Date: Thursday, 5/16/2024 4:30 PM

Alisú Schoua-Glusberg and Daniela Glusberg

Research Support Services Inc.

Preparing non-English versions of survey questions for limited English proficient (LEP) populations in the U.S. typically involve translating English questions, while trying to maintain the meaning and measurement properties of the original items. However, even using best practice for survey translation, complications can arise stemming from the fact that target language respondents will not necessarily share the historical or cultural context of English respondents.

Some immigrants  arrive in the U.S. with limited or no English proficiency. Language acquisition facilitates acculturation, as language allows to learn about the dominant culture and how it constructs certain concepts. Immigrants who complete surveys in non-English languages have typically not yet fully learned these concepts. 

Many demographic survey questions are deeply rooted in how we classify people in the U.S. along a number of characteristics, classifications not necessarily matching classifications in the person’s culture of origin. Race, gender, sexual orientation, educational attainment, among others, can be understood differently in another cultural and linguistic context, and this can lead to survey error. 

Translations often use cognates of English words with common etymology and apparent overlapping meaning, but cultural and historical context makes them different. For example, the word ‘race’ is usually translated into Spanish as ‘raza’. However, in context, race refers to a U.S. classification of people based on skin color and ancestry. ‘Raza’, on the other hand, is not used this way in most Spanish-speaking countries. In Mexico, it’s not used to classify a person but to describe  the collective ethnicity of a people born out of the historical mixing of indigenous peoples and Spanish conquerors. And yet, we have no better Spanish word for ‘race’. 

This presentation will discuss examples such as this and these cultural conceptual mismatches may lead to survey error in important background variables.