Today I have been reading commentary on the daily mail article about an experiment to have two child actors stand in a busy mall looking frightened and lost while their mother and researchers watched from a distance. The scandal was that no one stopped to offer help to the lost girls, despite their distressed appearance.
I question using this evidence in any way to characterize the nameless and briefly encountered people, or to characterize the broader culture they belong to. The commentary provides stories about fear of misinterpretation of actions, or of callous disregard of others, or even of fear of a possible sting operation.
I’d throw in my own theory based on the fact that it was a staged event and that people have an ability to detect staged or faked behaviors. They didn’t help because they understood that the girls weren’t really distressed. There is a lot of body language messaging that is not easily captured even on video. In person the event may have just had an aura that there was no need to offer help. There, I added my 2 cents.
But I add my commentary for the point that the data presented is not relevant to the conclusions we are trying to draw. The data is a staged event. From that data we can know a lot about the girls, the mother, and the researcher.
This concrete data is being reused to suggest something about the people who walk by.
This reuse of data is typical in big data analysis. It is reusing data outside of its original relevance. This makes the data dark. It does not really shed any light on the secondary questions being raised. But this is an excellent example how quickly we assume that the data is not only relevant but conclusive about the strangers who happened to cross the camera view.
I use the word dark data to be data with an interpretation that we readily accept but without good support. It is an abuse of data.
The same observational evidence can be used in a positive way by suggesting a hypothesis. This is what I call hypothesis discovery. This little experiment illustrates a hypothesis that strangers are afraid to assist when confronted with a child in distress.
The appropriate next step is to test that hypothesis with a controlled experiment that focuses on the strangers and their impressions of the situation’s call of their attention and the options they think available for them to pursue.
In this day of staged events with the high rewards for creating viral videos, it may be more reasonable to assume this scenario is a trick than it is a real call for help. With that bias, the people who were there may have been more alert to notice clues that it was all staged.
To me this example tells us more about the people who weren’t there at all. The people who read about the event and jump to conclusions that are no way supported by the evidence. That really scares me.