Justification & Description of Cognitive Testing of Spanish ACS CATI/CAPI Questions

The Census Bureau plans to conduct research under the generic clearance for questionnaire pretesting research (OMB number 0607-0725). Cognitive interviews will be conducted to evaluate select segments of the 2008 Spanish language contained in the CAPI/CATI version of the American Community Survey (ACS). We will also conduct a small number of English-language interviews on the same segments of the English version of the CAPI/CATI version of the ACS. This study is not designed to formally test the English language version but English interviews will be used as a baseline to examine whether respondents are interpreting the terms and questions in a parallel manner across the two languages.

The research will have several aims:

• To identify and document problems associated with specific Spanish-language terms and concepts contained in the CAPI/CATI version of the American Community Survey;

• To test differences between the “stateside” and Puerto Rico versions of the Spanish language ACS instrument with respondents from a variety of Latin American countries and from Puerto Rico;

• To test differences between the CAPI and CATI versions of the Spanish survey instrument. This will involve the testing of a few of the questions both with and without a flashcard;

• To identify potential alternative Spanish-language wording for the problematic terms and concepts;

• To test newly identified/revised terminology; and

• To examine how Spanish-speaking respondents understand and respond to the Spanish language questions in the context of the CAPI/CATI instrument.

The cognitive interview results will inform possible revisions to the Spanish CAPI/CATI version of the American Community Survey.

From early January, 2008 through the end of April, 2009, a contractor will carry out a maximum of 150 interviews and Census Bureau staff will conduct a maximum of 30 interviews in Texas, California, and Florida. The 2008 American Community Survey questionnaire has been divided into three segments for this testing and the project will be comprised of three phases. A copy of the questionnaire for each phase is enclosed. Each phase will contain two rounds of testing on one of the three segments of questions.

Phase 1: The first 30 interviews (25 Spanish and 5 English) will be conducted based on the first group of ACS questions. The contractor and the Census Bureau researcher will propose revisions based on the round one findings and the second round of testing will be conducted using revised versions of the same questions. We will again do 30 interviews (25 Spanish and 5 English) on the revised versions of the phase one questions. Final recommendations will come out of the second round of testing.

Phase 2: Round one will be comprised of 30 interviews (25 Spanish and 5 English) based on the second segment of ACS questions. The contractor and the Census Bureau researcher will propose revisions based on the round one findings and the second round of Phase 2 testing will be conducted with revised versions of the same ACS questions. We will do 30 interviews (25 Spanish and 5 English) on the revised versions of the phase two questions. Final recommendations will come out of the second round of testing.

Phase 3: Round one will be comprised of 30 interviews (25 Spanish and 5 English) based on the third segment of ACS questions. The contractor and the Census Bureau researcher will propose revisions based on the round one findings and the second round of Phase 3 testing will be conducted with revised versions of the same ACS questions. We will do 30 interviews (25 Spanish and 5 English) on the revised versions of the phase three questions. Final recommendations will come out of the second round of testing.

Recruiting for the Spanish-language interviews will focus on monolingual Spanish-speaking respondents and Spanish-speaking respondents with limited English language proficiency. Recruiting for the English-language interviews will focus on monolingual English speakers.

Spanish-speaking respondents will be from a variety of national origins and will vary by the amount of time they have lived in the United States. All respondents will vary by educational level and by age and gender.

Respondents will be recruited through flyers, newspaper advertisements, and through the contractors’ ties to local Hispanic communities in Florida, California, and Texas. Respondents will be paid an honorarium of $40 for each interview.

The cognitive interviews will be conducted using concurrent and retrospective cognitive interview probes and techniques. There will also be debriefing questions for the respondent at the end of the interview. All interviews will be tape-recorded with the respondents’ permission, to facilitate summarization of results. Participants will be informed that their response is voluntary and that the information they provide is confidential and will be seen only by employees involved in the research project.

We estimate that the length of each interview session will average between 60 and 75 minutes. Thus, the maximum estimated burden for the 180 interviews is 225 hours.

The contact person for questions regarding data collection and study design is:

Patricia Goerman

Center for Survey Methods Research

Statistical Research Division

U.S. Census Bureau

4600 Silver Hill Rd.

Washington, DC 20233

(301) 763-1819

[email protected]