I tend to grab the paper title and search for it as a literal string, and in the majority of cases, the only full version available is behind a journal paywall.

Sometimes, especially in CS, you get lucky and find a citeseer or personal-page copy, but anything IEE(E) is pretty rare IME.