Table 1.
Parameter | Mean | Median | 95% CI |
---|---|---|---|
(a) | |||
β0 | 0.79 | 0.81 | 0.13, 1.43 |
β1 | −3.24 | −3.25 | −4.71, −1.78 |
τ | 0.91 | 0.89 | 0.58, 1.30 |
(b) | |||
β0 | 5.04 | 5.02 | 3.52, 6.57 |
β1 | −7.65 | −7.64 | −14.52, −1.00 |
β2 | 1.12 | 1.16 | −5.09, 7.64 |
τ | 0.29 | 0.32 | 0.23, 0.36 |
Coefficients in (a) are for a normal linear-model fit and in (b) for a second-order polynomial fit using a log-Normal distribution; τ is the posterior precision; CIs are Bayesian posterior 95% credible intervals; β values not overlapping zero provide strong evidence of a positive or negative relationship.