The debate about randomised control trials (RCTs) is both intellectually interesting and important both for the developed and underdeveloped world (convention is to refer to the latter as "developing" but of course often its not). A comon argument that one hears,mostly from non-academics, is that RCTs are unethical because "you are playing with people's lives" , "denying people something etc". A conversation with a colleague, who I won't name lest I misrepresent them, put this in context: every time the government does a reform , say cut a tax or reduce child benefit, it is doing an experiment. It is playing with people's lives, it doesn't know what the effect will be but it won't learn anything: from it because of the design. With an RCT, you might learn something. So whats it going to be?
The debate about randomised control trials (RCTs) is both intellectually interesting and important both for the developed and underdeveloped world (convention is to refer to the latter as "developing" but of course often its not).
ReplyDeleteA comon argument that one hears,mostly from non-academics, is that RCTs are unethical because "you are playing with people's lives" , "denying people something etc".
A conversation with a colleague, who I won't name lest I misrepresent them, put this in context: every time the government does a reform , say cut a tax or reduce child benefit, it is doing an experiment. It is playing with people's lives, it doesn't know what the effect will be but it won't learn anything: from it because of the design. With an RCT, you might learn something. So whats it going to be?
The title of Imbens' paper sums up the views of many:Better LATE Than Nothing
ReplyDelete