Everything on this page is made up. The people are simulated, and we wrote the patterns into them ourselves, so these results say nothing about real people. It runs the real rules on invented data, to help you imagine the research. The door is not open yet.
One of these groups is smaller than 50 people, so nothing comes back: no share, no count, not even how small it was. The code refuses it; no reviewer has to notice.
You write down the question, the hypothesis and the decision rule first, and it is published with a timestamp before it runs.
It is reviewed by a person. Miss Aquarius℠ never chooses which questions are asked.
Only finished seasons are studied, so no answer is about the present.
It runs inside the service, every read logged. Only the comparison leaves.
Every group is at least 50 people, every answer carries noise, and each season has a fixed budget, so many small questions cannot add up to one person.