Monday, September 01, 2014

Heterogeneous Treatment Effects? The Case of Deprogramming British Radicals

If I'm reading this correctly,  the British Prime Minister will require that British citizens who have spent considerable time in Iraq and Afghanistan will have to attend "deprogramming" classes if they return to England.  The enormous economics literature on "treatment effects" has some insights to offer.

I have several questions;

1.  Do the treated subjects have any incentive to take the treatment seriously? Will they just sit there and meditate and ignore the talk and videos?

2.  What is the treatment's substance?  How do you deprogram a person?  How many hours of training are completed?  How will the subjects be evaluated to judge whether in the short run that they have been deprogrammed? Will it be a signalling equilibrium such as shaving a beard and wearing Western clothes?    Will MRI brain scans be used to monitor how the lobes of their brains light up to different stimulus such as a photo of ISIS vs. Hilary Clinton?  Such brain scans would be more expensive than simply surveying these individuals and would be less likely to yield "cheap talk".

3.  Given that people differ, what % of those assigned to treatment are at risk to be "deprogrammed"?  Who is at the margin?

4. Is there a risk of backlash so that by taking the treatment, a subset of those treated become more radical because they hate the idea of "Big Brother" mind control?

5.  Once these individuals are released back into their British communities, will their family report to the government how their day to day progress is going?  Will there be the equivalent of a parole officer?  How will he/she judge progress and worrying signs?

UPDATE:  My wife thinks that the British government may intend to use this program to profile this group to find out who are the risky ones among the bunch. It would interest me if the British government is shrewd enough to be able to achieve this partition of the entire sent into "risky guys" and "low risk" guys.