[Air-L] Are Twitter Handles identifiable information? (eProtocol section 13)

[Ye Na Lee]

Dear member of AoIR
I am currently going through IRB process (non-exempt) for my research which
involves analyzing tweets containing certain hashtags. In my opinion,
Twitter handles do not constitute as identifiable participant information.
However, I got a response from IRB, saying:

"Twitter handles may or may not be identifiable, depending on the user. In
some cases they are not identifiable, in many other cases they are
identifiable (i.e., twitter users' identities are known). Names and emails
are not the only variable that are considered "identifiable." Any
information that can lead to identification of an individual is considered
"identifiable." (See FAQ for further information:
http://cphs.berkeley.edu/faqs.html#e1). Please revise section 13a to
account for this. Ensure that corresponding updates are also integrated
into pertinent sections/documents."

Which means I need to remove identifiers, Twitter handles in this case from
tweets as soon as I collect them like item 13d says.

"Identifiers should be removed from data/specimens as soon as possible
following collection, except in cases where the identifiers are embedded
(e.g., voices in audio or faces in video recordings). If data are coded in
order to retain a link between the data and identifiable information,
explain where the key to the code will be stored, how it will be protected,
who will have access to it, and when it will be destroyed."

This will be very cumbersome for me because my research is not quantitative
and does not involve collecting massive amounts of tweets at once using
Twitter API. It is a qualitative research using virtual ethnography to
explore and understand the conversation around a certain hashtag movement,
which means I will need to constantly go back to certain Twitter accounts
and use snowball sampling to collect tweets like searching who they are
following and looking into these accounts as well.
(I plan to use NVivo 10 to collect and code tweets)

I assume IRB is taking more precaution about the security issues because my
research involves the risk of cyberbullying. I understand the risk and will
completely anonymize Twitter handles on my paper. However, I honestly do
not see the need to separate/remove identifiers in the process of
collection. I`d really appreciate if anyone who`s done a similar research
as mine could tell me about their experiences with the IRB process.
Thank you in advance.