For example: drug use. Many people use drugs, but what is the difference between them? Is one type of drug a 'good' drug, like caffeine or nicotine? Burning Man attendees may be good employees and may micro-dose with LSD while at work. Does this make them a good hire? Again, the issue is that there just isn't enough data to properly correlate.
Social media can be very 'noisy' and should be one of a number of factors that are used in decisions.
I do. I often interview candidates in my role as a software developer at my company. Finding them on social media would be trivial. But I don't look for them there. I don't care.
I don't even want to know their politics before giving my recommendation and I'm deeply suspicious of anybody who thinks they would want to know that information.
The time the average hire spends with the company is short enough that having to fire the occasional qualified employee early over "culture fit" really isn't a big deal.
If they're posting nazi stuff on Facebook, everybody knows they're nazis.
Companies aren't going to ignore "free information available at your fingertips" about the candidates they hire.
For example a problem candidate may have easily accessible "red flags" on their social media such as racist stuff and drug use.
They could "save" themselves some costs and future problems by knowing that ahead of time.
When companies are trying to optimise their hiring process they aren't going to ignore free additional information about the candidates, they want every last piece they can get their hands on so they can (at least in their mind) make the optimal decision.