing for the 2010 Census to (1)
implement operations to increase the response rate and control costs; (2)
use technology to increase productivity; (3) hire and train temporary
staff; and (4) plan an accurate census in areas affected by hurricanes
Katrina and Rita.

The testimony is based on previously issued GAO reports and work nearing
completion in which GAO observed recruiting, hiring, and training
practices in the 2006 test, and visited localities that participated in
the Local Update of Addresses Dress Rehearsal as well in the Gulf Coast
region.

[48]What GAO Recommends

At this time, GAO is not making new recommendations, but past reports
recommended steps for the Bureau to enhance the mobile computing devices
and promote an accurate census in areas impacted by hurricanes Katrina and
Rita. The Bureau generally agreed with these recommendations and has acted
to implement some of them.

The Bureau has made progress towards implementing a re-engineered census
design that holds promise for increasing the response rate, thereby
controlling the cost of the census while promoting accurate results. The
re-engineered design includes a short form only census designed to
increase the response rate by about 1 percent and a targeted second
mailing, which is expected to increase response by between 7 to 10
percent. Both of these initiatives are new, have been tested, and will be
a part of the 2010 Census design. According to Bureau officials, a 1
percent increase in the response rate can save $75 million, making these
initiatives critical to the new design.

Uncertainty surrounds a keystone to the reengineered census, the mobile
computing device (MCD). The MCD allows the Bureau to automate operations
and eliminate the need to print millions of paper questionnaires and maps
used by census workers to conduct census operations and to assist in
managing payroll. The MCD, tested in the 2004 and 2006 census tests, was
found to be unreliable. While a contractor has developed a new version of
the MCD, the device will not be field tested until next month, leaving
little time to correct problems that might emerge during the 2008 Dress
Rehearsal.

Timeline of Selected Key Decennial Events

Source: GAO summary of Census Bureau data.

The Bureau faces challenges in recruiting, hiring, and training an
estimated 600,000 temporary employees. For example, opportunities exist
for the Bureau to hone its recruiting efforts to identify individuals who
would be more likely to be effective at census work and willing to work
throughout an operation. Also, census workers indicated a need for
additional training on reluctant respondents as well as location-specific
challenges they encounter. The Bureau must also be prepared to accurately
count the population affected by hurricanes Katrina and Rita. The Bureau
has contacted local officials in the Gulf Area and is developing a plan
that includes workshops and special staffing considerations.

References

Visible links
  11. http://www.gao.gov/cgi-bin/getrpt?GAO-05-9
  12. http://www.gao.gov/cgi-bin/getrpt?GAO/GGD-00-6
  13. http://www.gao.gov/cgi-bin/getrpt?GAO-01-579
  14. http://www.gao.gov/cgi-bin/getrpt?GAO/GGD-00-06
  15. http://www.gao.gov/cgi-bin/getrpt?GAO-05-9
  16. http://www.gao.gov/cgi-bin/getrpt?GAO-06-444T
  17. http://www.gao.gov/cgi-bin/getrpt?GAO-05-9
  18. http://www.gao.gov/cgi-bin/getrpt?GAO-05-9
  19. http://www.gao.gov/cgi-bin/getrpt?GAO-06-272
  20. http://www.gao.gov/cgi-bin/getrpt?GAO-06-822T
  21. http://www.gao.gov/cgi-bin/getrpt?GAO-06-1009T
  22. http://www.gao.gov/cgi-bin/getrpt?GAO-06-272
  23. http://www.gao.gov/cgi-bin/getrpt?GAO-06-822T
  24. http://www.gao.gov/cgi-bin/getrpt?GAO-06-277
  25. http://www.gao.gov/cgi-bin/getrpt?GAO-06-444T
  26. http://www.gao.gov/cgi-bin/getrpt?GAO-06-465T
  27. http://www.gao.gov/cgi-bin/getrpt?GAO-05-661
  28. http://www.gao.gov/cgi-bin/getrpt?GAO-05-09
  29. http://www.gao.gov/cgi-bin/getrpt?GAO-05-86
  30. http://www.gao.gov/cgi-bin/getrpt?GAO-05-71
  31. http://www.gao.gov/cgi-bin/getrpt?GAO-05-82
  32. http://www.gao.gov/cgi-bin/getrpt?GAO-04-898
  33. http://www.gao.gov/cgi-bin/getrpt?GAO-04-470
  34. http://www.gao.gov/cgi-bin/getrpt?GAO-04-37
  35. http://www.gao.gov/cgi-bin/getrpt?GAO-03-605
  36. http://www.gao.gov/cgi-bin/getrpt?GAO-03-228
  37. http://www.gao.gov/cgi-bin/getrpt?GAO-03-227
  38. http://www.gao.gov/cgi-bin/getrpt?GAO-03-40
  39. http://www.gao.gov/cgi-bin/getrpt?GAO-02-956R
  47. http://www.gao.gov/cgi-bin/getrpt?GAO-07-779T
*** End of document. ***