Statistics and ethics: human volunteers edition

| 1 Comment

1 Comment

That's ridiculous, not to mention disgusting. Thank god for Human Subject Review.


However, HSR can be a bad thing, too. Some colleagues just took cookstoves to a refugee camp in Darfur, Sudan. In actual use in the camps, the stoves reduce fuel-wood consumption by about 50% compared to typical "three-stone fires" (build a fire in the middle of three stones, and balance the cookpot on the stones). This is a huge deal, since the women have to walk for hours to find sources of wood for fuel, which is a huge labor cost and also exposes them to the danger of rape. (The men can't go instead because they are simply killed).


Anyway, my colleagues took some stoves over, taught people how to use them, and monitored the fuel consumption while the stoves were used. They also wanted to know some relevant facts such as the amount of fuel gathered per household, the extent to which lack of fuel is a limiting factor in consuming the food (e.g. how often does the family skip a meal due to lack of fuel), and so on. After observing for a few days, they made up a little survey, then went hut-to-hut to ask their questions. They got a lot of useful information that will influence the supplies the refugees are given, the creation of a market in stoves (which will now be produced locally), and so on.


None of their results can be published. There was no practical way to have their survey instrument reviewed, nor to formally obtain informed consent from all participants. Even the cooking demonstrations probably violated human subject protocols. Some of the results will get around by word of mouth, and the NGOs on the ground in Darfur will put the information to work, but my colleagues cannot publish it for wider use and for archival purposes, and technically should not have performed this work at all.


I'm not sure there's a right answer here. Probably performing this work without Human Subject Review should be considered improper (as it currently is) and should not be published. But it's good that there are people, like my colleagues, who are willing to break those rules.


Of course, this is a far cry from the sort of stuff mentioned in the article that started this thread.

Leave a comment

Subscribe to Entry