Premium
Calculating Multirace Categories and Race‐specific Rates: A Cautionary Note to Rural Health Researchers and Analysts
Author(s) -
Hodgson Todd W.,
Farmer Frank L.
Publication year - 2003
Publication title -
the journal of rural health
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.439
H-Index - 57
eISSN - 1748-0361
pISSN - 0890-765X
DOI - 10.1111/j.1748-0361.2003.tb00555.x
Subject(s) - race (biology) , metropolitan area , demography , mortality rate , geography , population , ethnic group , medicine , biology , sociology , botany , archaeology , anthropology
Context: Revised federal standards for racial and ethnic data have resulted in 57 new self‐selected multirace categories that cannot be directly compared with traditional “one‐race‐only” categories. “Bridge methods” are available, but each produces different estimates of single‐race populations, raising concern about the potential influence on calculated rates for rural and nonmetropolitan areas. Purpose: to examine the potential impact of several bridge methods when calculating race‐specific mortality rates. Methods: Population and mortality data were collected for 75 counties in Arkansas, and race‐specific mortality rates were calculated using 3 deterministic bridging methods: (1) Equal Fractions, (2) Largest Group, and (3) Largest Group Other Than White. The mortality rates were compared at the state and metropolitan/nonmetropolitan levels, as well as for 4 rural substate regions. Findings: Variable rates were obtained from different methods. Estimated mortality rates for the largest single‐race groups, White and Black, varied only slightly by method at the state and nonmetropolitanl metropolitan levels. The rural substate regions with the smallest Black populations showed the greatest variation between method‐estimated rates for Blacks. Conclusions: The results suggest that the smallest minority race categories are exceedingly susceptible to the impact of bridge method selection, especially in nonmetropolitan substate regions.