The Sharpshooter Fallacy is a logical fallacy where patterns in random data are used to create a hypothesis after the fact. The fallacy gets its name from a hypothetical `sharpshooter` who shoots randomly at a barn. Even though he is shooting randomly, chances are some spots on the barn would be hit more than others. The `sharpshooter` then goes and paints a bulls eye around one of these spots that have received a much higher than average rate of hits. The effect is that he looks like an excellent marksman because so many of his shots are in the target. That is, the same data are used to create the hypothesis and to test it.
This fallacy has often been used by liberals looking to implicate a certain product causes cancer. If you were to test the rates of diseases in certain small communities, random chance would result in some communities have a much higher rate of a certain disease than other communities. Lets say community X has much higher rate of Lymphoma than average and also has a plant making chemical Y. Assuming that chemical Y causes the Lymphoma is an example of the Sharpshooter Fallacy. Chances are another small community producing chemical Y might have no Lymphoma at all.