ROI Data by Schools and Majors

A few universities/colleges that stuck out to me with disappointing results are Northwestern, Michigan, Pomona, Tufts and U Chicago … especially U Chicago.

1 Like

Chicago has extremely little engineering offerings, fewer CS majors than HYPSM… type colleges, and has become more selective (with higher academic quality of students) than certain earlier surveys used to estimate long term ROI. If you compare comparable students in comparable majors working in comparable areas, I’d be extremely surprised in Chicago is still far below peers.

CollegeScorecard has some majors specific controls. Example numbers are below, comparing Chicago to Cornell (Ivy with largest sample size). One college is not consistently higher or lower than the other. I’d expect the overall average are similar after controlling for major.

Computer Science – Chicago = $142k, Cornell = $158k
Economics – Chicago = $96k, Cornell = $83k
Political Science – Chicago = $51k, Cornell = $40k
Psychology – Chicago = $43k, Cornell = $51k
English – Chicago = $37k, Cornell = $39k

4 Likes

Here is the list of degrees that the Department of Homeland Security considers STEM:

1 Like

If the field isn’t challenging and the schools that offer the major in that field aren’t challenging, that means the barrier to entry into the field is too low. Even if the field pays well now, it isn’t going to last.

1 Like

I was fine until I got to Data Processing…

The entry level CS job market ranges between 50 and 500k

1 Like

Your upper end numbers are accurate, but not particularly useful for the vast majority of CS students because only a tiny percentage get hired at that level. Perhaps a few hundred per year, across all companies?

I feel something like a 95th percentile number is far more useful.

2 Likes

That is a fair point. This is a difficult number to estimate. We need the lowest pay of the top 5% of people. It takes work to estimate how many CS graduates there are in the US every year. This website (Computer Science | Data USA. ) estimates about 50k CS grads a year. That feels a bit low to me. So let’s figure 50k-100k. You need to know the lowest pay of the top 2500 to 5000 new grads. Let’s pick the “big four” employers to start – Google, Amazon, Microsoft, Facebook. Start with Google – they hire about 7k people globally (Here’s why you only have a 0.2% chance of getting hired at Google) – both laterally and new grad. Google is more US based than others such as Amazon. I would guess a large part of the hiring will be new grad. So conservatively 1000 in the US. Another way to think of this is that Google hires 50 APMs (Associate Product Manager) a year. The ratio of APMs to SWEs (software engineers) is 10 to 15 : 1 at Google. So perhaps hires 500 to 750 new grad SWEs a year. Also, industry attrition is high. Average tenure for a new grad at Google is 1.9 years. So the 1000 new grads are just replacement for all the attrition. I wouldn’t be surprised if the numbers are 2k a year. Amazon probably hires similar numbers. I think they had 1500 interns in the summer this past year in the US. Of those probably (guess) 1000 are junior interns. About 800 get full time offers. They probably have a yield of 50% – so 400 full time hiring from the internship. Maybe a comparable number directly after final year. Assume similar numbers for Microsoft and Facebook. So I wouldn’t be surprised if just these 4 companies hired about 4000 new grad SWEs. The number could well be 8000, because 4000 sounds low, simply because of attrition. Then there are another 15-30 companies that pay in the same ballpark – the likes of Linkedin, MongoDB, Lyft, Uber, Splunk, Cloudfare – whatever else… Some modestly lower than the big 4. Some higher. The big 4 are centered around 200k all in for first year. FB may be a bit higher at 220k or so. Amazon may be a bit lower at 190k or so. Bloomberg in NYC comes in at around 170k or so I think. I’ve seen 250K+ type numbers mentioned for Stripe as an example of something higher. Based on this back of the envelope estimation, I would suspect the 95th percentile pay is in the 200k-225k range.

1 Like

I will hazard a guess that your analysis is missing two huge employers of CS talent- the US Government, and the Banking/Insurance sector. There are the sexy and high paying jobs in these two sectors of course- and brilliant CS students take those jobs every year- but they represent a fraction of the roles that CS grads end up with.

Hence- I think your “back of the envelope” calculation wildly overestimates where the majority of CS grads end up for their first job, and therefore wildly overestimates the 95th percentile compensation. CS jobs in Charlotte, Nashville, Salt Lake, Tulsa, Dayton, Clovis, Los Alamos, et al aren’t paying 200K. Civilian jobs working for DOD have many fantastic benefits-- ditto FBI, CIA, NSA, etc. but starting salaries of 200K are not part of the deal.

Way, way off.

You can find some specific numbers on Glassdoor. New software engineer hires generally report salaries well below your numbers, and do not show a sharp distinction between "big 4’ and others. For example, some numbers I listed from another post are quoted below.

Google SV Software Engineers with <1 Year Experience
25th to 75th range for base salary – $129k to $152k
Average base salary – $138k
Average total compensation – $148k

Cisco SV Software Engineers with <1 Year Experience
25th to 75th range for base salary – $139k to $149k
Average base salary – $149k
Average total compensation – $149k

Along the same lines, while the maximum range listed among all grads in salary surveys sometimes exceeds $200k, these offers generally involve unique circumstances, such as starting with the equivalent of several years experience, so it is not an entry level position. Rather than focus on the maximum with unique circumstances, I think a 25th - 75th percentile + median type range is more useful, and many colleges report such numbers. For example, Berkeley reports the following for CS. Colleges that have a smaller portion of grads working in Silicon Valley or similar very high cost of living areas usually have lower averages. Colleges where a smaller portion of grads are high achieving students usually have lower averages.

Berkeley Outcomes Survey: CS Majors
25th Percentile = $110k
Median = $120k
75th Percentile = $130k

1 Like

I am basing this on the estimate that the total number of CS graduates in the US, annually is 50-100k. Based on the website I linked to. You can agree or disagree with the number the website had. The reason I left the banks and the US govt out of the estimate is because they are not at 200k, and I don’t need their numbers to estimate a 95th percentile pay number

That’s what many companies purport… that they couldn’t find the talent. Really? We all know that some if not many corporations have some work to do to improve on their Transparency. Real reason is cheaper labor (exploitation) even in the professional fields. Many candidates vying for student visas wouldn’t mind take a pay cut to get sponsored. Of course, there will be few cases where a highly specialized aerospace engineer or highly degreed molecular biologist may need to be imported; however, if they’re not careful with who they end up in their organization - profit or even non-profit - events like below will continue to be too frequent and damaging to US’ workforce let alone US. Period.

Emory professor hit with criminal charge, linked to Chinese government program (nbcnews.com)

These numbers don’t include signon, RSU grants that vest over 4 years, guaranteed cash bonus. You should look at Levels.fyi. Amazon pays 33k for just the 12 week summer. Other FAANGs are at similar levels for the summer. Last summer, if you could do low level systems work, Roblox paid 80/hr. They paid 100/hr for full time new hires. Last summer Lyft offered something like 9600/month for remote work. And the perception was that they were in significant financial stress due to Covid, and they went down from the prior year. So 3 month salary for a sophomore summer intern was around 29k.

There are US born and bred employees who have been charged with espionage, people who abuse their security clearances, etc. So is the answer not to hire Americans?

1 Like

You do understand that CS majors get jobs at banks, insurance companies, large supermarket chains, trucking and rail freight companies, infrastructure and construction companies, right? Nobody is getting RSU’s at these companies.

Your view is VERY skewed towards SV. Do you know anyone who lives in Wichita and has a degree in CS? Cincinnati? St. Paul? or even ANY of the other 48 states besides NY and California?

I understand all of that. @hebegebe just asked me for a 95th percentile new grad pay estimate, and I tried to estimate it. I also understand there are COLA differences.

I think that a 25-75 range is useful as well.

But given that the compensation skews upward as the percentile increases, I thought it would be useful to know what a very good CS student might hope to achieve. A 95th percentile number, if available, would provide that type of guidance. I suspect that @neela1 estimate of around $200k is close to accurate.

At the top research universities, attracting talented foreign students in STEM isn’t about making “even more $$ through the extra tuition”. It’s about keeping the US competitive. One of the biggest reasons US universities are what they are is because they attract talents globally. They attract not only foreign students, but also foreign scholars. Just take a look at faculty rosters (especially newer members) in STEM at top US research universities (fewer Americans have pursued PhD in STEM in recent decades).

5 Likes

I was referring to first year salary rather than summer internship. If you are including stock options, then I’d expect some of the highest earnings will occur at start-ups whose stock had a sharp increase.

In any case, my main point is in the earnings format/sources used in the linked ROI tables of this thread, college’s salary/outcomes report, and Glassdoor; earnings for CS majors are much lower. It’s not common for colleges that have a large portion of grads working in high cost of living areas to report starting median salaries of $100k for CS majors, but even 75th percentile ranges rarely exceed $150k. For example the full distribution for 2019 CMU SCS is below, which is one of the colleges with the highest reported CS salaries. The overwhelming majority of colleges in the US will be substantially lower. 99% of grads participated in survey. However, 44% of employed persons filling out survey did not report salary info. This likely presents a bias in the positive direction from actual. From a student’s perspective, I think it is more relevant to focus on the usual range of earnings, rather than the rare outliers with abnormally high earnings, often related to unique circumstances, such as substantial past work experience.

CMU SCS: CS Major 2019 Salary Distribution

Min = $50k
5th percentile = $70-$80k
10th percentile = $100-$110k
25th percentile = $100-$110k
50th percentile, Average = $110-$120k
75th percentile, Mode = $120k-$130k
90th percentile = $130-$140k
95th percentile = $160-$170k
Max = $175k

Most Common Employers

  1. Facebook – 13% of grads
  2. Google – 12% of grads
  3. Microsoft – 5% of grads
  4. Jane Street – 3% of grads

Most Common Locations

  1. Silicon Valley – 45% of grads
  2. New York City – 17% of grads
  3. Pittsburgh – 10% of grads
  4. Seattle – 10% of grads

You list Jane Street as 3% and you say the max salary out of CMU SCS is 175k. There is nothing more to discuss :-). Unless you are talking only about base salary. Then that is neither here nor there. Fundamentally there are things missing there. You should count how many FAANG placements CMU SCS has had in 2019. It is certainly more than 5% your data implies, unless you are counting only base pay.