Improve data-driven school selection

Launched in 2017, the National Education Management Information System (Nemis) has met the dream of managing and automating education data and other related administrative functions.

The main objective of the portal was to help the Ministry of Education to gather accurate and real-time information on learners and learning institutions.

Such a system is meant to gather statistics from schools by following people, models, methods, procedures, processes, rules and regulations.

It also relates with the emerging computer technology to get all these functions to work together to provide comprehensive, integrated, relevant, reliable, unambiguous and timely data to education leaders, decisions makers, planners and managers to efficiently achieve the set goals.

Nemis, as is typical of any new technological intervention that is deployed en masse, has had some bottlenecks ranging from data incompleteness to technical incapacity of users and lack of synchronisation with data on the Integrated Population Regulations Systems (IPRS).

LESSONS

The IPRS was launched in 2015 to store the data of all Kenyans at a central location for easy electronic access by institutions, including private corporations that provide crucial and sensitive services.

This year, the ministry activated a module for Form One admission to secondary school.

There was, however, a backlash and this offers valuable lessons as we drive forward in adopting a culture of data-driven decisions in both public and private corporations.

Lesson 1: Digital transformation is driven by consumer education.

The ministry directed that all admission letters for three categories of schools, apart from sub-county schools, be downloaded from the Nemis website.

Bearing in mind the remote location of many parents and their level of digital literacy, it will be worth the effort for the ministry to look at the data from the portal and ascertain how many of them actually downloaded it as directed.

There was a need to provide prior education through mass media on how to download the letters.

Lesson 2: Behavioural metadata is critical.

What story does the data speak? From the previous data in the ministry’s possession, what is the admission trend?

Do students report to the schools which they were selected to join?

For those who don’t, how do they settle on the school they eventually report to?

For schools with a substantial number of no-shows, how do they fill the void?

Who is the key decision maker in selections — the student or the parent?

Why do many parents prefer to take their children to schools of their choice? Who, then, should the selection target?

Lesson 3: Weigh algorithmic versus humagorithmic selection.

The ministry states that the selection is computer-generated.

Well, but the data is input by humans, and the algorithms are tuned by humans, too.

To what extent do the human and algorithms work together?

How optimised is the selection algorithm to avoid true positives and false negatives — that is, instances where one is posted to a school they didn’t select?

Does the selection criteria match the parents’ preferences? How involved are parents during the selection criteria?

Can the enforcement happen at this stage, which is usually at the beginning of the final Standard Eight term, where the consequences of selection are clearly spelt out to manage anticipations?

Lesson 4: Location intelligence is important.

How well does Nemis make use of geographical information mapping?

In the spirit of regional balance and cultural adaptation as a criteria, it would be worthwhile to analyse longitudinal impact of education performance of learners from different environments for future placement considerations.

There are instances where pupils were selected to join county day or mixed schools far away from home, practically forcing a parent to rent a house for them.

Lesson 5: Technology enables strategy.

Availability of data does not necessarily mean the data will speak for itself.

Despite the good work so far in setting up an efficient data collection system, it’s time the ministry invested in the right technology to scale the usage of the platform for multiple access.

Computer technology gives technical support to the education management information systems only when provided with the right people, with the right information and at the right time to make the best decisions, planning and monitoring in the best interest of organisation.

Mr Oriedo is a data scientist at Predictive Analytics Lab and an author. [email protected]