In research design, every comparison starts with two competing statements, one assuming nothing has changed and one assuming something has. Let's place each option into that picture.
- Hypothesis of association: Used when a study is checking whether two factors move together, not whether groups differ, so it answers a different question altogether.
- Null hypothesis: This is the default starting position in any statistical test. It says the groups being compared are, in truth, the same, and whatever gap we see in the data is random noise.
- Hypothesis of differences: This label suggests a difference is already assumed, which is the reverse of a "no difference" statement.
- Alternative hypothesis: This is what a researcher hopes to support with evidence, stating that a real difference or effect is present, again the reverse of what the question asks.
Every statistical test is built to try to reject the null hypothesis in favor of the alternative, which only makes sense if the null hypothesis is the one claiming no difference. The correct answer is Null hypothesis.