Characteristic |
Beta |
95% CI 1 |
|---|---|---|
| sex | ||
| sexMale | 53 | 9.9, 97 |
| election_age | -0.06 | -0.79, 0.66 |
| sex * election_age | ||
| sexMale * election_age | -0.79 | -1.5, -0.07 |
| 1
CI = Credible Interval |
||
Five Parameters
Using data from all deceased gubernatorial candidates in the United States from elections held between 1945 and 2012, we seek to forecast candidate longevity in state-wide US races post-election. There is concern that longevity for gubernatorial candidates will differ significantly from that for candidates in Senate and other state-wide elections. We are using a Bayesian regression model with the formula lived_after ~ sex * election_age to analyze the data, where lived_after is the dependent variable representing post-election life expectancy. The model reveals that the interaction between sex and election age has a positive direction, suggesting that as election age increases, the life expectancy difference between male and female governors tends to widen. A Quantity of Interest for this question is the difference in the number of years that the average male candidate lives after their election and those of an average female candidate. However, we don’t know if our approximation of average female candidates would be accurate, as our data tends to have fewer female candidates, all of whom are spread across a wide range of longevity.
\[ lived\_after_i = \beta_0 + \beta_1 male_i + \beta_2 c\_election\_age_i + \\ \beta_3 male_i * c\_election\_age_i + \epsilon_i \]
