See first, think later, then test. But always see first. Otherwise you will only see what you were expecting. Most scientists forget that.