I am trying to apply the method proposed by Chib in [Marginal Likelihood from the Metropolis Hastings Output](http://www.economics.uci.edu/~ivan/ChibJeliazkov2001.pdf) to calculate the marginal likelihood of a logit model the includes latent variables. Specifically, 
$Pr(Y=1)=\exp\frac{\beta x+z}{1+\exp(\beta x + z)}$
where an unobserved variable $z \sim N(0,\sigma)$. 
I will be very thankful for tips on how to calculate the marginal likelihood of the model!