Your Major For Life (Top & Bottom Twenties)

Will Truman

Will Truman is the Editor-in-Chief of Ordinary Times. He is also on Twitter.

Related Post Roulette

89 Responses

  1. Christopher Carr says:

    How would it figure in I wonder if you could factor in how well certain majors prepare you for professional education – i.e. how much more likely are history majors to be accepted to law school over some other major. How does majoring in English improve your chances of getting into medical school, etc.Report

  2. NewDealer says:

    Some observations:

    1. STEM majors probably pay more because they are notoriously brutal academically. Most people have trouble advancing in math beyond basic calculus including very intelligent people.

    2. For the bottom majors, I imagine a lot of people go into those fields knowing they are not well-respected or well-paying. How many would be clergy expect good salaries? Social Work and Education majors also get people who would rather do work they consider healing and making of the whole than well-paid.

    3. We need teachers, we need social workers, psychologists, and clergy (even though I am agnostic -atheist, I believe they can do good work). The low wages and disrespect that people receive is shocking.

    4. Wouldn’t more STEM majors reduce the salaries in the fields above?Report

    • trumwill mobile in reply to NewDealer says:

      1. Yup. And the jobs just aren’t attractive to a lot of people who could do it.

      2. Yup also. The low rates for general majors is concerning but those who went into teaching and social work knew what they were doing. Not sure about the psych majors.

      3. To be fair teachers make some of that back on the backend through pensions. But yeah, they should make more. Some places have a shortage and some places don’t, but even without shortages it would be nice if we could attract better people so that Co’s aren’t the dregs of academia (in the aggregate).

      4. Yes, though because of number one we don’t have to worry too much about it. Right now we’re importing a lot of talent.Report

    • Brandon Berg in reply to NewDealer says:

      Right. Ability bias is a big problem. Lots of people drop out of STEM majors because they can’t cut it, and wind up in lower-paying majors. It would be interesting to see these numbers controlled for SAT and high school GPA.Report

  3. Patrick Cahalan says:

    We’re number two! We’re number two!Report

    • Brandon Berg in reply to Patrick Cahalan says:

      Mathematics & Computer Science…I’m guessing that that’s because a lot work in finance?Report

      • Mike Schilling in reply to Brandon Berg says:

        The collapse was largely brought to us by ingenious though ultimately nonsensical mathematical models.Report

        • Barry in reply to Mike Schilling says:

          “The collapse was largely brought to us by ingenious though ultimately nonsensical mathematical models.”

          Nope, unless you believe that the VP’s and above in Wall St fell for models which promised high returns for low risk.

          Me, I prefer to believe that the top dogs, and went for things which gave them huge bonuses, and had the justifications drawn up accordingly.Report

          • Patrick Cahalan in reply to Barry says:

            Yes, it’s called “Whoo! Money!”Report

            • Barry in reply to Patrick Cahalan says:

              I always found that excuse beyond laughably thing; if there’s one thing that a 50-something year old Wall VP has seen for decades, it’s shiny, complicated finance models which made ridiculous simplifying assumptions and promised the world. In fact, the VP probably had to have those ideas slapped out of him when *he* was a bright shiny new graduate.Report

        • Kimmi in reply to Mike Schilling says:

          … that were written by people who hadn’t graduated college.Report

      • Angela in reply to Brandon Berg says:

        The Finance part is certainly an issue.
        However, the pay even for non-finance areas is also good.
        Frankly, the combination of big picture design with tedious exacting detail, added to interface issues with hardware / software / drivers, means there aren’t tons of people who are good at it. A lot of that can be “helped” with the more modern IDE and wizards, but then you need people to develop those. ahhh- job security!Report

  4. Anne says:

    yes! my two majors are not in the bottom 20!!! Boo not in the top 20 either
    Art History and Anthropology in case you are wonderingReport

    • Will Truman in reply to Anne says:

      Art History is #115, Antropology is #148, out of 173.

      My CIS major is #36. That might should make me feel good, but I was apparently seriously underperforming. I have my excuses reasons, though.Report

      • Jaybird in reply to Will Truman says:

        Theology would probably qualify as being its own reward. Philosophy probably falls into the same category, but it would have to pay more due to the higher probability of people not believing in such things as the afterlife.Report

    • Aaron W in reply to Anne says:

      Ha, I found it funny that “Art History” is usually the kind of major that people mock, but the unemployment rate is 6.9% and the median earnings are $45,000. That doesn’t seem so bad to me!Report

      • Brandon Berg in reply to Aaron W says:

        It’s not great, by college graduate standards. Only a handful of majors have a median salary below $40,000, and most of those are vocational training for low-paying jobs. I suspect that Art History’s saving grace is that it’s so impractical that it only qualifies you for jobs whose criteria include “A diploma in we-don’t-care, just to keep out the riff-raff,” so you don’t wind up in social work or preschool teaching.Report

      • NewDealer in reply to Aaron W says:

        I love art history but how many art history majors are doing work related to their field.

        Also working in a low-level gallery job, like the arts, or publishing, belongs to those with independent funds, the insanity to do weekend shifts in something else, or pure misfitism.Report

  5. Maribou says:

    Are these undergraduate majors? I always think of majors as being an undergrad thing, so I guess yes?

    Because majoring in “Library Science” for a bachelor’s degree is like… majoring in law for an undergrad. Ie, you’re getting a first-level college degree in an area in which ALL the well-paid professionals have professional degrees, and, people will suspect you of not having a broad enough education if you are that darn specialized.

    PS Anecdata: I make about 1/2 what people doing my same job who have the master’s degree I am working on make. Having a STEM degree as my undergrad could potentially make a huge difference to my earning potential once I have my masters, too.Report

    • Maribou in reply to Maribou says:

      (by “people” in that 2nd paragraph? i meant “grad school people” as well as “employer people”. That is like, the weirdest undergrad major ever, in the US.)Report

    • Will Truman in reply to Maribou says:

      Sorry, Mari, I missed your post. Yes, this is for undergrad. You’re right that some of them are going to appear poorly because higher education is considered part of the progression. I think that’s one of the reasons that the teaching jobs appeared as low as they did. Teachers don’t have to get higher degrees, but they get paid more when they do and a lot of them get their ME’s while they’re working. Which means that bachelor’s-only teachers are not only paid less because of their education level, but often paid less because they’re disproportionately just starting out.Report

    • Anne in reply to Maribou says:

      Art History is also a field where you are expected to get at least a Masters if not a PhD so maybe it would move up a list that included grad school
      I Have my Masters but my specialty field is Art Conservation in which unless you want to teach or be a Conservation Scientist a PhD is not an cost effective choiceReport

  6. KatherineMW says:

    But how long will this be true? All the top-level majors seem to be engineering and math, which will last until even smarter people invent computers that can replace you.Report

    • Angela in reply to KatherineMW says:

      The computers can automate the routine processing.
      Someone still needs to build and program the computers.
      Are you using computers more now than 50 years ago? Of course.
      Lots of other fields will be hit (hard) by AI and computer enhancements. Math / CS / Engineering will benefit.Report

      • BlaiseP in reply to Angela says:

        AI can only handle the rules. Humans will always create the policies and handle the exceptions. Humans get bored doing routine stuff and AI can only handle the routine stuff. Marriage made in heaven.

        Still surprised by how few CS types I see coming out of school with any practical AI on their course lists.Report

        • Patrick Cahalan in reply to BlaiseP says:

          You deal with more CS people on the ass end than most folk.

          You probably should write a post about “here’s what you should take in your undergraduate courseload if you want me to hire you when you graduate”.Report

          • BlaiseP in reply to Patrick Cahalan says:

            So I was given as a reference by an intern the other day. The firm called me up, asked me some interesting questions. Here’s kinda what I said.

            “Good coders are born, not made. She’s got the right stuff. Technology comes and goes but the problems don’t change much. She is thorough and deeply curious about the problems she faces. She understands process, gets along well with ordinary people and will give you the minimally complete solution every time out of the gate.”

            She got the job. I’m told she faced a great deal of competition for that job, too.

            Yeah, I could write such a post.Report

        • Patrick Bridges in reply to BlaiseP says:

          Part of that is how much background and experience practical AI (or practical OS or practical scientific computing or …) takes. Most undergrad AI classes are primarily AI survey classes so you get introduced to the ideas of support vector machines, alpha-beta and best first search, neural networks, and such. Applying it meaningfully and dealing with all of the real-world details (which you’re even more aware of than I am, Blaise) is hard.

          A lot of this has to do with how broad and deep computer science has become. As in Physics and Chemistry and all the other sciences, CS BS degrees are basic background breadth degrees. Students specialize either in their M.S./Ph.D. or in what they focus on continuing to learn after they graduate.Report

      • Kimmi in reply to Angela says:

        Java/Python lets stupider people code than Assembly.Report

        • BlaiseP in reply to Kimmi says:

          Two words. Shut up.

          And it’s not “Assembly” It’s assembly language for a given processor. As in 6502 Assembly Language. And when’s the last time you were called upon to write any — and for which processor?Report

          • Kimmi in reply to BlaiseP says:

            Blaise, I do image analysis for “fun”. Last code I was hacking (as in rewriting) was MMX and SSE (all int based work, most packing and unpacking).

            And all of it wrapped in decent C++, so I’m technically using assembly extensions.

            I don’t claim to be all that and a bag of chips. If I wasn’t doing graphics, I’d probably just use the Intel compiler and let it handle all the hardcore stuff.

            But starting a conversation with “shut up” seems like a nonstarter to me.Report

            • BlaiseP in reply to Kimmi says:

              So does saying a given programming language allows for stupider people to write code. I’ve seen horrible assembly language modules. I’ve seen entire shops over mainframe ASM macros.

              I am a programming language agnostic. If you’re wise, you’ll evolve the same strategy. The people I’ve met who write at kernel level all write in C these days and nobody opens an asm() block.Report

              • Kimmi in reply to BlaiseP says:

                oh, I don’t doubt it…
                But a good deal of the design decisions for Java were for safety and less “buggy” code. Also slower, and harder to customize (natural consequence of garbage collection).

                C’s grand for kernels, nobody’s sayin’ different (well, unless you’re writing a distributed kernel. in which case, you’re mad). Assembly’s good for what it’s good at, and it’s still damn nice in graphicsland (unless you’re playing around inside id’s Engine).Report

              • BlaiseP in reply to Kimmi says:

                Kim, I’ve been in Java since the night it was first available for download from Sun. I know more about writing Java applications than almost anyone alive and certainly more about how it operates on mainframes and servers than anyone I’ve ever met. I’ve had to live with those design decisions. I’ve managed just fine. It’s not always the most appropriate language for every application but no language is free of shortcomings and every one of them leaks if you don’t know what you’re doing.

                Java is evolving.

                And Python is a great language.Report

              • Kimmi in reply to BlaiseP says:

                Did I ever say I didn’t like python?
                nah, it’s swell.Report

              • BlaiseP in reply to Kimmi says:

                Glad to hear you think so. Guido didn’t whip it up for the benefit of stupid persons.

                I fire technology zealots, Kimmi. You wouldn’t last long in my shop, spouting off like that. When you see someone doing something wrong, that’s an opportunity for you to help, not to sneer.Report

              • Mike Schilling in reply to Kimmi says:

                But a good deal of the design decisions for Java were for safety and less “buggy” code.

                How silly of them.Report

              • Kimmi in reply to Mike Schilling says:

                Well, it’s better than big o notation, at any rate.Report

              • BlaiseP in reply to Kimmi says:

                Huh? Big O and computational complexity have almost nothing to do with performance tuning a Java application. This sort of compu-babble is why I end up having to un-train every CS graduate I have to manage. It’s I/O that matters, threading and thread safety, learning to instrument code, stop trying to do in code what the OS and the database and the services can do for you.Report

              • Kimmi in reply to Kimmi says:

                Yeah, I know, i bounced from one subject to a completely different one.
                Big O notation was grand when processing time was the bottleneck.

                And whether or not I sound like I know compubabble or not, I can assure you i’m no cs grad.Report

              • Mike Schilling in reply to Kimmi says:

                Big O and computational complexity have almost nothing to do with performance tuning a Java application.

                Only because the sort of applications where they are important tend not to be written in Java.Report

              • BlaiseP in reply to Kimmi says:

                I have to deal with computational complexity in PostGIS. And I write it in Java. Points, polygons, routing.

                There are two routes to “fast”. Fast in terms of time on the machine — or fast in terms of my time. If I can do it in memory, it’s several orders of magnitude faster than flailing on the disk. But trying to manage these stunts in someone else’s toolkits, thinking I might save some of my time, where the solutions are all generalised, it only looks like a time saver. Conversely, when it comes to my time, I quit worrying about truly optimal mathematical efficiencies once the ALUs got fast enough. Just wasn’t worth it.Report

              • Mike Schilling in reply to Kimmi says:

                We’re saying the same thing, but you’re saying it dismissively. You haven’t addressed big enough problem for O(N) to really matter, or, if you have, you’ve done it by using someone else’s already optimized algorithms.

                Try this: calculate the Fibonacci series in Java using pure recursion on

                F(n) = F(n-1) + F(n-2)

                and tell me how far you get. Now try it with an algorithm that remembers the previous terms instead of recalculating them each time. Last, tell me that inefficient algorithms don’t matter.Report

              • BlaiseP in reply to Kimmi says:

                If computing truck routing with PostGIS isn’t a big enough problem for O(n) to matter, I don’t know what is: sample problem: a truck leaves the Port of Los Angeles bound for Albany NY: compute the optimal route, working out how many miles were travelled in each state and compute the fuel taxes for each state. That’s a fairly significant problem, one you won’t get in a standard library.

                I constantly face O(n) over extremely large datasets. And yes, I’m going to start with my physical I/O as my first constraint. Every other consideration comes in a very distant second. That’s the real world, where the instrumentation rubber meets the road.Report

              • Kimmi in reply to Kimmi says:

                Linear search, done properly, is faster than binomial search. O(n) versus O(logn).

                And this is why I fucking HATE Big-O notation.
                Blaise is right — the bottleneck is NOT processing time. It’s memory if you’re lucky, I/O if you ain’t.

                (And that’s why I’m not a Java fan. If you can’t manage your bottleneck, you’re probably up shit creek. Java’s decent for things where memory isn’t the bottleneck. I do big datasets (graphics, I might have mentioned), heavy analysis…).

                Mike, it’s not to say that Big-O doesn’t have its place (mostly in “this wont’ get done before the end of the universe, stupid”), but… it’s just plain inaccurate for a lot of what’s taught.Report

              • Mike Schilling in reply to Kimmi says:

                Um, Kimmi, log N is *smaller* than N. Oh, and there’s no such thing as a “binomial” search. Well, maybe in Pascal. (That was a joke).

                Anyway, you and Blaise have fun.Report

              • Kimmi in reply to Kimmi says:

                ya, I know that log n is smaller than n(and yes, I knew I had the name wrong. ya can fix it for me, cause I forget. must be that whole “not a cs major”).

                I’m still not wrong. Because I know current computer architecture better than I know something folks came up in the 70’s-80’s when it was relevant.

                Do you know how big your L1 cache is?Report

              • BlaiseP in reply to Kimmi says:

                Well, Java on the server, especially on the mainframe, is a very different animal. I can literally bolt on a physical JVM and things change drastically. With an Azul Zing JVM, you get the same effect, with garbage collection happening under the covers.

                But I don’t architect systems which get into GC trouble any more. Part of code review is to look for instances of new anywhere and demand upfront initialisation and reuse where possible, especially via frameworks like Spring or CMP. It’s really no different than how you’d control memory leaks in any other language.

                Solve for I/O first. That will take your module’s grade from a D to an A- instantly. Sure, I don’t want to run inefficient algorithms any more than you do, Big O still matters.

                Here’s a case in point, where I will gladly trade iteration inefficiency for obviousness. I have an EDI 837 which I’ve parsed and objectified. Each developer is tasked with developing a validation rule. I gave him a little front end which he can point to an EDI file. The front end parses and objectifies it into some Java Collection and he writes his rule. He delivers his rule, it’s installed with all the others.

                That translates to a fair bit of extra string handling — but that’s trivial in the big picture. His rule is only one of thousands. Each rule will fire on a thread, reading the same object in memory and the results are damned near instantaneous. Seems inefficient until you realise how much time is saved in development. Processing power is time-cheap. Physical I/O is time-expensive. I route the troublesome EDI file to one of a dozen operators’ front end which, surprise-surprise, is pretty much the same front end the developer used to develop the rule, only pointed to the troublesome file, and now a human is doing the exception handling.Report

              • Kimmi in reply to Kimmi says:

                GC Trouble?

                … in a lot of what I do, routing anything to the user is a sign of an algorithm that needs improving.
                … well, the stuff I do in my spare time. The other stuff’s boring.Report

              • BlaiseP in reply to Kimmi says:

                No, Kimmi. They’ve submitted a claim which failed a policy rule. Ambulance rides for non-emergency treatment, that sort of thing. They can’t claim that.Report

  7. Barry says:

    “Mathematics & Computer Science ($99)”

    Sorry, I’ve got a math BS, and know better. That’s really reflecting Computer Science (plus grad school in other fields for Math BS holders).Report

    • Patrick Cahalan in reply to Barry says:

      It doesn’t say what you’ll be working *in*. My undergrad was in mathematics and my favorite stuff was abstract algebra and now I work with computers.Report

      • Barry in reply to Patrick Cahalan says:

        Which is important, in that there’s a lot of ‘math guys’ who would have been better off just majoring in CS.Report

        • Will Truman in reply to Barry says:

          And girls! Girls actually major in math in numbers not nearly as much lower as others within the STEM arena. Over 40% last I checked.Report

        • Patrick Cahalan in reply to Barry says:

          I dunno, I would have enjoyed CS courses that were analogous to theoretical math classes (computational complexity stuff) and disliked most of the programming ones, I’d guess. I almost killed my Fortran instructor.

          The meta point – and one germane to the OP, really – is this: if you’re not becoming an academic (see all the other posts in the symposium on why this is maybe not the most realistic of goals), and you’re not becoming a professional with a higher ed requirement (lawyer, doctor), then your undergraduate degree is only loosely coupled to your professional development thereafter.

          It’s about skills, not what’s on the paper. And really, most majors (including the STEM ones) provide you with a limited set of skills that can be abstracted outside their discipline. You pick up the skills through other methods more often than not.

          Best systems administrator I ever met was an English lit undergrad. Most of the better programmers I know were engineers or math majors, and not CS people, and they’re better programmers than your average CS folk for completely different reasons.

          Most of those upper-paying jobs require a metric short ton of process training. Most of those in the bottom chunk don’t.

          Generally, the ability to grok process is really important in earning more than the average office worker’s income. But you can learn how to grok process just fine without even going to college.Report

        • Patrick Bridges in reply to Barry says:

          I don’t buy that at all.

          A lot of computing is *very* mathematical, but many people with CS background don’t have the mathematical experience to handle it, because their are CS details they needed to specialize in learning. I do a lot of work with scientific computing and mathematicians who do numerical algorithms work, and the combination of their very strong mathematical background and good computing skills, and my strong CS background and decent mathematical skills (partially the result of a physics minor) pairs well.Report

  8. Barry says:

    Will, I don’t mean to harsh on you, but I’ve gotten to the point where I use people’s use of ‘STEM’ as a warning sign that they really don’t understand what’s going on.

    This acronym covers everything from math to computer programming to biology to chemistry to physics to engineering, and everything from a BS to a Ph.D. Anybody who actually knows what’s going on knows that the employment outcomes and salaries are incredibly different across those two ranges. And also that almost all of those are not in oversupply.Report

    • dhex in reply to Barry says:

      true. some of this is a side effect of the npr/ny times genre of weeper about the kid with the women’s studies degree or the lit degree from some top flight “known” fancy pants school (vassar, vanderbilt, something else that starts with a v) with 100k in debt, no job prospects, and living with their parents. plus the general bias from those not of the npr/nyt axis.Report

    • Will Truman in reply to Barry says:

      You’re absolutely right that STEM is not a thing, it is many things, and the situations between these things differ greatly. I use the term because that’s the context in which it came up here at The League.

      Whether there’s an oversupply depends on which are we’re talking about, of course, but depends on how you look at it elsewise. We are importing so much talent that I think it’s hard to say that we’re producing enough of a lot of these areas. It also seems to be the case that regardless of how the market for graduates may not as good as its boosters make it out to be, it’s still better than a lot of the alternatives.

      When I started, I had kind of suspected that engineering would, by and large, fare well. And that basic math and science degrees would fare more poorly. That seems to be the case. What’s funny is that in the other thread, I almost singled out petroleum and chemical engineering as something I might point out to my daughter. I’d no idea that the former would end up #1. (I’m not sure about the long-term prospects, so might encourage either chemical or environmental engineering even if they aren’t as high on this particular list.)Report

      • Kimmi in reply to Will Truman says:

        I’m not certain any of these are really all that and a bag of chips.
        I’m certain that a lot of these are skewed by large earners.

        I mean, lawyers with a JD make a median of $111,000 or so
        Physicists make median $105,000, but they get a free ride through post-BS.

        • Barry in reply to Kimmi says:

          One caveat is that many of these statistics might apply to what one does. A working physicist is a lucky person; most of them will work in other fields (probably programming).

          This applies to the law school thread as well; those who aren’t working as lawyers but found some job will be counted as either employed X or unemployed X, but not as unemployed lawyers. That’s how law schools can have such abysmal outcomes, but the unemployment rate ‘among laywers’ can be ~2%.Report

      • Barry in reply to Will Truman says:

        Will, thanks for taking my criticism well. In terms of ‘shortage’, remember that it’s a question of cost and working conditions. We import some people because they’re in the world elite, but far more because they’ll work cheaper (and if they’re H1B (?), they’re bound to their boss).

        As for petroleum engineering, I’ve seen this go up and down I believe three times now since the 1970’s. When oil is expensive, companies spend a lot of money in exploration and bringing in new fields; when it’s cheap, they cut those expenses savagely.Report

        • Will Truman in reply to Barry says:

          Barry, I’d actually had a paragraph in there where I talked about my views of STEM, but removed it because I wanted this to be more basic and not have the commentary. It actually included a bit about the variance within STEM.

          If what you say about PE is true, then it’s probably going to be a good field to go into in the future. I don’t see prices going down any time soon.

          I don’t know how much of the shortage is a question of cost, though. The numbers indicate that the pay is actually quite good, for the right majors (most engineering ones, for sure). Maybe not as good as you think it should be, but if someone is skipping this line of work because the pay isn’t good enough, I have to wonder what they’re actually doing that pays better.

          The only reason we do let these H1-B visa folks in is because the employers can at least somewhat credibly be able to claim that they can’t find Americans to fill these positions. Given the sheer number we do import, it tells me that we’re not making enough. There is very little in engineering I can look at and say “Well, no wonder they can’t find Americans willing to do that under those conditions for those wages.” Indeed, engineering is one of the fields that, unlike medicine and law (for instance), parents do very often recommend it to their children.

          We are agreed that there are some fields within STEM that have grim prospects. But there are, it seems quite clear, a number of fields where that isn’t the case. More in that ecosphere than just about any other I can find. (Again, for a variety of reasons, this doesn’t translate into “Everybody should go STEM!” for a variety of reasons, from vagueness to aptitude to circumstance.)Report

          • Barry in reply to Will Truman says:

            “The only reason we do let these H1-B visa folks in is because the employers can at least somewhat credibly be able to claim that they can’t find Americans to fill these positions. Given the sheer number we do import, it tells me that we’re not making enough. There is very little in engineering I can look at and say “Well, no wonder they can’t find Americans willing to do that under those conditions for those wages.” Indeed, engineering is one of the fields that, unlike medicine and law (for instance), parents do very often recommend it to their children.”

            Last I heard, from both qualifications and pay, H1B’s are not in the upper part of the skills distribution. They *are* brought in for pay (and, I believe, for bondage).Report

      • Lyle in reply to Will Truman says:

        Let me go back a bit to a recollection of my parents. When my dad graduated from Purdue in 1950 engineers had a hard time finding jobs, (the Korean War had not yet really cranked up), and a large number of guys who got technical training in the military went into engineering with the GI bill. Of course a few years later they were scarce as the military-industrial complex cranked up. I saw the same thing in petroleum engineering and Geology/Geophysics, after 1972 until 1981 was a time when the US industry was hiring in large numbers, then the price of oil crashed and for 20 to 30 years no one was hired, thus the number of folks in the majors declined. About 2000 major oil companies looked at the age profile of their workforce and discovered that within the next 10-15 years a large part would be at least retirement eligible, so they realized they needed to begin hiring to get folks trained to fill the positions being vacated in a few years. Then of course the gas/oil boom due to fracking hit the US and demand went sky high. So definitly fields in STEM go thru cycles. Now of course another question is what percentage by age 40 are managers and not really directly doing what they trained for?Report

        • Will Truman in reply to Lyle says:

          My brother was warned by everybody not to go into aerospace engineering because the market when he was entering college was just dreadful. It was the post-Cold War drawdown and a lack of enthusiasm for NASA on the part of the Clinton Administration. He didn’t care, though, because it was what he wanted to do. It turned out, enough people took that advice and voila by the time he graduated there was a market again.

          STEM fields definitely go through cycles, but even a down cycle in a lot of engineering majors isn’t what it is like elsewhere. And a lot of the times we have to ask ourselves, when a petroleum engineer can’t find work as a petroleum engineer, what do they end up doing? Oftentimes, the answer isn’t actually that bad because they have skills that are transferable to other things.

          I can’t remember the last time I saw widespread problems for employment in a lot of engineering fields. Petroleum engineering strikes me as one of the riskiest because you’re putting a lot of your eggs in the fossil fuel basket. But I actually think enough people might be staying away from it that it will continue to be good for the people that go into it (especially if it correlates with gas prices, which aren’t expected to fall).

          As I’ve said elsewhere, we can’t look at these numbers and say “We need to encourage everybody to go into these fields” because a lot of people aren’t suited for it, and oversaturation does occur (and occurs presently in some STEM majors and quite a few career tracks). I think we’re a long way off, though, from being able to argue that most STEM majors don’t have a notable leg up on most other majors.

          I have a really hard time with Mark’s liberal arts suggestion. I think it might be a better world if we encouraged more generalism and less specialty, but I don’t think that’s the economy we live in. It’d require more on-the-job training than a lot of companies are willing (or have) to do.Report

          • Barry in reply to Will Truman says:

            “STEM fields definitely go through cycles, but even a down cycle in a lot of engineering majors isn’t what it is like elsewhere. And a lot of the times we have to ask ourselves, when a petroleum engineer can’t find work as a petroleum engineer, what do they end up doing? Oftentimes, the answer isn’t actually that bad because they have skills that are transferable to other things.”

            (note: this applies to the law school thread as well, for those reading both).

            One of the things to keep in mind is that most people can find some sort of job, especially new grads. That’s not the problem; the problems are (a) you can spend four years working hard to acquire skills which are not that valuable and (b) crap jobs pay like crap, but your student loans are for life. The fact that field X has miniscule unemployment rates is more a function of the bureaustat fact that unemployed X’s find work doing *something*, and therefore are no longer unemployed X’s.Report

          • Kimmi in reply to Will Truman says:

            Engineering has been in a downcycle since the 1970’s.
            I’ve had friends get scooped up by Australia, because they’ll pay better.Report

  9. Barry says:

    Sorry – ‘almost all of those *are* in oversupply’.

    (I don’t really care what employers say, since their definition of ‘shortage’ means ‘no long line of MIT grads waiting for a chance to work for room and board’).Report

  10. So, here’s something that really jumped out at me – except for chemical engineering and the pharmaceutical sciences, all of the top 10 are extremely narrow majors that are offered on a limited basis (their all amongst the 40 least-popular majors) that are in especially high demand at the moment.* Even chemical engineering and pharmaceutical sciences are quite narrow majors, though they’re much more widely available and in the top 60 or so most popular majors. Obviously, they’re also in high demand, but the fact that they’re as popular as they are while also commanding significant wages and high employment suggests that their ranking is more sustainable. In other words, for the other 8 majors in the top 10, their relative success is likely owed to the fact that they’re in relatively short supply. This pattern largely holds for the 11-20 most successful majors as well, with only 3 of that group being in the top 100 most popular majors, and 6 being in the 40 least popular majors. The point being that for most of the top 20 most successful majors, that success is owed largely to the fact that the number of people graduating with those majors is vanishingly small. If any meaningful trend developed to open these majors up to more people, or towards more people having an interest in these majors, the bottom would quickly fall out on their success.

    The bottom 20 show a similarly interesting pattern – they’re also almost all highly specific majors (ie, they’re oriented towards particular career training rather than being liberal arts-style majors**). While many of them are also amongst the least popular majors (see note below again), 10 are in the top 90 most popular majors, and only 7 are in the 40 least popular majors. But even more important, 9 of the 20, and 10 of the 21, least successful majors are in the field of education. This group of 10 includes the 8,9,48,52,and 58th most popular majors overall. That’s an awful lot of people majoring in a field that only has relevance to one possible career path, which also happens to be a career path that can be somewhat easily entered – and often is entered – by people who do not have a major in that field. Yes, it is a career path for one of the largest professions in the country, but it seems pretty clear that there are still way more people with education-related majors than there are educator jobs, much less educator jobs that require an education-related major; even if we significantly increased education funding, this would still likely be the case.

    I strongly suspect that if as many people majored in, say, Petroleum Engineering as in General Education, Petroleum Engineering’s employment numbers in this survey would look an awful lot like the numbers for General Education.

    My point here being that, unless you’re majoring in something available to only a comparatively small number of students, and in a field that you can reasonably expect to grow rapidly or continue growing rapidly upon your graduation, you’re probably better off going with a more typical liberal arts major than with a major that is intended to train you for a specific career.

    This dynamic is perhaps most apparent by contrasting the employment numbers for political science majors (in my experience, the most popular major amongst actual law students) with the employment numbers for pre-law majors.

    Pre-law majors rank a comparatively abysmal 127th based on Will’s methodology, with an almost 8% unemployment rate and a 25/75 range of $32k/$69k. Perhaps more importantly, the median wage -$45k – is a lot closer to the bottom of that range than it is to the top. By contrast, Political Science ranks a quite respectable 62 (and the often closely related international relations major ranks 51), with only a 6% unemployment rate and a 25/75 range of $38k/$91k. The median wage -$57k- is still a lot closer to the bottom than the top, but it’s also nearly $20k higher than the 25th percentile, where the median wage for a pre-law major is only $13k higher than the 25th percentile.

    Ultimately, the lesson I draw from all of this is that, unless you’ve got a unique line on a major that is in short supply but high demand, your best bet is to go with something in the liberals arts that gives you an opportunity for a plan B if you’re first choice career path doesn’t work out.

    *That they are so comparatively rare also raises a sample size question, but for my purposes, I’ll assume the sample size issue would not significantly affect the results here.
    ** Interestingly, the two liberal arts-style majors in the bottom 20 are both interdisciplinary majors. I strongly suspect this has to do with the underlying personality traits of someone who just decides to major in everything.Report

    • You’re right about the narrowness of a lot of the all-star majors. Where I take issue with your analysis is that almost all of the engineering majors do quite well. There are no examples of engineering majors falling down to Gen Ed.

      It’s true that these majors are getting the opportunities they are precisely because their ranks are so thin, but that also demonstrates, to me, that there is room to grow here. Due to the natural limitations on the number we can produce, I continue to believe it’s about as safe as you can go in this uncertain economy and world.

      If, that is, you can do it. And that’s where a lot of STEM-boosters are just wrong. A lot of people lack the aptitude. Others may have the aptitude, but lack the interest (and it’s not the sort of thing you can do when you’re disinterested). Also, as I said to Barry, your major within the STEM umbrella matters a great deal. Just because a major has the word “science” on it doesn’t mean it will actually pay well.

      So what should those who don’t have the aptitude do? Well, as mentioned, STEM covers a large area and if you can’t do petroleum engineering, you may be able to get a CIS degree. Of course, even that’s not right for some people. So the best advice would be… have an idea of what you want to do when you get out, and ask people in that industry whether they’re happy and (if there’s no direct major feeding into it) what you should do.

      It’s funny how we scoff at the idea that 18 or 19 year olds are able to find a suitable spouse, yet saddle them with the need to choose a lifelong career. But… that’s the world we live in.Report

      • Point taken, particularly your point about fallback options for engineering majors, which allows even narrow interests to open up broader career paths. If I can refine my thesis a bit, I guess the point would be that, whatever one’s temperament and skill set, the most important thing is to choose a major that opens up a lot of fallback options, ie, is readily transferable to a wide array of other career paths should your ideal path be unsuccessful.

        As you say, all the engineering majors rank highly both because they’re in high demand and also because they provide for a good degree of flexibility if Plan A doesn’t work out, but the same is not true of the other STEM fields. A good number of “hard” sciences do quite poorly – most everything in the biology arena is in the bottom half of the rankings, for example, and even biological engineering does significantly worse than the other engineering fields.

        Regardless, I think the main point I’m trying to get across is that the most important difference between majors that are likely to put you on a course for economic stability or prosperity and those that are not has only a little to do with STEM vs. not-STEM. Instead, it’s a function of demand for a given student’s Plan A combined with the flexibility the major gives the student to develop a Plan B. Majors that help teach generic skills like critical thinking and problem-solving do reasonably well even where the number of jobs in that specific field are small or almost non-existent; majors that seek to teach specific skills for a particular career path, not so much.

        Basically, if you think you want to be an English teacher, major in English rather than Education; if you want to be a math teacher, major in math rather than Education; if you want to be a lawyer, major in polisci rather than pre-law; if you want to be a businessman, major in economic rather than business management; etc., etc.Report

        • Patrick Cahalan in reply to Mark Thompson says:

          I guess the point would be that, whatever one’s temperament and skill set, the most important thing is to choose a major that opens up a lot of fallback options, ie, is readily transferable to a wide array of other career paths should your ideal path be unsuccessful.

          The biggest open fallback option is the willingness to fallback.Report

    • Barry in reply to Mark Thompson says:

      We’ve seen this in nursing; the present nursing ‘shortage’ is long gone.
      Except, of course, in the sense that hospitals who treat their nurses like sh*t find out that they lose nurses.Report

      • Will Truman in reply to Barry says:

        If there are more places looking to hire nurses than there are nurses looking for work, there is a shortage of nurses. The reason that a particular place might have more difficulty finding nurses than another which is fully staffed could be related to working conditions or pay, but that doesn’t really change the formula.

        As it stands, nursing graduate unemployment sits at 2.2%. And a whole lot of hospitals are hiring MA’s when they would prefer to be hiring nurses, which if anything obscures the degree of the shortage.Report

        • Barry in reply to Will Truman says:

          Will, what I meant was that I’ve heard of hospitals now refusing to hire nurses out of school. Only experienced nurses need apply. Which means that there’s no shortage of nurses.Report

  11. Alex says:


    The very fact that Will broke down STEM by major tells me that he’s not just blowing smoke about a generic STEM shortage. I notice that certain STEM majors are conspicuously absent from the list.

    And I don’t like the grouping of math/computer science either. If computer science were just Turing Machines and all that then this grouping might make sense, but nowadays many CS graduates might be accurately described as software engineers, which is very different from what math graduates learn. The grouping of math/computer science is something that stopped making sense in the 70’s (if not sooner).Report

    • Barry in reply to Alex says:

      The reason I’m sensitive is that I graduate with a math BS in 1989, and found out that it was worthless; after graduation I was working as a security guard.Report

    • Lyle in reply to Alex says:

      Actually one needs to seperate computer science into two fields the actual coding work, and the system analysis to find out what the customers needs are and figure out how to meet them with a software system. The latter is what industry wants, not coders. Determining the problem and finding a solution is a valued skill. In fact there is now an area called enterprise IT architecture that covers just this.Report

      • Will Truman in reply to Lyle says:

        Two fields, perhaps, but the question is academic tracks. If Computer Science folks fill both, then it should be baked into the numbers. That’s one of the reasons that I consider the numbers important. They tell us where people end up. In some cases, even if they don’t do precisely what they went to college to do, they still do well. My post-college roommate got a degree in physics. He was hired by a chemical plant because his science background was sufficient (he left to get his MS/PhD in astrophysics). On the other hand, Barry ended up as a security guard. This is among the things we need to consider.

        (And while I consider posts like this to be okay, in more serious and specific conversations, we really do need to delineate between fields within STEM. All of which are not created equal. My roommate’s former job aside, I’d respond differently if Lain said she wanted to major in physics than if she said she wanted to major in mechanical engineering.)Report

        • Patrick Cahalan in reply to Will Truman says:

          Computer Science can’t fill both; you actually need three.

          Computer Science is about understanding computation. Software engineering is about building software systems. Information Systems is about designing human/machine interactions.

          The best degree programs will introduce you to all three concepts, but it’s hard to get a solid grounding in any of them without a full major.Report