Language bias and self-rated health status among the Latino population: evidence of the influence of translation in a wording experiment

This research uses a translation experiment to assess the Spanish translation of the fair response in the self-rated health measure among a representative study of the Latino population in the USA. Using a unique Latino-specific survey (n = 1200), researchers built in a split sample approach in the self-rated health status measure where half of the Spanish-speaking respondents (n = 600) were randomly given regular and the other half were given Mas o Menos in translating the English fair response. We first estimate a logistic regression model to estimate differences across language categories on the probability of reporting poor and fair health and then estimate a multinomial logistic regression to test whether respondents who took the survey in Spanish and given regular are more likely to rate their health as fair compared to English speakers and Spanish-speaking respondents who are given the Mas o Menos version. From our logistic regression model, we find that Spanish-speaking respondents given the regular response are more likely to report poor health relative to English-speaking respondents and Spanish-speaking respondents who were randomly given Mas o Menos. The results from our multinomial logistic models suggest that Spanish respondents provided with Mas o Menos are more likely to rate their health as good relative to the base category of fair and relative to both English and Spanish speakers given regular. This research informs the study of racial and ethnic disparities by providing a detailed explanation for mixed findings in the Latino health disparities literature. Researchers interested in self-rated health should translate the general self-rated health option fair to Mas o Menos as our wording experiment suggests that the current wording regular overinflates the reporting of poor health.