Website Of The Week Website of the week

Teach yourself statistics

BMJ 2001; 322 doi: (Published 24 February 2001) Cite this as: BMJ 2001;322:498
  1. Christopher Martyn (cmartyn{at}
  1. BMJ

    A glance at this week's (or any week's) BMJ will show that medical research inevitably involves statistics. In his self help book of the 1930s, Mathematics for the Million, Lancelot Hogben included a chapter on statistics entitled “The Arithmetic of Human Welfare”—a warm, almost touchy-feely description of a subject that many doctors still regard as repellently algebraic. Unfortunately the chapter doesn't live up to its title. The explanations are long winded and often tough going. Technology, however, has made things easier for Hogben's successors. The possibilities created by Java applets for interactive tutorials make the web a good place for the statistical novice to look.

    The Rice Virtual Lab in Statistics ( provides HyperStat, an online statistics book. Its strong point is not the text but the excellent simulations that accompany it. There are clear demonstrations of many of the things that most of us struggle to grasp—the central limit theorem, regression to the mean, the binomial distribution, etc. Especially good is the display of how transformations by, for example, square roots or logarithms can affect the interpretation of a relationship between two variables. The site also has case studies to take you through statistical analyses of real data and an analysis lab, where you decide how to explore example data sets.

    At, John C Pezzullo, a professor of pharmacology and biostatistics, has compiled a list of more than 600 links to web based statistical resources. They are grouped into sensible categories, and it's not hard to find interesting and useful stuff. Several sites allow you to enter your own data and analyse it. Others provide such things as glossaries of statistical terms ( and, random number generators (, power and sample size calculations (, and printable statistical tables in pdf format (

    View Abstract

    Log in

    Log in through your institution


    * For online subscription