It's a technique that seems obvious once you've heard the idea but I'd never thought to test in this way. Time to think through a pilot test to see if this is as solid a match for testing new features in our systems as I think it is.
Feeling a lack of natural scrolling
To Brodigan's point on giving sufficient time for testers to get used to the new feature before taking it away: It took me a while to get used to OS X Lion's "natural scrolling." It took me time to reverse years of muscle memory that told me the new method was backward. Once I got to the other side on muscle memory I found it as natural as the marketing term claims. It's now to the point that using the traditional scrolling gesture in Ubuntu feels so frustrating that I took to the Internet to learn how to get natural scrolling working in Ubuntu as well. Without giving sufficient time for users to adjust to the new feature we won't get accurate information about whether it hurts to take the new feature away.
Libraries using this technique for existing services
In a parallel to Brodigan's point that "deprivation studies actually happen a lot in real life," the Queens Library is using deprivation methods to help the communities they serve understand what life would be like without the features they provide through mock funerals like the one they held at their Pomonok branch in Jamaica. This may seem an odd move for a library to make but given that their key argument in encouraging community action is that the libraries "cannot sustain cuts of these proposed levels at a time when they are experiencing some of the highest number of visitors in history," it may prove effective motivation. My heart and mind respond more clearly to losing something I like or depend on than hearing that an abstract percentage of funds won't be available.