SignificanceVolume 10, Issue 6 p. 44-45 Editorial/LettersFree Access Editorial/Letters First published: 18 December 2013 https://doi.org/10.1111/j.1740-9713.2013.00713.xAboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinked InRedditWechat Birthdays and anniversaries bring reflection; it is 10 years since the first issue of Significance came out. Ten years is a long time. I have edited Significance for eight of those years – my predecessor, Helen Joyce, did the hard work of setting the magazine up and establishing its style. Those 8 years have taught me at least a couple of things about statistics. One is just how wide their applications are. In the course of those years we have covered shipwrecks, penguins from space, voyages to Mars, Viking warriors, hobbit men, how starlings flock and whether nightingales sing in tune, even how many fish there are in the sea and whether we are likely ever again to see a Caribbean monk seal (Significance, December 2010). Also how to save orang-utans, the safety of Prozac, how much salt we should eat, whether organic food really is better for you, the 5% significance level, climate change, bees, feeding the world, whether country or town is greener, the authorship of books of the Bible, the archaeology of Pompeii and of the woods of southern England, Hollywood films, women's clothing sizes, the crash of 2008 – you get the idea. Statistics is relevant to pretty nearly everything under the sun. This is perhaps the great secret than non-statisticians do not realise. The public image of statistics is of lists of numbers. The public do not always grasp that those numbers are about something. They come together to tell stories. The statistician's job is to decipher those stories; and the stories can be about just about anything you like. That is one reason why statisticians are lucky in their profession. They do not have to be stuck in one groove. If the statistics of health and medicine fascinate, excellent; they can specialise in that, and do a huge amount of good in the world and save lives. But if they want to look at school exam results one week, and downtown traffic jams the next week, and what people are saying on Twitter the week after, that option is also available to them. Statisticians can be as varied in their professional lives as they want to be. The second thing I have learned is that statistics is changing – and very fast indeed. Ten years is a short time – on our fold-out timeline of statistics (see the centre pages of this issue) it hardly shows. But in not much more than that time new insights – Bradley Efron's bootstrap (Significance, December 2010) is just one – have come. Computing tools such as R have eased the traditional workload and transformed what is possible in statistics. Big Data has arrived, and has become mainstream and is producing a whole new category of specialisms and job titles: high-flying graduates now call themselves “data analysts” and “data scientists” rather than statisticians. But as the first statistician-as-celebrity Nate Silver (see page 36) has put it: “I think data scientist is a sexed-up term for a statistician.” And, not least, statistics has begun to acquire a public image. People are beginning to realise that it is not just numbers. Journalists are beginning to become aware of statistical issues in the stories they cover, and are asking sensible statistical questions and conveying the answers to their readers. That is partly due to people like Nate Silver. But that, too, was one of the aims of the people who started Significance. If the magazine has contributed a little to the wider appreciation of the value of statistics to society, then its first 10 years will not have been wasted. If it has given some enjoyment, and some information about fields other than their own, to statisticians and to those who are not statisticians alike, then it will also have served a purpose. Julian Champkin A meaner ox M. H. Robson (Letters, October 2013) returns to Francis Galton's analysis of the entries in a weight-judging competition at a country show in 1906, published in Nature (March 7, 1907). He notes that the mean estimate of the dressed weight of the ox, given in a later letter by Galton, is closer to the outcome than the median estimate that Galton used in his initial article. However, on examining worksheets and other materials in the Galton Archive at University College London, I found that there are small errors in the numbers in the printed article, and that correcting these makes the story even better. The median estimate is the 384th in Galton's ranked list of 787 estimated weights, equal to 1208 lb, and the outcome, supplied in a letter from the organiser of the competition, was 1197 lb. The Nature article gives these as 1207 lb and 1198 lb respectively, and the corresponding error as 9 lb. The final digits of the two key numbers were somehow transposed, and the true error in the median estimate should have been 11 lb, as given by Galton in some earlier correspondence. Two weeks later Nature published a letter by R. H. Hooker, who returned to the mean-versus-median debate and wished that Galton had also calculated the mean. Without the actual figures, but judging from the table of centiles in Galton's article, the mean seemed to him “to be approximately 1196 lb, which is much closer to the ascertained weight (1198 lb) than the median (1207 lb)”. A week later, as Robson notes, Galton reported the actual mean, which is 1197 lb. Robson goes on to say that Galton did “concede rather grudgingly than the mean was in fact even closer” to the outcome than the median, but Galton in fact made no such comparison, merely observing that the closeness of Hooker's estimated mean demonstrated the utility of his table of centiles, despite its compactness. He suffered the defeat of the median estimate in silence. But armed with the correct outcome, we now see that the mean estimate was spot on! It is a striking example of the gains from forecast combination, long before the classic article by Bates and Granger. Kenneth F. Wallis University of Warwick A Snowball for Significance's tenth birthday In 1960, a group of (mainly) French-speaking writers and mathematicians started Oulipo1 (Ouvroir de littérature potentielle; roughly translated as “workshop of potential literature”). One of their goals was to promote new forms of constrained experimental textuality. Perhaps the best-known type of such constraints is the palindrome, a sentence that reads the same backwards as forwards. An example is “Stats O Do Stats”. Another textual form promoted by Oulipo and of particular interest to mathematicians is the snowball2, a poem in which each line is a single word, and each successive word is one letter longer than the previous word. I wrote this one with 22 lines to celebrate Significance's tenth birthday. Mario Cortina Borja London Overconfidence intervals Mannes and Moore (Significance, August 2013) are overconfident in asserting that “If you are perfectly calibrated … you would get 7 out of … 10 [assessments of 70% confidence intervals] correct”. The number would be a binomial random variable based on p = 0.7 and the number of questions attempted. The probability of getting exactly 7 from 10 is 0.27, and nearly as high for 6 and 8. A perfectly calibrated statistician would be surprised only if she scored fewer than 5, and would attribute a score of 10 to superior general knowledge. Peter Oppin London Innocence and guilt I read with interest the article (Significance, October 2013) with the subtitle “What is the significance of the 5% significance level?” So I thought readers of Significance might be interested in the following quote from G. K. Chesterton's Father Brown detective stories (The Innocence of Father Brown: The Blue Cross, 1910): “‘Proof!’ he cried. ‘Good God! the man is looking for proof! Why, of course, the chances are twenty to one that it has nothing to do with them. But what else can we do? Don't you see we must either follow one wild possibility or else go home to bed?’” In detective fiction, therefore, it seems we have 5% significance or a “wild possibility”. S. R. Wilson Canberra and Sydney Ouch Regarding the call for a Bayes’ theorem tattoo “on a part of the anatomy that we can decently reproduce” (Significance, October 2013), wouldn't the obvious choice of anatomical part be a posterior? Larry Lesser El Paso, TX Letters should be sent by e-mail to significance@rss.org.uk, or by post to: Significance Letters Page, Royal Statistical Society, 12 Errol Street, London, EC1Y 8LX. They should be short (preferably under 250 words), may be edited for length and should clearly indicate whether or not they are for publication. They must be received by January 15th, 2014, in order to be considered for publication in the February issue. References 1 The Oulipo group is still active; see http://www.oulipo.netGoogle Scholar 2 A study of the snowball, by David M. Reider, can be seen at http://www.bgsu.edu/departments/english/cconline/rieder/2i.htmlGoogle Scholar Volume10, Issue6December 2013Pages 44-45 ReferencesRelatedInformation