It's worth pointing out that the majority of bystander effect studies are lab-based, and involve asking WEIRD [0] undergraduate students what they'd do in a hyp

It's worth pointing out that the majority of bystander effect studies are lab-based, and involve asking WEIRD [0] undergraduate students what they'd do in a hypothetical situation. The cited study in the grandparent comment is an analysis of security camera footage in Amsterdam, Cape Town (!), and Lancaster, UK, and so is not synthetically asking people about what they reckon would happen, but is instead an examination of what does actually happen. Not that street fights in two very (and one relatively) typically Western cities don't represent a biasing limitation of sample, but it's still more applicable, IMO. [0] https://en.wikipedia.org/wiki/Psychology#WEIRD_bias https://en.wikipedia.org/wiki/Psychology#WEIRD_bias