Hibiscus A
Messages for Gaining Respondent Cooperation in Multiple Languages (303150)
Yazmin Garcia Trejo, U.S. Census Bureau*Patricia Goerman , U.S. Census Bureau
Mikelyn Meyers, U.S. Census Bureau
Alisu Schoua Glusberg, RSS
Mandy Sha, RTI International
Keywords: cross-cultural, multilingual surveys, multiple methods of evaluation
This paper examines methods for evaluating the effect of interviewers’ introductory messages on respondents’ willingness to participate in censuses and surveys. While there is considerable research dedicated to interviewer effects on survey participation, we know little regarding evaluating and designing introductory messages for use across languages. This paper looks at results of three U.S. Census Bureau studies focusing on verbal messages for use with non-English speaking respondents. The three studies look at the issue through different methodologies: 1) field interview observation; 2) expert review; and 3) focus groups. During the 2010 Census, researchers observed non-English interviews in the field. At that time, interviewers had not been trained on what to say when approaching non-English speakers but improvised their own messages. Researchers identified a number of issues with the messages being used. In 2015, the Census Bureau, along with RTI International and RSS, conducted expert reviews of verbal messages for use at the doorstep in Chinese, Vietnamese, Korean, Russian and Arabic. As a starting point, the reviewers looked at direct translations of English-language messages, with little or no cultural adaptation. Reviewers then proposed new wording that they deemed more culturally appropriate. The third method was focus groups with respondents who spoke English, Spanish, Chinese, Vietnamese, Korean, Russian and Arabic. Standard English-language introductions were translated and modified, using the expert review recommendations, into the six languages. Researchers created videos of interviewer and respondent interactions, which were shown to the focus groups to elicit feedback. This paper discusses findings obtained from the three different methodologies and provides recommended messages for use in each language. We conclude with a discussion of next steps towards developing and evaluating introductory survey messages tailored to different language groups.