I think there is something to be said about provenance of the degree. For example there's been quite a lot of expansion in number of colleges or even community colleges expanding their systems while actual prestigious colleges themselves have only expanded so much.

Here are the stats for Harvard enrollment of undergrads (1,3), along with US population (2,4) and percent Harvard student (not sure where I get number of people in the workforce with harvard degrees data but maybe this is a decent proxy):

Year - ugrads - US population - % of US pop at harvard

1990 - 22,851 - 248,709,873 - 0.0092%

2000 - 24,279 - 281,421,906 - 0.0086%

2010 - 27,594 - 308,745,538 - 0.0089%

2025 - 24,519 - 343,000,000 - 0.0071%

1. https://nces.ed.gov/programs/digest/d13/tables/dt13_312.20.a...

2. https://www.census.gov/data/tables/time-series/dec/popchange...

3. https://www.harvard.edu/about/

4. https://www.census.gov/popclock/

Let me just write this down… Just for illustration, assume average lifespan of poor person is 60 and average lifespan of rich person is 80, 30% of the population is from the rich person group and rest from the poor person group, and these two facts hold for current time and the 90s.

Let’s say currently, every rich person goes to college, so college to non-college lifespan is 80:60.

While in the 90s, let’s say 20% goes to college and every college going person is rich. Then the lifespan of college going person would still be 80 and non-college going person would be more than 60.

So, another way of looking at it is that the non-college going population is getting to be the special demographic whose statistics are getting skewed, though I’m not sure that’s the correct way of looking at it.

It might be a skewed distribution where life expectancy drops off rapidly below the median but isn’t that different at the top. So it’s not a big difference when it’s the bottom 90% and the top 10%, but it is when it’s the bottom 60% versus the top 40%.

Well, sure, but generally when your hypothesis demands a highly non-linear distribution function to make sense, it's just wrong. That might be true; the math could be made to work. But if it were, that is the result the study would be pushing and not the bland thing about smoking.

It's a smell test thing, basically.

Why shouldn't it be non-linear? Non-linear is what I would expect--the educated know enough to avoid a range of stupid behaviors that lower life expectancy. You see few of the life-destructive behaviors in the degreed group. You also would have seen the same effect in many of the non-degreed individuals but the data didn't separate them. More degrees, more of the careful people move from the non-degreed pool to the degreed pool and the gap between the pools rises.

There's also the factor that simply getting a degree screens out many of the people that engaged in such behaviors.

I think that's basically correct, but you don't even have to put it in quite those terms. Life expectancy measures are dominated by early deaths. The group of people without a college degree probably includes most of those who are likely to die early (teen moms, structurally unemployed, lumberjacks or electrical linemen, etc.) It also includes a bunch of people without those risk factors (administrative assistants, etc). So as more of the latter group get a degree, the high-risk population comprises a greater and greater share of the group without a degree.

I don't think it would require a weird distribution. Average life expectancy is dominated not by people living a really long time, but by a minority of people dying really early (due to child/maternal mortality and work-related or accidental deaths). If those people are concentrated among those without a college degree, then you'd see the life expectancy of people without a college degree declining as a greater share of the population goes to college.