Re: Off diagonals in $TABLE

Hi Paul, One cannot have a correlation based on a single individual, so there is no such number that could be output in the TABLE file. However, if you are interested in the correlation between EBE of etas (across all subjects in your analysis data) you can just output the etas using TABLE, then read into R, take one row per subject (assuming this is IIV, no IOV), and calculate the correlation, e.g. between ETA1 and ETA2. For most eta diagnostics/processing I would suggest you set any etas that are exactly zero to NA. Otherwise, you will have misleadingly lower magnitude of correlation, e.g. because half you the subjects received placebo and have the eta on Emax as exactly zero. Likewise, one would get higher shrinkage for this eta if including these subjects with zero eta value, and for most purposes that is not how I would prefer to use the term shrinkage. But there are exceptions so I do not mean to start a separate discussion on this - sometimes it may be useful to include also the zero etas in calculations, graphs, etc. Or did you mean the uncertainty in the individual EBEs, and the correlation between these (EBE) estimates? In that case the above is irrelevant, but I do not think that is what you mean. Best regards Jakob Jakob Ribbing, Ph.D. Senior Consultant, Pharmetheus AB Cell/Mobile: +46 (0)70 514 33 77 [email protected] www.pharmetheus.com Phone, Office: +46 (0)18 513 328 Uppsala Science Park, Dag Hammarskjölds väg 36B SE-752 37 Uppsala, Sweden