In the last post, regression models for school funding and men’s and women’s income were presented as predictors for COVID case mortality (COVID deaths/COVID cases). A Facebook reader asked me to explain the models and measures. Case Mortality is an estimate of the risk of dying if one catches COVID. The regression model estimates how changes in a predictor variable correspond to changes in Case Mortality. Models with median household income, child care availability rate, traffic volume and the percentage of home ownership at the county level as predictors of COVID case mortality for the 10 county region.
Median Household Income
Overall, median household income was a weaker predictor of COVID case mortality than men’s or women’s median income levels accounting for 41.4% of the variability in case mortality. Conversely, men’s median income accounted for 52.5% and women’s median income accounted for 64.9% respectively. The regression model states that for every $10,000 increase in median household income, there is a predicted 0.005 decrease in the risk of death is one catches COVID. The graph for women’s earnings as a predictor of COVID case mortality is presented below. It shows that the counties are more tightly clustered around the regression line than they are for median household income,.
Child Care Center Availability
County child care center availability is the number of centers per 1,000 children under age 5 in the county. Childcare is a negative predictor of COVID case mortality. This association accounts for 50.3% of the variability in case mortality. Every unit increase in the child care center rate predicts a 0.00158 decrease in the case mortality rate. Cambria County is an outlier with a high center rate and a high mortality rate. The child care center rate was positively associated with county level vaccination rates.
Traffic Volume
County Health Rankings defines traffic volume as the average traffic amount on major roadways per meter in the county. It is a negative predictor of COVID Case Mortality accounting for 41.2% of the variability. For every 100 fold increase in traffic rate, there is a predicted 0.003 decrease in COVID case mortality.
% Homeownership
The percentage of housing units owned by the occupants is a positive predictor of COVID case mortality accounting for 48.2% of the variability. For every 10% increase in ownership, there is a predicted 0.005 increase in the COVID case mortality rate. This appears to contradict the relationship with income and mortality. Income and home ownership are not significantly correlated for these counties. Counties with high home ownership and low incomes may have lower property values.