It is customary to formulate a "null hypothesis", H0, the logical opposite to the proposition under test, HA.
"A null hypothesis test is a ritualised exercise in devil's advocacy" and is really designed to eliminate the simplest criticism of any study, "it could be explained by chance alone". Ideally you should test your hypothesis directly; ie if the effect size is sufficient and the data adequate then direct inspection would show if the null hypothesis was falsified. The null hypothesis can only be falsified and the failure to do so does not "prove" the hypothesis.
Statistics is useful in those situations where data is reliably random and the discrimination between positive and negative outcomes is not clear. See the Normal Distribution. It is useful to remember that "statistics don't prove anything, they just show how surprised you are"! Tossing a coin 1000 times might result in 1000 heads; you may be so surprised you test the balance of the coin but the outcome of the coin tosses alone does not "prove" the coin is biased. The null hypothesis and statistics are ways of exploring the problem and must always be related to the real world; don't lose track of what you are really wanting to do.


