Compare distributions of parameter estimates across chains

[danlwarren]

Something Rob brought up in conversation: it would be nice to have some way of comparing distributions of parameter estimates across chains to see whether they're significantly different.