Wednesday, January 22, 2014

STEM pays

Peak annual earnings for engineering graduates were even higher at $98k. Source: How Liberal Arts and Sciences Majors Fare in Employment: A Report on Earnings and Long-Term Career Paths.


When I was a new faculty member at Oregon I was told with great earnestness by our President (a JD with limited understanding of STEM or serious academic research) that liberal arts majors made more than science and engineering graduates in the long run. This encounter catalyzed my suspicion that those who run the world often don't understand how it actually works.

Better figures here.



26 comments:

Iamexpert said...

And this probably applies at the high end too if Bill Gates is any indication.

Hacienda said...

20k split at the of SIXTY! between humanities and science. When you balance out quality of life, eg who make
better friends and company. Who gets the good looking chicks. Who gets to live spontaneously and with some
adventure. Who gets to keep their basic human instincts intact. It's not even close. Humanities, all the way!

Richard Seiter said...

I wonder if the JD really believed that or if s/he was poorly stating a belief that JDs made more lifetime (i.e. lifetime earnings made up for the delayed start). Or was it more a matter of "More of the successful people I know have liberal arts majors than STEM"?

I think it is worth mentioning that the title on the first page of the linked brochure is "Liberal Arts Graduates and Employment: Setting the Record Straight." There is a definite agenda here and I would be suspicious of where this lies on the information/propaganda spectrum. A good sign of this is the lack of engineering in the graphics (except for proportion of people in all fields/social services). Have you seen the original report (source of information for this brochure which was titled as you stated)? Available at http://secure2.aacu.org/store/detail.aspx?id=E-LASCIEMPL

By any chance was the data broken out by individual majors? That would be very interesting. Just knowing which majors fell in which category would also be interesting (many are obvious, but perhaps not all). I suspect there would be wide variation in the humanities majors.

A frequent topic in this blog is differing ability levels between students in STEM and other majors. Has anyone done an analysis factoring that in? That might make an even more compelling case for the financial value of majoring in humanities. ;-)

pat said...

I have some difficulty with this data. First of all my experience is that any doofus should make $100K these days. Secondly it looks like only those employed are counted. A reasonably successful lawyer should make $200K a year but most lawyers hardly ever earn a fee at all. Without some kind of variance or inequality measure like a Gini index these data are not very illuminating. For example, almost all graduate engineers are employable somewhere but marketing majors certainly aren't.

David Coughlin said...

I see that turn at 51+ happening all over the place. Lots of guys who didn't make it very far in management, and whose technical skills can be replaced by younger engineers. They get laid off and take moderate to big pay cuts to be apparatchiks.

Robert Buttons said...

Everything you need to know in one chart.

larrydarrell1917 said...

any doofus "should" make $100K these days


but some doofuses DO and some DO NOT.

BobSykes said...

Only 1% of college-age people have the IQ and work ethic required to complete a STEM degree.


That's 1%. What do you intend to do for the rest of the population? Gas chambers? Gulags? Plantations?


Currently, we have substantial overproduction of STEM degrees. At the BS engineering level, a significant fraction of graduates never work in their discipline, although they may work in related areas like technical sales. At the PhD, level, 75% of all engineering candidates are foreigners, almost all of whom go home. The US economy has no place for them.


STEM degrees (like law and medicine) are very vulnerable to automation. In 1966, when I graduated with BSCE degree, every engineering company had armies of junior engineers with slide rules and mechanical adding machines doing design calculations, and rooms full of draftsmen and accountants. A survey team was at least 3 and usually 4 men. Now, all that's gone. And the internet allows companies to coordinate design teams across the entire world.


If the Senate immigration bill passes, almost all the PhD candidates will stay in the US, and STEM salaries will be crushed.


Promotion of STEM for everyone has got to be the greatest academic fraud of all time.

Hacienda said...

Newton is smarter than Shakespeare?


That'a a tough, tough call. Shakespeare's pretty damn underestimated by mathematicians, philosophers, and
the general population.


And Newton was prone to some stupid errors, in addition to being a general +sshole that nobody liked.


Modern day, I don't believe things have changed that much.

gs said...

I calculated the average yearly raise and got the following (Being one
of those rare people who are equally bad at arithmetic w/wo a computer,
I'm showing the inputs):

{humanities, professional, science & math} =

{(66.2/26.3)^(1/35), (64.2/31.2)^(1/35), (86.5/26.0)^(1/35)} =
{1.02673, 1.02083, 1.03494}

Science salaries might barely keep ahead of inflation. I doubt the other two do. The AACU report does not mention inflation.

Patrick said...

No .. that would be college for everyone.

larrydarrell1917 said...

"Only 1% of college-age people have the IQ and work ethic required to complete a STEM degree."

thank you for sharing that figure. but didn't you make it up?

"At the BS engineering level, a significant fraction of graduates never work in their discipline..."


and in math and nat sci it's much higher. a problem is non-stem people, politicians, think of math and science and engineering as all the same thing. an msc in geology has infinitely more market value than an msc in math or biology.
that uni ed hasn't been matched to what the market needs is a failure of the free market.

larrydarrell1917 said...

both were english. both had red hair. did one more important than both, maudslay, have red hair too. i have an irish setter, but otherwise red hair is gross.

i've
always thought little of shakespeare, but if there were iq tests back
then i'd expect both shakespeare and newton to score very high, but with
different 'profiles'.

but in terms of their importance or in terms of their effect newton is shaq and shak is a very short midget.

Hacienda said...

Are you an engineer?


Biotech and bioengineering will/are changing the world. Apple was Job's company. Silicon valley is a silly place where nothing really ever has been accomplished. All basic innoviations are government/military/university. We'd all be happier if Apple/Google/Microsoft never existed.

Rudel said...

I disagree with the basic premise other than that IQ plays a role in max income that totally overwhelms the degree of educational attainment as a factor. i.e. Degrees attained are only are rough surrogates for intelligence. I earned amounts several standard deviations above those average peak earner graphs with only a B.A. degree (albeit it one with calculus, formal logic, and statistical course work.) I am certain it was a direct result of my I.Q. also being several standard deviations above the norm and this was true for many of my contemporaries also. I know of more than one successful engineer (i.e. one with multiple patents) from different fields without completed degrees at all.

Despite the Tiger Mom bias of this blog I wouldn't trade a single point of I.Q. for another degree.

larrydarrell1917 said...

no, and you're right innovation starts in govt funded research more often than not. that's one of the reasons people like paul ryan are so laughable and despicable. angela merkel is a phd in p-chem. ryan isn't and such a person could never become president in the us or the uk. austria's legislators are the best educated, or so i've read. america's the least, in the developed world.

should shockley be given as much credit for intel as moore. maybe. people remember rockefeller, who wasn't a technical person. they forget maudslay. perhaps it should be the other way round.


what most historians and literary critics call history is the history of famous people. it is just a sophisticated version of people magazine and celebrity biographies. try something like the history of machine tools to find real history.

Hacienda said...

Depends on what is meant by biotech, of course. Just browse through youtube and you
can see what's in the pipeline. I particularly like thought control, dream-recording.
Real Harry Potterish, Marvel comics stuff.

Richard Seiter said...

Good points. I wonder if the focus should be on aspects of the high school curriculum? As an example, my fantasy world would add something like:
- Critical thinking (e.g. reasoning/logic, evaluation/use of evidence, and presenting/evaluating a supported argument)
- Basic statistics - focused on understanding and evaluating statistics used in public discourse
- Practical math - enough to support critical thinking and basic statistics
- Practical STE - enough to support critical thinking (e.g. what is/not the scientific method, technology trends and their implications, tradeoffs in engineering)

I think a MOOC targeting these to moderately capable (say at junior level) high schoolers and early undergrads would be perfect. Supplement that with in class work and discussion if possible.

Something like this would be good preparation for college and would also help those not attending college.

P.S. I would add to your arguments the difference between the value of a degree in a world where 25% have them compared to a world where 40% have them. Wild guesses at numbers now and 40 years ago based on enrollment stats from http://www.pewsocialtrends.org/2009/10/29/college-enrollment-hits-all-time-high-fueled-by-community-college-surge/

MUltan said...

MOOCs are the hot new thing, but I prefer books, websites, pdfs. There is a place for lectures, but something more along the lines of a television show such as NOVA would take better advantage of the online medium. The main advantages of physical lectures are that they force students to go through each one in a single sitting according to a schedule, there is a social element, and reality is always more engaging than a screen. They are cheap compared to tutorials, but expensive compared to books or videos. MOOCs don't provide the same benefits as live lectures, they're basically just a series of videos.

My daydream for how a replacement system might work: If online courses allowed more teacher and small-group interaction, they would be more like tutorials, but the cost advantage would go away.On the other hand, the system could facilitate getting together with other students of similar levels and paces without much expense, and an hour or two per week of one-on-one time with a teacher would be affordable yet would give more personalized attention than classes allow. Individuated courses of study are possible with a smorgasbord approach - the student has a buffet of options of suggested material and can choose what to study and in what order, with occasional suggestions from teachers and peers. This requires individualized assessment.

Assessment is where there is the most room for improvement over traditional schools. Validating questions and determining their difficulty, gathering large question banks, sharing them and securing them can now be done mostly automatically using Rasch measures and item-response theory, giving much better assessments than grades and more adaptable and quicker measures than standardized tests. Open collaboration in question production and vetting could replace high-stakes tests, in-class tests and most homework with more accurate, tailored to the student, and richer metrics.

Your curriculum looks OK, but I'm not sure statistics is central. Maybe from a Bayesian and applied point of view. Persuasive speaking and writing is a more important subject.

The relationship between math and physics is the crucial feedback loop for the technical part of the curriculum. Math needs a motivation to make it stick, and physics provides that motivation as well as many analogous situations and mental models that train the intuition. Then the math is applied to things that are less intuitive at first, but become so in time. Using hydraulic analogies to understand electrical circuits, which in turn give a model for EM fields and quantum mechanics, which serve as mental models for the math that describes them, is a good example of the sort of reversal of the usual order of presentation I'm talking about. Using this approach, I taught electronics classes for bright 9-12 year-olds which also successfully conveyed much physics and math along the way.

The current math curriculum is an historical relic. I'd make real-valued Clifford / Geometric Algebra the "one division ring to rule them all" for teaching. Nearly everything relates to it, it's simple enough to re-derive nearly anything, yet it's structure is such that it always has intuitive geometric content and can represent any physical situation. The same methods work in any dimension of any signature, the same notation in every area of physics. I'd also make programming/testing/debugging key parts of computer simulations of problems the students' usual mode of work rather than doing paper problem sets. The simulations should be a long-term project or series of projects for the students, and would be supplemented with practical mental math, such as Fermi estimation problems.

larrydarrell1917 said...

if only the majority of men weren't sheep this all would have happened a long time ago. i've had no use for a classroom since i was 15.

if people did not seek and did not accept professorships other than the purely research variety these solutions would appear almost immediately. just as if people held tax advising in disdain taxes would almost immediately be simplified.

but most men are sheep. they cannot tell the difference between the way things are and the way they should be, between the approbation of the crowd and doing what is right.

Richard Seiter said...

An interesting observation from another discussion of this work: "Excluding the graduate-degree holders, humanities and social-science majors earned less than professional and pre-professional majors." http://chronicle.com/article/How-Liberal-Arts-Majors-Fare/144133
(I'm assuming they did a fair comparison and removed all graduate degree holders, not just for one group?)
P.S. The correction at that link is a good indicator of how hard some folks are working to spin the results. Note that it applies to the first graphic above. They cherry picked 56-60 for the comparison. Look at the detailed bar graphs at my link.

larrydarrell1917 said...

ol's depends need changing badly.

he thinks of iq as if it were magic. it isn't. it's the only thing which might make psychology a science, but it fails even at that.

so he and others use "iq" as proxy for "native intellectual/cognitive ability, smartness, etc.". and all of these "concepts" vaporize under scrutiny.

for ol ideology is reality. it's just talk, but he can't see it as such. because the effect of the jive talk is real he takes the jive talk as representing reality.

larrydarrell1917 said...

ol keeps repeating his rot and jive talk. i guess old dogs can't learn new tricks.

once the stupid are excluded 100% of the iq income correlation is mediated by formal education credentials. http://www.halfsigma.com/2006/06/high_iq_does_no.html. the wordsum used as iq proxy is better than this blogger knows. vocab is 1. the most g-loaded 2. the most reliable 3. the most heritable of the wechsler subtests for both children and adults. http://blogs.scientificamerican.com/beautiful-minds/2013/10/17/the-heritability-of-intelligence-not-what-you-think/.

furthermore the neo-liberal syllogism: iq is correlated with income, iq is heritable, therefore class is the result of genes is INVALID. only 1/20th of intergenerational income elasticity can be explained by the heritability of iq. children are not clones of their parents. their may be pure breeder, but they're rare. http://www.hks.harvard.edu/inequality/Seminar/Papers/BowlesJEP.pdf.

once the stupid are excluded, say iq < 100. the iq income correlation is still present but it's pathetic < .3.

the self made rich are smarter than the rest but how much smarter depends a lot on how they made it. ol's own sector likely requires more ability than most, but i've known many illiterate techies.

larrydarrell1917 said...

ol keeps repeating his rot and jive talk. i guess old dogs can't learn new tricks.

once the stupid are excluded 100% of the iq income correlation is mediated by formal education credentials. http://www.halfsigma.com/2006/06/high_iq_does_no.html. the wordsum used as iq proxy in the gss is better than this blogger knows. vocab is 1. the most g-loaded 2. the most reliable 3. the most heritable of the wechsler subtests for both children and adults. http://blogs.scientificamerican.com/beautiful-minds/2013/10/17/the-heritability-of-intelligence-not-what-you-think/.

furthermore the neo-liberal syllogism: iq is correlated with income, iq is heritable, therefore class is the result of genes is INVALID. only 1/20th of intergenerational income elasticity can be explained by the heritability of iq. children are not clones of their parents. their may be pure breeder, but they're rare. http://www.hks.harvard.edu/inequality/Seminar/Papers/BowlesJEP.pdf.

once the stupid are excluded, say iq < 100. the iq income correlation is still present but it's pathetic < .3.

the self made rich are smarter than the rest but how much smarter depends a lot on how they made it. ol's own sector likely requires more ability than most, but i've known many illiterate techies.

larrydarrell1917 said...

http://www.hks.harvard.edu/inequality/Seminar/Papers/BowlesJEP.pdf
http://www.halfsigma.com/2006/06/high_iq_does_no.html
http://blogs.scientificamerican.com/beautiful-minds/2013/10/17/the-heritability-of-intelligence-not-what-you-think/

the above links show that 100% of the iq income correlation is mediated by formal education. AND that only 1/20th of intergenerational income elasticity is due to the heritability of iq. from retarded to brilliant the correlation is < .4. when only those with college degrees are looked at the correlation vanishes entirely. the gss's wordsum is the best short iq test possible as vocab is 1. the most heritable, 2. the most g-loaded, 3. the most reliable of all wechsler subtests for children and adults.

Carpet cleaning Sydney said...

I really like this statistics for understand a theme of anything clearly. A statistic can say everything about a project. So anybody can understand easily.
End of Lease Cleaning

Blog Archive

Labels