When you see that an editor has changed a clue, the first thing you should ask yourself is whether the new clue is better. Try to be honest with yourself. If your answer is that it is better then you've discovered the reason for the change.
If you feel that the new clue is not an improvement, then here are some of the things you should check about your original clue/clues:
1) Look for part-of-speech mismatches. If you've clued HUNGRY as [Wants food], your clue will be rejected every time. Try the substitution test. He ___ enough to eat a horse. [Wants food] clues IS HUNGRY, not HUNGRY. This is the single most common mistake I see from rookies.
2) Check if your clue is quite long. Editors have space constraints. There's sometimes a tendency for newcomers to write mini-essays for clues.
3) Check if your clues are too short. A lot of one-word clues makes for a boring set. Consider the oft-appearing entry EAT. [Consume] is probably the most boring clue there is. [Chow down/Have a bite/Fuel up], etc. are all livelier.
4) If you thought your puzzle was a Wednesday/Thursday offering and it winds up running on Monday or Tuesday, chances are over 50% of the clues will have been changed to ease things up. Conversely, if you guessed Monday/Tuesday and your puzzle runs on Wednesday or later, many of the clues will have been toughened up.
5) Ask yourself if your clue was pedantic. Clues taken straight from the dictionary are often stiff and formal.
6) Have you gone overboard on fill-in-the-blank clues?
7) Have you gone overboard with question mark clues of the starter/ender/opener/closer, etc. variety. Too often clues of this type are poor. Say you clue BOARD as [Room starter?] What does "Room starter" mean? Your clue needs to make some kind of sense for it to work. Now if you used [Room opener?], there is some sense, but these types of clues can be overused. Try not to get carried away with them.
8) Even for easy puzzles, editors don't like the clues for longer entries to be gimmes. Did you blow it?
9) Was your clue absolutely accurate?
10) Was your clue convoluted? Many new constructors have trouble putting themselves in the solvers' shoes. They expect solvers to be mind readers. If it requires a complex chain of reasoning to unravel a clue, odds are that clue will never see the light of day.
11) Was your grid name-heavy? If so, an entry like BOB or HOPE that you clued as a name might be changed to a non-name clue.
12) Did you have a bunch of tricky clues in the same sector? For Monday-Thursday puzzles there is such a phenomenon as too much of a good thing.
These are just off the top of my head. If you approach an edited set of clues with the idea of improving your skills, you'll do that. I'm not trying to tell you that you'll never disagree with an editor's decision. That happens to all of us. But more often than not, the new clue will be an improvement.
Try to remember that you are only one among a slew of constructors that an editor is dealing with at any given time. If you expect a dialog over every clue change then, to be honest, you'd better find a new hobby. (Also remember how many puzzles an editor looks at when you correspond with an editor about your puzzle. Don't even assume that the editor remembers the puzzle much less every last detail about it.)