I'm currently reading Iain Gately's Drink: A Cultural History of Alcohol, an astonishingly thorough history of the world through the eyes of drinkers. In particular, my eyes were drawn to this passage regarding 13th Century England:
Ale was so vital to the very existence of the third estate that its price and quality were regulated by law. In 1267, King Henry III issued a pioneering piece of consumer protection legislation—the Assize of Bread and Ale—which set the maximum retail price of town-brewed ale at one penny for two gallons; the same penny bought three gallons from a country brewer. Prices were to be reviewed each year and could be adjusted in accordance with fluctuations in the cost of grain.
How times change, eh? Henry III would surely approve of Cooking Lager's campaign to 'Make it the Maximum'.
12 comments:
Glorious!
PS, it's "mediaeval", or be really fancy and do that joined together ae thing.
You mean mediæval.
Both are perfectly acceptable, but I'm afraid medieval is the preferred spelling these days. Google 'professor of mediæval history' and see how you get on.
I agree that mediæval has a nicer feel to it though.
My problem is I use a Mac, so to type 'æ' I need to hold down the alt' key and press keys at random until it shows up, in this case the key with the speechmarks.
BTW, don't you mean "Ye Olde Profeffore Of Ye Olde Mediæval Historie"?
PART1
I, too, would like to thank Carl for an insight into the “epidemiologic mind”. However, my thoughts will be somewhat more scathing of a branch of enquiry that has long been dangerously out of control.
All of Carl’s posts are a meandering through loose philosophical ideas of causation and a flipping between statistical and causal frameworks without seeming to bat an eyelid. These are very different frameworks, involving very different assumptions and argumentative requirements. This is not Carl’s fault per se. He is only repeating the mish-mash that he has been taught.
When asked a straightforward question about a standard for causal inference regarding the “relative risk” statistic, he could not provide one. This is not surprising because there isn’t one.
Very briefly, consider what are known as the “hard sciences”, e.g., physics. A critical function of science is pinpointing; the attempt to discern high-level predictors (antecedents) for specific consequents. An understanding of the laws of motion, etc, etc, allows one to predict from a position on the earth and with near-perfect accuracy the timing of sunrise/sunset and eclipses, for example. An understanding of underlying causal processes allows for the highly accurate pinpointing of a consequent, i.e., a priori, from an antecedent or cluster of antecedents. And this accuracy is not in relative risk terms but in absolute probability terms. High absolute predictive strength is ultimately a main “gold standard” of causal inference.
Some examples of the nonsense that is produced by a reliance on the relative-risk statistic: Consider a control group and an experimental group of 100 people each. (it doesn’t even matter what the experimental measure is). In one experiment there is 1 (baseline) person in the control group and 7 in the experimental group demonstrating a particular measured consequent. In another experiment, there are 10 (baseline) persons in the control group and 70 in the experimental group demonstrating a particular consequent. In both instances the RR will be 7.0. Yet the two situations represent very different circumstances. In the first instance, the absolute predictive strength of the experimental variable for the consequent is 6% above a baseline; in the second, 60% above a baseline. There is very good basis for causal speculation in the second circumstance that is not so for the first. It would be impossible to claim any causal understanding when, by using a particular predictor, you would be wrong 94% of the time, i.e., as in the first case. Alternatively, epidemiology would jump to causal conclusions entirely on the basis of a high RR in both circumstances. Consider another example, where there are 45 (baseline) persons in the control group, and 90 in the experimental group demonstrating a particular consequent. The RR would be 2.0 – far lower than the 7.0 in the very first instance. Even though the RR=2.0, there would be a good basis for beginning to entertain a causal basis for the relationship between an experimental variable and a particular consequent because of an absolute predictive strength of 45% above a baseline.
Relative risk is not – should not be – a basis for causal argument. RR only includes positive consequents following an antecedent. It does not account for negative consequents (non-occurrences of the consequent). Therefore, the conditional-probability context along which relative risks occur is hidden. A proper measure of absolute predictive strength is, say, a correlation co-efficient such as Pearson “r” which accounts for both positive and negative consequents in the presence of a specific antecedent(s).
PART2
A critical scientific goal, then, is the identification of antecedent(s) that uniquely predict a specific consequent with 100% accuracy. Understandably, in some realms of enquiry the subject matter and complexity make this sort of accuracy an impossibility. However, the goal is still to strive for predictors with a [absolute] predictive accuracy at the upper-end of the conditional probability scale, i.e., high-level accuracy.
Relative risk is the statistic of choice for conmen. It would not be surprising that the vast majority of epidemiologists have never heard of absolute predictive strength. Epidemiology runs almost exclusively on the relative risk statistic. And we have seen how misleading it can be.
The psychological danger of abuse of relative risk. Consider the very first example where there was 1 and 7 per 100 in each of control and experimental groups for a particular consequent. Ninety-three percent of the experimental group (risk factor) is comparable to 99% of the control group, i.e., they do not demonstrate the particular consequent in question. With an RR=7.0, epidemiology will tell the entire experimental group that they are “at risk”; the typical advisement being that they (all of them) should eliminate the “risk factor” to eliminate their risk, i.e., causal implication. This claim actually contradicts the very data that is being referred to. Ninety-three percent of the experimental group does not demonstrate the consequent in question. Therefore, eliminating the risk factor will not modify their risk at all – they were never “at risk”. For the other 6% above a baseline, the absolute predictive strength is so poor that it provides no basis for causal argument. Remember that for causal argument, there must also be an accounting for why a consequent does not occur for an antecedent(s) in 93% of cases, i.e., the preponderance of cases are counter-examples or anomalies to the causal proposition. You will notice that the epidemiological “reasoning” is upside-down: It makes the atypical appear typical, and the typical to appear atypical or non-existent. Telling people that they are “at risk” with causal implications when in fact they are not, is an assault on psychological health which will then have further detrimental social, moral, political, cultural, and physical consequences.
It needs to be considered that the vast majority, if not all, identified “risk factors” for heart disease, cancer, etc, have an absolute predictive strength of 10% or less; for dietary epidemiology, the predictive strength in many cases is barely above 0% accuracy. These are on the bottom(wrong)-end of the conditional-probability scale. Second-hand smoke claims are barely above 0% accuracy, i.e., if you use secondhand smoke as a predictor for a particular [disease] consequent, you would be wrong just short of all of the time. Deranged, then, is epidemiology’s use of the term “cause” in referring to a near-0% accuracy level in the same way that the “hard” sciences refer to near-100% accuracy. This relative-risk/”causal” insanity is being inflicted on societies around the world on a daily basis. It is a constant terrorizing.
Consider the three major risk factors for early-onset heart disease – hypertension, high cholesterol, and smoking. Remember that the word “major” is a relative term. In terms of absolute predictive strength, they are not major at all but are very poor predictors, each being under 10%. Yet, all persons that have these “risk factors” are aggressively advised that they need treatment to reduce the “risk”. It should be obvious by now that most are not “at risk” and there is no basis for a direct causal argument. Obvious, too, is that the treatment of these risk factors is lucrative for the pharmaceutical cartel, specifically, and the medical production-line, generally. Even more perverse is that while these people are being treated, they are actually well. It is an excellent money spinner. You can be medically treating well people for 10, 20, 30, 40 years with the promise that it will alleviate the potential for a detrimental consequent a long way in the future and when most of those being treated do not even demonstrate a statistical association, let alone a causal basis, with the consequent in question. Worse still is that the medication has potential side-effects.
The question then arises as to how epidemiology has come to be in this sickly condition. The simple answer is eugenics and statistics. It should not be surprising that the originator of the term “eugenics”, Francis Galton, was a statistician. Others that advanced the idea were also statisticians, e.g., Karl Pearson. Eugenics is a materialist view of the world. It views humans in the same way that it would view a herd of cattle. And the eugenicists believe that they can breed or foster a superior “human herd”. In this sterile materialist framework, there is no mind, soul, spirit, God, freedom. The eugenicists see themselves as an [self-installed] elite separate from the herd and “perfectors” of the herd. The eugenicist is obsessed with such things as smoking, alcohol, diet, exercise. A particular statistical trick, aligned with eugenics and allowing eugenics to proliferate, is population(herd)-level statistics. The population or “the herd” is viewed as “one organism”. The relative-risk statistic, in this contorted view, seems to make sense. You compare different groups on different measures, and all that you’re concerned about is significant trends indicating differences between the groups (herds). This is why relative risk differences are erroneously assigned to the entire membership of an experimental group. The critical problem is the construct of “one organism”. A group of people is not “one organism” but a collection of discrete individuals and particular measures occur in individuals. Further, the “group trends” indicated by relative risk does not translate at all into high predictive strength at the individual level. As seen earlier, relative-risk differences are usually produced by an atypical (small) subgroup within the overall group. As soon as you ask “how well does an antecedent predict a consequent at the individual level”, the epidemiology house of cards comes crashing down. Remember that epidemiology does not ask, conveniently, for predictive strength at the individual level.
PART4
What should be noticed is that epidemiology advances numerous causal claims but is very light on actual coherent causal explanation. It does this in at least two ways. Given that there is no consideration at all of absolute predictive strength at the individual level, a gathering of “eminent” epidemiologists can agree to conclude that a particular RR represents a causal relationship between a risk factor (an antecedent) and a specific consequent. This can be termed causal argument by consensus. An RR could be barely above 1.0 or 5.0, etc., etc.. There is no fixed standard. Only the consensus of the group matters. A second way is to invoke the “precautionary principle” (also Rose’s Paradox). Even for low RR’s (e.g., barely above one), epidemiologists can argue that although the causal status of a statistical association is in question and although the RR difference is small, it involves sufficiently large numbers at the population level that policy changes should occur as if the statistical association is causal as a precaution. Both “causation by consensus” and the “precautionary principle” are deceptive ways of bypassing the requirements of coherent causal argument, and that will involve absolute predictive strength. This conduct is delinquent and fraudulent, to say the least, with many dangerous ramifications. The quickest way to bring epidemiology into check is to require that those claiming causation between factors bet all of their worldly possessions on their ability to accurately predict an outcome (absolute predictive strength) in specific individuals on the basis of their nominated antecedent(s). The causal claims would immediately stop.
Eugenics also has inbuilt prejudices, such as antismoking and anti-alcohol. Consider the well-worn claim: “Smoking causes lung cancer”. For smoking and lung cancer the RR is 10.0. For every one in a hundred in a control(non-smoking) group, there are 10 instances of lung cancer in a group of lifelong heavy smokers. The RR seems high, but the absolute predictive strength is poor (10%). In other words, if you use lifelong heavy smoking as a predictor for lung cancer, you would be wrong 90% of the time. As indicated earlier, epidemiologists do not comprehend the need to explain why 90% of lifelong heavy smokers do not develop lung cancer, i.e., the preponderance of evidence is for the non-association – even in just statistical terms – of heavy smoking with lung cancer. Therefore, relying on the RR and the antismoking bias in eugenics, a “causation by consensus” produced “Smoking causes lung cancer”. You can search the medical literature of the last half century and you will not find one example of a coherent causal explanation of the above claim. The simple reason is that such an explanation (referring to both positive and negative instances), based on the available evidence, is impossible. So acute is the antismoking fixation within eugenics that any statistically significant RR, regardless of how low, is referred to in causal terms (causation by consensus).
PART5
And, so, there is now a plethora of diseases “caused” by smoking. The eugenicists have even tabulated the “toll”. The Center for Disease Control has produced a program called Smoking Attributed Disease & Mortality and Estimated Cost (SAMMEC). It is entirely a statistical exercise. Relative risks for a variety of diseases/mortality are converted to attributable numbers on the basis of assumed attributable causation. The numbers are summated to produce an overall “death toll” from smoking. Via another set of questionable assumptions, the disease and death toll can be converted to a monetary “cost to society”. Worse still is that, in statistical terms, RRs for specific disease/mortality may involve cross-correlations with many other factors (confounders). These confounders are not partitioned for the toll associated with a specific risk factor (e.g., smoking). Consequently, and as indicated by other posters, if we were to summate the toll from all un-partitioned risk factors, we would end up with an annual statistical death toll far in excess of the actual annual death toll. So, this conduct is not only misleading but it is maximally misleading. Within all of this nonsense, the eugenicists can play all manner of other statistical shenanigans, e.g., claims that reducing this risk factor by X% will reduce a particular consequent by Y%, etc, etc. and goes by the name of “preventive medicine”. We hear these sorts of claims on a daily basis. Yet nowhere in all of this obsession with numerics, statistical gymnastics and cost/benefit analyses – this sterile physico-arithmetic fantasy world - that eugenics is notorious for is there any actual coherent causal explanation or predictive strength at an individual level. It is all population(herd)-level statistical derangement. Further, nowhere in this insanity is there any consideration of the detrimental consequences of this conduct along psychological, social, and moral dimensions: these dimensions do not figure in the materialism of eugenics. The madness of these madmen is not part of these madmen’s definition of madness.
Another excellent example is “exposure to ambient tobacco smoke causes Sudden Infant Death Syndrome”. Fortunately, SIDS is a rarity. SIDS is a “syndrome” defined by exclusion: When no other cause of death can be attributed, it is assigned to this unknown syndrome category. The relative risk for exposure to ambient smoke (indirectly measured) and SIDS is around the high 1.0’s. However, the absolute predictive strength is barely above 0%. Epidemiologists would have us believe that an unknown aspect(s) of tobacco smoke “causes” an unknown syndrome via an unknown process(es). This sort of inflammatory trash, i.e., non-explanation, with considerable psychological, emotional, social, and moral consequences is what passes for “causal explanation” in eugenics-driven epidemiology. Obviously, the causal claim has been produced by delusional “consensus”.
So, it can be said that epidemiology reasonably employs aspects of the scientific method, particularly in data collection. However, at the interpretation level, the conduct is squarely anti-scientific and fraudulent. It routinely violates edicts of statistical and causal inference. And this incredible problem is endemic. Those that are attracted to epidemiology and Public Health are probably already of a materialist persuasion. They are then further trained in this delusional thinking. And it is not as if epidemiology has recently gone off the rails. It was never on them from the outset. Epidemiology is a dangerous scientific masquerade.
PART6
What has allowed this grand, dangerous fraud to be perpetrated on a global scale? (very briefly)
At the turn of the last century, eugenics became mainstream in the USA, the UK, and a number of Scandinavian countries. The USA appears to be the most prominent. The mega-wealthy in the USA (e.g., Rockefeller, Carnegie, Ford, Kellogg) were supporters and funders of eugenics (and antismoking, anti-alcohol). Rockefeller and Ford were also prominent supporters of Nazi eugenics. Rockefeller and Ford had trade agreements with the Nazis through the 1930s. There was also a very intimate relationship, through treaties, between Rockefeller’s Standard Oil and Germany’s IG Farben.
Hitler was a student of American eugenics. National socialism was “applied biology” and a continuation/extremizing of American eugenics. The American eugenicists saw Hitler as the “darling” of the movement, a leader that could apply eugenics principles unfettered by, say, constitutional concerns. During the 1930’s, the Nazi regime was actively promoted in the USA, particularly California. IG Farben, the German petro-chemical giant, is considered to have financed Hitler and the National Socialists into power, and with whom it partnered in its invasions, i.e., industrial socialism (see Dr. Rath website).
A well-buried, little-known, fact is that in late-1933/early-1934, just after the fascists were installed to power in Germany, there was an attempted coup in the USA intending to install a fascist government. This is known as the “Business Plot”. A congressional investigation concluded that there was an attempted, but failed, coup. The wrong man was picked to lead the coup, who then turned whistleblower. No prosecutions followed (possibly deals). However, the attempted coup involved a “who’s who” of the mega-wealthy in the USA. (google “Business Plot”. Some documentaries have been done on it. If anyone wants links, I can provide some).
At the end of WWII, and given the atrocities of the failed Nazi regime, the American eugenicists and mega-wealthy went to considerable lengths to distance themselves from their connection with the regime. There were a number of headline prosecutions at the Nuremberg trials. But many Nazis were given access to, and freedom in, other countries (Operation Paperclip). The leaders of IG Farben, instrumental in many atrocities, were given light sentences and then went right back to their business dealings. The giant IG Farben was ordered to be split up. A number of these split entities are now high-profile pharmaceutical companies, e.g., Bayer, Hoechst (now Sanofi-Aventis??).
The point of the above is that eugenics did not die with the defeat of Nazism. Given its association to horror, eugenicists laid low for a while post-WWII. They then started up again in the 1950’s, careful not to use the “E”[ugenics] word. However, the biological reductionism, the healthism, that has gathered momentum over the last four decades, its obsession with smoking, alcohol, diet, exercise, disease – to the exclusion of all other dimensions of the human condition, is eugenics by other names. The obsession with statisticalism and population(herd)-level control to the detriment of individual autonomy, is all eugenics. Eugenics is dictatorial/fascist. Public Health is a key instrument of eugenics. The mega-wealthy have had a century-long interest in eugenics. They are particularly interested in this deranged ideology because it allows the appeasing of a number of their other obsessions – greed, avarice, megalomania. The mega-wealthy, not all, have an utter contempt for ordinary people. Their interest is in installing a eugenics “elite”, which they control, to oversee the remainder of the population. This remainder is a human labor force/herd that are viewed as no more than cogs running the industry that profits the mega-wealthy. This is what was seen in Nazi Germany.
PART7
Through Public Health and Health and Safety, through the constant terrorizing of the public with phantom dangers, eugenics has essentially medicalized the human condition, medicalized ordinary life. It controls the thinking of the masses. It is all about population control.
Having digressed, we return to epidemiology. Why is epidemiology, or Public Health, the way it is? Because that’s the way the elite (controlled by the mega-wealthy) want it. If you adhere to the rigors of scientific enquiry, and statistical and causal inference, you cannot control the population. Rather, through a masquerade of scientific enquiry, a bastardization, “science” can serve the State – “science” working to an agenda. The mega-wealthy, through philanthropy, through “charitable” organizations, through political and educational organizations, have dictated over the last three decades what and how people are taught from primary through to tertiary education, and the types of people that make it through to high administrative positions and political office. Through their influence over government, they dictate what themes will be funded in academic research (and which will not). They dictate that the tiny enquiry of epidemiology wields extraordinary public-policy power and that it remains exactly as is. The Public Health bureaucracy has been burgeoning over the last number of decades. Contemporary Public Health (and preventive medicine) is population control. Rockefeller, an early benefactor of eugenics, was the primary force behind allopathy (cut, burn, poison) being installed as the primary medical establishment. He also created the American Cancer Society, the American Medical Association, the American Heart Foundation, the American Lung Association early last century. These are essentially protection fronts for the allopathic medical establishment. This allopathic medical model has been replicated in countries around the world. Remember too that the medical establishment was a frontrunner in the Nazi regime. There are progressively more and more of these disease and dismembered-body-organ organizations. There’s even a British Liver Trust. (Why do we need these dismembered-body-organ organizations at all?) There are numerous other “charitable” organizations and NGO’s. They all toe the Public Health line and collectively they constitute a formidable, ever-growing, eugenics network, whether many are aware of it or not. And they are all directed by the eugenics-driven medical headquarters - the World Health Organization, a post-WWII creation. Through this network, societies and cultures around the world are being overthrown, converted to sterile, uniformity. Many countries are already resembling low-security prisons. They can get much worse.
(see the book Rampant Antismoking Signifies Grave Danger where many of the epidemiology issues above are considered in greater detail. www.rampant-antismoking.com )
@ Anon 1. Thank you, thank you, thank you!
A brilliant series of posts that proves a lot of what I'd long suspected. Some of the information I already knew but I'd never been able to place all the pieces together. Fantastic stuff.
May I reproduce these posts on my (rarely updated and little read) blog? Please contact me if you are against this.
Thank you again.
Post a Comment