Here is a ‘discipline neutral’ way of understanding policy research. This stuff comes out from a course on Introduction to Policy Research that I take this term. Interestingly enough this is the first time that we have a specific course titled as that – Public Policy, as a part of any graduate program in an Indian university. What follows is a way to simplify what policy research is for the uninitiated. It can possibly help in mapping the landscape of the several different kind of policy studies that exist. This is closer to what we have often done – applied and functional to get going in a space. The policy professor here whose own work looks at Indian politics, policies and processes from an applied perspective, helped folks understand policy research area through the schematic that I’ve cleaned up and represented below.
On public policy, he argues that it is about creating change by making hard choices between competing values. It is about approaching problems and these approaches vary with different interest groups. I am attracted to this explanation because it is not theory heavy and emerge from a gaze set on real world and unlike other insular theories in public policy that I have come across.
The schematic came about from discussion on India’s transition in public policy space and various studies like Rob Jenkins’ on what made the 1991 reforms work in India, Achin Chakrabarty on reform debates and more importantly Pranab Bardhan who articulates the Indian state as “predatory” during the Indira Gandhi period. The variety of public policy study and practice as seen in the West is relatively new in India and therefore the extra effort in explaining its direction, intention and methods. Public policy study is done in two ways –
1) Policy Research – which is a post facto analysis of what happened with policy X in place and why.
2) Public Policy Analysis – which is conducted before or towards developing and implementing a new policy.
And hence the following schematic to understand public policy research –
Contested Realities 
IN your post you mention this:
1) Policy Research – which is a post facto analysis of what happened with policy X in place and why.
2) Public Policy Analysis – which is conducted before or towards developing and implementing a new policy.
Is it not the other way around? i.e. Analysis being post facto and research being towards developing a new policy.???
Another Q i had is with your schematic, the bottom right box in the matrix again looks deductive. From large theory to individual specifics. The Banerjee and Duflo study.. Is it the other way round then?