Statistical Modeling, Causal Inference and Social Science

Andrew Gelman, a statistics professor at Columbia, has run this blog for close to two decades, and it's become required reading in parts of academia for its running commentary on bad statistics, failed replications, and sloppy causal claims in published research, often naming the specific paper or news story it's skewering.

The tone is blunt and occasionally combative, which has made it both influential and controversial among the researchers it criticizes. Best for people who work with data and statistics professionally and want a sharp, if sometimes cranky, watchdog on how research actually gets done.