2007 Methods Panel OMB Supporting Statement Pta 092606

SUPPORTING STATEMENT

1. Justification

1. Necessity of the Information Collection

The U.S. Census Bureau requests authorization from the Office of Management and Budget (OMB) to conduct the American Community Survey 2007 Methods Panel tests.

Given the rapid demographic changes experienced in recent years and the strong expectation that such changes will continue and accelerate, the once-a-decade data collection approach of a census is no longer acceptable as a source for the housing and socio-economic data collected on the census long-form. To meet the needs and expectations of the country, the Census Bureau developed the American Community Survey (ACS). This survey collects detailed socioeconomic data every month and provides tabulations of these data on a yearly basis. The ACS allows the Census Bureau to provide more timely and relevant housing and socio-economic data while also reducing operational risks in the census by eliminating the long-form historically given to one in every six addresses.

Full implementation of the ACS includes an annual sample of approximately three million residential addresses a year in the 50 states and the District of Columbia, and another 36,000 addresses in Puerto Rico. A sample this large allows for annual production and release of single-year estimates for areas with a population of 65,000 or more. Lower levels of geography require aggregates of three and five years’ worth of data in order to produce estimates of comparable reliability to the census long-form. However, an ongoing data collection effort with an annual sample of this magnitude requires that the ACS continue to research possible methods for maintaining if not reducing data collection costs. If costs increase, the ACS would have to consider reductions in sample thus reducing the reliability of the data as compared to the reliability of the census long-form, especially at lower levels of geography.

One of the tests included in the 2007 Methods Panel addresses a method for potentially reducing data collection costs. In this test, we will implement the same mailing strategy as ACS production and send each sampled address a prenotice letter (see Attachment A), an initial questionnaire packet, (see Attachment B), and a reminder postcard (see Attachment C) and for those who haven’t responded by a certain date, we will send a second questionnaire packet. However, for this test we will send the prenotice letter using standard postage. Current ACS production procedures send all mail pieces using a first-class postage rate. Using standard postage rather than first-class postage for this mail piece can potentially save the ACS approximately two hundred and thirty thousand dollars in data collection costs each year. The test will evaluate whether the use of standard mailing for the prenotice letter impacts mail response rates.

A second test included in the 2007 Methods Panel addresses another aspect of ACS data collection relative to the census. Both the ACS and the census collect a core set of basic demographic questions (age and date of birth, gender, relationship, Hispanic origin and race). However, the 2010 Census will use a different format (similar to the format for the 2000 Census) from the format used by the ACS for collecting this information on the mail questionnaire. The census format, referred to as a sequential person design, creates a column for each person that includes each question and associated response categories (see Attachment D). The ACS format, referred to as the grid design, lists the names of all persons down the left side of the form, the questions across the top of the page, and the response categories fall in the ‘cells’ created by crossing the person names by question (see Attachment E).

This second test will compare the sequential person and grid formats for collecting the basic demographic information to measure the impact on data quality, specifically unit and item non-response rates, response distributions, and within household coverage. The outcome of the test will determine whether the different formats might contribute to differences in the estimates for the basic demographic questions. If the format does influence how people respond to these basic demographic questions, the Census Bureau will decide whether the ACS should alter its format of the collection of these data items to more closely reflect the census style format prior to the 2010 Census.

The 2007 Methods Panel may also include a third test contingent on the funding allocations in the President’s budget for 2007. This third test will measure and compare the data quality between two versions of new content proposed by the National Science Foundation for inclusion on the ACS. The proposed content asks about the major field in which a person received his or her bachelor’s degree. In this test, half the sample will answer an open-ended question reporting the actual degree he or she received. The other half of the sample will provide their field of degree information by answering a series of yes/no questions. The test will assess which, if either, version results in data of sufficient quality for inclusion on the ACS.

Given that the ACS collects data every day of the year in every county in the U.S. and in every municipio in Puerto Rico, the ACS provides an opportunity to produce data not available from any other source or survey at the same low levels of geography. The Census Bureau, in conjunction with the Office of Management and Budget, has a policy for determining whether new content or questions will be added to the ACS. As part of the content determination process, the Census Bureau must test the proposed content to determine whether the ACS can produce data of sufficiently high quality for the proposed topic. In all likelihood, this test will fold into the grid versus sequential form design test noted above in an effort to reduce cost and burden. The Census Bureau and OMB will consider these results in deciding whether to include the new content, per the Census Bureau’s Policy on New Content for the ACS (see Attachment F).

The Census Bureau collects data for this survey under Title 13, United States Code, Sections 141, 193, and 221. All data are afforded confidential treatment under Section 9 of that Title.

2. Needs and Uses

In order to provide data of comparable reliability as the census long-form at low levels of geography (e.g., census tract level) or for characteristics of special, small populations, the ACS must collect data on a continual basis and aggregate three to five years worth of data. Essentially the ACS collects data every day of the year, either by mail, telephone interviews or personal-visit interviews in order to have an adequate number of interviews to achieve estimates with comparable reliability to the census long-form at low levels of geography. Federal agencies use the ACS data to determine appropriate funding for state and local governments through block grants. State and local governments use ACS data for program planning, administration and evaluation. Thus, the reliability and the quality of the data must remain high in order for the users to rely on the data for funding decisions.

Similarly, the federal government as well as state and local governments uses the core, basic demographics collected as part of the census for funding and programmatic decisions. With full implementation of the ACS, those same data are available every year. From a data user’s perspective, large differences in the estimates for those core data items between ACS and the census can be problematic in terms of funding and program decisions. Since the ACS is a sample survey rather than a census we expect some differences in results between the two. However, there are many other factors that contribute to different results, such as differences in the interviewing staff, social relevance of the census versus a current survey, and even form design.

Thus, the 2007 Methods Panel will investigate ways to reduce or at least maintain data collection costs so the Census Bureau can continue to provide data of comparable reliability as the census long-form did. Additionally, the 2007 Methods Panel will test whether differences in form design between the census and the ACS may contribute to differences in results for the basic demographic items used by federal, state and local governments for funding and programmatic decisions. Lastly, funding permitting, the Methods panel will test proposed content regarding major field of study for a person’s bachelor degree in order to provide the National Science Foundation and the National Center for Education Statistics with current information regarding estimates of types of fields in which people receive bachelor’s degrees.

Information quality is an integral part of the pre-dissemination review of the information disseminated by the Census Bureau (fully described in the Census Bureau’s Information Quality Guidelines). Information quality is also integral to the data collection conducted by the Census Bureau and is incorporated into the clearance process required by the Paperwork Reduction Act.

3. Use of Information Technology

The first two tests included in the 2007 Methods Panel are mail-only tests, meaning the ACS will only implement the mail phase of data collection. However, the mail phase includes providing assistance to respondents via the Internet as well as by calling a toll-free number answered by a census interviewer using an automated instrument. For both of these tests, several of the mail pieces (the questionnaires, the cover letters, the reminder postcard, and the brochure) will include a toll-free number to reach staff at the Telephone Questionnaire Assistance (TQA) center. In the TQA operation, census interviewers can complete the ACS interview using an automated data collection instrument. These TQA interviews will contribute toward the measure of unit response for both tests. Additionally, several mailing pieces will also include a URL for the ACS where respondents can go to obtain more information about the ACS.

The third test, a test of new content, will include Computer Assisted Telephone Interviewing (CATI) and Computer Assisted Personal-visit Interviewing (CAPI) data collection, in addition to the mail data collection. Thus, if funded, this test will also include the TQA operation, provide a URL for general information about the ACS, as well as collect data via CATI and CAPI.

4. Efforts to Identify Duplication

The ACS removes the need to collect the detailed housing and socio-economic data as part of a long-form sample in the census. The basic demographic questions asked as part of the ACS as well as the census appear in other demographic surveys, but the comprehensive set of questions do not duplicate any other single information collection.

5. Minimizing Burden

The proposed data collection consists of questions asked of a small sample of households, not businesses or other small entities.

6. Consequences of Less Frequent Collection

The tests within the 2007 Methods Panel represent one-time, special tests with a defined period for data collection.

7. Special Circumstances

The Census Bureau will collect these data in a manner consistent with the OMB guidelines.

8. Consultations Outside of the Agency

The first two tests reflect internal Census Bureau needs, thus we have not involved other federal agencies in the design of the test. Knowledge from outside of the Census Bureau primarily comes from review of survey methodology and form design literature.

Subject matter experts from the National Science Foundation, the National Center for Education Statistics, as well as academics in survey methodology participated in the question development and pretesting efforts for the third test of new content.

We published a notice in the Federal Register on May 16, 2006 (vol. 71, pg. 28302), inviting the public and other federal agencies to comment on our plans to submit this request. We received one comment in response to that notice which the Census Bureau felt was not relevant to these proposed data collections.

9. Paying Respondents

We do not pay respondents or provide respondents with gifts.

10. Assurance of Confidentiality

The Census Bureau collects data for this survey under Title 13, United States Code, Sections 141, 193, and 221. All data are afforded confidential treatment under Section 9 of that Title.

In accordance with Title 13, each household will be assured of the confidentiality of their answers. A brochure is included in the questionnaire packet that is sent to sample households and contains this assurance (see Attachment B). This brochure also explains that participation in the ACS is mandatory.

11. Justification for Sensitive Questions

Some of the data we collect, such as race and sources of income and assets may be considered to be of a sensitive nature. The Census Bureau takes the position that the collection of these types of data is necessary for the analysis of important policy and program issues and has structured the questions to lessen their sensitivity.

12. Estimate of Hour Burden

In February 2007, we plan to mail questionnaires to a sample of 20,000 households to measure the impact on unit response of using standard postage. We plan to mail questionnaires to a sample of 40,000 households in March 2007 for the test comparing two different questionnaire layouts. The Census Bureau estimates that for the average household each version of the questionnaire used for the two tests will take 38 minutes to complete, including the time for reviewing the instructions and answers. If funding for the third test comes through, we will embed the test of the proposed new content in the test comparing the two questionnaire layouts, maintaining the sample at 40,000. However, if the third test receives funding, the test will also include a Content Follow-Up (CFU) Reinterview. The CFU Reinterview operation includes contacting all responding households for whom we have a phone number. Based on previous testing, we expect to have a phone number for about 80% of the original sample.

The estimate of the average length of time reflects the same estimate used for ACS production. The test comparing the standard versus the first-class postage will use the ACS production questionnaire, and the test comparing two different questionnaire layouts use the same content as ACS production, though the format for the basic demographic section differs. The proposed third test, examining the new content, will use the same content as ACS production plus one additional question only asked of people who have a bachelor’s degree or a higher degree. Thus the estimated average length of time for each of these tests remains at 38 minutes.

The CFU Reinterview component of the third test is estimated to last 15 minutes on average since that interview will only ask the basic demographic questions, up to five housing questions, the two questions preceding the education questions, and the education questions, including the proposed field of degree question. (The BOC will provide the instrument specification once completed.)

The total number of respondent burden hours if we do only the first two tests is estimated at 38,000. The total number of respondent burden hours if the ACS also includes the test of new content is estimated at 46,000 to account for the CFU Reinterview. Table 1 below shows the burden hours by test.

Test or Test component	Estimated Average Burden Hours
Standard versus First-class postage	12,667
Grid vs. Sequential questionnaire layout (tested alone)	25,333
Grid vs. Sequential questionnaire layout combined with test of new content	(Included in the above.)
Content Follow-Up Reinterview component of test of new content	8,000

13. Estimate of Cost Burden

There are no costs to the respondent other than his/her time to respond to the survey.

14. Cost to Federal Government

The total estimated cost of the 2007 Methods Panel test comes to approximately $450,000. The Census Bureau will pay all costs of the 2007 Methods Panel tests.

15. Reason for Change in Burden

This collection is being submitted as new.

16. Project Schedule

The test of the standard postage will essentially follow the same mail-out schedule as the February 2007 panel of ACS production with the exception that the prenotice letter will be mailed-out with standard postage and will go a few days earlier to account for potentially slower mail delivery. See table 2a below for the anticipated dates of each mail component.

Table 2a: Schedule for delivery of mail pieces for test standard postage¹

Mail Piece	Date
Prenotice letter	1/22/07
Initial questionnaire packet	1/29/07
Reminder postcard	02/01/07
Replacement questionnaire packet	02/23/07

The test of the sequential versus the grid format of the basic demographic section of the ACS questionnaire will follow the same mail-out schedule as the March 2007 panel of ACS production. See table 2b below for the anticipated dates of each mail component. Data capture of the completed questionnaires will continue through the end of April 2007.

Table 2b: Schedule for delivery of mail pieces for test of questionnaire format

Mail Piece	Date
Prenotice letter	02/22/07
Initial questionnaire packet	02/26/07
Reminder postcard	03/01/07
Replacement questionnaire packet	03/19/07

The test assessing the quality of the data for the proposed field of degree question will use the same timing for the mailing pieces as the test of questionnaire format. However, this test will also include data collection through CATI, CAPI, and CFU Reinterview. See table 2c below for the anticipated start dates for each phase of data collection.

Table 2C: Start dates for each phase of data collection for the test of new content.

Contact Method		Date
Prenotice letter		02/22/07
Initial questionnaire packet		02/26/07
Reminder postcard		03/01/07
Replacement questionnaire packet		03/19/07
Content Follow-up Reinterview	03/16/07
CATI data collection		04/01/07
CAPI data collection		05/01/07

17. Request Not to Display Expiration Date

We will display the expiration date on the form.

18. Exceptions to the Certification

There are no exceptions to the Certification for Paperwork Reduction Act submission.

1 The ACS mails the prenotice letter and the initial questionnaire in the last few days of the prior month (e.g., January) so that the sampled household actually receives their questionnaire during the first few days of the sample month (e.g., February).

8