I've been involved in a few tests of this nature over the past 35 years, and the one thing thats constant is the people involved are not random, they know why the test is being done and are working very hard to make it work. Its when you turn the whole population loose with it that the problems start, reminds me of speed limits they works great in a traffic engineers office but then you can just drive down I-5 to see how it really works.