We present a new framework, using a model-based bootstrap with pre-whitening and post-blackening, to analyze the robustness of overall controller synthesis methods to model uncertainties. This considers both the system identification and the subsequent controller synthesis together as an overall synthesis method to be analyzed. This framework is extremely flexible, can be used with any identification or synthesis method, and can easily be extended to many other types of problems. As a very encouraging first step, we see a dramatic confirmation of the dangers associated with certain synthesis methods.