Today is the new Hijri year 1443. I would like to discuss a topic that has troubled me for many years, namely research tradition and publications.
The week before last, I attended two talks on publishing. I normally treat these talks as purely motivational rather than content-driven as most of the information is available somewhere on the web (in fact there are guide books available) if one cares to find them. As such the talks are probably suited for those who are at the beginning of academic career. My worry is actually more on the values that these talks may carry (do note whatever I will say below equally applies to my own self). Good publications come from good science and not the other way round. Many times in the past, I have mentioned about Goodhart's law, which states that "when a measure becomes a target, it ceases to be a good measure". This is because that once we are given a numerical target (considered as measure of quality), people tend to game the system to achieve such numbers without direct reference to the intended quality itself. Thus, their meanings get lost. For measure of scientific productivity, the number of publications has been always the easy target irrespective of what their contents are and hence manipulable. Even the number of citations (and hence h-index) can be manipulated. To correct this, one could eliminate self-citations (see Schreiber) or complement it with s-index which is an index of self-citations (see Kacem, Flatt & Meyer) but I would go further. Eliminate as well citations from our co-authors as well and that will lead to a much reliable number whose citations come from other authors (not within our influence). Removing self-citations can in fact reduce the index from 13 to 46% (see Schreiber). The only problem with this is that the index calculation is actually time-consuming.
Now, no accusations of manipulations or specific claims are being made here but merely raising the complications that can arise. It is also not in my nature (or anyone reasonable) to go around and finding faults of others nor do I want to make excuses for my own shortcomings. Naturally, every academic has his or her own yearning and ideals for achieving scientific progress but without delving into details, gross injustice can occur and may be dentrimental to scientific progress itself. One can't deny that administrators/managers tend to take the easy option of making blanket KPIs across different disciplines and a researcher in a discipline with slower paper production rate will be at a disadvantage. During a Q&A session with the speaker, I raised the concern for local theoretical physics researchers who are not able to be on par with researchers in some fields who can produce around twenty papers a year (about two papers a month). Personally, I think this is mind boggling; research ideas often take time to develop, mature and worked out completely (and some might not even get published). Having too many publications a year may also raise suspicions of not adhering to the Vancouver protocol of authorship. Retractionwatch has also highlighted such problems of prolific researchers (see also here). Being prolific is not the concern here but it is the variance across different fields and subfields (see e.g. here). One reaction I get during the Q&A is that someone knew a theorist who used experimentalist data of others to produce 70 papers (unsure whether this is for a year or spanned over several years). I know not of the name mentioned in this response and certainly the said author is not within the subfield I am interested in. What then would be a sensible rate of publication for theoretical physics (again this depends on the subfields)? I could not find a study that actually shows a number for theoretical physics but if one identifies theoretical physics as part of mathematics, one could use this statement made by American Mathematical Society: average of two or fewer publications per year. In any case, I dislike the idea of making scientific publications as a kind of sports.
So how does one proceed given it is highly unlikely that there will be any change of policy in the near future? The speaker replied is to do more networking and do (genuine) collaborative work, which is sensible. The explosive growth of knowledge in every field requires researchers to work on harder scientific problems as time goes by and given the limited capacity of an individual researcher, it requires researchers to work together in working out solutions. This is reflected in the study made by Grossmann for mathematical research: an increase of collaborating authors from 28% in 1940s to 75% in 1990s. My own simplistic approach within our group into this expansion problem is to have reading groups and have members/students read different articles and have discussions. The bigger the group, the better (the reason why our group take up more students over the years). Whatever forthcoming collaborations there would be, will be natural if everyone keeps some overlapping interests. There was this suggestion that we should look into collaborating with experimentalists, probably look into some form of modeling work. If this can really be done, it would be good but it is often harder to realise. My own early experience after coming back from my PhD was indeed to consider theoretical condensed matter research (that led me to work on punctured surfaces). I essentially bought these two books of Fradkin and Goldenfeld in Seoul to get me going (see pic below) for which I started a reading group to study these books, but after a few attempts, people lost interest.
Perhaps the two books chosen above was not suited for experimentalists. I made another attempt to work with experimentalists when I first joined ITMA. I bought one of my most expensive books ever, namely that of Dagotto on CMR (after the two-volume elephant book), to work on theories of colossal magnetoresistance. I even had a student to work on related aspects of Jahn-Teller effect with some diagonalization problems. He later gave up, and that too, failed. Thereafter, I focus what I know best - mainly theories of more mathematical nature. Perhaps the other instance of collaboration with experimentalists, was to set up a quantum information lab in ITMA and then in the Physics Department more recently. The former got shelved due to the institute restructuring. The latter was also not successful as the intended funds (two different sources) was not forthcoming. So now, whatever idea of collaborating with the experimentalists will be on the shoulders of my junior colleagues, as I will be retiring. Even if I decided to continue work elsewhere after retirement, I will stick to my own theoretical interest.
There was another remark made by the speaker that I responded in the Q&A. This is about why did we join the university and I replied indeed why should we join the university rather coyly. I think we have our own reasons (besides 'makan gaji' - the lowest level of motivation) and ambitions when we joined the university. I for one, wanted to make a tradition of theoretical physics in UPM. There were others before me, of course, who did theoretical physics at UPM. Our first head of department, Dr. Osman Ese was a theoretical condensed matter physicist but he passed away due to stroke a few years after. One can see his publication here. Prof. Karen Badri made the remark that if he was alive, he would probably made the path for theoretical physics easier. The other was Dr. Zainul Abidin Hassan who works on the Hubbard model (see here). He opted for early retirement to focus on his business. Prof. Mohd Yusof Sulaiman (nuclear physics experimentalist and solar physics) also dabbled with theoretical nuclear physics with Prof. M.A.K. Lodhi when he was around. Right after me, was Dr. Chan Kar Tim, formerly working on cusp forms and now on density functional theory and complex networks, and also Dr. Nurisya Mohd Shah on mathematical physics. Next in line will be Umair Abdul Halim who is working on geometric quantum mechanics. I do hope they will carry on with strengthening the theoretical physics tradition in UPM. In fact one should establish more theoretical work in these pandemic times as people are finding hard to get back to their labs. Perhaps rather than the theorists looking over into the work of experimentalists as suggested in the Q&A, maybe the experimentalists should consider the theories in their respective fields to work on. Perhaps then, the collaboration will be easier.
May Allah guide us in what we do in the future.
No comments:
Post a Comment