There are several ways. Which one you use depends on exactly what
you need. If you wish to run only the one analysis on
non-smokers, then this approach is a good one:
temporary.
select if not smoker.
UNIANOVA ...
The temporary selection of non-smokers is only in effect for one
procedure--in this case, your UNIANOVA. If you want the selection
of non-smokers to last longer, you could set a filter. E.g.,
compute f = not smoker.
filter by f.
Now whatever you do will be done using non-smokers only. To use
all cases again, turn off the filter:
filter off.
--
Bruce Weaver
bwe...@lakeheadu.ca
www.angelfire.com/wv/bwhomedir
"When all else fails, RTFM."
The question that comes up for me is why you would not include smoker as
a design factor in the analysis.
Art Kendall
Social Research Consultants