Basically a guy set up a test page with certain unique words either hardcoded in HTML (as a control) written to the page using JavaScript's document.write() function, and written to a page using JavaScript in a externally referenced file. Here are hist results:

I then searched for each of the six words at Google.

* The two HTML words both generated a search result that included the page.
* The two words inserted by a JavaScript in the page generated no search results.
* The two words inserted by a remotely sourced JavaScript generated no search results.

Which are utterly unsurprising if you think about it. Google's crawler doesn't implement a JavaScript interpreter. Plain and simple. Because it doesn't have to.

As someone whose career is researching, designing, and developing advanced web crawlers, I can tell you JavaScript parsing/interpretation is a giant pain in the ass and a big performance killer. Plus things like client side validation and image pre-loading (things that most crawlers don't care about) also gets in the way and slows you down. From a shear cost vs. gain, it currently makes no sense for Google to interpret or index JavaScript. Ajax apps only makes crawling much harder.