To assess the impact of parameter fungibility, we need to compute group-level correlations (only in the latent-trait model).
Technically, we need to extract the samples for the group-level means (mean=pnorm(mu)) separately per between-subject group and then compute the correlation across samples:
To assess the impact of parameter fungibility, we need to compute group-level correlations (only in the latent-trait model).
Technically, we need to extract the samples for the group-level means (
mean=pnorm(mu)
) separately per between-subject group and then compute the correlation across samples:cor(mcmc_mean)