IMS member Jonathan Skinner has taught a free statistics class (“without math”) as part of the offerings at his local food coop, in order to spread the good word, for the past several years. He writes:
I am a retired statistician with a PhD in mathematics (1970). It struck my fancy to offer statistics in the classes at our local Honest Weight Food Coop, as a change from those on meditation and cooking. My monthly sessions aim to provide a chance to talk about statistical issues—and science in general—and to generate a spark of thoughtfulness for people outside academic circles. I advertise: “No math needed for or used in this stand-alone session.” This is not strictly accurate, since what I omit is algebraic notation, and when possible I describe the computations in words. I think, if Euclid can describe, with proof, the generation of perfect numbers from powers of two before the invention of algebra, I can define variance without using a frightening capital sigma, with indices, indicating summation. The expected attendees are people interested in a bit of intellectual zing, who would like to learn something about statistics, who don’t mind digressions into general science, and who have little or no background in mathematics. Some who actually attend claim to suffer math-fright or total ignorance of statistics, although this later claim is usually false. Others reluctantly admit to some statistical background but say they have forgotten it all. All of them seem to relate easily to graphs and diagrams.
Before each session, I email a script that I will work from to all attendees for whom I have an email address. Two to four people show up, so our meeting is like a long conversation about statistics.
My outlook is based on Fisher’s definition of statistics on page 1 of his Statistical Methods for Research Workers (now almost 100 years old and still in print): “Statistics may be regarded as (i) the study of populations, (ii) as the study of variation, (iii) the study of methods of the reduction of data.” [Emphasis in the original.] The second and third items differ little in substance from the definition of D.J. Hand in his recent article on administrative data in the JRSS (Series A, February 2018): “statistics is the technology of extracting meaning from data and of handling uncertainty.” So I always include Fisher’s definition with explanation, since I think everyone needs to recognize Fisher’s contribution to our subject. I also try to distinguish statistics from mathematics, saying that in math a number is just a number, while in statistics, a number is always a number of something and has a story behind it. If you don’t know the story of the data and its treatment, how the study of the population was carried out, then you don’t understand the statistic. But it’s a mistake to remember the story and forget the magnitude—you can’t have one without the other—even though popular mentions of a statistic rarely give its story.
I work in a bit of the history of statistics, to the effect that statistics dates back 5,000 years, when it was a matter of counting and tabulation. Probability is equally old, with dice-like bones found in archaeological digs. The two subjects joined about 200 years ago, although our modern subject, unlike the Average Man who was born from Quetelet’s work in 1831, lacks a birth announcement. Now, models are all-important, so they, as well as variability—Fisher’s item (ii)—are mentioned in every session.
Other than this, the content of the class changes every month, with minimal repetition. I draw statistical stories and topics from wherever I can. In addition to the work of Fisher and Professor Hand mentioned above, I am especially indebted to history books such as those by Stephen Stigler (Seven Pillars of Statistical Wisdom) and Theodore Porter (Trust in Numbers: The Pursuit of Objectivity in Science and Public Life) and to books on models such as those by Paul N. Edwards (A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming, with its motto, “without models, there are no data”) and by Virginia Eubanks (Automating Inequality: How High-Tech Tools Profile, Police, and Punish the Poor, with details of the misery arising from mis-analyzed administrative data). The recent book by Bradley Efron & Trevor Hastie (Computer Age Statistical Inference: Algorithms, Evidence, and Data Science) and the older book by C.R. Rao (Statistics and Truth: Putting Chance to Work) have provided examples. Margo Anderson’s The American Census, Michael Lewis’s The Fifth Risk, Emmanuel Todd’s Who Is Charlie?, etc.—too many books related to or using statistics to mention. I also include classical references and often end with a poem. My favorite poem is “The Three Goals” by David Budbill, about the difficulty of relating individuals to populations. [You can read this poem here — Ed.]
Do you put your statistical expertise to good use, outside your work? Tell us about it! Email firstname.lastname@example.org