There was this thing on TV where a black male actor was cutting a bike chain with a saw or something like that. Many people confronted him. He said "The bike has been here for a long time." And they told him: "It's not your bike. You can't just take it."
Then they showed a white female actress doing the same exact thing. A couple people asked her what she was doing. She said to one person "I'm stealing the bike." They offered to help her.
I don't trust those things on TV. They seem so easy to slant in a direction that suits the producers. That's not to say that a bias wouldn't exist in that scenario. However, it's not like you can conclude that most people would stop a black man from stealing a bike, but help a white woman do the same, from the sample set that they CHOSE to show.
Among many issues:
* contrary to the narrator's statement, the two males are not dressed similarly
* they speak differently
* there is no black female control
And of course, being a TV show, there is no systematic statistical analysis.
Now, TBH even with more rigorous protocol I suspect the conclusions drawn would be similar, but as it stands, this "experiment" is not altogether useful.
It sounds like he's saying "people are more likely to intervene if they think they could take the other person in a fight". If you're afraid because of a size differential, you're more likely to stay silent (not "you" specifically, but the societal "you")
Then they showed a white female actress doing the same exact thing. A couple people asked her what she was doing. She said to one person "I'm stealing the bike." They offered to help her.