voodoo no more - April 14, 2009
In January, an in-press article criticizing the statistical methodology of social scientists using fMRI to back their hypotheses caused a major stir. The no-holds-barred paper claimed that many published papers are worthless because brain imaging data have been inappropriately analysed.
Those scientists personally attacked were not only upset by the content. They were also offended by the title which referred to ‘voodoo correlations’. Many argued that they were not only well aware of the importance of good statistical practice, but also used it.
The editor of the journal in question, Perspectives in Psychological Research, has now said the final version to be published in May will appear with a more circumspect title: ‘Puzzlingly High Correlations in fMRI Studies of Emotion, Personality and Social Cognition’. The authors, Edward Vul from MIT and Harold Pashler from the University of California, San Diego and their colleagues also answer the specific criticisms they received in response to the widely circulated original version in an accompanying paper.
