John,

I think the software issue is less of an issue than the conceptualisation.

Regarding the items (below) that you detailed, it is hard to see how they
could be used in a grid:

1. decreasd productivity of smokers
2. potential medical benefits
3. State's rights versus federalism
4. right to privacy
5. jeopardizing minors
6. tax revenue
7. respiratory problems and increased cancer risk
8. driving safety
9. "gateway" drugs
10. less destructive than alternative "recreational" drugs such as alcohol

For example, if they were elements - try (a) picking groups of three and
asking how 2 are alike and different from the third, then (b) rating the
items above on the resulting constructs. It is even more difficult to
consider 1-10 as constructs, I suspect people would find some of these 'as
constructs' very difficult to apply.

One approach would be include 'legalising cannabis' as one of several
topical behaviours (including other drugs) in a grid and use these broader
topics as elements. Another approach, would be to use as elements people
who do certain things, and see how they are construed -
e.g a cannabis user, a drink driver, a male homosexual, an atheist, a
businessman, a cigarette smoker, a paedophile, graffiti 'artist', a
prostitute - you get the picture

Alternatively, you might seek to collect data that is more qualitative in
nature.


Bob