Statistics is an often quoted and misquoted field, used to support differing arguments entirely with seemingly the same evidence according to the will of the person making the statistical calculations and observations. What follows is one simple way to get away with this. One simple way to do this and to get away with it is if by utilising a factor with an extremely low R2. Your R2 should determine how much your independent variable explains the effect in your dependent variable (Murray, 2). For instance you could, publish a finding saying that over-fishing is responsible for an increase in shark attacks and do so with an R2 of just 0.1. You can hide this fact buried in the paper with ease, and while you state that over-fishing is the case, it doesn't explain 90% of the increase in shark attacks. This is a commonly employed tactic and has been since the birth of modern econometrics (McCloskey, 7). If you misquote an R2 value, much like claiming that a study with very few participants has a very high significance to the population at large, you are able to produce what will seem to be a very thorough and convincing work. The result will even be true, if you want to interpret it that way, over fishing has had some effect on shark attacks. What that effect is, how much of an effect it is and whether it has caused an increase in shark attacks are all up for debate and not proven, but it does have an effect so long as no one looks too deep in to the methodology.
Bibliography
McCloskey, D., Why Economic Historians Should Stop Relying on Statistical Tests of Significance and Lead Economists and Historians Into the Promised Land Newsletter of the Cliometrics Society 2:2 December, 1986. Print
Murray, C., How to accuse the other guy of lying with statistics Statistical Science, 20:3, 2005, 239-241. Print
