Funny that you mention up the Canada Loon series. When I read the initial post, this series was the first thing that sprung to mind.
Most of those puzzle were given lower difficulty ratings, mostly 2* or 2.5*. However, out group consensus was that the amount of time and effort required for those caches was highly disproportionate to the ratings. This is why we chose to solve the puzzles as a group rather than individually.
The puzzles were quite labour-intensive, but the hides are all fairly easy to the extent that if you're looking for more than five minutes, either your co-ordinates are incorrect or the cache is missing. To me, 1-1.5* should have been added to the difficulty levels of the majority of that series.
For me, the difficulty portion of a puzzle's rating is more about the puzzle than the hide, as the rating should be proportional to the feeling of accomplishment after solving it. With a tricky hide combined with a relatively simplistic puzzle, it is appreciated when the owner states as much on the page. Again, the Binthairs are prime examples of this. Some of his puzzles were quite difficult, but, with a few exceptions, once you were within 15m of GZ, you knew where the cache was. Challenge caches are a whole other story, as the difficulty rating is more about the actual challenge that the hide, or the puzzle involved.
GC3H6VJ is a good example of the difficulty rating being more about the hide than the puzzle. The puzzle was dead easy, but the hide was a magnetic nano attached to a fire hydrant. For the record, the hide really wasn't all that difficult.
Given the variety of puzzle types out there, and the variety of local caching cultures, there will never be any consensus on how to properly rate the difficulty of a puzzle. In my opinion, a 4* puzzle in Oakville is roughly analogous to a 2.5* puzzle in Ottawa. That's not a comment degrading Oakville's caches compared to ours, it just is what it is.