...
> For example, would you expect two elements rated 'not applicable' on
> all constructs to be maximally close, or maximally distant?
Brian,
I've got it down on my "todo" list to parameterize the missing values in
WinGrid (ingrid99 for Windows) and you have given me some insight by
your use of the term "meta data".
Currently WinGrid handles missing values the way Patrick Slater and I
agreed back in 1980. Namely, the question mark that is entered into a
cell assumes the average value of the remainder of the cells for that
row. What I'm hearing from you is that there are at least three other
different types of missing value. I have normally experimented with
missing data in grids by doing a search and replace on several copies of
the grid, sometimes changing the question mark to a zero or sometimes
its opposite. But it would be nice to have several types within one
grid.
In referring to one of your terms, can I assume that if you wanted the
missing value of an element to be "maximally close" to the other
elements for a given construct would that be the same as pre-assigning
it the average of that construct's non missing elements?
Regards,
j.Maxwell Legg
PS. When I'm a bit closer I'll port your RepGrid and HTML file formats
to WinGrid to get a clearer idea of how you handle this.
PPS. I'm looking for a MacHead to take on the WinGrid source code and
bring it up as VBscript inside Microsoft's Office98 for the MacIntosh.
Do you by any chance have that software?
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%