Approximately 85 percent of primary school-age children in western Kenya enrol in the first year of primary school, but only about a third of them stay on the final year. In 2001, the Year 6 dropout rate for girls was 10 percent (the rate for boys was significantly less at 7 percent).
Policymakers in Kenya have frequently focused their attention on increasing school enrolment or improving the attendance rate of schoolteachers. However, there was also a need for them to address the high levels of students leaving education by the age of 10.
In 2001, the Dutch charity, ICS, introduced a merit-based Girls’ Scholarship Programme (GSP) to help fund and encourage girls’ education.  It was trialled in two rural Kenyan districts, Busia and Teso. The procedure was that “girls who scored well on academic exams at the end of 6th grade had their school fees paid and received a cash grant for school supplies over the next two years”. Scoring well, in this case, meant being in the top 15 percent. 
The broad objectives of the project were to:
- Promote female education.
- Improve teacher attendance.
The more specific objectives of the GSP were to:
- Estimate the impact of a scholarship programme on girls in Kenyan primary schools.
- Encourage female pupils to study harder, assuming that they were motivated to strive for scholarship opportunities.
- Help the families of high-achieving girls to cover these costs.
The scholarship schools were randomly selected from among a group of candidate schools. The difference in educational outcomes between the GSP schools and the comparison schools were assumed to result from the incentive of the scholarship.
For the first two years of secondary school, the GSP-winning girls received:
- A grant of KES500 (US$6.40) to cover school fees, paid to her school.
- A grant of KES1,000 (US$12.80) for school supplies paid directly to her family.
- “Public recognition at a school awards assembly held for students, parents, teachers and local government officials”. 
The public impact
The evaluation found that the existence of the scholarships boosted classroom effort and raised the levels of pupils’ exam scores. It also found that there was a positive impact on the girls with low pre-test scores, who had a less realistic chance of winning the scholarship, so it was not just the likely winners of the scholarships who benefited from the GSP. There was also a reduction of about 25 percent in school absenteeism among pupils and an impact on overall teacher attendance, which increased by 7.6 percentage points.
Stakeholder Engagement Good
The major stakeholder was the Dutch NGO, ICS, and its staff. Other stakeholders included the schools, teachers and students in the target areas, and the ICS donors from the Netherlands and elsewhere. In March 2001, after a random assignment of schools into the GSP and comparison groups, ICS staff met with school head teachers to invite schools to participate, which all of them chose to do.
Political Commitment Good
There has been enthusiastic commitment from successive Kenyan governments to the policy of widening children’s access to education. In 2003, the Kenyan president, Mwai Kibaki, ended the practice of charging fees for primary education, and paid subsidies to schools in order to fund free schooling.
Clear Objectives Good
The objectives if the GSP were set out clearly, both the high-level objectives, such as encouraging more Kenyan girls to progress to secondary education, and the specific aims, such as encouraging pupils to study harder and attain higher scores in their academic exams.
ICS-Africa has been operating in western Kenya since 1996 and had run numerous educational programmes. It was in essence a pilot programme, so its purpose was experimental, although it used a well-established method of randomised control testing to evaluate the performance of the GSP schools against that of the control group. The results of the GSP provided a strong platform for the policy of offering scholarships to Year 6 girls in Kenya.
The NGO had sufficient funding to carry out the GSP. It had experienced staff working in the area. They had established good relationships with the local schools in the region, through previous ICS programmes and other programmes being carried on at the same time.
It is a good example of measurement, since the results were monitored constantly and the results of exam scores were measured effectively to measure the success of the GSP. General academic achievement was captured in exam results, which were considered likely to be a sound objective measure. The programme raised exam scores by 0.19 standard deviations for the girls who were enrolled in schools that were eligible for the scholarship.
The programme also had a large impact on overall teacher attendance: in the combined Busia and Teso sample there was a 4.8 percentage point increase.
There was good alignment between ICS, the participating schools, policymakers and pupils. The local schools were very cooperative: there were 127 sample primary schools in the pilot areas, 64 of which were invited to participate in the GSP. A randomised selection of schools was made as to which schools went into the GSP and which into the control group.
The teachers’ attendance levels increased, partly because parents visited the school more often in order to assess the performance of students, and they started pushing teachers to increase their focus on their children's education.
ICS awarded scholarships to the highest scoring 15% of Year 6 girls in the GSO schools in each district, and there was a general enthusiasm for the principle of promoting female education.