You need to do all the background research to be reasonably confident that your hypothesis doesn’t suck. And more research to set up your experimental procedure and statistics so you can actually get a useful answer.
This is where many published studies fall short. They simply have not used the correct experimental methods or statistics to draw any meaningful conclusion from their research.
This happens far more that you would guess.
Make no mistake, just because a piece of research is published, doesn’t mean it’s correct, or even well done.
Behind every research paper are people, and people no matter how well trained make mistakes, and sometimes these mistakes end up in published research.
Retesting your hypothesis is the key to determining a real answer. If you can get a REPEATABLE result then you’re on the right track. If you can never re-create the effect then your hypothesis needs work and something has to change.