W3C home > Mailing lists > Public > www-forms@w3.org > February 2002

RE: refusing to answer required questions

From: Micah Dubinko <MDubinko@cardiff.com>
Date: Fri, 1 Feb 2002 12:28:31 -0800
Message-ID: <E840F0B7E6189547BDB91DA8BF2228AB28C1C9@csmail.cardiff.com>
To: "'Thomas M. White, MD'" <tw176@columbia.edu>, www-forms@w3.org
Hi Thomas,

The Working Group has solicited input from a number of experts in the  areas
of research and professional survey/polling areas.

The general direction we've taken is for a general-purpose solution that can
be used to build more specialized tools.

As a specific example, I believe the scenario you outlined below could be
constructed out of the existing components in XForms.

For instance, the XML instance data to capture the input data could have
separate 'slots' for the user choice-from-the-list and reported answer...

<research>
...
  <question number="4">
    <userChoice/>
    <reportedAnswer>N/A</reportedAnswer>
  </question>
...
</research>

This represents a question where the user didn't select a predefined choice,
but rather chose the option "not applicable".

Thus, an element <userChoice>, mapped directly to a list, is not required,
but all the elements <reportedAnswer> are required. This could be done with
a single line...

<bind ref="//reportedAnswer" required="true"/>

Another option is to use an XML Schema union, as you suggested.

For the user interface, a form control <selectOne> could present the
specific choices for that question, while other user interface, say <button>
with a <setValue> XForms Action attached, could ensure the reportedAnswer
gets populated.

Thank you for your time. This message is just a personal opinion, not a
formal response to your Last Call message. If this sounds agreeable,
however, let us know.

Thanks!

.micah


-----Original Message-----
From: Thomas M. White, MD [mailto:tw176@columbia.edu]
Sent: Friday, February 01, 2002 11:00 AM
To: www-forms@w3.org
Cc: www-forms-editor@w3.org
Subject: refusing to answer required questions


Hi- 
I'm new to this forum, but I've been addressing similar concerns within the
realm of psychiatric and medical informatics via our own approach to
implementing forms. 
Paul Sagi's concern re privacy is especially true for forms and interviews
within medicine,  psychiatry, and epidemiology research.  In each of these,
people need to be given the opportunity to refuse to answer questions, even
if they are "required" by the form or underlying data schema. 
We have found that, in general, all required elements might need to support
these additional options for "not answering" a question:  (1) refused (e.g.
for people who refuse to answer a sensitive question), (2) don't know (e.g.,
"do you have a family history of <rare disease X>"), (3) huh? (e.g. the
person doesn't understand the question), (4) NA (e.g. if the question isn't
applicable for the person, even though the programming logic suggests that
it is -- an error in the logic).  Survey / polling research often also needs
(5) no opinion.  We have associated each of these with separate comment
fields so that if any of these options are selected, the user is able to
optionally specify their reason for refusing, etc.  In fact, the committees
that oversee research often require that subjects be allowed to refuse to
answer questions. 
From an XForms perspective, I'm not sure how this would be addressed.
XML-Schema can be used to generate <union> structures which support these
additionally allowable responses for each data element, but that seems
potentially burdensome upon the authors.  Moreover, these exceptional
answers should really have distinct user interface features from the main
options (for example, we implement them as buttons to the right of each
question; and the visibility of each button is determined by the privilege
level of the users -- e.g. interviewer vs. interviewee).  Thus, not only is
this a compound data type, but it is a compound user interface element with
refuse, etc. buttons whose visibility are controlled separately from the
visibility of the main data element. 
Can this type of functionality be supported by XForms?  If not, I recommend
that it be added, otherwise XForms might have limited use within medical
research settings. 
/Tom 
-- 
Thomas M. White, MD, MS 
Assistant Director 
Bureau of Evidence Based Medicine and Practice Guidelines 
New York Sate Office of Mental Health 
330 Fifth Ave, 9th Floor 
New York, NY  10001 
212-330-6358 (T,W,Th) 
212-543-6724 (M,F) 
  Get your free Australian email account at http://www.start.com.au
Received on Friday, 1 February 2002 15:31:21 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 10 March 2012 06:21:50 GMT