It Depends What You Study, Not Where

The big thing for me with top schools is the "baseline quality of the degree". I didn't go to a top school, but spent my time studying everything math and CS related I could, and managed to get an awesome education. At the same time, I am fully cognizant of the fact that peers who were less driven likely were more easily able to skate by, for the most part. So when a resume comes across my desk from a guy from my alma mater, I don't exactly know how to evaluate it -- there's no "well this guy went to CMU or MIT, so I can assume a modicum of strengths". I wouldn't turn down a candidate for an interview because of their school (or lack thereof), but coming from top schools generally frames you in a different way, in my book. If anything, I tend to gauge the difficulty up slightly more quickly in my questions.

The figures described seem to argue the opposite of the headline. If you get a 12% return on the cost of your education regardless of where you go, then that means that more expensive colleges are providing a commensurate increase in returns.<p>The fact that they can maintain 12% return is fantastic, and you should be spending every last cent you can afford on that. After all, if I offered you a 12 cents a year if you give me a dollar today, and someone else offered you $12000 a year if you give them $100k this year, you&#x27;d probably take the latter.<p>And this is not even counting the fixed cost of the time you spend going to university. The cheap and the expensive both charge you the same in time. But the latter gives you a phenomenally higher return on that time.

These ROI calculations on degrees are more of a (flawed) basis for encouraging (less often, discouraging) college. They are not useful. It may seem like a good idea to develop an &quot;ROI for college&quot;, but it distracts from what&#x27;s important.<p>First, much of the value of a degree is in getting your first job (networking at elite schools must be helpful however). As your career progresses, the importance of having a degree diminishes. Leaving aside Med, Law, etc.. where you have to have it.<p>Second, your pay is not a pure output of your degree. What field you&#x27;re in and the salary prospects in that field, how hard you work, how much you negotiate you salaries, how often you change jobs, what part of the country you live in, (I would argue) simply how smart you are - some would say even your race and gender.<p>Another metric could be, cost of tuition vs. one year avg. salary in your field. If I went to college 20 yrs. ago for computing and the tuition was 5k a year, and the average salary is $100k - and go to college now, the cost is $20k - do the math on college value. Kind of a joke right now I can get the Android dev job for $100k prob without the degree.

When will journalists finally learn that Payscale is absolutely useless for this kind of analysis? The Payscale sample is solely graduates who never got a grad degree.<p>All this data suggests is that STEM majors are more likely to get high-paying jobs with only their undergrad degrees. It completely omits the many humanities&#x2F;arts majors who go on to graduate study and thence to perfectly lucrative careers.

Previous discussion:<p><a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=11627901" rel="nofollow">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=11627901</a>

I don&#x27;t know this survey. But I&#x27;ve read similar and recently published surveys on the subject. Notably, if you stratified the sample to remove computer science students the conclusion is reversed.

ROI is kinda useless in this context.<p>i.e., they&#x27;re essentially saying, if the uni is very selective (e.g. Harvard), the ROI is similar to say a state college.<p>But they&#x27;re disregarding that Harvard, whoever subsidises you, is a $100k investment or w&#x2F;e, and state colleges are $10k. Just making up ballpark numbers here.<p>Now the ROI might be the same, i.e. they might both end up earning roughly 1.5x their investment in annual salary. But seen this way, everyone appreciates there&#x27;s a gigantic difference in where you study, even though percent wise the ROI is exactly the same.

As if future income is the only &quot;return on investment&quot; in an education... smh

The data actually support the exact opposite conclusion from the headline: if the IRR of a degree is constant, then the payoff is proportional to the cost (i.e. the amount you invest). More selective schools tend to be more expensive. So pay N times more for tuition, expect to earn N times more in the long run.

Neither majors nor schools are randomly assigned to students, so I don&#x27;t think that there&#x27;s actually much to take away from this.

It says &quot;return on degree&quot;, so is the income measured relatively to what they spent on the degree? If so, it obviously depends still very much on where you study, as top schools cost much more.

There is another HN article that shows salary for STEM majors at mid career was equivalent, irrespective if school. According to that article, at 44, I&#x27;m making more than the average MIT grad of the same age, and I went to KU.

Not true all the time. There is a state university in the Philippines called Rizal Technological University, located in Boni Avenue, Mandaluyong City. In that university, there&#x27;s a college called CEIT or College of Engineering and Industrial Technology. One of the courses offered by that college is Bachelor of Science in Computer Engineering and in that course there is a subject called Advanced Programming. Guess what is taught in the subject called &quot;Advanced Programming&quot;? Ready?... they are teaching HTML. You read it right, HTML or Hypertext Markup Language.

Not how I think about it.<p>It&#x27;s looking at ROI. And the cost of education typically correlates with quality and long-term earnings. (Put your pitch forks down, not all schools, just generally.) So, if I think about this on a 20 year horizon. I have a $50K degree program at a state school vs a $200K degree from Harvard (numbers rounded). Lets just use 10% as the ROI. At 10% over 20 years, the return on the state school is ~$336,000. For Harvard, this yields ~$1,350,000. So, the ROI might be the same, but you basically get an extra million over 20 years for getting into Harvard.

<i>&gt; compare the career earnings of graduates with the present-day cost of a degree at their alma maters, net of financial aid</i><p>Woah, what? So a degree that cost $1 but raised your career earnings by $100 would be at 100%? 1000%? Not sure how they&#x27;re doing the math, but it completely dwarfs the rest. And shows how obviously flawed the measure itself is.<p>One of the key things left out is the baseline cost of time. A few years of your life is not free and probably makes for a pretty high baseline cost.

For most people I agree. However, as someone who has graduated from a top school (U. of Michigan), I can&#x27;t tell you how many times my credentials have got me both interviews and jobs. Where you go to school CAN make a big difference, but it depends on the school, the person, and the field of study.

It seems to say that where you studied matters if it&#x27;s not a STEM degree.<p>For me, the explanation is that it&#x27;s relatively harder to determine whether someone has learned their stuff if it&#x27;s Math or Science, but quite hard if it&#x27;s creative writing or history. There&#x27;s an xkcd about this.<p>That&#x27;s why something like an online coding test might work: someone clueless will not be able to write that palindrome quiz. You can have a zero false positive rate at the expense of a false negative rate (competent people can freeze in exam situations). With essay subjects, it&#x27;s pretty hard to separate the BSers from the real deal. You&#x27;ll always have some false positives.

Not that the headline is only true for CS&#x2F;Math. For humanities degrees, there is a clear trend correlated with where you study.<p>An interpretation is: for humanities, who you know matters more than what you know (ie, the connections you make).

This is only about earnings (the study is from PayScale), and does not take into account all the other advantages from elite schools, not all of which can be measured easily. Like money&#x2F;happiness metrics, eliding those advantages may cause one to come away with a simplistic understanding of the relationship between school and outcomes. America cares about elite schools for a reason:<p><a href="https:&#x2F;&#x2F;medium.com&#x2F;@spencer_th0mas&#x2F;why-america-cares-about-elite-schools-411798386ba7#.cxq6ugexr" rel="nofollow">https:&#x2F;&#x2F;medium.com&#x2F;@spencer_th0mas&#x2F;why-america-cares-about-e...</a>

this is actually a little disappointing form the Economist which usually has great high quality content. Do we really need to visualize statistics to prove that comp sci pays more than a degree in library studies? how about if you get your MBA from a community college vs NYU or Wharton? Once you compare apples to apples you&#x27;ll certainly see a statistical difference in ROI based on the WHERE.

We all listen to what we want to hear, don&#x27;t we.

It&#x27;s the connections you make at the top schools vs the lower tier. The educations are similar, people you make friends with are not.

If you want to get a job based on what you learn at college, you can do that at a lot of places.<p>If you want to get a job because your roommate started a billion-dollar tech company, or inherited a billion-dollar company, or has a parent who is the President of a foreign country, you had best go to Harvard (or maybe a Princeton&#x2F;Stanford type).

So best choice is to probably to invest in S&amp;P while learning to program on your own.

Note that at the most-selective universities there does appear to be a relationship between selectivity and education ROI.<p>The trouble with linear methods is that they often fit to the bulk of the population and have poor fit at the extremes.

Isn&#x27;t this a recent dupe?

For intelligence. Society, on the contrary, is build upon status, papers and titles - a hierarchy in general, so in the social aspect &quot;where&quot; matters the most. Ask any British about Oxford.)<p>Btw, studying CS in MIT (a school which traditionally taught principles and big ideas) and in, say, third-rate school that taught Java coding makes a huge difference.<p>It is not a coincidence that it was MIT and Stanford from where the best things in CS came from (and we should not forget the &quot;parallel universe&quot; of Standard ML and eventually Haskell, which came from UK schools).

It would be interesting to see returns broken down by quartile instead of just the average.

What is this &quot;it&quot; which depends? Is the curious language in this headline trying to say: &quot;it <i>matters</i> what you study, not where?&quot;<p>Ironically, <i>where</i> did this author learn grammar?

Where is medicine and law in these diagrams? Arts?

Line of best fit is wrong and a terrible approximation here.<p>If they used a curve fit they&#x27;d have seen the line curves up as Admission rate decreases.

It depends grammar, not accidently

MBA depends on where you study