How far is the girls’ school?

In the quarter of districts with the smallest girls’ share of government schools from middle grade upwards, by official designation, the average district enrolment rate for girls aged 5–16 is 35.1 per cent. This finding combines enrolment from Pakistan’s 2023 Census with school-location data for 123 districts. In the quarter with the largest girls’ share, the average is 71 per cent, narrowing the average gap to boys’ enrolment from 12.9 percentage points to 2.1.

Government schools are designated for boys, girls or mixed enrolment, which allows the two gendered networks to be mapped separately. Across Sindh, however, many government schools are mixed in practice, so the official designation understates the number of schools serving girls. The opening comparison is sensitive to how Sindh’s schools are counted.

The difference extends beyond school numbers to how far the two networks reach. In 101 of the 132 mapped districts, the nearest girls’ school teaching middle grades or above is farther away than the boys’ equivalent. At village level, the shortage is harder to miss: 33,859 of the 45,351 populated rural mouzas covered by the data have no girls’ middle school of any kind.

Building the school map

Pakistan has no single public map of its government schools. Each province maintains its own public repository, but the level of detail varies sharply.

Balochistan publishes school coordinates directly. Sindh and Khyber Pakhtunkhwa provide locations through their education departments’ school records, while Punjab and Gilgit-Baltistan publish lists lacking coordinates, requiring school locations to be reconstructed from their recorded settlement names.

Together, these sources locate 118,677 government schools across 132 districts, covering 94.5 per cent of the schools listed by the provincial records. Private and non-formal schools are excluded on account of uneven public coverage across provinces: Punjab documents much of its private sector, while other provinces publish little or none. Including them would risk comparing differences in record-keeping, as opposed to differences in access.

The 2023 Census has wider reach than the school-location data. It records enrolment among children aged 5–16 in every district, including the seven ex-FATA merged districts, where school locations are not published. The Census captures their enrolment, but the absence of school locations leaves them out of the distance analysis.

That matters because some of the country’s widest gender gaps are there: 34.4 percentage points in Orakzai, 32.6 in Mohmand and 30.2 in Bajaur. Without their school-location data, we cannot tell whether including them would strengthen or weaken the relationship between distance and enrolment.

Measuring the journey

For every populated one-kilometre grid cell, the analysis calculates the distance to the nearest girls’ and boys’ government school teaching middle grades or above. District figures are weighted by population, so a sparsely populated area does not count as much as a populated settlement.

The interactive map also estimates walking time, using a model that accounts for the difficulty of travelling through different terrain. Straight-line distance carries the national comparison, while walking time is also shown on the map and in Figure 7.

The focus on middle grades and above also matters. Gender labels can be especially loose at primary level: in Sindh, schools designated for boys commonly enrol girls. The split between the two government networks becomes more meaningful higher up the school system, but Sindh remains an exception throughout. Girls are enrolled in 6,269 of the province’s 6,998 government schools teaching middle grades or above, although only 1,609 are designated for girls. When Sindh schools are counted according to whether girls are enrolled, instead of how they are designated, the difference in distance to girls’ and boys’ schools falls close to zero in every district.

Same district, different access

Across the 132 districts with mapped school locations, girls’ middle-or-higher schools are farther away than boys’ schools in 101. Girls’ schools are closer in 30 districts, while Lodhran records an exact tie. The median district-level gap is 0.6 kilometres, with girls’ schools farther away.

But the median across mapped districts conceals much wider differences. Several of the widest district gaps are in Balochistan. In Zhob, the nearest girls’ school is 16.3 kilometres farther away than the nearest boys’ school; the gap is 12.2 kilometres in Washuk and 10.4 in Kohlu.

At provincial level, larger distance gaps go hand in hand with smaller girls’ networks (Figure 2). In Balochistan, the population-weighted median distance to a girls’ school is 5.7 kilometres, against 3.7 for boys. In Punjab, girls’ schools are marginally closer: 2.1 kilometres, against 2.3 for boys.

Remoteness does not explain the full difference. Boys and girls within the same district share much of the same terrain and settlement pattern. Comparing the two networks narrows the question from remoteness in general to how government provision differs between them.

Distance is one way to see the gap. The village-level Mouza Census gives a more immediate measure: whether a school of any kind exists within a village at all.

Of the 45,351 populated rural mouzas covered by the Mouza Census, 33,859 have no girls’ middle school. In Balochistan, 90.1 per cent have none; in Sindh, 80.9 per cent; and in Khyber Pakhtunkhwa, 76.1 per cent.

The disparity appears even before middle school in Sindh. About four in ten populated rural mouzas have no girls’ primary school, compared with roughly one in ten without a boys’ primary school. By middle grades, the share of mouzas without a girls’ school rises to four in five.

The two measures answer different questions. A village without a girls’ school may have one nearby, across its boundary. The distance measure captures that wider reach; the Mouza Census shows when the school is absent from the village itself.

Why the road to school ends sooner for girls

The distance gap on the map reappears in household accounts of why rural children leave school. Cost barriers are reported at nearly even rates for girls and boys: “too expensive” is cited for 16 per cent of girls who left school and 15 per cent of boys. Distance, parental permission and the absence of female staff are far more heavily concentrated among girls. “Too far away” is cited for 12.6 per cent of girls, more than twice the 4.9 per cent for boys. “Parents did not allow” rises to 13.7 per cent for girls and barely registers for boys, at 0.1 per cent; “no female staff” is cited for 6.5 per cent and 0.2 per cent respectively.

These figures cover rural children who had been enrolled and later left school, and are not available district by district. Among children still attending, girls and boys report similar travel distances. That apparent parity may partly reflect who remains enrolled: some girls have already left school because the journey was too far. The journey looks more equal among the children who are still making it.

Where access shows up in enrolment

Across districts, the median enrolment rate is 48.8 per cent for girls and 60.6 per cent for boys. The median gender gap is 9.4 percentage points. The widest gaps sit where girls’ enrolment is already low. In Hangu, just 38 per cent of girls aged 5 to 16 are enrolled, compared with about 71 per cent of boys. A five-point gap can accompany various overall enrolment rates, from near-universal enrolment to a district where fewer than half of girls are in school. Figure 5 therefore keeps the underlying overall enrolment rate alongside the gaps.

Figure 5 · Map · static

Where the girls are missing from school

Enrolment rates by district, ages 5–16

District	Girls enrolled (%)	Boys enrolled (%)	All children enrolled (%)	Gap (pp)
Abbottabad	86.6	87.8	87.2	1.2
Attock	80.8	82.2	81.5	1.4
Awaran	46.8	49.4	48.1	2.6
Badin	29.6	46.5	38.6	16.9
Bahawalnagar	66.1	65.1	65.6	-1.0
Bahawalpur	59.2	59.4	59.3	0.3
Bajaur	20.5	50.6	35.7	30.2
Bannu	39.2	61.5	51.2	22.3
Barkhan	24.4	56.8	41.7	32.3
Batagram	41.2	64.3	53.2	23.1
Bhakkar	67.8	74.0	71.0	6.3
Buner	50.3	74.7	63.1	24.4
Chaghi	27.8	32.3	30.1	4.5
Chakwal	83.9	85.6	84.8	1.8
Charsadda	64.7	75.6	70.4	10.9
Chiniot	67.5	73.8	70.7	6.3
Chitral	80.8	85.0	82.9	4.2
Dadu	47.9	56.9	52.5	9.0
Dera Bugti	20.1	33.5	28.2	13.5
Dera Ghazi Khan	47.5	57.7	52.7	10.2
Dera Ismail Khan	47.2	60.7	54.3	13.5
Faisalabad	79.6	79.5	79.5	-0.1
Ghotki	31.8	55.5	44.3	23.7
Gujranwala	81.4	76.3	78.8	-5.0
Gujrat	87.2	82.3	84.8	-4.9
Gwadar	61.3	69.7	65.8	8.3
Hafizabad	79.2	76.2	77.7	-3.0
Hangu	38.5	71.5	55.9	33.0
Haripur	84.0	87.4	85.7	3.4
Harnai	24.7	34.1	29.6	9.4
Hyderabad	65.7	66.7	66.2	1.0
Islamabad	84.7	83.1	83.9	-1.6
Jacobabad	30.1	45.3	37.9	15.2
Jaffarabad	30.0	40.9	35.6	10.9
Jamshoro	49.2	52.0	50.7	2.9
Jhal Magsi	26.2	40.7	33.6	14.5
Jhang	69.6	74.8	72.3	5.2
Jhelum	86.7	86.4	86.6	-0.3
Kachhi	19.3	26.7	23.3	7.4
Kalat	29.1	42.2	35.9	13.1
Kambar-Shahdadkot	38.6	49.6	44.5	11.0
Karachi Central	77.4	72.5	74.8	-4.9
Karachi East	74.0	67.7	70.6	-6.3
Karachi South	71.0	72.9	72.0	1.9
Karachi West	57.5	59.5	58.5	2.0
Karak	63.8	79.3	71.9	15.5
Kashmore	23.8	35.8	30.0	12.1
Kasur	76.1	74.6	75.4	-1.5
Keamari	54.3	58.3	56.4	4.0
Kech	51.6	59.7	55.8	8.2
Khairpur	48.3	62.5	55.5	14.2
Khanewal	69.9	73.1	71.5	3.2
Kharan	31.8	40.6	36.7	8.8
Khushab	74.8	79.8	77.4	5.0
Khuzdar	26.6	37.3	32.6	10.7
Khyber	29.3	63.2	47.3	33.9
Killa Abdullah	17.1	37.2	28.3	20.0
Killa Saifullah	22.2	49.7	37.0	27.5
Kohat	55.8	80.3	68.7	24.5
Kohistan	12.1	26.1	19.4	14.0
Kohlu	28.7	28.2	28.4	-0.5
Korangi	67.0	66.1	66.5	-0.9
Kurram	33.7	54.4	44.8	20.7
Lahore	79.0	76.3	77.6	-2.7
Lakki Marwat	43.7	68.8	56.4	25.1
Larkana	49.2	60.6	55.3	11.4
Lasbela	36.3	44.9	40.7	8.6
Layyah	73.9	78.6	76.3	4.7
Lodhran	60.6	62.7	61.7	2.1
Loralai	25.8	51.2	39.2	25.4
Lower Dir	65.6	78.9	72.5	13.3
Malakand	75.5	81.3	78.5	5.7
Malir	56.3	57.8	57.1	1.5
Mandi Bahauddin	77.8	76.0	76.9	-1.8
Mansehra	72.7	84.2	78.6	11.6
Mardan	66.1	77.6	72.1	11.4
Mastung	37.4	50.9	44.8	13.5
Matiari	44.8	54.6	49.8	9.8
Mianwali	67.1	78.9	73.2	11.8
Mirpurkhas	39.6	56.8	48.8	17.1
Mohmand	19.7	52.2	36.1	32.6
Multan	66.2	69.6	67.9	3.4
Musakhail	42.1	53.2	48.2	11.1
Muzaffargarh	51.8	58.6	55.3	6.7
Nankana Sahib	75.4	76.2	75.8	0.7
Narowal	86.3	84.0	85.1	-2.3
Nasirabad	21.6	31.4	26.7	9.8
Naushehro Feroze	57.2	63.9	60.6	6.7
North Waziristan	18.5	45.6	32.5	27.0
Nowshera	66.9	79.3	73.2	12.4
Nushki	48.5	57.8	53.4	9.3
Okara	72.8	73.3	73.0	0.6
Orakzai	26.4	60.8	45.0	34.4
Pakpattan	69.5	72.2	70.9	2.6
Panjgur	35.0	37.2	36.1	2.2
Peshawar	55.8	73.2	64.6	17.4
Pishin	30.9	46.6	39.0	15.7
Quetta	55.4	64.3	59.9	8.9
Rahim Yar Khan	51.8	55.3	53.6	3.5
Rajanpur	38.9	47.8	43.4	8.9
Rawalpindi	83.8	82.5	83.2	-1.3
Sahiwal	70.5	75.0	72.8	4.5
Sajawal	18.0	31.0	24.9	13.0
Sanghar	36.1	49.7	43.0	13.6
Sargodha	80.3	79.4	79.9	-0.9
Shaheed Benazirabad	47.4	61.5	54.5	14.1
Shangla	34.9	58.3	47.3	23.4
Sheikhupura	77.0	75.2	76.1	-1.8
Sherani	14.1	27.3	21.8	13.2
Shikarpur	32.7	41.1	37.0	8.4
Sialkot	88.6	85.4	86.9	-3.2
Sibi	34.2	40.3	37.3	6.1
Sohbatpur	38.0	51.7	45.0	13.7
South Waziristan	27.6	39.8	33.9	12.2
Sukkur	47.5	59.7	54.1	12.2
Swabi	67.0	80.8	74.2	13.8
Swat	57.6	75.2	66.9	17.6
Tando Allah Yar	35.6	45.1	40.4	9.4
Tando Muhammad Khan	30.7	43.1	37.3	12.3
Tank	33.1	57.4	45.5	24.3
Tharparkar	29.3	47.4	38.9	18.1
Thatta	21.2	26.8	24.0	5.6
Toba Tek Singh	79.5	78.2	78.8	-1.2
Tor Ghar	26.5	60.4	44.2	33.9
Umerkot	34.5	53.9	45.0	19.4
Upper Dir	53.5	73.0	63.6	19.4
Vehari	66.2	67.9	67.1	1.6
Washuk	22.3	32.4	27.9	10.1
Zhob	38.4	50.7	45.2	12.3
Ziarat	32.6	41.3	37.1	8.6

The geography of enrolment closely tracks the geography of access. Across the 123 districts with both measures, girls’ enrolment is lower where the nearest girls’ middle-or-higher school is farther away (Figure 6). Each doubling of distance goes with about 12 percentage points fewer girls in school; for boys, the corresponding figure is about 9 points.

The analysis cannot distinguish the two slopes reliably enough to say that distance weighs more heavily on girls than boys. Household accounts show a different pattern, where distance is cited more than twice as often when girls leave school. Because the comparison runs across districts, distance is tangled with other forms of deprivation that also depress enrolment.

The next comparison calculates the gaps within each district, then compares them across districts. Boys and girls then share the same terrain, remoteness and much of the same local context, while the distance between their school networks can still differ. Where girls’ schools are farther away than boys’ schools, the enrolment gap is also wider (Figure 7). The relationship remains after accounting for material deprivation, although a rank-based check gives more weight to poverty than distance.

The result remains suggestive. Social norms that keep daughters closer to home may shape both whether girls remain in school and where girls’ schools are built.

Who the school network is built for

The strongest relationship in the analysis appears when districts are ranked by the girls’ share of the middle-or-higher government network. In the lowest quarter, girls’ schools make up an average of 21.2 per cent of the network and the average district enrolment rate for girls is 35.1 per cent. The progression runs through all four quarters: average district enrolment rises to 46.2 per cent, then 56.3 per cent, before reaching 71 per cent in the quarter where girls’ schools account for an average of 55.4 per cent of the network. Over the same progression, the average gap with boys narrows from 12.9 percentage points to 2.1.

The ordering is striking, but school supply and demand are difficult to prise apart here. Districts that devote more of the government network to girls may also be places where families, officials and local politics are more supportive of girls’ education. The association is the strongest in the analysis, but the chart cannot tell whether greater provision raises enrolment, stronger demand produces more provision, or both.

Sindh complicates the ranking. If schools are counted by whether girls are enrolled instead of their designation, the ordering across quarters breaks. Yet within Sindh, the relationship between distance and enrolment remains under either definition.

Across the different measures, the imbalance appears at several scales: in the share of schools designated for girls, in the distance to the nearest school, in whether a village has one at all, and in enrolment itself. No single measure settles why girls are missing from school. Together, they show how closely the geography of the school network is bound up with who reaches the classroom.