Education pays. Or being clever gets you more education, and also pays. Or being willing to work hard does those things. Or coming from a family with more money, or with a particular kind of attitude (curiosity? go-getting? ...), does those things. Or there's a lot of prejudice out there (on the basis of sex, skin colour, or whatever) that affects both how a person's education goes and how their later career does. Or any combination of those. Or any number of other possibilities.
All together, now: Correlation is not the same thing as causation.
I'd love to see Wages = b0 + b1 * Bachelor's degree dummy + b2 * parents income + b3 * IQ + b4 * [thing I can't say] + b5 * [other thing I can't say] + b5 * work ethic measure (HS GPA?) + e There are a million things wrong with doing it this way but it would be an interesting addition to looking at that graph by itself. I half-heartedly tried to Google a decent paper but didn't have much luck with anything newer than the 70's
I propose a businessman whose output cannot be measured very well earns p/r+log(e)+k, where p is the profit the company makes, r is his rank within the company (CEO is 1, CFO/CMO is 2, etc), e is the number of employees beneath him, and k is the base salary at that firm.
The idea here is that upper management at certain companies effectively own the firm and distribute profit accordingly, while middle management usually advances in income by increasing empire.
Measuring the effect of education is one of the most difficult econometrics tasks due to the way that most societies pursue education -- one year at a time until you're done with it. There's not enough variation.... I was about to launch into a detailed technical explanation, but I think I'll just leave it at the intuitive one.
It's much worse than that. The linked article is about the 'median wage' at each of those education levels. I can't imagine a metric less interesting to HN than 'median performance.'
Show me the top 10% of earners, then give me the breakdown of how much traditional education those people consumed. Hustlers aren't getting by on their credentials.
It's wrong because average people are more likely to do average things in their life, like pursuing a college degree because that's the common thing to do.
Outliers on the other hand sometimes choose their own path. Sometimes they drop out of college because of financial problems, but still attend courses that interest them. Sometimes they choose to learn on the job, by making mistakes and fixing them, by screwing around disassembling stuff, by learning from mentors they admire, by trying out new things and so on and so forth.
Saying that Gates, Jobs, Zuckerberg are college dropouts really ignores that these people are/were more educated than the hordes of people coming out with master degrees. And this I think is the biggest problem with our educational system, the fact that these alternate paths aren't properly recognized.
Disclaimer: I'm also a college dropout, I regret it and intend to finish college, on the other hand I have a monthly revenue at least 3 times bigger than the software industry's medium salary in my city for senior developers.
What are you suggesting as a preferred measure of central tendency if not median? The usual suspects include arithmetic mean (strong outlier bias, especially for incomes), or mode (which has other problems).
Or are you suggesting that measures of central tendency are inappropriate for this sort of analysis altogether? In which case, what population/sampling moments would you use instead?
I'd suggest that 2 is false. For a policy analysis, you're interested in either total (mean) or typical (median) outcomes. Probably median. You might toss in some floor/ceiling limits as well (if a particular course of action has a highly bi-modal tendency, with one mode being particularly low, it's not so hot).
As to my own feelings on the matter: a good college education is really valuable. I've certainly met people who are brilliant and/or successful who haven't finished, or even in some cases attended college. One of the biggest problems (and it ultimately does become a problem) with the latter is what I call the "chip on the shoulder" issue: it's an awareness that the person has that they haven't been to college and a mix of compensation and confidence issues resulting.
Much of what you learn in college comes from the context: who you're exposed to, what you're exposed to, library access (much less significant today than 20 years ago, given the Internet and its resources, but there are still huge amounts of information not readily available online). You absolutely can experience much of this outside the academic environment, but you've got to work at it.
At the same time, there are many schools (and student bodies) which are uninspired, and programs which do a poor job of preparing students for the outside world. There were holes in my education. There was much that I learned in college which was self-directed, or incidental.
One of the bigger lessons was the degree to which you might want to discount some random professor/researcher's claim of something ;-)
I'd be surprised if those numbers weren't even more skewed--- high-profile entrepreneurs like Bill Gates are surely outnumbered by the vast hordes of MBAs, engineers, etc. Recall that 10% of the American working population is around 13 million people! And the cutoff is around $100k in income. A large portion are white-collar office workers at the upper-end of their companies' pay scales, or not even that upper end for technology firms, and people with no degrees aren't commonly found there (take a look at the top 500 employees at Boeing, Exxon, or Goldman Sachs, for example).
All such studies suffer from the same debilitating flaw—there are almost certainly confounding variables at work. To reach any firm conclusion, you would need to compare degree recipients with a cohort of statistically equivalent people who could have earned degree X but did something else instead. All these "education pays" charts show is that people who earn advanced degrees earn more money. But such people are also probably smarter and more determined than average. You don't need a Ph.D. to know that correlation doesn't imply causation.
P.S. This is not sour grapes. I have degrees in physics from Harvard (A.B.) and Caltech (M.S., Ph.D.), and I make well more than the median earnings for Ph.D.s.
P.P.S. I know the real reason these sorts of studies proliferate, but this comment is too small to contain it.
Stories like this pop up frequently on HN. But I'm intrigued by your post post script. What do you think is the reason for the proliferation of these studies, and why do they so often ignore the correlation-does-not-prove-causation fallacy?
In this case, I see two possibilities. The Bureau of Labor Statistics author is too dumb to understand the fallacy. The second is that she does understand the fallacy, but thinks that at least some Americans won't. If the latter, what purpose would that serve?
I think it's because not because these numbers are imperfect, or because many things can be read into them that shouldn't, that they are completely useless. In the end you need some data that is intuitive to understand, even if there are many ifs and buts involved.
You say this in way that seem to mean that for you education do not "pay". But I really don't know how one could affirm that education do not "pay".
Heavy focus on deep, large, long and high-quality education to their offspring is the most important reason, if not the only one, for elite people to stay in the elite. It is also the most rational explanation to why some people are often perceived as more successful than other, eg Chinese and Jewish.
It is easy to come up with counter examples of some guy who did not have the best traditional education cursus and did manage to be successful despite this. But exceptions are not the rule. The mainstream, average rule is that people in power have had good or excellent education.
So this little chart about "education pays" is just a confirmation, among others, to a well-known fact. It has a lot of shortcomings, and shortcuts (like using the income as a measure for success) but it is some value, at least as a reminder.
NB: I may seem hot on the topic. Yes, I am hot on this. Why? Maybe because I live in China, where I see everyday a lot of people who struggle extremely hard, save each yuan they can, in order to push their offspring to the best universities.
What is not clear to me is why people with degrees are always trying to promote that others also get degrees. The exclusivity of the degree is what makes it interesting and valuable.
Not knocking the education component, education is always a good thing (whether it comes from school or some other avenue), but the perception of success you talk about is because it is still somewhat of an exclusive club; probably doubly so in developing nations. However, thanks to the pressure on students to go to college, that gap is continually shrinking and constantly devaluing the value of the degree.
If you buy a Ferrari, you are perceived as more successful than others. You are going to be afforded luxuries that others do not have, just because you are signalling that you are "better". If we did some studies, I bet we would even find that Ferrari ownership, like a degree, pays – through better contacts and opportunities to make money because of the signalling. But I don't see Ferrari owners out there telling everyone that they need to own a Ferrari. They understand the importance of being part of a select few.
If education pays, you would think you would want to keep that under wraps. Given that, it always feels like there are some insecurities at play every time the topic comes up.
This topic reaches the frontpage on HN at least once a week. Can we avoid rehashing the same arguments? Maybe just vote these up when something original comes up.
So, does education make you better at your job? My experience says that there is very little correlation between having a MS vs a BS vs no degree in the quality of work people get done, at least as a professional coder. In fact, to me, a Masters is a red flag, it causes me to give a person additional scrutiny, because of the huge glut of awful coders with Masters that I have run into.
So, in my experience, having a more advanced degree could only lead to higher income because of big companies with big HR departments, the government, and other clueless employers giving a pay differential.
I would love to see the numbers comparing someone with a PHD from Shit State, compared to a person with a BS from Princeton, compared to a person with a genius level IQ who dropped out of Berkley. Or a PHD in History vs a BS in Math.
I'm always suspicious of PHD's in industry. You don't get a PHD to go work for GE, you go work for GE when you have a piece of paper with PHD on it but can't get an academic position.
So, what I'm saying is, while education level may be correlated with income, it's not a cause, except where employers are clueless.
I occasionally run into this "I'm suspicious of Ph.D.s in industry" phenomenon and frankly it's baffling. (disclosure: I am obviously biased by having a Ph.D. and no longer being an academic)
Ph.Ds. in industry create lots of amazing things: UNIX, 747s, airbags, wireless protocols, manufacturing robots, search engines, and logistics optimization systems are just a few that almost everyone depends on every single day (indirectly in some cases). These are not systems that can be designed without a large amount of focused domain knowledge, which is precisely what the Ph.D. entails.
Sure, there are some people with Ph.D.s who are impractical and unsuited for front-line "get things done" work. I can assure you that is an intrinsic personality trait and not something created by intensely focused study on interesting ideas, which is the defining characteristic of the doctorate.
There is a big, big, big problem when comparing degrees: People with PhD's are older and therefore have more experience. Everything in this entire chart could theoretically be attributed entirely to experience and not education. Of course, it could also all be education, but as they say, Correlation does not imply Causation.
This is a specious argument by the bls. The question isn't whether education pays or not; the real question is whether or not education increases a person's productivity.
I'd argue not. It all depends on the degree. College is a form of job training for people like engineers, in which case it has a high return on investment. However, many [if not most] degrees have no relevance to their college degrees.
As many as 63% of college graduates occupy jobs that don't require a college degree! (Uncited, lazy, sorry.)
So does college "pay"? Sure. Does it increase one's skills and productivity? Factoring the opportunity costs, it's hard to tell. I'm in the camp of people who argue that the returns to higher education are due to what's called a "sheepskin", or "signalling", effect in economics.
Interesting that doctoral degrees do not pay more on average than professional ones! Perhaps it is because most PhD's enter the field of research as opposed to industry? I haven't met that many PhD's in industry but there are a few.
It doesn't say "PhDs", it says "Doctoral degrees", which may mean (it isn't clear), that EdD (Doctor of Education, a common degree for public school administrators), DBA (Doctor of Business Administration, usually from Harvard), EDB (Executive Doctor of Business, typically offered to already successful business people) and other similar specialized but not necessarily "professional" degrees (which I interpret to mean MD, JD, DDS, chiropractic, etc.) would be in this group.
So they are "professional" because the are part of the standard career track that eventually leads to someone getting a formal "professional" qualification? That would suggest that an MBA is not a professional degree, although it is a vocational degree, because MBA graduates don't typically go on to become formally professionally qualified (e.g. unlike lawyers, medical doctors, etc.).
[NB I'm in the UK and while I know plenty lawyers/chartered engineers etc. I've never heard anyone talk about a "professional degree"].
No they are professional because they are required* as part of the license in that profession. To be an RN you have have to have a Bachelors of Science in Nursing. A CS major couldn't become an RN but an RN can become a software developer. It is the same for accountants, doctors, surgeons, etc.
Professional degrees are very specific and single focused. An MBA is just an MBA, no career requires it, but it is a nice to have. It is not considered a professional degree although many professionals do have it.
[*] and when I say required it is a felony to work in that field without the proper degree or license.
I didn't think UK CAs have to have a degree in accounting, I'm also pretty sure that you can become a CEng without an engineering degree, it's just very difficult.
I know that in law in Scotland you need two academic qualifications (a law degree, and a postgrad diploma in legal practise) then two years training before you qualify as a solicitor. However, if you want to go to the bar you only need a law degree - but then you have to do a specialised training course and then devil for a year before you can be called.
I still get a kick out of people who talk about how education pays but don't the whole correlation/causation thing. Apparently, education for them does not include basic stats.
What it says is that somebody who got a medical/law degree 20-30 years ago, did very well over the last 20-30 years.
It says nothing about whether somebody leaving school today will do better in 20-30 years having spent 10x as much as their parents generation on college - rather than becoming a plumber or electrician or founding a startup
And it's even worse than that - 20/30 years ago most of those none law/medicine degrees would have got somebody a VP type job. Now a non-professional degree gets you an entry level secretary job.
If you compare the lifetime earnings of a rare graduate 30years ago with those of a new 'social studies' type grad today it will be a lot worse
I agree with your primary point, but your follow-up example isn't the greatest... Flossing your teeth can have a real impact on health -- gum disease (which flossing is an excellent preventative agent against) can cause all sorts of serious health problems like increased risk of heart attack and stroke. People with periodontal disease have a 2x greater chance of coronary artery disease than people without periodontal disease. That's a pretty big deal.
The first point (gum disease can cause health problems) is most likely true, even if just intuitively, but they would have to be pretty serious to reduce life expectancy by 7 years - even smoking reduces it by 'only' 14 years, and I'd be hard pressed to believe that not flossing is half as dangerous as smoking.
The second argument is a classical cause/effect thing. I can't imagine that tooth diseases increase heart diseases by 100%, but I can very easily imagine that people with certain habits (eating lots of cheap unhealthy food with much fat and sugar) suffer from both.
Then again I don't know anything about this field, if there is any research that confirms your arguments I'd be surprised but interested.
All together, now: Correlation is not the same thing as causation.