To save the “solutions” i.e. the point estimates of location parameters, you need one or two options in the parameter file. The behavior is common to the latest version of GIBBSF90+ (GIBBS2F90/GIBBS3F90/THRGIBBS1F90). Please use either of the following options.

• Simultaneous sampling with variance components:
  • OPTION solution [effects..]
• Compute the “solutions” with fixed variance components i.e., no (co)variance estimation:
  • OPTION fixed_var [effects..]

For example, you save the posterior mean and SD for effect numbers 1, 3, and 4, use OPTION solution 1 3 4. If you want to save all the effects, please enumerate all effect numbers.

There are other options to save *all* the solutions in all rounds: OPTION solution all or OPTION fixed_var all. It will create a huge file and will be very slow. Do not use these options unless you really want to have all the solutions for all rounds and/or you already know the number of burn-in samples.