Steven Novella at NeuroLogica Blog has a great post explaining why ID can’t meet the criterion of falsification. How does one conclusively disprove the existence of the Great Vorlon?
I would add two points. First, a good trick that intelligent design creationists play is that they subtly make their ‘hypothesis’ (such as it is) the null hypothesis. That means evolution by natural processes must always make the affirmative case, or overturn intelligent design creationism. Because ID creationism is so well established. Or something. I call bullshit.
Second, it always interests me how, when engaging in ‘formal’ discussions of hypothesis testing, we always use the Popperian framework of falsification, when, in fact, many times scientists use an approach that is essentially a likelihood approach: judging which hypothesis best fits the data, as opposed to falsifying a null hypothesis.
Of course, from a likelihood perspective, intelligent design creationism is bullshit too.
More about likelihood and evolution here.