confidentiality

[Kathy Hirsh-Pasek]

Brian,  I have been working on the NICHD Study of Early Child Care and some
of these issues have come up in our discussions.  We have a huge database on
children 0 - 3 years that will be made public next January.  My sense of the
group discussion is to take confidentiality very seriously.  With respect to
your levels, that would translate to at least your restrictive level 5.
Yet, I wonder whose responsibility it is to release data -- the researcher's
or the subject's. In your write up of these levels,  you suggest that it is
the researcher's.  Rather, I think that the only one who can grant
permission to use data publicly (or to academicians) is the subject,
through a consent form.  If we do want to use this data more widely, than we
all have to review the consent forms that go to our IRBs.  And, as we have
repeatedly found in the NICHD Study, opinions on these matters vary widely
across  IRBs from different universities.   Only if we explicitly ask
whether data can be shared with others, should we have the power to release
it.  My vote, therefor is to err on the side of caution.   Kathy Hirsh-Pasek
----------
>From: Brian MacWhinney <macw at cmu.edu>
>To: info-childes at childes.psy.cmu.edu
>Subject: confidentiality
>Date: Thu, Feb 18, 1999, 12:19 AM
>

>Dear Info-CHILDES,
>  As we move into the era of video linked to annotations and
>transcriptions, issues of confidentiality become an increasing concern.
>People who would be happy to donate their transcript data to CHILDES
>might have serious second thoughts about donating the related audio or
>video data.  How can we deal with legitimates concerns about speaker
>confidentiality and still maintain international scientific
>collaboration for the study of verbal interaction?
>   I would like to propose an approach that focuses on levels of
>confidentiality with the strictest level being no access at all and the
>the loosest level being full access.  In particular, I think we could
>distinguish 6 major levels.  I would like to get people's comments on
>this idea and whether it would work to successfully address the
>confidentiality issue.  Feel free to think in terms of all sorts of
>perspectives, including scholars, subjects, government officials,
>citizen advocates, lawyers, humanists, and the like.  Please post your
>comments directly to info-childes, unless you think it is not
>appropriate to do so.  Does this proposal succeed in "solving the
>problem".  Do we need additional mechanisms?   Here is the specific
>proposal:
>
>Level 1:  Data would be fully public domain (CNN, public speeches,
>public interviews, etc.) and generally viewable and copyable over the
>Internet.
>Level 2: Placing data on this level would open general viewing and
>listening to the public across the Internet, but would block copying.
>Level 3:  This level would restrict access to academic researchers who
>had signed a non-disclosure form.  This form would set tight standards
>regarding avoidance of use of personal names when required.  It would
>allow some temporary copying or downloading of the data for local
>analysis, but would require that downloaded files be deleted after a
>specific period and never further copied or distributed.
>Level 4:  This level would restrict access to academics who had signed
>non-disclosure forms.  In addition, it would totally disallow copying.
>Level 5:  Data on this level could be viewed only after the original
>data collector had given approval over the Internet for the particular
>researcher.
>Level 6:  This level would only allow viewing and listening in
>controlled conditions under the direct on-line supervision of the
>particular researcher.
>Level 7:  This level would only allow viewing and listening in
>controlled conditions under the direct, in person, supervision of the
>particular researcher.
>Level 8:  These data would not be viewable, but would be
>archived in the format of the general system for use by the original
>investigator only.
>
>I wonder if this level system would not only work to maintain
>confidentiality, but also to support the "legitimate interests of the
>original data collector."
>
>Please comment on this important issue.  If you can think of other fora
>for discussing this issue, that would be good to mention too.
>
>--Brian MacWhinney
>
>