>Is it possible to set up Leverage to display certain fields dependent
>on the response to a previous field?

The simple answer is "no".  It is possible to define a Leverage db so
that the answers to fields (the values in a field's pop-up list) vary
depending on the values in other (previously filled in) fields.  I
believe it is possible to control which field opens next, depending on
answers, but it is not possible for fields to appear conditionally.

However,
1) Leverage is programmable, and
2) I can think of a way to do something similar (indistinguishable,
   perhaps) -- it fits in with,
3) my experience is that for surveys, audits, inspections, and similar
   tasks where there are many questions (scores or even hundreds) one's
   initial idea of how to set up the database is not the best way (at
   least not in Leverage).

I would (in your case -- it wouldn't be this complicated for the general
case of a survey without the conditional branching element, but it would
be similar) set up three or four databases.
Hmm.  Four.

I don't know exactly what you're doing, so I'm using a patient interview
as something we can perhaps both appreciate.

The first database would contain all the questions probably identified by
an index/number, one question per record, only two fields per record:

id  question
--  --------
1   sex
2   pregnant
3   month
4   any miscarriages
5   allergies
...

The next db would contain all the question/answer pairs, one pair per
record.
It would also indicate which question to ask next, for this answer:

id  question         answer     next question
--  ---------------  ---------  -------------
1   sex              male       5
1   sex              female     2
2   pregnant         yes        3
2   pregnant         no         4
3   month                       4
// blank answer means to go to next question for all answers
4   any miscarriages yes        ...
...

The third db would be the one actually active while the survey was being
taken. Perhaps curiously, it would have only a single record (this is
because of some limitations in the high level programming primitives
Leverage supports -- it's much handyier with lower level stuff).  Once
the record was filled it would be stuffed into the fourth and last
database, the one that would contain the actual survey results, and the
next question would be retrieved and displayed. (It might be best to
think of the third database as a one question "window" into the actual
survey.)  The structure of the third and fourth dbs is almost identical:

id  question      answer     next
--  ------------  ---------  ----
...

Of course, the values would vary as the survey was taken.  The "id"
field could be linked to the first db or one could always start with
question 1 or whatever. In any case, choosing a question id would fill
in the question text. This would only have to be done once at the
beginning.  After that the questions would always be fetched to match
the previous answer.

The answer field would be linked to the second db with a condition
limiting its pop-up list to the answers for the current question.  Once
an answer was either chosen from the list or written in (I can imagine a
refinement where text could be entered only if the "answer" field in the
second db was empty) the user would tap the "next" field.

This field would be programmed to log/copy the current record (i.e.,
question and answer) into the fourth db (which would lack the "next"
field, but would otherwise be identical to the third db) and pull the
next question from the second db based on the current answer.

Thus, if the answer to question 1 was "female" question 2 would be
displayed asking if the patient was pregnant.  If the answer to question
1 was "male" the next question would be question 5, etc.


Since this would make a good, generally useful sample for other users,
and would be an interesting exercise, I wouldn't mind setting it up for
you (perhaps using your structure, but probably not using all your
questions -- entering all your questions wouldn't be interesting).

The first db is not strictly necessary, but makes it much easier to
populate the second db if it is being done on the Newton.  If the
questions and answers are being imported from a desktop machine it's
less important, but doesn't do any harm beyond using a little extra
space.

Let me know.  Thanks for a stimulating question.

-- John

p.s.  In case it wasn't clear from my description, all this can be done
in Leverage as it comes out of the box -- no special tools or additional
software is required.  On the other hand, it probably would not appear
on a quiz in Leverage 101.