How to win a "scientific" argument if the science doesn't support you.

Science has a very simple rule: you have to be certain of what you say. It doesn’t allow arguments based on opinions. Science requires evidence and deductive reasoning. Something that is infuriating for those who want to use science to say something it can’t or shouldn’t say.
That is why the “precautionary” principle has proven so useful. In essence, the precautionary principle is a rewriting of the rules of science. Without the precautionary principle, one has to say: “the evidence shows”. With the precautionary principle you can say: “the evidence could lead us to believe”. Bang goes all the burdensome necessity of proving what you say with evidence to call it science, in comes the precautionary principle which has no limit to the unscrupulous. You can take any bit of information, add a few “scientific facts” show how it “could lead to some kind of tragedy” and then claim: “science tells us that we must …<add whatever opinion you like to stop the supposed tragedy>”.
Indeed, in many cases you can totally bypass the step of adding any facts. Global warming will be harmful (but Age Concern: 23,000 people a year die due to cold in the UK) therefore “science tells us” we must stop global warming. Well, actually, the best evidence would suggest that if anything we should encourage warming because the empirical evidence suggests warm periods are better for us humans.
It would be a bit like a judge saying to a jury: you must weigh the evidence and only convict if you are certain “beyond reasonable doubt”. At which point the prosecution says: “but shouldn’t we take precautions in case this individual is a latent mass murderer” … to which the judge says: “of course we can’t let off a mass murderer, certainly if any of you have any doubt whatsoever of the innocence of this defendant then you all must convict”.