Colleges With Highest New Grad Median Earnings By Major

The CollegeScorecard database (https://collegescorecard.ed.gov/ ) added median earnings of new grads by major and school today, for students who received federal FA. The earnings are for the first year in which the student is not enrolled in school My interpretation is “earnings” may include sign-on bonuses and other non-base salary cash compensation. A summary of the 5 colleges with the highest listed median earnings for different majors is below. College-major combinations with an insignificantly small sample are not available.

I believe that earnings by major has far more to do with the individual student and location than the college attended. Nevertheless, I find this type of information interesting to review. The highest earning college-major combinations all involve computer science, often at highly selective colleges. Earnings for other majors do not strictly follow STEM vs non-STEM, although there is a clear correlation. Instead some STEM fields have higher earning such as most fields of engineering, while others have lower earnings, such as biology . Similarly a few non-STEM fields had higher earnings, such as economics, while most had lower earnings. Major generally appeared more correlated with earnings than college attended.

Computer Science and Highest Overall

  1. Caltech – $153k
  2. Brown – $141k
  3. CMU – $139k
  4. Penn – $135k
  5. Harvard – $129k

Engineering Majors (including Computer)

  1. Berkeley (EECS) – $117k
  2. MIT (EECS) – $117k
  3. UW: Seattle (Computer) – $112k
  4. UW: Bothell (Computer) – $112k
  5. CMU (EECS) – $99k

Engineering Majors (excluding Computer)

  1. Maine Maritime (Marine) – $93k
  2. Excelsior (Nuclear) – $89k
  3. UAlabama (General) – $89k
  4. Mass Maritime (Marine) – $87k
  5. SUNY Maritime (Mechanical) – $83k

Nursing

  1. Somoma State – $110k
  2. CSU:EB – $109k
  3. CSU:S – $107k
  4. Sam Meritt – $102k
  5. Mercy-- $99k

Mathematics

  1. MIT – $120k
  2. Brown – $92k
  3. Harvard – $80k
  4. Johns Hopkins – $79k
  5. Cornell – $76k

Economics

  1. Duke- $90k
  2. Dartmouth-- $84k
  3. Chicago – $83k
  4. Villanova – $83k
  5. CMC – $82k

Physics

  1. UIUC – $74k
  2. MIT-- $74k
  3. RPI-- $65k
  4. UArizona- $63k
  5. UWisconsin-- $61k

International Business/Relations

  1. UoSC:C – $67k
  2. Northeastern – $66k
  3. Georgetown – $63k
  4. Villanova-- $59k
  5. GWU – $58k

Biology

  1. UCSD – $68k
  2. UMinn – $65k
  3. Thomas Edison – $55k
  4. UMissouri – $48k
  5. Harvard – $48k

Education

  1. NYU – $52k
  2. College of NJ – $51k
  3. SIUC – $49k
  4. College of NJ – $49k
  5. Kean – $49k

History

  1. Duke – $51k
  2. Penn – $51k
  3. Harvard – $48k
  4. Rice – $48k
  5. Dartmouth – $46k

Psychology

  1. Harvard – $47k
  2. Columbia – $45k
  3. Tufts – $45k
  4. Barnard – $44k
  5. Cornell – $43k

English

  1. Monmouth – $42k
  2. Columbia – $41k
  3. Santa Clara – $41k
  4. U Houston – $41k
  5. Dartmouth – $41k

Theater/Drama

  1. UColorado – $32k
  2. Notre Dame – $31k
  3. UNLV – $30k
  4. UNC:SoA – $30k
  5. CUNY:Tech – $29k

Lowest Overall

  1. U Puerto Rico: Ponce: Health – $4k
  2. U Sagrado Corazon: Journalism – $5k
  3. Carribean U: Ponce: Health – $5k
  4. Carribean U: Bayamon: Health – $5k
  5. U Ana G. Mendez: Comm. Disorders – $6k

What is missing is the number graduating. According to Brown, in 2018 they awarded 248 BSs in CS. Since it started from 48 in 2010, we can figure that there are no more than 800 CS graduates of Brown who are more than a year after graduation. ABout 17% of Brown students are on Pell grants. However, in 2012 it was 14%, and in 2008, it was 12%. So perhaps 15% of all Brown CS graduates received federal support. So that median represents maybe 120 students, and, considering the rate of response to these questionnaires, we’re talking MAYBE 50 graduates, more likely closer to 30. So that median is meaningless.

The same is true for UPenn, which has about 400 graduating students from their entire engineering program, so likely fewer than 100 are CS, and, of them, some 15 have receive federal grants.

Harvard is perhaps twice those numbers, but even 60 isn’t a lot.

Going back before 2010, and Harvard had maybe 100 graduates a year, and the other two had fewer than 50. Students who received federal aid at any of these colleges in the early 2000s were even fewer than today.

So those statistics are extremely unreliable.

Similarly mathematics. None of the Ivies have enough students on federal aid who graduated with math degrees to produce data which would allow for any meaningful statistics.

That, in general, is the problem for looking at the data at the level of major, since, for so many majors, the number of students on federal aid who graduated from private “elite” colleges (which generally award only about 1,500-2,000 undergraduate degrees a year) is too small for the statistics to be meaningful.

What would interest me, though, is to see for which majors there is a significant difference in income for students who received federal aid, compared to those who didn’t, and compared to students from the top 1%. Basically - for which majors do family connections make the most difference?

You don’t need to guess. The sample sizes for each group are listed in the database. For brown CS, the listed earnings are the median of 59 students.

59 students is obviously not a precise value for the full class, but it’s not meaningless either. Instead the sample size limits how you would interpret the results. For example, the results are highly suggestive that Brown CS majors tend to have far higher starting salaries than nearly all other majors at Brown, which is consistent with other sources for recent Brown grad earnings. The precise median for the full class may be a significant distance from the listed CS median among this sample size of 59 students, but the gap between other majors is large enough to be confident that there is actually a large gap between CS and the vast majority other majors. The rest of the top 5 for Brown were (Math: $92k, Engineering: $74k, Business: $64k, and Economics: $58k)

The results are also suggestive that CS majors at highly selective top CS type colleges tend to have substantially higher earnings than the overall average for CS majors (the rest of the top 10 were Stanford, Mudd, Yale, MIT, and Cornell), more so than most other fields. This is again consistent with other sources. However, the sample size is not large enough to assume that Brown CS majors are expected to earn more than HYPSM… CS majors… . Instead I think it is likely that the small sample size decreases accuracy, and the colleges whose median earnings errors more on the high side are most likely to be the ones flagged among the top 5. Brown likely errored on the high side.

However, my point of the post was not to draw sweeping conclusions or say with confidence that these 5 colleges are the ones that will give you the highest earnings, significantly more than other similar colleges. It was that I find the information interesting. If you do not, don’t use the database.

As stated, the combinations with very small samples are not available. This was true of math majors at Ivies. Several had too small sample sizes to be published in the database. Among the 3 listed Ivies in the original post, Cornell had the largest sample size, which was 43 students.

There is enough information to estimate federal aid vs no federal aid, but again, it won’t be precise values… less precise than the numbers from the original post. So if you think the original post is “meaningless”, then there is little point to making such an estimate.

The estimates are meaningful for colleges for which there is a decent sample size, such as the UCs.

As for comparison in Brown between majors, the differences between CS and economics may be enough, but the differences between CS at Brown and CS at CMU is not, so ranking these as the top 5 is difficult to justify.

They should wait to use the statistics from colleges with small sample sizes until they have those sample sizes. In some majors, such as CS, colleges like Brown are rapidly expanding these majors (they number of graduates almost doubled in 2019, compared to 2018), and in a few years there should be sample sizes large enough to make these comparisons meaningful.

As for the comparison of with and without federal aid, I just wrote that it would be interesting, but I agree that we could not make a meaningful camparison at this point.

Is the data self reported or does it come from tax returns related to the federal aid?

Some colleges have their own career surveys, so if the numbers from those career surveys are very different from those for federal FA recipients, that may be an indicator of what you are asking here. But note that colleges’ own career surveys may have much less than full response rate, or may report differently (median versus mean), or may get pay information differently, so that is not a perfect comparison.

Very interesting @Data10 . Even at Harvard there is a big difference in earnings of STEM graduates and other fields of study. In many cases, the major chosen by students matters more for earnings than the school attended.

Within STEM majors, there can be big differences. Compare computer science to biology, for example.

Top 1% does not translate to family connections. We are top 1% and unless one of our children is looking at a very specific employment sector (which I doubt will be the case), we can’t help them. Top 1% definitely helps in other ways though.

Agree with techno. If your kid is majoring in ed it’s hard to see how a family connection is going to help. Teacher hiring is pretty transparent, easy to understand, and there is a lot of information out there in the universe as to which school districts are hiring and which are not. And if your kid is not qualified to teach HS chemistry, no amount of “family connections” are going to get him/her certified.

I think there was a study several years ago that showed the nepotism/family connections was actually a much bigger problem clustered in low paying fields-- Dad works for the public works department in a small town, kid gets hired for the department of recreation. Mom works in a school cafeteria, kid gets hired as a custodian. I’m not sure the children of the 1% are clamoring for these jobs. And the high end professions- medicine, corporate law-- either have high barriers to entry (sure kid, go get board certified in neurology and then you’ll have a leg up getting a job in mom’s hospital) or actual (and working) rules against nepotism.

At my current company, they would not hire the children of employees for internships, summer jobs, etc. Not a good look.

Top 1% helps in other ways for sure- but I think the role of “family connections” in fields like banking is highly overstated on CC. The days where the idiot nephew got a job in private wealth management at Credit Suisse to make some MD’s grandmother happy are over. Too much exposure and too much risk when the idiot nephew screws up. Did it happen in the 1950’s? Probably.

A sample size of 59 from a population of 800 provides a 90% confidence level of a margin of error of less than 9%.

This range spans barely two positions in the table. I wouldn’t call this data meaningless.

Within Computer Science there is a great dispersion in earnings among schools. At MIT Computer Science majors median earnings are $117k. At Stony Brook, the median earnings for CS majors are $71k.

Yes, the new grad from a wealthy family can have family support for an extended job search for the optimal job, for unpaid internships, and/or for up front relocation costs. The one from the poor family needs a job immediately to start paying off student loans, does not have the option of unpaid internships, and may find it difficult to afford relocation if not reimbursed or reimbursed later instead of paid up front.

1 Like

The reported earnings are based on tax returns, as described below. The field of study is reported by the college. A more detailed summary of the methodology is at https://collegescorecard.ed.gov/assets/FieldOfStudyDataDocumentation.pdf .

“The earnings measurement represents the sum of wages and deferred compensation from all nonduplicate W-2 forms and positive self-employment earnings from IRS Form 1040 Schedule SE (SelfEmployment Tax) for each student measured. Earnings values are presented in 2018 inflation-adjusted
dollars”

I wonder whether Students who received FA might have less tendency to pursue post-graduate options, like business, law, medicine, and graduate schools. These data look at immediate returns, while long term (life-time) earnings of people with professional trainings (those biology pre-med or history pre-law) could be much much higher.

I noted a similar pattern. For example, a comparison of earnings between top 5 USNWR ranked colleges and top 5 largest colleges is below. At the top ranked colleges, CS is almost always the highest earning major, usually by a wide margin. However, at the larger publics that are not especially high ranked in CS, the highest earning majors are usually engineering fields. CS often lags behind engineering and is well below the top ranked USNWR colleges. National salary surveys such as Payscale, follow a similar pattern to the larger colleges, with CS lagging behind some popular engineering fields.

I expect there are many factors that contribute to this effect. One may be a limited supply of especially high earnings CS positions that tend to be concentrated in a few specific high cost of living areas, such as Silicon Valley. One related factor is student selection. I’d expect top USNWR ranked grads to be more likely to target such positions even if on the other side of the country; and I’d expect large public grads to be more likely to target local positions in their home state, which generally does not include the SV or similar very high salary area. Another factor is employer selection. The hiring process for such positions is competitive, and I’d expect grads from top USNWR positions are more likely to do well in interviews, more likely to have the skills and knowledge that the employers are looking for, and often have better networking/connections.

CS Earnings for Top 5 National in USNWR

  1. Princeton – NA
  2. Harvard – $129k
  3. Columbia – $96k
  4. MIT – $118k CS / $117k EECS
  5. Yale – $120k

CS Earnings for Top 5 CS in USNWR

  1. CMU – $139k
  2. MIT – $118k CS / $117k EECS
  3. Stanford – $126k
  4. Berkeley – $115k CS / $117k EECS
  5. UIUC – $92k

CS Earnings for Top 5 Largest Colleges

  1. UCF – $58k (CS is 5th highest earning major)
  2. Texas A&M – $71k (CS is 6th highest earning major)
  3. Ohio State – $64k (CS is 8th highest earning major)
  4. FIU – $55k (CS is 7th highest earning major)
  5. Florida – $74k

The sample is for graduates from 2015 and 2016, with earnings measured in 2016 in and 2017. According to the database, Brown had 124 CS grads in 2015 and 144 CS grads in 2016, including both fed FA and non-FA, for a total of 268 CS grads during this period. The fed FA sample of 59 CS grads represents 22% of the total 268 CS grads.

Given the tax return nature, I’d expect nearly all fed FA to be represented. The bigger issue may be how well the ~22% fed FA represents the ~78% who are not. Some sources suggest similar earnings for both groups. In any case, I agree it is not a meaninglessly small sample. By some metrics, it is more meaningful to compare students with similar fed FA backgrounds at different colleges like this, rather than compare overall Brown grad earnings with median parent income of >$200k to earnings from students at various other colleges where few come from wealthy families.

Here is an example, using UCB because it has a relatively detailed career survey at https://career.berkeley.edu/Survey/2018Majors . However, note that the NCES numbers are median, while UCB numbers are average (mean), which is often a bit higher due to the occasional high end outlier which can be higher by more than low end outliers can be lower. The UCB career survey includes students with more than one major in all of their majors.



Major                   NCES median     UCB average
(engineering and CS)
Chemical Engineering     72700           66902
Civil Engineering        66000           65801
Computer Science (L&S)  114800          107302
EECS                    116600          111168
Mechanical Engineering   71900           77042
(other pre-professional)
Architecture (not BArch) 47900           56760
Business                 76600           75604
Social Welfare           36300           38736
(arts)
Art Practice             28800           56463
Music                    N/A            112900*
(humanities)
English                  32100           50389
History                  26700           51140
Philosophy               25100           46511
Rhetoric                 34900           48100
(social studies)
Anthropology             25500           54234
Economics                63000           70179
Ethnic Studies           41600           46962
Media Studies            43800           49035
Political Science        40100           51157
Psychology               35700           56931
Sociology                38700           46966
(sciences)
Applied Mathematics      61500           92694*
Mathematics (pure)       47300           72500*
Biology (probably IB)    33100           42818
MCB                      34900           49229
Chemistry                39800           61827
Physics                  45800           75985*
Statistics               63700           80871*

*Job titles in career survey suggest significant number with second major or minor in CS.


There appears to be a pattern in that, for engineering, CS, and other pre-professional majors (except architecture), the NCES numbers for federal FA recipients are similar to those from the UCB career survey. But liberal arts majors typically show significantly lower NCES numbers for federal FA recipients than the UCB career survey numbers.

So if they enter that specific employment sector, you could help them, no? That is EXACTLY my point. That is why I spoke of different majors - are there specific majors which feed into employment sectors in which the wealthier people can help their kids in ways that poorer families cannot.

Would a family in the bottom 20th percentile have the ability to help them in ANY employment sector except in rare cases?

Huh? Exactly how are you calculating the margin of error without knowing the variance of the sample? The margin of error for a mean value is x̄ ± t* s / (√n). Since you have no idea what the variation (s) is, you cannot know what the confidence limits actually are.

So I have no idea where that 9% comes from.

Of course this equation assumes that both the sample and the population have a close to normal distribution, and that those 59 were randomly selected. Since salary data is ALWAYS highly skewed (which is why medians are used), even this calculation is not useful

Also - the data that @Data10 presented are medians, not means. You cannot establish a confidence limit for a median without actually using the data, and even then, with a distribution that is far from normal, this is extremely difficult to calculate.

If the sample and the population had a more normal distributions, the actual CL would be from the 22nd lowest salary submitted and the 38th lowest salary submitted. The median is the 30th lowest salary. The 95% margin of error is 8 salaries, and the 95% confidence limits are 16 salaries. Which covers some 27% of the sample.

95% CL of the median are median ± 1.96 √(nq(1-q))
q is the percentile, which is ranked 0.5 x sample size, and n is the sample size.
The CL is in ranks of the data points, not their values.

Let’s suppose the reported information is based on solid data (large sample sizes, reliable sources, etc.) Even then, it would be hard to tell what exactly it is measuring or the implications.

I would expect that for exactly the same job at the same company, 2 different candidates with the same qualifications will tend to be offered pretty much the same starting pay, pretty much regardless of where they earned their degrees. If one gets a better package than the other, that may have more to do with negotiating skills than the diploma. All else being equal (which of course they never really are). Don’t expect top companies to pay a Brown bonus.