Methodological Influences on Comparability of Race Measurements:
Several Cautionary Examples
Elizabeth Martin and Eleanor Gerber
KEY WORDS: question wording effects, response error, mode effects, interviewer effects, reliability, race measurement
ABSTRACT
Several recent developments, including immigration from Central and South America and changes in official racial classifications used in government surveys, raise concerns about the quality and consistency of racial statistics. This paper reviews methodological problems affecting race measurements, including wording and conceptual issues, mode effects, and interviewer effects. We document and comment on the possible impact of variations in wording of alternative questions on race currently being used in Census Bureau household surveys and censuses.
We illustrate the impact of different measurement methods with four recent examples of data anomalies apparently produced by differences in the methods used to collect race data. These include differences in race data collected using 1990-style and Census 2000-style mail short forms; by the Accuracy and Coverage Evaluation survey and Census 2000; by Census 2000 and the American Community Survey; and by the Census Quality Survey and Census 2000.
Finally, we suggest research that is needed to improve the quality of race measurement methods, and better understand methodological sources of variation in race measurements.
CITATION: Elizabeth Martin and Eleanor Gerber. 2003. “Methodological Influences on Comparability of Race Measurements: Several Cautionary Examples.” Proceedings of the American Statistical Association (Survey Research methods Section): 2697-2704.
