MSU rurality index: Development and evaluation

Authors

  • Dr. Clarann Weinert SC, PhD, RN,

    Associate Professor, Corresponding author
    1. Department of Mathematical Sciences, Montana State University
    • College of Nursing, Montana State University, Bozeman, MT 59717
    Search for more papers by this author
  • Robert J. Boik PhD

    Associate Professor of Statistics
    1. College of Nursing, Department of Mathematical Sciences, Montana State University
    Search for more papers by this author

Abstract

The construction and validation of a residence-based index of the degree of rurality are reported. The MSU Rurality Index is a locally normed index and assigns a quantitative measure of rurality to each participant in a study. Data required to compute the index are minimal; only two variables need be measured. Statistical analysis effort required to compute the index is somewhat more intensive than usual, but this tradeoff between data collection and data analysis seems sensible. Data collection is generally more expensive than data analysis. The two required variables, county population and distance to emergency care, are optimally transformed to achieve normality and weighted to achieve validity. A measure of departure from normality is automatically obtained while constructing the index. A positive score reflects a rural residence and a negative score reflects an urban residence relative to the group under study. A score of zero reflects average rurality. Reliability and construct validity were examined using two data sets. The MSU Rurality Index has high test–retest reliability, is appropriately related to a set of ordered categories concerning place of residence, and is as or more strongly related to these categories than alternative indices computed from distance or population alone. The validity of the MSU Rurality Index does not appear to be compromised by its parsimony. Even though the MSU Rurality Index is based on only two variables, it was found to be associated with various health care variables as strongly or more strongly than an 11 variable county-based index developed in Texas. ©1995 John Wiley & Sons, Inc.

Ancillary