In this talk I will review the many number of ways my colleagues and I, especially George Robertson, have chosen to evaluate novel information visualization designs and the lessons we learned from them. Often, paradigms borrowed from Psychology were pivotal in discovering breakthroughs in our designs and guided us forward in our research. In addition, controlling for various variables such as gender, spatial ability, visualization ability, computer experience or gameplay made a huge difference in our findings and subsequent explorations. I will begin by giving an overview of many of our early experiments and then provide a more recent overview of information visualization research currently ongoing at Microsoft Research. Finally, I will close with new, on body and automated methods for tracking emotion, such as frustration, interest and engagement that might provide a new, useful direction for information visualization research.