[lapsedacademic]

> Stop this.... I question the statistical literacy...

Ugh. Why?

You asked for a statistically grounded conversation, so let's do that.

Let's start, for example, with pdf page 25 (and surrounding context) of https://davidcard.berkeley.edu/papers/causal_educ_earnings.p...

The analysis done in that paper is a good starting point for a productive conversation. We could discuss the bounds on various coefficients and decide whether the conditional statements about those coefficients made in the paper have clear answers in either direction. Or we could critique the various modeling assumptions. Etc.

[hooande]

Section 3.6. "Family background" of the pdf you linked to discusses the impact that parent's educational attainment has on how much money someone earns as a result of going to college, not whether they attend college. It's not used as a control in a way that's relevant to this discussion.

Do you have another study?

Either way, it's a fact that parental income is the best predictor of future income. Not educational attainment.

[lapsedacademic]

> best predictor

But why? What are the CAUSAL relationships between parental earnings, educational attainment, and child earnings? The children of doctors are more likely to become doctors, but saying that educational attainment is therefore less related to doctoring than parental occupation is obviously a bit absurd. Just because a parent paves the path doesn't mean that educational attainment is irrelevant to walking that path. And anyone who makes it through med school and residency has the option to enjoy high earnings, regardless of parental income.

The MD example, for the curious and humble reader interested in Truth rather than Winning, makes it abundantly clear why section 3.6 of the linked paper asks a question that's directly relevant to untangling these causal links.

> Do you have another study?

There's an entire literature base on exactly this question. "lifetime earnings parental earnings education" returns 130K results on Google Scholar. But, to be blunt, I don't think you're interested in learning anything. I think you're interested in Winning the thread. So I'm not posting for your benefit; that would be futile. I'm posting for the benefit of intellectually curious readers.

[hooande]

This is a particularly pernicious misunderstanding because it leads people to believe that they have to take out loans to go to college or they will earn less money. Saying "People who do X make more money" can have consequences if that statement isn't necessarily true.

What you want is a study that shows that people from lower income quintiles that go to college have a higher lifetime earning than people from the same quintile that didn't go to college. Maybe that exists? if it did, I'd imagine the pro college people would be waving it around everywhere.

Using Google Scholar to find relevant research is a great habit. but you really have to read it to make sure it says what you think it says

[lapsedacademic]

> What you want is a study that shows that people from lower income quintiles that go to college have a higher lifetime earning than people from the same quintile that didn't go to college. Maybe that exists? if it did, I'd imagine the pro college people would be waving it around everywhere.

Yes, there is a large college wage premium for students in lower income quintiles. The most that can be said is that it's smaller, but still quite large.

I assumed the point of contention was a more nuanced question about causation, since the above is just a simple factual question that can be checked without any sort of analysis.

[caddemon]

I agree "people who go to college make more money" is not a helpful thing to be telling kids, but I think it would be much more fruitful to pose the question as comparing the outcomes of different fields of study (which could also include specific trades), rather than questioning the utility of college entirely.

[shkkmo]

> What are the CAUSAL relationships between parental earnings, educational attainment, and child earnings?

Social network, safety net, family experience with college, etc.... There are plenty of reasons why class mobility is imperfect. [Edit: I, for example, had access to summer jobs in highschool through my parents' professional network that were not as easily available to other people.]

> There's an entire literature base on exactly this question. "lifetime earnings parental earnings education" returns 130K results on Google Scholar.

Yes, but you chose a specific article to post to refute a specific claim. The article doesn't address that claim, so it is entirely reasonable to ask for a citation that does actually back up your argument. Your response here amounts to: "just go read the all the literature until you see I'm right" and is not constructive, even without the name calling.

Edit: You seem to have substantially edited your comment. Thanks for removing the name calling but generally ghost edits like this are frowned upon here.

[lapsedacademic]

> Yes, but you chose a specific article to post to refute a specific claim.

Yes it does! I think you're misreading OP's post.

What was OP's claim?

>> The most accurate predictor of a person's lifetime income is the income of their parents. Children of wealthy parents are more likely to go to college. It's like saying "People who drive expensive cars in high school make more money over their lifetime, period".

OP's assertion about "best predictor" is true but irrelevant. The interesting question is why?

OP asserts that the answer to that question is literally "for the same reason that rich kids drive BMWs".

OP is asserting that college has the same causal effect as a parent purchasing a BMW for a child. I.e., none at all, it's just a proxy for parental wealth.

That strikes me as an unlikely causal hypothesis.

Could there perhaps be a reason other than parent income that the child of an MD drives a BMW to school? Probably not.

But could there perhaps be a reason other than parent income that the child of an MD does well in their premed program? Seems likely.

And indeed, the above article establishes a causal link that's directly relevant to falsifying that assertion, that college == bmw in terms of causal effect.

Elsewhere, OP asks if the college wage premium persists across family backgrounds. I think perhaps something related to that question is what you perhaps read into their post. But that's not actually the claim they are actually making in that post.

(BTW: CWP and PEP are positive for students from low income backgrounds... these are just numbers you can look up... why am I the thread secretary for basic statistics?)

[shkkmo]

I simply do not see a anything in that study that refutes that college is just a proxy for parental income. The study only discusses parental background in terms of parental education and I don't see any controlling for parental income (though those two factors are clearly correlated, but are not identical and conflate them in several places.)

In reality, a significant part of the correlation between of college and is indeed due to college being a partial proxy for parental wealth. At the same time a significant part of the correlation between parental wealth and child income is to the that same proxy.

Even when you control for parental wealth, there are large heterogeneities in the effect of college on income in different groups. This makes it hard to argue for a simple, direct causal link between college and income.

While I think you and me tend to agree on this subject, I think you should focus less on being the "thread secretary" and more on understanding the opposing argument and clearly explaining your argument rather than posting dense statistical papers with no analysis and using abstruse acronyms.

[DangitBobby]

Let's say there's a social norm to "go to college if you are smart enough or hard-working enough" for lower income families and "no matter what" for the wealthy (since all you need to be successful anywhere is wealth). If that were the case, they would be self-selecting into college on the basis of their own perceived ability to succeed there, confounding other related measures.

[8note]

I'd guess that the why is irrelevant too.

Why? The parent got their child a tutor. Why can't poor kids have that? They're parents can't afford it.

Over and over again, better food, better housing stability, not having to work while doing school, etc.

The cause still comes back to some parents being wealthier than others.

It's classist to think wealthier parents are more virtuous in some way than their poorer peers

[mlyle]

I think focusing just on resources misses part of it.

Upper class parents know how to raise children to present as higher class because they have the benefit of having been raised and lived in that class.

Being able to spend time on my kids helped, but they also entered school at a high level in math and reading and with the diction of a higher class because I knew how to teach this to them.

Some of the knowledge of how to succeed in education and develop children's minds is unevenly distributed, and it's not something easily fixed by just committing resources. (Though committing resources surely helps).

[darkerside]

Just have to say, what a shitty way to end your comment. You've poisoned the conversation, and I think you read into something that wasn't there. The other commenter took the high road by ignoring it.

[shkkmo]

> Either way, it's a fact that parental income is the best predictor of future income. Not educational attainment.

Sure, but they are interrelated factors and they way they effect the distribution is complicated. This study was linked elsewhere and does control for parent's income: (I didn't vet the methodology or data, just looking at what their reported results say.)

https://research.upjohn.org/empl_research/vol23/iss3/1/

One of the reasons that parental income is such a strong predictor of child income is because parental income has a strong effect on how much college will increase your income.

Interestingly enough, that effect is quite disparate based on more than just parental income.

The study says that low income whites see only a 12% boost to income from college while high income whites see a 131% boost to income from college. Interestingly, blacks show an even higher boost to income from college, 175%, and parental income had no statistically significant effect on this boost.

Also interesting is how those effects play out when you look at different parts of the income distribution. Parental income increases the average effect of college, but doesn't significantly affect the median effect. Thus a lot of the increase to the effect of college on average incomes [edit: for children of higher income parents] is from gaining access to the long tail of very high income outcomes.

So the answer is if you are a poor white male, college is far less valuable than if you are female, rich or black (in increasing order of college effect size.)