I spoke with Gita Sen after the final panel yesterday about the example she mentioned of a woman who had been beaten by husband for participating in their survey. The survey was not asking anything that anyone would be offended by - he apparently just didn't want her talking to them. They learned about it from others in the village. The take away she mentioned in her talk was that the concept of 'informed consent' was insufficient in such a case. I asked her more about the situation, what options there are, and how to start trying to resolve a situation like this.
Option: get the husband's permission to talk to the wife. That doesn't exactly empower women, does it? It strengthens the existing power relationships or can even create them if they weren't there before.
Option: as part of getting informed consent, ask if the woman feels there is a likelihood of her husband beating her for talking to us and if she is willing to take that risk. But if we as researchers are told that there is a risk of our participant being beaten for helping us, are we sure we still want to go through with it? Do we believe the fruits of this research justify asking someone to put themselves at risk?
Option: avoid the situation by not talking to those people at risk. But they may be providing an essential viewpoint and may in fact be the target audience the research is intended to help. Given the different rebuttals to this point and the last one, I don't think there will be one answer that works for all situations.
Does it change a researcher's moral culpability or duty if this is a spousal relationship characterized by violence (he would have beaten her anyway, we just provided the excuse de jour) or if this is an isolated, unusual event where we are a causal factor?