1. The halo effect. It’s common to believe that someone who is good in one regard is good in all. So the belief that Polanski is a good film-maker leads people to think he is a good man. Investors in Bernard Madoff’s funds might have made a similar error - they inferred from evidence that he was “a great philanthropist, a pillar of the community” that he was trustworthy.
2. Ingroup bias. Kieran has called it right: a lot of the defences of Polanski are no more than “He’s one of us.“ But it’s a common tendency for us to be favourably disposed to people like ourselves. For example:
4. The fundamental attribution error. Having decided that Polanski is a good man, people refuse to believe that he can have done a bad thing: “It wasn’t rape-rape” says Whoopi Goldberg.
The error here is to exaggerate the extent to which character determines actions, and to fail to see that good people can do bad things under certain circumstances. It’s in this context that Polanski’s being a victim of the Holocaust is ironic. That episode teaches us that ordinary people can do atrocious things.
My personal way of trying to correct for this error is to regard people not as “characters”, but as bell curves; we see only a sample drawn from such a curve, and we shouldn‘t draw inferences about the whole curve from one sample. For this reason, I would no more call Polanski a rapist than call Tony Blair a guitarist.
5. Misperceiving punishment. It’s common to regard punishment as a judgment upon an individual. Polanksi‘s defenders recoil from doing this.
But we shouldn’t see imprisonment in such morally-laden terms. Instead, we should take the title of Gneezy’s and Rustichini’s famous paper (pdf), “A fine is a price” one step further; a prison sentence is just a price. Polanski’s supporters expect him to pay for his suits or his dinners - so why shouldn‘t they expect him to pay for his rape?
I don't say this to criticize Polanski’s supporters - loyalty to one’s friends is to be admired. I do so to point out the ubiquity of cognitive biases.