Experiments, Data, and the Scientific Ecosystem.

On the Facebook ethics discussion and noisemaking. 

Today, Michael Bernstein posted a bit of a discussion on the Facebook ethics and their PNAS article. He called for social computing researchers to shout a bit as the “silence is incredibly dangerous” but actually the conversation is happening in a few facets.

In fact, I find the explosion of conversation itself curious. We have seen it already. Target knows you’re pregnant from your shopping. Hospitals know you’re sick by your credit card. And Facebook knows the contagion of your posts. The counter here is Target doesn’t try to make you pregnant…or does it? Advertising and marketing can make us do a lot of things but Target, Visa, and the like don’t publish their findings in scientific journals.

The subject domain is of no real surprise here either but still generates discussion. In 2010, Adam published “An unobtrusive behavioral model of ‘gross national happiness’.In 2011, the American Psychological Association said in a post, “Social psychologist Adam D.I. Kramer, PhD, has an enviable subject pool: the world’s roughly 500 million Facebook users.” And Adam himself downplayed the PNAS contribution but obviously stands by it else he would have retracted the article from the journal.

Personally, I think there is a solid question which I don’t see being discussed. There’s enough data in this world of big data so why was there even a controlled experiment? So while we can point away from the legacy protocols and towards “rethink(ing) our assumptions about how human subjects research plays out,” I can’t help to wonder why this study was run (read: how did it advance previous work?).

What has been covered extensively is the research ethics conversation. Academic and industry scientists work in different environments. Academic researchers have institutional review boards. Industry researchers have lawyers. Not all IRBs or lawyers alike allow the same thing and researchers are not required to prove their ethics beyond the often routine description of their experiment. So we rely on our own views and discussions. In many a program meeting, I’ve seen an ethics assumption be applied (where one assumes they received approval).

It’s unfair to say SIGCHI researchers have been silent. The conversation has been happening across many channels—public, semi-public, privileged, and private—in many contexts. And it’s most likely that this article’s stronger contribution is that it has us rethinking the scientific ecosystem in which we work.