PREPARED STATEMENT OF

Mr. Chairman, Mrs. Maloney, and Members of the Subcommittee:

Once again, I am pleased to be here to report on the status of Census 2000 activities. The Census Bureau has now completed well over 99 percent of the nonresponse followup workload nationally. We have completed nonresponse followup in 437 of the 520 local census offices and have completed 90 percent or more of the workload in all but two offices. I believe we will complete nonresponse followup everywhere in time to meet the schedule for subsequent operations.

At a press briefing 2 weeks ago, I called Census 2000 "the Good Census," and I believe that accurately describes what the Census Bureau has achieved. When you initiated this series of operational hearings back in early February, just a little over 4 months ago, we were cautiously optimistic because we knew that planning and early operations were on schedule. We had done everything up to that time that needed to be done. But the road ahead was full of obstacles and of potential problems that could put the census at risk. Here are some of them and the Census Bureau's successes in meeting them:

Hiring was a challenge. But the Census Bureau was successful in recruiting over 2.6 million qualified applicants and was able to hire and retain sufficient staff to complete every field operation thus far.

Paying staff on time was a challenge. But there have been no major problems with our payroll systems and during nonresponse followup the Census Bureau issued an average of nearly half a million paychecks each week.

Completing early field operations on schedule was a challenge. But the Census Bureau completed on or near schedule the enumeration of remote areas of Alaska, update/leave, update/enumerate, list/enumerate, group quarters enumeration, and service based enumeration.

There was a risk of an unexpectedly low mail response rate. On the contrary, the country slightly exceeded the mail response rate for the 1990 census, exceeded the target in our budget, and reversed a 30-year decline in this area. This is a serious achievement.

There were concerns about how well the Census Bureau's data capture systems would work. The Census Bureau's data capture systems have performed better than expected. Our four data capture centers have received and scanned over 120 million questionnaires; and the accuracy rate for optical character recognition, used to capture handwritten responses, is over 99 percent.

There was a question about whether the Census Bureau's programs to provide questionnaire assistance and multiple response options would work. They did. We provided assistance to about 6 million callers over the telephone, more than 546,000 people at walk-in questionnaire assistance centers, and received and processed more than a million foreign-language questionnaires. We received approximately 580,000 Be Counted forms and about 66,000 responses over the Internet.

As we entered the nonresponse followup operation about 2 months ago, the Census Bureau was faced with its most serious operational challenge yet and risks that could have affected data accuracy, data quality, and budget. Would we have enough staff and would they be highly productive? Would the public cooperate or would there be resistance? The great success the Census Bureau has had in the nonresponse followup operation is due to the dedication, enthusiasm, and resourcefulness of the census workers. They have taken their jobs seriously and worked heroically to help their communities have the best count possible, braving tough neighborhoods and, in a very few instances, tragic circumstances.

I believe the success of the nonresponse followup operation is also due to a residual effect from our extensive advertising campaign and the efforts of our more than 100,000 census partners. As a result, census workers experienced limited outright hostility or resistance, although there were isolated instances and some attempts to organize resistance. We continued to advertise right through the nonresponse followup operation and stepped up efforts in some localities that were

lagging. We also continued to receive tremendous support from concerned mayors and other local officials who took special steps to encourage cooperation with the enumerators.

The repertoire of tools for hard-to-enumerate areas made available to the local census offices also help in completing nonresponse followup so successfully. I testified on these tools at a hearing in May, so I will just summarize here. Using a variety of demographic and operational variables from the 1990 census, Census Bureau staff identified 1990 census tracts that were hard to enumerate, difficult to recruit in, or had some other special situation that would require the application of special enumeration tools. Because the data in our planning database were nearly 10 years old, experienced field staff made the final determination of which tracts would pose problems and, more importantly, what tools would be needed to overcome them. The tools available for use included establishing Be Counted and Questionnaire Assistance Center sites, providing bilingual enumerators and/or local cultural facilitators, using special enumeration procedures, such as update/enumerate, and having enumerators work in pairs or larger teams to conduct their work.

Here are some specific examples of applying special tools to hard-to-enumerate areas in the Denver and Philadelphia regions:

The Las Vegas, Nevada, local census office identified a very hard-to-enumerate census tract--low 1990 response rate, low income, lots of subsidized rental housing, many shelters and soup kitchens, and over 80 percent of the households spoke Spanish as their primary language in 1990. The local census office formed an entirely Spanish-speaking crew of enumerators, paired them into teams, and moved them about from one area to another as necessary. They achieved outstanding results. They completed their task in record time, with no significant incidents, and with an extremely low rate of refusals or partial interviews.

The Denver, Colorado, local census office worked very closely with the Denver Public Housing authority to gain access to and find sufficient local enumerators to conduct the count in public housing complexes. These special efforts included obtaining testing and training space at the complexes and special efforts to hire sufficient bilingual enumerators. As a result of these efforts, the enumeration of the public housing areas was substantially improved.

The Great Falls, Montana, local census office used the hard-to-enumerate planning process to identify more than two dozen Hutterite colonies within the office area that would need special attention. Hutterite colonies are religious communal groups that are closed to most outside contact. Most have a "boss" who maintains contact with the outside. The Great Falls office acquired a directory of names and phone numbers for the Hutterite "bosses," contacted them, and worked with them to establish ground rules for how the enumeration could be conducted. The "bosses" were sworn in and acted as cultural facilitators, in most cases accompanying the enumerators door to door on the colonies to complete the enumeration. Through advance planning and meticulous follow-through, the enumeration of the colonies was successful.

The Las Cruces, New Mexico, local census office developed a Spanish-language Fotonovela, an 8-page comic book featuring real celebrities, to discuss the confidentiality of the data and the importance of answering the census. In partnership with the Catholic Church and the Las Colonias Development Council, staff distributed this promotion tool throughout the heavily Spanish-speaking colonias near the U.S./Mexico border and other areas. These areas had been identified through the hard-to-enumerate planning process.

One of the District of Columbia offices worked with property managers, advisory neighborhood commissions, appropriate members of census complete count committees and other civic leaders, and "orange hat safety patrols" to build support in certain high crime areas for a blitz enumeration. The blitz technique, which involves having a large number of enumerators work in a specific area, increases enumerator safety, productivity, and visibility in the community. In this case, the blitz enumeration technique was an effective way to get better cooperation and complete the enumeration in less time than it normally would have taken.

The Baltimore local census offices made extensive use of language and cultural facilitators to enhance the enumeration of certain areas. Language facilitators were used in areas that were designated as having high concentrations of Spanish- or Russian-speaking individuals. Cultural facilitators, who are individuals who are trusted in their communities, were used to reduce barriers that would have prevented enumerators from obtaining a successful interview.

Thanks to the efforts of our enumerators and partners and the early work the Census Bureau did to encourage support for the census and to identify special tools for hard-to-enumerate areas, nonresponse followup has been a great success.

The Census Bureau has already begun to document what worked well and suggestions for future improvements. We have debriefed recruiting staff and partnership specialists and will conduct additional debriefings over the next several weeks. This is standard procedure at the Census Bureau so that we can learn from our experiences and build in improvements for the next census.

While we are pleased with progress thus far, there remain several operations that will improve what is already a good census. I have said numerous times that any national statistic--including the census count--is an estimate of the truth. The challenge is to get that estimate as close to the truth as is possible. That is, there is a "true" count of the resident population of the United States on April 1, 2000. Were we to conclude the census with the completion of nonresponse followup, we would provide an estimate of that true count. It is unlikely that that estimate would be absolutely accurate, that is, identical with the true count. We believe the estimate can be moved closer to the truth.

Census 2000 will improve the census estimate with three other operations. The coverage edit followup does so by reconciling population count discrepancies. The coverage improvement followup operation does so by enumerating housing units added to the address list too late to have been included in the initial nonresponse followup operation. And the Accuracy and Coverage Evaluation (A.C.E.) does so by dual system estimation in a procedure that measures the number of persons missed and the number erroneously included in any of the prior census operations.

Coverage Edit Followup.--The coverage edit followup operation, which began in early May and will be completed in late July, is an important coverage check being conducted by telephone from 13 calling centers around the country. The trained telephone agents are calling two types of households for which we have the telephone numbers. The first are those households for which responses were received by mail or Internet and there are what we call "population count discrepancies." A "population count discrepancy" occurs when there is a difference between the number of persons in the household that the respondent reported in Question 1 ("How many people were living or staying in this house, apartment, or mobile home on April 1, 2000?") and the number of persons for which data were reported. We are also using this process to follow up on households with more than six people. The census questionnaire only has room to report data for six people, so it is important that in this operation we collect data for the additional members of these large households, so their characteristics can be included in the census.

Coverage Improvement Followup Operation.--In 342 of the local census offices where we have completed nonresponse followup--this is the number that made a cutoff date of June 15--we are now beginning the coverage improvement followup operation. Selection and training for supervisors is going on on a flow basis now, and enumerators will be trained and in the field the first week of July. We are conducting this operation in waves. As additional offices complete nonresponse followup, they will enter the next wave, which will be in the field 2 weeks after the first wave. A third wave is planned, if necessary.

The coverage improvement followup operation can be thought of as a final, essential cleanup operation that rechecks some of the enumerators' earlier work and enumerates some housing units that were added to our address list too late to be included in earlier operations. For example, enumerators will recheck some housing units that were identified as vacant or that were deleted during nonresponse followup. Census Bureau experience from previous censuses indicates that some enumerators erroneously classify occupied housing units as vacant or delete them, so this is an important step to assure a full and accurate census. Enumerators will visit for the first time housing units that were added in our "new construction" program, which was a partnership effort with local governments conducted in the spring of this year. And they will enumerate some housing units added during update/leave or the appeals process for the local update of census addresses, as well as households that submitted blank forms.

The coverage improvement field work is scheduled to last 3 weeks in each wave. Procedures and quality controls are very similar to those for the nonresponse followup operation, including the requirement that enumerators make up to six attempts to find someone at a housing unit that appears to be occupied.

Personal Visit Interviewing for the Accuracy and Coverage Evaluation.--As I mentioned at the May 19 hearing before this Subcommittee, to provide sufficient data to compare the A.C.E. to the initial census, the Census Bureau must conduct interviews to collect data from each of the housing units that were independently listed in the A.C.E. listing operation. We initiated the interviewing with a telephone phase in late April and completed about 90,000 interviews by telephone, or about 29 percent of the total workload of about 314,000 cases. We did not start personal visit interviewing during nonresponse followup to preserve independence between the A.C.E. and the initial census enumeration.

Personal visit interviewing is now beginning in those local census offices that have completed nonresponse followup. Personal interviews are conducted only with a household member during the first 3 weeks that the case is available for interviewing. If an interview with a household member is not obtained after 3 weeks, interviewers will attempt to interview another knowledgeable person. During the latter part of the operation, the best interviewers are used to convert as many noninterview cases as possible to completed interviews, either by talking to a household member or another knowledgeable person. This nonresponse conversion has been planned to improve the completeness of data for matching. We expect to complete personal visit interviewing in early September.

Although we are now moving into other field operations in local census offices that have completed nonresponse followup, we are committed to fully applying our procedures to account for every remaining address in the local census offices that have not yet finished. Those procedures, as I have testified previously, are extensive and include making up to six attempts to complete the enumeration of each housing unit. The procedures also include extensive quality assurance procedures and supervisory controls.

Daily production levels begin to decrease toward the end of nonresponse followup. Sometimes enumerators completed the easier cases first, finished the work closest to their homes first, or believed that the quicker they finished their assignment, the sooner they would be out of work. In order to bring the operation to closure within the scheduled 10 weeks, we look at areas within each local census office that are lagging behind. We also implement a procedure known as "final attempt." When the area covered by a crew leader has completed 95 percent of its workload, "final attempt" begins and the crew leader consolidates the remaining work and gives it to the most productive and dependable enumerators. When an entire local census office reaches an average 95 percent rate of completion, the Regional Director instructs the office to begin "final attempt" within 2 days. During "final attempt" enumerators then make one final visit to each outstanding address that has already been visited at least two times and to some of the housing units for which only minimal data was earlier collected to complete as much of the questionnaire as possible. If an address has only been visited once, an enumerator will make up to two additional visits during "final attempt." The intent of "final attempt" is to resolve all outstanding cases within a few days, but nonresponse followup is not over until a questionnaire is completed and checked into the local census office for every unit.

You asked in your letter of invitation about any serious problems in any of our offices. I will be happy to answer any specific questions you may have about that.

I would like to return now to the earlier discussion about "the Good Census." A good census involves three elements. The examples of successes I mentioned earlier relate to two of these, namely operational robustness and a dedication to including everyone in the count. The third element of a good census involves openness, transparency, and public scrutiny, which in turn can lead to public trust in the process. This has been the most open and transparent census in history; every detail has been and is being scrutinized. We welcome that scrutiny because we believe it will dispel any notion that there is or could be any political manipulation of the final results.

At a press conference last week, in the spirit of openness, I made public a document entitled "Accuracy and Coverage Evaluation: Statement on the Feasibility of Using Statistical Methods to Improve the Accuracy of Census 2000." This statement was prepared after extensive discussions with the Census Bureau's senior staff and after review of all relevant documents. The document sets forth the rationale for the Census Bureau's preliminary determination that (1) statistically corrected census data can be produced within the time frame required by law and (2) statistically corrected data will be more accurate.

At the same press conference, Commerce Department General Counsel Andrew Pincus described the Secretary of Commerce's proposed regulations that would delegate to the Director of the Census Bureau the authority to make the final decision on whether to use statistical sampling to correct the census counts (published June 20, 2000 at 65 Federal Register 38370). This proposal will help make certain that the final decision is fully informed by the statistical experts at the Census Bureau. The final decision will follow a review of the A.C.E. A committee of distinguished senior career professionals will study the results of the A.C.E. and make its recommendation to the Director. This recommendation will be made public. After considering the recommendation, the Director will make the final decision.

Mr. Chairman, I will now answer any questions.