Special issue: Are you behind the times?
Articles by Pasco Fearon (resampling), David Clark-Carter (effect size), Jeremy Miles and Mark Shevlin (structural equation modelling), Andy Field (meta analysis), and Daniel Wright and Sian Williams (reporting results).
18 December 2003
Statistics, like any branch of human knowledge, is open to debate and subject to change.
However, many psychologists have been happy to stick to a set of techniques that they learned in their first year as undergraduates and that have been around since the mid-1930s.
Statisticians have continued to explore those techniques and create new ones, and the availability of ever-faster computers has meant that aspects of the techniques can be tested in ways that were never available to those who originated them.
In 1900 when Pearson devised the c2 test, which is often considered the first 'modern' statistical test, computer and calculator referred either to a pretty rudimentary mechanical device or even just a person with a sharp pencil and a large piece of paper.
Things have moved on. Nonetheless, numerous attempts to persuade psychologists (and those in many other disciplines) of the need to take account of the new findings and approaches have fallen on stony ground.