Human Subjects Review and Maintaince of Audio/Video Tapes

[Jeff ALLEN]

At 09:45 03/05/99 -0700, Lynn Santelmann wrote:
>I have recently applied for human subjects review at Portland State
>University 

Yes, Applied Ling dept, up one level and down the hall from Modern
Languages in Neuberger Hall.   I did my BA in French and MA in TESOL/App
Ling there.

I've heard a few stories about the Human Subjects review committee being
quite difficult at PSU.  You are not the first one from there to complain.
I finished my graduate work there nearly 10 years ago and remember comments
from grad student union meetings, etc. about dealing with Human Subjects
Review.
Granted, it is a safety factor for the subjects themselves, but when it
starts squelching valid research, that is not good.

>  and was surprised to receive a query from the committee as to
>"how long will the videos and records be stored and how and when will they
>be destroyed?" 
>
>I had carefully explained in my application that any publications of
>results or  transcripts would use only code names, so as to protect the
>privacy of the participants, and that the audio and video tapes were to be
>kept in locked cabinets in my office/lab. 

Watch out with such conditions.  In my data collection efforts of very
precious and
hard-to-come-by data (recorded data of 150 Haitian Creole speakers and of
300 Korean speakers for speech recognition systems), I have always kept the
digitally recorded data on one hard disk, with daily back-ups on 8mm tape
(back-ups done overnight).  I take one 8mm tape home each day, alternating
the tapes each day so that the one in my briefcase is yesterday's back-up
material.  8mm tapes are good for sound files since you can store 2- 5 Gb
of data on them, maybe more now.   Then you burn CDroms every so often as
additional back ups.  Granted, you are dealing with video tapes, which are
also subject to breaking, stretching, cracking, etc.    Multiples copies
are necessary for avoiding a major disaster to your data if Neuberger Hall
unfortunately goes up in smoke.

>To be honest, I hadn't considered
>destroying the original tapes, 

Either would I.    I still have all of the audio taped conversations of the
interviews for my doctoral work.   it is the only proof for verifying data.
 Once the tapes/data are gone, there is no way that you or anyone else can
check your work, from a scientific standpoint.

>and I don't want to be required to do so. I
>have often found it necessary to go back to the original tapes to clarify
>issues that come up later, or to answer questions that arise later (such as
>might stem from peer review).

Exactly.    

>This certainly touches on the issues of privacy and confidentiality that
>were discussed on this list recently, 

An interesting thread on "anonymity" in corpora studies and the
anonymisation of data appeared on the Corpora List recently. I have the
whole thread of messages if you are interested.  I also know different
folks working in corpus linguistics who specifically deal with this issue
all the time for the use of medical records and police records.  I can
refer you to them for how they have dealt with this in project proposals
for universities and govt-funded research grants.


>but at the same time, I find this
>request to be excessive, especially since I am not proposing to make the
>recordings public, except through transcripts.
>
>My questions for the community are: 
>1. Has anyone else run up against this kind of request to destroy the
>original tapes? 

>2. Any suggestions on how to respond to this committee?

Yes,

The data is for research purposes only.  There is no intent for public or
research dissemination (although you might want to be careful not getting
yourself to stuck with this).  State it so that you will not
disseminate/distribute the data without prior consent of the subjects, and
that non-disclosure statements would be signed by any parties that receive
the data.

Also carefully state that in the published results, you will follow
existing criteria for anonymisation.  The thread of messages on this topic
from Corpora will point you to more experts in this field.   

State that if there is any need for external intervention of other experts
with regard to this data, you will require them to sign a non-disclosure
statement.  It would help to have an example of it for the HSR committee to
look over as an accompanying annex to your reply.

Lastly, destroying the original data, including copies, is removing all
potential data for any scientific benchmarking, verification, and
validation issues.   

Any comments from other people on the list?

Jeff



=================================================
Jeff ALLEN - Directeur Technique
European Language Resources Association (ELRA)  &
European Language resources Distribution Agency (ELDA) 
(Agence Européenne de Distribution des Ressources Linguistiques)
55, rue Brillat-Savarin
75013   Paris   FRANCE
Tel: (+33) 1.43.13.33.33 - Fax: (+33) 1.43.13.33.30
mailto:jeff at elda.fr
http://www.icp.grenet.fr/ELRA/home.html