Although there is no mathematical way to test this, it has been my experience that your Chess.com rating and your 'real' rating are often very similar, with your Chess.com rating occasionally being inflated slightly. Of course for your first few games (20, using other rating systems) your rating will be terribly innacurate anyways so don't worry about it.
-matt
Your rating here does depend a lot on how much time and effort you put into your games. Clearly if you are studying every move in depth, checking openings databases, using the analysis board etc, and your opponent is nonchalantly glancing quickly at the board and firing back the first move he/she thinks of, then you will quickly gain a higher rating relative to "true" ability.
Its not really possible to discover how much effort the average player here puts into their games, but it is probably true that if you are concerned about your rating then you will be making more efforts in games to try and improve it - and it will become inflated as a result.
The only way to find out is to get playing over-the-board and see how you get on.
Thank you for your explainations
cuandillar mentioned that opening books and databases are used her, is that really allowed? I at least would never come up with the idea to use an opening book when I am playing against an opponent.
mvh Fredrik
From what I have seen a 2000 player on this site often plays like a 1600 or 1700 postal player (USCF rating), which would correspond to a 300 or 400 point 'inflation' factor.
Players on chess.com are having a LOT of fun, but are often more careless than players with similar rankings that I have encountered in, for example, a USCF Golden Knights or Electronic Knights. I get a lot of free material from blunders and oversights that I would never get in a Golden Knights game. Not that I'm complaining 
I see your point ih8sens but I think what they are meaning is that not all use opening books etc and that people doesnt put the same amount of time into the moves.
This must be good for me as I do not use opening books etc.
mvh Fredrik
I just wanted to say that I found some sort of rules at support.chess.com/index.php?_m=knowledgebase&_a=viewarticle&kbarticleid=17
that says
You many only have ONE Chess.com member account. You may NOT get any help from any person or any chess engine that analyzes your specific position, including tablebases. You MAY use books, magazines, or other articles. You may also use computer databases (including Chess.com's Game Explorer). EXCEPTION: If both players agree for the use of a chess engine in an UNRATED game then it can be allowed.
Makes sense I suppose. I know my live chess (on this site) rating is considerably lower as the time controls speed up (I've become accustomed to having forever to figure things out and end up timing out a lot) but I'd have to argue that my CC level would probably hold true or nearly on any site. Of course, this can't be applied to everyone based on what you just said...
really a fascinating discussion. I guess the only way to figure it out for sure would be to join an 'official' correspondence chess association.
Which reminds me... anyone know how to get into National or International correspondence chess? ...
my chess.com rating is 300 lower than my USCF rating. my live.chess blitz rating is 300 lower than my WBCA rating.
so in my opinion, i would say that we are making a comparison between apples and oranges here. there are too many different factors involved for them to be comparable.
Just thought I'd add my two cents, since this seems to have not been mentioned yet in this thread.
The chess.com ratings use the Glicko system, which is intended as an "improved" version of the Elo rating system. In both systems, as I understand it, a rating of 1200 is considered the average rating. Note that this means that a player that rates 1200 is an average player for that specific population of chess players.
So, to put it another way, your rating on chess.com will differ from your score on FIDE (or wherever) by virtue of the fact that your ratings only reflect your position in the respective group against which you are measured.
As mentioned by other posters, the type of play used within any group of players (tournamet play, corsespondence play, etc) will affect the type of chess player population, hence also affecting the meaning of any particular players rating.
I play at four websites and have four ratings 1478, 1464, 1738 and here at chess.com my rating is an embarrassingly low 1082. I feel these ratings are a true reflection of how serious I play at these different websites. For example - at the website where my rating is highest, I play only in tournaments (as opposed to individual games). I try very hard to win the tournament - not just the games. I play very cautiously, study and take my time to be satisfied with each move I make. At the other two sites I play more casually. Here at Chess.com I've only played a few individual games and (sorry) but I haven't been very serious about them. Hopefully that will change.
As for an "official" rating? I have a USCF postal rating of 1366 and a OTB rating of 1200+ from 27 years ago. All these ratings make me an average player. I have no need for the official stamp of approval. I play for fun and it's unlikely I will ever be the champion of anything. But I beat everybody where I work except Bruce. (But he got fired - hehe)
I would recommend that you consider how your rating compares to your opponent's average rating. For example, the player rated 1900 who's average opponent's rating is under 1400 isn't as significant as a 1900 rated player with an average opponent rating above 1700. I believe this is why you find on this site that some 1300 and 1400 players play as well as 1600 and 1700 players. More than likely, these lower rated players have a higher average opponent rating than some higher ranked players. Of course, this is all just speculation. I have no hard data to back it up. 
And if what I believe is true, then I would suggest once your rating exceeds your average opponent's rating by 200 points, you may then want to seek out some higher rated players.
I try to play against as highly rated opponents as I am but as I am rahter new to the site does my rating still increase which means that my avarage opponent is a bit lower than my rating. I though believe that there is not such a big difference as 200... I am going to check that.
A good rule is to not play opponents under your current rating.
mvh Fredrik
I am pretty new to chess and I do not have a "real" rating, only the one given to me at chess.com (around 1650). Is this rating somewhat real or is it much too high? I really hope that it is not too wrong, and believe so as I have problem finding good opponents which might be explained through that there are not that many chess players in Sweden.
I would be very grateful if someone with a real rating could compare it with his or hers chess.com rating.
mvh Fredrik