Click to confirm you are 18+

Lies, Damned Lies and ummm, those other things

basil_hallwardJun 29, 2017, 5:24:15 PM

One of the most unfortunate aspects of modern scientific work, apart from the deification of science by people who don't really understand what it is, is the tendency to conflate science with mathematics, which is not truly a science but an art - that is an artifice, in the original sense of the word something not of nature but created by humanity. Worse than that, they conflate statistical data with facts and thus trick themselves into believing mathematical models generated by computers are more reliable indicators than empirical evidence. This article was inspired (triggerd?) by a science writer (who admits he is not educated as a scientist, but in humanities, but nonetheless is a science fan who is besotted with statistics and the idea that machines are intelligent and will soon be more intelligent than humans.<a href="http://blogs.telegraph.co.uk/news/tomchiversscience/100217094/depressing-just-nine-per-cent-of-britons-trust-stats-over-our-own-experience-though-most-of-us-wont-believe-that/">Tom Chivers had posted a blog in the Daily Telegraph website in which he expressed despair at the reluctance of the public to accept the evidence of statistical surveys when determining what to believe.</a>Ironically, his evidence for this comes from a statistical survey. Only 9 per cent of the sample declared themselves swayed by such data.Tom is a good sort of guy (I believe he works for buzzfeed now), he was always courteous and witty in his responses to people who called him out for his dogy reasoning and blind faith in science. But unfortunately he was and probably still is absolutely besotted with the idea that 'Big Data' held the would give us the answers to every question, and thanks to the ability of digital technology to collect vast amounts of data, and using 'Artificial Intelligence' parse it incredibly quickly and throw up the answer to any question almost instantly, making us as omniescent as God.

That faith is misplaced of course, no matter how much data we collect and how detailed our analysis is, we cannot know the answers to the big questions, 'how far does infinity stretch?'; 'what is the meaning of life?'; is there a supreme being?'; and 'which horse will win the three – thirty race?'?Were I able to muster the kind of blind faith Mr. Chivers and other "science" fanboys place in statistics I would invest it in one of the traditional religions and at least be able to convince myself there is a chance of being resurrected into the life eternal. As it is, unlike Science fans who proclaim their atheism but have a zealot's faith in science, I am a true sceptic. That is, I am as sceptical of those theories about the universe beginning with a Big Bang as I am about a God having created it all. And while I believe analysis of data can give us a lot of answers, like anyone who worked in Information technology in the 1960s and 70s (and up to the 90s), I am very sceptical of predictions based on the conclusions drawn from it, which can only ever be subjective. Why it should perplex science worshippers and commentators with very strong liberal world views that data doesn't make everyone share their liberal world view.The answer to that, I can tell you as someone who has a qualification in statistics in addition to almost thirty years working with computers, lies in the old adage beloved of systems analysts and programmers; "If you torture data enough it will give the answer you want."Here (as everywhere) it pays to remember the theory of Bayesean Inference, that in any mathematical experiment conclusions will always be influenced by prior experience, current knowledge and expectations. The Rev. Bayes postulated that the probability calculus tells us that the purpose of data isn't to tell us what to believe; its purpose is to tell us how to modulate that which we already know.Put more simply, data and statistics can modify what you already believe about a hypothesis by weighting that belief by the data's plausibility.'<a href="http://www.bloomsbury.com/uk/why-truth-matters-9780826495280/">Why Truth Matters</a>' is a interesting book that offers some genuinely philosophical ideas on what a fact is, how it can be said to be true, and how 'facts' arise in the first place. And the deep problem is, that facts are ultimately relative to a worldview. If someone else doesn't use that worldview, then any 'facts' in it are of zero truth content.The whole point of sales marketing and politics is to persuade people to a worldview in which the salient issues are the ones the people doing the leading want you to embrace.Nigel Farage, leader of UKIP and leader of the Brexit campaign, dealt very niftily with an ambush lain by left wing activists who wanted to accuse his party of homophobia (he had already beaten off accusations of racism) when on taking a question from the audience on a television debate as to where he, and UKIP stood on gay marriage. Farage replied 'I actually don't care very much, UKIP doesn't care, and most of the people who support UKIP don't actually care, beyond not wanting to ram one view of a relationship down everybody's throat.'We are bombarded almost daily with statistics from public opinion polls conducted among very small samples of the public about how an overwhelming majority of the general public as a whole are wildly enthusiastic about legalising same sex marriage. Go into any working class pub anywhere in Britain but outside the more fashionable suburbs of London and you find people think there are far more important problems facing the world than the whines of the gay rights lobby.

The world is not divided into those who support same sex marriage or are against it as the statisticians and social scientists would have you believe, it is in fact divided into those who think same sex marriage an important issue, and those who could not give a flying f what bearded queens in bridal gowns do, if only they would just shut up, bugger off and get on with it. We must stop wasting the time of governments that have more important business to attend to and demanding that everybody should 'care' about the sex lives of irrelevant minorities.What modern scientists and their left-liberal-progressive fans have done, in their desperation to reduce everything to a binary, is dismiss diversity. Thus if you read a few web comment threads on news site where, for example, abortion is discussed you will find that complex subject is reduced to pro-choice versus pro-life. Criticise any pro - choice view, even those of extremist feminism which hold that terminations should be available right up to 974 weeks after conception*, argue quite reasonably that a limit of 22 or 24 weeks should be set, except in cases of medical necessity, and you are lumped in with those who would impose a total ban on religious grounds.Do 99.99% of scientists believe in climate change, i.e. that the climate has and does change? Of course. the historical evidence of climate change is everywhere. That is a more complex and nuanced question. First you have to define who is a climate scientist and who isn't. Next certain questions must be asked: Why have the climate scientists prophecies of doom failed to manifest? There is still snow, the ice caps have not disappeared, the 50 million refugees displaced by rising sea levels must be hiding under a bush. Yet the Warmageddonists continue to argue that their projections, based on adjusted data and statistical output from mathematical models cannot be questioned because "the science is settled."That is where those with an agenda stop.The science worshippers do not ask the really relevant questions. Like do you think that world is still in a warming phase, a neutral phase or a cooling phase? The evidence gathered from real world sources is its somewhere between neutral and slight cooling. So you might get a fairly even mix of answers.Or do you think that the main cause of late 20th century warming was in fact CO2 in the atmosphere? Here I would expect that you would get a 60/40 split in opinion.Or finally do you think, given the fact that China is opening a new coal fired power station every week, that it's worth western governments bankrupting their nations trying to set a moral example to countries whose interests would be well served by western nations causing their own national economies to collapse, rather than protecting their citizens from the possible effects of a slightly warmer climate?To that question the rational answer its overwhelmingly 'no'. But science has seldom been rational.The answers to those questions are never put before the public because the difficult questions are skillfully evaded. To present a picture of 99% of scientists believing in climate change and that man has an affect on climate, and therefore we must have windmills, smart grids and die of fuel poverty, only the data which supports those conclusions may be presented. And the global view thinkers get really pissed of when forced to accept that Joe and Jane Public are savvy enough to see through the statistical deceptions and understand that localism might hold better answers than the globalists 'one-size-fits-all' solutions.Perhaps the real political question should be: Are you of the opinion that a guaranteed lower standard of living is preferable to a 5% chance that you may suffer some effects (possibly adverse, possibly beneficial) from climate change in 50 years time that your sacrifice will be powerless to affect anyway?Statistics are not facts, and facts are not necessarily truth. A statistic is something measured, something counted, enumerated. What the numbers may mean is a matter of subjective interpretation (sic). For example in Manchester, England, rain falls on 205 days a year on average and there is an average of about 70 inches of rainfall. The truth of that to someone who lives in West Texas is Manchester has terrible weather. In Cherrapunji, India, there are over 330 days on which rain falls and the total annual downpour is almost 500 inches on average with the record standing at over 900 inches. To people in Cherrapunji, Manchester has an almost arid climate.Of deeper importance than the raw numbers is why one seeks to make said measurement. The reason is embedded in a theory or world-view of the person making the measurement. To them the measurement has utility: it potentially leads to "knowledge" (a relative concept) informative to decisions.Statistics (the discipline) may be thought of as an extension of two-truth-value (binary) formal logic into multi-value logic with degrees of certainty between zero (equivalent to definitely false) and one (equivalent to definitely true). The "degrees" of truth and falsehood referred to here are merely transmitted from starting assumptions and have no independent meaning. But another view of statistics is that it is more closely related to the dark arts than the sciences.Statistics is a powerful tool for reasoning, in the context of uncertainty for plotting trends and for forecasting the possible outcomes of sporting events. It's frequent misuse or abuse as a means of presenting subjective interpretations of data as "facts" does not invalidate that. A prime example is the "Smoking causes lung cancer meme." About twelve and a half percent of smokers develop lung cancer therefore statistics show that smoking does not "cause" lung cancer in the way that ingesting cyanide, or jumping off a high building causes death. On the basis of a single statistic we can say that smoking may be a contributory factor in many cases of lung cancer. But cancer is scary and so invoking it's name has an emotional effect. Put together all the health problems that smoking can be linked to and you have a strong statistical case against smoking but no single indisputable fact to provide a statement like “cigarette smoking causes lung cancer”.It is not "unscientific" to be sceptical about statistics, is is very scientific and extremely sensible.

all-rights-reserved