This sounds rather similar to the form of scientific fraud where you first create a conclusion, then invent/manipulate the data until it supports your conclusion.