September 29, 2013

Chabris on Gladwell in the WSJ

In the Wall Street Journal, psychometrician Chris Chabris dumps on Malcolm Gladwell in a review of Gladwell's latest book.
Book Review: 'David and Goliath' by Malcolm Gladwell 
Malcolm Gladwell too often presents as proven laws what are just intriguing possibilities and musings about human behavior. 
By CHRISTOPHER F. CHABRIS 
... The idea that difficulty is good when it helps you and bad when it doesn't is no great insight.
In a recent interview, Mr. Gladwell suggested that the hidden weakness of "Goliath" enterprises is their tendency to assume that the strategy that made them great will keep them great. But there are prominent examples of companies that failed after not changing direction (Blockbuster and Kodak) as well as ones that succeeded (Apple deciding to stick with a proprietary operating system rather than shift to Windows). There is no prospective way to know which is right, despite what legions of business gurus say. Sticking with what has worked is far from irrational; indeed, it is the perfect strategy right up until it isn't.
One thing "David and Goliath" shows is that Mr. Gladwell has not changed his own strategy, despite serious criticism of his prior work. What he presents are mostly just intriguing possibilities and musings about human behavior, but what his publisher sells them as, and what his readers may incorrectly take them for, are lawful, causal rules that explain how the world really works. Mr. Gladwell should acknowledge when he is speculating or working with thin evidentiary soup. Yet far from abandoning his hand or even standing pat, Mr. Gladwell has doubled down. This will surely bring more success to a Goliath of nonfiction writing, but not to his readers. 

Interestingly, Gladwell revives Thomas Sowell's 1970s critique of the tendency of affirmative action to mismatch students. (I wonder if Gladwell left out the affirmative action part.) Chabris writes:
This is an entertaining book. But it teaches little of general import, for the morals of the stories it tells lack solid foundations in evidence and logic.
One of the longest chapters addresses the question of how high-school students choose colleges. The protagonist is a woman with the pseudonym of Caroline Sacks, who was at the top of her class in high school and had loved science ever since she drew pictures of insects as a child. She was admitted to Brown University and the University of Maryland; she went to Brown, her first choice of all the colleges she visited, with the goal of a science degree. 
Ms. Sacks ran into trouble early on in her science courses and hit a wall in organic chemistry. There were students in her classes who seemed to effortlessly grasp concepts she struggled with, and she got discouragingly low grades. She switched her major and looks back with regret, saying that if she'd gone to Maryland, "I'd still be in science." 

Is Brown really a Caltech-style sink or swim school? It may well be in the harder subjects, but I just don't know.
In this conclusion she may be right. Mr. Gladwell reports data showing that, no matter what kind of college students attend, those who start a science major in the top third of the ability range of students at their own college (judged by their SAT scores) are much more likely to graduate with a science degree than those in the bottom third—the odds are about 55% versus 15%. 

I used to be a 100% true believer in Sowell's argument, but now I can also see a lot of advantages of going to an elite rich school if you can get in. A lot of state flagship schools are sink or swim in the STEM fields, while rich private schools (Brown is the least rich Ivy) have more hand-holding resources.

31 comments:

  1. Berkeley is utterly brutal for the STEM kids. I was a dopey lib-arts major but I roomed with three STEM guys. They had their asses kicked every day.

    Just in terms of scheduling it was brutal. For pre-med, freshman year, everyone had to take chem and o-chem. The former was at 8 am M-W-F, the latter at 8 AM Tu-Th. That’s not counting lab and section.

    Meanwhile we dippy lib arts could schedule only afternoon classes and take Fridays off.

    Still and all, for everyone, Berkeley was sink or swim. There was no grade inflation. They had to clear the decks for the incoming raft of community college students, guaranteed spots according the California “master plan” for higher ed. The way they made space was by flunking out half the kids admitted two years before. Everyone soon learned the brutal truth: frosh-soph years were designed to kill you. If you made it that far, it got substantially easier.

    But not as easy as Stanford. We called that “the promised land.” All you had to do was show up and you’d be guaranteed an A-minus. And your degree was worth more on the back end.

    ReplyDelete
  2. Mismatch is probably the most interesting effect of Affirmative Action, and one of the few that will get mainstream Liberals to acknowledge that AA might not be quite the panacea it was intended to be. Typically mismatch is not as exaggerated at the elite Ivy's, Stanford, etc. but becomes painfully obvious at second tier universities like Duke or UT, for instance, and gets progressively worse as you go down the chain. This is partly for the reasons Steve mentions (the Ivy's throw more money at the problem and do quite a bit more hand-holding), but also because AA students at these elite institutions are, if not quite on par with their cohorts, exceptionally bright, motivated, and capable. The fact is, O-Chem at Duke is no less difficult than O-Chem at Harvard, but the AA student taking O-Chem at Duke is not the same caliber of student as his counterpart at Harvard. By the time you get to the AA student taking O-Chem at Penn State, you can pretty much forget about it.

    This is called the Cascade Effect, and is something like an academic Peter Principle: "Affirmatve Action students tend to rise to their level of incompetence."

    ReplyDelete
  3. a Newsreader9/29/13, 7:49 PM

    If I remember correctly, Brown doesn't even hand out grades. All classes are Pass/Incomplete (failing is so gauche).

    When it comes to mismatch, it depends on what the student's goal is. If it is to become a working scientist or engineer, then mismatch might be detrimental to his career. However, if he just wants to be a professional minority for hire, then what does it matter. Indeed, just getting into a highly ranked school is sufficient for this career path.

    ReplyDelete
  4. For what it's worth, Harvard is tough as nails in STEM subjects, and Caltech is hard but not absurd. I was an undergrad physics major and my sister was biology pre-med; in both cases the demands were severe. (We both did well, but only because we worked our asses off.) Caltech, though more brutal, still isn't cruel. When I taught freshman physics at Caltech, typically a student or two would fail each term, but even then they often got a second chance (i.e., if they passed term 3, they would retroactively pass term 2). And Caltech had some hand-holding; I was brought in as a special-agent tutor for some especially marginal cases. Such cases are unavoidable, because Caltech has a unified core curriculum—even biology (or English!) majors have to learn Maxwell's equations. Every major at Caltech is a STEM major.

    On balance, I'd guess that other ultra-elite schools (Ivies, Stanford, etc.) are similar to Harvard and Caltech in the rigor of their STEM programs: the general theme is that STEM majors are brutal everywhere. If Harvard seems less brutal than state schools (even a great one like Berkeley), it may just be that its students are better able to clear the high bar. Although Harvard has plenty of marginal students due to various admissions preferences, they tend to steer well clear of STEM, and the top 50% (at least) of each class is freaky good. I'd wager, for example, that Harvard gets more math, physics, and chemistry olympiad gold medalists each year than all state schools combined. Those guys aren't going to fail out of a STEM major.

    ReplyDelete
  5. I took a large-enrollment biology class at my California junior college and set the curve. The experience made me think about minoring in biology.

    Then, I transferred to Rice and took the large-enrollment introductory biology class there. I barely got a C. The experience was sobering. No bio minor for me.

    ReplyDelete
  6. If you want to go to medical school and don't want to study hard in college, go to a place like Stanford and major in sociology or something easy. Afterwords, gradually chip away at post-bac premed classes -- one or two at a time -- while gaining "life experience" doing something fun and not too stressful. You should will have a transcript full of As.

    Source: seen it done.

    ReplyDelete
  7. It's was that way when I went to a UC in the 70's but I never understood why. I read recently that Engineering schools take a perverse pride in running off all but the top students so they ones that aren't winnowed out make the schools look good. Here is one reason why the country is short on STEM grads.

    ReplyDelete
  8. a Newsreader said...
    If I remember correctly, Brown doesn't even hand out grades. All classes are Pass/Incomplete (failing is so gauche).


    You remembered incorrectly. In the first place, students at Brown since "The New Curriculum" came about have been able to choose between lettered grades (A, B, C and no credit) and pass/no credit. There is no "incomplete." The choice has to be made early in the semester.

    Academic advisors at Brown typically discourage students from taking more than one course per semester pass/no credit. Certainly graduate programs would not look kindly upon an applicant from Brown with a large chunk of his classwork with passes only.

    The idea of the system is to encourage students to take chances and try courses that they otherwise may not. For example, someone studying political science may take a visual arts class pass/nc.

    Also, there is no "fail" at Brown. If a student performs worse than C, he gets no credit and the EXTERNAL transcript would not show the coursework.

    Many students consider the engineering major (called "concentration" at Brown) to be the most rigorous with heavier-than-usual requirements. Unlike the typical undergraduate engineering work elsewhere, however, Brown's engineering education is fairly conceptual -- many Brown engineering graduates go on to work in fields like financial services, law, business and medicine rather than engineering.

    Brown's computer science department is also fairly high quality and its students are recruited by the usual suspects in the IT industry.

    One thing useful to keep in mind is that Brown and its nearest competitor Dartmouth are unique among the Ivies and other high quality universities of similar rankings in that they are essentially undergraduate teaching universities. Accordingly professors there are geared toward teaching rather than publishing research. They are also expected to be available and accessible to undergraduates in ways they generally are not at places like Harvard or Stanford. So undergraduates get unusual amounts of interactions with top professors.

    The down side is that it has a very samll graduate school and no professional school (except a medical school geared toward producing primary care doctors rather than specialists). One can easily imagine the negative impact that has on the endowment (I think Harvard Business School alone may have a larger endowment than Brown as a whole, but that may be apocryphal -- in any case the joke is that Harvard's rowing team has a bigger endowment than Brown does).

    That does not seem to hurt Brown's popularity as it seems to attract a lot of children of the wealthy and the famous and the students are reportedly ranked as the happiest among all top-tier schools.

    Of course, it is, as an institution, exceedingly leftist and its faculty has the highest Democrat-leaning voting percentage and contribution rate among the Ivies, I believe.

    In essence, it is SWPL university par excellence... especially for those who find the likes of Harvard and Stanford "too institutional" and "too oppressive and stuffy."

    JN

    ReplyDelete
  9. Douglas Knight9/29/13, 9:52 PM

    Years ago I read an econ paper that I can't find anymore that claimed that comparable students were more likely to fail out of Michigan State than U Michigan.

    ReplyDelete
  10. You remembered incorrectly. In the first place, students at Brown since "The New Curriculum" came about have been able to choose between lettered grades (A, B, C and no credit) and pass/no credit. There is no "incomplete." The choice has to be made early in the semester.

    Academic advisors at Brown typically discourage students from taking more than one course per semester pass/no credit. Certainly graduate programs would not look kindly upon an applicant from Brown with a large chunk of his classwork with passes only.

    The idea of the system is to encourage students to take chances and try courses that they otherwise may not. For example, someone studying political science may take a visual arts class pass/nc.

    Also, there is no "fail" at Brown. If a student performs worse than C, he gets no credit and the EXTERNAL transcript would not show the coursework.

    Many students consider the engineering major (called "concentration" at Brown) to be the most rigorous with heavier-than-usual requirements. Unlike the typical undergraduate engineering work elsewhere, however, Brown's engineering education is fairly conceptual -- many Brown engineering graduates go on to work in fields like financial services, law, business and medicine rather than engineering.


    That sounds about right. Thanks for the elucidation.

    ReplyDelete
  11. as a fairly recent graduate from Case Western, in Cleveland OH, we were a fairly good engineering school not quite the level of Rose Hulman or Caltech, but we liked to say we were as good as MIT, but with 10x the work. Looking back now, I realize that the undergrad coursework is really preparation for post-grad degrees and research work, especially material science and bio-engineery type stuff (helpful to have two major hospitals nearby). With only 3k undergrad to 6k graduates, its pretty obvious what the point was now. I assume we had AA students as well, but its hard to say because of the international congigent and the fact that we had a 1/3 transfer/drop ratio, and of the engineers, maybe 50% transfer or change majors. Fortunate in being smart, but I took a few classes beyond my competence as well - dynamic mathematics and an optimization theory class. The memory stings.

    ReplyDelete
  12. Using Rice as a benchmark, they might push you out of the Science and Engineering schools, but they want to see you graduate within however many years for the US News rankings.

    Plenty of cupcake Humanities/Social Science stuff that might be difficult to get As in, but easy for Bs.

    ReplyDelete
  13. Serious question you ought to explore: are STEM classes that different between schools (i.e. between elite and average universities)?

    I was an engineering major at a very average land grant school. The stuff was undeniably hard (and I personally was lazy), but it seemed pretty standardized.

    In other words, chemistry and physics, calculus and so on, seemed pretty scripted in what they had to cover. Even into upper level classes (thermodynamics, statics and dynamics, statistics, intro to electrical engineering, and so on), there really seemed like a basic body of knowledge that had to be covered in 14 weeks: you couldn't not cover some of it, or do anything more than what it was (my thermodynamics book, in particular, seemed like a tome of basic knowledge: it had what you needed to know, it was complete, it seemed, in 1985, like a basic book of knowledge typset in 1955).

    So do elite schools do more in freshman calculus, or basic thermodynamics? I had assumed the difference was that 1) elite schools just have smarter guys who learn the same material better, and 2) elite schools have more and better upper grade electives. Not having attended an elite school, I have no way of knowing.

    anon

    ReplyDelete
  14. "So do elite schools do more in freshman calculus": foreigners often find that funny, the idea that engineering undergraduates study school-level calculus at College.

    ReplyDelete
  15. "So do elite schools do more in freshman calculus, or basic thermodynamics?"

    You don't take "freshman calculus" at an elite university.

    http://www.math.harvard.edu/pamphlets/courses.html

    Harvard, for instance, offers a few "freshman calculus"-type classes, but note that if you actually want to major in math, (1) they expect that you already got a 5 on the BC AP, obviously, and 2) you have to take a bunch of harsh classes in proofs before they let you take typical "upper division" subjects:

    Each year several first-year students ask to skip the Math 25/55 level and start with Math 122 or another 100-level course. The Department, based on many years of experience, strongly discourages this. Even if you have taken several years of math at another university, even if you have seen every topic to be covered in Math 25 or 55, you will not be bored in these accelerated courses. The topics covered in Math 25 and 55 are not as important as the level and the depth of mathematical maturity at which they are taught. Taking Math 25 or 55 is the most intense mathematical experience you are going to have in any Harvard course, shared with the most talented of your peers. You may learn more advanced material in other 100- and 200-level courses, but never with the same speed and depth as in Math 25 or 55. These courses are not taught in any other university because no other university has the same caliber of first-year mathematicians.

    IOW: prepare your body. It will be a rough ride. And it will be nothing like taking differential equations or real analysis at a state school.

    --bbtp

    ReplyDelete
  16. Ok I am baffled. Steve do you really think hand holding helps AA students? I thought the whole premise of your blog was statistical disparities in intelligence between large families ( your definition of race) was insurmountable.

    ReplyDelete
  17. The MS degree from Michigan back in 1981 was ninety percent review material for those is us who graduated from IIT in India. In fact the math book, Complex Variables and their Application by Ruel V Churchill was covered in my second year in the undergraduate program. The real difference in STEM programs in the US are the labs where the superiority in cash shows through

    ReplyDelete
  18. Mike in Boston9/30/13, 8:49 AM

    I can also see a lot of advantages of going to an elite rich school if you can get in. A lot of state flagship schools are sink or swim in the STEM fields...

    I have been suspecting this for a while, and I am definitely going to keep it in mind for my own kids.

    One anecdote: I was E.E. at MIT, my colleague at Oklahoma State. Don't get me wrong, we both busted our hump in upper-division classes. The difference is in the engineering fundamentals: statics, dynamics, mechanics, etc. He knows that shit cold. I can only imagine how much his ass got kicked.

    By contrast, at MIT EE's could skip some of those or take a watered-down version. Similarly, mechanical engineers could take a watered-down version of circuits classes. No one was trying to weed you out on the fundamentals. There were lots of review sessions and tutoring, and freshman year F's were wiped off your transcript.

    So you got through the basics, and then if you couldn't hack the upper-division EE classes you switched to applied math or business or something and probably still graduated in four years. No picnic, but a lot less sink-or-swim than it could have been.

    ReplyDelete
  19. I agree with what Mr. Lomez wrote about AA- that the not-elite schools are where mismatch seems to thrive. Right now I am pursuing a second degree, and having taken courses at both the community college and the nearby State U., I must say that I am mightily unimpressed by the quality of the black and hispanic students. For the community college, I think of it as a sort of taxpayer-funded adult babysitter to keep them from getting pregnant or arrested. Of course, with super high unemployment and no good-paying jobs available, perhaps it's cheaper than prison and TANF?

    ReplyDelete
  20. Is Brown really a Caltech-style sink or swim school.... in the harder subjects?

    First of all, no top school is REALLY sink or swim. If you are failing, they will try to help. 2nd, I'd venture to say that the STEM material taught at Brown is similar in scope and difficulty to that at other top schools as well. The real difference is between STEM and non-STEM courses at any school. If someone is not cutting the mustard at Brown, the usual (and probably wisest) course is to switch to a non-stem major, not to transfer to a state school. In the end, if higher math is beyond you, it's going to be beyond you at the state school as well.

    ReplyDelete
  21. Right. When I was at a state school in engineering, guys who couldn't hack the math switched to marketing or management and cruised, while girls switched to education. Math is hard everywhere.

    ReplyDelete
  22. Caltech has changed tremendously since the 70s and 80s. As schools became ranked by graduation rates, Caltech had to find a way to make it less brutal. Back in the "bad" old days, a third of the entering class would transfer out. I knew guys and gals who ended up at Chicago, Harvard, and Stanford, and all thought it was loads easier. Easily 50% of the class failed a midterm or final in the required frosh math or physics classes. Nearly half of the entering class got an F in at least one subject before graduating or transferring. Nowadays, there's grade inflation even at Tech but the Core minimum is still tough. Compare Math 1. Caltech uses Apostol which is what the most advanced section of MIT's first year calculus course often uses. The only diff is MIT has three diff levels of calc for frosh. Caltech has only one. No rest for bio or business majors. But grade inflation that only leads to one or two failures makes Tech more bearable today.

    In contrast, let me quote a math prof from Michigan: "If only I could use Apostol for any freshman calc at Michigan..." Then she rolled her eyes.

    ReplyDelete
  23. Just in terms of scheduling it was brutal. For pre-med, freshman year, everyone had to take chem and o-chem.

    No one takes General Chemistry and Organic Chemistry the same year, much less as Freshman. You don't know what you are talking about.

    ReplyDelete
  24. Actually, the hand holding is cruel.

    But as weak as a student is when they are AA students at MIT, if they can hang on long enough to enter an easier major, in the end, they graduate with their brass rat. Everywhere n the world, that ring is recognizable as a sign that you are brilliant. MITers know better but no one else does.

    Same is true for women.

    My story: math prodigy without any mentoring got me into UCSD math in high school. Multivariate calc, Lin alg, diff was when I was 15.

    Went to MIT at 16. Bad idea. Stupidly they skipped me out of All of the math but Lin alg. MIT's linear algebra covered UCSD' in the first five weeks. I nearly flunked as I didn't understand a word even though I had set the curve at UCSD.

    I was a physics major. I was nowhere near as good as the other students in the major and I was too stubborn to switch. All the girls I knew did and several of the men. I did not learn my lesson. I earned Cs which was depressing because being alive and known by the prof earned you a C.

    The upper division courses in that major were the grad level courses everywhere else.

    I graduated in math where I was a B student instead.

    ReplyDelete
  25. Anonymous said...
    Serious question you ought to explore: are STEM classes that different between schools (i.e. between elite and average universities)?


    The curricula are not all that different, but the pace and intensity are. The latter are made possible by the peer groups that are highly select.

    The dropout rates at elite universities (STEM or otherwise) are not high, but when students do drop out, typically it is because of demoralization rather than inability to absorb the substance of the curricula per se.

    Very few students from highly urban areas or affluent suburbs drop out -- they are typically from intensely competitive high schools in the first place, both public and private, and are well prepared to handle the intensity (indeed, some select high school graduates find the competition level at elite universities something of a let-down from their arguably even more competitive high schools).

    The very few who drop out are typically from "small towns." Often they were valedictorians or salutatorians from their home town high schools and considered themselves one of the handful of the smarted pupils there. Imagine their shock when they arrive at an elite university and encouter legions of students from elite high schools who seem so confident and even seem to thrive, in fact, with the pace of elite level university education. Some of these peers may have even taken courses at elite universities already while in high school.

    On top of suffering from a bit of a "fish out of water" syndrome, it is easy for students like these to be demoralized by how hard the work is and how easily their peers from "big ponds" seem to handle the pressure.

    A small number of such students decide that they do not like the "environment" after all and transfer to another university near their homes, typically public ones.

    So, it is not so much the work load that weeds out "the weak," if you will, but the peer group.

    JN

    ReplyDelete
  26. The students at aforementioned uni are all wealthy preppies so rigor isn't a huge concern there. Much to certain profs' dismay.

    There was one engineering group project comprising myself, some athlete chick, and two rich Malaysians who couldn't really write sentences in English. Never got their opinion on whether it's a sink-or-swim school

    ReplyDelete
  27. Not surprised you didn't talk about this article in the same WSJ - as it is a thorough and effective refutation of HBD http://online.wsj.com/article/SB10001424052702304213904579095303368899132.html

    ReplyDelete
  28. If you're not interested in a STEM major and a STEM job, going to an "elite" university probably is better for your career prospects, provided you have the social skills to capitalize on the social networking opportunity you get at an elite school. You don't necessarily learn more Sociology at Harvard, but the people you meet in your classes (and your TAs and professors) will hook you up with better jobs where all that's needed is a sociology degree. Or you'll get into a better law school because your degree is from Princeton than from Portland State, and you'll get better job offers because you went to a "better" law school.

    ReplyDelete
  29. The difference between Berkeley and Stanford is, if you're struggling and not an affirmative-action admit, at Berkeley, you'll be told you're having trouble, and that there are lots of resources for you to go check out, and that's that. If you don't follow up, they'll let you fail out. At Stanford, or if you're an AA admit (or football player, etc.) at Berkeley, they'll follow up to make sure you're getting the help you need.

    ReplyDelete
  30. Maxwell Power said...
    The students at aforementioned uni are all wealthy preppies so rigor isn't a huge concern there. Much to certain profs' dismay.


    While elite universities have predominantly affluent student populations, I would hardly say that they are "all wealthy preppies."

    These elite universities are more diverse economically than prep schools.

    Anthony said...
    If you're not interested in a STEM major and a STEM job, going to an "elite" university probably is better for your career prospects, provided you have the social skills to capitalize on the social networking opportunity you get at an elite school. You don't necessarily learn more Sociology at Harvard, but the people you meet in your classes (and your TAs and professors) will hook you up with better jobs where all that's needed is a sociology degree. Or you'll get into a better law school because your degree is from Princeton than from Portland State, and you'll get better job offers because you went to a "better" law school.


    This isn't the 1920's and most Ivy Leaguers today do not get their first jobs from their friends or fathers of their friends. If anything, their peers are likely to be competitors for the most desirable jobs.

    However, it is true that elite markers are used as pre-filters by some employers. Certain major law firms, for example, only take applications from those who graduate from top 20 law schools. It's not that someone from a second- or third-tier law school couldn't be a great lawyer. It's just that this type of law firms don't want to take the time to separate the wheat from the chaff.

    Where the elite networking and friendships come handy is later on -- once one starts to enter the "player" stage and one's friends may also be near the peak of their careers.

    JN

    ReplyDelete
  31. I'm not really impressed that you consider them "diverse." They are, predominantly, wealthy preppies. You sound like some intern from the admissions office

    ReplyDelete

Comments are moderated, at whim.