If a clinical trial didn’t pre-register their stats plan and then publish their exact methods, they would be rejected by any reputable journal. That pollsters don’t do that undermines any claims of being “scientific”