One loves the protagonist's arc. Another thinks the ending needs work. One loves the comic relief; another wants to shoot that character in the face. How do you know who's right? What's an author-in-training to do?
FIRST: Walk away. Do something else for a couple days and don't, DON'T e-mail them back to explain how the description of the weather in the beginning of your novel is actually an ironic metaphor for life.
NEXT: Take each comment one at a time. You'll probably have one of three reactions requiring different responses:
Reaction 1) You agree. This is easy: make the change and move on.
Reaction 2) You don't care about the change one way or the other; it looks the same either way. I'll deal with that in a second.
Reaction 3) You disagree. In this case, you have to take into account the experience of the beta reader and how strongly you feel about what you wrote. Is the beta reader your editor? Then make the change unless you feel VERY strongly about it. Is it your good friend who hates your genre but wants to do you a favor? Consider the change, but don't feel bad about leaving it either.
Here's a chart to help.
(Mom, I'm just trying to be funny. I love your input, really. Please send cookies.)
Now what if it's write on the line, or you don't care about the change one way or the other (Reaction 2)? The easy thing is to let it go, leave it unchanged. After all, every comment you listen to means more work, right? But that's wrong, WRONG. Bad author! If you're not sure one way or the other, you should fix it. I'll explain why.
So, I used to work in the computer games industry (hence the charts and analytical...ness). On one project we had a bug that crashed the game. It happened randomly, but not very often, which made it very hard to track down. We developers figured it wasn't a big deal, since it only happened like once every 10 or 20 games. Our boss – a wise man – felt differently. “It's true,” he said, “that only one tester out of ten gets the bug. But what if 100,000 people buy our game? That's 10,000 people for whom the game will crash. 10,000 people who may never play our game again.”
Needless to say, we fixed the bug.
It's the same thing with your beta readers. Maybe you have 2 betas, or 5, or 10. If something bothers one of them, a good percentage of your readership may be bothered by the same thing. So unless you're very sure the beta reader is wrong, make the change.
(Note: This is also why, if two beta readers actually DO agree on something, you should make the change immediately. Assuming it's not the end of the world of course, which may also be true.)
So to sum up: if you agree with a comment, make the change. If you're not sure about a comment, make the change. If you disagree with a comment, make the change unless you're very sure it's good as is. And if two comments contradict each other, choose to agree with one of them.
You want to write the best story you can – why else would you be asking other people what's wrong with it? That being the case, do yourself a favor and listen.