This explains some aspects of bullying, but not all. Specifically, quoting the article, why "those who are not necessarily friends, but who share many friends i

This explains some aspects of bullying, but not all. Specifically, quoting the article, why "those who are not necessarily friends, but who share many friends in common — are also more likely to bully or otherwise victimize each other"? Maybe people you don't have enough friends in common with feel like a different herd, and people have instict to bully the weakest member of their own herd, but don't care about other herds as much. Maybe it's because bullying a member of your herd allows you to rise to the top of your herd (which is the assumption of the article), while bullying a member of a different herd does not. Or maybe because when you start bullying a member of a different herd, the whole victim's herd may collectively turn against you ("yeah, he may be a loser, but he is our loser, so back off"), while bullying a member of your own herd is safe in this aspect. (Unless your herd has a strong leader who opposes internal fighting -- rarely happens with kids.) So the testable prediction would be like: If you have a classroom with e.g. 29 white students and 1 black student, the white students will bully the black one. (Unless there is someone more visibly different, e.g. handicapped or speaking a foreign language.) But if you have 15 white students and 15 black students, more likely the white students will pick a white victim to bully, and the black students will pick a black victim to bully. (Unless the class is more split along some other trait than race. In which case e.g. the English-speaking students would pick an English-speaking victim, and the French-speaking students would pick a French-speaking victim.)