Why girls fall behind in maths in adolescence: what the Indian evidence says

Conclusion. In India the gender gap in maths is small in primary school and opens in adolescence. The Indian studies point to four causes, each checked against the source:

Teachers: teachers’ gender stereotypes.
Family spending: families spend less on girls’ education.
Household norms: gender norms at home.
Stream choice: fewer girls choose science in Classes 11 and 12.

International studies add two more: girls’ lower confidence in their maths ability, and competitive test settings.

This note complements The Lost Girls of Navodaya (../lost-girls-of-navodaya/README.md). That paper measures where girls fall away in the Navodaya system. This note summarises what other research says about why. Every claim below was checked against a downloaded copy of the source. Where the working summary this note began from went beyond what the source says, the correction is noted.

The gap opens in adolescence

A study that followed children in Andhra Pradesh and Telangana found gender gaps in learning are “either absent or small in absolute magnitude” at ages 5 and 8, and widen “particularly between the ages of 12 and 15” (Singh and Krutikova, 2017). Household survey data on over 2.3 million rural children aged 8 to 16 find a maths gap, but not a reading gap, that grows with age (Das and Singhal, 2023). The gap varies sharply by state: northern states lag, while some southern states show a reverse gap.

Six explanations

#	Explanation	What the evidence shows	Source
1	Teachers’ stereotypes	In Class 9 in Andhra Pradesh and Telangana, girls taught by maths teachers with more biased beliefs about gender and ability fell further behind boys over one school year. The effect is strongest for middle-performing students and in classes where most students are boys. It holds for male teachers; for female teachers it is not statistically significant. Girls also came to like maths less, but that explains only a small part of the effect.	Rakshit and Sahoo (2023)
2	Less family spending on girls	In the Andhra Pradesh and Telangana sample, families spent about twice as much on boys’ education as on girls’. At 15, 74% of girls were still enrolled against 81% of boys, and boys were more likely to be in private school.	Singh and Krutikova (2017)
3	Which school a child attends	Earlier scores, enrolment, spending and the particular school a child attends together explain about 60% of the maths gap at 15 in India, leaving about 40% unexplained. Across the study’s four countries, half to two-thirds is explained.	Singh and Krutikova (2017)
4	Household gender norms	The size of the gap is linked to household practices and gender norms in the district.	Das and Singhal (2023)
5	Stream choice	Nationally, girls are about 20 percentage points less likely than boys to study science or commerce in Classes 11 and 12, and the gap remains after controlling for past test scores.	Sahoo and Klasen (2021)
6	Confidence	Across PISA countries, high-performing girls do worse than high-performing boys in maths, and girls believe less in their maths ability even when they perform as well. The high-achiever gap disappears when girls and boys report similar confidence and anxiety.	OECD (2015)

Two further findings, both from outside India:

Chile: the maths gap nearly doubles between Grades 4 and 8. Parental background, classroom factors and teacher gender do not substantially explain it. Girls rate their own maths ability lower than boys with the same scores (Bharadwaj et al., 2012).
Competitive testing: the shortage of girls among top maths scorers may partly reflect how boys and girls respond to competitive test settings. The evidence is mainly from laboratory experiments (Niederle and Vesterlund, 2010).

Corrections to the working summary this note started from:

The teacher effect: it rests on a broader index of teachers’ beliefs about gender and ability, not a single “boys are better” question. It holds for male teachers only.
The grade 4–8 doubling: this is Chile only. Bharadwaj et al.’s multi-country evidence is PISA at age 15. The paper also says its explanations do not substantially account for the gap; the summary had them explaining “some” of it.
The unexplained 40%: “the school a child attends” means the particular school, not school type.
Competitive testing: this is hedged in the source (“in part may be explained”).

What this means for the Navodaya findings

The Indian evidence fits the pattern in The Lost Girls of Navodaya: girls are strong at entry but thinner at the top by Class 10, and fewer strong girls choose engineering. Two of the causes are ones a programme can act on directly: teacher beliefs, and girls’ confidence at equal performance. The wise interventions described in that paper target the second.

Measuring socio-emotional skills

Define the skill first, then pick how to measure it. Self-reports, teacher reports and performance tasks are each imperfect in different ways, and none is yet fit for comparing schools (Duckworth and Yeager, 2015).

Resource	What it covers
Kautz, Heckman, Diris, ter Weel and Borghans (2014), Fostering and Measuring Skills (OECD; also NBER WP 20749)	Review showing that achievement tests miss character skills, whose predictive power rivals that of cognitive skills.
Kankaraš and Suarez-Alvarez (2019), OECD Education Working Paper No. 207	Framework for the OECD Study on Social and Emotional Skills (Big Five based; ages 10 and 15).
Duckworth and Yeager (2015), Educational Researcher 44(4)	Compares self-report, teacher report and task measures, and their limits.

References

Sources were downloaded and read before being cited, on 30 September 2026. [download] after a reference is the copy that was read; it opens only for someone signed in with an avantifellows.org Google account. Checksums are in SOURCES.md.

Bharadwaj, P., De Giorgi, G., Hansen, D. and Neilson, C. (2012). The gender gap in mathematics: evidence from low- and middle-income countries. NBER Working Paper 18464. download
Das, U. and Singhal, K. (2023). Solving it correctly: prevalence and persistence of gender gap in basic mathematics in rural India. International Journal of Educational Development 96, 102703. download
Duckworth, A.L. and Yeager, D.S. (2015). Measurement matters: assessing personal qualities other than cognitive ability for educational purposes. Educational Researcher 44(4), 237–251. download
Kankaraš, M. and Suarez-Alvarez, J. (2019). Assessment framework of the OECD Study on Social and Emotional Skills. OECD Education Working Paper No. 207. download
Kautz, T., Heckman, J.J., Diris, R., ter Weel, B. and Borghans, L. (2014). Fostering and measuring skills: improving cognitive and non-cognitive skills to promote lifetime success. OECD; NBER Working Paper 20749. download (OECD), download (NBER)
Niederle, M. and Vesterlund, L. (2010). Explaining the gender gap in math test scores: the role of competition. Journal of Economic Perspectives 24(2), 129–144. download
OECD (2015). The ABC of Gender Equality in Education: Aptitude, Behaviour, Confidence. PISA, OECD Publishing. download
Rakshit, S. and Sahoo, S. (2023). Biased teachers and gender gap in learning outcomes: evidence from India. Journal of Development Economics (read in its IZA DP 14305 working-paper version). download
Sahoo, S. and Klasen, S. (2021). Gender segregation in education: evidence from higher secondary stream choice in India. Demography 58(3), 987–1010. download
Singh, A. and Krutikova, S. (2017). Starting together, growing apart: gender gaps in learning from preschool to adulthood in four developing countries. Young Lives Working Paper 174. download