“Science has taken a turn towards darkness”

These are quotes from the editor of The Lancet, the premier medical journal in the world, reported by Steve Hayward at Powerline. My link is through Hayward’s post, but any of us who have dealt with global vermin will know what he means. And that is not where it stops.

“A lot of what is published is incorrect.” I’m not allowed to say who made this remark because we were asked to observe Chatham House rules. We were also asked not to take photographs of slides. Those who worked for government agencies pleaded that their comments especially remain unquoted . . . this symposium—on the reproducibility and reliability of biomedical research, held at the Wellcome Trust in London last week—touched on one of the most sensitive issues in science today: the idea that something has gone fundamentally wrong with one of our greatest human creations.

The case against science is straightforward: much of the scientific literature, perhaps half, may simply be untrue. Afflicted by studies with small sample sizes, tiny effects, invalid exploratory analyses, and flagrant conflicts of interest, together with an obsession for pursuing fashionable trends of dubious importance, science has taken a turn towards darkness. . .

The apparent endemicity of bad research behaviour is alarming. In their quest for telling a compelling story, scientists too often sculpt data to fit their preferred theory of the world. Or they retrofit hypotheses to fit their data. Journal editors deserve their fair share of criticism too. We aid and abet the worst behaviours. Our acquiescence to the impact factor fuels an unhealthy competition to win a place in a select few journals. Our love of “significance” pollutes the literature with many a statistical fairy-tale. We reject important confirmations. Journals are not the only miscreants. Universities are in a perpetual struggle for money and talent, endpoints that foster reductive metrics, such as high-impact publication. . . nobody is ready to take the first step to clean up the system. [Emphasis added.]

Scientific research in some areas has become like the media. There are only certain conclusions you are allowed to reach and if you want that next grant, or that next publication, you had better make sure you are well within the acceptable range of opinion.

And here is the direct link to the editor of The Lancet where you can read his comments in full.

This entry was posted in Cultural Issues, Ethics and morality. Bookmark the permalink.

42 Responses to “Science has taken a turn towards darkness”

  1. Peter OBrien

    Where have I heard something like this before? Aaah yes….!

  2. Mayan

    McCloskey and Ziliak wrote an educational book on the perils of blind faith in statistical significance, urging researchers to pay more attention to effect size:

    http://www.amazon.com/Cult-Statistical-Significance-Economics-Cognition/dp/0472050079/ref=sr_1_1?ie=UTF8&qid=1432787729&sr=8-1&keywords=cult+of+statistical+significance

  3. Tom

    Science and the academy, like politics, have been captured by what could be described as a fantasy of fashionability: the utopian dreaming of a technocratic ruling class which requires – almost demands – that all of the achievements of Western civilisation and The Enlightenment are a great evil that must be wiped from our imagination in a quasi-religious holocaust.

    In their quest for telling a compelling story, scientists too often sculpt data to fit their preferred theory of the world.

    What Richard Horton is describing The Lancet is the West’s descent into madness after the invasion of science by Malthusian savages who have been de-educated over the past two generations by our schooling system.
    The simultaneous removal of moral certainty through the collapse of religion and the family has also been accompanied by the suffocation of liberty and enterprise through the explosion in the size of government, which, in the West, adds yearly to its prohibitions on personal behaviours as the size of the statute books grows and now confiscates 30-60% of citizens’ earnings, compared with around 10% a century ago.
    These developments virtually ensure that the systemic virus can’t just be modified or corrected to put us back on the right path. A major social and economic crisis in the West is now virtually inevitable.

  4. JC

    The cat has been saying this for years.

  5. mundi

    Engineering has become far more important than science – which is usually government backed.

    If you look at all the successful engineered tech, very little of it is coming from breakthroughs in academia. With the exception of perhaps medical, all the technological scientific advancement is coming out of the private sector.

  6. Sydney Boy

    There is far too much of a self-licking ice cream in research (predominantly university research). Researchers need to get grants and funding to pay for the research, and they can only get grants and funding by being published in said journals. Regardless of whatever crap that is that they publish. Of all the university staff (academic staff) now employed at universities, fewer than 50% actually teach. Many of them are just involved in their own little projects of minuscule significance – just trying to get published so they can get more funding to get published again. It’s just a make-work scheme much of the time.

  7. Gilas

    Statistical significance at the p<.05 is just that ie. indication that a positive (non-null) result has a 1 in 20 probability of being positive by chance. Almost 100% of acceptable medical evidence has that degree of significance, and all decently trained doctors know that.
    There is no absolute certainty in medical (or any scientific) research. What Richard Horton (at the link) is arguing isn’t about the 0.05 level of statistical significance, but the practices of innumerate pseudo-scientists who are highly dependent on funding institutions for their next meal, as we see in the Great CAGW scam.
    He then goes on to fkuc it up by referencing high-energy particle physics, arguing that a much higher level of certainty against the null-hypothesis be required in publishable research. If that were accepted NOTHING, nada, zero, zilch in medicine would pass the test, all medications and procedures currently used would be proved useless, and with rolled-gold bells on.

    Oh, and just forget applying this standard to any “settled” science!…

    Interestingly, The Lancet was the first “reputable” Medical Journal I noted to have been captured by the leftards. Just reading their editorials about Global Health, Inequality, Social/LGBT justice and Rights for All (think of the children!) would make any Cat green with bile.

    Hypocrisy at its purest.

  8. Jim

    Academics are faced with an incentive structure (publish in top journals to get ahead) that simply channel their efforts into increasingly useless areas of research.

    To get published, you need to come up with something that is so theoretically clever, so narrow, or so complex, it is probably of little (if any) use to practitioners / society. The gap between research and real-world needs is widening, and the gap seems to be filtering down to undergraduate teaching.

    I host a few undergraduate economics students each year that are required to do an “industry placement / internship” as part of their 3rd year. They all seem to have an amazing capacity to do complex proofs (useful if you want to do a PhD in North America I’m sure), but no capacity to actually apply some basic economic principles to a practical problem.

    Isn’t it time the incentive structures were changed for universities and researchers, so that some of the greatest minds in society actually did something of value?

  9. Art Vandelay

    much of the scientific literature, perhaps half, may simply be untrue

    The Lancet knows all about publishing articles which are garbage.

  10. Zatara

    (useful if you want to do a PhD in North America I’m sure)

    Mexico or Canada?

    These days students seem more interested in complex poofs.

  11. goatjam

    In future ages (if there are any) this era will be referred to as The Endarkenment

  12. jupes

    Our love of “significance” pollutes the literature with many a statistical fairy-tale.

    No shit Sherlock.

    How many people died in Iraq again?

  13. Some History

    much of the scientific literature, perhaps half, may simply be untrue

    That could be very flattering. When it comes to Public Health, most specifically “lifestyle epidemiology”, the bulk of what passes for “science” could be deemed [agenda-driven] trash.

    Lifestyle epidemiology (LE) relies almost exclusively on the flimsy statistic of “relative risk” (or variations thereof) which can be very deceptive. In LE, control-group baselines are typically very, very low/small, in many instances not too far above 0%. Multiplications of these tiny baselines reflected in “relative risk” are themselves tiny. At best, statistically significant relative risk is tapping very low-level, atypical trends. And, of course, statistical significance of itself is not an indicator of underlying causation, if any.

    This sort of “featherweight” research would require highly tempered language in arriving at conclusions. But not so in LE. Absolute risk or the extent of a baseline are rarely if ever mentioned. It’s probable that “researchers” are not even familiar with these basic concepts. Statistical significance (of relative risk) alone has come to mean everything and the term “causation”, with a strong implication requiring of strong evidence, is flung about the LE literature with reckless abandon. Highly-atypical trends are spoken of as highly-typical trends and in strong, direct causal terms. Results are not only misrepresented but misrepresented in the extreme.

    This is the norm in LE. Highly flawed research that produces conclusions that do not coherently follow from results rarely, if ever, attract critical scrutiny in the literature. It simply becomes part of “mounting evidence”. The “mounting evidence” subsequently becomes “settled science”. The “settled science” then becomes highly influential in public policy.

    It’s like comedy capers with science but there’s no comedy here. This has been going on for more than half a century in LE. It’s devoid of scholarship. Over that time, some bright minds would have forced required changes to how research is conducted. Rather matters have only worsened.

    This bastardized “science” run by the shallow and incompetent is addling people’s minds. A few days don’t go by that there’s some new scare coming from LE where results are terribly over-interpreted and further “catastrophized” by media headlines. We’re in another period of superstition masqueraded as “science”. Unfortunately, the incompetence and fraud are to the benefit of the Public Healthists. It promotes the impression to the public that there are all sorts of new, earth-shattering discoveries being made by Public Health almost on a daily basis. It provides Public Health with a highly influential status and capacity to attract funding that is unmerited and fraudulently acquired. Public Health is demonstrably incapable of scientific self-scrutiny. The correction must come from outside PH. But here academia generally has also been found wanting.

    There is a particular theme in these shenanigans – agenda. As long as conclusions fit an agenda, e.g., antismoking, research, no matter how flawed, is considered “wonderful”. If conclusions buck the agenda, the research, no matter how exemplary, is swarmed upon with voracious critique. Again, contrary to scientific ideals, Public Health students are taught agendas. They are also taught that they should be changing…. “fixing up”… society/world in line with these agendas. It’s a dangerous mix. We have people that are ideologically aligned who are responsible (in the monopolizing sense) for what will count as data, the acquiring of data, the analysis of data, and the interpretation of data. In terms of coherent enquiry, this arrangement is woeful.

  14. Some History

    A brief example of what passes for “science” in LE/PH (apologies for the sensitive nature of the example but it makes the point clearly).

    Over the last three decades, the antismoking of Public Health has produced many abominable claims. And right up there in the abominable category is the inflammatory claim that smoking/SHS “causes” sudden infant death syndrome (SIDS).

    SIDS in particular presents a peculiar situation. Unlike other mortality, it has no identified disease precondition. It is a syndrome defined by exclusion rather than demonstrable, specific pathology. When no other pathology explanation is possible, it is labeled as the unexplained category of SIDS: “SIDS is defined as the sudden death of an infant under 1 year of age that cannot be explained after a thorough case investigation, including a complete autopsy, examination of the death scene, and review of the clinical history…”

    If there is any “positive” in the circumstance is that SIDS is very rare, highly atypical. There has been much research into SIDS producing a very long list of “risk factors”. The problem is that, given that SIDS is rare, in Relative Risk assessments, the baseline is tiny, barely above zero (a fraction of 1%). Even RRs of 5.0, 10.0, 20.0 don’t necessarily mean anything because they are multiplications of a tiny baseline. Causal argument requires a reasonably high consistency of effect in absolute terms. In absolute risk terms, the probability of SIDS NOT occurring where secondary smoke is not present is (from memory) ~99.8% compared with ~99.5% where secondary smoke is thought to be present: The vast, vast, vast majority of children, whether secondary smoke is “present” or not, DO NOT develop SIDS. The very high degree of consistency is that SIDS is NOT statistically associated with smokefree or “smoke present” (whatever that means) environments.

    Pushing the idea of SHS “causes” SIDS is particularly repugnant in attempting to force antismoking conformity. It’s not just the small relative risk over a tiny baseline (i.e., no consistency of effect). In the case of SHS and SIDS all sides of the causal equation are, and remain, unknown – even after “causal attribution”. According to Public Health, an unknown attribute(s) of unmeasured highly-dilute remnants of tobacco smoke produces an unknown condition through an unknown process that results in mortality. Despite a barely-above-zero baseline, small RR, and all aspects of the causal equation being unknown, zealots will shamelessly claim that exposure to SHS “causes” SIDS. It’s utter absurdity. Yet this is what passes for “causal argument” in lifestyle epidemiology. And it’s parroted by a plethora of medical, Public Health, government, and “charitable” organizations as “settled science”. Again, the reliance solely on statistically-significant relative-risk allows this inflammatory, agenda-driven nonsense to proliferate.

  15. Nato

    mundi @ 1634848

    Engineering is what people think of when they close their eyes and think of science.

    Thoroughly tested and shown to work the same way every time the experiment is repeated.

    The problem is that when a theoretician imagines a mathematical solution that we are unable to experimentally test, this is also referred to as ‘science’ and not ‘philosophy’.

    Gilas @ 16694892 I agree.

    5 Sigma is the cut-off for a discovery. eg in 2012 ATLAS and CMS gave a 5.9 and a 5.0 sigma confidence that your machines getting a reading at 126 GeV is actually a particle.

    Declaring that particle is, in fact, the legendary Higgs Boson is less certain.

    Sildenafil could compare to your 5 sigma. C22H30N6O4S was a novel chemical formula and does actually provide vasodilatory effects. Declaring it is the treatment for angina pectoralis, is less certain.

  16. Empire

    Old news.

    Would a sane person expect otherwise? This is the inevitable result of socialised “healthcare”.

    Well said @Some History.

  17. Entropy

    I dunno if I can take anything seriously from the lancet, given its history contributing to the very problem so described.

  18. Sydney Boy

    @Some History – I was appalled at the contents when I took a Masters subject in public health a few years ago on “qualitative research” – which sounds just like the LE you are describing. It actually advocated developing a bias to conduct research in certain circumstances. It actually contained concepts in research such as “Hermeneutics”, “Participatory Action Research”, “Symbolic Interactionism” and other loads of crap. And yes, it was all about sympathising and promoting those left-wing agendas.

  19. Tel

    Nothing is wrong with science, but some or our modern practitioners aren’t quite what earlier generations were. Just like government sponsored charity and government controlled industry don’t deliver good results, so too government sponsored science hasn’t delivered what citizen science did a century or two ago.

    The case against science is straightforward: much of the scientific literature, perhaps half, may simply be untrue. Afflicted by studies with small sample sizes, tiny effects, invalid exploratory analyses, and flagrant conflicts of interest, together with an obsession for pursuing fashionable trends of dubious importance, science has taken a turn towards darkness. . .

    It’s always been the case that scientific *THEORIES* contain a lot of stuff that is simply untrue. There’s nothing amazing about this, time winnows out the good from the bad. Theories are put to the flame of experimental reproduction, and harsh critical logic. At some stage we picked up the blockhead idea that “peer review” by publishing houses sending a few advance copies around to randomly chosen readers turns a mere theory in THE TRUTH. Of course it bloody doesn’t, use yer noggin.

  20. Marko

    Agree with Gilas and Art: The Lancet is the worst offender in medical science in its promotion of pseudoscience, dogged promotion of fictitious research (the MMR vaccine scare) and right-on leftism with a ‘Public Health’ angle. The Australian MJA is not much better. They are no longer scientific journals.

  21. alexis

    There is no absolute certainty in medical (or any scientific) research. What Richard Horton (at the link) is arguing isn’t about the 0.05 level of statistical significance,

    I frequently see papers where numerous end points are measured, each at the 0.05 level, which means that the chance of one of them being positive by chance alone is actually far more than 1 in 20. I used to be able to calculate it exactly. In the worst case, you could analyse ten or twenty endpoints and reach a chance of a positive result, in the absence of a real effect, of 100%.

    And that’s the research that actually appears. Imagine all the research that was abandoned and the papers that weren’t accepted because the results didn’t suit. Ioannides nailed it.

  22. I’ve been fascinated by the contradiction that is the Islamic Golden Age & why it ended around the 12th century & turned into the repulsive anti-science clericism.

    The Golden Age on reflection is real, but mostly an accident of history & geography. The culture of the arab lead muslims did not enable it, rather they ran the assets they inherited into the ground.

    Al-Mamun picked up the pro-science torch lit by the second caliph, al-Mansur, and ran with it. He responded to a crisis of legitimacy by attempting to undermine traditionalist religious scholars while actively sponsoring a doctrine called Mu’tazilism that was deeply influenced by Greek rationalism, particularly Aristotelianism. To this end, he imposed an inquisition, under which those who refused to profess their allegiance to Mu’tazilism were punished by flogging, imprisonment, or beheading. But the caliphs who followed al-Mamun upheld the doctrine with less fervor, and within a few decades, adherence to it became a punishable offense. The backlash against Mu’tazilism was tremendously successful: by 885, a half century after al-Mamun’s death, it even became a crime to copy books of philosophy. The beginning of the de-Hellenization of Arabic high culture was underway. By the twelfth or thirteenth century, the influence of Mu’tazilism was nearly completely marginalized.

    In its place arose the anti-rationalist Ash’ari school whose increasing dominance is linked to the decline of Arabic science. With the rise of the Ash’arites, the ethos in the Islamic world was increasingly opposed to original scholarship and any scientific inquiry that did not directly aid in religious regulation of private and public life. While the Mu’tazilites had contended that the Koran was created and so God’s purpose for man must be interpreted through reason, the Ash’arites believed the Koran to be coeval with God — and therefore unchallengeable.

    Draw what you wish, I fear the same process is at work today.

  23. Michaelc58

    I unconsciously gave up on medical research 20 years ago when I saw that the ‘salt causes hypertension’ dogma was based on careers grown from experiments feeding mice near their bodyweight in salt. (high salt diets are however bad for other reasons).

    I can think of a number of causes of this corruption of science:
    1.An entitlement mentality to a ‘research career’ by masses of mediocre people
    2.Sheer volume of nonsense pressed into publication by the said masses.
    3.Ever growing acceptance by the public and authorities of pure bullshit for reasons beyond me – health funds paying for Homeopathy comes to mind.
    4.Loss of basic reasoning skills regarding cause and effect, action and consequences, in part due to an increasingly magical world of unfanthomable technology – who can any longer say with certainty what is and is not possible – consider entanglement, worm holes, time travel, climate control, annual public health scares, who knows, so why not homeopathy?
    5.Computer modelling – the biggest research frauds I have seen involved computer modelling – so sexy, so convincing, so expensive and yet so misleading in complex non-linear biological systems. And that was long before climate models.

    Just saying.

  24. egg_

    1.An entitlement mentality to a ‘research career’ by masses of mediocre people
    2.Sheer volume of nonsense pressed into publication by the said masses.
    3.Ever growing acceptance by the public and authorities of pure bullshit for reasons beyond me – health funds paying for Homeopathy comes to mind.
    4.Loss of basic reasoning skills regarding cause and effect, action and consequences, in part due to an increasingly magical world of unfanthomable technology – who can any longer say with certainty what is and is not possible – consider entanglement, worm holes, time travel, climate control, annual public health scares, who knows, so why not homeopathy?
    5.Computer modelling – the biggest research frauds I have seen involved computer modelling – so sexy, so convincing, so expensive and yet so misleading in complex non-linear biological systems. And that was long before climate models.

    Yup, per upthread, likely all the money and brainpower has gone into Engineering.

  25. rickw

    Engineering has become far more important than science.

    The great thing about Engineering is that if you are bullshitting, reality will point out the error of your ways pretty darn quickly.

  26. The great thing about Engineering is that if you are bullshitting, reality will point out the error of your ways pretty darn quickly.

    Yes, engineers face the consequences of their mistakes and can gain direct benefits.

  27. Lawrie Ayres

    I don’t suppose this fellow is referring to climate science as well. The in crowd at climate science central do not have to show their workings or their data because they are just right. They also do not have conflict of interest either judging by Flannery’s connection with geothermal research and Rudd’s $90 million. I mean they are scientists aren’t they?

  28. J.H.

    The reason is easy to explain….. They are doing Politics and not Science.

    It is, that simple.

  29. Alexis

    And 6. The fact that doctors need to research, maybe do a PhD just to qualify as specialists even though they might be far better off just treating; leading to redundant research by possibly completely unsuited minds.

  30. Tel

    Computer modelling – the biggest research frauds I have seen involved computer modelling – so sexy, so convincing, so expensive and yet so misleading in complex non-linear biological systems. And that was long before climate models.

    Computer modelling is an excellent tool to boost the power of your thinking process. You can explore larger data sets, more autonomous agents, bigger everything than what you can do on paper. However, it won’t do your thinking for you. If you are an idiot, the computer will make you a bigger and more dangerous idiot.

    That said, there are some really good things about computer models (if the full source is published):
    [1] they are highly repeatable
    [2] they are subject to many incremental improvements
    [3] they are subject to independent scrutiny

    If you don’t release your computer models (in full) then (pretty much by definition) you have something to hide. From a commercial perspective this might make sense, if you think your model is giving you an edge somehow. From a scientific perspective it never makes sense, because the objective of science is truth not profit.

  31. wreckage

    The difference between science and a publicly traded company is that the publicly traded company has to publish its data.

  32. Roger

    In their quest for telling a compelling story, scientists too often sculpt data to fit their preferred theory of the world.

    You don’t say.

  33. john malpas

    And yet , and yet – people are living longer in spite of all. How can this be if everybody is wrong?
    Or maybe us oldies are really dead.

  34. Caroline Storm

    How extremely worrying to find out that The Lancet and Wellcome Trust can in any be seen to have any common interest.

  35. Some History

    Speaking of junk, here’s an interesting read:

    I Fooled Millions Into Thinking Chocolate Helps Weight Loss. Here’s How
    http://io9.com/i-fooled-millions-into-thinking-chocolate-helps-weight-1707251800

  36. Gilas

    john malpas
    #1695339, posted on May 29, 2015 at 9:04 am

    And yet , and yet – people are living longer in spite of all. How can this be if everybody is wrong?
    Or maybe us oldies are really dead.

    Improvements in public health, namely sanitation, sewerage control, vaccination, improved hygiene due to the Enlightenment and increasing wealth (just compare to that eternal shit-hole known as Africa).

    And some truly novel treatments such as aspirin, synthetic morphine, antibiotics, anaesthesia and surgical antiseptic protocols. Most of these were discovered by truly inspired geniuses (results of a decent education and personal ethic) and adopted during war emergencies, ie. without the self-serving, cancerous government/professional bodies/bureaucratic oversight so essential today.

    I estimate that the summit of human achievement was the century between 1850 and 1945. It’s been downhill ever since.

    Unless the left is massacred, our descendants are lost.

  37. Gilas

    Withour referencing mine, the comments in this thread are quite impressive. A higher standard of decorum and knowledge than many Cat-fights…

    They should be collected.

  38. Ralph Seccombe

    Sorry, I couldn’t get past the Reagan quote about government burueas and programs lasting forever. If this statement had any validity, it hasn’t been true since the days of Thatcher and Reagan himself. Sackings of public servants are now routine, and didn’t Abbott close down something like 100 agencies?
    Not impressed!

  39. walter p komarnicki

    I wrote to Deutshe Welle’s ‘In Good Shape’ health program the other week about how acupuncture helped me stop smoking, is easing the pain of arthritis and sciatica (I’m 70 and a half), and is helping me recover from a recent motorbike accident.
    They said the researchers at Harvard have been studying acupuncture, but most studies have not been of a positive nature. I can only go by my own experience, and certainly do not consider acupuncture a placebo.

    I then asked about a previous program I had seen about how the majority of Germans held strongly to views that homeopathy was effective – despite all evidence to the contrary – but probably because it had been thought up by a German in the 19th century and so this made it all go down so well.
    Their Medicare system even subsidises it!

  40. Æthelred

    Gee, if only half of the published results were wrong that would be fantastic because least the other half would be right.

    In cancer research (= fundamental biology) about 90% of the conclusions published in top journals are wrong.

    Begley & Ellis (2012), Raise standards for preclinical cancer research. Nature 483:531

Comments are closed.