This is mostly correct, but you have misunderstood one detail. You are correct that the null hypothesis is accepted without evidence rejected with strong evidence supporting the alternative. But there is no requirement about which statement will be the null and which the alternative. Which hypothesis is the null and which the alternative is at the discretion of the researcher conducting the test.
While this is in fact true, it's not the entire truth. The whole point of a null hypothesis is that one must demonstrate that there's something worth explaining prior to offering any explanation--you have to prove that there's something there before you talk about what that something is. This means that some null hypotheses are better than others. For example, if I give you population curves over geologic time, a good null hypothesis is that they're driven by random chance, that all we're seeing is statistical noise. Once we demonstrate that it's NOT statistical noise, we can move on to discusing what is actually happening. A bad null hypothesis is that the population curves are driven by predation pressures.
Using "God exists" as a null hypothesis is even worse than that, because it presumes the very hypothesis we're trying to determine the validity of. This is nothing short of begging the question. Don't get me wrong, there are some cases where this is useful. For example, we could say "Let's assume that the Muslim god exists. What would our universe be like?" This is no longer a null hypothesis, however; it's into Strong Inferrence at this point. We have two competing hypotheses (the Muslim god exists, or it doesn't), which are mutually exclusive, cover the entire range of possible outcomes (meaning that the Muslim god can either exist or not; proof that the Muslim god doesn't exist does not necessarily have any bearing on any other god), and there are tests which can clearly show which is correct. This has nothing to do with null hypotheses, however; it's a completely different type of hypothesis-generation.
The only good null hypothesis in the case of "____ exists" is "____ does not exist". Otherwise, again, you're simply begging the question.
Malerin said:
The theist will say that evidence of God's work is all around them, since a foundational belief of theirs is God created the universe.
And this shows why Beth is wrong. As soon as you say "The evidence supports my possition" you are by definition NOT talking about a null hypothesis. You're arguing that the null hypothesis is wrong, and that your possition is correct (two different arguments, but they are both assumed).
The atheist would say that there's no evidence for God since a foundational belief for them is the universe was created through natural processes.
There's no belief involved. We can conclude that there are no gods because there has not been enough evidence presented, of sufficient quality, to cause us to reject the null hypothesis. This has nothing to do with belief; it's a simple matter of not begging the question.
Lamuella said:
the thing is, for me at least, the skeptical position isn't "I have seen no evidence for X, therefore it does not exist". It's more like "I have seen no evidence so far that suggests X, therefore there is no reason yet to act like X exists"
Exactly.
.