How to state a testable null hypothesis?
From what I understand, a null hypothesis (unlike an alternate hypothesis), is a statement that reflects the status quo. In other words Ho (the symbol) is a statement used to determine whether the question need to pursue some alternate reasoning or hypothesis, but if the Ho is fine, then the original statement stands true. Though, the null is quantitative, so you have to compare it with something else. Null could mean that today is indeed nice, but how do we define nice? Accepting null could mean that today turns out to be nice and/or today is even better than nice. So, because this is a statistical question, there is always a second side to the answer.
