Justification & Explanation of 2008 Methods Panel Testing

SUPPORTING STATEMENT

Justification

1. Changes to Approved 2007 ACS Methods Test Data Collection

The U.S. Census Bureau requests authorization from the Office of Management and Budget (OMB) to modify the approved American Community Survey 2007 Methods Panel, OMB Control Number 0607-0936, originally approved on December 18, 2006. A change request to this original package was approved on April 19, 2007.

This change submission describes two tests currently planned for the ACS Methods Panel in 2008. Implementation of these tests is dependent on receiving funds through the Census Bureau’s FY08 appropriation. If FY08 funds are not received, these tests will be delayed until funding becomes available. Both projects test methods to increase response rates and cost-efficiency. None of the testing involves changes to the ACS questionnaire. Given the relatively low risk of the methods we plan to test coupled with the need for large samples for the tests, the tests will be conducted on 2008 ACS production housing units using the production ACS questionnaire. The tests will be conducted separately, over a period between one to four months (depending on the test).

These two 2008 Methods Panel tests will be conducted during the mail phase of the production operations. Computer-assisted Telephone Interviewing (CATI) and Computer-assisted Personal Interviewing (CAPI) operations will be conducted as part of the production interviewing process. The CATI and CAPI data collection phases do not include any Methods Panel test activities.

The first project is a test of mailing a multi-lingual brochure in the pre-notice or initial questionnaire mailing as an attempt to increase response from foreign-language speaking households prior to personal visit interviewing. The second project tests whether sending an additional mailing piece to mail nonrespondents for whom we do not have a telephone number can increase response during the month when other mail nonrespondents are being contacted for a telephone interview.

Multi-lingual brochure test

The ACS questionnaire includes a small message on the cover that informs respondents how they can obtain assistance in English or Spanish. There are no other messages on the questionnaire or any other mailing piece that explain how non-English speaking households that require language assistance can request help. Therefore, the ACS often collects data for these populations during the telephone and personal visit follow-up operations, particularly for Spanish-speaking households.

The multi-lingual brochure test is aimed at increasing ACS data collection for non-English speaking households prior to nonresponse follow-up activities, thus potentially providing a less expensive means for collecting their data and reducing any confusion resulting from the lack of instructions in other languages. We will determine if we can increase response from Spanish-, Korean-, Chinese-, and Russian-speaking households by mailing a single brochure in these languages (and English) that provides information about how to obtain assistance by telephone. Assistance may include helping respondents fill out the mail form, fulfilling requests for a Spanish questionnaire, or, most likely, conducting the interview in these languages over the phone. We will also test whether the placement of the brochure, with the pre-notice letter or the initial questionnaire package, impacts response from these households. See Appendix A for both versions of the brochure, the one mailed with the pre-notice letter, and the one mailed with the initial questionnaire package. Both versions will undergo cognitive testing before the field test.

The selection of the languages for this brochure was related to their prevalence in the ACS telephone workload as well as the current language capabilities in the call centers. One notable language that is missing is Vietnamese, which is also prevalent, but challenging to implement at this time given the difficulty in staffing an interviewer that speaks Vietnamese. While we recommend using these four languages in the test, we remain optimistic about adding more languages in the future should this test prove successful.

We anticipate mailing the multi-lingual brochure to 2008 ACS production housing units over the course of four production months. The monthly mail sample (approximately 230,000) will be cut into thirds (roughly 77,000 each), where one-third will receive the brochure with the pre-notice letter, one-third will receive the brochure in the initial questionnaire package, and the remaining third will not receive a brochure. Since we have a very small target population, we will combine the measures across four months. The test will assess the success of including the brochure in the two mail pieces by measuring the percentage of non-English speaking and linguistically isolated households interviewed in these languages prior to nonresponse follow-up activities. We will also assess which placement of the brochure generates the most response. We plan to mail brochures to production cases for the August, September, October, and November sample panels in 2008, pending funding.

Additional mailing to mail nonrespondents without phone numbers

This test involves determining whether the ACS can increase mail response by sending an additional mailing piece to mail nonrespondents for whom we do not have a phone number and thus, cannot include in the telephone nonresponse follow-up operation. This study will test two different types of mailing pieces (a postcard and another questionnaire package), each mailed after the replacement questionnaire package. We will measure which type yields the highest increase in response for the non-telephone eligible universe relative to the costs associated with each piece. Note that we will be cognitively testing two messages for the new postcard, only one of which will be used in the field test. See Appendix B for the revised cover letter for the third questionnaire package and both versions of the postcard. All of these materials will undergo cognitive testing before the field test.

We will test these treatments on production cases that have not responded by mail, and for which a phone number is unavailable. Based on ACS production data, we expect to have roughly 45,000 housing units each month that are mail nonrespondents without a phone number. Thus using one month of ACS production data, we plan to split that sample in thirds, where 15,000 housing units will receive an additional reminder postcard, 15,000 will receive a second replacement questionnaire, and the third will not receive any additional mailing. The test is scheduled to be conducted during the May 2008 production sample panel, pending funding.

2. Changes to the Estimate of Hour Burden

The total number of respondent burden hours approved in the OMB clearance for the 2007 ACS Methods Panel test (per the April 2007 revision) was 44,000. The 2008 tests add no additional burden since all tests use 2008 production cases for which the burden is covered in the production ACS OMB clearance.

3. Reason for Change in Burden

Changes to the approved Methods Panel OMB clearance illustrate an extension of the Methods Panel testing plans for 2008. The 2008 ACS Methods Panel tests relate to one of the overall goals of the Methods Panel, to improve the efficiency of data collection operations.

4. Test Project Schedule (Pending funding)

Table 1. Start dates for Multi-lingual Brochure Test

Contact Method	Dates
Pre-notice letter	July 24 (for August panel) August 21 (for September panel) September 18 (for October panel) October 23 (for November panel)
Initial questionnaire packet	July 28 (for August panel) August 25 (for September panel) September 22 (for October panel) October 27 (for November panel)

Table 2. Start dates for Additional Mailing Test

Contact Method	Dates
Production Pre-notice letter	04/24/08
Production Initial questionnaire packet	04/28/08
Production Reminder postcard	05/1/08
Production Replacement questionnaire packet	05/22/08
Additional mailing	6/5/08

4