I like to refer to the “Well, duh” test. A reasonable person has to be able to disagree with it. If every intelligent reader will agree with you, you're not doing your job. Here's a good test: imagine reading your thesis statement to your class and asking for a show of hands on who agrees. If the audience is unanimously in your favor before you've presented any evidence, you have a bad thesis. (Of course, if you start with a divided audience and bring them all around to your point of view after you've laid out your argument, you're in good shape.)
Finding a controversial thesis takes some work. For starters, you can't really argue about most matters of fact. If it's something that's widely known on which everyone agrees, there's no point in arguing it again.
Neither, however, can you really argue about matters of opinion. Controversial isn't the same as subjective. If your paper is simply about a matter of taste or preference, there's no way to argue about it: de gustibus non est disputandum, “There's no accounting for taste.” That's one reason that evaluative theses are to be avoided — how can you prove that The Handmaid's Tale is better than Beloved?
So: if you can't argue about facts, and you can't argue about opinions — what's left? Matters of interpretation, things that (at least in principle) can be right or wrong, but aren't obvious on first glance.
We Americans aren't really accustomed to this kind of argument. On the one hand we're taught that it's impolite to question someone's beliefs — “Everyone's entitled to an opinion” is a modern mantra. On the other, crappy daytime television shows suggest that “argument” is the same as shouting at one another. Even our so-called presidential “debates” are just joint press conferences: there's no real engagement in anything substantive. But one of the purposes of an English class is to teach you to make arguments about things that, at least in principle, can be settled.
Here's a tip, and one of the best I've got in my entire guide: write on something you don't understand on the first reading. It sounds odd, doesn't it? — we want good grades, and of course we're drawn to the things we get right away. I mean, if you have a choice of problems on a math exam, you don't pick the question you understand the least, do you?
But it's paradoxically good advice on an English paper, because it almost forces you to come up with a controversial thesis. Now, by “don't understand” I'm not talking about “don't understand the words” — I assume you've made at least a minimal effort to get the literal meaning. What I mean is that you should find things that puzzle you, that don't seem to fit into your understanding of the book, because the best theses are always lurking there.
Once you've got a puzzle, something that bothers or confuses you, you're on the trail of a good, controversial thesis. Try to answer it to your own satisfaction, probably by resorting to close reading. Why does the author start using big, complicated words here when he used simple ones everywhere else? — what's the point of concealing the heroine's name until the very last page? — why does the narrator keep mentioning his dead wife? Once you answer it to your own satisfaction, you've got a good, controversial thesis, and you're on your way to a good grade.
For more insights, check out the swell guide by my old pal, Erik Simpson, who's really too smart for his own good.