Forum: eRm 0.9.1 is now available on CRAN
- Package "sna" not included anymore since it is not checkable within reasonable time. Thus, only a relevant C-routine is picked out to check for ill-conditioned data.
- If complete 0/full responses: First items are deleted and then it is checked for ill-conditioning.
- Ill-conditioning check for LLTM added (for each measurement point separately).
- If models with more mesurement points and items have 0/full response, estimation is stopped and user has to delete the corresponding items within the other measurement points separately.
- Waldtest improved: Critical items (0/full responses) are deleted for all subgroups such that the test can be carried out for the remaining items.
- Bug in "person.parameter" and "Waldtest" removed.
- In "LRtest" the eta's and their standard errors are returned. In addition, a list of the subgroup data matrices is returned.
- Check for RSM whether the number of categories are the same for each item.
- Check for PCM if there are missing response categories.
- Allowing for the specification of a vector with starting values for the eta parameters.
- Check if there are persons which have been presented one item only (the remaining are NA's)
- Return table of coefficients in Waldtest
- Person labels added in personfit
- Print out NA patterns S3methods for objects of class ppar