Change Request: allow non-integer sampling weights ("MULT")

[Bill Raynor]

The current implementation appears to only allow integer values for the MULT variable ( TETRAD crashes if you give non-integer values).

 

Please adjust that so that we can add non-integral weights.

 

Why: Those of us who deal with real-world (sample) data often end up with sampling weights, or shrinkage estimates, or (proper) priors. These all show up as data with non integer weights.