The file drawer problem

The Economist's recent article How science goes wrong is a useful reminder of publication biases in science. It recently came up in a discussion at work, which took me back to Rosenthal's paper on 'the file drawer problem' (null results in science are seen as boring and unpublishable, so tend to get filed away, while the study or its analysis is tweaked to produce something that shows a significant effect). Rarely mentioned in research methods teaching.

Rosenthal, R. 1979. The file drawer problem and tolerance for null results. Psychological Bulletin 86/3, 638–641