Who reads the small print?

As a proud ESRI Developer Centre we were disappointed to see the map on the front page of ESRI‘s recently promoted (via twitter) Geomedicine website. The front page (screen shot below) shows an interactive map of the USA. On it users can overlay heart attack rates (per 100,000 of population) and alongside data showing the locations of reported incidents from the Environmental Protection Agency’s Toxic Release Inventory. Both imply a clear east/ west split across the US with the east fairing worse in both cases. An instant and perfectly reasonable reaction to this visualisation is that heart attacks are caused by the release of toxic material. I felt encouraged to make this assumption by the questions above the map: “What environmental exposures have you personally experienced in your lifetime?” and “Would any of this geographically rich information be useful at your next physician visit?”. Of course, medically, the natural conclusion from the map is spurious, so why have ESRI published it? Scroll to the bottom of the page to the disclaimer which states that no causal link between the datasets is implied and that their intention is “present some of the data about geography as it relates to human health”. By the time you make it that far down the page and squint a little it is too late. I think ESRI could have put a little more thought into this. At the very least they could have shown more data sets, being careful to avoid those that appear to correlate. Changing the “small print” at the bottom to larger more prominent text would also improve people’s perceptions of this map.