Advanced PageRank Analysis by Dan Petrovic
Private testing of this seemed to confirm his claim. The addition of a nofollow tag to some links did not improve the PageRank of other pages linked to on a site, however the PageRank could have changed, just not by a whole number, so that the change wasn’t visible (such as going from 4.5 to 4.75).
Though that did seem to confirm Matt’s claim, another observation that was made didn’t. If you have 10 links on the home page of a site, they usually go down by one PageRank value (example: 5 to 4). Having 100 links on the home page meant that the pages they pointed to couldn’t get higher than a PageRank of 2 or 3. This contradicts Matt’s claim that having less links on a page doesn’t make the PageRank of those other pages go up, and that the number of links doesn’t make a difference at all. This makes no sense, and it would be great to get clarification from Matt and to hear his view on this issue.
The conclusion from all of this is that Google does, in fact, follow nofollow links, but they just don’t take them into consideration in determining a page’s PageRank. “No follow,” therefore, is more like “no vote.” It is worth bearing in mind, however, that Google uses many other factors to determine a PageRank, like on-site positioning, navigation depth, and off-site links and contextual references. This means that it is much harder to tell what factors are affecting a PageRank and that our job is that much harder.
The conclusion from all of this is that it is not worth all of the time to try to place internal nofollow tags, but rather to spend your time creating more content and generating more links to your site.