Accident stats don't support the image of fragile tin can some people hold dear.
Physics cuts both ways. Smaller cars in a single car accident have less energy to dissipate and are less prone to roll overs. Roll overs do tend to cause higher fatality rates in SUVs/CUVs than in cars for those types of accidents.
That has nothing to do with survivability. The car can be structurally rigid and the shell may be engineered to sustain great loads but the human body has a limit at which brain and internal organs turn into a mush. It's not the fall that kills but the sudden stop, and the rate of deceleration is much higher in a smaller car.
Rollovers are different in nature and, unfortunately, many choose not to wear a seatbelt thinking they are invincible in a large SUV whereas the opposite is true. In a rollover, the larger taller vehicle generates much higher centrifugal forces causing more damage to the occupants, especially if they are not wearing a seatbelt. I'd be curious to see the stats on that and see the percentage of fatalities with occupants not wearing the seatbelt.
If the fatality is caused by ignorance and not wearing the seatbelt then it's not the car's fault, and the statistics are just that.
Can you produce a report that shows the statistics that apply specifically to me and my driving habits? Guess not.
Since I always do wear my seatbelt, and will not move the car until every occupant in my car is strapped, I feel safer in a larger vehicle.
That's the choice I've made based on what I think is common sense and my personal interpretation of all available statistical information out there.
If someone never wears a seatbelt and never drives on two way streets or highways then they actually may be better off in a small car with 11 airbags according to statistics.
I hate statistics, they're like a prostitute that politicians pay to sway public opinion for their own benefit. They can even "prove" that a small car is safer than the big car! How 'bout that!