It seems to me that you have cause and effect backwards. I think the most likely chain of events is something like this: Member gets sick Member gets rejected

It seems to me that you have cause and effect backwards. I think the most likely chain of events is something like this: Member gets sick Member gets rejected from the group (because for each individual in the group, it's better to avoid the sick member - no group selection necessary) Member dies So the member doesn't get sick because they're rejected - they get rejected because they are sick. This would have obvious benefits for each individual that rejects the sick member, regardless of the behavior of the rest of the group (https://esajournals.onlinelibrary.wiley.com/doi/abs/10.2307/1941192 https://esajournals.onlinelibrary.wiley.com/doi/abs/10.2307/...)