Justification for ACS Testing of Revised Reminder Postcards & Letter

The Census Bureau plans to conduct cognitive testing of postcards and a revised cover letter for a proposed additional targeted mailing for the American Community Survey (ACS). This additional mailing is one of the proposed experimental panels for the ACS Methods Panel that is currently being planned for the spring of 2008. The postcards and revised cover letter will be targeted to ACS sample households who do not send in their mail questionnaire by the end of the first month of data collection and for whom we have no phone number listings. Lacking their phone numbers, we cannot include these households in the second phase of ACS data collection by telephone (the CATI nonresponse followup). They remain in the nonrespondent pool, but have approximately a 1/3 chance of selection for inclusion in the final stage of data collection (CAPI nonresponse followup).

The Census Bureau hypothesizes that sending an additional reminder postcard or an additional replacement questionnaire to these households during the telephone phase of data collection could boost their mailback response rates, improve the quality of data, and possibly reduce the costs of personal visit followups (a separate OMB change request for this mailout test will be submitted later this fall).

We plan to conduct cognitive testing of two versions of a new postcard and a cover letter for a third replacement questionnaire package to choose and refine the materials that will be included the 2008 ACS Methods Panel mailout test. These three documents are enclosed. One of the postcards uses a conciliatory and altruistic approach, gently asking respondents to help their community leaders by mailing in the completed form, thereby providing vital information to help plan services of benefit to respondents. The other postcard uses more forceful language to note that participation is required by law (an approach that Don Dillman found to be successful in raising census response rates). We will use the results of cognitive testing to decide which one of the two postcard approaches will go forward into the split-panel test, and to revise the language, as necessary. We are also testing a revised version of the cover letter for use in an additional mailing with a third replacement questionnaire package that will also go forward into the split-panel test. The letter combines these “carrot” and “stick” approaches.

From late August to early December, staff from SRD will conduct a maximum of 15 cognitive interviews in the Washington, DC metropolitan area, varying the order in which respondents are presented with the 3 experimental pieces, and conducting a followup debriefing.

Our plan is to recruit respondents who are at risk of falling into the category of mailback nonrespondents without telephone numbers. To determine those respondents at risk, we conducted a statistical analysis comparing the characteristics of 2006 ACS respondents for whom we did not have telephone numbers with 2006 CAPI respondents for whom we did not have telephone numbers. We found that the CAPI universe had disproportionately higher concentrations of renters, people in multi-unit structures, crowded households, movers, Spanish-speaking households, seasonal residents, households using public assistance, and households with no telephone service. We will attempt to target people who have one or more of these or related characteristics for cognitive interviews. Respondents will be recruited through flyers, community organizations, personal networks, and our recruiting database. Interviews will be conducted in SRD’s cognitive laboratory or in places more convenient to respondents.

Prospective respondents will be informed that participation is voluntary and information that could identify individuals will be held strictly confidential. The information will be used only for research purposes. Interviews will be audiotaped with the permission of the respondents.

Respondents will be paid $40 for their participation.

The estimated time for completion of the interview is 45 minutes. Thus, the maximum estimated burden is 12 hours.

The contact person for questions regarding this data collection is:

Laurie Schwede

Center for Survey Methods Research

Statistical Research Division

U.S. Census Bureau

4600 Silver Hill Road

Washington, D.C. 20233

301-763-2611

Laurel.K.Schwede@census.gov