The Power of Many Small Sex Differences — Cognition, Personality, and Interests¶
Source: Nature Scientific Reports Date Published: May 27, 2026
Study Design¶
A large-scale Swedish study (N = 2,767, aged 35–45) investigated whether the combination of many small individual-level sex differences can collectively produce substantial predictive power.
Key Finding¶
The core result: 80% of individuals were correctly classified as male or female by combining just 13 measures spanning cognitive abilities, personality traits, and vocational interests. The authors frame this as "many small differences can collectively create a large effect."
Effect Sizes¶
| Domain | Effect Size (Cohen's d) |
|---|---|
| Interests (things) | 1.11 (large) |
| Spatial ability | ~0.70 (medium) |
| Other cognitive measures | 0.10–0.44 (small) |
| Personality measures | 0.10–0.44 (small) |
Progressive Classification Accuracy¶
The logistic regression model showed progressive improvement as more domains were added:
- Cognition only: 67% accuracy
- + Male-typed tasks: 70%
- + Personality: 74%
- + Interests: 80%
Occupational Implications¶
The combined measures predicted 22% of the variance in occupational sex segregation — a substantial figure given that occupational choices are influenced by countless social, structural, and economic factors beyond individual differences.
Interpretation¶
The study provides evidence that small individual differences, while negligible in isolation, are collectively powerful predictors of categorical membership. This challenges the narrative that sex differences are "trivially small" by showing that multivariate combinations reveal a far more differentiated picture than any single measure.