@garyackerman How would you know the data is invalid? You will never have perfect knowledge of any situation, so you can only do the best you can with what you have.

It’s easy to say the data was invalid in retrospect. The eternal challenge is in knowing the data is invalid *beforehand*. The difference between the two is the difference between a novice and an expert.

@garyackerman I guess it depends on the alternative. I prefer date driven decisions even ones where the data is no better than random chance to decisions made based on what will move the most money into the pockets of the powerful. On the other hand if the graft is the reason the data is skewed, it is better to case lots than use it.