Effect of Language on Mental Health Reporting Among Hispanics
Flicker, L., Dean, E. F., Quiroz, R. S., & Wickelgren, M. (2006, May). Effect of Language on Mental Health Reporting Among Hispanics. Presented at American Association for Public Opinion Research 61st Annual Conference, Montréal, Canada.
Mental health indicators vary across cultures in terms of both service utilization and reporting of symptoms. Several factors may contribute to the variation. First, North Americans distinguish mental and physical illness more than other cultures, and Hispanics report physical symptoms when depressed. Second, idiomatic expressions describing mental health symptoms (such as feeling blue) are not universally understood across cultures. Third, symptoms of mental health conditions do not carry the same stigma across all cultures, thus affecting reporting. These three factors raise concerns about the reliability of multilingual surveys assessing mental health.
Our research examines mental health items in two surveys administered in Spanish and English. The first is the World Trade Center Health Registry (WTCHR), a population-based study of 71,437 persons exposed to the World Trade Center attacks of September 11, 2001. The second is the National Survey on Drug Use and Health (NSDUH), an annual household survey that collects data on drug/alcohol use and mental health symptoms from approximately 67,500 respondents across the country.
Both surveys employ the K-6, a mental health screening tool. Preliminary analysis of WTCHR data found that Spanish-speaking respondents report more mental health symptoms and conditions in the past month when compared with other groups. The 2004 NSDUH reported lower estimates of serious psychological distress (a composite measure for feelings of sadness, hopelessness, enervation, etc.) for Hispanic youth and young adults.
We analyze each studys data in more detail to explain the discrepancy in mental health status among Hispanics in the two studies. In addition to comparing point estimates for mental health status, we will predict mental health symptoms and status by ethnicity, language of survey administration, and demographic variables in separate models for each study. We expect an interaction between Hispanicity and language of survey administration in the reporting of mental health symptoms.