Contents - Index

MCMC Step Size


Default SD of normal jumping distribution.  For beta parameters that are not a part of hyperdistributions, the default step size to generate the next value of the parameter is generated as a random normal variable with a SD specified in this edit box.  Typically, you want to accept the new parameter value about 45% of the time, so you can adjust this SD to approximately obtain this acceptance rate.  For the hyperdistribution parameters, means are also sampled with a normal jumping distribution where the standard deviation sets the potential size of the changes.  For the standard deviation parameter of a hyperdistribution, the distribution of tau = 1/sigma^2 is sampled on a log scale with a random jump sampled from a uniform distribution.

This input is required as part of the MCMC estimation procedure.  However, the MCMC algorithm has been modified to adaptively modify the step size so that new parameter values are accepted about 45% of the time.  As a result, the value of the step size specified is just used as the initial step size, and is updated as the Markov chain progresses.