Its good to see that people are still doing research on this. This study provides some useful feedback as well as a number of references I was unaware of.
This article in the Journal of the American Society for Information Science and Technology directly challenges the notion that the age of an edit can be used to evaluate it's reliability. I'd like to be able to read this article (particularly as I'm referenced in the abstract) but like a lot of scientific scholarship the article is not available to the general public.
On the other hand, this study seems to indicate that the age of an edit could be used to evaluate its reliability.
I think the reason for the split in the results might relate to how easy it is for a layperson to identify that a given piece of information is incorrect. Some kinds of vandalism are more subtle than others. Subtle vandalism is more likely to survive.
RE: Evaluating WikiTrust: A trust support tool for Wikipedia