I'll throw in two cents worth:
The difficulty of a puzzle is tough to measure objectively. Subjective measures are far more, well, accurate.
A generation or two ago, a large part of the difficulty of crosswords (and I'm referring primarily to the New York Times puzzle here, which was then and is now the king of the hill) was in an obscure, crosswordese vocabulary of words like ANOA and ETUI. Obviously, the more obscure the vocabulary, the more difficult the puzzle.
Today, crosswordese is kept to a minimum by the NYT and pretty much everyone else below them on the food chain. The primary measure of difficulty today in the NYT and puzzles of similar stature is in the cluing. I'm going to steal an example here from Amy Reynaldo's book: AREA might be clued on Monday or Tuesday as "Region" or "District". Midweek, the clue might become "It may be gray" or "Geometry calculation". For Friday or Saturday, it might be a bit more obscure ("Purlieu"), or vague ("Country statistic"), or involve wordplay ("Side by side, maybe?").
[In case I've gotten ahead of myself, NYT puzzle difficulty increases from the easiest puzzles on Monday to the most difficult on Saturday].
Generally, the top solvers are ranked by speed. There is an annual American Crossword Puzzle Tournament (
http://www.crosswordtournament.com/), and you can find year-by-year standings on their website. Points are awarded for both accuracy and speed, but nearly everyone that finishes in the top 10 in a given year completes each puzzle perfectly, so the distinction between the best of the best comes down to speed alone. If you've seen "Wordplay", you 'll recognize some of the names; if not, see "Wordplay"!
Hope that answers the gist of your question.
