The aim of this evaluation project was to qualitatively evaluate in-person doorstep interactions with multilingual and/or complex households during the Census Bureau’s 2020 Census Nonresponse Followup operation. The initial focus of this research was on identifying reasons for nonresponse as well as gauging receptivity to different messaging and materials across languages. Other topics were added to the scope of this project that included investigating the potential for mistakenly omitting or including household members on the roster collected at the doorstep (especially children), in addition to better understanding the impact of the COVID-19 pandemic on doorstep interviewing done during the 2020 Census.

This work was undertaken by staff at the Census Bureau’s Center for Behavioral Science Methods (CBSM), RTI International (RTI), and Research Support Services Inc. (RSS). RSS staff conducted and analyzed 28 focus groups, each group comprised of either English-speaking monolingual, or bilingual English and Spanish-, Arabic-, Chinese-, Korean-, Russian- or Vietnamese-speaking enumerators. All groups were moderated in English by experienced moderators who were also fluent in the non-English language of the group, except for the Korean groups, which had a Korean-speaking notetaker in each group. Enumerators were recruited by Census Bureau field staff at the regional offices with guidance and oversight from CBSM researchers. Groups consisted of one to six enumerators, with a total of 111 enumerators participating, and took place between September 16 and 30, 2020.

Some key recommendations from this evaluation are as follows: provide key respondent facing wording (such as doorstep messages and census questions) in more languages; emphasize the importance of verbatim interviewing in all languages (especially when a formal translation exists); provide enumerators with more non-English doorstep messaging practice in training; optimize assignment of language barrier households to enumerators who speak the language of the household; include more information and explanation in census materials, enumerator training, and public campaigns about the importance of including young children in the census; and provide more information about confidentiality to respondents, building managers, and landlords answering as proxies to help minimize omissions at the doorstep that stem from concerns about there being more residents than allowed in housing units.