2 ms·
The post I was replying to has no statistical claims. My example was illustrative; the broader point is that to advance a hypothesis that riots accompany withdr
by _delirium 15y ago
The post I was replying to has no statistical claims. My example was illustrative; the broader point is that to advance a hypothesis that riots accompany withdrawal of social safety nets, but not their absence, you would need to show a general lack of riots in periods where social safety nets didn't exist in the first place. But there are thousands of examples of riots in countries and periods with no social safety nets; the Haymarket riots are just one random example.
Perhaps it is possible to perform a statistical analysis that shows that the absence of social safety nets does still, in general, produce stability, but I haven't seen one put forth. The linked article certainly does not include any data relevant to that hypothesis.
- yummyfajitas 15y agoThe original article claims that a % change in social benefits is a strong predictor of riots. I'm assuming that by choosing to focus on this factor, the article intends to imply that it is the strongest predictor within their data set. If that is the case, then the number of riots is related more strongly to withdrawal of benefits than to their base levels. This would support joshuaheard's claims, though of course it doesn't rule out the possibility that riots(social benefit levels) is a nonlinear function, increasing near 0 and flat in the regime where the original dataset lived.