Sure, kids with married parents appear to have better outcomes by some measures. But a narrow reading of the data ignores strong evidence about the viability of alternative family structures.
We all know that most children of poverty who make it through college will end up better off than their parents, but what about the kids who start at the top?
Measuring the development of patches of Earth seems ridiculous. But that's exactly what we do. How might things differ if we measured income per natural instead of income per resident?