If you're dealing with an unknown, confusing phenomenon, you don't know what the underlying hypothesis is. You don't know yet what constitutes rigorous testing

If you're dealing with an unknown, confusing phenomenon, you don't know what the underlying hypothesis is. You don't know yet what constitutes rigorous testing and what is just wandering off into the weeds. So you need to start with seeing if the thing happens at all. You do in fact need to know whether the previous result was garbage before it can be productive to do a wide battery of tests. And "wide" depends on how sensitive to conditions your effect is, which you also don't know yet. Your closest thing to an anchor is the initial claimed setup, so start there.