.. non-hypothesis-based data analysis. Take, for instance, some health data that was overlaid with GIS data at a health authority I worked at a few years back. One of the more curious combinations that popped up was the fact that there were a higher number of Leukemia cases in households that lived under high power electricity lines. We were all about to unanimously blame the cancer on the long-period-exposure to EMF, when someone who lived in the area brought up a good point... properties under the lines were less expensive that those properties away from the lines, and therefore more attractive to lower-income families - the very same lower income families whose eating habits are generally lower on the nutrition scale. ie: higher carbs & fats, with less roughage, antioxidants, etc, which all helps to contribute to poorer health.