How do I design a study that actually answers my question?

You can analyze data once you have it. Designing the study that produces the right data is a different and harder skill, because it happens before there is anything to crunch. The analysis works with what exists. The design decides what will exist, and that is where the real thinking lives. The heart of good design is not technique. It is the question. Most weak studies come from a question that was never made sharp enough to answer. If you cannot say exactly what result would count as yes and what would count as no, no amount of clever method will save it. Clarity about the question comes first. So start there. Write your question in a single plain sentence, then ask: what would I need to observe to answer it clearly, and what could fool me into the wrong answer? Design backward from that. The method follows the question, not the other way around. Get the question honest and specific, and the shape of the study begins to reveal itself.