Hacker News new | past | comments | ask | show | jobs | submit login
In the Salary Race, Engineers Sprint but English Majors Endure (nytimes.com)
42 points by tlb 22 days ago | hide | past | web | favorite | 53 comments

> Computer science and engineering majors between the ages of 23 and 25 who were working full time earned an average of $61,744 in 2017, according to the Census Bureau’s American Community Survey. This was 37 percent higher than the average starting salary of $45,032 earned by people who majored in history or the social sciences (which include economics, political science and sociology). Large differences in starting salary by major held for both men and women.

> Men majoring in computer science or engineering roughly doubled their starting salaries by age 40, to an average of $124,458. Yet earnings growth is even faster in other majors, and some catch up completely. By age 40, the average salary of all male college graduates was $111,870, and social science and history majors earned $131,154 — an average that is lifted, in part, by high-paying jobs in management, business and law.

Uh, yeah. From 23-25 you’re comparing engineers vs random 4-year degrees. At 40 you’re comparing engineers vs a group that suddenly includes lawyers and MBAs.

You’d find similar results if you compared baristas to premed students. At 23 the typical barista is earning at least minimum wage while the typical premed student is earning literally nothing (unless moonlighting as a barista...). Suddenly their fortunes reverse at 40 when a significant portion of premed students have turned into established doctors.

23-25 is the time that all the advanced degrees are being earned. If you want a fair comparison you have to compare equivalent education in each age group. You can’t suddenly change the cohort at 40.

> At 40 you’re comparing engineers vs a group that suddenly includes lawyers and MBAs.

Sure, the fact that the career paths diverge is... the whole point. The cohort didn’t change, one part of it went to law school and the other didn’t. Your analogy is totally off base.

When you explicitly select people who are between 23 and 25 and have entered the work force, you are intentionally selecting for people who do not have a JD and are not actively earning one. The set of 23 year old lawyers is essentially nonexistent. The set of 23 year olds working full time and also enrolled in law school is hardly much larger.

Ah, I see what you mean. I’ll grant that the “currently employed” requirement artificially inflates the starting salary figures, probably more so for history majors than engineers. The broader point remains, though, that history/social science degrees can lead to paths that, if pursued, end at higher salaries.

Also, I expect salaries are also higher for engineers/developers with a masters or PhD than just a BS, so there’s some of the same effect in that population too. (Although I suppose it’s more likely that engineering students seeking advanced degrees, especially PhDs, are also working full time).

Where is the evidence that they selected for employed grads?

It’s in the part of the article I quoted.

“Computer science and engineering majors ... who were working full time earned an average of $61,744 ... higher than the average starting salary of $45,032 earned by people who majored in history or the social sciences...”

Plenty of engineers get JDs and MBAs too. It sounds they are selecting for engineers that didn't get postgraduate degrees.

Article is paywalled for me, but that’s not what the quote says.

I think it is a fair comparison, because this is measuring career progression. Social science professions progress toward management positions or law. What would be a similar path for engineers? If that doesn't exist, then it is a problem that needs to be dealt with.

It isn’t measuring career progression. That’s the point. It is conflating entirely unrelated career progressions. The set of people who have a sociology degree at 23 and are working as salespeople or office managers or whatever else are almost entirely disjoint from the set of people who have a JD or MBA at 40. Largely that second category is made of people who were not working at 23 because they were actively pursuing a postgraduate degree.

There is actually no career progression that results in a post graduate degree being earned. That is simply not part of career progression. Earning an advanced degree is something that one can choose to do instead of investing in one’s present career. No amount of hard work at your job will turn a four year English degree into a law degree or MBA.

>are almost entirely disjoint from the set of people who have a JD or MBA at 40

My S.O. has a sociology degree, and worked at JC Penny's after college. She now has an MBA and a JD, and is COO for a healthcare delivery network of several hospitals and a whole load of clinics.

I doubt she is the only one like that.

I’m fairly confident that your SO’s case and educational path is extremely atypical. Even if it weren’t, it wouldn’t say that a sociology degree is a good way to earn a high income or become a COO. It would say that a sociology degree followed by a JD and MBA is a good path for those things.

That's nonsense. 90% of MBAs had jobs at age 24, before MBA school.


Even if this is true, it is very misleading to conflate this with 4-year liberal arts degrees being lucrative long term. “English majors out-earn engineers at 40” is an extremely different statement than “English majors who also earn a JD out-earn engineers at 40.” If you have to get a second degree to make a good living as an English major, that says that the first degree itself isn’t very lucrative.

This is not misleading when you realize that people earning a JD today are the same people who some time ago graduated with a liberal arts degree. They're not engineers or mathematicians.

The point isn’t that they aren’t engineers. The point is that the 4-year liberal arts degree isn’t what earns them a high income. It is extremely misleading when you claim that “X leads to Y” when the reality is that “X+Q leads to Y” and the X is virtually irrelevant.

The article claims that what leads to higher earnings are the “soft skills” that liberal arts majors learn that somehow part of later in their careers, but the data they rely on doesn’t point to that at all. What helps liberal arts majors’ average earnings in the future is almost entirely attributed to the sudden inclusion of people holding advanced professional degrees.

Your contention that a liberal arts education doesn't contribute to someone who later in life becomes a lawyer is unfounded. It is similar to say that high school doesn't contribute to higher earnings: of course it does, how would someone otherwise go to college and make those big bucks? If X is a requirement to do Y, then X certainly contributes and must be counted among the factors that lead to high salaries for people who have Y.

>it is a problem that needs to be dealt with.

Well, I think it's definitely a problem. Maybe while engineers and techies are young and working and making all that money, you use some of it and get an MBA or law degree or some other professional degree? Even better, go to med school. (That one's really hard though, only do this if you really want to be a doctor.)

In any case, education like that would deal with the mid-life income growth issue fairly effectively. (And in the tech industry, with all the ageism, it can be a mid-life "having any income at all" issue. Which is even worse.)

See my top level comment, but you don't have to explain the 'facts' in this article since it appears to be based on useless data, and it doesn't even present that useless data correctly, so there aren't any facts.

But the large majority of eng lit and history etc grads don't get high flying jobs.

But it does show that compared to peer professions Engineering is badly paid even more so in places like the UK and EU

$141000/year is a shockingly large income for any broad group of people in the US. According to the ACS, the median earnings for all bachelor's degree holding full time make workers is $61k. That's for the full population, so it should in fact be centered somewhere near age 35-40.

I don't know where the author is getting the $112k number for all men at age 40 is from, but it seems impossible from my knowledge of the ACS statistics. Perhaps the mean drifts that high with super star executive pay? Regardless, it makes me suspicious of the other numbers in this article- anyone have an idea of how the author could get these earnings statistics from ACS data?

I was thinking the same thing, this doesn't seem possible.

I'm studying machine learning as a PhD student at Caltech. I have heard of graduating ML PhD students from my department making $400-500k/year as a starting salary. Heck, they could probably retire by 35. I doubt any English major could match that... I wish the NYT article also looked at starting salaries of fresh PhDs, not just new college grads.

> Heck, they could probably retire by 35.

As a PhD student (PhD students are underpaid to an almost criminal extant), you might think that, but start with half of that going to taxes, a $2M home, and escalating expenses. You can make it happen, but you get a bit of a culture shock when you go from making that much money to living off $50k in interest and living in Toledo.

'FIRE' pursuants do not generally buy $2M homes, or no, they wouldn't manage it!

The thing is this, that's nice for people who go to CalTech and get PhD's. What about the guy with a Master's from University of Illinois, Georgia or Michigan? What about the guy who only has a BS?

None of those guys are getting 500 grand.

The point is that people like that should probably consider a long term plan to keep their incomes growing at the mid-life stage.

Well sure and there are soccer players without any form of degree who make 10x that, what's your point?

My guess is this has more to do with technician roles having an acceptable baseline and the personality type not wanting to go into management. If you sampled just managers only, I doubt you'd see the same difference because technical expertise is extremely valuable in management.

Regardless of the nuance of the stats though, the idea that you'd choose between wildly differing majors based on watered down averages strikes me as a really poor strategy. The variance on salary progression is huge, and it's not what you learned in school but what you do on the job that fuels your advancement. IMHO you're much better off starting by optimizing what you are naturally interested in at a young age, develop your skills along the way, but stay open and ready for new opportunities as you progress through your career.

I've never been able to break into management and now I'm 40. Compared to most programmers I'm personable and outgoing. Getting a leg up into management is very difficult. In 13 years as a programmer I've never worked with anyone who achieved it. Don't bank on it being possible. If you want to get into management you need to take radical action. Don't drift along like I did and be bitterly dissapinted that you're stuck in development probably forever, like I did.

> One of the largest and most popular courses in the Stanford computer science department is CS229 — Machine Learning, taught by the artificial intelligence expert and entrepreneur Andrew Ng. This course did not exist until 2003, when Professor Ng taught it for the first time with 68 students, and very little like it existed anywhere on college campuses 15 years ago. Today, the machine learning courses at Stanford enroll more than a thousand students.

Machine Learning was a standard offering in CS departments in the early 2000's. It was already a venerable subfield with much of the basic theory having been developed in the 1970's and 1980's. I remember thinking it was something of a curiosity and not terribly practical but it's absurd to imply it sprung from nothing.

I was also surprised by that bit. Tom Mitchell's influential textbook [1] came out in 1997, and was already being used by quite a few courses within a few years of release.

[1] http://www.cs.cmu.edu/~tom/mlbook.html

That was my textbook at York Uni back in 2000, it was a great book. I recently got a new copy and quickly realised that I have forgotten most of the maths that I used to know, so first I need to do a refresher on that :)

That book taught me an important lesson that I wish I'd learned earlier, that for many of my courses reading the textbook would have been a good idea. My lecturer for the ML course was unbelievably tedious, and I realised I'd absorbed almost nothing. I sat down with Mitchell's book for a week before the exam, did all the exercises, and got the highest mark in my degree.

I often wonder how much better I could have done, and how much more I'd have learnt, if I'd put that sort of work into all my other modules earlier in the course.

How does your assertion contradict the quoted excerpt? It merely states that Ng's course began in 2003 with a relatively small number of students, and that few other colleges had taught such a course 15 years ago. Not that Ng invented the field of machine learning.

Specifically this part:

> This course did not exist until 2003, when Professor Ng taught it for the first time with 68 students, and very little like it existed anywhere on college campuses 15 years ago.

I took a Machine Learning course as part of my degree in 2001/2. About half of my year in the department attended. The course had been running for many years and I understood it to be a standard fixture in CS department curricula across the UK at the time.

Thanks, that's all I was asking for (I won't demand that you do an empirical study across all colleges this early in the morning ;) )

Bold claims with almost no citations? Thanks for the opinions, but I think I'll pass.

We should have levels.fyi for more things in life. It really blew my mind how much the tech companies pay. Made me realize I needed to prioritize those companies for my job search.

> Made me realize I needed to prioritize those companies for my job search.

Ehh, or just do less work at a place that pays less.

It's a common misconception that working at these companies requires selling your soul. Some company cultures are better than others. Some teams are better than others.

Lots of people at these companies work a strict 40.

I've been a SWE at Google for a while, been promoted a couple times, been a tech lead a couple times. Pretty much always worked 40 hours a week, and I don't think this is unusual. There might be some folks working 50 hours, but I suspect in most cases that's because they want to work 50 hours.

Google has a reputation for being fairly laid back.

32 hours is currently my goal—that’s the cutoff for healthcare.

So for nearly two decades take a pay cut of tens of thousands of dollars just to catch up at age 40?

Even if there were any citations to back up the claims in this article, it wouldn't make any sense!

Not just that, it didn't call out opportunity cost.

Can't read the article, even through the web link above.

It makes me think based on the title that lots of new college hires must be making insane amounts of money right out of school. What happens say, 10-15 years down the road when being a good developer won't cut it and you either go into management or IC it somehow?

I don't think everyone can do that because as you move up there are fewer positions. Also, there are new college grads right behind you willing to do what you're doing for way cheaper. Even if you've kept up with your skills it seems like it's easier than ever to pick up new skills due to democratization of learning.

it really depends on the industry.

It seems like median would be a much better metric than mean here.

the effect of compounding is notably absent in the discussion.

Author doesn't know what that is. :-)

(sigh) They're comparing averages, so the whole theory is suspect.

Averages are a terrible way to do a comparison. Averages aren't terribly useful unless you know that your data is Gaussian (or a sum of IID Random Variable => see the CLT). Folks love to erroneously assume Gaussian.

In this particular problem, I would expect to see a multimodal distribution for the liberal arts majors.. (i.e. some folks have massive salaries (famous authors) and some have small salaries (high school English teachers)).

In such a case, the average can end up at a value that basically doesn't exist in the underlying distribution. So you end up comparing nonsense.

> Between the ages of 25 and 40, the share of STEM majors working in STEM jobs falls from 65 percent to 48 percent. Many of them shift into managerial positions, which pay well but do not always require specialized skills.

While technically true, I would argue that many of the most successful managers are at least in part successful because they have a really strong background in the field they are managing. Most people want to be managed by people who understand some of the nuances of what they're doing.

I speculate the cause is self selection. People who know they don't have what it takes to become manager or any job that depends on soft skills are unlikely to major in liberal arts and more likely to choose engineering. Hence, fewer engineers move up the ladder. I can speak from experience. My parents prevented me from choosing a masters in science management because of the management.

The title of the article says "English majors", but the body says "Liberal Arts majors".

"Liberal Arts" isn't a category of undergraduate major. It is it's own category, which is generally only offered at 'Liberal Arts' colleges which offer only that major. However, if you take 'Liberal Arts' to be a 'category of categories', it could potentially include:


  Humanities - English Literature, Modern Languages, History, and Philosophy
  Social Sciences - Anthropology, Economics, Geography, 
  Political Science, and Sociology
  Creative Arts - Fine Art, Theatre, Speech, and Creative Writing
  Sciences – Astronomy, Biology, Chemistry, and Physics

Ok, nearly any major that existed before 1940 apart from Mathematics and Engineering.

What did the survey say exactly? Well, first of all, from everything I can tell it only includes 4 states, Colorado, Michigan, Texas, and Wisconsin. Not exactly representative, probably, but what schools are included?

School list (for Colorado):

Well, I've never heard of any of those schools, but whatever, maybe most people go to schools I've never heard of so that's not exactly evidence of anything.

I didn't want to go through every school, so I just picked University of Colorado, Denver, since it seems like it's probably one of the bigger schools?

I used https://factfinder.census.gov/faces/nav/jsf/pages/searchresu... to compare all the majors. I couldn't find a single non-STEM major that had better incomes than even a single STEM major. I assume this is the dataset he is taking his data from since he links to it.

So I guess what I'm saying is, this article seems to be predicated on a tiny dataset from a few select schools from a few select states, does not include flagship schools, (in fact it seems to only include 3rd and 4th tier schools, no Anne Arbor, Reed, University of Portland, University of Oregon, Rice, UofT Austin, Texas A&M, U of Wisconsin, Lawrence, Milwaukee School of Engineering, Marquette, etc etc etc.

So before you debate 'what this means' you should know that, based on the above, it means nothing, and no general facts can be drawn from it. I couldn't even figure out what methodology he used to come to his conclusions (if any), or even what he means by 'liberal arts' since that's not a category in the data.

All that, and the data I could find completely disagrees with his analysis, and in fact shows the complete opposite.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact