|
|
Running linear regression with listwise exclusion of missing cases and
with cases weighted by weighting variable for population
representativeness, and requesting full set of residuals and diagnostics
to be saved as variables to data editor results in data for all cases in
the analysis for stnd pred, stnd resid, Mahalanobis, and leverage, BUT
sysmis for some of the cases in the regression on DFBETAs, deleted resid,
cook's distance etc. Why should this happen? Can I do something to get
these diagnostics for all cases in the model? Thanks in advance.
|
Locked
