Questionnaire Effects on Measurements of Race and Spanish Origin
Nancy Bates, Elizabeth A. Martin, Theresa J. DeMaio, and Manuel de la Puente
Although race may be regarded as a straightforward and objective characteristic, the accurate measurement of race and ethnicity for the U.S. population poses difficult challenges. Popular racial classifications vary according to immigration patterns, societal changes and political forces. The result is a complex concept subject to many interpretations.
This article analyzes and summarizes research results on the decennial census race and ethnicity questions conducted or sponsored by the U.S. Census Bureau Center for Survey Methods Research. The article analyzes multiple replications of an experiment conducted to determine effects of question order on item nonresponse and reporting of race and origin. In most of the experiments, placing Spanish origin prior to race successfully reduced nonresponse to Spanish origin while increasing nonresponse to race only slightly, if at all. The effects of reordering questions on race reporting among Hispanics are also examined, including an exploration of interactions with Hispanic subgroup, place of birth, and language. Recent results from in-depth interviews and focus groups with Hispanic respondents also supplement the findings from the questionnaire design tests.
Batch Sampling; composite sample; Bayes estimate; MLE; bias correction.