Scientific Rubbish

Just when I thought that I had probably said enough about bad science, along comes another paper in Nature, ( https://doi.org/10.1038/d41586-021-03035-y) where it was noted that hundreds of junk-science papers have been retracted from reputable journals after fraudsters manipulated the publication process. Many of these were caught in “special issues”, where in some cases the whole issue was rubbish. A special issue is often published when someone suggests that a collection of papers on a specific topic would be helpful. Thus one of those attacked was Springer-Nature’s Journal of Nanoparticle Research. A groups of what appeared to be eminent computer scientists and engineers from well-known institutions in Germany and the UK wrote to the journal’s editors suggesting a special issue about the role of nanotechnology in health care. The editorial board agreed with the proposal and created a special issue entry in its editorial management system and apparently authorised access to three group members so they could handle the manuscripts. Reliance on sloth of others always pays dividends!

It appeared that months later some members of the editorial board argued the papers were of poor quality and they investigated. However, some of the papers had already been published. The investigation revealed that the original proposers were not who they claimed to be. Now this strikes me as evidence of particular slothfulness on the part of the journals. If someone claims to be a senior person at a university, they will be listed on the University’s web page, usually with evidence to support the glowing claims. That is because if the person is any good, the University wants to take what credit it can. Now, if only I knew who these editors were, why I have a very important bridge to sell them.

Which raises the question, why do people do this? One reason suggested by the article is that the scammers offer a service to researchers who are not doing very well. For a payment, they will put the name on the paper. Oddly enough, a paper with many contributors from many places is often considered to be very good, because a lot of people will have sorted out the bad stuff. Nobody checks to see if the names really knew about the paper, so a genuine “big name” can be added. Now the researcher gets another paper from a “reputable journal” to add to their CV, which means they get help for their funding applications, or even keep their jobs. One criticism of that theory raised in the linked item was, “The papers are so obviously terrible, so why would you want them on your CV?” That reasoning is wrong because it carries an inherent assumption: those reviewing the fund application or the promotion/appointment lists actually read the papers. The CV lists a title. It may seem incomprehensible, but on its own that happens with a lot of reputable papers to those not directly involved in the field. As an example, here is the title of a paper “Single ion thermal wave packet analyzed via time- of-flight detection.” That, I must add, is a perfectly respectable paper, but how many readers would know what it was about just from the title? You would have to read the paper to know whether it is respectable, and you would need to know some physics. It would not be that difficult to write something about nanotechnology that looked vaguely respectable to those completely outside the field. All you have to do is take an existing paper and change some words, mainly nouns, but keep the verbs and important keywords.

So what should happen to stop this happening? The first question is, why are Springer and Elsevier being attacked? The answer is these are big commercial publishers, so it is money. The special issues make money without the need for particular effort. But I think the second issue is to examine why people do it? The procedures of funding research or employment must change so that the number of papers is meaningless. The third issue is that you hear that scientific papers are peer reviewed and hence have real value. What this farce shows is that peer review is a farce in many cases. But maybe that is for another time, but not next week.

Advertisement

An Ugly Turn for Science

I suspect that there is a commonly held view that science progresses inexorably onwards, with everyone assiduously seeking the truth. However, in 1962 Thomas Kuhn published a book “The structure of scientific revolutions” that suggested this view is somewhat incorrect. He suggested that what actually happens is that scientists spend most of their time solving puzzles for which they believe they know the answer before they begin, in other words their main objective is to add confirming evidence to current theory and beliefs. Results tend to be interpreted in terms of the current paradigm and if it cannot, it tends to be placed in the bottom drawer and is quietly forgotten. In my experience of science, I believe that is largely true, although there is an alternative: the result is reported in a very small section two-thirds through the published paper with no comment, where nobody will notice it, although I once saw a result that contradicted standard theory simply reported with an exclamation mark and no further comment. This is not good, but equally it is not especially bad; it is merely lazy and ducking the purpose of science as I see it, which is to find the truth. The actual purpose seems at times merely to get more grants and not annoy anyone who might sit on a funding panel.

That sort of behaviour is understandable. Most scientists are in it to get a good salary, promotion, awards, etc, and you don’t advance your career by rocking the boat and missing out on grants. I know! If they get the results they expect, more or less, they feel they know what is going on and they want to be comfortable. One can criticise that but it is not particularly wrong; merely not very ambitious. And in the physical sciences, as far as I am aware, that is as far as it goes wrong. 

The bad news is that much deeper rot is appearing, as highlighted by an article in the journal “Science”, vol 365, p 1362 (published by the American Association for the Advancement of Science, and generally recognised as one of the best scientific publications). The subject was the non-publication of a dissenting report following analysis on the attack at Khan Shaykhun, in which Assad was accused of killing about 80 people with sarin, and led, 2 days later, to Trump asserting that he knew unquestionably that Assad did it, so he fired 59 cruise missiles at a Syrian base.

It then appeared that a mathematician, Goong Chen of Texas A&M University, elected to do some mathematical modelling using publicly available data, and he got concerned with what he found. If his modelling was correct, the public statements were wrong. He came into contact with Theodore Postol, an emeritus Professor from MIT and a world expert on missile defence and after discussion he, Postol, and five other scientists carried out an investigation. The end result was that they wrote a paper essentially saying that the conclusions that Assad had deployed chemical weapons did not match the evidence. The paper was sent to the journal “Science and Global Security” (SGS), and following peer review was authorised for publication. So far, science working as it should. The next step is if people do not agree, they should either dispute the evidence by providing contrary evidence, or dispute the analysis of the evidence, but that is not what happened.

Apparently the manuscript was put online as an “advanced publication”, and this drew the attention of Tulsi Gabbard, a Presidential candidate. Gabbard was a major in the US military and had been deployed in Syria in a sufficiently senior position to have a realistic idea of what went on. She has stated she believed the evidence was that Assad did not use chemical weapons. She has apparently gone further and said that Assad should be properly investigated, and if evidence is found he should be accused of war crimes, but if evidence is not found he should be left alone. That, to me, is a sound position: the outcome should depend on evidence. She apparently found the preprint and put it on her blog, which she is using in her Presidential candidate run. Again, quite appropriate: resolve an issue by examining the evidence. That is what science is all about, and it is great that a politician is advocating that approach.

Then things started to go wrong. This preprint drew a detailed critique from Elliot Higgins, the boss of Bellingcat, which has a history of being anti-Assad, and there was also an attack from Gregory Koblentz, a chemical weapons expert who says Postol has a pro-Assad line. The net result is that SGS decided to pull the paper, and “Science” states this was “amid fierce criticism and warnings that the paper would help Syrian President Bashar al-Assad and the Russian government.” Postol argues that Koblentz’s criticism is beside the point. To quote Postol: “I find it troubling that his focus seems to be on his conclusion that I am biased. The question is: what’s wrong with the analysis I used?” I find that to be well said.

According to the Science article, Koblentz admitted he was not qualified to judge the mathematical modelling, but he wrote to the journal editor more than once, urging him not to publish. Comments included: “You must approach this latest analysis with great caution”, the paper would be “misused to cover up the [Assad] regime’s crimes” and “permanently stain the reputation of your journal”. The journal then pulled the paper off the publication rank, at first saying they would edit it, but then they backtracked completely. The editor of the journal is quoted in Science as saying, “In hindsight we probably should have sent it to a different set of reviewers.” I find this comment particularly abhorrent. The editor should not select reviewers on the grounds they will deliver the verdict that the editor wants, or the verdict that happens to be most convenient; reviewers should be restricted to finding errors in the paper.I find it extremely troubling that a scientific institution is prepared to consider repressing an analysis solely on grounds of political expediency with no interest in finding the truth. It is also true that I hold a similar view relating to the incident. I saw a TV clip that was taken within a day of the event where people were taking samples from the hole where the sarin was allegedly delivered without any protection. If the hole had been the source of large amounts of sarin, enough would remain at the primary site to still do serious damage, but nobody was affected. But whether sarin was there or not is not my main gripe. Instead, I find it shocking that a scientific journal should reject a paper simply because some “don’t approve”. The reason for rejection of a paper should be that it is demonstrably wrong, or it is unimportant. The importance cannot be disputed, and if it is demonstrably wrong, then it should be easy to demonstrate where it is wrong. What do you all think?