County characteristics affecting registration share
Regression analysis
When analyzing county level factors, we limited data to September, 2018.
Our final model included the following variables: organ procurement organization, percent of population enrolled in Medicare, percent of county that is males, percent of county that is White, percent of adults over 25 with a high school diploma, percent of adults over the age of 25 that have a college degree, and an interaction between the percent of the county that is white and the percent of adults over 25 with a high school diploma.
| Term | Estimate | Std. Error | Statistic | p-value |
|---|---|---|---|---|
| (Intercept) | -118.033 | 29.318 | -4.026 | 0.000 |
| opoCenter for Donation and Transplant in New York | 8.206 | 1.893 | 4.334 | 0.000 |
| opoFinger Lakes Donor Recovery Network | 13.194 | 1.932 | 6.830 | 0.000 |
| opoUNYTS | 13.706 | 2.336 | 5.867 | 0.000 |
| percent_medcr_2015 | 0.732 | 0.238 | 3.072 | 0.003 |
| percent_male_2015 | 1.620 | 0.422 | 3.835 | 0.000 |
| percent_white_2015 | 0.661 | 0.162 | 4.067 | 0.000 |
| percent_hs_diploma | 2.156 | 0.659 | 3.274 | 0.002 |
| percent_college | 0.381 | 0.110 | 3.463 | 0.001 |
| percent_white_2015:percent_hs_diploma | -0.034 | 0.008 | -4.270 | 0.000 |
We found that type of OPO that conducts the registration affects registration rate. Percentage of medicare enrollment, percentage of persons finishing high school education and finishing 4-year college education are positively associated with registration rate. This finding agrees with our hypothesis. Also, the sex male and race white are positively associated with registration rate.
Here we compare the candidate models:
| Model | Adjusted r-squared | AIC |
|---|---|---|
| Final model | 0.81 | 331.97 |
| Model 1 | 0.76 | 345.46 |
| Model 2 | 0.74 | 348.62 |
A P8105 Final Project by Nick , Jack , SiJia , Alina and Shu