Playing Possum: The Variability Hypothesis

A new analysis published on LessWrong revisits the controversial 'variability hypothesis'—the idea that males exhibit greater variation in physical and cognitive traits than females—using an unexpected dataset: Tasmanian devil whisker measurements. The research, led by user 'rba', examines whether X-chromosome inactivation in placental mammals (where females randomly deactivate one X chromosome, potentially averaging gene expression) could explain greater male variability. To test this, the team turned to marsupials like Tasmanian devils, which lack this X-inactivation mechanism, analyzing longitudinal data tracking whisker lengths across 600+ individuals from a 2022 study by Attard et al.

The analysis found that Tasmanian devil whisker length distributions showed complex patterns that don't neatly support simple genetic explanations for sex-based variability differences. The researchers examined effect sizes using Cohen's d and found the data challenges straightforward applications of the variability hypothesis across species. This work represents how AI and rationalist communities are applying data science approaches to revisit contentious scientific debates, using publicly available datasets and modern analytical frameworks to test hypotheses that have been politically charged since Larry Summers' 2005 Harvard remarks sparked controversy.