Detail from "the second line," a painting by Bob Graham. For more about the artist, click here.

Online Program

Beyond Words: In Search of Effective Methods For Translating Survey Letters
View Presentation View Presentation

*Yuling Pan, U.S. Census Bureau 
Marissa Fond, U.S. Census Bureau 
Mandy Sha, RTI International 

Keywords: survey translation, survey letters, multilingual, linguistic and cultural minorities

Translating survey letters is a method that survey organizations use to reach hard-to-survey populations such as those who speak languages other than English. A common practice is a word-for-word translation of an English letter into target languages, following the information sequencing of the English text. Numerous studies pointed out the inadequacy of this approach: speakers of target languages could understand the translation word-for-word but not the communicative intent of the letters. This could affect their cooperation or survey participation.

This paper demonstrates why this approach fails. Data are 204 cognitive interviews from a two-part multilingual project conducted by the U.S. Census Bureau to test translations of survey letters in five target languages (Spanish, Chinese, Korean, Russian and Vietnamese). We analyzed the discourse structure (e.g. information sequencing) of the English letters and the respondents’ reactions to the translated survey letters.

Findings from Part I of the study show that differences in discourse structure between English and target languages could hinder respondents’ comprehension of the letters’ intended messages, while improving translations at the word or sentence level did not greatly enhance readability. Next, in Part II, we proposed an alternative approach to translating survey letters, based on theories of discourse analysis and social interaction, encompassing attention to communication channel, relationship, and topic introduction (Scollon 1998). Part II cognitive interview results demonstrate that the new translations increased respondents’ understanding of the intended messages and the likelihood of their survey participation. This finding has methodological implications for achieving cultural equivalence in survey translation, with the hope that translated survey letters will achieve their intended purpose of encouraging survey participation from non-English speaking respondents.

ASA Meetings Department · 732 North Washington Street, Alexandria, VA 22314 · (703) 684-1221

Copyright © American Statistical Association.