Saturday, January 31, 2009

$10 billion lost

The economic times has an article, titled "Students' exodus costs India forex outflow of $10 billion." The article starts off by informing that 500,000 students choose to go abroad every year for higher studies. Assocham president Sajjan Jindal says,

The primary reason why large number of Indian students are forced to opt for foreign universities is that Indian institutions have high capacity constraints. This trend can be reversed by opening a series of quality institutions with public private partnership by completely deregulating higher education.

Also vocational education percentage in India is at meagre 5 percent of its total employed workforce of 459.10 million as against 95 percent of South Korea, 80 percent of Japan and 70 percent of Germany. China has over 500,000 vocational schools while India has less than 3,000 such institutions. This makes a strong case for India to allocate a substantial percentage of its budgetary allocations to promote vocational education to make the country a manufacturing hub.
I have no idea by what it means. Vocational training is primarily imparted in ITIs. Is Assocham wanting to open more of them? Fine.

However, opening a series of quality institutions for higher education is not so easy. Most of the private schools (BITS is an exception) have not been very successful. If IITs are facing a problem in getting good faculty, then the problem will even more acute for a private institution. Further, I think the numbers are hugely inflated. Several students go abroad for undergraduate and graduate education abroad even if they get admitted in IITs. This is because the career opportunities are better. To claim that the exodus of students will stop if we have more colleges/ institutes / universities is foolhardy.

Friday, January 30, 2009

Life in academia

The chronicle has an interesting article titled "Why are more and more graduate students turning away from careers at research universities?"The article starts off with the remark that faculty who graduate students hear a lot these days (according to the article) "I don't want to live your life." I do not have such a problem because I do not have a life :-)

Academia in research institutions was never meant to be a 9-to-5:30 job. While a day in the life of a graduate student can be grueling (or look like grueling), the numbers of hours one puts in during this period is not significantly higher than someone who practices cricket every day to join the Indian cricket team.

Intelligence alone is often insufficient to succeed in any job. In the book, "Greatness: Who makes History and Why?", Simonton quoting the famous Anna Pavlova says "Success depends in a very large measure upon individual initiative and exertion, and can not be achieved except by the dint of hard work." He further says,

Only clock watchers and time card punchers put in just 40 hour weeks. Creative output in the science correlates positively with the number of hours devoted to scientific research. One study found that distinguished researchers in the physical and social sciences worked 60-70 hours per week for virtually the entire year..and all exhibited a driving absorption in their work.....To convince the skeptical sloths, let us move to a higher plane of evidence. In lieu of specific cases, suppose we examine group statistics. Wayne Dennis studied contributors to disciplines as diverse as music, linguistics,...and chemistry. He found that the top 10% who were the most prolific in each field could be credited with about 50% of all contributions to the field. In contrast, those among the 50% who were at the least productive were responsible for only about 15% of all the work. There is even a law that expresses how much the elite monopolizes a creative endeavor. According to the Price law, if k is the number of persons active in a discipline, then the square of this number approximates the size of that subset who produced half of the contributions.

A senior colleague of mine has done an interesting study. He has taken the 7th standard marks of all the professors working in the university and tried to find a correlation. The correlation coefficient was less than 0.1, indicating that performance at the school is not necessarily a good indicator of whether the person will become a professor in an academic university.

The article seems to imply that professors liking their jobs and working hard has resulted in students turning away from academics. A youngster can not complain that he has to practice for ten hours in the sun to make it to the cricket team. Sachin Tendulkar still practices as hard as anyone else in the cricket team. A youngster should instead marvel at Sachin working so hard (even after achieving so much) rather than complain ! Further, after completing graduate school, one need not work in a research university. One can work in teaching oriented universities, in industrial labs and so on.

In some instances, the requirement of working hard may also dissuade non-serious science "enthusiasts" Recently, in IISc, we admitted two graduate students for the master's program who had great ideas on winning the Nobel prizes during their course program. When it was pointed out to them rewriting the second law of thermodynamics will require them to first understand thermodynamics thoroughly (taking a course, for example) and then working hard, they left the program. I do not think, however, this was bad.

However, all this is moot. Science is supposed to be entertaining and engaging. In IISc, for example, you can find faculty working for 30 hours or 100 hours a week and get paid the same. Many of us work for long hours because one finds pleasure in doing this work rather than anything else. If graduate students are turned off because professors enjoy their jobs so much that they work hard, so be it.

Wednesday, January 21, 2009

IISc college

In an interesting article on IISER, Shobo Bhattacharya makes very valid points, most of which I agree with. However, when the discussion on IISc comes up, here is what he has to say,

Aware of this conundrum, the Planning Commission of India recommended recently that eminent research institutes like the Tata Institute of Fundamental Research (TIFR) and the Indian Institute of Science (IISc) embark upon undergraduate education by establishing, for example, a TIFR-College and an IISc-College. The idea has had little traction with institutes however and has produced additional skepticism. Some of the researchers are concerned that teaching duties will seriously impair their ability to conduct world-class research, despite overwhelming evidence to the contrary.

The overwhelming evidence that teaching aids research is based on university systems in USA and elsewhere. It is a forgotten fact, however, that the tutorials and grading of assignments for the undergraduate courses in US are exclusively done by graduate students. The professors are hardly involved in these activities and this saves considerable time that can be devoted for research.

People who are favor of IISc college extensively argue that this will result in better graduate students doing doctorates in the country. Institutes like IITs, where faculty are involved both in undergraduate teaching as well as research, find that practically none of their undergraduate students go on to do a doctorate in India. Therefore, if IISc starts an IISc college dedicated for undergraduate teaching (which I think it will) and all the active research faculty teach in this college, how many of these students will go on to pursue a doctorate in India? My guess is less than 5%. Therefore, the input quality of graduate students to IISc will largely remain the same even if IISc and TIFR open colleges.

Sunday, January 18, 2009

IISER, NISER

In his blog, my colleague talks about setting up IISERs and NISERs. He mentions that the previous government mentioned that IISERs were supposed to be setup in Allahabad, Bhubaneswar, Chennai and Kolkata. The current government overruled it and said it will be set up in Kolkata, Chennai, Mohali, Bhopal and Thiruvananthapuram while a NISER would be setup in Bhubaneswar. Are the IISERs and NISER different? Obviously not, except that NISER is funded by DAE. Except for IISER-Chennai, all the others have been setup and their directors have been appointed. The following link even compares the academic credentials of the directors. Thus, the only city that was dropped was Allahabad and it becomes obvious why if you consider where the previous government's HRD minister got his doctorate from. People in the "know" seem to explain why IISER Chennai has not been setup, yet. The point of Abi's blog, however,is that the policies of the government has been changed by its successor is well taken but this is not surprising at all.

Regarding the question whether IISERs/NISER are successful, it depends on the criteria you set. I do not know how one evaluates an academic institution in India. I do know, however, that atleast two of the above IISERs have received more than thousand applications each for their faculty positions in the four science subjects. A large number of students write the entrance exam and are interviewed. Their faculty have got funding from DST etc and have started publishing papers. However, infrastructure problems remain but are hopefully, teething. Even an established institution like IISc, which got 100 crores in 2007, will complete all its buildings only in 2009. Anyone who works in administration will tell you that getting hundred crores from the government is much easier than spending it following all the myraid (and, maybe, archaic) rules.

However, I do think that the success of an institute is judged mainly by its students. If world-class motivated students join and their motivation is sustained, this will result in an excellent institution. Despite flaws, exams like JEE/CAT/GATE that filter out 95% of the candidates result in world class students. However, the motivation in them has to be sustained. [The fact that B.Tech students who do not get any jobs or admitted for higher studies teach other B.Tech students in many private engineering colleges demotivates most of the incoming undergraduate students.] The incoming students will remain motivated only if the faculty are world-class and motivated.

With the current scarcity of good faculty, a rapid recruitment of faculty just to fill up the position will result in a bleak future for these institutions. Even if world-class faculty are recruited by IISERs, these faculty will remain motivated if and only if it is augmented by excellent infrastructure like lab facilities , efficient administration that is not required to follow outdated arcane rules, merit based evaluations for all promotions/ awards, ability to expand without the mentality that small is beautiful, unfettered research grants, emphasis on excellence rather than social networks and, finally, money. However, granting only the last (money) makes that the required condition for excellence is met but sufficient conditions for creating and maintaining excellence (in academics) are not met.

Thursday, January 15, 2009

Research Double

One of my former students recently wrote to me and congratulated me on the "double" of 200 papers and 2000 citations. The term double comes from cricket where 100 wickets and 1000 runs is the start for a good all rounder.

In research, however, citations are of more importance than the number of papers. I know of faculty in USA who have published less than 50 papers but have more than 2000 citations.For example, the editor of the journal (in which I regularly publish) has less than 75 papers but she has more than 4000 citations. However, there are no such exceptional cases in India. In India, the number of papers and citations seem to have a direct correlation. Some statistics based on the "double" for IISc (includes retired faculty),

Number of faculty who have > 1000/10000: 1 (Prof. CNR Rao)
Number of faculty who have > 400/4000: 2 (Prof. P.Balaram)
Number of faculty who have > 300/3000: 5 (0 in engineering)
Number of faculty who have > 200/2000: 26 (2 in engineering)
Number of faculty who have > 100/1000: 74 (7 in engineering)
In cricket, there is a treble: 100 wickets, 1000 runs and 100 catches. In research, we may have 200 papers, 2000 citations and h-index of 20. The above numbers do not change for the treble.

Okay, how this compare internationally? Let us take Purdue university. How many faculty of Purdue have > 300/3000? The answer is 300.

Tuesday, January 13, 2009

Faculty Pay based on teaching

My colleague provides some interesting links about better and effective teaching. Taking this to an extreme, my alma mater, Texas A&M University, has recently instituted a system where a bonus pay of $10,000 per year will be given based on teaching evaluations by students. This seems to be a blatant approach of consumerism to higher education. While student evaluations of faculty are important, it is well known that "good" student evaluations can always be obtained if the faculty gives out easy grades and does not give out hard assignments. A major study by Ohio State University in 2007 found absolutely no correlation between student evaluations and actual learning.

But Aleamoni said that even if his research suggests that some student evaluations — designed in ways that differ from the Texas A&M approach — can be reliable, he has always stressed that these evaluations should never be the sole basis for a decision about the quality of someone’s teaching. “Students are only in a position to judge performance in the classroom,” Aleamoni said.

Any real evaluation of teaching, he said, must include peer analysis of such issues as, “How well was the course designed? Are the materials current and up to date? Have they set up the right kinds of standards for the students?” And students aren’t in a position to judge these things, he added.


Of course, one does not have to worry about this in IISc, where teaching is minimal. At one time I used to teach three main courses and now I have decayed to teach three students in one course.

However, this is not dissimilar to the schemes instituted by the Government of India wherein one gets paid various amounts of money (sometimes, concurrently) if faculty get fellowships like the Swarnajayanthi, JC Bose etc., or awards like the Bhatnagar or are elected as Fellows of Academies. As Professor Balaram notes in his recent editorial, most of these are given based on promise and patronage rather than performance.

Wouldn't it be better if the rewards are given without personal money but with huge research or teaching grants? This would result in the faculty actually doing science rather than writing proposals to various funding agencies or lobbying to get themselves nominated for these fellowships.

Monday, January 5, 2009

Excellence in Research

Prof. C. N. R. Rao, in his inspirational talk today titled "Doing Science in India: How to Succeed!" gave four main tips for succeeding in doing research in India. They are picking the right problem, working hard, perseverance and publishing the work. He also mentioned the need to think big by graduating more doctorates (he mentioned that all IITs together graduated only a total of 598 doctorates in 2007) and publishing more papers.

Sunday, January 4, 2009

Humanities Department in IISc

Recently, the student council petitioned that IISc should start a department of humanities and there was subsequently a meeting to discuss this. With the support of the Deans ("best courses I have taken are in these areas"), departmental chairmen ("best idea I have heard in IISc") and the alumni, it may become a reality.

While humanities may be very useful as part of an undergraduate education, I am not sure what major role it will play in the education of a doctoral student in science and engineering unless the department is at a world class level in terms of research. A humanities department with one faculty each in history, english, economics and sociology will not become world class due to the lack of a critical mass. Instead, it may be better to open a department with focus on subjects like anthropology or archeology and recruit ten world class faculty in this area. Of course, getting faculty will be a problem even though many people believe that the very mention of IISc opening such a department will make the faculty working at other institutions to leave their institution and join IISc.

Instead of becoming all-rounders, shouldn't IISc focus on doing better on something that we already do reasonably well ? Regarding the comment that the IITs have humanities department, IISc should not be compared with IITs but with IIMs, which also offer only post-graduate education. Maybe IISc should open a humanities department as soon as IIM-Ahmedabad opens a chemistry department !

Saturday, January 3, 2009

Grad School to Professor

Recently, one of the students met me on the road and asked me to compile a list of links that may be useful to graduate students. 

If you are already in IISc, then the essay on whether you should attend graduate school  is probably not useful to you. After entering graduate school, the first major priority is that of choosing an appropriate research advisor

While in grad school, the most important aspect is to write research papers. Professor Whitesides, who has the highest h-index among all living scientists, says "If your research does not generate papers, it might just as well not been done. Interesting and unpublished is equivalent to non-existent" Professor CNR Rao, who has the highest h-index among all scientists in India, gave me sagely advice, "Do not be afraid to publish" 

Whitesides has written a wonderful article in Advanced Materials on how to publish a research paper. Whitesides says that one should start writing a paper at the same time one starts a research project. In fact, this was the same advice my postdoc advisor gave me and he used to say that by the time the experimental work and analysis was over, the paper should also be nearly ready. 

Recently, Prof. Kamat from University of Notre Dame, who is an editor of Journal of Physical Chemistry, recently visited IISc and gave an excellent talk on writing research papers and the use of figures in the papers and ethics in research. There is also an excellent guide on how to publish in top journals. In addition to writing papers, giving good talks to a wide variety of audience is essential in graduate school. 

For people like me whose first language is not English, books on writing are very useful. My postdoc advisor gave me an excellent book - The Elements of Style - by Strunk and White and it is perhaps the best known book on writing well. It is available online for free now. The best book on improving English that I have read is "On Writing Well". It is a great book that deals with grammar and other semantics in an entertaining and humorous way. However, if your English is really poor, then the above books may be too heavy and, unlike other people who recommend Wren and Martin's classic book, I recommend the book - Practical English usage - by Swan. The 2006 edition costs Rs. 260 and is available locally in IISc.

After you finish graduate school, you should be successful as a post doc.  At this point, you will start receiving requests for reviewing manuscripts. I follow these tips seriously and also try to limit myself to review only three times the number of papers I publish.

Finally, you may land up with a faculty position. Becoming a successful faculty member at a research university is not easy. The book—The Effective, Efficient Professor, by a famous chemical engineering professor Phillip Wankat—is a treasure trove of information on the strategies and techniques that would make a faculty member life successful and is a must-read for all new faculty.

Friday, January 2, 2009

Removing GATE

A recent article in NY Times discusses the trouble with using SAT as the sole criterion for admission. This is actually one of the followups from several articles published on this issue in the Chronicle of Higher Education. This is based on the report by the National Association for College Admission Counseling. which discusses that tests like the SAT and ACT were never meant to be viewed in isolation but considered as one of the factors that include grades, essays and so on. But with the college ranking systems considering SAT as one of the main criteria for ranking, the colleges seem to have preferred the exam as the one of the major criteria of admission until recently. Now many colleges in USA have made SAT optional, though the major institutes like MIT and Harvard still require SAT.

This backdrop is interesting considering that IISc is planning to drop GATE as one of the requirements for admission to its research degrees (masters and doctoral programs) in engineering. Under this proposal, the main criteria is the percentage of marks obtained by the candidate in the B.E/B.Tech examination and anyone who secures more than 70% marks in their undergraduation will be eligible for admission.

My opinion on the above has always been that we should not obsess with admissions tests like GATE for doctoral programs and we should take an expansive view of merit that would include GATE, the undergraduate scores, communication skills and motivation. However, for masters program, we need an all-India entrance exam just to screen the huge numbers. With nearly six undergraduates in engineering (and 2.5 lakhs of them expressing an interest in higher studies by writing GATE) with wide variations in the undergraduate marks awarded in each university, it would be nearly impossible to screen them only by interviews.