Ratherfancy posted a link to this article on cxpartners:
The myth of the page fold: evidence from user testingPeople tell us that they don’t mind scrolling and the behaviour we see in user testing backs that up. We see that people are more than comfortable scrolling long, long pages to find what they are looking for. A quick snoop around the web will show you successful brands that are not worrying about the fold either...
Particularly revealing is their eye-tracking heat map of two alternate designs for the same page:
They also note that "Stark, horizontal lines discourage scrolling." Presumably because it cues the user that there's nothing more to see, or that what's down below is something completely different.
This seems to jive with anecdotal evidence from eBay and Vimeo.
eBay's text ads performed better at the bottom of a page below a list of auction items because the user was scanning the list of items available on eBay, and having exhausted that list, opted to try a non-eBay merchant to get what they were looking for.
Vimeo's medium rectangle performed better below the fold next to the comments. One hypothesis was that the flashy image-rich display ad didn't have to compete with the video for the user's attention. An alternative hypothesis might be that the user's intent is to watch the video, then glance through the comments for something interesting. The display ad simply isn't compelling enough to deflect the user from his intent, so offering it as something to click after the user has accomplished his goal increases its performance.
The danger for product and design, of course, is when we substitute blindly following one truism for another. You still want to put the "good" stuff above the fold to tell the user that this page is worth their time. Once you've got their attention, they have to be encouraged to scroll to see the "pretty good" stuff through a combination of good IA and design.