Women underestimate their spatial intelligence—even when they perform just as well as men

Women tend to underestimate their spatial intelligence compared to men, despite performing just as well on tests, according to new research published in Sex Roles.

People’s beliefs about their abilities often do not align with their actual skills. One area where this misalignment may matter is gender differences in self-assessments; women consistently rate themselves as less capable than men, even in the absence of true ability gaps. This tendency—called the hubris-humility effect—suggests men overestimate their abilities (hubris), while women underestimate theirs (humility).

Gabriela Hofer and colleagues explored this effect within spatial intelligence, a domain where men are commonly believed to excel, and which is critical for success in science, technology, engineering, and math (STEM). Yet prior studies rarely included both self-assessments and performance measures, making it unclear whether the effect reflects male overconfidence, female underconfidence, or both.

In addition to gender, personality traits like narcissism and honesty-humility may influence how people evaluate their own abilities. Grandiose narcissism, for example, is linked to inflated self-views, whereas honesty-humility reflects modesty and sincerity. These traits, like spatial self-estimates, are also gendered; men tend to score higher in grandiose narcissism and women in honesty-humility. Because self-beliefs predict career interests more strongly than actual performance, these patterns may contribute to persistent gender gaps in STEM.

The study included 208 German-speaking participants (103 women, 105 men; aged 18–37) recruited primarily from university settings. Participants completed a 90-minute in-person session where they first completed six spatial intelligence tasks, then estimated their own performance both for each task (proximal self-estimates) and overall (distal self-estimates).

These tasks covered three major facets: spatial visualization, mental rotation, and visuospatial memory. Each facet was measured with two well-established tests taken from standard cognitive assessment batteries. For example, in the paper folding test, participants mentally visualized how holes punched in folded paper would appear when unfolded. In mental rotation tasks, they determined whether rotated 2D and 3D figures matched a target figure. Visuospatial memory was assessed by asking participants to memorize and redraw maps with specific routes or marked buildings. All tasks followed standardized formats with strict time limits.

After the tasks, participants completed personality questionnaires measuring honesty-humility (via the HEXACO scale), grandiose narcissism (via the NPI-d), and vulnerable narcissism. They also completed a validated vocational interest inventory focused on Realistic and Investigative interests, two domains closely linked to STEM. To examine gender and personality effects, the researchers computed misestimation scores (difference between self-estimated and actual performance).

Contrary to popular belief, women and men performed equally well on spatial intelligence tests. However, women gave significantly lower self-estimates than men on both the overall (IQ-style) and test-specific performance ratings. This suggests that women underestimated their abilities, particularly in the mental rotation task, while men’s estimates aligned more closely with their actual performance. Thus, the data support female humility but not male hubris.

Grandiose narcissism correlated positively with performance overestimation, meaning individuals high in this trait were more likely to rate themselves higher than they performed. Honesty-humility was associated with more modest estimates, but only for test-specific (not global) estimates. Even after controlling for personality traits, gender remained a significant predictor of underestimation: women were still more likely to underestimate themselves than men.

In terms of career interests, men reported higher Realistic and Investigative interests, than women. Importantly, self-estimated spatial intelligence had stronger associations with STEM interests than actual test scores did. When controlling for both performance and self-estimates, gender still explained a significant amount of variance in STEM interests.

The authors note that their sample, though reasonably balanced and comparable to national statistics in Austria, was limited to university students, which may limit generalizability to broader or older populations.

The research, “Women’s Humility and Men’s Lack of Hubris: Gender Biases in Self-Estimated Spatial Intelligence,” was authored by Gabriela Hofer, Marla Hünninghaus, Jana Platzer, Sandra Grinschgl, and Aljoscha Neubauer.