Diagnostic accuracy of self-reported racial composition of residential neighborhood

• Published in: Annals of epidemiology 2015-08, Vol.25 (8), p.597-604
• Main Authors: Hidalgo, Bertha, PhD, MPH, Kaphingst, Kimberly A., PhD, Stafford, Jewel, MSW, Lachance, Christina, MPH, Goodman, Melody S., PhD, MS
• Format: Article
• Language: English
• Subjects: Adult; African Americans - statistics & numerical data; Blacks - statistics & numerical data; Cross-Sectional Studies; Diagnostic accuracy; Diagnostic Self Evaluation; Ethnicity - statistics & numerical data; Female; Hispanic or Latino - statistics & numerical data; Humans; Internal Medicine; Male; Middle Aged; Neighborhood racial composition; New York - epidemiology; Racial segregation; Racism - statistics & numerical data; Residence Characteristics; Self Report; Social Environment; Socioeconomic Factors; Surveys and Questionnaires; Whites - statistics & numerical data
• ISSN: 1047-2797; 1873-2585
• DOI: 10.1016/j.annepidem.2015.04.003

Abstract Purpose To examine the diagnostic accuracy of self-reported measures of individuals' perceptions of the racial and ethnic composition of their communities with objective data (i.e., census) as the criterion standard and assess differences in concordance in subjective and objective measures of segregation by race and ethnicity. Methods We examined data from 943 adult community health center visitors in Suffolk County, New York to assess differences between self-reported racial composition of current neighborhood and 2010 U.S. Census data. A cross-sectional convenience sample was obtained; questionnaires were used to compare participant responses about the racial composition of their current neighborhood and their town of residence. Results Respondents who self-identified as white were more likely to self-report racial composition of their neighborhood consistent with 2010 Census estimates. Relative to census estimates, 93.1% of blacks overestimated the proportion of their current neighborhood that was black, and 69.8% of Hispanics overestimated the proportion that was Hispanic. Conclusions There were statistically significant differences between the participants' self-reported neighborhood racial composition and census data across race and ethnicity groups. Future studies are needed to validate self-reported measures of individuals' perceptions of the racial and ethnic composition of their communities to examine the association between individual segregation experience and health.