Let's say society prospers when there's more stuff and withers when there's less stuff. The traditional way to measure the amount of "stuff" is by looking at GDP (on a societal level) or GDP per capita (on an individual level). None of this has any obvious effect on the population of the country, so they're the same thing.
I'll also make the assumption that your lifetime earnings are a multiple of your entry-level salary, just to have an anchor for the numbers. Equivalently, assume that whenever you get a raise, it's defined as a percentage of your current salary.
With that assumption and the GDP definition of value-to-society, we can immediately see that it's 50% better to, by intervention, raise someone's starting salary from $50K to $80K than to raise a different person's starting salary from $10K to $30K. But it might be cheaper and easier to produce a 60% improvement in the $50K guy than a 200% improvement in the $10K guy.
I'll also make the assumption that your lifetime earnings are a multiple of your entry-level salary, just to have an anchor for the numbers. Equivalently, assume that whenever you get a raise, it's defined as a percentage of your current salary.
With that assumption and the GDP definition of value-to-society, we can immediately see that it's 50% better to, by intervention, raise someone's starting salary from $50K to $80K than to raise a different person's starting salary from $10K to $30K. But it might be cheaper and easier to produce a 60% improvement in the $50K guy than a 200% improvement in the $10K guy.