The Emperor Has No Clothes

Reviews the book, Reliability in Cognitive Neuroscience: A Meta-Meta-Analysis by William R. Uttal (see record 2012-22023-000). Uttal begins the book with a thoughtful short chapter explaining the raison d’être for meta-analyses of a body of experimental data, in this case focusing on the cognitive neuroscience literature. Those who are familiar with modern statistical methods will understand; those not-so-well-versed neuroscientists would do well to review carefully the statistical issues. Uttal’s step into meta-meta-analysis is a bit controversial, although no one could fault his goal, which is to attempt to reduce the variability in the data from brain-imaging experiments. Meta-analysis and meta-meta-analysis increase the variability, leading Uttal to conclude that the answer to the question implicit in the title of his book is, no, the data are unreliable. Uttal’s point should be taken seriously by anyone engaged in research that uses changes in brain images as the dependent variable in a cognitive experiment. (PsycINFO Database Record (c) 2016 APA, all rights reserved)