THE SPIRIT LEVEL REVISITED REGRESSION LINES, CORRELATION, OUTLIERS and MULTIVARIATE ANALYSIS A plain man’s guide to statistical inference in THE SPIRIT LEVEL

One such critique, a paper written by Peter Saunders (PS), has been published by the think-tank Policy Exchange. It is entitled Beware False Prophets and is currently readily available on the Internet. Wilkinson and Pickett (W&P) have responded robustly to their critics and their comments are also available on the Internet [1]. In defence of their thesis that income inequality correlates (in some cases strongly) with various social problems they point out that their selection of the dataset from which these results were obtained, was based on sound principles and decided before the data relating to each was analysed. This is in sharp contrast to the datasets preferred by Saunders who has added and subtracted countries to and from the dataset in an effort to get the relationships he prefers.