The COVID epidemic is a wonderful example of what quantitative data can do. I think that some of the doctors who prescribed ivermectin or hydroxychloroquine for their patients with COVID genuinely believed it was helping, because they saw so many people on those drugs do well. But the quantitative actual studies on both show them to be ineffective. (It is also clear that some of the doctors were pushing a political agenda and didn't particularly care at all about data, or about patient outcomes.)

All of us are too good at fooling ourselves, perhaps especially those that don't think that they are being fooled. We strongly register information that agrees with our views and disregard contradictory information. Treatments for mental health conditions are also particularly prone to placebo effects and other factors that make quantitative datas from double blinded studies essential if we are to make meaningful recommendations for treatment.