Thursday, 30 September 2010

Ethics, Experimentation, and Social Services

From the New York Daily News:
In an attempt to test how well a one-stop assistance program is working, the Department of Homeless Services split 400 struggling families into haves and have-nots.

The "haves" get rental assistance, job training and other services through a program called Homebase.

The other half wasn't so lucky.

Those people - chosen at random - were dubbed the "control group" and shut out of Homebase for two years. Instead, they were handed a list of 11 agencies and told to hunt for help on their own.

The city will still be watching them, but for a whole different reason.

For two years, researchers will track the separate groups by their Social Security numbers to see how each "uses city services ... such as shelter, public assistance, foster care," city documents state.
This is how you run an experiment. You try to randomise the data, you pick a control group that matches those being experimented on, and you track what happens. You "shackle" all other variables, so that the two groups are identical - except in the one area under examination.

But first... when dealing with humans.... you get clearance from an ethics panel. They're there to make sure that you don't do things like leaving gangrenous limbs untreated, to see what improvement various medical interventions make. That particular experiment would be quite useful, by the way. Circumstances such as Earthquakes, War and so on that prevent effective treatment also prevent good monitoring and gathering of statistics. Much medical treatment is poorly evidenced, and it's possible we may be doing less good than we think by them.

Nonetheless... it's inhuman to do this. Anyone capable of doing it has let their (desirable) objectivity and (commendable) wanting to help more people get in the way of being human. Not a monster.

This experiment is a perfect example. The experimenters expect that by denying aid to some, it will cause those experimental subjects misery. They want to see how much difference - if any - their new intervention makes.

Perhaps this could be justified if it was for a short period, and if the control group were then provided additional resources afterwards, to make up for the hardship. Moreover, although it pollutes the data by introducing another variable, all controls should have been volunteers - told that in exchange for initial difficulty, they'd get compensation in the long term.

But they weren't given the option; the period of the experiment is two years, far too long; and there's no hint of compensation. No competent Ethics panel could have cleared it, and there's no sign that any were consulted. Such sloppiness gives me no confidence that the experimenters have good performance metrics for measuring effectiveness (or otherwise) of the new facility.
"These are real parents and children, not rats in a lab experiment," said Patrick Markee, a senior analyst with the Coalition for the Homeless.

City Councilwoman Annabel Palma, who heads the General Welfare Committee, said vulnerable New Yorkers "should never be steered away from the services and benefits for which they are eligible."

Starting in August, letters went out to 400 families and individuals applying to join Homebase, telling them they would be part of the study. Those picked make up about 5% of the 7,700 families that used the program last year.

The city defended the plan.

"We serve thousands of people through this program, and this study is only looking at 400 people," Deputy Homeless Services Commissioner Ellen Howard-Cooper said.

Researchers will watch both groups to see if the "control group" families find help through other agencies or wind up in shelters, officials said.
I propose another interesting experiment; find all those involved with conducting this "experiment" - who are bound to be of above average intelligence and education, compared to the indigent experimental subjects - confiscate all their assets but the clothes on their backs in lawsuits, saddle them with debt, a short prison sentence so they have a criminal record - then see how they fare, without assistance. Just to determine the effects of education and intellect in the programme.

Of course, such an experiment would never pass an ethics panel review, and I'm too good a scientist not to insist on one as a precondition.

She who fights monsters runs the risk of becoming a monster herself, you see.


Anonymous said...

Interesting how you post this then the issue with the STD tests in Guatemala are exposed.

There is no ethics in the government. Doctors and scientists who work for the government have no true control over what they are going to investigate. If the government is willing to do heinous things in the name of pursuing knowledge then the people they hire to do the work will also be unethical.

Then there is the entire ethics are relative arguement.
What is ethical is only held as ethical because someone says it is thus.
Ethics stem from honor. Both of which are constructs of the human mind.

That is anouther discussion in its own right.

Have a good day,

Zimbel said...

Here's a U.S.A. Today article on the Guatemalan experiment. I thought that the last couple of sentances were very much on point:

Collins says. "Today, the regulations that govern research by the U.S. government, whether funded domestically or internationally, would absolutely prohibit this kind of study."

Reverby says those protections may not be enough. "We all realize how much research on drugs is done overseas," she says. "The question is how do we keep on top of what's going on now?"