I was really struck by Lua Wilkinson’s recent post on the ethics of photography in low income countries. She included a fascinating anecdote about Dorothea Lange’s famous photo of a California sharecropper and her children, which later turned into one of the iconic images of the Great Depression:
According to reports, [Florence] Thompson and her family were not happy that they had unknowingly become the poster children for poverty. They voiced concern that the information surrounding the photo wasn’t even accurate… Thompson gave permission for the photo to be taken but was under the impression that the photos would never be published publicly, and even Lange herself notes that she failed to send the final prints to the family as promised.
The point is this: Lange asked Thompson if her photo could be taken and Thompson said yes. The problem was not lack of permission. The problem was that Lange had control over that photograph, and there was little communication about what that actually meant. Lange controlled where the photo ended up, as well as how Thompson and her family were portrayed.
The Thompson family was turned into a symbol of poverty. Lange, widely respected for her work in social justice, likely saw the family as such and knew the power that symbol would have on policy-makers and humanitarian aid. But in doing so, by turning the Thompson family into a symbol, she also took away their power. She controlled their identities. They had just one story, and that story was of poverty, whether they (as the story’s central characters) agreed or not.
There are some obvious similarities here with the process of gaining informed consent in survey research in low income countries. Every researcher I know is scrupulous about getting consent before beginning an interview, but I don’t think most respondents really know what will happen to the data they provide after the interview is finished. The consent form never mentions that information about their household and others like it will be discussed by academics and policymakers at high-flown conferences around the world in a few years. Call it partially informed consent – lots of information provided about the circumstances of the interview, but very little about who controls the data and how they interpret it afterwards.
Should Lange not have published that photo? Its publication did help spur an increase in food aid to depressed areas of California, and it’s a truly striking piece of art as well. But it remains the case that Thompson’s consent was only partial – to have the photo taken, but not to have it published. Lange ought to have asked for her permission before releasing it.
And what of informed consent, then? Admittedly the parallels are not exact. Thompson and her family were easily identifiable and upset precisely because they could be identified in a way that they didn’t want, while IRB requirements mean that survey data is kept rigorously anonymous, and that respondents are promised that they won’t face any other problems as a result of participating in the study. The fact that there’s a power imbalance between the respondents (who provide the information) and the researchers (who control its use) doesn’t inherently mean that the interaction is unethical. Indeed, there’s a strong ethical case for research that aims to reduce poverty. But it’s still important to think critically about the power dynamics present in this work, and particularly the question of whose voices get heard in development. What might respondents say differently if they knew that their information wasn’t just of interest to the researcher, but might be presented on a global stage?