supporting statement part A.wpd

SUPPORTING STATEMENT

U.S. CENSUS BUREAU

2007 CENSUS BILINGUAL FORM STUDY(CBiFS)

A. Justification

1. Necessity of the Information Collection

The Census Bureau tested a bilingual census form in the 2005 National Census Test (NCT). The bilingual questionnaire had a “swim lane” design that provided two response columns, one in English and one in Spanish, each containing the same questions and response categories. This form was mailed to a randomly selected set of 10,000 housing units across the United States.

Results from the 2005 NCT show that the bilingual form significantly increased the self response rate nationally (by 2.2 percentage points for paper response, and 1.1 percentage points for total response), and more specifically, in areas where there is a high concentration of non-White and Hispanic populations (Bouffard and Tancreto, 2006). Moreover, the bilingual form resulted in a higher proportion of response from Hispanic persons than the English-only form. However, item nonresponse rates for the bilingual form were higher for all household level items and Hispanic origin compared to the English-only form (Bouffard et al., 2006). There are many potential reasons for these item nonresponse discrepancies, including forms design, question wording, translation, and differences in the responding population.

The purpose of a follow-up test of the bilingual form is to discern if the item nonresponse issues can be resolved by improved form design and utilization of the questions that will be on the form in the 2008 Census Dress Rehearsal. Moreover, this test provides the opportunity to study the impact of the bilingual form in areas that contain a heavy concentration of Spanish-speaking people with limited English proficiency. Furthermore, this test provides the opportunity to verify the 2005 NCT finding of increased response to the bilingual census form¹.

Title 13, United States Code, Section 141 authorizes the Secretary of Commerce to conduct a decennial census of the population, and Section 193 authorizes the Secretary to conduct tests (such as the 2005 NCT) to gather supplementary information related to the census.

2. Needs and Uses

Mailing Strategies

This study will include one control panel and two experimental bilingual form panels. Each panel will have the same questionnaire content (i.e., the content selected for the Dress Rehearsal). The control panel will be the English-only form. The two experimental panels will contain variations of the bilingual form design to try to improve on the design that was used in the 2005 NCT. One of the experimental panels will be the bilingual form that will be used in the dress rehearsal, while the other will contain an experimental format alternative to the first panel. We expect to mail each form to roughly 10,000 housing units, for a total of 30,000 housing units in the study.

In early June 2007, the Census Bureau will mail an advance letter to every housing unit in this study. The advance letter will explain why we are conducting the 2007 Census Bilingual Form Study (CBiFS). The letter also will inform households that they will soon receive a request to complete a questionnaire. The advance letter for the two bilingual panels will have a dual language design.

The second mailing will be the initial questionnaire package. Housing units will receive a paper questionnaire and a first-class postage-paid return envelope. Included in the mailing package will be a letter from the Census Bureau’s Director that encourages households to respond. Households selected for the bilingual panels will receive one of the bilingual questionnaires, and households selected for the control panel will receive the English form. Respondents will be asked to mail back their completed questionnaire by Census Day, July 1, 2007.

The third mailing will be a reminder mailing. The reminder mailing asks households to respond to the CBiFS if they have not already done so. The reminder mailing will be an English-only postcard for the control panel and a letter for the bilingual panels since the Spanish translation requires more space than is available on the postcard.

Beginning July 2007, a replacement questionnaire will be sent to all housing units that have not responded by a pre-determined date. Accompanying the questionnaire will be a letter from the Census Bureau’s Director urging response. The replacement questionnaire will be the English form for all panels, including the bilingual panels. There wil be no field follow-up for the 2007 CBiFS. Sample questionaires are included in the attachment A, forms DE-1, DE-1(E/S), DE-2(E/S).

3. Use of Information Technology

We will be conducting this study in the paper self-response mode only, since we are seeking information on bilingual design alternatives specific to that mode.

4. Efforts to Identify Duplication

To our knowledge, this effort does not duplicate information collected by any other agency. Also, to the best of our knowledge, there is no directly comparable information assessing the effect on census response rates of using various contact strategies, large-scale replacement mailing changes to the coverage questions, residence rules instructions, and relationship categories, or the effects of response bias and reliability in the revised Hispanic origin race and questions that will be tested.

5. Minimizing Burden

In order to minimize the burden placed on the households selected for the 2007 CBiFS, we will attempt to obtain the required information from only one person for all the members of the household. Since we plan to use questionnaires that are similar to the 2005 NCT, which had an average response time of 10 minutes per household, we estimate the response time for the 2007 CBiFS to be approximately 10 minutes per household.

The 2007 CBiFS questionnaires incorporate a variety of user-friendly design features, such as embedded instructions that are intended to help respondents understand the intent of the questions. The total response burden will be approximately 5,000 hours for all 30,000 housing units selected for the test.

6. Consequences of Less Frequent Collection

The proposed 2007 CBiFS is a one-time study. It is the last of several Census mailout operations intended to finalize content for the 2010 Census. This is our final opportunity to collect the data necessary to:

• Develop the optimal bilingual short form questionnaire (including instructions, response options, and examples) for delivery to areas that contain a heavy concentration of Spanish-speaking people with limited English proficiency for the 2010 Census. Failure to implement the 2007 CBiFS treatments may impair our ability to develop the sophisticated instruments that we will need to capture the increasing diversity of the Nation. The final decision may affect all federal statistical and administrative data collections and tabulations based on the answers, and may be applicable to other Census Bureau surveys and census program as well;

• Discern if the item nonresponse issues (identified in the 2005 NCT) can be resolved by improved form design and utilization of the questions that will be on the questionnaire in the 2008 Census Dress Rehearsal

7. Special Circumstances

No special circumstances exist.

8. Consultations Outside the Agency

The Census Bureau published a notice (See Attachment) in the Federal Register on August 30, 2006, inviting public comment on our plans to submit this information collection (Vol.71 pg. 51571). We received a single email from a concerned resident stating that this study was unnecessary and suggested we cut the program.

The Census Bureau has historically conducted research to evaluate results for the decennial census. In order to develop the race and Hispanic origin items for both the 1990 and 2000 decennial censuses, we benefited from consultation with a variety of data users, including, but not limited to, academicians, national researchers, community leaders, and the Race and Ethnic Advisory Committees. Members of these committees, as well as the members of the Decennial Census Advisory Committee and the Professional Association Advisory Committee, are well-known scholars and social or political activists, and are respected as spokespersons for their communities and organizations. Both now and in the past, we have followed up on advice obtained through this ongoing consultation process.

In addition to the committees cited above, the Census Bureau has consulted with outside experts, including members of the National Academy of Sciences (NAS) and Dr. Donald Dillman of Washington State University. The NAS Panel on Residence Rules made extensive contributions to the development of the panels that are planned to test residence rules instructions and coverage questions.

9. Paying Respondents

Respondents participating in this survey will not receive any form of compensation for their participation.

10. Assurances of Confidentiality

Each component of the 2007 CBiFS respondent mailings and communications explains that participation is required and includes the Census Bureau’s assurance that the personal information requested will be treated as confidential, as stipulated by Title 13, United States Code.

All information that could identify individuals will be held in the strictest confidence under applicable federal statutes. In accordance with Title 13, United States Code, each household contacted will be assured of the confidentiality of its answers. This survey complies with the requirements of the Privacy Act of 1974 and the Paperwork

Reduction Act.

11. Justification for Sensitive Questions

The Census Bureau perceives no question as being sensitive.

12. Estimate of Hour Burden

A sample of approximately 30,000 households will be interviewed. To calculate the burden hours, we assume a theoretical 100 percent response rate and an estimated completion time of ten minutes per household based on results from the cognitive study. Consequently, the estimated total annual respondent burden for the 2007 CBiFS is 5,000 hours.

13. Estimate of Cost Burden

There is no monetary cost to respondents, only for the time it takes to respond to the questions.

14. Cost to the Federal Government

The cost of the study is covered under funding for decennial testing. The Census Bureau’s Decennial Management Division allocated resources for this study which includes Census Bureau staffing and overhead costs, as well as contracted services for translation, printing, and postal mailing costs.

15. Reason for Change in Burden

This is a new data collection effort.

16. Project Schedule

These dates are tentative.

Task	Estimated completion date
Draft study plan	7/2006
Finalize questionnaires	1/2007
Census Day	7/1/2007
Response data file	9/2007
Present preliminary results	1/2008
Final report	4/2008

17. Request to Not Display Expiration Date

No exemption is requested.

18. Exceptions to the Certification

There are no exceptions to the certification.

1 The 2005 NCT was a national test, while this follow-up study will target areas where the dual language design is likely most needed. While the results from 2005 cannot be directly generalized to this universe, we have no reason to suspect that the bilingual form would not maintain higher response than the English-only form among the population we are targeting.

6