Selected article for: "linear regression and variance analysis"

Author: Chambless, L.
Title: Why do per capita COVID-19 Case Rates Differ Between U.S. States?
  • Cord-id: 09ty9tgk
  • Document date: 2020_10_20
  • ID: 09ty9tgk
    Snippet: Background: The popular press has explored the differences among U.S. states in rates of COVID-19 cases, mostly focusing on political party differences, and often mentioning that political party differences in health outcomes are confounded by demographic and socio-economic differences between Democratic areas and Republican areas. The purpose of this paper is to present a thorough analysis of these issues. Design and Methods: State-specific COVID-19 cases per 100,000 people was the main outcome
    Document: Background: The popular press has explored the differences among U.S. states in rates of COVID-19 cases, mostly focusing on political party differences, and often mentioning that political party differences in health outcomes are confounded by demographic and socio-economic differences between Democratic areas and Republican areas. The purpose of this paper is to present a thorough analysis of these issues. Design and Methods: State-specific COVID-19 cases per 100,000 people was the main outcome studied, with explanatory variables from Bureau of Census surveys, including percentages of the state population that were Hispanic, black, below poverty level, had at least a bachelor's degree, or were uninsured, along with median age, median income, population density, and degree of urbanization. We also included political party in power as an explanatory variable in multiple linear regression. The units of analysis in this study are the 50 U.S. states. Results: All explanatory variables were at least marginally statistically significantly associated with case rate in univariate regression analysis, except for population density and urbanization. All the census characteristics were at least marginally associated with party in power in one factor analysis of variance, except for percentage black. In a forward stepwise procedure in a multivariable model for case rate, percentages of the state population that were Hispanic or black, median age, median income, population density, and (residual) percentage poverty were retained as statistically significant and explained 62% of the variation between states in case rates. In a model with political party in power included, along with any additional variables that notably affected the adjusted association between party in power and case rate, 69% of the variance between states in case rates was explained, and adjusted case rates per 100,000 people were 2155 for states with Democratic governments, 2269 for states with mixed governments, and 2738 for Republican-led states. These estimates are based on data through October 8, 2020. Conclusions: U.S. state-specific demographic and socio-economic variables are strongly associated with the states' COVID-19 case rates, so must be considered in analysis of variation in case rates between the states. Adjusting for these factors, states with Democrats as the party in power have lower case rates than Republican-led states.

    Search related documents:
    Co phrase search for related documents
    • additional variation and local state: 1
    • adjacent state and local state: 1