TABLE 4—
Regression Estimates for Characteristics Associated With the Availability and Perceived Effectiveness of Public Health Activities
| Availability of Activitiesa | Perceived Effectiveness of Activitiesa | |||||
| Characteristic | Coefficient | SE | Relative Effectb | Coefficient | SE | Relative Effectb |
| Population size (log) | 0.0606 | 0.0298** | 0.0539 | 0.0346 | ||
| Population density (per square mile) | −0.0000389 | 0.0000378 | −0.0000227 | 0.0000438 | ||
| Percentage of population below poverty level | −0.1088 | 0.0372*** | −0.0989 | 0.0431** | ||
| Percentage of population non-White | −0.00214 | 0.00146 | −0.00347 | 0.00169** | ||
| Hospital beds per 100 000 residents | 0.0000067 | 0.0000283 | 0.0000201 | 0.0000328 | ||
| Physicians per 100 000 residents | 0.00096 | 0.00330 | 0.00098 | 0.00382 | ||
| Local health department spending per capita (log) | 0.0672 | 0.0254*** | 0.0559 | 0.0294* | ||
| State–local administrative relationship (0, 1) | ||||||
| Centralized authority | −0.1377 | 0.0609** | −0.130 | −0.1144 | 0.0705 | −0.110 |
| Decentralized authority | −0.0326 | 0.0506 | −0.0333 | −0.0244 | 0.0587 | −0.0258 |
| Reference: shared or mixed authority | ||||||
| Type of governmental jurisdiction (0, 1) | ||||||
| County or joint city/county | 0.0207 | 0.0564 | 0.0193 | 0.0243 | 0.0653 | 0.0224 |
| City or township | 0.0954 | 0.0909 | 0.0955 | 0.115 | 0.107 | 0.116 |
| Reference: multicounty or regional | ||||||
| Local board of health with policy authority (0, 1) | 0.0945 | 0.0445** | 0.0980 | 0.136 | 0.0518*** | 0.144 |
| Categorical services offered by health department (0, 1) | ||||||
| Comprehensive primary care services | 0.0234 | 0.0392 | 0.0229 | 0.0666 | 0.0455 | 0.0678 |
| Environmental health services | −0.0243 | 0.0450 | −0.0250 | −0.0252 | 0.0523 | −0.0262 |
| Intercept | 2.765 | 0.393*** | 2.281 | 0.456*** | ||
| Adjusted R2 | 0.17 | 0.14 | ||||
| No. | 352 | 352 | ||||
aThe dependent variable for each regression equation was transformed via the natural logarithm to reduce skewness and the influence of outlier values.
bFor each dichotomous variable, the coefficient estimate was transformed into a relative effect indicating the percentage change in the dependent variable associated with a change in the dichotomous variable from 0 to 1. The transformation method provided by Kennedy was used for this purpose.50
*P < .10; **P < .05; ***P < .01.