Hi, The attached pdf file shows the bizarre translation of DV var levels into the histogram graph. All the 10 continuous levels of the dependent variable soctrust are occupied, but the histogram shows unoccupied classes. The diagram is misleading to me. I then repeated the example with three three-groups and got the same pattern.
To me this looks like a bug. I work with the latest version dated 26 May, 2015, under windows. I can send you the data file, if necessary, though it is big. Regards, ftr
output5.pdf
Description: Adobe PDF document
_______________________________________________ Pspp-users mailing list Pspp-users@gnu.org https://lists.gnu.org/mailman/listinfo/pspp-users