This proposal seeks to employ self-report and auxiliary data sources to conduct a systematic, comprehensive assessment of how respondent race/ethnicity and culture influence survey-related behaviors. Although most acknowledge the important role that culture can play in determining survey-related behaviors, much of the available survey evidence comes from post hoc and secondary analyses that have limited scope and generalizability. In addition, most available survey research focuses on racial and ethnic classifications of survey respondents and fails to consider underlying cultural orientations and values that may mediate observed variations. To address these limitations, this study will collect data from a sample of 1,200 adults representing four distinct cultural groups in the U.S.: African Americans, Korean Americans, Mexican Americans, and non-Hispanic Whites. Three of these groups represent the largest racial/ethnic populations in the U.S. Korean Americans are included to represent the country's rapidly growing East Asian culture and to insure maximum variability in the cultural orientations and values to be assessed. Sampled adults will be further stratified by gender, age, education, and language preference groups. Interviews will be audio- recorded and videotaped. These recordings subsequently will be behavior coded to identify verbal and nonverbal markers of cognitive processing difficulties, such as problems with question comprehension, memory retrieval, response mapping, and socially desirable responding. The response latencies associated with each answer also will be electronically recorded. Using these data, hierarchical linear modeling will be employed to investigate research questions concerned with (i) the effects of race/ethnicity on respondent verbal and nonverbal behaviors as indicators of the response process in survey interviews;(2) the effects of race/ethnicity on survey response styles, such as acquiescent and extreme responding behavior, non-differentiation, no-opinion responding, and response order effects;(3) the degree to which the effect of race/ethnicity on cognitive processing is moderated by question design features, such as topic and format;and (4) the degree to which race/ethnicity effects can be explained by individual differences in cultural value dimensions. This research will provide the opportunity to explore the magnitude of racial/ethnic variability in cognitive processing difficulties across a variety of common survey question features. It also will address why racial/ethnic differences in cognitive and response effects emerge by examining the extent to which such differences can be attributed to systematic variations in cultural value orientation and other individual differences. Results will lead to recommendations regarding best practices in the design of survey questions that minimize cross-cultural variability in cognitive processing difficulties. This is an important goal, as health surveys are one of the key yardsticks by which public health disparities are measured.
