The Economist writes about "The Kindness of Crowds" with empirical data extracted from CCTV footage in UK here. The conclusions drawn are around crowds, voilence and its self management only.
The point to note is that the data shows crowds alleviating volatile situation through external intervention by a by-stander in most cases. Should the by-stander be considered part of the crowd? I do not think so. A by-stander chooses to help or not help.
Please see my previous post "Debunking "Wisdom of Crowds" Myth - Two Examples" here.