This study represents the U.S. Census Bureau’s first attempt to apply known pre-testing techniques as developed in the monolingual English context to the translation of one of its demographic surveys. The Spanish translation of the American Community Survey (ACS) Computer-Assisted Personal Interview (CAPI) instrument was used to conduct cognitive interviews. The research question was to determine how Spanish speaking respondents understood key terms and questions, and whether those interpretations were consistent with the meaning of the English questionnaire. Spanish questions whose meaning reflected the English content were considered to be functionally equivalent. That is, they were considered to be measuring the same construct without cultural or linguistic interference (Smith 2002).

The report documents linguistic and questionnaire design challenges that are unlikely to be unique to the ACS Spanish instrument; rather they are hypothesized to be consistent across Spanish questionnaire translations. It is also hypothesized that these issues do not result from inadequate translations or poor translation techniques, but rather stem from lack of pretesting.