Shortage of Scientific Research in Turkey

Shortage of Scientific Research in Turkey

Dr. Esen Ercan Alp
Argonne Distinguished Fellow, Argonne National Laboratory
and SESAME Scientific Advisory Committee Chairman

We know very well that there are first-class scientists in Turkey in almost every field. Some academics are not only capable of following the most recent developments, but also able to contribute with original research. We come to know these individuals from time to time. Well, if that’s the case, why do we have a significant shortage of world-class research output? We will try to answer this question from a very limited lens of high impact publications. In particular, we will seek answers to the following questions:

1) Why Turkey’s share in high-level scientific research papers so low?
2) Where are we in terms of other universities worldwide?
3) Are there research organizations or private companies doing high-level research?
4) What can and should we do?

Let’s be very clear about one thing: The evaluation of the scientific output of Turkey should be best done by research funding organizations like TUBITAK, in coordination with TUBA (Turkish Academy of Sciences), YÖK (Higher Education Authority), and Science Academy Society of Turkey. Indeed, TUBITAK shows some interest and sensitivity in recent years regarding this topic, yet we don’t have a reliable and transparent index or metric that we can look up to measure Turkey’s contribution to universal scientific research. TUBITAK has instituted an incentive program for high-level scientific research (UBYT). I am a firm believer in incentive programs. I hope the scientific success metric becomes part of this incentive program.

1) Why Turkey’s share in scientific research papers in high impact journals is so low?

Turkey is in the top 20 nations (somewhere between 17-19) in terms of the size of its economy. It always has been. Yet, Turkey is 39th when it comes to scientific research output, as measured by “Nature Index, 2020”. Figure 1 shows, on a country basis, the shares in international research. We can immediately make the following observations: 16 countries that are in G20 are also in the top 20 when it comes to high-level research. The four exceptions are Mexico, Indonesia, Saudi Arabia, and Turkey. For Turkey to be in the top 20, it has to pass Denmark, a country with a population of 6 million. To achieve that, Turkey must increase its publication count from 65 to 387. This factor of six is impossible to overcome within the current system. It is especially noteworthy to mention that Denmark puts 3 % of its GDP on R&D, while Turkey hovers around 1 % in the last decade. In other words, the gap is increasing, not decreasing.

We have reached this point at the end of 97 years since the declaration of the Turkish Republic. So the blame goes to the current government as much as to its predecessors. We should keep our faith in the power of science; we should let the scientists work in an environment free of political pressures, and adopt a more modern approach to organize the entire scientific research enterprise.

In 2000, as we entered the third millennium, Turkey was able to set aside 0.46 % of its GDP to R&D effort. This includes the government and private sectors combined. Twenty years later, this number is still under 1 %. The current disputable number is 0,96 % but includes VAT and Customs Tax for imported research items. In other words, the average increase in the R&D share is limited to a mere 3.5 % per year for the last 20 years. The implication is clear; if you aim to double the research output, one would need 20 more years. Recall that the difference between Denmark and Turkey was six-fold, not two-fold.

These days, there is always talk of “2023 goals”, which seems to be an empty political agenda, rather than a concrete plan. I say this because quite often, I hear from the ministers that by 2023, Turkey will allocate 3 % of its GDP to R&D, which is close to the OECD average. Unless one increases the R&D budget by 44 % every year for three years in a row, 2023 goals go up in the air into oblivion. Well, a 12-fold increase is not realistic to expect or achieve. With my meager effort, I can see that Turkey will not reach a decent level of R&D activities even by 2050.

Figure 1: The top 40 countries' research output in Nature Index journals. Turkey is at 39th position with 65 papers per year, while Denmark, with a population 15 times smaller than Turkey, has 387 papers. Compared to Germany, with a similar population has 4546 publications, 70 times more than Turkey.

It may be a good idea to explain what we mean by High Impact Journal article.  As a major science publisher, the Nature Journal keeps track of all publications in the top 82 journals in the World. The list is given here ( https://www.natureindex.com/faq#journals).
While the articles published here comprise only 5 % of all scientific articles in refereed journals, the number of citations reached 30 %. In other words, these articles are six times more likely to be quoted or cited in future scientific work. So, many organizations or countries make the distinction between journal articles and high impact journal articles. Let be clear, a publication in these journals does not necessarily mean that they are more valid or reliable than the other scientific work. However, the editing and refereeing process is more robust and stringent. So, in general, there is a tendency among researchers to publish their work first in one of these journals.

To make this point clear, I would like to point out three major scientific accomplishments in the last decade, and where they were published.

i) Experimental Observation of Higg’s Boson

Higg’s boson is one of the fundamental particles predicted to exist according to the prevailing Standard Model of the particulate universe. It helps to unify electromagnetic, weak and strong forces. The existence of such a particle was predicted in 1964 by Peter Higgs. To observe those particles, a large machine called TEVATRON (with a circumference of 6.3 km) was built at Fermi National Accelerator Lab near Chicago, and later a machine called Large Hadron Collider (with a circumference of 27 km), at CERN, Switzerland. The first concrete observation with sufficient statistical accuracy was published in 2012 in Physical Review Letters, one of the high impact journals I have mentioned above. This paper had 5154 co-authors, including several from Turkey. To find this elusive particle, sometimes called “God’s Particle” in the popular press, billions of dollars were spent over 48 years. Since then, there are at least 2725 high impact journal articles published relating to this discovery.

During this period, Turkey managed to become an associate member of CERN and immediately saw the benefits. The most cited paper of Boğaziçi University is related to Higg’s boson discovery; 3.35 of the total 3.87 articles came from Higg’s related research. It would be wise to elevate Turkey to full membership of CERN and complete the first stage of the TARLA electron accelerator project at Ankara University.

Figure 2: ATLAS detector at Large Hadron Collider and artistic rendition of the collision process.

ii) Observation of Gravitational Waves by LIGO
The Laser Interferometer Gravitational-Wave Observatory, LIGO,  made its first observation of gravitational waves reaching earth caused by the collision of two dark holes in 2016, exactly after 22 years since its construction started. This observation was first published in Physical Review Letters and cited over 4360 times. 

LIGO observatory consists of two facilities in the United States and one in Italy. They are thousands of miles apart, yet they work in connection with each other. These three detectors are going to be complemented in 2024 when India completes its facility. The origin of the observation is based on measuring the change in curvature of space when the gravitational wave passes through the arms of the interferometer, which are 4 km long. The sensitivity and vibrational isolation are so good that a change in length of one part in ten thousand of the size of a proton over 4 km apart is observable. Since 2016, more than 50 observations have been recorded and published. As far as I know, there is no single Turkish University involved in the LIGO project out of 94 institutions.

The infrared observatory being built as part of the Eastern Anatolia Observatory should provide some hope that Turkey will join some of these extensive collaborations around the World.

  
Figure 3:  One arm of LIGO Livingston, Louisiana, USA, and the vibration isolated mirrors of the interferometer.

iii) Atomic structure of COVID-19 virus

The chaos created by COVID-19 has affected everything and everybody throughout the World. Everybody agrees that anything short of the development of a vaccine and therapy will prevent a return to everyday life. I have no intention to go into vaccine and therapy efforts here. Yet, it is imperative to understand that successful development requires an in-depth understanding of the structure of the virus and how it attacks human cells. The genetic sequence of the virus was published by February 2020 in the journal of Nature. A race immediately started to solve its atomic structure. Fort his task; one needs several sophisticated tools like a cryo-electron microscope or synchrotron radiation x-ray source. Besides, one needs modern mathematical computational tools like CHARMM working on a supercomputer, or x-ray small-angle scattering tools to determine the overall shape and structure, as well as hundreds, if not thousands of experienced researchers. In fact, in the last six months since January 2020, at least 351 structural study has been completed and deposited to the Protein Data Bank, PDB, at the time of this article (1 September 2020). I am sorry to report that there is not a single institution in Turkey equipped with any of these capabilities or involved in solving protein atomic structures. Yet, there are several young researchers, as part of their Ph.D. or postdoctoral work, were engaged in this line of research before returning to Turkey. Nevertheless, one should not expect any meaningful contribution from Turkey in this field, since there are no such program or appropriate research tools.

There is a common theme in these three examples I have picked. None of these publications could have come from Turkey, nor any meaningful key contribution would have originated from Turkey.  Such facilities like Large Hadron Collider, LIGO, or an electron synchrotron do not exist now, and it is not expected to exist in our lifetime. The machines I mentioned here are already in their fourth or fifth generation cycle. They have been built since the 1930s. Turkey did not have any such effort until the 2010s.  They need money and manpower that the meager 1% of the allocation of GDP would not enable them to accomplish such tasks.

As a result, Turkey remains mostly a spectator. Many talented Turkish scientists travel to these places for their Ph.D. work, and when they come back to Turkey, they can no longer pursue such basic studies. Interestingly, when the opportunity presents itself, Turkey is apt to miss it. One such example is a project called SESAME. Upon unification of Germany, the Helmholtz Association decided to rebuild a new synchrotron machine in Adlershof Park in East Berlin and retire the existing BESSY, Berlin Electron Synchrotron. Several countries in the Middle East were offered a chance to host it in 1999. Jordan, with a new King at that time, offered needed support and agreed to host the facility, while Turkey was lukewarm. Luckily, Turkey has signed up as a founding member, and ever since paid his share of yearly dues as an exemplary country. There are now more than a dozen research groups in Turkey submitting research proposals to utilize this new synchrotron radiation facility. More recently, a historic opportunity exists to build a macromolecular crystallography beamline, precisely suited for protein structures like COVID-19, EBOLA, HIV, SARS, and such. We will see….

 Figure 4: Atomic structure of CoVid-19virus (A), Cryo-electron microscope (B),  Advanced Photon Source, Argonne National  Laboratory, Argonne, Illinois USA (C), and DIAMOND synchrotron facility in Oxford, England (D). Currently, these two facilities lead the World on CoVid-19 related structural studies.

2.  Turkish universities with respect to the World?

Almost all countries in the World use several different scales to measure the success of their Universities. There are private companies, state-supported agencies, and international organizations performing and publishing their findings. However, there is no universally agreed upon “scale,” “index,” or “metric.” One of the recognizable metrics is the “Nature Index,” which I have chosen to exploit. It is comprised of 82 High Impact Journals in various fields of science and medicine. One can argue that this is a very narrow measure, and I will not disagree with that. Yet, it has been updated frequently during the year, and it has established a baseline, so one can follow if there is an improvement or not. These 82 journals publish about 60,000 papers annually. Among them, Nature and Science have an impact factor of 43 and 41, respectively. Nature, being published since 1869 have 3 million readers around the World. Nature Index is freely accessible since 2015.

Another index I will make use of is the “QS-Index” or Quacquarelli Symonds. Formerly known as Times Higher Education (THE) index, I followed the progress of Turkish universities in the last five years.  QS index uses 5 different criteria to rank the universities: Academic research (40 %), faculty/student ratio (20 %), number of citations per faculty (20 %), Institutional Reputation (10 %), number of foreign students (5 %), and number of foreign faculty (5 %).

According to the QS index, there is no Turkish University in the first 400. While there were five universities in the first 500 in 2015, presently, there is only one. In other words, we can say that whatever we have tried in the last five years, it didn’t work. That can be seen clearly In Figure 5.

This, however, does not mean that Turkey isn’t trying. Just to the contrary, there is a program by TUBITAK, 2232, which managed to attract 127 young scientists to Turkey from abroad in 2019. Most of these researchers returned to a university of their choice, but also private companies, as well. However, there are lingering questions: Will the program continue in 2020 (it turns out, yes), will further the devaluation of the Turkish Lira affect this program adversely? What is the satisfaction level of the first returnees? Will the program be enhanced in the coming years? These questions must be answered positively for further success.

Let’s take a closer look at the situation of individual universities. The top ten universities in Turkey seems to be always the same: Bilkent, Koç, ODTÜ, İTÜ, Sabancı, Boğaziçi, Ankara, Hacettepe, Istanbul ve İzmir Yüksek Teknoloji Enstitüsü. Independent of their relative positions to each other, we are concerned with their overall standing related to the rest of the World. The best university in terms of research published in high impact journals is Bilkent. Yet, they regressed from 411 to 551 in the last five years. That is despite the fact that two major nanotechnology centers, which are relatively modern and successful on their own, have been built. As a result, they managed to increase the overall scientific output of the university. Yet, we understand that this is not enough to gain ground among the other top research universities. Why is that? Why is it that despite sizeable investment in research centers, very accomplished faculty, and the best students selected from a nationwide competition, the relative position of Bilkent went backward, along with all other Turkish universities? I should remind that the three examples I gave earlier in this article, namely Higg’s boson, gravitational waves, and CoVid-19 research. The key to the answer is hidden there; the rest of the World is not sleeping.

In 2017, there were five Turkish universities in the top 500. They were Bilkent-Koç-ODTÜ-Sabancı-Boğaziçi. In 2020, we only see Koç at 465, and that’s it. It is clear that even the top universities have regressed in their relative positions. To place at least one university from Turkey into top 100 by the year 2023 is one the goals explained by TUBITAK’s president. In order to achieve this, Bilkent, must publish 117 high impact articles from the current level of 13. This nine- fold increase with the current state of affairs is nothing but impossible.

Figure 5.  The relative position of the top 8 Turkish universities according to the QS index. As all of them regressed, it is clear that something else must happen to change the situation.

Figure 6. Most successful universities in Turkey and the World. The average difference is 75-fold. To reduce that vast difference, some new mechanisms and policies have to be initiated and adhered to, including investment in large scale scientific infrastructure, more autonomous universities, and a stronger, secular  K-12 education. 

3. Outside the academia: Public and private research organizations and their contribution to the research output of Turkey

One of the topics often ignored is the situation of public and private research institutions and their ability to publish in high impact journals. As far as I could determine, there is not a single article published by a non-academic research organization in Turkey in these top journals. In contrast, in Switzerland, two companies, La Rouche and Novartis, were able to produce more research papers than all of the Turkish universities combined. While this is an interesting observation, it also heralds another dilemma: Turkey’s plans to increase the share of R&D budget in GDP counts on half of the investments coming from the private sector. The fact that there is no culture of high-level research, the assumption that half of the investment will come from the private sector is just unrealistic.

Figure 7. Companies publishing in scientific research journals use the allure of scientific research to attract top talents into their research labs.

4) What can be and should be done?

First things first: We must admit that we have a serious problem at hand. Since the problem persists from the beginning of the Republic, the solution cannot be just political. It is not enough to say that we respect science and agree that it is crucial. We need to accept the consequences of agreeing with the observations we made above. We have to increase the share of R&D by up to 3 % of GDP in a reasonable time frame of a few years. We need to abandon the non-scientific, undemocratic practices in governing the universities. We have to agree that the universities are autonomous in their own fiscal and administrative governance, including choosing their presidents, deans, and department chairs, admission policies, and determining their curriculum.

Turkey has taken some practical measures towards creating a scientific infrastructure. The law 6550 that was enacted in 2015 foresees forming a dozen or so institutions in various fields ranging from fundamental physics to biology, technology, and social sciences. Yet, five years from the passage of the law, none of the centers have managed to get their full budget allocation on time and thoroughly. Many of the centers are under undue pressure from incompetent administrators of respective universities and in bureaucrats in ministries. There is no grand master plan on how to connect these centers and connect them to existing infrastructure. Research clusters, a popular form of creating synergies among the existing laboratories, are unheard of in Turkey. The workshops organized by TUBITAK to bring the top 100 scientists from abroad for a few days should turn into a permanent structure to follow up on the discussions and turn them into action plans. Turkey immediately should draw up plans to join several large-scale European research organizations like EMBL, ESS, CERN, and ITER as a full member. Turkey should integrate its research capacity to European infrastructure as soon as possible.

***The original article has been published at https://www.herkesebilimteknoloji.com/ in Turkish***