United !States General Accounting Office /LJ/7U ‘I, - Testimony lllllllllllllllI 141717 For Release 1990 Census: Status of Questionnaire on Delivery Follow-Up Efforts Expected at 9:30 a.m. CDT Monday July 2, 1990 Statement of L. Nye Stevens, Director Government Business Operations Issues Before the Subcommittee on Census and Population Committee on Post Office and Civil Service House of Representatives GAO/T-GGD-90-52 GAO Form 160 (12/87) . \ 1990 CENSUS: STATUS OF QUESTIONNAIRE FOLLO~,~UP'EFFORTS .8. >. SUMMARYOF STATEMENT OF < ;/ L. NYE STEVENS DIRECTOR, GOVERBMEBTBUSINESS OPERATIQNS ISSUES GAO's May 21 testimony before the Subcommittee focused on two major sources of concern for the successful completion of census follow-up efforts: (1) incomplete cost and progress reporting was hampering the Census Bureau's efforts to monitor and manage the census and (2) the Bureau did not have sufficient field staff. Follow-up efforts appeared to be lagging significantly in many of the Bureau's 447 district offices and census progress could be seriously impaired if the Bureau did not take aggressive action to correct these deficiencies. Today, GAO reports that since the May hearing the Bureau has made major improvements that appear to have put the census generally back on schedule. First, the Bureau successfully addressed problems in its management information system: district offices are reporting key progress data, and the software problems GAO found generally have been resolved. The Bureau also has taken action to address its staff shortages. For example, the Bureau increased pay rates for enumerators and other field staff in about 31 percent of its offices and expanded its incentive pay program nationwide. In addition, the Bureau has taken other actions, such as a second mailing to over 350,000 households in New York, to further assist with the completion of follow-up efforts. In anticipation that many offices would not complete the follow- up efforts on schedule, the Bureau's planning assumptions contained about a 3-week period between the scheduled end of the activity, June 6, and the beginning of subsequent operations. Most of the Bureau's district offices completed follow-up efforts within this 3-week grace period. Overali, about 90 percent of the Bureau's offices are projected to have completed follow-up efforts by the end of June or the first few days of July. Despite the progress the Bureau has made since the Subcommittee's hearing in May, major challenges still remain before the census will be completed. Follow-up work in about 10 percent of the offices will continue beyond the 3-week extension. The Bureau's New York and San Francisco regions do not expect to complete follow-up efforts until July 15 and July 7, respectively. The timely completion of follow-up is important to allow the Bureau's Post Enumeration Survey, which is critical for the question of adjusting census counts, to be completed on time. The Burwu also needs time to review the extent to which its address list is complete to ensure an accurate enumeration of the Nation's population. Mr. Chairman and Members of the Subcommittee: We are pleased to be here today to discuss the status of the 1990 decennial census, particularly the Census Bureau's nonresponse follow-up operation, which seeks to obtain completed questionnaires from households that did not initially respond to the census. My comments are based on our ongoing effort, as requested by the Subcommittee, to monitor census operations at Bureau headquarters and in the field. Our last testimony before this Subcommittee was in New York on May 21, about mid-way through the nonresponse follow-up 0peration.l We identified two major sources of concern for the successful completion of nonresponse follow-up: the status of the Bureau's management information system (MIS) and enumerator staffing. We reported that MIS problems including incomplete reporting were hampering the Bureau's efforts to monitor and manage the census. In addition, we found that at least 20 percent of the offices were understaffed in 9 of the Bureau's 13 regions. In the Dallas region, which covers Texas, Louisiana and Mississippi, about 74 percent of the region's district offices were understaffed early in nonresponse follow-up. Moreover, we reported that the Bureau's staffing statistics did not differentiate between part-time and full-time staff and l-progress of the 1990 Decennial Census: Some Causes for Concern (GAO/T-GGD-90-44, May 21, 1990). 1 therefore masked the full extent of the Bureau's staffing problems in certain locales, including New York. Nonresponse efforts appeared to be lagging significantly in many offices and the progress of the census may have been seriously impaired if the Bureau did not take aggressive action to correct these deficiencies. ACTIONS TAKEN TO IMPROVE NONRESPONSEFOLLOW-UP Mr. Chairman, I am pleased to report today that since the May hearing the Bureau has made a number of major improvements that appear to have generally put the census back on schedule. First, the Bureau has successfully addressed problems in its management information system and district offices are reporting key progress data. Census managers in headquarters and in the field now are using MIS reports to monitor and manage the census, especially where it takes place--in the field. For example, on the basis of MI@ data and other information, the Bureau moved a total of 33 managers from headquarters, and in some cases other regions, into selected New York and San Francisco district offices to provide technical guidance and support to speed the completion of follow-up efforts in these regions where problems were particularly acute. According to the Bureau, these managers are reviewing cost and progress reports for their assigned offices to identify problem 2 . areas and the necessary corrective actions to resolve the problems quickly. Changes to the Census Pay Program The Bureau also has taken action to address its staff shortages. For example, the Bureau modified its pay program. We noted earlier this year that the Bureau's decision to implement geographic wage rates was an important achievement; we also observed, however, that the original geographic pay rates may not have been competitive and could need to be raised during the censuso2 On June 3, the Bureau increased pay in 140, or about 31 percent, of its district offices to accelerate the completion of nonresponse follow-up and meet the Post Enumeration Survey (PES)3 schedule by attracting new employees and motivating the existing workforce to continue working. About 56 percent of the offices that received a pay increase are located either in California or New York. None of the offices 2The Decennial Census: Potential Risks to Data Quality Resulting From Budqet Reductions and Cost Increases (GAO/T-GGD-90-30, Mar. 27, 1990); 1990 Census: Costs Are Uncertain Because Wage Rates May Be Uncompetitive (GAO/GGD-90-78, May 1990). 3The PES is a matching study in which the Bureau interviews a sample of households independent of the census. The persons enumerated in the PES are matched to census records to determine whether each person was correctly counted or missed in the actual census. See Critical Issues for Census Adjustment: Completing Post Enumeration Survey on Time Wh-1 P t t‘ g (GAO/T-GGD-90-15, Jan. 30, 1990), forlaedi~~u~~i~~ oyatthae 3 that received the pay increase are in Texas, because of the progress made in that state in completing nonresponse follow-up. In those areas where pay was increased, pay rates for enumerators and other field staff were raised by between $.50 and $2.00 per hour. Enumerators in New York City, where the Bureau is offering its highest enumerator wages, now receive a base pay of $10.00 per hour. Regional management staff in New York and San Francisco, the two regions with the lowest nonresponse follow-up completion rates, believe the increased pay levels attracted additional staff and are assisting them in completing follow-up efforts. The Bureau also expanded its supplemental pay program for field staff in district offices nationwide. For example, under the revised program, in addition to the hourly wage, enumerators will be eligible to receive $1.50 for each census case they complete. The Bureau believes the revised supplemental pay is necessary to help retain and motivate its most productive employees to work the additional weeks until nonresponse follow-up is completed. Local Actions to Complete Work In addition to the changes to its pay program, the Bureau’s regional and district off ices also have acted to overcome intractable staff shortages. For example, the Bureau has moved field staff from neighboring offices into offices that are 4 unable to attract sufficient staff from within their own areas. In Texas, about 60 to 70 field staff from the Arlington and Northeast Dallas district offices assisted with completing the enumeration in the Central Dallas office. Likewise, staff from other district offices assisted with the enumeration in the Austin and Denton district offices--two other areas where attracting staff to work full time was very difficult. Another initiative to complete work expeditiously, which according to the Bureau was particularly effective, was implemented in Austin as well as other places. The Bureau used between 175 and 200 of its most productive enumerators in the Austin area to complete about 2,000 cases in a single day in a community that had proven hard to enumerate. A regional official said the intensive effort, with the active involvement of the local member of Congress, was very successful to increasing the willingness of enumerators to work in the area and, more important, to encouraging residents to answer the census. Follow-Up Mailing in New York The Bureau also remailed a census questionnaire to all households in 15 of the New York region's 28 district offices from which it had not been able to obtain a completed census form. These offices were selected because they had at least 20 percent of 5 their nonresponse workload outstanding as of June 21, and many of their uncompleted cases were in multi-unit buildings where the Bureau has had difficulty obtaining access to gather census information. The Bureau mailed over 350,000 new questionnaires to New York households over the period June 25 to 27. Follow-up mailings have proven to be a cost-effective means of increasing response rates in other surveys. We recommended in 1982 that the Department of Commerce test the feasibility of using follow-up mailings to reduce the need for personal follow- up visits during the 1990 census.4 However, the second mailing in New York is the first time the Bureau has attempted such a mailing during a decennial census activity, according to Bureau officials. We believe that by evaluating the effectiveness of follow-up mailings during the test cycle preceding the 1990 census, the Bureau would have been better prepared for a follow- up mailing in New York. According to the Associate Director for Decennial Programs, the second mailing provides an important test for future census efforts. We agree and urge the Bureau to design an evaluation strategy to ensure that the results of the follow-up mailing in New York can be used for future decision-making. 4A $4 Billion Census in 19901 Timely Decisions on Alternatives to 1980 8rocedures Can Save Millions (GAO/GGD-82-13, Feb. 22, 1982). 6 MOST OFFICES HAVE COMPLETED NONRESPONSEFOLLOW-UP WITHIN THREE WEEKS OF SCHEDULE Nationally, of the Bureau's 447 district offices that had nonresponse follow-up work, most did not complete the field activity by the scheduled June 6 deadline. In anticipation that many offices would not complete the operation on schedule, the Bureau's planning assumptions contained about a 3-week period between the scheduled end of nonresponse follow-up and the beginning of PES interviewing and field follow-up, which includes a Bureau check of housing units identified as vacant or nonexistent during nonresponse follow-up. Nonresponse follow-up progress thus far demonstrates that the additional time was needed. For example, only six district offices, or 1.3 percent, finished on or before the scheduled completion date. However, since June 6, the Bureau reports that follow-up efforts have progressed significantly. For example, while none of Texas' I 31 district offices reported completing nonresponse follow-up on schedule, the Tyler and San Angelo offices reported finishing work on June 8 --2 days after the scheduled completion date--and all Texas offices reported completing nonresponse follow-up by June 28. Nat ionally, as of June 28, the most recent data available at the time we developed our statement, about 62 percent, or 275, of the 447 offices had completed nonresponse follow-up. An additional 134 offices were in the final stages of 7 the field activity and the Bureau expected that they would be completed within a week. Overall, therefore, about 91 percent of the Bureau's district offices should complete nonresponse follow- up by the end of June or the first few days of July. This would place the Bureau well ahead of its 1980 progress at this point, when about 56 percent of the Bureau's offices had completed follow-up within 4 weeks of the scheduled completion date. REMAINING CHALLENGES Despite the progress the Bureau has made since the Subcommittee's hearing in May, a number of major challenges still remain before the census will be completed. First, nonresponse follow-up is still incomplete in some key urban areas even though it is now over the three weeks since the scheduled end-date of the activity. In our April report on census recruiting, we noted that completing fieldwork expeditiously is a key ingredient to ensuring a high-quality census. Efficient fieldwork provides the Bureau 'lith time to review census counts and resolve apparent discrepancies. As of June 28, however, the New York region reported that it had completed 83 percent of its nonresponse follow-up workload; and San Francisco reported that it had completed 87 percent of its workload. According to Bureau officials in New York and San Francisco, the regions hope to complete nonresponse follow-up by July 15 and July 7, respectiively. 8 In addition to district offices in these two regions, some district offices in other regions also are experiencing difficulties finishing nonresponse follow-up. For example, as of June 28, the two district offices in Boston reported that they completed a combined total of about 88 percent of the city's nonresponse workload. The Washington, D.C., district offices also reported that they completed about 88 percent of their combined workload. The timely completion of nonresponse follow-up is important to allow the PES to be completed as planned and within schedule. PES interviewing, whereby enumerators visit households to collect PES data, was scheduled to begin in all district office areas on June 25. It actually began in some areas during the week of June 18 and will start in the remaining areas as offices complete nonresponse follow-up efforts. The Bureau will mot be able to confidently estimate when it will actually complete PES interviewing until the second week of July. The interviewing phase of the PES originally was scheduled to last until July 27, with an additional week to complete field quality control efforts. The Bureau is under an extremely tight time frame to meet the July 15, 1991, deadline for the Secretary of Commerce to decide 9 i. whether to adjust census counts. At this point, the Bureau does not expect the delays in completing nonresponse follow-up to significantly impede subsequent census operations, particularly the PES. We have no basis at this time for confirming or questioning the Bureau's belief. We note, however, that the 2- week delay in completing nonresponse follow-up during the 1988 dress rehearsal contributed to the delay in completing the PES. Finally, we should note that the Bureau's reported census completion figures to-date are based on the housing units in its address list. Completing fieldwork expeditiously provides the Bureau with additional time to review the extent to which its address list is indeed complete before census population counts are delivered to the President on December 31, 1990. This is particularly important in light of events that indicate potential problems with the Bureau's address list. Some of these problems include the many households that reported not receiving a census questionnaire-- and an indeterminate number that did not receive a questionnaire but failed to report it, the lack of certainty that changing i.;structions to field staff on how to handle reports of nonreceipt of questionnaires or receipt of duplicate questionnaires enabled accurate corrections to the address list, and a higher-than-budgeted number of additional housing units identified during the final Postal Service address check. 10 we mm em -- In closing, M r. C h a i r m a n , while nonresponse follow-up has been delayed a n d virtually n o o ffices finished o n schedule, most of th e district o ffices n o w either h a v e finished or e x p e c t to finish shortly. A t this p o i n t, th e B u r e a u d o e s n o t a n ticipate th a t th e delays will significantly i m p e d e th e progress o f th e census, p a r ticularly th e P E S , which begins as o ffices finish nonresponse work. H o w e v e r , while th e B u r e a u 's progress since th e S u b c o m m i tte e 's New York hearing is c o m m e n d a b l e , th e B u r e a u n e e d s to c o n tin u e its e fforts to e n s u r e th a t its remaining o ffices c o m p l e te follow-up e fforts as expeditiously as possible in order to m e e t tig h t census a n d P E S tim e tables. This conc:lu d e s m y prepared statement, M r. C h a i r m a n . My colleagues and I would be pleased to respond to q u e s tio n s . 11
1990 Census: Status of Questionnaire Follow-Up Efforts
Published by the Government Accountability Office on 1990-07-02.
Below is a raw (and likely hideous) rendition of the original report. (PDF)