Five Parameters
Using data from governor candidates in the US between 1945 and 2012, we seek to predict the longevity of candidates in state-wide US races after election. There is concern over the assumption of stability because life expectancy has been increasing over time. We used a Bayesian gaussian regression model with sex and election age as predictors of the number of years lived after an election. The interaction between sex and election age is positive. As election age increases, the longevity difference between male and female governors widens. A quantity of interest is the difference in number of years alive after an election for male candidates compared to female candidates. Although it seems that males tend to live longer post-election, there are few female candidates leading to a wide range, which puts our approximation into question.
\[ lived\_after_i = \beta_0 + \beta_1 male_i + \beta_2 c\_election\_age_i + \\ \beta_3 male_i * c\_election\_age_i + \epsilon_i \]
Characteristic |
Beta |
95% CI 1 |
|---|---|---|
| sex | ||
| sexMale | 54 | 14, 95 |
| election_age | -0.05 | -0.72, 0.65 |
| sex * election_age | ||
| sexMale * election_age | -0.81 | -1.5, -0.13 |
| 1
CI = Credible Interval |
||