Evidence, of course. The thing is that the George Williams wing claimed that apparent altruism could not exist because their model precluded it. So I was hammering on the model.
Without a model that generates testable hypotheses (hopefully framed in numerical terms), the evidence isn’t actually evidence.
I will admit that I am only talking about groups that are made up of individuals. (Again, of course.) Some groups of individuals act one way. Some act another. Over the last some hundreds of thousand years, groups of individuals that showed solidarity tended to live and those that splintered died under selective pressure that included inter-tribal warfare.
I started out to write a book about all this but got saved from the effort by actual pros. I suggest checking out Unto Others by Sober and Wilson and the work Sam Bowles at the Sante Fe Institute. (There are, also, some sources in the sidecar article below.)
Some of my favorite non-human evidence comes from parasites. (Though accounting for human’s extreme eusociality without other species as evidence would be enough.) Parasites are good because folks tend to believe that the negative has to be more real than the positive and altruism is too peace, love, and understanding for true hard-butt scientists types to consider. Parasites are icky. That’s a good counterbalancing force.
There are parasitic infections that cycle through ants and cows. One of a genetically isolated group (haystack = parasites inside the ant all hatched from a single egg mass) sacrifices itself for the group and takes over the brain of the ant causing a suicidal mission by that ant to get itself and the rest of the group of parasites eaten by a cow completing the parasites’ life cycle. During that process, the haystacks are melded back into a single large population hence matching the structure that the model predicts as necessary for this type of behavior to evolve.