考研必读经济学人杂志精选:Makingsenseofnumbers

文章作者 100test 发表时间 2007:06:11 11:47:27
来源 100Test.Com百考试题网

Damned lies and statistics

　　Making sense of numbers

　　Jan 6th 2005 From The Economist print edition

　　CHARTS rear up in the best and worst publications， from doctoral theses to the tabloid press. Yet the concept of drawing a graph from empirical data is a relatively recent phenomenon. Howard Wainer nominates William Playfair as the father of "modern graphical display". Playfair， something of a neer-do-well， Mr Wainer maintains， was a Scottish draughtsman for James Watt. By producing an "atlas" of Britain in 1786-44 charts， no maps-Playfair demonstrated that the presentation of evidence could have a beauty all its own. （The Economist began using charts only in the late 1920s， though it was a heavy user of data from its first issue in 1843.）

　　In his personalised and readable jaunt through the history of charting over the next couple of centuries， Mr Wainer introduces a varied cast of characters. They include Thomas Jefferson， who drew a chart of vegetable supply in the Washington， DC， market during 1802， John Tukey， the developer of exploratory data analysis， and Edward Tufte， arguably the best-known of current writers on visual display.

　　A statistics professor at the Wharton School of the University of Pennsylvania， Mr Wainer also underlines possible pitfalls in graph production. As examples， he gives irregular time periods， broken scales and alphabetical rather than ranked ordering.

　　In similar fashion， Jane Miller， an academic at Rutgers University who trained as a demographer， warns against common charting errors. Hers， much more a textbook， is clearly written， with a checklist at the end of each chapter， invaluable for students. It should be required reading for journalists and politicians.

　　Data need a context： a figure or two tells you little. The fundamental questions of journalism-who， what， when and where-have to be answered in charts too. Although Ms Millers book is chiefly concerned with writing about numbers， the last chapter gives advice about speaking with numbers. In presentations using visual aids， she says， use no more than one slide a minute.

　　The core of Joel Bests sequel to his popular "Damned Lies and Statistics" is how numbers can be used to confuse public issues. Mr Best， a professor of sociology and criminal justice at the University of Delaware， looks first at missing， confusing and scary numbers. He moves on to authoritative， magical and contentious numbers， and throughout takes particular aim at figures used in the public domain. Pithy examples abound， such as the 150 people allegedly killed annually by coconuts， even though no one has actually made a tally. At the end， he examines the chances of teaching statistical literacy to a wider audience， arguing that "when everyones numbers come under.