Being a geek is a reason to find yourself at the bottom of the hierarchy, but so is being overweight, wearing glasses, being shy, being in a religious minority, etc. The particular reason doesn't matter to the bullying process, it's a free variable. I'm pretty sure that a school full of geek kids will find their own reasons to bully some of its students.
First, note that the standard the school has to meet isn't "no bullying", but rather "significantly less bullying". I presume this is what you mean.
Anyway: What is your evidence for this claim? Have you ever spent time in such a school and compared it to a regular school?
Data point of N=1. I went to a special "school for smart people only". There was dramatically less bullying than the regular-people school I came from.
But anyway the researchers are providing actual scientific evidence of this effect, and you appear to be arguing against it using, apparently, intuition based on no evidence whatsoever. "Yes, these experts say X, but that sounds unbelievable to me; therefore they must be wrong."
Surely you do not endorse this sort of reasoning.
So, if we trap the geeks where they can be bullied and beaten, they'll understand what the real world is like better? I'll be real with you, I've never been hated for expertise or diligence in real life, not like the way bullies work today, which is way worse than it was for me. And I know where some of those bullies end up. They did not benefit from their jeering and hatred. Can we stop using this to justify the torture of society's most promising youths?