It's also important to keep track of how the players are being used. When the coach is only putting you on the ice for offensive zone faceoffs, your +/- may be skewed in your favor. Meanwhile, the guys who are getting all the defensive zone faceoffs have many more opportunities to develop a lower +/-. Likewise, as a 4th liner, your goal isn't necessarily to score goals, so they play more defensive minded. More or less if you can get it to center ice and dump it in, it's a successful shift, so I would expect to see +/- closer to 0 for grinders than for top liners, whom are expected to take more chances and score (also more vulnerable to goals against in many cases). These guys I would expect to see more of a variation.
I think many advanced stats can be extremely misleading as well. They have to be looked at together and with the eyeball test.
Plus-minus I think is really only useful then you're comparing players on one team. If one has an extremely high or extremely low +/-, there's something going on there. But it still doesn't say much beyond that.
So I don't know if it's fair to use +/- as a main argument to compare players who are used for different roles.