Christian Rudder is the president
of OkayCupid. He also is the “approver of
company research and its chief interpreter. “ He has been posting his findings from the
data he extracts on his blog. This is one of the ways that he has been driving
traffic to the site. Rudder is a staunch supporter of using the data extracted
from people’s browsing on the internet as a research tool (in other words, data
mining). For example, in the interview he states that “As a group, for
instance, Latino men rated Latinas as 13 percent more attractive than the
average for the site, while they rated African-American women 25 percent less
attractive. In fact, Mr. Rudder reports, black women on the site receive about
25 percent fewer first messages than other women do. ” He concludes that
this tells a story of racism. He says that “this chapter shows racism isn’t a
problem of outliers, it is pervasive.”
For the most part, I am okay with my browsing data
being used as data in some kind of social experiment. The only way I am okay
with this is if it is explicitly stated and completely anonymous. By explicit,
I do not mean that it’s hidden in the terms of service that we all just skip
through. I want it to be explicitly said in a pop-up of some kind that there’s
no way anyone could miss it. I am fond of what little privacy I have left, and
I would like to keep it that way. Privacy aside, I still do see a problem with
using these research statistics. The problem is simply this: correlation does
not always equal causation. These statistics cannot be taken at face value.
There must be some kind of further research to back it up outside of browsing statistics. For example,
saying that everyone on your website is racist because they are more attracted
to their own race more than others is a large claim. But this is also mentioned
in the article and acknowledged that this is most likely a poor example. The
article also brings up the issue that these statistics could be used to support
some kind of secondary claim, outside of the purpose of collecting the data.
It appears that way no matter what measures are taken.
We are going to be researched in some way, so to an extent I do support some of
Rudder’s views; especially that he wants companies to share even more of the
behavioral research findings with the public. This I can back up one hundred
percent, because if I am going to be inevitably researched, I would like to
know what it is going towards, but again this goes back to my issue of correlation
and causation. If these studies are going to be published and there is going to
be some kind of conclusion drawn, I think there must be some kind of other data
or study to back it up. If the company just releases the data and says here it
is, then fine. I think that the biggest issue these companies will be having is
that people are inevitable going to have a problem with, for example, Netflix
using your watch history to tell you about your daddy issues. It’s just not
that believable.
No comments:
Post a Comment