[Air-L] on the Wayback Machine (was public/private [part 1 of 2])
Lois Ann Scheidt
lscheidt at indiana.edu
Tue Aug 14 07:03:48 PDT 2007
> But to complicate your argument below (and Hughie's in a separate
> thread), a couple of thought experiments:
>
> Say I place content on a "publicly accessible" webpage without
> creating any incoming links or notifying anyone. Web crawlers won't
> find it. A search engine won't index it. While on the open and public
> Internet, unless a random URL-generator happens to guess the precise
> address of the page, no one will ever read it. Is this content "fair
> game for researchers"?
I don't profess to be a "hardware" or a "search engine" expert, but I
can tell you that the above thought experiment only works if you have
complete control of the hardware and software involved in the webpage.
Many hosts webpages link to pages of folks hosted by their service.
It's a term and condition of service...so it's not nearly as easy to be
completely off the search engines as one may think.
Beyond that point, for those of us that do "in the wild" research...how
would we find your thought experiment if it's not indexed? I guess it
would be possible to have a crawler that would randomly generate URLs
and check their validity. That might be useful for some research
question, but in that case I would assume the unit of analysis would be
the URL, at most, and more likely it would be a proof of concept test
of some kind - so no human subject would be involved.
Now beyond that, I would argue that from a research perspective the
blog was still open to study, if it can be found. I say that because
the best terrestrial analogy I can think of to your thought experiment
is someone who posted a broadsheet on a pole but with a cover over the
broadsheet...lots of things could pull that cover off...the wind, rain,
or a passing person who didn't like the color of the cover. In short,
a person who would go to that much trouble to hide in plan sight...well
it's still plan sight.
I think that's part of why this "what if" game gets bogus, a person who
would make an effort to do the above, isn't all that likely to be
"real" and I study real people. "What if's" are valuable in these
discussions but again we have to remember that we can dream up far more
elaborate scenarios then we are ever likely to see in the wild.
> Consider a different scenario: Walking down the street with a friend,
> I comment on the fact my rent check bounced, and she offers
> condolences. We're in public. We make the utterances loud enough that
> someone next to us could easily hear it, but not loud enough that
> someone across the street can. Now, this is a public utterance - we
> didn't have the conversation behind locked (password-protected)
> doors. Can this be archived for research? Better yet, can we setup
> microphones to automatically record every conversation uttered in
> public? Does the fact these conversations *can* be recorded and
> archived mean they, by default, *should be*? I feel we're saying the
> same about utterances on the Web - so what's the fundamental
> difference from a user perspective between the street and the Web?
> Both are a place where we engage in conversations and maintain
> relationships....
>
> I fear we're confusing what the Internet *is* and *is capable of*
> with *how it is used*. Yes, comments left on a personal blog are open
> for anyone to see. Yes, discussion board conversations can be
> archived. But that's not, I suspect, going through the mind of many
> casual users of this technology. It is a medium for communication,
> for connecting with people. I leave messages on my neighborhood
> parents discussion board because that's how I connect with that
> community. I use my real name because I'm among friends. Does that
> mean my comments are de facto fair game for any researcher who wants
> to scrape the database?
Actually I think the "confusion" is at a different level. Your second
thought experiment gets at it fairly well. When I am designing
research I can only know the "potential" harms, not the actual
ones...potentials are before the fact, actuals are retrospective.
Let's change your experiment a bit. You are home and in the shower
getting ready to meet your friend. You think about the stuff you are
going to tell them and you decide to mention your bounced check. And
maybe it crosses your mind that you need to be careful where you say
that so fewer people over hear.
When you decide to tell your friend you don't know who is really
around...or what someone who overhears might do with the utterance.
You only know that you are going to do your best to make sure that you
are careful.
Same with research design and human subjects work.
We have to learn from actual happenings but if we sit and do
statistically, oh my a qualitative researcher just used the s-word,
improbable "what-if's" then no research will happen at all.
Personally I believe the following -
- Research is never risk-free. Neither is life. If you are working
toward absolute zero risk in your research you will either give up the
pursuit, or build a protocal for your research that lies to you. Yes
there are ways to protect YOUR subjects but most of them can "endanger"
others in the process (more on this below). Merry Christmas that is
not zero risk, it's just minimal risk for the research and probably a
heck of a lot of your work that didn't buy anyone more "safety."
All of the aforementioned "harm" and "protective" terms were used only
as part of the discussion, because what is the real chance of harm from
most of our research, one-in-a-million, one-in-a-billion,
one-in-a-trillion? Medical experiments don't get to absolute zero
risk...they just shift the potential risk - which may be minimal as
well - from the researcher to the patient...unless something unexpected
happens and then the risk is back on the researcher. And how do they
"know" the potential risks? From computer simulations - based on facts
known from old (and often questionable) human experiments, and/or from
animal trials. So their "potential" risk knowledge comes from
analyzing actual risk after a trial, not from "what if's."
- I design research...I predict the potential harm from my
"intervention" and I work to minimize the risk ("minimal risk" is the
operative term). I do not mind read or assume my position as
researcher places me in a superior position and that my subjects need
special protections from me.
Sorry folks but I truly find that last concept to be pretty funny. I
am not injecting people with foreign substances, I look at artifacts
they produced and placed online. Personally, I have never had a
artifact's author/artist come back to me after I published and
complain. I have had a single author question my use of his work
without permission - I work with teens remember - and so we talked and
by the end of the discussion he understood and told me he was pleased
his work was in the paper he just thought he HAD to give permission
first. Plus he found out by tracking his visitor logs and finding I
had visited repeatedly, so it was not because of an unpleasant
experience on his part.
Now before I sound like a totally heartless researcher who only sees
her subjects as things...when nothing could be further from the
truth...if even one of the kids I work with were hurt by my research I
would have a crisis of conscience. But again I am not God, I know that
if I had done the best job I could do working through my research
design and trying to predict all the potential harms likely to happen
with my intervention then I have done everything I can do.
Some years ago, I wrote a classroom paper that discussed why I don't
pseudonymize my subjects in my chatroom research. The simple reason is
that there is no way I could come up with nicknames that would cover
the participants and not potentially deflect on to another chatroom
user. I'm good but I'm not THAT good.
So if I did pseudonymize I might protect my participants but in that
one-in-a-billion case where someone was in harm's way from something I
did...I might well have created the situation rather then hidden it.
I'll let you "what if'ers" out there run that one through the filters.
Lois
More information about the Air-L
mailing list