Marsabit & Mandera Families Defy Challenges to Invest in Education



An analysis of education outcomes in Kenya shows that just three factors, unemployment rates, electricity access and internet access, are correlated with rates of access to education in Kenya. In fact, these factors can predict 95% of the variance in the rate of access to education for counties in Kenya. While this may not be surprising, what is surprising is that Marsabit and Mandera, both semi-arid counties, have a marginally higher rate of access than their socioeconomic factors would suggest.

Additionally, looking into dropout rates and illiteracy rates reveals that some counties with the lowest access to education seem to have lower dropout rates for those students who do attend school.

Surprisingly for me, counties where access to education is the most difficult, that is, very few people have ever been to school, have lower dropout rates. For instance, Garissa, Wajir, Mandera and Isiolo, despite being among the most marginalised in terms of access to social facilities, have drop out rates nearly as low as Kiambu, Mombasa and Nairobi's where access to education is higher.

HOW TO READ THESE MAPS

Illiteracy rates

Illiteracy

Dropout rates

Dropout rates

Both metrics

Illiteracy

Dropout rates

Illiteracy

Dropout

Garissa, Mandera, Wajir, Marsabit and Turkana have a high rate of people who have never been to school.

West Pokot has the highest dropout rate. Counties such as Garissa, Wajir, Mandera and Marsabit which have among the highest number of people who have never been to school have low drop out rates relative to the rest

In counties with the lowest education access, people who do go to school are then more likely to finish school

Among people who have ever been to school in the counties with low access to education, people were less likely to drop out once they began school than in the counties where education may be more accessible. That means, people in these counties with low education access who have been to school are then more likely to finish school.

So far, there has been coverage on the counties with the highest levels of illiteracy. Much of the coverage is academic and involves surveys which the researchers have done. There is opportunity for focused analysis of the universe of government-provided data to report on why Marsabit and Mandera appear to have higher access to education than suggested by speaking to teachers, education officials and parents from the region. There is also opportunity to present my new findings in a way that is more reader-facing than what has been published by academic experts and with more of a robust foundation in data than is the case in news articles such as this one about the difficulty of access to education in some counties.

You can view the distribution of the counties of interest against these socioeconomic factors one by one in this notebook.

Access to education rates in the counties are low compared to many other parts of the county but they are still higher than expected given the socioeconomic factors which the model reveals as important for predicting education.

Two education experts say the disparity is worth looking into. To their knowledge, given these areas are arid and semi-arid areas and also struggle with insecurity issues from terrorism and banditry, students have a difficult time accessing education. If the data indicates that the dropout rates are lower, it could be attributable to two major things:

  • Efforts by non-governmental entities to support the retention of students in schools
  • Given students in these communities face difficulties accessing food at home, they could be more incentivised to stay in school despite the hardships. For instance, some studies in urban slums in Kenya have shown that more children attend and stay in school is food is offered. It would be worth looking into whether this is a factor in this case too

In more exploratory interviews, a teacher in a semi-arid county, told me that they struggle with transition rates from one level of education to the next, e.g. primary school to high school. However, once the child is enrolled at the new level, they are likely to complete. He says some reasons for this are:

  1. Devolution: The decentralization of the Kenyan government a decade ago in 2013 has led to increased power for each county to shape its educational landscape.
  2. The work of NGOs and human rights activists
  3. Increased economic opportunities: With more economic development, parents have seen the impact of education on people’s lives and are keen to have their children benefit from the same opportunities.
  4. Reduction in nomadic practices allows children to stay in school. He says as recently as 2014, students would register for their examinations, go away to herd cattle and then return to do their exams. But that has changed today

Other experts speculated that higher than expected access to education could be because students in semi-arid counties are more dependent on food provided by schools and thus more incentivised to return than in the counties where access to food is easier. Secondly, they posited that there are lower opportunities for entrepreneurship in the arid and semi arid areas, meaning children have less opportunities for what to do outside of school.

A Kenya-based education expert also flagged that Coastal communities could be facing the opposite problem: access to education is lower than expected given the socioeconomic factors. Indeed, the model reveals the same, with Mombasa County, a major city in the country, appearing among the counties with lower access rates than expected.

It is possible from this analysis that for the portion of the population that gets a chance to attend school in semi arid counties, more complete their education than in majority of Kenyan counties. On the other hand, that means, people in these counties with low education access who have been to school are then more likely to finish school.