I wonder how you could design such a study. The population, even on this board, that is following any semblance of paleo is so scattered with so many different versions, that that would have to be part of the mathematical function. What would the control variable be? What sort of design for tracking would you put in place? In the larger scheme of things, IMO, a very worth-while subject to study with a really good logistical road-map.
I do think that the impact of this diet and way of living can be at least tracked with simple metrics regarding common disease incidence in Paleo vs. SAD, since an improved immune system is one of the benefits. We talk most frequently here about the big five (cancer, diabetes, hyper-tension, heart disease and high cholesterol), but what about the common cold and flu? For years, I had maybe one cold a year and in the last two years, I guess, I caught everything I came in contact with until the paleo benefits kicked in in January. I've been incredibly well since. I think with a good control group, a study of this sort is totally doable.
That being said, what is your threshold for "longevity"? I have a 99 year old aunt who smoked, ate SAD and bowled for exercise. I have a friend who's father is 105 and just had a second hip replacement. His diet most likely leans vegetarian. In both cases, getting to 85 in the family is perfectly normal. I suppose I'm asking regarding the initial hypothesis, is genetics going to be considered, because we all have family that abuse the heck out of their bodies and somehow live to tell the tale, albeit on anti-hyperintensives, chemo, insulin, etc.
I hope you do start such a study that tracks 2-3 generations on paleo or at least grain, legume and simple sugar (for the most part) free. I think the results would be fascinating.