I spent the last few days reviewing Graduate admission files, even though I am not on the admission committee. Mostly out of my own curiosity, but also because a number of students applying to our department emailed me personally expressing interest in joining my group, and so I decided to review their files, along with ~500 other files. Since it's all done electronically now, I can do it from my laptop far away from my office.
Is it possible to evaluate application in a fair, objective and unbiased way? Or do we always subconsciously favor certain candidates (domestic vs. foreign, male vs. female, experimentalists vs. theorists)?
Letters of recommendation are typically very subjective and difficult to evaluate. You have to read between the lines, and that is very difficult - even if you know the recommender well. You have to parse "enthusiastically" from "excited" from "exceptional" from "outstanding". Which one of the words mean exactly what? Is "truly outstanding" a lot better than "exceptional"? Is "it is my pleasure to recommend" much worse than "I would like to enthusiastically recommend..."? The recommenders are asked to mark whether the student is top 1%, top 5%, top 10%, top 20%, top 40% etc.
But these are very subjective too. Even people in the same department often mark the same person as ranging from top 1% to top 20%. What does it mean? I think it reflects more on the recommender than the student. Besides, being top 1% in physics classes at small college full of liberal arts majors may be equivalent to being merely top 20% at MIT or Caltech where everyone is a nerd (more on that later).
Same goes for GPA - the variations are as large as error bars.
GRE test: Verbal is pretty meaningless, unless someone wants to major in science journalism. Analytical writing doesn't have much meaning either. And Quantitative section should be maxed out at 800 or so for most physics majors.
I only wish there was some sort of objective physics test that was administered evenly across different schools, countries and continents - and it was tough enough to differentiate between candidates!
Fortunately there's just such a thing - and it's called GRE Physics. For those who are not physicists, GRE Physics is a 3-hour test consisting of 100 multiple answer problems (assuming the format still the same as when I was applying). Maximum score is 990, which is about 97% percentile, and the score of 900 or higher means you are in top 10-11% of people who took the test.
However, GRE Physics scores are often downplayed or downright ignored by some committees. Partial reason for that is that the foreign applicants tend to ace GRE Physics test, and most americans tend to do poorly. As a result, our department routinely declines multiple (foreign) candidates who scored 990 (maximum) on GRE Physics - and in addition have great GPA, are in top 5% of their class and may even have done some top-notch research - in favor of domestic candidates who may have scored 600 or 700 on GRE Physics.
There are multiple reasons for that - for whatever reasons most interest in astronomy, plasma, nuclear and high energy physics comes from domestic students, while foreign students are more interested in condensed matter, bio, materials etc.
To a lot of schools, foreign students are more expensive, and the language barrier is often quite significant in considerations of hiring first-year graduate students as TAs.
Some faculty members tell me that it's "difficult to evaluate students from China", and some of it is certainly true - the letters are from unknown to us professors, their research statements are worded poorly, etc. But I also heard people imply that the fact that most foreigners ace Physics GRE should be viewed with a grain of salt. I certainly don't think that GRE subject score defines everything you need to know about a student, but it DOES tell us something - and I would argue it's the only *subjective* measure that is uniform across the board. I think it's as good measure as any of overall undergraduate coursework preparation across major areas of physics, and probably correlates very strongly with grades in graduate-level classes and ability to pass qualifiers. Correlation with research may be less strong though.
One graduate student was telling me that a certain highly ranked department makes the first cut by eliminating any application with GRE Physics score less than 900 - and after some selections are made, they go back and fish out students who scored below 900 but are outstanding in other ways. This may be an urban legend, since average GRE subject score for even the most selective top-10 schools is somewhere around 850-900.
So here's some statistics from the current pool of applicants at my department:
Typically our department gets about 500 applicants, give or take. We make offers to 100-150, and about 20-40 accept. It's a difficult game of guessing how many students can be supported, how many students will accept, which sub-fields need to be represented, etc.
Last year's entering graduate class composition was 23% female and about 40% foreign students. This is probably quite typical across other departments.
25% of all applicants have GRE Physics scores of 900 or higher. 12% of all applicants (or about half of 900+ group) have scored the maximum - 990.
Women make up 10% of 900+ group and 8% of the 990 group.
Students listing their specialty as condensed matter experiment make up 51% of 900+ group. Condensed matter theorists are 19% of 900+ group, HEP theory and experiment are 7% and 4% respective. Plasma is 5% and bio is 11%.
Americans are about 5% of the 900+ pool, and that's counting two canadians. All of the 900+ americans are male. None of the americans who scored 900+ (and there are several 990 among them) are from what I would call an elite university (top-10 ivy league-like schools like Harvard, Princeton or MIT) - but it's also possible that american students from elite schools who scored 990 are confident enough to apply ONLY to elite top 5 or so universities for graduate school.
More than 90% of those who scored 900+ are from China/Hong Kong. There are a couple of students from India, Korea and Taiwan and a few from Western Europe. Very few candidates are from Eastern European countries.
I guess the part that bothers me is something that probably won't surprise most people reading this blog - there are different standards set for domestic and for foreign students. Domestic students who aced GRE Physics and have reasonable research experience and letters will have an easy pick of accepting offers from most selective departments. Domestic students who struggle in GRE Physics can compensate by being active in research, or have very good letters, or - better yet - reasonably good letters from someone who committee members know well. However, for international students, especially students from China, the bar is raised quite high - even acing Physics GRE and being involved in some undergraduate research may not be sufficient - these students are trying to differentiate themselves from literally a hundred of other students from China with perfect Physics GRE.
I am still unclear as to reasons why americans can't ace Physics GRE. Is it that undergraduate preparation in China is so much superior to US? Is there a cultural difference in amount of diligent and single-minded focused studying required? How important is the timing element - 100 questions in 3 hours is a pretty fast pace, as far as I recall.
A lot of american students, supposedly coming from top 1% of their classes, struggle not only with Physics GRE, but also quantitative section of general GRE test, which I always thought was extremely easy - by struggle I mean scoring 700 or 720, as opposed to 790 or 800. Why is that?
On Cobert report last night there was a guest who was arguing that IQ tests don't measure intelligence, but instead they measure test-taking ability. I think it measures some linear combination of the two, but from what I heard variations of IQ tests are strongly correlated, which implies that they DO measure something meaningful.
Wednesday, January 23, 2008
Subscribe to:
Post Comments (Atom)
35 comments:
IP, something I've always wanted to know about your American tests (GRE, SAT): Why are they set up in a way to make time such a very critical factor? In most of my physics exams, time was not all that important, it rarely happened that students would have got much better grades if they'd had a bit more time. It went much more along the lines of "either you know how to solve such problems or you won't figure it out even if you have half an hour longer." Are university exams different in the US?
Additionally, I'd argue that time is not so critical in research: It's much more important to keep long-term goals in mind and to work steadily towards them than to be very fast for 90 minutes.
So, what's the idea behind these racing exams?
schlupp - the timing factor may have something to do with the style of ETS - the Educational Testing Service. There are some logistical restrictions - how long can a test last? 3 hours is already pretty long. 4 hours? 5 hours? But they do want to cover a wide range of questions from all areas of physics. I could also imagine that tests become almost linearly more expensive if you extend the time.
However, I don't see why they can't administer a shorter, 50-questions test in the same 3 hours, which will hopefully reduce timing as an issue.
Playing devil's advocate - time may be a useful factor here - certain things they test you for, you either know it, or you don't. IQ tests use timing too. People who are really smart and well-versed in the material can do problems faster than people who aren't.
I should point out, however, that a lot of exams in US educational system - especially at graduate level - are open-book, take-home exams that last 24, 48 hrs or even longer. These come fairly close to simulating the kind of problems you solve when stuck in your experimental routine and need to figure out a solution - use books, computers and whatever you want, and have hours and hours at your disposal.
In my personal experiences, these numbers like GPA or GRE score say nothing about any applicant's research potential. I have seen stellar undergrads who failed in grad school and seemingly mediocre others who turned out to be excellent researchers. Grad school is based on research almost exclusively and I would be more interested in learning these students' undergraduate research projects if there is any. If you look at Princeton's admission site, they out some statement along this line. There will be plenty of time during grad school years to master the material that hasn't been mastered as an undergrad. What matter is one's ability to carry out research.
A friend of mine was the valedictorian of our undergrad class. He had excellent scores on all tests. He did his PhD with John Schwartz at Caltech on superstrings and he is a priest in Colorado now...
The statistics are really interesting :)
I'm in the graduate program of astronomy in a large public university, but I hear tons of rumors about those Chinese applicants with 990 on physics GRE.
Most professors have some sort of consensus that many of those students are doing cheating in any sort... If not cheating, anyway my department has very severe criterior for the admission of chinese students too. They complain that chinese students often quit school and get job in the us...Maybe this is more critical point than the test score.
I need to point out to be fair that the us undergraduate physics education is said to be 1~2yrs behind compared to asian (China, Japan, Korea, etc...) physics education by the time of graduation. I've studied in the normal public university in the us, and really agree with that statement. 1st year (and some of 2nd year physics) was marely the reproduction of the high school physics in Japan).
masatoshi - I heard professors implying there's cheating going on too - which is part of "we don't know how to evaluate Chinese applications". I personally think this is BS - there's no cheating, at least not much more than in US.
I don't think it's that difficult for a talented, well-rounded physics majors to score 990 or close. So the real question is not "How is it that 500 chinese students can score 990?", it's "How come so few supposedly *gifted* american students can score 900 or more?".
The real issue is not even whether GRE Physics is a good predictor of research performance. It's not like american students deliberately underperform on GRE tests in protest. It could be a stupid test, but if it benefits you to do well, you do your best to ace it. Just like driving license test may not be the best test of your driving skills - that's not a good enough reason to fail it.
IP - I think the amount of effort that students expend on each part of their grad school applicants is correlated to the weighting of each part by the admissions committee. Since the GRE tends get less weight than recommendation letters, research experience and GPA, students will put just enough effort to do decently but not make the extra effort to do spectacularly. Conversely, since these three factors are difficult to assess for international students, they're much more strongly motivated to score highly.
There is a history of rampant cheating by international students on the regular GRE, which is why the exam was restructured several years back. People would compile a complete set of GRE test questions off the exam computer and sell the solutions to students for memorization.
Anonymous - sorry, but I am just not buying your justification.
I don't think that students have a very good idea as to relative weight assigned to various aspects of their application package - and I can bet you that all 5 or so domestic students applying to my school who scored 900 or more will get accepted - and probably at other schools too. So GRE Physics score still matters. In a strange way, the high GRE score helps domestic students more than foreign students. So it totally would make sense for domestic students to ace it, if they can do so.
Secondly, I now strongly suspect that the weight of various parts of application is adjusted as to produce a desirable demographic distribution, as opposed to the other way around. It's the tail wagging the dog problem.
Finally, the idea that evil foreigners get 990 through cheating and honest domestic students get 600s only because they are not cheating is a myth. Not every Chinese student gets 900 either, by the way, and there are plenty of americans who get 990. I scored well on GRE, and so did many of the people I knew - and none of us cheated. Watching a lot of foreign graduate students study and take tests, I am absolutely confident that a lot of them DID get 990 honestly. This was of course back then - before cell phones and other "cheating" technologies.
And I don't think having your friend recite the typical problems they had in their tests helps you much anyways - chances are - you are going to get different problems. I also suspect american students talk to each other about their GRE experiences too, so it's not like China has a monopoly in this area.
IP, although cheating happens in all countries, I think China has a big problem with it. A few years ago, a chinese classmate showed me a huge book with many GRE questions, not only Physics, but also verbal and qualitative (from when it used to be multiple choice as oppose to essays). This chinese subject told me that whenever someone took the exam, they would memorize as many questions as they could and then write them down as soon as the test was over to later share them via internet. Due to their language structure, memorization is an easy task for them (at least they have an advantage memorizing huge amounts of info).
I got an MS degree at a, mmm say top 100 university and then move to a top 20 for my PhD. At the top 100 univ, chinese students led the class. Now, at the top 20 one, the smartest guy in my class is indian. I remember when we had our first midterm in electrodynamics, a chinese student got the highest grade and all the problems came straight out of Jackson (I realized this after the test) but during the second midterm when the problems were not from Jackson or any reference book used in the class, the same student got an significantly lower grade. This behavior happened frequently when the problems where from books. We later found out they have ALL the solutions to Jackson, Sakurai, Goldstein, etc.
Also, another chinese classmate once told me that because of the large population in China and the limited number of student visas the US gives to them, only the very top (in paper at least) are able to come to the US. Apparently a significant number of accepted students don't get the visa and therefore have to reapply or give up. This situation "promotes" cheating in order to get out of China.
In any case, they deserve some credit for being able to keep all that info in their heads.
From the standpoint of education, I gotta say that undergrad education in the US has a lot of areas to improve on. That said, the fact that foreign students are usually the top of their classes at top universities makes the comparison nonlinear.(I got my BS from the #2 univ in my country, and now I have classmates from universities ranked in the 60s, in my case it didn't work out since I am definitely not smarter or better than all of the americans, ha).
Also, getting a student visa is not so easy. In countries like China and India, some average students are rejected on the basis of deficient academic background. They also have an interview to determine the level of english they have and many of them get rejected by this. There is some filtering process and the best students are able to get through, for the most part.
Seems it's became a big argument about the international students. As a representative of those students, I'll put my comments. I'll be careful to be fair regardless of any personal political or emotional situation, but still the comment might be subject to my personal bias, so please make sure to suspect what I'll say.
First, regarding to the cheating by the chinese students, my advisor visited China this winter for the conference, and said that he found the evidence of the existance of the cheating syndicate (or the equivalence); probably from the personal conversation with some chinese professors. It is also famous among Japanese and Korean students that the ETS has given up giving GRE tests in Korea due to the suspects of the organized cheating by some company. Nowadays, all most Koreans have to come to Japan to take GRE tests. This year, physics GRE problems were on the internet before U.S. students took the exam on the scheduled time by those who took the test beforehand in their countries.
You may have heard about this, but the enthusiasm for the entrance exams in Asian countries are much more serious problem compared to the situation in the U.S. like I personally felt that the U.S. students are thinking nothing about the entrance exam even they're doing there best. This is the cultural difference. They're much more serious about there future scientific carrier because they don't have bright future plan if they stay in their countries.
Justifying my home country, Japan has good resources nowadays, and in fact, a significant fraction of talented students choose to stay in Japan, but still, some of them are attracted by the plenty of research opportunities and the positions at the worlds' best universities in the U.S.
This is no longer the problem of the science, but more like a political and fiscal problem. People do whatever things to make their dreams come true whether it's really scientifically motivated or not.
I have negative feeling to those who cheat for the exams cause they're not the ones needed for science nor suitable to do science. Those who sincerely doing their bests on the exams can be easily harmed by the systematic faults of the admission processes (i.e., lack of time for the truly beneficial and optimized selections of the candidates).
I don't dare to say myself as benefit to the science, but as being Asian, I'm put on the same criterion as those possibly cheating serious Asian students.
I've studied in the state university in the U.S. for 4 years and my GPA was 3.99 for both major and total. Got 900 on the physics GRE without cheating and have plenty of research experiences in Astronomy. If I were dealt as a domestic student, my future might have been deferent.
I'm really comfortable with my research environment now, but that kind of situation can easily happen regardless of the nationality.
I personally cannot trust current testing system, and hope the ETS pays more attention to those unwelcomed visitors to the world of science.
IP - talking about foreign admissions, is there any unwritten rule on foreign students who come to the US and initially attend a not so well ranked university (which is a lot more easy to get admission to), gets a Masters and then try to "trasfer" to a top school?
I was in that situation, although my initial attendance to a not so good university was "forced" by an exchange program between my country and the US. Anyways, even though my record is/was good, I had a hard time transfering. I only made it into the universities where 1-2 professors there knew me (not friends).
I have friends that have gotten rejections when attempting to transfer and they have good publications and grades. Does the home department also object (if yes, how much?) to students leaving to a better program?
I wonder about the comparison of Europe to America, or Europe to China?
IP - In my experience the physics GRE, as practiced today, is a lousy test, in the sense that it doesn't measure what you really want it to. It has sooo many questions that time is a hugely important factor, and those questions often seem to be a poor test of actual physics ability. I agree with you that takehome exams tend to give a much better sense of what people can actually do in realistic situations. On the physics GRE, I want to know if an applicant is in the top 25 percentile, or the bottom 25 percentile. Other for the people in the middle 50, I don't think that the physics GRE tells much. Contrary to Sylow's experience, I find that GPA, particularly in physics and math courses, usually is a reasonable indicator. Overall GPA also says something about the organizational skills of an applicant.
As far as taking foreign scores with a grain of salt, this is not necessarily based on prejudice or xenophobia. I can tell you of many examples of applicants from, e.g., China who have perfect physics GREs, perfect verbal GREs, can't pass the TOEFL (test of English as a foreign language). How can someone who can't read English well enough to pass the TOEFL get a perfect on the verbal GRE? This certainly looks very odd, and some of my Chinese colleagues have matter-of-factly told me not to trust scores in such cases.
Wow! After reading the comments - perhaps I should be more skeptical about foreign applications. I have taken GRE in stone age, before the internets made sharing information easier. But I should also note that I was always very impressed with the academic abilities of foreign students - of course those are the ones who were selected by the committee for one reason or another.
I cannot imagine what it's like to be a candidate from China (or India or Taiwan or Korea): nothing you do seems to be good enough - acing GRE doesn't count, letters are meaningless, GPAs are high across the board, and proof of research experience could be questioned too.
I came to the US as a grad student from India. In my experience, cheating on the GRE/TOEFL does not happen at least on a regular basis. The reason that Indians (and Chinese) do much better on these tests (especially the physics GRE) is because they take these tests much more seriously. Each of these tests cost ~ $100 and that is serious money for Indians and Chinese. My room and board at college in India cost about that much for a semester. So taking the test twice is not an option. Try making the price of the test $10,000 in the US and see what happens to the scores. Studying for a month or two for these tests is pretty standard for us poor people. I remember learning some words for the verbal GRE that I have never used since the test. For the physics GRE, we do practise doing problems quickly and trying out old tests etc. Those of you alleging cheating should be aware of these facts.
It is more relevant to ask whether a high score on the test means anything. The way I look at it is this - getting a high score is not a guarantee of success in research, but it shows that the student is dilligent enough to make the effort to get the score. We in India and China know that if we get a score of 700 on the Physics GRE then we are done for. So we make sure that we get the 990 or whatever. And I think that counts for something. Think back about all the students you knew who had very high scores on these tests. Sure, most of them did not do especially great on research, but I bet that a large majority of them went on to a good career in something. And thats all we can ask for.
Finally, there is no way to judge who is going to be good at research. All of us who are academic professionals know that good research requires a dash of intuition, an unwillingness to give up and a big dose of luck. Some make it, others don't.
Hey - There is no doubt that many many foreign students are extremely well prepared for graduate study in physics, and I'm sure the large majority of the high test scores are legit, particularly on the subject test. It's when there's a huge mismatch somewhere, particularly on language skills, that it raises eyebrows. As anonymous mentioned up-thread, the regular GRE scores have become much more in line with things like the TOEFL now that the exam has been restructured.
In all cases, prior research experience seems to be the best indicator of future success. We definitely sit up and take notice if an applicant has a publication in a known journal!
Doug, it is very easy to get perfect score in verbal GRE. Total number of words used in that exam (most of them are skesperean english words) are around 2500 or so and there is a booklet which contains all of them and chinese students sit and memorize the meaning of those words. It was shown to me by a chinese grad student a while ago. There is no cheating going on there. However, in TOEFL, you MUST know the language to pass it because they dont ask you equivalence questions (e.g. what is the equivalent of assault? answer: attack) thus a good command of the language as a whole is requires.
I personally do not understand this whole fuss about the english abilities of chinese students. I do not think the french have any better command of the english language in all honesty (I know lots of french and chinese people personally).
Besides, a chinese grad student I used to know of could write and speak english as well as I do. When he first came to USA, he could barely speak. Our advisor made him practice by asking him to make oral presentations every week. After so many years, he is in excellent shape now. However, you can very rarely improve someone's ability to do research. To me, it is like being a painter. You cannot teach someone to paint like Degas or Van Gogh after the age of 20. You either have it or not. The other things you guys are concerned about can all be fixed easily (like written and spoken language, knowledge of textbook material etc.) therefore if a student had coauthored a paper as an undergrad, I would accept him no matter where he comes from and what his test scores are.
There are have been some quantitative studies as to whether GRE scores are good predictors of future success on the PhD track. The major findings are that standardized tests (including the Physics GRE) do correlate well (correlation factor ~0.4) with graduate class grades but do not correlate as well (correlation factor <0.2) with success as defined by time in PhD, citation rates, or oral exam scores, etc. Additionally, these standardized tests are inherently biased against under-represented minorities and women with differentials of 150 points on average.
GRE scores are a good measure of certain attributes that may lead to success, they are not the entire story. And putting a hard floor requirements for a GRE physics test score would appear to bias against potentially successful applicants in the pool of women and under-represented minorities. Since the Physics has some of the lowest ratios of female graduating PhDs (<15%), this should be taken seriously.
Here are some references:
Ref 1
Ref 2
Ref 3
Ref 4
As an American student who just took the GREs I feel I should comment on my preparation. It consisted of doing two or three of the old tests all the way through under test like conditions and then going over the answers. (I got an 800 so I'm pulling the class average at grad school down) Most of my friends spent a similar amount of energy on the test. It is not surprising that students who spend, as the posters above say, months preparing. It isn't that I didn't care about the test, there were just more interesting things to do (a majority of me time was going in to a research project at this point).
In the end it seems to have worked for me as I am at my first choice grad school and quite happy.
Part of the quality issue, at least according to a professor I know whose research and position at our institution are so solid that he can speak his mind without loss of credibility, or really any other significant repercussion, is that unlike incoming American students, many Chinese students can hit the ground running because they have more research experience and training, in many cases this being a research-and-publish type master's in hand. The way this professor put it, given the immense pressures faculty face, you can't hold it against faculty who will choose the student who can get you material for publication during the first year, when most American students are busy with classes. These students have earned their place, so I respect them for their accomplishments.
But that doesn't change the facts that there is a trend of Chinese students cheating, whether intentional or by dint of not realizing that something was forbidden when in their experience it has always been acceptable. And not just at my institution, it's something that comes up whenever people interested in academic politics are gathered. It breeds distrust of Chinese students in general, and the way they Balkanize off away from the rest of the department amplifies that.
Ignoring the utility of standardized tests to measure suitability and aptitude for a moment, when so many of them get caught cheating (in the cases I personally know of or experienced, collaborating on pledged homework and getting busted stealing a labmate's homework solutions and using them to work on homework with a whole room of fellow Chinese students), it becomes particularly easy to think that maybe a lot of them cheated just to get here in the first place.
And it makes it rather worrisome to walk into an administrator's office and see a tally indicating that, of this year's offers, at least half were made to students of that nationality -- how many of them really deserve to be in an elite American institution when our own system is failing to prepare many Americans to compete for the same opportunities...
Frankly, reading the post and responses, I can't agree with either the premise that performance on these metrics is the only thing that matters, or with the premise that somehow doing well entitles you to have a significant part of the cost of your education paid for by the taxpayers of another country. As a taxpayer who is close enough to the system to see what's going on, I feel that there are things which must be considered other than excellent performance on a collection of physics trivia, such as the fact that picking good Chinese students means that those resources are not available to support American with equal potential but who have not had the fortune to have access to the same resources as those coming out of the Ivies and MIT-quality schools and the Cooper Union and such.
*Try testing for that idiom on the TOEFL.
My experience: I was a late comer to physics and had a very limited liberal arts physics education (by a couple of dedicated professors). Not only could I not compete with international students, I couldn't compete with American students who had a more complete B.S. in physics that started from their first day of college. I spent the year after undergrad doing research with a professor at a ~top 20 school and developing a relationship, and he more or less sponsored my way into grad school despite my dismal ~600 physics GRE score. He thought I was a good researcher and things went fine until I came up against the department's PhD written qualifying exam, which may well turn out to be the end of the physics road for me.
Or do we always subconsciously favor certain candidates (domestic vs. foreign, male vs. female, experimentalists vs. theorists
You'd never want to subconsciously favor by sex, or race, or anything like that, but I don't think "domestic vs. foreign" and "experimentalists vs. theorists" fit in this list. There are very different implications for accepting foreign students, and there's a disparity between supply and demand when it comes to theorists.
Have u try the Physics online discount bookstore
http://www.cocomartini.com/
I get all my textbooks for this semester from this bookstore. All are brand new and 60%off discount from normal price. See if any help.
hehe ^_^
I have more thoughts on all of this, tangentially related to admissions.
I will try to remember them and spill them out as a separate post.
For example - is the "taxpayers paying for foreign student" a good argument for favoring domestic students? In other words, let's assume that the foreign student IS actually more qualified by an objective metric, better trained, more hard-working/efficient and will be a better scientist in the long run. But because we use federal or state funds, are we supposed to heavily favor less qualified (hypothetical example!) domestic student?
If so, can this argument be extended for postdocs? Hiring of faculty?
Here's another question that always bothered me: suppose you break down performance as due to 1. natural talent and 2. hard work.
In accepting graduate students, making hiring at postdoc or faculty levels - or speaking more broadly, in how society rewards performance - should we favor talent over hard work?
For example, if a student is extremely talented but seems to be slacking off, yet still gets good grades, does well in his research - getting this student may be risky (too flaky for projects that require a lot of hard work) - but at the same time IF the student decides to put in hard work, in combination with natural talent, this would make them a research super-star.
OR, you could have a student who may struggle academically but is a real workhorse in the lab. Experimentalists may have seen plenty of examples of students who are "doers", rather than "thinkers". However, often times there's a hard limit to what those students can do in terms of overall research performance - especially if they lack the talent to have a broader view of the field, and some theoretical background that is often a necessary component of what we call "research instinct".
This nature vs. nurture debate is purely academic, since in most cases you have no idea how any specific student will perform, until they spend a couple of years in the lab. But it's an interesting question nonetheless. Should we reward talent or hard work?
IP, Taxpayer's money and domestic/foreign students: The German Max-Planck society used to argue that way and pay German students and postdocs better. And I think you would find a lot of Europeans agreeing with the idea. (Not me, though. I'd rather say that the taxpayer's money should buy the best researcher the taxpayer can get.) Anyway, the European Parliament and Commission came after them a few years ago, and they officially abolished the practice. Or at lest abolished the official practice.
The talent/work question buggers me, too. I come from a culture where we would traditionally praise brilliant failure over plodding success, which leads to a certain could-have-would-have mentality that is not too healthy. I think it comes basically down to the question of promise/achievement: You might favour a brilliant but lazy student, because you hope that (s)he will wake up, start working and be successful. But you wouldn't tenure people because they might have done great things if they'd bothered.
Schlupp - I am in full agreement with you with respect to "get the best researchers". I just noticed that Max Planck is under attack for taking away top people from universities, I wonder how that plays out in the near future.
As to talent vs. hard work - by the time of tenure decisions, and even tenure-track hirings, it's a combination of both talent and hard work. In admissions for graduate students you may get many cases where it's one or the other. Some students may not have had much of a chance to demonstrate effectiveness in research (I would argue that summer REU programs are severely limited because of their short-term nature). But they may excel academically - so they have talent and potential to succeed, but their hard work is a question mark.
Then you may have students who get so-so GPA and so-so test results but have slaved away in the lab, with stellar letter of reference from PIs.
So I though I would weight in on the GRE physics marks issue... Where I come from you have to write a 4 hour multiple choice comprehensive test to get into university for undergrad. You study for a 1-1.5 years for 10 hours a day and you ace it. You train and you really don't learn anything in depth. You learn how to answer MC questions fast. In most other countries PGRE is treated the same way and since most domestic students don't train as hard and the test does not really measure any deep knowledge of any topic then it is not really a good, objective measure
I went to physics grad school with a lot of Chinese, Indian, Russian students. I was a US undergrad. The Russians were miles ahead of us in terms of what they'd learned as "undergrads," but everyone expected that. At least for the Chinese students, we didn't think that they got high scores on the physics GRE by cheating - rather, we thought they had a special class to prepare for it (I'm fairly sure that my fellow Chinese students told me that). Cheating was more of an issue with solutions manuals for the problems in Jackson; these are widely available outside the US.
Of course, American undergrads have neither the background nor the time to take special classes; they have to do stuff like take history classes and go to parties on Saturday (both of which they _should_ be doing, anyway). I improved my own GRE score by 90 points two years later (I took time off after college) with more time to study.
The problem with test scores is that if you hand physicists a number they will use it beyond the realm of its applicability. I might believe that a 300-point GRE difference tells you something about physics knowledge, but I believe that a 100-point difference is in the noise. Furthermore, some people are just better at taking tests. If you made an entire class of high GRE scorers, you'll get a lot of students who want to do theory, and you'll weed out promising experimentalists. This is especially a problem since experimenters generally need and can pay for more students.
Finally, in the long run, it's hard to evaluate this with my class since so many people leave the field. But while we were there and now 10 years after, the Americans with their lower test scores did as well or better at doing research and becoming productive members of the discipline. This is not a slap at the foreign students, many of whom succeed in the face of language and culture barriers. It's just that the GRE and problem sets test the ability to regurgitate knowledge, while being a good researcher requires a set of independent motivation and communication skills that the test does not measure.
I am not too sure about Americans, but where I'm from (W Europe) my profs always told us never to remember anything that's not worth remembering. Our E&M prof was proud of the fact that he didn't know Maxwell's eqns by heart. "If I need them I just derive them quickly". "physics is all about insight" and "you're not studying medicine".
Sure, some phys GRE questions can be solved by a quick dimensional analysis, but many, many equations have to be learnt by heart, especially if you want to finish them within the given time span.
Also, sort of a self-fulfilling prophecy is going on here. Statistics and blog posts like this leave a lot of domestic students to think they do not have to study that hard, since 'everybody knows' the curve is set by Asians anyhow and therefore their scores are interpreted differently. Asians, on the other hand, know that they'd better take 3 months off just to study the test (some of them do this) otherwise they don't have any chance to ever leave their country (generalising a bit here...). More speculating now; couldn't it be that education in Asia is far more aimed at learning by heart? For sure, this will help enormously on the phys GRE.
Personally, I found studying actual physics for this test is completely useless, you really have to study the test. Knowing and understanding all the relevant physics behind a double slit experiment will get you the right answer to the question, but takes too much time. Remembering the one important formula, without a remote understanding of the physics involved will let you answer the question immediately.
But yes, it measures something. People with a high IQ will likely score higher, because they are better able to understand what is needed for the test and will remember more of those equations and how ETS wants you to put them to use. People with a high IQ will on average have done better at university as well, so there is a correlation.
A high score also shows that you are willing to spend a lot of time and effort on something irrelevant, which might just be the characteristic universities are looking for in a grad student (just kidding, but you get the point I guess)
A high score also shows that you are willing to spend a lot of time and effort on something irrelevant, which might just be the characteristic universities are looking for in a grad student.
(Chiming in slightly late here) Have you heard of the book "Disciplined Minds" by Jeff Schmidt? One of the point he made in the book is exactly that. Not kidding at all.
Mind one thing that international student from non east Asian countries (like South Asia, South America, Eastern Europe) are not considered to be taking unfair advantages at test. However, a few days back I was speaking with my undergrad adviser who also is in the grad admissions committee, and he explained to me that if they notice that an international applicant took the Physics GRE after finishing his Masters / taking years off / after doing a job, generally they assume that this guy spect years acing the test etc. Also, Chinese applicants now have a near legendary reputation (!) for their general lack of knowledge about what the heck is "doing research", specially experimental work. Neither of these plays fairly with them.
In fact, our physics graduate program (I am in astro) has two sub-committees: one for Chinese students and one for everyone else, domestic or international. The points are that half the applicants are from China anyway, and they have unquantifiable qualities (potential to live in lab for 7 years to do PhD etc.) and weaknesses (research experience etc.). At the end, what happens is that a top few Chinese apps are accepted straight out, and the rest after April 15 in case th department is running short somehow on a given year.
As a last point: a few schools indeed have funding issues. The Univ of California Astro programs accepts roughly 1 intl student every two years, heaven knows why else than funding, whereas a typical astro program (public or private) have between 1/3 to 1/2 accepts from non-US citizens / perm residents (including US school graduates who are not citizens etc.). I never got to understand though why Illinois or UT or OSU may have enough money for intl students but not the better reputed UC system schools.
Have u try the PHYSICS online bookstore Cocomartini
http://www.cocomartini.com/
I get all my textbooks for this semester from this bookstore. All are brand new textbooks and half price discount textbooks.
Good luck and wish some help.
hehe ^_^
Information about GRE (Newbie)
http://career-assessments.blogspot.com/2008/01/graduate-record-examination-gre-newbies.html
GRE Aspirants
http://career-assessments.blogspot.com/2008/01/graduate-record-examination-gre.html
I have an undergrad degree from a Canadian institution and did my PhD work at a fairly "elite" private American university.
1) "Elite" private American universities baby their undergrad students. Even their physics students. I've taught a number of standard courses (Griffiths EM, undergrad quantum etc). Not very much is expected from the students on the exams. An understanding that I would qualify as merely "adequate" is enough to get most or all of the problems right.
2) Bigger public universities are much harder on their students. To get "A"s the kids seem to have to learn more, so there is much more incentive for learning at a high level.
3) This second approach is much more similar to what I experienced in undegrad (large public university in Canada). I walked out with a 3.2 GPA, but I had a better education than all but one of the non-Asian people in my entering graduate class (so a comparison group of 20 or so). I also had the highest GRE score of any non-Asian student (910). The Asian students' scores are unknown to me, as is the quality of their undergrad education. There is often not much interaction between Asians and non-Asians.
4) There are whispered rumours that Asian GRE scores cannot be trusted and are the result of widespread cheating. I am loathe to make a judgement on this without further evidence.
5) Despite their superior "on-paper" academic status, Asian graduate students take longer to graduate and go on to inferior post-doctoral careers. This may be a language barrier or it may be something else. I'm not sure.
I came across this blog while searching for the phrase "perfect GRE score odds." I took the general GRE test a couple weeks ago because, as an unmarried 34 year old woman I think I'd better start making some wiser career choices if I want to survive into old age, and it seems like Grad School may be a good way to make up for twidling my thumbs since I graduated from undergrad.
I just got the official score report yesterday:
V: 800
M: 780
A: 6.0
Now, the reason I'm researching the phrase "perfect GRE score odds" is because I want to know how uncommon it is to get near perfect on the whole exam, and whether I can leverage this in some way to get into a great school, despite the fact that it took me 5 years to get my undergrad degree finished and I ended up with approx 2.5 GPA. It was dismal. There were a couple semesters where I got Ds and Fs, and a lot where I got Cs and C-s. What's worse is that I haven't really done anything particularly impressive with my adult life, either. Does anyone here know of anyone like me that got into a good school? Will a school look fondly on me based exclusively on this one achievement? Are there certain types of programs that will value this score more than others? Anyone affiliated with any grad programs that can give me some insight? I just want to know whether this score can actually be leverage to help me get my life turned around in some way, or whether it's really just a hurdle to have gotten past and I should still just go to the local MBA program.
nice post
Thats a very interesting post. I have been inspired. Thanks. Web Designer
Post a Comment