Rstanarm joint model 95% CI

Modeling
1

In using Joint model, and found some inconsistence in the result output for mean mean/sd for 95% CI.

For example, in the example data, etavalue for

Long1|etavalue   1.341     0.240       3.823 

However, when use command summary(mode, probs=c(.025, 0.975))

the results show as:                              mean            sd      2.5%       97.5%
Assoc|Long1|etavalue                           1.353       0.241    0.933        1.849  
Assoc|Long1|etavalue

manually 95% CI calculation, 2.5%: 1.353 - 1.960.241 = 0.88, and 97.5%: 1.353 + 1.960.241 = 1.825

I am not quite sure this is correct or not? Could you please explain it how the rstanarm package to calculate 95% CI?

Thanks in advance.

Please share your Stan program and accompanying data if possible.

When including Stan code in your post it really helps if you make it as readable as possible by using Stan code chunks (```stan) with clear spacing and indentation. For example, use

model {
  vector[N] mu = alpha + beta * x;
  y ~ normal(mu, sigma); 
}

instead of

model{
vector[N] mu = alpha+beta*x;
y~normal(mu,sigma);
}

To include mathematical notation in your post put LaTeX syntax between two $ symbols, e.g.,
p(\theta | y) \propto p(\theta) p(y | \theta).

2

It does not calculate a credible interval as mean \mp 1.96 * sd because it is not a confidence interval. Rather, the lower boundary is calculated as the value such that the proportion of posterior draws less than it is 2.5% and similarly the upper boundary is calculated as the value such that the proportion of posterior draws less than it is 97.5%. It should match what is produced by posterior_intervals with prob = 0.95. However, the endpoints of 95% credible intervals are not estimated particularly precisely with the default settings, so you should consider shorter credible intervals and / or increasing the number of post-warmup iterations.