Let us prioritise investment in Education Management Information Systems
My work, in an information-oriented Education consulting company, has allowed me the opportunity to view EMIS work, in Africa and elsewhere, sometimes contributing to specifications and design but more frequently as a user, asking questions (reasonable questions, I think). So, for example:
We work in evaluation and reviews using standard system-wide data on progress: What is happening to enrolment? completion? performance (on what measures?) PTR etc. etc.
We work with decentralised administrations asking for disaggregated versions of the same data for their District or Province and working data that can help prioritise support or inspection, and with increasing decentralised responsibilities, inform the same sort of planning at that level.
We have worked alongside administrations to improve efficiency, decrease waste, and reduce the risks of malpractice, for example in ordering and distributing text-books and materials or managing the deployment and careers of teachers. The role of improved information system in reducing teacher lateness, absenteeism and ghosting is well known. But so is the problem of counting students to inform the calculation of per capita grants or textbook supply.
From Ministry down to the schools from aggregates statistics to counting students there are common issues about getting data right, or close enough (how close is close enough for capitation grants?),and in good time (or can we use last year’s?).
I would like to pursue two thoughts in front of the concerned community, hopefully controversial enough for this medium but based on these experiences.
The term EMIS has become an obstacle to sensible thinking and development of information systems to support education in general, and in developing countries in particular.
Whilst this may be a forum in which the shorthand acronym can be deconstructed sensibly and with understanding, I encounter the term most often in the practice of education development work. In that context the term has become a black box possibly an off-the shelf-solution or one-off procurement that will then squirt statistics into the Planning Department. It hides the day-to-day complexities of collecting data and making things work, but also buries the developmental challenges of building the culture and practices where information is feeds into policy, planning, management and administration.
Education development discourse is littered with portmanteau words, and there is a general risk that they shift in meaning across different languages and cultures: “EMIS” seems particularly vulnerable to this so I urge you all to eschew it. Take a sentence that say “what information?”, “for who?”, maybe “why?”
Administrative sub-systems have potential for development pay-off and provide the best source of reliable statistical data. Get them talking to each other.
I would urge more support for information systems that address essentially administrative tasks but that typically involve important sets of data in a context where they are regularly maintained and validated. Consider the system that might sit in a central or sub-national administrative office keeping the data required to pay teachers. Such systems have been prioritised to assist reliable and timely payment and have helped some countries weed out ghost teachers.
They are conceptually simple but the data are of such importance to those concerned teachers who want paying that there is an imperative to keep the data up to date. These are almost certainly the most reliable data available on teachers, if you want to count them. But in too many cases there will be a parallel exercise, perhaps based on a school census paper form that asks the principal to count teachers and their attributes to feed into the national EMIS exercise.
Collecting, and storing, the same data twice is obviously inefficient and the source of inconsistencies, but is remarkably common. I have recently been working in countries that send per-capita grants (i.e. grants based on the number of students enrolled) directly to schools. Despite their remoteness and other challenges the Ministry of Finance has undertaken an annual data-collection exercise, duplicating that of the national EMIS in the Ministry of Education. The point is not just the wasted effort but the resulting arguments caused by the inconsistencies (which result, usually, in lower than expected payments). Reconciling the data, or the Ministry of Education using the Finance data for its national returns, seems impossible!
Other administrative task in which we see ad hoc initiatives to computerise the core processes, and which collect, use or generate important sets of data include: ordering and distributing textbooks (schools, numbers of students by year and subject); examination administration (schools, students by subject); in-service professional development (teachers by subject and qualification), etc.. That is before we consider situations in which individual schools use a school management systems which typically have records of students (by age, year, subjects studied etc.), teachers and so on.
I argue for prioritising, within development context, investment in information systems for such administrative tasks, not only for the potential gains in efficiency and transparency of these important processes but also as part of a strategy for systemic information that serves the wider information needs, especially those often served by a stand-alone EMIS.
The reality is often that computerisation of administrative tasks is piecemeal, where creative and inspired people want to make their job easier and build something that works for them: a tower of Excel spreadsheets instead of paper, but getting a job done. We have small scale projects that build something more professional, such as a Higher Education Management Systems, which colleagues have been implementing to look at HE staff, their teaching loads, performance etc..
What seems to be lacking is the overarching view of education sector information uses and users that takes account not only of the need for the statistical outputs of the typical EMIS but also of all the others users and generators of information, including these administrative tasks, and at all levels from institution upwards.
We have tried to work in this way recently to develop “EMIS” specifications (that was the name of the contract), seeking to define the needs of all the identified stakeholders, not only those that need management reports and statistics but also those that need to administer services. This exercise started with a clean sheet of paper, which helps. In the more complicated, piecemeal, situations that normally pertain, I would argue the need to audit and understand the information needs of the widest section of stakeholders and the ways in which current actors, most obviously the administrative actors discussed above, collect their data. Mapping all the users (or potential users) from the Planning Department to the teacher taking a daily register of attendance, allows us to look at system-wide data requirements and sources or interchange between agencies.
I am not underestimating the difficulties, technically or in the cultures that like to keep information secret and resist sharing with others. And having an overview does not imply rushing to implement all the parts at once, we recognise the risks of that. Rather it allows manageable, piecemeal sub-systems to be developed, tested and embedded in ways that utilise and/or contribute to the “data landscape”.
Two thoughts/points are well taken. The key to EMIS success is to let data in a readable form reach as many educational stakeholders as EMIS and then let their feedback loop back to support the next round of data collection and reaching out. Even data is questionable, I would have it shared. Let EMIS team be embarrassed. My experience has told me that EMIS data if published only reaches the statistical elite, domestic or international. Synthesis report based on EMIS data is rarely done, therefore not even reaching anyone. Sad as it is, Billion dollar investment in EMIS in the past twenty years globally has made little difference in improving system management, administration or policy development or transparent M&E. Let us encourage a massive campaign of let "data out" and let feedback loop circulate.