Hello!

While not something you should reference directly, the XNA Wikipedia article gives a nice overview of the product and version history.
I would take it as a first step in understanding the product, and begin to jot down and do some detailed research of the technologies surrounding XNA.

While I would certainly reference Microsoft sites, realize that any company will promote their product to a certain degree, so it would be a good idea to gather perspective by understanding competiting technologies and reports coming from experts not directly affiliated with Microsoft.

You could also approach the guys who make the Torque engine, as they've written tools to work with and without XNA and maybe could shed some light on the differences.
You also have a community of developers here that use XNA and have developed successful games with it.  Try getting some interviews.

Note that most use XNA and XNA Game Studio interchangably, but in trade shows Microsoft has stated that XNA is a 'brand' for all their development tools, not just those for the indie community.

Seems to me Microsoft is proposing a scenario in which:
- The develper wins because they can develop games faster and spend less time testing (typical testing for XBLA games tpycially run in the $100K+ area I've read)
- Microsoft wins because they have more control of the "sandbox" the games can run in - cutting testing time down and creating a layer where an OS can not fail due to a rogue piece of software
The biggest snag being performance of the framework, and the 'gotcha' of having to use C# (in an industry that mainly uses C and C++)

You may want to try making something with XNA to further understand how it works and how it ties into Microsoft development tools.
Pong seems to be a popular first effort ;)

Good luck with your paper!