Epidemiology, Diet Soda and Climate Science

Guest Essay by Kip Hansen – 6 October 2019

featured_image_epidemiologyEpidemiology is the study and analysis of the distribution (who, when, and where), patterns and determinants of health and disease conditions in defined populations. “

“It is the cornerstone of public health, and shapes policy decisions and evidence-based practice by identifying risk factors for disease and targets for preventive healthcare.”

That’s the Wiki speaking on the subject.

And that is precisely where the broad field of epidemiology has gone wrong.

WARNING:  This is a long essay.  Not a quick read — about 5,000 words — a twenty minute read for most.  Put it aside for reading at your leisure.  Let me know in comments if you find it was worth your time.

If we ask what epidemiologists mean by “risk factors” we find the fatal flaw:

“Risk factors or determinants are correlational and not necessarily causal, because correlation does not prove causation. For example, being young cannot be said to cause measles, but young people have a higher rate of measles because they are less likely to have developed immunity during a previous epidemic. Statistical methods are frequently used to assess the strength of an association and to provide causal evidence (for example in the study of the link between smoking and lung cancer). Statistical analysis along with the biological sciences can establish that risk factors are causal.

Some prefer the term risk factor to mean causal determinants of increased rates of disease, and for unproven links to be called possible risks, associations, etc.”  —  from the Wiki

Why do I describe the above as “where the broad field of epidemiology has gone wrong?”  First, statistical analysis  cannot, ever, establish causality in this field.  To see why this is so, let’s look at what John P.A. Ioannidis (see here and here)  wrote just last year about nutritional epidemiology in the Journal of the American Medical Association (JAMA) — an editorial piece titled: “The Challenge of Reforming Nutritional Epidemiologic Research” [ pdf courtesy of Columbia University ]:

“Some nutrition scientists and much of the public often consider epidemiologic associations of nutritional factors to represent causal effects that can inform public health policy and guidelines. However, the emerging picture of nutritional epidemiology is difficult to reconcile with good scientific principles. The field needs radical reform.”

 “In recent updated meta-analyses of prospective cohort studies, almost all foods revealed statistically significant associations with mortality risk.  Substantial deficiencies of key nutrients (eg, vitamins), extreme over consumption of food, and obesity from excessive calories may indeed increase mortality risk. However, can small intake differences of specific nutrients, foods, or diet patterns with similar calories causally, markedly, and almost ubiquitously affect survival?”

Ioannidis’ findings and his question are both very apt.  What is he saying here?  He is saying that when they looked at epidemiological nutrition studies, almost every food item looked at had “statistically significant associations with mortality risk” or in other words, everything we eat is killing us faster and sooner, or making us live longer, and he asks if that is really possible.

So, what is going on here?  The first thing that is going on is that epidemiologists are being lazy — by this I mean that in many of these studies, the study design involves looking  a single dietary factor (sometimes a single dietary item) — almost always from some broad general health study database such as the European Prospective Investigation into Cancer and Nutrition (EPIC)  or, in the United States, the Nurse’s Health Studies (NHS) — and comparing that dietary factor to “All Cause Mortality”.  All Cause Mortality means simply death by any cause.  There are a lot of causes of death — the official list is called ICD-10-Cause of Death.  [ pdf ]   So, in the following example from Ioannidis,  a study looked at a database that included the self-reported daily/weekly/monthly the dietary intake of hazelnuts by a huge number of people who filled out dietary surveys at some time many years ago (maybe even only once)  and then checked death indexes (in the U.S., they use the Social Security Death Index) to see which individuals had died and when.  The epidemiologists then used statistical analysis techniques to determine how those hazelnuts affected life-span.  The results of such studies?  (quoting Ioannidis as above):

“Assuming the meta-analyzed evidence from cohort studies represents life span–long causal associations, for a baseline life expectancy of 80 years, eating 12 hazelnuts daily (1 oz) would prolong life by 12 years (ie, 1 year per hazelnut), drinking 3 cups of coffee daily would achieve a similar gain of 12 extra years,  and eating a single mandarin orange daily (80 g) would add 5 years of life.  Conversely, consuming 1 egg daily would reduce life expectancy by 6 years, and eating 2 slices of bacon (30 g) daily would shorten life by a decade, an effect worse than smoking. Could these results possibly be true?

Of course they can’t!  One of the reasons is that hazelnuts do not prevent , for instance, accidents — which are the Number 3 cause of death in the United States and represent about 6% of all-cause deaths.  It is hard to imagine any plausible way in which eating hazelnuts could even contribute to the prevention of accidents (they could cause choking, if not chewed properly,   but that is a separate cause of death).   There is no biologically plausibility to the idea that eating hazelnuts somehow prevents both (or either)  heart disease and cancer — the Number 1 and 2 killers (though they may be a possible contributory factor to either a benefit or a harm, through some unknown pathway).  These three Causes of Death alone account for 50% all “All Cause Mortality” in the United States.  Any study that analyzes individual food items or diet components  against All Cause Mortality is flawed before they are begun because they are looking at an end-point that is known  not to be caused by the intervention (food item).  You’ll see the significance of this later….

“In 2017, the 10 leading causes of death were, in rank order: Diseases of heart; Malignant neoplasms [ cancer ] ; Accidents (unintentional injuries); Chronic lower respiratory diseases; Cerebrovascular diseases [ stroke ]; Alzheimer disease; Diabetes mellitus; Influenza and pneumonia; Nephritis, nephrotic syndrome and nephrosis [ kidney diseases ]; and Intentional self-harm (suicide). They accounted for 74% of all deaths occurring in the United States.”   — CDC [ pdf ]            [bracketed explanations for clarity — kh]

So, if an epidemiologist is weighing consumption of hazelnuts against All Cause Mortality then (quoting Ioannidis again):

 “These implausible estimates of benefits or risks associated with diet probably reflect almost exclusively the magnitude of the cumulative biases in this type of research, with extensive residual confounding and selective reporting.”


“Almost all nutritional variables are correlated with one another; thus, if one variable is causally related to health outcomes, many other variables will also yield significant associations in large enough data sets. With more research involving big data, almost all nutritional variables will be associated with almost all outcomes. Moreover, given the complicated associations of eating behaviors and patterns with many time-varying social and behavioral factors that also affect health, no currently available cohort includes sufficient information to address confounding in nutritional associations.” [ emphasis added – kh ]

Bottom Line:  Given the extremely complex and only vaguely understood details of human nutrition, and the correlations between those innumerable diet variables, these types of studies will find correlations and associations between [almost] all the variables and all possible outcomes.  In other words, these Big Data nutritional studies are magical — and can be used to produce almost any outcome for any variable.

We already know, from long experience, not to rely on single studies and thus we can avoid the “Single Study Syndrome”.   The oft provided solution to the Single Study Syndrome is to do meta-analysis studies — a study in which the findings, both qualitative and quantitative, of many studies on the same topic are combined “to develop a single conclusion that has greater statistical power.” [ source ].  Sounds like a grand idea, doesn’t it?  We look at lots of studies on hazelnuts or coffee or mandarin oranges and combine their results and re-do the statistical analyses and see what’s what.

But, Ioannidis has this to say about that:

“…In an inverse sequence, instead of carefully conducted primary studies informing guidelines, expert-driven guidelines shaped by advocates dictate what primary  studies should report.  Not surprisingly, an independent assessment [ pdf here ] by the National Academies of Sciences, Engineering, and Medicine of the national dietary guidelines suggested major redesign of the development process for these guidelines:  improving transparency, promoting diversity of expertise and experience, supporting a more deliberative process, managing biases and conflicts, and adopting state-of-the-art processes.”

So, here we find that the studies being done find “non-science-ical”  results — obviously invalid findings — and yet they still get reported and published in the journals.  These individual studies, that are capable of finding almost any outcome desired (or outcomes that the author’s biases lead them to) are then combined into meta-analyses which end up being simply a reflection of the biases in field of nutritional epidemiology, much of it driven by advocates which dictate what primary studies should be about and what findings they should report.

Does this strike you as Sound Science? 

It is not sound science — it is a mockery of sound science.

“Beyond food studies, results of single-nutrient studies have largely failed to be corroborated in randomized trials. False-positive associations are common in the literature. For example, updated meta-analyses of published data from prospective cohort studies have demonstrated that a single antioxidant, beta carotene, has a stronger protective effect on mortality than all the foods mentioned above. The relative risk of death for the highest vs lowest group of beta carotene levels in serum or plasma was 0.69 (95% CI, 0.59-0.80). Even when measurement error is mitigated with biochemical assays (as in this example), nutritional epidemiology remains intrinsically unreliable.  These results cannot be considered causal, especially after multiple large trials have yielded CIs [confidence intervals] excluding even a small benefit.” [emphasis added — kh ] (quoting,  again, Ioannidis)

Although I could really use those extra 12 years that are to be gained by eating hazelnuts, I have to admit that It Is Not True in the Real World.

Diet Soda?

The mass media — and all those health advocates and advocate-journalists of various stripes — have been agog with the news of a new study out of Europe.  Oh, yes, a BIG study —  451,743 people — a big cohort study (exactly the type being discussed by Ioannidis above).    The details of the study are just too ….. I was going to say “imaginary”  but thought better of it.  Instead, here they are in capsule form:

“Key Points

 Question:  Is regular consumption of soft drinks associated with a greater risk of all-cause and cause-specific mortality?

 Findings: In this population-based cohort study of 451 743 individuals from 10 countries in Europe, greater consumption of total, sugar-sweetened, and artificially sweetened soft drinks was associated with a higher risk of all-cause mortality. Consumption of artificially sweetened soft drinks was positively associated with deaths from circulatory diseases, and sugar-sweetened soft drinks were associated with deaths from digestive diseases.

Meaning:  Results of this study appear to support ongoing public health measures to reduce the consumption of soft drinks.”          — [ source ]

In this study, we see that Ioannidis is proved correct in all aspects of his criticism of Nutritional Epidemiology.  The “question” (hypothesis) of the study is pre-determined by current advocacy against “soft drinks”, a large and varied general category of popular carbonated beverages.   Sure enough, because currently practiced nutritional epidemiology using large cohort studies allows the finding of [almost] any association desired, they find that “greater consumption of total, sugar-sweetened, and artificially sweetened soft drinks was associated with a higher risk of all-cause mortality.”  When they then shine their statistical packages on more general classes of cause of death, they still find “artificially sweetened soft drinks … positively associated with deaths from circulatory disease” and “sugar-sweetened soft drinks … associated with deaths from digestive diseases”. 

[ As an interesting note, these associations, after being “adjusted” for a dozen or so possible confounders, are non-linear — that is J-shaped, low consumption appearing to  improve survival.  The abstract of this study is here — one needs to get a copy of the full study pdf and download the supplemental information to see the non-linear graphs — note that the first two comments, which appear under the abstract, agree with Ioannidis. ]

Thus, we find Ioannidis’ statement that “almost all nutritional variables will be associated with almost all outcomes” seem to be validated.  Further, “Results of this study appear to support ongoing public health measures to reduce the consumption of soft drinks” indeed [as per Ioannidis]reflect[ing] almost exclusively the magnitude of the cumulative biases” and are simply in support of “expert-driven [already existing] guidelines shaped by advocates dictat[ing] what primary  studies should report.

Further, Ioannidis states:   “Individuals consume thousands of chemicals in millions of possible daily combinations. . . . . Disentangling the potential influence on health outcomes of a single dietary component from these other variables is challenging, if not impossible.”

Nutritional Epidemiology has been outed by Ioannidis and we now have an explanation of the “nutritional science whipsaw effect” —  which we see as the always popular: “One week drinking coffee is good for you, and the next week it is lethal”.    This effect is so prevalent  that Ioannidis concludes “Nutritional research may have adversely affected the public perception of science.”

The important point of all this about nutritional epidemiology is not just that the findings published in our local newspapers, heralded in the TV news and echo-chambered ad nauseam on the ‘Net are at best almost entirely misleading (to avoid the simpler  word “wrong”), it is the reason that these results fail to inform us of anything true about nutrition — what we should eat or avoid eating:

The scientific and statistical methods used in today’s nutritional epidemiology are not capable of correctly informing us of the truths they are claiming — the causal relationships between dietary items and health outcomes.  

The proponents of this type of nutritional epidemiology are fooling themselves, particularly by looking at end points (effects) that are not directly and causally biologically connected to the intervention (dietary food item) — far too often focusing on All Cause Mortality or vague large classes of disease (such as cardiovascular disease or cancer).   When some dietary item is then statistically found to be beneficial or harmful, these effects are often justified by researchers with Kiplingesque “Just So…” stories  to explain the finding.

 For diet soda:

“Experimental evidence conducted in animals and humans has shown that artificial sweeteners disrupt the composition of gut microbes (that is, the gut microbiota) in a direction that could lead to obesity, glucose intolerance, diabetes and, ultimately, cardiovascular disease. Artificial sweeteners may also cause biological changes in the brain that influence satiety and weight gain.”                        [ source ]

And, finally:  Climate Science?

How does a better understanding of the problems found in nutritional epidemiology offer us any insight into the field of Climate Science?

At the core of both fields is the issue of causality.

Causation indicates that one event is the result of the occurrence of the other event; i.e. there is a causal relationship between the two events. This is also referred to as cause and effect.   [ source ]

As Ioannidis has pointed out in nutritional epidemiology: the specific scientific methods (large cohort studies based on food frequency surveys) and resultant statistical analyses are fundamentally incapable of ferreting out the individual effects on human health of individual, or classes of, dietary factors — they cannot discover causality.   This is both a methodological problem and a result of the object of the study — human nutrition.  The complexity of, and incredible variation in,  human diets and the interplay between dietary intake and the myriad positive and negative health effects of those dietary components and their interaction with each other, as well as a near infinite number of environmental, genetic, and societal factors make it very hard for nutritional epidemiology to discover all but the largest of effects (which are seen in poisoning by strychnine or the development clinical vitamin deficiencies).

 Similarly, for climate science, the object of study, the Earth’s climate system is not only exceptionally complex, but also chaotic.  First, we have to understand that, as we see in nutrition science, climate is comprised of hundreds of interacting components, each changing on time scales ranging from seconds to centuries, each being integral influencing and causal factors for the others — all correlated in ways we often (almost always) do not fully understand.  And, as in nutrition science, almost all climate variables are correlated with one another; thus, if one variable is found to be correlated to some weather/climate  outcome, many other variables will also yield significant associations in the huge present-time and historical data sets relating to Earth’s weather and climate.

Thus we find the situation, unacknowledged by most of the climate science field, that [paraphrasing Ioannidis] “Disentangling the potential influence on medium to long range climate outcomes of a single climatic factor, such as atmospheric GHG concentrations,  from these myriad other variables is challenging, if not impossible” based simply on the complexity of the climate itself.

This is further complicated by the fact that the climate system itself is known to be chaotic and thus highly resistant to the prediction of future states.

Edward Lorenz’s  work on the topic culminated in the publication of his 1963 paper “Deterministic Nonperiodic Flow” in Journal of the Atmospheric Sciences, and with it, the foundation of chaos theory. He states in that paper:

“Two states differing by imperceptible amounts may eventually evolve into two considerably different states … If, then, there is any error whatever in observing the present state—and in any real system such errors seem inevitable—an acceptable prediction of an instantaneous state in the distant future may well be impossible….In view of the inevitable inaccuracy and incompleteness of weather observations, precise very-long-range forecasting would seem to be nonexistent.”

That is to say, the physics of the climate system themselves are chaotic (in the special sense used in the field of study known as Chaos Theory).   Further, the Earth climate system comprises two coupled chaotic systems — the atmosphere and the oceans.  This is not controversial at all, but rather well-known and widely acknowledged by nearly everyone in the field:

“The climate system is a coupled non-linear chaotic system, and therefore the long-term prediction of future exact climate states is not possible.”

And then because the complexity and chaotic nature of the dual system prevents the normal course of science inquiry — the discovery predictable effects of known of causes — IPCC-style climate science calls for:

“Rather the focus must be upon the prediction of the probability distribution of the system’s future possible states by the generation of ensembles of model solutions.”

[ quotes from:   IPCC WG1 TAR ]

What does it mean that the climate system is chaotic?  It means that small changes in one place or one climate component can cause small or large changes in another.  Volcanic eruptions in Southeast Asia can change next year’s weather in Europe.  Large wildfires in America’s Northwest can change weather in Australia.  Sun spots, or the lack of them, may or may not be changing the climate now.

What does chaos have to do with numeric climate models?    Climate Models are numeric representations of little pieces of the climate, which all feed into one another.  The models themselves are fantastically, possibly preternaturally, complex.   Many, possibly most, of the mathematical formulas necessary to simulate the relationships and interactions between the many components of the climate system are nonlinear differential equations which do not lend themselves to solution, thus must be simplified before being used in the climate model.  These simplified formulas are mere approximations of the real relationships.

Nonlinear differential equations are often extremely sensitive to initial conditions…thus one gets the results seen in NCAR’s 40 Earths Project.  NCAR claims that the results are  “a staggering display of Earth climates that could have been along with a rich look at future climates that could potentially be.”   What their study  actually demonstrates is that Edward Lorenz was absolutely right — climate models are and will always be extremely sensitive to initial conditions  and will produce hugely different results over mid- to long-term runs even when starting points are as small as less than one one-trillionth of a degree different (in this case, in the global average surface temperature).

The solution of the IPCC and most climate modelers is to focus on “the prediction of the probability distribution of the system’s future possible states by the generation of ensembles of model solutions.”  This sounds like very sound science — but is, in light of Chaos Theory, nonsensical, and offers no real world prediction or projection at all.  Dr. Robert G. Brown, physicist at Duke University, explains why (at length) in Real Science Debates Are Not Rare. [For his discussion of the point about climate models, start reading at his sentence “At the moment, I’m reading Gleick’s lovely book on Chaos”.]

There is much, much more to this idea of the complexity and chaotic nature of the climate system and what that means for climate models.  Some of this has recently been allowed to come to light in a book (in Japanese, with an introduction and appendix in English) by Dr. Mototaka Nakamura, himself a career-long climate modeler,  in his new book  “The Global Warming Hypothesis is an Unproven Hypothesis“. (an eBook version is available for 99 cents)  From the appendix in English:

“All climate simulation models have many details that become fatal flaws when they are used as climate forecasting tools, especially for mid- to long-term (several years and longer) climate variations and changes. These models completely lack some of critically important climate processes and feedbacks, and represent some other critically important climate processes and feedbacks in grossly distorted manners to the extent that makes these models totally useless for any meaningful climate prediction.”

Donahue and Caldwell (2018) explain what happens when the order of processing is changed in climate models — you get different results!   They have a cute PowerPoint that illustrates the problems.    Erica Thompson and Leonard Smith, at the London School of Economics’ Centre for the Analysis of Time Series  have discussed The Hawkmoth Effect in terms of climate models.  They too have a poster.

“What is the Hawkmoth Effect?  The term “butterfly effect”, coined by Ed Lorenz, has been surprisingly successful as a device for communication of one aspect of nonlinear dynamics, namely, sensitive dependence on initial conditions (dynamical instability), and has even made its way into popular culture. The problem is easily solved using probabilistic forecasts. [ a point with which  I disagree — kh ]  A non-technical summary of the Hawkmoth Effect is that “you can be arbitrarily close to the correct equations, but still not be close to the correct solutions”. The less media-friendly hawkmoth does not get as much attention as its celebrated butterfly cousin. However, it is not yet accounted for by modern methods. Due to the Hawkmoth Effect, it is possible that even a good approximation to the equations of the climate system may not give output which accurately reflects the future climate. Climate decision-makers and climate model developers must take this into account.”

Thompson and Smith are willing to let “probabilistic forecasts” be a handling for The Butterfly Effect [it is not really, see Dr. R. G. Brown above — kh] but there is no getting around The Hawkmoth Effect.

So we see climate models facing a set of scientifically strong arguments against their efficacy:

  • Lorenz and The Butterfly Effect — extreme strong sensitivity to initial conditions.
  • The Processing Order problem — “There is no “correct” process ordering… and process order has a big impact on model behavior”.
  • The Hawkmoth Effect — “Structural instability of complex dynamical systems” — tiny differences in the equations used in models result in different model output (forecasts).
  • Missing and/or mathematically misrepresented processes or feedbacks in climate models result in meaningless predictions.
  • The complexity and internal correlation between the myriad components of the climate system itself inhibit the discovery of the causalities of individual components of the climate system.

None of these five factors actually prevent us from improving our understanding the climate of today or how it works.  A great deal of very good science is taking place in an attempt to figure out how the climate works, what the relationships exist between atmospheric and oceanic modes and their cycles, how clouds form and why, the relationship between the Sun and atmospheric phenomena and many other important questions.  They just make it harder.

Each of these five factors has direct bearing on the question of causality in climate science — what causes what and when.  IPCC-style climate science is focused almost in its entirely on one single climate causal factor:  greenhouse gas concentrations in the atmosphere.  This single cause is then hard-coded into climate models to produce “projections” of possible future climate factors.  These predictions/projections are then presented as proof of the necessity of implementing the proposed social and political solutions that preceded the science by decades.

Like Nutritional Epidemiology, we see, in IPCC-style climate science,  a system which thus reflect[s] almost exclusively the magnitude of the cumulative biases” of the field and  are simply in support of “expert-driven [IPCC reports and policy recommendations] guidelines shaped by advocates [including the IPCC itself among many others] dictat[ing] what primary  studies should report.”   Because climate science is such a young field, and so much is still unknown, the field has been driven by policy advocacy, funding and publication bias and social pressure on climate scientists to conform,  and models which are known to be unfit-for-purpose have been used to reinforce the “necessity” of  the social/economic/political solutions proposed by the IPCC by predicting catastrophic futures including the imminent demise of human civilization.

Also like Nutritional Epidemiology,  the difficulties in discovering causality in climate science has led experts to make to loud public policy pronouncements and predictions, not based on science but on preferred policy outcomes,  which have, with the passage of time, repeatedly and consistently failed to come to pass.  In nutrition, this is the whipsaw effect:  butter is bad, eat margarine — oops! — margarine is bad, eat butter.  In climate science, we have had:

  1. “James Hansen of NASA’s Goddard Institute of Space Studies beginning in 1988 predicted major droughts and up to six feet of sea level rise in the 1990s. One reporter recalled that in the late 1980s, he asked Hansen in his Manhattan office whether anything in the window would look different in 20 years. Hansen replied, “The West Side Highway [which runs along the Hudson River] will be under water. And there will be tape across the windows across the street because of high winds.”” [ source ]
  1. “Al Gore predicted in 2009 that the North Pole would be completely ice free in five years. A U.S. Navy scientist in 2013 concluded that the Arctic’s summer sea ice cover would all be melted by 2016.” [ source ]
  1. “ABC News ran a segment in 2008 promoting a movie called Earth 2100. Some predictions to scare us to buy the propaganda were gas reaching $9 per gallon, $12.99 cartons of milk, and New York City — engulfed by water in 2015.” [source ]

The failed climate predictions are so ubiquitous that they have become a standing joke among the general public, at least in the United States.  The constant drumbeat of catastrophic climate change predictions has, again as with nutrition science, in all probability harmed the public perception of Science in general, and continues to do so in the present.


 There are many working in the Sciences to try to bring about changes in the way science is done and the way it is reported.  These efforts to bring about corrections are often fought by the purveyors of the science field’s status quo. 

It is unfortunate that many of those fighting against needed changes are government agencies and professional science and medical associations that would have the most to gain from better science.  Like the advocacy groups that have staked out positions on various topics in nutrition science, and used their trusted positions to influence policy makers to create public health guidelines in keeping with their advocacy planks, the IPCC and associated social and political advocacy groups have seized control of public climate policy advocacy and are demanding that governments set policy conforming with their advocated social and political goals.   Not only is this confounding of science with social politics bad for science — it is bad for public policy.

Daniel Sarewitz, a professor of science and society at Arizona State University’s School for the Future of Innovation and Society, and the co-director of the university’s Consortium for Science, Policy, and Outcomes wrote in an article in The New Atlantis  (Spring/Summer 2016)  titled “Saving Science”:

In the future, the most valuable science institutions will be closely linked to the people and places whose urgent problems need to be solved; they will cultivate strong lines of accountability to those for whom solutions are important; they will incentivize scientists to care about the problems more than the production of knowledge. They will link research agendas to the quest for improved solutions — often technological ones — rather than to understanding for its own sake. The science they produce will be of higher quality, because it will have to be. The current dominant paradigm will meanwhile continue to crumble under the weight of its own contradictions, but it will also continue to hog most of the resources and insist on its elevated social and political status.

There have been some efforts to accomplish the ideals set out by Sarewitz in various fields, in addition to those of the CSPO.  Ocean Acidification has had several efforts to correct the methods and reporting of OA research (pdf here and pdf here, reported by me here and here ) and  Social Psychology has seen similar efforts (examples here and here and here; and in book form here ).  Another paper, “False-Positive Psychology: Undisclosed Flexibility in Data Collection and Analysis Allows Presenting Anything as Significant”,  “demonstrate[s] how unacceptably easy it is to accumulate (and report) statistically significant evidence for a false hypothesis” and offers some possible solutions.  The Royal Statistical Society (and its American counterparts) has called for reform of the use of statistics in scientific research.  Devang Mehta has called for reform of research publishing  in Nature.  These combined proposed methodological solutions can be applied to many fields.

And, of course,  John P.A. Ioannidis, whose work led to this essay,  has been working tirelessly in the field of medical  and clinical research.

Reading the popular science press, we see that, in real practice, many fields of science are still in the stage wherein the  “current dominant paradigm …. continue[s] to hog most of the resources and insist on its elevated social and political status.

Whatever your relationship is with Science — be it in research, education or science journalism — you can support good careful and rigorous science; you can tactfully call-out  poor science and bad science reporting; and you can  lend your efforts and your voice to the task of reforming the Sciences and restoring their proper practices and returning them to their proper place in society.

# # # # #


Author’s Comment:

We can learn by comparing the problems in one scientific field to the problems in another — and hopefully see a way forward through the obstacles and impediments to the discovery of the underlying truths of the world around us.

Both Nutritional Epidemiology and Climate Science are filled with honest hard working thinkers and researchers.  Still, the challenges presented by the need to publish, to get funding for research, to be accepted by their peers and achieve tenure and security in employment that will allow them to support themselves and their families can push them to produce results that, in the end, do not lead to real advancements in their field.  We see this played out when those who retire from the academic field of battle, only then, become very honest and open about the problems with biases being enforced in their research topic.

Share your experiences in the comments, if you can.

# # # # #

0 0 votes
Article Rating
Newest Most Voted
Inline Feedbacks
View all comments
October 6, 2019 6:24 pm

“statistical analysis cannot, ever, establish causality in this field”

I am not sure that statistical analysis is used to “establish causality” in any field. But once a theory of causality is proposed, statistical analysis can be used to determine whether the data are consistent with that theory.


Reply to  Chaamjamal
October 6, 2019 6:49 pm

Chaamjamal ==> “…in this field.” Statistical analysis can be used to find support for a hypothesis, but this type of nutritional epidemiology mostly find vague and tiny associations/correlations and then pretends that this means causality — bolstering their pretense with statistics. Thus Ioannidis’ study.

Leo Smith
Reply to  Kip Hansen
October 6, 2019 9:55 pm

A simple example may clarify.

Poor people eat junk food get fat and die young. Because junk food actually does cause mortality.
Rich people eat better food and have access to medical care at a higher level too. They live longer.,
Poor people don’t eat hazelnuts.,
Ergo eating hazelnuts is correlated with an increased life expectancy.
But eating hazlenuts wont make you live longer if you also eat junk food and have rubbish doctors.
Neither will becoming rich if you still eat nothing but pizza
Any more than here in the UK giving everyone a degree in some useless subject has increased their earning capability.

Ken Irwin
Reply to  Leo Smith
October 7, 2019 1:57 am

IIRC a UN study found there was no such a thing as a healthy food (ie Hazelnuts will make you live longer). There are only unhealthy diets and lifestyles.

Try telling a starving man that a hamburger and fries will be bad for him – it won’t be.

But then recently there was that UK kid who ate only potato chips and French fries, lost both his sight and hearing due to malnutrition (vitamin deficiency).

I read a story once that Frank Sinatra was diagnosed with malnutrition because of his habit of eating steak to the exclusion of most everything else.

With the exception of those with specific dietary needs or deficiencies, most diets and dietary advice is bunkum.

Moderation in all things.

My doc asked me how I lost (and kept off) 20kg – my answer “eat less and exercise more.”

You don’t need a dietary guru, but you do need self discipline.

Reply to  Ken Irwin
October 7, 2019 8:10 am

Sheri ==> Thanks for the link to a far better explanation of the bogus “junk food makes boy blind” nonsense.

Readers influenced by the original report (either here or in the mass media0 should read the link provided by Sheri.

Ken Irwin
Reply to  Ken Irwin
October 7, 2019 9:19 am

Sheri, Thanks for the link – next time I quote an MSM article without having delved further – please slap me – hard !

I should know better. Regards – Ken

Reply to  Ken Irwin
October 7, 2019 6:51 pm

Sheri- as the article says the boy involved was not properly or effectively treated for many problems. The mind can have many strange disturbances that are not always fully treatable.

But in the end, the proximate cause of the blindness was severe deficiencies caused by a severely restricted diet. And, it has been know for centuries that not eating a balanced, nutritionally sound diet is the cause of many bodily harms. Based on all the various cultures around the world it doesn’t make much difference exactly what you eat as long as you get the nutrients required in the appropriate amounts.

October 6, 2019 6:28 pm

“In 2017, the 10 leading causes of death were, in rank order: Diseases of heart; Malignant neoplasms [ cancer ] ; Accidents (unintentional injuries); ”

Per JAMA (1999) and John Hopkins (2017) studies, the medical establishment is the 3rd leading cause of death in the US. IMO it’s #1 because those studies only accounted for deaths in hospitals.

Reply to  icisil
October 6, 2019 6:52 pm

icisil ==> There is a lot of talk of ” iatrogenic cause of death” caused by doctors or medical treatment.

I would be interested in links to the studies you cite.

Reply to  Kip Hansen
October 7, 2019 2:52 am

I think the JAMA study is Journal American Medical Association July 26, 2000;284(4):483-5. Link appears to be paywalled. Lots of other sites talk about it.

John Hopkins study

12-minute podcast British Medical Journal: Medical error—the third leading cause of death in the US (podcast)

Reply to  icisil
October 7, 2019 5:31 am

icisil ==> Interesting — controversial when published — iffy, contested data — made a big splash in the popular media, but in the medical media (journals) there was a much more measured response. Reading broadly across the literature, the idea that survived the critiques is that there should be a better way of counting causes of death, particularly those directly caused by or involving “medical errors”, should be improved.

Reply to  icisil
October 7, 2019 3:57 am

yeah and today we had an aussie chat session on aunty abc with a socalled superscience medico who was damned determined to not admit iatrogenic death tolls but real determined to accuse anything non pharma as bunkum even when stats of 20k or more proven pharmas deaths/events and 200 supplement suspected and only 2 deaths
and yes I also did a mental compare to the koolaidKlimate claimants total refusal to consider theyd be a tad wrong

first q to ask..whos got a vested profit interest?
second is wheres the trials data in full how many how long what grouping used etc
and yes I am a boring old fart who susses the details on pharma and alternatives
whos resistant to the media hypes over foods being bad, but do my own research and make my own decisions.
I will say if its processed much I do avoid it

Reply to  ozspeaksup
October 7, 2019 5:39 am

ozspeaksup ==> Thee have been some very good studies done on the question of the efficacy of both multi-vitamins and specific supplements. This JAMA editorial gives a good outline of the current (2016) situation.

Alfred (Cairns)
Reply to  ozspeaksup
October 8, 2019 8:48 pm

Why don’t you write proper English. Are you trying to put down all Australians?

October 6, 2019 6:30 pm

Good points. Mr. Steve Milloy, the Junk Science guy, has been all over this for years. Read one of his books.

Reply to  eck
October 6, 2019 6:54 pm

eck ==> Milloy has done a lot of good work over the years….but Ioannidis gets to the root of the problems .

Reply to  Kip Hansen
October 7, 2019 12:22 pm

Check out the PDF that is supplied for the Ioannidis paper “The Challenge of Reforming Nutritional Epidemiologic Research” – right hand column, page 2: “Resources for some of these studies could have been better spent on unambiguous, directly manageable threats to health such as smoking, lack of exercise, air pollution, or climate change.”

While I don’t disagree with Ioannidis’ main premise of his paper, I think Socrates said something along the lines of true wisdom comes with knowing what you don’t know. If you are an expert on one thing, that doesn’t make you an expert on all things.

October 6, 2019 6:36 pm

Correlation in a complex data set is far more unlikely to be a relation than is non-correlation if you cannot prove a connection through use of the scientific method.

Epidemiology is the study of the *topography of transmission* within a population based on vectors as well as immunity. It specifically is NOT Bacteriology, Virology, Hysterology, Geneology or Spermatology or Histerology. Something must be an epidemic with clear vector transmission patterns to be epidemiology’s territory and Diet Soda is a POINT SOURCE POLLUTION not a transmissible disease.

Do NOT attempt to tell me what words mean in order to alter their meaning to prove your point or theory.

Reply to  Prjindigo
October 6, 2019 6:55 pm

Prjindigo ==> Not quite sure if you are objecting to something I wrote or what…try again?

David Riser
Reply to  Prjindigo
October 6, 2019 7:24 pm

The Dictionary definition of the word Epidemiology: the branch of medicine which deals with the incidence, distribution, and possible control of diseases and other factors relating to health. It is bigger than you think.
Dave Riser

Reply to  Prjindigo
October 7, 2019 12:11 pm

It’s called mission creep in layman’s terms.

October 6, 2019 6:36 pm

The entire field’s a huge problem and riddled with holes, uncertainties and lies, especially when so-called “health” foods and marketing are taken into consideration. Apparently, the “5 a day” marketing from UK grabbermint health was just a figure plucked from thin air because, well.. who the hell REALLY knows?

There are too many variables and individual susceptibility to any kind of disease/chemical being just one. Although smoking may not be the healthiest thing to be doing, several times there have been 100 year old’s still puffing a pack a day. Statistically, if they weren’t, at that very moment they might have been jogging healthily along and been run over by a bus. I know it’s a stretch of the imagination but altogether possible.

Meanwhile, it’s against my doctor’s orders to be run over, hit hard by anything or crash aeroplanes which is probably why I can eat whatever I like and still survive. Screw the health industry, I’m doing okay and not panicking one iota.

Leo Smith
Reply to  ЯΞ√ΩLUT↑☼N
October 6, 2019 10:37 pm

I have been on and am still attending a course intended to provide information of staving off type II diabetes.
I t is NHS funded and has been vetted.
There is an epidemic of type II diabetes.
It is associated with ‘metabolic syndrome’ . See wiki
Basically I have all the bloody symptoms.
Now what is at the heart as the causal agent is allegedly chronic over stimulations of cells by insulin caused by an unremitting high carbohydrate diet and/or chronic stress (adrenaline-> insulin to release fight or flight glucose that isn’t then needed) and lasck of exercise. Especially rapidly absorbed and processed carbohydrates like wheat flours and other starches., Oddly sugars are slightly less bad. As are ‘whole grains’ as they take longer to break down.
Associated with this syndrome and with reasonable causal links being established is
High blood pressure
High cholesterol levels
Coronary artery disease.
and the derivative effects of those.
What transpires is that almost everything you knew about a healthy diet has been found to be wring.
Fat doesn’t make you fat. In fact fat makes you feel satiated. So you eat less.
Carbohydrates satisfy but the the glucose levels spike and crash and you feel hungry all over again.
No one needs carbohydrates to live. Carbohydrates are seriously ADDICTIVE. Carbohydrates – especially milled grains – are cheap, but lethal. Especially in a low exercise society.

Cholesterol in food has no relationship whatsoever to cholesterol levels in your blood. Eat all the eggs you want. And eat all the full fat cheese and yoghurt you want. It will make you less hungry and it doesn’t need added sugar to make it taste nice.
Eat rarely and well. If you can fast for a day, do it
Fruits and veg are not necessarily good for you. Roughage they have and water, but they also have sugars and starches. Green leafy stuff is better than seeds fruits or tubers.
Food labelling is mostly wrong. Like climate change they aren’t trying to make you well, just sell you product.
All that matters is the extremely small letters where they state ‘carbohydrates per 100g’
The government NHS guidelines state
“Base meals on potatoes, bread, rice, pasta or other starchy carbohydrates ”
It is COMPLETELY CONTRADICTED by the course I went on, also sponsored by the NHS.
Recommended carbohydrate intake is around 1kg a day
I am losing weight and feeling great on less than 150g a day. I target 50g.

The problem is that there is an epidemic, and unfortunately the nutritional arena is full of half baked studies. And a lot of food manufacturers lobbying..

I trust the course I went on because it wasn’t just correlations, there were actual biochemical pathways showing how increased glucose transport messes with lipid transports leading to ‘bad cholesterol’ and how all this led to a particular form or weight gain and to other issues like fatigue, that seem to have gone for now.

I think the lessons to be learnt here are that statistics doesn’t prove a damned thing BUT that a correlation may indicate something worth investigating. And if the biochemistry suggests a causal mechanism, then and only them its it time to get rather excited and do more studies.

And finally, everyone has skin in the game. Imagine the disaster if everyone stopped eating bread. And statins. How much money is in those?

Or freshly squeezed orange juice and just whipped up a spinach drink instead. Yuk. But that is probably far more ‘healthy’.
Big business wants you to eat junk food take stains and metformin. Cos they make more money out of you than if you don’t eat anything but a fresh grilled steak, mushrooms fried in butter and a salad on the side, but no fries. Followed by strawberries and cream and some Stilton (but no biscuits)

Reply to  Leo Smith
October 7, 2019 1:43 am

Spot on. The top heavy NHS has not caught up with front line ( correct ) advice.

Reply to  Leo Smith
October 7, 2019 3:16 am

“I have been on and am still attending a course intended to provide information of staving off type II diabetes.”

Research diabetes + bacillus coagulens

Example: https://selfhacked.com/blog/b-coagulans/ (has lots of good references)

1) Decreases Insulin
B. coagulans containing a synbiotic decreased blood insulin levels; HOMA-IR and HOMA-B in pregnant women [1].
2) Improves Blood Lipid Profile
B. coagulans containing a synbiotic reduced TAG, VLDL and elevated GSH levels in pregnant women [2].
B. coagulans reduced total blood cholesterol, LDL-cholesterol, and marginally increased HDL-cholesterol in a small-scale clinical study [3, 4].
3) Is Beneficial in Diabetes
Consumption of synbiotic bread with B. coagulans reduced insulin levels, improved blood lipid profile and increased good cholesterol (HDL-C) in type 2 diabetes (T2D) patients [5, 6, 7].
Similarly, consumption of the synbiotic bread with B. coagulans improved NO and MDA levels in T2D patients [8].
Synbiotic containing B. coagulans improved insulin, hs-CRP, uric acid and plasma total GSH levels in diabetic patients [9].
B. coagulans, inulin and beta-carotene coadministration decreased insulin, HOMA-IR, HOMA-B, triglycerides, VLDL-cholesterol levels, and total-/HDL-cholesterol ratio. This treatment also elevated plasma nitric oxide (NO) and glutathione (GSH) [10].

Reply to  Leo Smith
October 7, 2019 5:23 am

CARB HATERS ANONYMOUS really needs to be started. It’s pathological at this point.

Reply to  Sheri
October 7, 2019 6:32 am

Folks really should make a distinction between simple and complex carbs. There is a significant difference how they’re metabolized. Complex carbs, good; simple carbs, not so much.

Reply to  icisil
October 7, 2019 8:25 am

Please, please tell me that donuts and pastries are complex carbs. Pretty sure they must be…just look at how complicated it is to make a croissant. Like, seriously complex!!!


A C Osborn
Reply to  Leo Smith
October 7, 2019 5:28 am

Government sponsored epidemics.

Jeff in Calgary
Reply to  Leo Smith
October 7, 2019 11:08 am

While I don’t doubt the majority of what you are saying,…

Especially rapidly absorbed and processed carbohydrates like wheat flours and other starches., Oddly sugars are slightly less bad. As are ‘whole grains’ as they take longer to break down.

How exactly does sugar take longer to break down into …sugar… than carbohydrates take to break down into sugar?

October 6, 2019 6:58 pm

At least with epidemiology, the researcher can study a great number of samples in a population, within a relatively small period of time, and calculate statistics based on the study. But in climate science, there is only one Earth, and “climate” is defined as an average of some set of state variables measured over very long periods of time – at least 30 years for example. Thus, to even begin to observe a sufficient number of “samples” of climate change so as to calculate enough relevant statistics to be comparable to an epidemiological study, the the climate researcher would have to wait a period of time measuring in centuries, if not millennia. However imprecise epidemiology is, Newtonian physics is to epidemiology as what epidemiology is to climate science.

This is why climate “scientists” must resort to computer models. It’s not that the use of computer models makes logical sense – a computer simply does what you tell it to do. Climate “scientists” use computer models out of desperation, pretending as if the output really is data because that is the only thing they can do. Doing it the right way would take too long.

Reply to  Kurt
October 6, 2019 7:27 pm

Computer models give the false impression on the masses that because they’re computers, they must be smart and so are their programmers. It’s simply argument from authority. Unless the homeless guy sleeping under the overpass knows precisely what’s going on in the computer model, how could he hope to argue? Same with everyone else that doesn’t have that special degree – they just believe it’s true and often because they’re lazy, but it doesn’t take much to comprehend the basics.

Although it’s a stretch using The Goreacle’s claim: “You don’t have to be a Climate Scientist™©® to see Climate Change™©®, you just have to look out the window”, it’s not altogether difficult after having studied the history and actual data the Green Blob refuses to show you to see right through the fog of lies.

This is why advertising over the last 20-40 years was so geared toward “clinically/scientifically proven/analysed/tested” sputter because they wanted everyone to simply believe their product claim was of the highest authority, in essence making most believe the hype right off the bat when the truth is usually the opposite.

Reply to  Kurt
October 7, 2019 7:21 pm

Climate models have been questioned since they were invented, and with good reason. But Edward Lorenz had pinpointed the the total failure more or less on the first try. Systems of partial differential equations often cannot be solved be cause no solution can be found. The result is that tiny errors compound and can lead to entirely different answers. The system is chaotic but it can never be exactly replicated. Even computer simulations started with the exact same number can end up with entirely different answers. His beautiful butter fly rendition demonstrates that the result never exactly follows the same path twice.

The Hawkmoth idea and Christoper Essex’s “computer epsilon” demonstration show that every computer has a limit where two numbers cannot be able to be shown to be different- the measurement scale can’t resolve the difference. With a chaotic system of partial differential equations it is inevitable that the math model cannot repeat and can never give any valid predictions.

The idea of using multiple runs of a faulty model to make a prediction is just ridiculous. Statistics deals with things that can measured and have random errors. Any analysis of a bunch of unrelated results from multiple runs of unpredictable models can’t give any useful results.

October 6, 2019 7:13 pm

One big take-away is the telling quote from “Saving Science” (Go back and read the whole thing.)

In the future, the most valuable science institutions will be closely linked to the people and places whose urgent problems need to be solved; they will incentivize scientists to care about the problems more than the production of knowledge. They will link research agendas to the quest for improved solutions — often technological ones — rather than to understanding for its own sake. The science they produce will be of higher quality, because it will have to be.

Sounds great, who would be against better science, and using it to solve real-world problems.
But again, it is unfortunately, nonsense. If you want to solve a problem, you use engineering and technology. But that is too easy. In the larger sense of society at large, problem solving is in the domain of public policy and politics. So politicians put in place policies which feel good and garner votes. When it all goes south, they blame scientists, and call for improved scientific research, as to imply there was something wrong with current research. Then they call for evidence based policy making. Sure.
Case 1: California needs water to support a growing population, provide for agriculture, and enhance flood control.
Solution A) Continue building dams for reservoirs and flood control projects.
Solution B) Halt all current projects, cancel all projects in the planning stages. Divert water from agriculture to the Pacific ocean to “save” a bait fish. Blame the disaster on “Global Warming” or “Climate Change”.
Case 2: It is thought that the continued use of fossil fuels could cause a major problem and cannot continue.
Solution A) Nuclear Power. Nuclear power has problems? We already know how to approach upcoming problems. Improved scientific research, evidence based policy making!. (sound familiar?) Solve the problems and get on with it.
Solution B) “Environmentally correct”, which is to say ideologically correct solutions which cost billions, and do nothing. But these solutions sound great while enriching well connected crony capitalists.

So we have gone nowhere. The problems in public policy, like nutrition science, like epidemiology, like climate science, are all the same. Who is giving out the money, when, where and why, to whom exactly, and in exchange for what, exactly. Here in the US, the politicians control the money. The results are what you would expect.

A Counterpoint For Comparison:
Problem: Modern airliners use too much fuel, cost too much, and are too expensive to build.
Solution A) Scientists do their thing. Engineers do their thing. Boeing 777.
Solution B) Boeing 787.
There is nothing wrong with the science!

Reply to  TonyL
October 7, 2019 2:57 am


I too read that excerpt and did not recognize the description as science. It seemed to me to be almost the opposite of what I understand by the word.

A C Osborn
Reply to  TonyL
October 7, 2019 5:34 am

Not just Government are dishing out the cash, Big Pharma and Big Food (diets/fads etc) are also controlling the cash and research.

Reply to  TonyL
October 7, 2019 11:17 am

This stuck out for me too. The ‘Science should be focusing on real world problems’ approach is not science. It is application of known parameters to solve problems. Under this approach, we would not have progressed from taking the handle off the pump to control cholera. From a lot of further research, we got treatments for cholera, vaccine, safe water supplies and so on. The guys who found microbes and studied them just because they were new and interesting were behind this.

This is calling what my brother was doing in refugee camps science: he was making a cheap simple handwashing station to put by the bogs to control GI diseases. The science preceded the practical application.

Reply to  Kip Hansen
October 7, 2019 7:30 pm

Kip, presently the problem of how to help the poor in Africa and Asia is being handily solved by the people themselves, where they are allowed to. Primarily coal fired power stations and limited range power grids are already starting to work effectively, if they are not stopped by by the UN propaganda or the false promises of more “help” from other people seeking power.

October 6, 2019 7:18 pm

“That’s the Wiki speaking on the subject.”

If you mean Wikipedia, say so. “The Wiki” is ambiguous.

Reply to  jorgekafkazar
October 6, 2019 7:26 pm

jorgekafkazar ==> In ,my defense, it is clearly A WEB LINK, which leads to a specific Wikipedia article, thus not ambiguous. Moving one’s cursor over the link in most browsers reveals the link URL.

Nicholas McGinley
October 6, 2019 7:23 pm

One of the basic flaws in just about every study relating to what people eat, is that just about all of them rely on self reporting, which is by definition highly subjective.
People are notoriously bad at accurate self reporting of even the most basic details of their diets.

Reply to  Nicholas McGinley
October 6, 2019 7:35 pm

Nicholas McGinley ==> Yes, that is clearly correct — a point that Ioannidis (and many many others) point out repeatedly. It is very difficult to collect real data on individual diets — even when they collect them on a daily basis….consider “Lunch: Ham sandwich”. (two slices of Wonderbread and a single slice of lunch-meat ham? or a 12″ sub from your favorite deli with 1/4 pound of sliced smoked ham) Lots of problems.

Of course, this essay covers that in another way — the topic — diet — is very hard to collect a good data set on….like climate science….

David Riser
October 6, 2019 7:26 pm

Good article!
Dave Riser

Reply to  David Riser
October 6, 2019 7:35 pm

Dave ==> Thanks, glad you liked it.

October 6, 2019 7:38 pm

Nice article. For years I have considered nutritional studies to be the worst amongst a group of very suspect disciplines (e.g., alternative health care, the art of inventing diseases to create markets for new medications, AGW, the study of marginally dangerous substances). I understand a little better why nutritional studies are continually generating such contradictory nonsense. Thanks.

Reply to  BCBill
October 6, 2019 7:40 pm

BCBill ==> Glad you found this useful — most of the credit goes to John Ioannidis (and others before him, Milloy, Aaron Carroll, etc)

October 6, 2019 7:39 pm

I have long suspected that artificial sweeteners cause AUTISM.
It was never vaccines-it’s chemicals babies are exposed to both BEFORE and after birth.
P.S. Just in case you have been confused by modern pop culture:
Those things (zygotes) that grow inside the uterus after fertilization which later ‘turn out to be babies’ are in fact: human fetuses. They are not tumors, growths, polar bears or otherwise undefined ‘mysterious and threatening’ alien bodies.
To believe otherwise is to be ANTI-SCIENCE.

Reply to  Barbee
October 6, 2019 8:50 pm

As I understand it, I don’t think it was the vaccines that was ever thought to be the problem, but rather the aluminium adjuvants they use to trigger the immune system, thus increasing the efficiency of the vaccine. Aluminium is considered to be a neurotoxin.

Reply to  sparko
October 7, 2019 3:57 am

Plus mercury (another neuro-toxin).

Aborted human fetal cell lines were authorized for use in vaccines about 1979. Autism incidence began to noticeably increase in 1980. Two concerns are insertional mutagenesis (fetal human DNA incorporates into the child’s DNA causing mutations) and autoimmune disease (immune system attacks own body). Italian researchers recently found complete human genome DNA in a MMR vaccine.

Complete Human Genome DNA Found in Contaminated MMR Vaccine by Italian Researchers

Autism incidence also has increased co-incident with the CDC’s expanding vaccine schedule. Does it make sense to give babies under 1-year old 26 vaccines with toxic adjuvants? Any sensible person can see the potential for problems. The CDC is a private company (not a government agency) with more than 20 vaccine patents providing profits from vaccine sales. How did the system get so corrupt that they are the ones recommending vaccine schedules?

Examining RFK Jr.’s claim that the CDC “Owns over 20 vaccine patents.”

Jeff Alberts
Reply to  icisil
October 7, 2019 6:40 am

How does that relate to the expansion of the autism spectrum, resulting in increased reporting?

Natalie Gordon
Reply to  icisil
October 7, 2019 6:57 am

DNA is found everywhere and in anything. We shed it constantly. Sweep your desk and search for DNA, you’ll find DNA. Our bodies are delightfully primed to hunt down and destroy foreign DNA. I cannot imagine how DNA in vaccine could cause anything unless it is hidden inside a functioning virus. It would be like blaming random roadside debris from other old cars on the side of the road for design flaws of car being made in Chicago. I blame Star Trek for the general lack of understanding what DNA is and what it can and cannot do.

Reply to  Natalie Gordon
October 7, 2019 8:16 am

Natalie ==> Yes, in addition to innumeracy, there is a great deal of un- and mis-education involved in the nutty ideas that sprout and take root on social media and the Internet.

This one problem in society in general is shocking to me . . . . just how misinformed people are, even those with advanced degrees can lack even basic understanding of scientific issues.

Reply to  Natalie Gordon
October 7, 2019 8:29 am

“Our bodies are delightfully primed to hunt down and destroy foreign DNA.”

Not necessarily the case with foreign human DNA. The WHO recommends no more than 10 ng of human derived fetal DNA contaminants per vaccine dose. The rubella portion of the MMR vaccine contains around 175 ng.

Reply to  Natalie Gordon
October 7, 2019 9:17 am

Multiplying that kind of foreign human DNA contamination according to the number of vaccines made from aborted human fetal cell lines (see list below) potentially can create real problems in developing less-than-year-old toddlers, IMO.

comment image

Reply to  Natalie Gordon
October 7, 2019 10:25 am

Merck’s MMR II vaccine (as well as the chickenpox, Pentacel ,and all Hep-A containing vaccines) is manufactured using human fetal cell lines and are heavily contaminated with human fetal DNA from the production process. Levels in our children can reach up to 5 ng/ml after vaccination, depending on the age, weight and blood volume of the child. That level is known to activate Toll-like receptor 9 (TLR9), which can cause autoimmune attacks.

To illustrate the autoimmune capability of very small amounts of fetal DNA, consider this: labor is triggered by fetal DNA from the baby that builds up in the mother’s bloodstream, triggering a massive immune rejection of the baby. This is labor.

It works like this: fetal DNA fragments[i] from a baby with about 300 base pairs in length are found in a pregnant mother’s serum. When they reach between 0.46– 5.08 ng/mL, they trigger labor via the TLR9 mechanism[ii]. The corresponding blood levels are 0.22 ng/ml and 3.12 ng/ml. The fetal DNA levels in a child after being injected with fetal-manufactured vaccines reach the same level that triggers autoimmune rejection of baby by mother.

Anyone who says that the fetal DNA contaminating our vaccines is harmless either does not know anything about immunity and Toll- like receptors or they are not telling the truth.


October 6, 2019 7:45 pm

“Nutritional research may have adversely affected the public perception of science.”

Most likely it has caused public perception of science to more closely resemble the actuality.

Global Cooling
October 6, 2019 8:38 pm

Words mean a lot in the world of propaganda. I don’t like the word chaos here, because it is connected to randomness. A system can be fully deterministic but statistical analysis can’t reveal how it works. See for example double pendulumcomment image.

Someone in this forum said that climate models don’t use initial conditions. Problem is not just initial conditions. In complex systems a small change in inputs, feedback or calculations can result in a large change in the output. This is a property of non-linear functions. For example, if the output is a function of input’s forth power, the curve in the visualization does not show fitness in simple linear statistics.

Climate system has many significant contributing factors. Still we try to use linear correlations between CO2, sun spots, .. to global average temperature as a proof to our propaganda.

Reply to  Global Cooling
October 7, 2019 5:50 am

Global Cooling ==> “Chaos”and “chaotic” in the essay are specifically flagged as having the special meaning used in Chaos Theory and linked to simple-ish explanatory Wiki articles. It is unfortunate that the field of the study of nonlinear dynamical systems with extreme sensitivity to initial conditions (and many other shared pathologies) was named “Chaos”. There is no going back now….language change sand new words and new meanings for old words is part and parcel of the evolution of language.

“Someone in this forum said that climate models don’t use initial conditions.” Someone did say that — but it is entirely false. See the links to the NCAR 40 Earth’s project and my essay discussing the implications of that project at Climate Etc. .

Peter Hannan
October 6, 2019 8:49 pm

I’ve been vegetarian for nearly 45 years, so I’ve probably paid more attention to nutrition and health than many laypeople. I’ve seen the fad diets and the scares come and go. A pretty obvious point about the hazelnuts idea is this: what kind of diet do people who eat hazelnuts regularly have? I’d guess that many will have a fairly high proportion of plant foods, other nuts and seeds, grains and pulses, fruit and vegetables – a good intake of fibre, vitamins and minerals. Is that controlled for? Hazelnuts may well just be a proxy for such a diet. Your comparison with climate science is very helpful, thank you.

Reply to  Peter Hannan
October 7, 2019 5:55 am

Peter ==> You have the right idea — who the heck actually eats 12 hazelnuts a day? “Hazelnuts may well just be a proxy for such a diet.” And regardless, does any well-rounded, well-varied diet that supplies adequate (but not too many) calories and all necessary vitamins, minerals and trace elements improve or detract from health status compared to another differing but similarly complete diet?

Reply to  Kip Hansen
October 8, 2019 8:17 am

Hazelnut consumption could be a proxy for many things, including culture, which often means groups sharing common genetics.

It was once thought the longevity of the Japanese was due to their heavy diet of fish. I think most would now say that the longevity was more closely linked to the genetics of the Japanese.

Reply to  Peter Hannan
October 7, 2019 11:30 am

Yes, and it seems that hazelnut consumption might also be a proxy for geographic location too:
Italy : 1.75Kg/yr/capita
Russian Federation: 0.15 Kg/yr/capita
Brazil : 0.05 Kg/yr/capita

The wonders of the internet (and courtesy of the ‘International nut and dried Fruit’ org) ! :

Reply to  Kip Hansen
October 9, 2019 1:49 pm

Yes, but they have a special use in flavouring dishes of tripe 🙂
In the UK, at the moment, this recipe is being lapped up the journalists:

Reply to  Peter Hannan
October 8, 2019 12:50 am

When I was a graduate student in Oregon, I ate a lot of hazelnuts (aka filberts, not be be confused with Dilbert), because they were free to harvest from the uni farm and made a decent pesto. Irritating hairs in the bracts and tough to crack (but nothing compared to macadamia nuts) made this diet contingent on low grad student salary. I’m struggling, though, to see how any average US citizen or resident could keep track of their filbert dosage. They are pretty rare in mixed nuts and not commonly available on their own, at least not compared to almonds and other mass produced nuts. I think this is probably a non-proxy of any sort and an amusing example of self-reported diets.

John Pickens
October 6, 2019 9:04 pm

Excellent article. I have read several reports of studies showing that consumption of drinks with artificial sweeteners cause weight gain.
My immediate reaction was that people who are drinking artificially sweetened drinks were probably trying to reduce caloric intake because they have a weight problem. Causation was simply reversed in these studies.
They were nonsense.

Reply to  John Pickens
October 7, 2019 6:00 am

John Pickens ==> Reverse- and retro-causation are not rare in the sciences today — for the very reasons pointed out by Ioannidis in Nutritional Epidemiology and in Climate Science by myself (here and here).

The telling of “Just So Stories” is also rampant — read almost any evolutionary biology paper.

Crispin in Waterloo
October 6, 2019 9:13 pm

““Risk factors” or determinants are correlational and not necessarily causal, because correlation does not prove causation.”

What is required in this context is the term “attributed”. The point of the Global Burden of Disease (GBD) exercise is to attribute certain risk to certain factors. It is based on checking the condition of people who have died before attaining the age of 86 (which means prematurely).

Health policy is guided by attributable risk, not analysed causes. “Cause” is what you discuss in the field of medicine and murder investigations. Public health and medicine are only connected by telephone lines.

This unsubtle differentiation is usually lost on the public which is highly concerned with medicine (their own health) and not managing public health policy. GBD is useful for shifting budgets to priority areas without too much concern for the disease consequence of any particular individual of doing or not doing so.

Consider then, the impact of having an attributable risk assigned to possible contributing causes, v.s. having a medical diagnosis of a cause of disease. They are simply not on the same page, or in the same chapter of the book of life.

For Joe Public the fact that a hurricane is attributed to the elevation of CO2 concentration in the atmosphere is not more helpful and no more truthful than attributing someone’s heart attack to “air pollution”. You can make such an attribution but it doesn’t mean squat.

Removing 100% of all air pollution will no doubt have some beneficial medical consequences for the whole population, on average, but it can no more prevent some particular person having a heart attack than mounting all staircases with the left foot first instead of the right. There may be a correlation, and one can make an attribution, but it ends there. No one can prove that someone’s heart attack was caused by air pollution.

If you really want to do something fundamental about diet and health, eat only things for which your teeth are adapted, which is nuts, grains and fruits. At least that would be a diet selection based on physiology. Humans are not naturally meat eaters which is why we don’t have the teeth for it. We are not even adapted to be omnivores. Check the dentition of a real one.

When we get the food right we will all live happy and healthy to be 150 years old.

Reply to  Crispin in Waterloo
October 7, 2019 6:08 am

Crispin in Waterloo ==> Quite right — Global Burden of Disease (GBD) is a public health concept…and in my book, is often illusory and not based on reality. It is the problem of “attribution” in which the scientific/medical standards for determining attribution are vague and highly dependent on opinion and bias — Ioannidis concept that they are “reflect[ing] almost exclusively the magnitude of the cumulative biases” and are simply in support of “expert-driven [already existing] guidelines shaped by advocates dictat[ing] what primary studies should report.”

Thus we see that massive international public health push against PM2.5 — an unproven risk with an unknown effect.

Reply to  Crispin in Waterloo
October 8, 2019 6:03 am

Crispin- GBD is primarily a pretext for distributing money. Whether by insurance companies trying to estimate risks or governments to fund politically useful projects such as food for the poor or medical research. The attribution of an effect through invalid statistics is a highly political process.

“eat only things for which your teeth are adapted, which is nuts, grains and fruits”- humans and all the related great apes ate meat(including insects) when they could get it. It is a part of our diet and essential for easily getting small amounts of important nutrients that are not found in a wholly vegetarian diet. Even in a vegetarian diet many of the important micro nutrients come from insects included with the vegetables. Flour is a good example. After it was discovered that flour contained insect fragments, and people were wealthy enough to pay for more pure flour, some people, primarily children, started to develop nutritional deficiencies. The cure was a search for supplements and vitamins to take instead of eating “meat” was on in full, non-scientific, charge.

Reply to  Crispin in Waterloo
October 8, 2019 8:49 am

Our teeth are very well adapted to rip meat and chew it. No problem at all for the vast majority of us.

If humans were not intended to eat meat, our bodies would produce B12. They don’t. It is very difficult to obtain enough from vegetarian diets, and our ancestors would have had serious health problems attempting to do so. Most vegans must take B12 supplements. I know of no vegetarian cultures in history, except for a few unsuccessful attempts based on religious beliefs. None because it was their ‘natural’ diet.

Secondly, our digestive systems, including the stomach (and having only one) and the length of our intestinal tract, are not optimized for vegetarian diets.

We are omnivores, and are not the only primates who are.

Len Jay
October 6, 2019 9:33 pm

What doesn’t seem to be taken into account here (or I may have missed it) is the pressure bought to bear by purveyors of a particular product to promote that product as healthy, or the competitors of the product to make it a killer in the public’s eyes.
A case in point is tobacco smoking. I doubt any thinking person would deny that inhaling hot poison laden smoke into one’s lungs several times a day is a healthy practice, but “thank you for smoking”.
That any attempt to convince the public that it is a healthy practice can only be made by those with an interest in selling the stuff.
Great wars rage over such matters. As mentioned above coffee versus no coffee – that is coffee producers versus say tea producers. Such wars must have a substantial effect on the perceptions of the public at the time that research is done.

Reply to  Len Jay
October 7, 2019 6:13 am

Len Jay ==> If it were only that simple, economic gain as the main driving factor, we could sort these things out easily. Unfortunately, there are many more factors involved — and each side in these individual “food fights” accuses the other side of being in it for monetary gain!

How science fields go awry is a very interesting and important question in modern society — we see this in crazy things like the anti-vaccination movement, in climate science, in nutritional epidemiology, in social psychology — and when some brave soul steps up to the plate to try and put things right — she is viciously attacked by otherwise brilliant practitioners of science representing the status quo.

Reply to  Kip Hansen
October 8, 2019 1:18 am

Hi Kip – If you are willing to be so bold, I would be interested in seeing your take on the vaxers. WUWT probably wouldn’t be the best place for such a discussion, but I find myself more and more skeptical of the vaccinate everyone for everything school. It seems the Measles Vaccine isn’t very good and one reason for the resurgence of measles. Chicken Pox may be better endured as mild disease as a child for a lesser chance of shingles when you are older, or at least the British seem to think so. What’s up with whooping cough I’m not sure, but a suboptimal vaccine seems to be part of it. Like climate, though, vaccination is totally polarised and you are either for the science or a denier. I’m no vaccination denier, though, so forget what I just said. I deny I said it. As long as science is a political football it can’t be reasonably discussed.

Christopher Hanley
October 6, 2019 9:55 pm

Health risks and ‘climate change’ are frequently epidemiologically linked, just recently: West Nile virus, dengue fever, Lyme disease, Rocky Mountain spotted fever, dangerous bacterial infections, increase in Type 2 diabetes, respiratory problems and stroke, even more frequent car crashes (duh).
The term ‘carbon pollution’ is a deliberate conflation of CO2 and genuine air pollution (particulates and chemicals) to discombobulate the general public.
A straight report in the Independent (December 2018) claimed: ‘climate change likely to make us more stupid’, over a large photo of a number of cooling towers belching steam — they obviously have the direction of causality arse-about.
The causal relationship between cigarette smoking and lung cancer is often used as an analogue for a supposed causal and direction relationship between CO2 and ‘climate change’.
However epidemiologically lung cancer was a relatively rare disease before tobacco, particularly cigarette, smoking became common:
Of course the climate has been changing forever and it is impossible to identify the final effect of increasing human CO2 emissions — if any — and the IPCC proclamation of causality is an example of the single cause fallacy and pure guesswork.

Reply to  Christopher Hanley
October 7, 2019 6:21 am

Christopher Hanley ==> I have written several times on the falsity of the claims that climate change will increase the spread of “tropical” diseases. ( here, here, here, and here ).

Bogus correlations can be found anytime two things seem to be increasing coincidentally.

October 6, 2019 9:59 pm

Hazelnuts are quite formidable food. They are one of the lowest in PUFA , with 6 gr/100gr nuts (only macadamia lower among commonly sold). They are 44.8 gr/100 gr Monosaturated fat & 3.7 gr/100 gr saturated fat.

Their 17.5 gr/100 gr carbohydrate are 14.2 gr complex carbs & 3.3 gr “sugar”. Fiber is 14.2 gr/100 gr insoluble & 6.5 gr/100 gr soluble fiber.

Why can they be conducive to survival?

For one thing they have 13.8 – 22.2 g mg boron/ kg. This boron is complexed with some of the hazelnut “sugar’s” fructose, as fructo-borate. Fructose improves boron entry into cells via a sugar transporter & avoids the way we pee away 90% of non-complexed boron.

In human well-being terms decent boron levels from fructose-borate lower several inflammatory factors (ex: IL6, IL1B, CRP).

Boron upreguates the transcription factor p53, which allows cells to undergo programmed cell death (apoptosis) & restorative processes get going. Cells normally cycle p53 with it’s degradation counterpart (Mdm2) about every 20 minutes; it (p53) doesn’t stop cell division or inordinately induce apoptosis (unless stabilized by it’s response element, the isoform p72 Mdm2).

When p53 properly active (not hyper-active) it also upregulates our useful DNA damage checkpoint, which stymies abnormal cells (they get stuck in G1 cell cycle, before mitosis sets up). Functional p53 boosts DNA base excision repair machinery to deal with aberrant insertions/deletions/mismatches.

Fructose- borate’s boron stops the breakdown (catabolism) of vitamin D3 (& boosts 25 hyroxylation); so inside cell’s get more D3. The import of D3 here is that it arrests cell cycles in a way that allows programmed cell death (apoptosis).

I eat hazelnuts & buy them preferentially over other nuts.

Rod Evans
Reply to  gringojay
October 7, 2019 12:55 am

gringojay, I too eat hazel nuts but not as a presupposed desire for some health benefit. I eat them when my hazel trees provide them. A time of harvest, very briefly around September here in the UK. I have to say I like the activity of picking them and breaking out the nuts. Sadly I am always beaten to the best and the bulk of the natural bounty by the squirrels (greys) they manage to have over 90% of the crop.
Now the question I have is this.
With all the benefits naturally contained within the hazel nuts you list, why don’t my squirrel friends live forever?
Squirrels also do a great job of planting hazel. That squirrel activity alone, provides me with continuous extra exercise, weeding them out of the flower beds. That is another benefit of the hazel nuts, not covered in your nicely summarised list of hazel nuts positives.
I guess my point is, you can not eat yourself to immortality, but you can eat yourself to an early death.
My simple advice is, eat when you are hungry not when the clock says it is time to eat. Home grown and home prepared food, is best. Preparing and cooking your own food provides additional exercise, plus you know what you are ingesting.

Reply to  gringojay
October 7, 2019 6:41 am

gringojay ==> Well, at least there is something that hazelnuts might do to improve longevity.

However, “Calcium fructoborate (CF), a natural sugar-borate ester found in fresh fruits and vegetables, is a source of soluble boron.” (NIH). Well rounded diets including one’s fruits and veggies should supply adequate dietary amounts of CF.

Reply to  Kip Hansen
October 7, 2019 1:34 pm

Kip, Hi – Although raisins have 28 ppm total boron only 5 ppm of that is fructoborate, so 17.9% of boron. Compare them to dry figs with 35 ppm total boron having 15 ppm of that as fructoborate, so 42.9% is fructoborate.

An example for “tonic” herbs is dandelion root. It has 200 ppm total boron & although only 80 ppm of that is fructoborate, that means 38.3% s fructoborate.

Another food I have data about is tomato paste. It has 20 ppm total boron of which 7 ppm is fructoborate, equal to 35%. So, when we hear the Mediterranean Diet touted in terms of fat/carbohydrate/protein as healthy maybe the fructoborate ingestion is another factor.

Sorry, but no comparable ppm data for hazelnuts.

Reply to  gringojay
October 7, 2019 2:23 pm

Again Kip, since you read own posts:

The traditional Japanese diet is often cited for exceptional survival. Miso has an average of 5.17 mg boron per kg; while Natto averages 7.35 mg boron per kg.

In the dynamic of healthy longevity boron’s dynamic with transcription factor p53 is non-linear. There are states (ex: short telomeres &/or old age elevated IGF1) that naturally raise p53, which p53 due to feedback loops in turn provokes extra (up regulation) of it’s degradation isoform p90 Mdm2 ; so the trend ends up as depressed p53.

Results of depressed p53 in the aged with untimely high IGF1 means increased cell division. In the context of sparse p53 triggered programmed cell death (apoptosis) there are many old cells that secrete what amount to signal proteins to other cells/organs (ex. metallo-proteins that degrade human tissue).

Traditional Japanese diet’s benefit confounds diet analysis that looks at it’s high rice carbohydrate content as proportion relative to protein & fat content. I think nutritional composition does not adequately explain healthy aging.

Reply to  grngojay
October 8, 2019 9:52 am

So Kip, – as for why, if eating hazelnuts/produce with fructoborate boosting p53 cellular dynamism, how can people still get ill.

For staters cancer afflicts p53. Cancer cells do not undergo programmed cell death (apoptosis) because their p53 dynamic is uncommon. For starters cancer can cause p53 to “pool” in a non-functional state & thus continue it’s life.

Virus genes afflict p53. Most often viruses induce over expression of the normal p53 degrader Mdm2. Some viruses secrete a protein that binds p53 with other proteins causing an accelerated degradation rate of p53. Viruses knock down the functional level of p53 & they can continue along.

Mutation of p53 also occurs in the human germ line of some of our general population. Mutant p53 coagulates & contact with certain protein blocks it’s (p53) linking to DNA. The result is reduced switching on of key genes.

Depending on the existence of normal type of p53 being around (since a single gene has 2 copies/alleles) that can compensate for mutant p53 activity. However, quite often in the course of time (ex: as person ages) the normal p53 loses out being expressed &, in the case of cancer then the cancer gets it’s way more.

I’ll skip p53 & RNA details for brevity. And point out that cancer’s mutant p53 being unable to program it (cancer) for cell death (apoptosis) that 21 of 36 chemotherapy drugs knock down biogenesis of ribosomes.

Biogenesis of ribosomes (up to 10 million per mammalian cell; humans make about 400 ribosomal RNA copies from about 80 ribosomal proteins using our 4 ribosomal RNA). It is biogenesis of ribosomes with increased protein synthesis that boosts cell growth/proliferation (of cancer in this context).

In other words ribosomal biogenesis gives cancer cells their advantage. Based on the number of ribosome/ cell there are different messengers mRNA translated & lots of mRNA with easy translation code are oncogenes, cell cycle regulators, survivin, & growth factors.

Nuts/produce with high fructoborate sustaining a normal p53 housekeeping function works to reduce ribosome biogenesis when there it reaches a threshold (stressing natural dynamics). But of course this too is not a linear constant. Forexample, if among the mRNA downregulated functional tumor suppressors can also be lost; while at the same time feedback of reduced ribosomal number naturally induces a reduction in normal p53 messenger mRNA.

Unedited. Pardon any errors.

Beta Blocker
October 6, 2019 10:00 pm

Research has demonstrated that if you continue to breath air containing 415 ppm carbon dioxide, sooner or later, you’ll be dead.

October 6, 2019 10:59 pm

Breaking news from epidemiological research:

Everyone who has ever eaten carrots dies.

Jeff Alberts
Reply to  stinkerp
October 7, 2019 6:45 am

Birth is the leading cause of death.

Jon Jewett
Reply to  stinkerp
October 7, 2019 1:39 pm

And much worse for broccoli!

October 6, 2019 11:26 pm

We are all living longer now than we use too compared say to 100 years ago.
So overall we must be eating better food?.

Reply to  Mark.R
October 7, 2019 6:47 am

Mark -==> A not trivial idea — but mostly it is believed to be the improvement in “enough” food and a wider range of foods. For the very poor, it is an increase in proteins (mostly meat) adequate for development as children and to maintain adult bodies in good condition. In richer countries, over the last century or so, people have had adequate protein but have added a wider variety of vegetables and fruits not only in season but throughout the year.

The advances in public health, sanitation and the defeat or control of so many common diseases has had an even larger effect.

Rod Evans
October 7, 2019 1:13 am

Thanks Kip a good read, and remakes the point about focused studies always seek to establish the presupposed causal links, think Co2. What is worse, is when the advocates of advanced dogma, again think Co2 focus so narrowly on their preferred cause contender they ignore the real candidates. The result is what we are seeing now in the climate change industry.

Reply to  Rod Evans
October 7, 2019 6:50 am

Rod ==> Spot on — focusing on finding the “WITCHES” in the village blamed for deaths of children prevented seeing that the privy was next to the community water well.

Jan de Jong
October 7, 2019 2:19 am

Say there are some 10 million scientists (and ‘scientists’) in the world who each must produce a paper once a year. No wonder the vast majority of those papers are BS and the result of ideology and anecdote or faulty statistics on big data.

Rod Evans
Reply to  Jan de Jong
October 7, 2019 2:35 am

I think the phrase that covers the majority of papers is “never mind the quality, feel the width”
One sobering thought, there are doctoral papers written but never read, in one of them there may be the essence of some scientific revelation, something we would all benefit from. Unfortunately it is like the final scene in Raiders of the lost Ark” that crucial document is well hidden among the millions and millions of other, never to be read papers.

Reply to  Jan de Jong
October 7, 2019 6:53 am

Jan and Rod ==> The pressure to publish — “publish or perish” — has been a negative factor in academia for a generation or two. The absolute flood of journal papers in almost all fields is a polluting factor with the GOOD being lost in the flood of the mediocre.

Don K
October 7, 2019 3:41 am

Kip – Another excellent article. A couple of possibly substantive points.

1. I’m sure that both you and John Ioannidis are aware that there is another problem with statistical analysis of human populations over and above those cited. Unlike electrons or other obscure tiny objects, people are heterogeneous. It’s perfectly possible for a food or other lifestyle element to be beneficial to most, but detrimental or even lethal to a few. An example would be peanuts. They are likely beneficial and nutrient rich for most. But a small percentage of the population is terribly allergic. BTW, if a “nutrient” kills people quickly and efficiently, do its victims get counted as negative results in nutrition surveys? It’s not like they’ll be answering questionnaires. Anyway, I’m thinking that what works in statistical mechanics may not scale down to human populations.

2. While the concept that order of calculation can seriously affect results sounds plausible, I can’t think of an obvious example. I’m thinking that in practice maybe it’s uncommon and is treated as a defect when found. In my working days, I looked at thousands or maybe tens or thousands of software bugs. Except for a handful of cases of inadvertent major loss of precision, I can’t recall any case where mathematically/logically correct operations yielded order dependent results. Not saying it can’t happen. Maybe my frequently faulty memory isn’t working right.

Quibble: The Hansen, Westside Highway story is, so far as I know, correct. But it’s not clear what road Hansen was talking about. The modern “Westside Highway” along the Hudson (parts of West St, 11th Ave, 12th Ave) is mostly only a few meters above sea level. But there was an elevated Westside Highway built in the 1930s that wasn’t completely removed until the late 1980s. Could it still have been in place at the time of the interview?

Don K
Reply to  Don K
October 7, 2019 3:48 am

Drat. I meant to delete the Westside Highway quibble. Doesn’t matter. Hansen was clearly really wrong about future sea level rise. It doesn’t matter whether he was very wrong and even more than just very wrong.

Reply to  Don K
October 7, 2019 7:02 am

Don K ==> Read the Donahue and Caldwell paper on order of processing in climate models (they are specifically referring to GCMs). You can just look at their little PowerPoint if you wish.

The Order of Processing problem arises because all the elements of the climate system are co-dependent — humidity at the ocean/atmosphere level depends on winds (or lack of), temperature, salinity, air pressure — each of which may depend on the others simultaneously. Temperature depends on winds, winds depend on temperature gradients, etc etc. GCMs have to tackle one element first — but which one?

Don K
Reply to  Kip Hansen
October 7, 2019 12:10 pm

Thanks Kip. Actually, you don’t HAVE to process the elements sequentially. But if you process them “in parallel”, you have to figure out how to combine the results of disparate computations That’s not a big deal for predicting orbits or trajectories, etc. So “they” do “parallel computation” in those and many other models. But, based on the Donahue and Caldwell paper, it is a problem in climate modeling. Sequential processing seems to finesse the combination issue at the cost of computing some state parameters based on a different state than others. I suspect that’s not an especially good idea. But what do I know?

Alan McIntire
October 7, 2019 4:07 am

After reading an article about the “Arcsine Rule” on William Briggs’ blog, I finally downloaded the free “R” program and started learning a little about handling linear regressions.




Beware of ANY data relying on cumulative sums of anything- cumulative number of people hit by cars while running across freeways versus cumulative sales of cottage cheese over that same period, and you’re LIKELY to get a “highly significant” correlation- either positive or negative. If the data is “detrended”- instead of cumulative totals, you compare yearly or monthly totals, the correlation will drop drastically.

The same works with CO2 and Tempratures- both are “cumulative” statistics, and naturally show a “significant” correlation. Jamal Munshi has published several papers on this very topic, showing relationship between fossil fuel emissions and CO2 in the oceans is virtually zero when the cumulative sums are “detrended”, and the correlation between year to year changes is run.

A C Osborn
Reply to  Alan McIntire
October 7, 2019 5:37 am

Alan, the only correlation between CO2 & Temps is after the miriad of adjustments have been made.
Which is precisely what they are for.

Reply to  Alan McIntire
October 7, 2019 7:09 am

Alan ==> Thanks for the links on time series analysis. You might want to check out LSE’s CATS program.

October 7, 2019 4:23 am

The most influential epidemiology study was the link between high fat diets and heart disease. It lead to the food pyramid emphasizing grains in the diet and replacement of fat with carbs in a host of proceeded foods. (Low fat sugar rich Pop Tarts would be classified as heart healthy food.) Soon after, the rate of type 2 diabetes went up by more than 50% along with rates of obesity. Turns out that while the high fat diet epidemiology revealed heart disease was down, mortality rates were unchanged and “deaths at the hand of another were increased which compensated the heart disease affect. Which made me conclude there really may be some wisdom in that old expression, “fat, dumb and happy”.

Natalie Gordon
Reply to  Sean
October 7, 2019 6:51 am

There is an additional confounding factor in that. Regan was presented with the high fat = heart disease “proof” and at the same time a “high sugar” = heart disease “proof” its about equal proof. The corn farmers in the midwest were suffering horribly and glucose-fructose was a corn product that potentially could save their farms. Regan chose the high fat route to protect the corn farmers.

October 7, 2019 4:54 am

““James Hansen of NASA’s Goddard Institute of Space Studies beginning in 1988 predicted major droughts and up to six feet of sea level rise in the 1990s. One reporter recalled that in the late 1980s, he asked Hansen in his Manhattan office whether anything in the window would look different in 20 years. Hansen replied, “The West Side Highway [which runs along the Hudson River] will be under water. And there will be tape across the windows across the street because of high winds.”” [ source ]”

Except the answer was in response to what would happen if CO2 levels doubled in 20 years. Usual selective reporting on WUWT.

Reply to  B63
October 7, 2019 7:15 am

B63 ==> Interesting idea — do you have a link to the original interview transcript? The quote is from the Washington Examiner, not something here at WUWT.

Of course, Hansen would have been equally wrong “if CO2 doubled in 20 years”…..

Larry Geiger
October 7, 2019 5:13 am
Reply to  Larry Geiger
October 7, 2019 7:18 am

Larry ==> everyone should read Briggs regularly, and buy his book “Uncertainty” as well.

October 7, 2019 5:31 am

I think the entire 5000 words were lost on many, many commenters (not all, of course). The idea that WE CANNOT KNOW is just soooo offensive that people reject reality and follow fantasy all the time. It was a great article, I’ll mark it for later reference, but it’s really like trying to talk people out of their most sacred, yet completely wrong, beliefs. Artificial sweetners will still cause autism, carbs are horrible, and the latest diet craze is absolutely true. Human beings have to KNOW, even if they know 100% wrong material. It’s just how humans are, it seems.

Reply to  Sheri
October 7, 2019 7:22 am

Sheri ==> You are obviously a kindred spirit. People do insist on knowing even the unknowable — and pride themselves in being “possessors of the great secret”….that special bit of knowledge that must be shared with others (even if there is no evidence for it being true at all). Thus we get nutty Health Advocates pushing ju-ju berries as a superfood, anti-vacc activists, chem-trail nuts…….

Natalie Gordon
October 7, 2019 6:47 am

I am an experienced genetic epidemiologist. My entire professional life was spent applying the principals of epidemiology to disease with a genetic predisposition. There are three major issues with epidemiology that affects both nutritional studies and especially climate science.

1) A good epidemiologist understands that any form of extrapolation beyond the area where you have solid data for, is extremely dangerous because your conditions can change. The example I was taught was that in Alberta in the 1950s the population of the Hutterites (an anabaptist sect that lives communally) was growing at such a rate that if all factors remained the same, Alberta would be 100% Hutterite by 1980. Obviously that did not happen because of changes in the assumptions that went into the extrapolation. (Birthrates among Hutterties dropped. There was an influx of nonHutterite immigrants coming for oil jobs. Land was cheaper in Saskatchewan so new colonies were established in Saskatchewan instead.) I was taught anyone who does extrapolation as truth is a moron not a scientist. It seems to me the entire field of climate science is based on precisely this kind of dangerous extrapolation.

2) A good epidemiologist never assumes causation just because of a statistical correlation. Rather a correlation means you have a clue to figuring out what is going on and then you use information to follow up and do more research. An example is the correlation of starvation in WWII Holland due to the Nazis cutting off the food supply. The result was a large increase in the number of babies born with neural tube defects like spina bifida. One could simply say Nazis cause neural tube defects. get rid of Nazis and neural tube defects will stop. (Obviously the win by the allies didn’t end neural tube defects.) However a correlation provided the critically important clue about a possible nutritional deficiency causing neural tube defects. Decades more of careful research eventually traced the cause in 70% of cases to a lack of folate in the diet during a critical time in pregnancy among those with a genetic predisposition in the form of variant genes for less efficient enzymes involved in folate metabolism. This causes failures of microtubule functioning in developing neural cells. Most nutritional related neural tube defects are now prevented by the simple expedient of adding folic acid to food supply. The nutritional field has often seems to stop at the “Nazis cause neural tube defects” stage of research (read hazelnuts) with no follow up.

3) There is widespread misunderstanding of the use of statistics. I recall speaking to a climate change modeller who told me in all honesty that he knew what the correct answer was and he simply kept playing with the parameters of input into his model until he got the correct result. (He called it data adjustments.) He had no understanding of why that was not something you do nor why I had a problem with it. No one had ever taught him about things like ascertainment bias and the importance of understanding each parameter beforehand and setting them so you can test your hypothesis, not get a p value that is significant. The most infamous case of this is the hockey stick. Mann just kept changing his inputs until he got the output he knew was right. Naturally anyone who took his input data and redid his work would get the same result, ‘proving’ how correct he was while completely ignoring the fatal flaw in his result. The entire climate change field seems to have their understanding of modelling and statistics ass backwards like this.

4) Beware the confounding factor. I recall a climate change alarmist article that “proved” climate change would lead to less sex for men. (How alarming is that!?!) The researchers looked at birthrates in Australia during a heat wave and showed there were statistically significant fewer births nine months later. Heat wave = fewer babies = less sex = less sex for men due to climate change. They entirely ignored confounding factors such as when you are relatively hot and sticky you just don’t feel like sex but if you are always hot and sticky you’re going to install air conditioning or acclimatize to the heat. The whole conclusion was just plain stupid because of ignoring confounding factors. Headlines about heat waves in Alaska that ignore the installation of asphalt near the sensor is the most recent confounding factor I can think of.

Our society as a whole is largely innumerate. Journalists are particularly bad for being generally innumerate. The result is bad science and worse journalism. The only “cure” is to stop rewarding those who are innumerate with grants and publications.I have no idea how we go about doing that.

Reply to  Natalie Gordon
October 7, 2019 7:47 am

Natalie Gordon ==> Thank you for your professional input — All very good and valid points.

It seems to be an endless battle against innumeracy, statistical naivety and the pressure to publish (anything!).

Jim Gorman
Reply to  Natalie Gordon
October 7, 2019 1:03 pm

Natalie –> Great comment!

With the recent threads on random errors versus uncertainty I must agree that too many climate science modelers and even “scientists” have too much of a background in math, statistics, and computer science. They have no concept of errors versus uncertainty in the real physical world. They are convinced they can simulate the climate in a computer without ever stepping out into the good old world and make actual, real measurements for proving an hypothesis.

A good example is “homogenizing” temperature data. Who in their right mind would mess with perfectly good raw data just to make it better fit their computer program needs? If I would have done this I would have been kicked out of class or lost my job without keeping detailed how, why, when, and detailed notes on each and every change. And, it would have required a permanent record in a lab notebook. No writing a program and then claiming that the program was correct in every mathematical and logical way, so therefore, the changes the algorithm made were also correct.

That’s the same as simply saying I’m correct and you can’t prove me wrong!

Reply to  Natalie Gordon
October 7, 2019 1:16 pm

A1+++ comment

One of the problems I had recruiting volunteers for studies was that those who did not have the target condition (eg, headache, low mood) simply did not bother. It was necessary to go back to the ethics committee and put out new adds, and even then, the healthy and happy are hard to get. Of the ones I got, they may well not be representative. I often wonder about other people’s control groups. Another problem I have seen is that there was a group of unemployed, rather on the druggy side, individuals who were the regular volunteers for studies run by the pharmaceutical industry. How this biassed results is not known.

An anecdote: We ran a study that involved administering morphine to young men, We argued (and the ethics committee accepted) that the risk of addiction was lowest in those who did not use any recreational drugs. Urine samples from potential subjects were tested and it became apparent that our budget was going to be used up testing rejects. The ‘clean’ ones seemed to come from engineering, so we advertised only in the engineering schools in 2 Uni’s, and got the sample with no difficulty. We got the first systematic data on the psychological effects of morphine in a non-drug experienced group, although I’m sure the engineers were not a representative sample of humanity. The overall response was that an opiate was unpleasant, with no positive effects on mood or emotion, in stark contrast with NIDA’s findings in ‘post-addicts’.

October 7, 2019 7:12 am

Comes down to this: Our modern, processed diet of refined grains and added sugars is not only unnatural for OUR species, it’s unnatural for ANY species. One cannot require what one cannot obtain in a state of nature. Northern Europeans adapted to carnivory during the protracted glaciations as grass-eating megafauna were the primary food source. One look at our gut to brain ratio when compared to herbivores and the evolutionary fact is obvious. Can some people thrive on large amounts of plant material as well? Depending on motivation and adaptation, yes they can, but it’s chemically a vastly inferior food source and poorly absorbed, utilized by primitive man only when meat was unavailable. Cooking and processing rendered plant products more palatable (unto addiction, even) but didn’t make them good for us as we have ZERO evolutionary adaptation to dealing with what we’ve been eating since WWII, including polyunsaturated veg. oils. These things are the source of the present burden of chronic disease, which was ABSENT before 1920 or so.

The state of the best research today clearly shows that if you wish to reduce your burden of disease and live a hale and hearty old age, the key move is to reduce or eliminate your intake of ALL refined carbohydrates (grains and sugars) which entails eliminating virtually ALL processed, packaged, manufactured food. Then you can eat whatever combination of animal and green vegetable makes you feel good; but the most important point is to reduce the refined carbs, which reduces blood glucose, which reduces circulating insulin, which reduces excessive fat storage as seen in obesity. People who’ve done this can pretty much chuck all their pills, and their outdated doctors along with them, because the entire paradigm of our current disease-management system is predicated on the default of this unnatural diet of grain products, nutrient-blocking soy, and refined sugars.

The state of both nutritional and “climate” science is FRAUD; it is pseudo-science being press-fit to service an underlying agenda common to both, and that agenda is CONTROL over the populace.

October 7, 2019 8:45 am

Diet Dr. Pepper is the best diet soda, and Diet Mountain Dew is a close second. I drink lots of both, and I am skinny and healthy. NutraSweet is a combination of two amino acids found in most cells of all of us. The studies purporting to show toxicity are just ludicrous.

Reply to  Michael Moon
October 7, 2019 9:14 am

I’d say that saccharin is harmless too, other than it is sodium-based (diet Pepsi is best — :)).

October 7, 2019 8:47 am

We already know, from long experience, not to rely on single studies and thus we can avoid the “Single Study Syndrome”.

Sounds familiar, no? CO2 control knob….

October 7, 2019 10:23 am

The idea that CO2 radically affects global temperature has no more logic to it than malaria is caused by bad air or a miasma (http://broughttolife.sciencemuseum.org.uk/broughttolife/techniques/miasmatheory ) Earth’s concentration of CO2 is just to low to have a appreciable effect on global temperature, it is lost in the thermal noise.

JDD Ohio
October 7, 2019 10:38 am

I drink a lot of diet coke and have had stage 1 colon cancer. (Cancer confined to interior of polyp and hadn’t spread anywhere.)

In doing research for my own health, I found it amusing that so many people automatically assumed that a study tending to show that diet soft drinks were preventative of colon cancer was automatically dismissed by many people who almost certainly wouldn’t have dismissed an opposite conclusion. See study here https://www.nhregister.com/health/article/Yale-study-Diet-soda-may-keep-colon-cancer-from-13090038.php I realize that it is just one study and doesn’t show causation, but for me, being a heavy drinker of diet coke, it was mildly assuring.


Jon Jewett
October 7, 2019 2:22 pm

We grew up with my mother’s belief that butter and animal fats were bad, but margarine and hydrogenated vegetable oils were good. Fortunately, she went on to whatever the Lord had for her before she found out that it was the opposite. She would have blamed herself for my father’s series of strokes at 65 that left him in the California Veterans home for the next ten years.

(At least they weasel-worded it, which I have come to respect.)

“In the past, saturated fats were considered the most harmful type of fat for heart health. Saturated fat sources like butter were replaced with foods made from hydrogenated plant oils which were a trans fat source.

What we now know is trans fats are more harmful for health than saturated fats. Therefore, replacing butter with margarine made from hydrogenated oils is not recommended.”

Paul of Alexandria
October 7, 2019 3:33 pm

I was wondering if you were familiar with the website of statistician William Biggs
https://wmbriggs.com/post/28123/ – “The Epidemiologist (P-value + Ecological) Fallacy Exposed”
He has some interesting things to say on statistical analysis.

Michael S. Kelly LS, BSA Ret.
October 8, 2019 3:30 am

“WARNING: This is a long essay. Not a quick read — about 5,000 words — a twenty minute read for most.”

Reminds me of the college football quarterback. The coach takes him aside during practice on day, and says “Son, if you don’t pass English, I’ll have to cut you from the team.”

To which the QB says “But Coach! In order to pass me, the English teacher wants me to write a 500 word essay!”

“So?” said the coach.

To which the QB responded “I don’t know 500 words!”

October 9, 2019 2:24 pm


The nutrition whipsaw effect — what we should eat and what we shouldn’t — what’s good for us and what’s bad for us — changing every few months or couple of years has damaged the public reputation of Science with the general public and made “nutrition recommendations” a standing joke. Climate Sience has done the same thing with clear thinking people with their constant barrage of false predictions that don’t come to pass.

Ioannidis has exposed why this is so — it is not a new discovery, gthere have been manhy voices pointing out the same thiungs for years. Those who folloew this field know Steve Milloy and Aaron Carroll and the others.

The important point is this: Nutrition Science Cannot Find Causes of Health Harms or Benefits in Individual Diet Items, or even types or classes of diets.. It is a methodological problem and a problem of the subject’s complexity and the inter-correlation of thousands of factors, which the current methods cannot disentangle.

Bluntly, I believe the same is true for Climate Science — and that the current generation climate models, for some of the same reasons, cannot separate out causes and effects for mid- to long-term climate features — nor can they accurately or reliably make sensible projections.

If you have questions that I don’t answer below, you can email me at my first name at i4.net.

# # # # #

Reply to  Kip Hansen
October 9, 2019 2:26 pm

….excuse my sloppy fingers….” there have been many voices pointing out the same things for years. Those who follow….

October 9, 2019 4:21 pm

Bookmarked for future reference.

Great post, Mr Hansen.

Reply to  Kip Hansen
October 11, 2019 3:15 pm

10-4. I followed Steve Milloy’s junkscience.com for many years, til he disallowed comments. He was the leading edge of epidemiology critique.

Johann Wundersamer
October 15, 2019 1:13 am

And, finally: Climate Science?

How does a better understanding of the problems found in nutritional epidemiology offe us any insight into the field of Climate Science?

At the core of both fields is the issue of causality.

Causation indicates that one event is the result of the occurrence of the other event; i.e. there is a causal relationship between the two events. This is also referred to as cause and effect. [ source ]

As Ioannidis has pointed out in nutritional epidemiology: the specific scientific methods (large cohort studies based on food frequency surveys) and resultant statistical analyses are fundamentally incapable of ferreting out the individual effects on human health of individual, or classes of, dietary factors — they cannot discover causality. This is both a methodological problem and a result of the object of the study — human nutrition.

The problem with nutrition is

– metastudies comparing existing studies about discrete, concise formulated questions

– instead of metastudies comparing existing studies about discrete, concise formulated questions on selected, at least similar cohorts

– not to speak of double-blind studies

And, finally: Climate Science?

– similar problems with the nutrition metastudies

– leave alone double-blind; where’s the studies of similar planets.

Johann Wundersamer
October 15, 2019 1:44 am

“Whatever your relationship is with Science — be it in research, education or science journalism — you can support good careful and rigorous science; you can tactfully call-out poor science and bad science reporting; and you can lend your efforts and your voice to the task of reforming the Sciences and restoring their proper practices and returning them to their proper place in society.”

Astounding, who. Is it a bird, is a plane or is it


[] can support good careful and rigorous science; tactfully call-out poor science and bad science reporting; and lend his efforts and voice to the task of reforming the Sciences and restoring their proper practices and returning them to their proper place in society.


%d bloggers like this:
Verified by MonsterInsights