Just speaking as a PhD Social Psychology student, stress reduces the cognitive resources available in an emergency situation. Some people handle it better than others, and maybe have more cognitive resources or had experienced previous emergencies which would have served as an "inoculation" for stress. This would mean that they handled the stress better and did not lose as many cognitive resources as others. The behavior of the captain also aided in lowering the level of stress for many of the passengers. He displayed calm and concern and so calmed the passengers he came into contact with. There is another dynamic at work where people look to the behavior of others as a cue to how they behave. Early in the disaster, when nobody knew what was going on, they were behaving more like a herd, waiting for the triggers for any other behaviors such as panic or orderly evacuation.

Addressing bsmith, when people are in groups, especially large ones, and they engage in some sort of ritualistic behavior, such as chanting at a game, they tend to feel more a part of the group and they lose their individual identity to an extent. Then when certain triggers occur, such as one person instigating violence against another group or breaking windows, they all feel the urge to join. It is even worse when you have some real instigators who elicit a response from law enforcement, which polarizes the group against the police and can cause more violence. At this point, more level headed people will try to escape, removing any positive influence from the group further radicalizing it.

There are observable and provable reasons for much of this behavior, but much depends on a great number of possible variables. This makes predicting behavior of large groups difficult unless you can identify some of the possible influences or catalysts. This is also why leadership and a strong individuality can influence things in the less violent direction.

Just my two cents (or maybe 10 cents in this case.