Un blog sur l'éducation dans les pays du Sud – A blog on education in the developing countries

15 octobre 2018

Learning Letter by Letter in Mahbubnagar District, India

Filed under: Uncategorized — Étiquettes : , , , , — education_south @ 18 h 33 min

Learning Letter by Letter in Mahbubnagar District, India

 “Human beings are born survivors, but not born readers” writes Radhika Iyengar, Director of Education at the Center for Sustainable Development, The Earth Institute, Columbia University. In this Blog post, she explores the patterns and strategies children adopt to learn to read in a rural district in Telangana State in India.

There are many opinions on how children learn to read. Many of them are based in experiences of teaching kids from typical middle-income families in the west. I started reading stories aloud to my girls when they were about 4 months old. I live in the United States and the school that my child goes to has a fully stocked library which she has been exposed to since she was three years old.  Our kids are exposed to books early on. By the time our middle-class kids are about 6 months old they already know which side of the book opens and pretend to read, using their little fingers to trace the action of the story. Soon they even start copying the sound (of reading), repeating the sentences on the pages that they’ve heard so many times.

However, these experiences of my kids and others from similar middle income families do not speak to the reality that millions of children from low-income backgrounds are faced with. The chart below, representing several countries in sub-Saharan Africa, shows that at the end of 4thgrade, fewer than 30 percent of the children can read a paragraph (except Tanzanian children in Kiswahili). This implies that these children have been in school for 3 years or more and can’t read simple sentences in their local languages.


Source: World Bank (2018). Facing Forward: Schooling for Learning in Africa. Regional Study on the Quality of Basic Education,. Tokyo, 3rdSeptember 2018.  Accessed http://pubdocs.worldbank.org/en/832221535674096676/090318-TICAD-seminar-14-Sajitha-Bashir.pdf

Grade 4 curricula in the countries listed in this table aren’t so different from the curricula used in our high-income country classrooms – children are expected to be analyzing texts, doing comprehension exercises, and writing short essays. Therefore, the benchmark for this chart should be that 100% of children are able to read paragraphs since in Grade 4, like our middle income children. Yet as we see here, this is far from the case.

Among many things, we blame the non-availability of textbooks, lack of motivated and properly trained teachers, poor school attendance among students, and the list is endless. Everyone is busy conjecturing here.

Maybe it’s time to return to what science tells us about how children learn. Human beings are born survivors, but not born readers.  Our brains need to be taught how to read. Unfortunately learning is a painful exercise.


Source: Abadzi (2014). Reasoning with Cognitive Neuroscience for Better Learning Outcomes (part I): Learning Basics. Comparative International Education Conference. Pre-conference Workshop, Toronto, Canada.  

Also discussed inAbadzi (2016). Training 21st– century workers: Facts, fiction, and memory illusions. International Review of Education. Volume 62. Number 3.

Cognitive neuroscience tells us that each skill has a hierarchical process of learning.  These processes are sequenced based on the complexity level. The learning curves was first discussed by Speelman & Kirsner (2005, p. 122-123). This was further elaborated in the context of early readers by Abadzi (2016). The circles in the diagram above represent groups of processes involved in performing particular tasks. The S shape represents realities of preparing for a new task.

As the graph suggests, one learns speech comprehension first which could be followed by learning letters. Therefore children need to be taught one letter at a time. Letters then can be blended together to form words followed by connected texts or sentences. These letters need to be taught consistently and slowly since our working memory is limited and can hold only limited number of items at one time. Of course practice is key here. The more we practice, the less time we take to perform each reading task. The less time we take to perform each task, the easier it becomes to decipher meaning from text and develop reading comprehension skills.

What are the common opinions that we hear around us- that children need to understand the context, that some children are visual learners, that we need to start with sight words. But these opinions don’t fully account for the reality that many children in grade 1 are asked to learn an official language – often French or English – which they may have never heard before, accustomed to speaking in their mother tongue at home. Previous research shows that learning in local languages first help to acquire literacy skills faster that can be transferred to pick up other languages down the road. Fortunately, cognitive neurosciencetells us that there is only one way that our brain perceives information and processes it. Thanks to scientists we know which parts of the brain get activated first by the reading process, and where that visual information travels to connect with parts of the brain that process language. Therefore all we need to do is teach in a way that lets the brain do its job and avoids these conjectures.

We need to feed the brain little chunks of information and provide enough practice so that build to a point where we remember it instantaneously.  With corrective feedback and lots of practice we can train our brains to recognize and read text. We need to ensure that recognizing print gets embedded in our long-term memory and never gets lost. The earlier the better! With age our ability to grasp new languages reduced drastically. The brain of a 6 year old is more adept at learning languages than a 14 year old.


Source: Abadzi (2014). Reasoning with Cognitive Neuroscience for Better Learning Outcomes (part I): Learning Basics. Comparative International Education Conference. Pre-conference Workshop, Toronto, Canada.

Our brains are made to forget things. The more time elapses, the more we forget. What can save us from forgetting how to read is getting enough practice very early on so that reading comes automatically to us at an early age. The skill of reading then becomes like one of tying our shoelaces or riding a bike – very hard to forget even if we haven’t done it in a while. We do not have to make a conscious effort. Abadzi (2014) translates the cognitive psychological literature like the “forgetting curve” in the context of early readers. She elaborates that due to the limited working memory much information will get lost and if so, discusses how can brains be trained to retain information in this case becoming literate. In summary, by providing enough practice, children do not forget and they become automatic readers.

These principles are being used in a rural district in Telangana State in India. This academic year more than 2,000 children are reading one letter a day, and getting enough practice with ongoing corrective feedback from their teachers. The environment is the same, very challenging as always. The teachers do not have to invest time to create Teaching Learning Materials. The workbook created through multiple stakeholder meetings (including linguists, literacy experts, monitoring staff, teachers, textbook writers and others) makes things easier for teachers who face anywhere between 40-90 plus students in their classes.  The teachers follow 3 simple steps “I do”, “We do” and “You do” to make the children progress in their reading. Below given is a page of the book that everyone uses to practice reading. It’s a plain black and white book. We do not have big distracting color pictures in the book that cover most of the page in the pretext of making reading interesting for the children. The children are excited enough to come to school and practice to read in the limited time that they have. We do not ask children to write, since writing without understanding what the letters mean renders the exercise little more than an art activity. Children are better served learning to write once they know what it is they are writing. Here is a page from the Telugu book in which the letters are of different font sizes, generously spaced and helps to provide enough practice to children to read.


Source: A page from the Telugu Literacy Workbook.

Early results show that there has been improvement in the key indicators such as the correct letter sounds which shows major improvements in the schools that this reading program was implemented in. The graphs below show the differences in the treatment group versus control group per grade.



There is one Champion in this entire process and without mentioning him this write-up will not be complete – The District Collector, Mr. Ronald Rose.  He decided that since children are not learning “adequately”, they should be given a simple textbook which helps them to get “concepts” easily. The workbook serves this purpose. Yes context matters and so the focus is first learning the local language – Telugu – and not English. Once children have mastered the concept of connecting letters to sounds, learning in additional languages becomes much easier.


Social Media:

Radhika Iyengar, Director of Education at the Center for Sustainable Development, The Earth Institute, Columbia University in which she explores the patterns and strategies children adopt to learn to read in a rural district in Telangana State in India. Read more here:

@earthinstitute’s @radhika_iyengar on the patterns and strategies children adopt to learn to read in a rural district in Telangana State #India










10 novembre 2016

Interview with Luis Crouch on early grades reading projects

Filed under: Uncategorized — Étiquettes : , , , — education_south @ 20 h 35 min

There are 97 articles on the blog. To get to 100, interviews of senior international experts are conducted.


Interview with Luis Crouch, Chief Technical Officer, International Development Group, RTI International.

Interviewer : In the last few years, there was a shift from donors to concentrate on reading in the early grades in the developing countries . Interventions and measurement tools such as EGRA have been developing rapidly. What are the reasons of this success? To what demand this endeavor has responded?

Luis Crouch : I think that initial entry into school, and even primary completion, had improved, by the early 2000s, many people started to wonder whether this was sufficient. It was clear that many countries were providing their children with about as many student-years of schooling (not grades of education) as, say, France or Finland in 1970. Yet, to common sense and anyone with experience observing classrooms, and based on measurements such as PASEC or SACMEQ, it was quite clear that children in, say, Uganda or Zambia, were not learning at anywhere close to the level of French or Finnish children in 1970. This was an efficiency concern, in some respects. It was also clear that, by the late 1990s, access to knowledge was much more inequitably distributed than access to entry into school, using any measure of inequality—so there was an inequality concern as well. But, what to measure in order to track whether learning (and the equality of its distribution)? At that point a debate started as to what should be some ways to track, and have advocacy, around learning. Some advocated for a measure around 15 years of age, since that is the legal age of entry into the labor force in many countries. Some advocated that if measurement at 15 is what drives the policy agenda, and that this is what helps countries focus, then focusing on, say, achievement by Grade 9 is insufficient. It was argued that traditionally, in most developing countries, the problems start with the foundation years. This was evident from very high rates of both formal and informal repetition in Grades 1 and 2, and from casual assessments that started to be conducted in the middle of the first decade of the century—around 2005.  In the event, assessments such as ASER (from Pratham in India), EGRA (at first associated with RTI), and then Uwezo and others, started to gain popularity—not as an alternative to assessment in later grades, but as a complement. These typically restricted themselves, or at least made possible, assessment in Grades 2 or 3 (ASER goes on to later grades).

 The demand might be rather originated from donors than from the governments themselves. What is the level of appropriation of theses tools by the governements in developing countries? Can you cite an example country where the ownership of these initiatives was high?

I would say that there is little doubt that donors initiate a lot of this. This is somewhat understandable, as donor agencies respond to parliamentary constituencies that being to ask whether the money spent on foreign aid in education does much good. DFID and USAID were subject to parliamentary reviews in the mid 2000s that, basically, asked, what is our money doing to help children learn? In addition, some multilateral organizations such as the World Bank, which are ultimately ruled by their members, started wondering the same thing. (Actually, the World Bank started thinking about these things earlier than DFID or USAID.) Speaking generally, one could say the level of appropriation is not that high—yet. However, while the tools themselves may not be appropriated, it is clear that the tools influence national assessment approaches and systems. They awaken interest, at least in some countries, in measurement (in all grades, including the early ones). In some cases the assessments are used to evaluate countries’ own projects and efforts, without donor support. Cases that could be cited are Uganda, Egypt, South Africa, and Mexico. One should also note that sometimes the appropriation is not by governments, but by local NGOs and civil society, who use the assessments to pressurize governments around the issue of quality. It is also important to note that in some cases there might be a backlash against some of these ideas.

Small scale interventions and experimentation have been multiplicated. However, there seems to be little coordination. For instance, Mali has had around 10 EGRA evaluations. The way the donors and NGOs operate in this field is not optimal in terms of political economy. What could be done to improve the coordination of early grades projects?

I think in some ultimate sense the responsibility is with the countries themselves, ultimately. If there is too much assessment going on, and it is interfering or causing confusion, the countries themselves ought to, in some sense, resist. However, that is easier said than done. Donors can be very persuasive, or can even withhold their enthusiasm for a country, or even the funding, if there is not some assessment performed. So, to some extent it can also be the collective responsibility of the donors to coordinate with each other. This can best be done at the country level, instead of having the home offices in Washington, London, or Paris (to name the cities where the donors with most responsibility for assessment sit) coordinate with each other. Finally, the NGOs themselves could push back a little. I think it is their professional responsibility. That too is easier said than done, but is at least worth mentioning.

Moreover, small-scale projects even when really successful are not always rolled/implemented on a large scale. Let’s take the example of deworming pupils in Kenya. It is a well-known example of successful measure, however at national level, there is still large proportion of pupils not benefiting from deworming (I could throw in data if you want). Small-scale interventions are relatively easy to implement, that’s the south face. Convincing a govt to change the national syllabus, textbooks and teachers training is something completely different, like climbing the north face. How do you achieve this, how do you ensure transitions from small scale to large-scale ?

That might be the most difficult issue of all. I would suggest a few things. First, don’t necessarily try to take the intervention (a particular implementation of reading in the early grades, say), but instead take to scale the fact that children should be learning to read by the end of Grade 2, and that there are specific things that can be done that do not necessarily require the specific “Brand” of NGO intervention. (E.g., make sure reading is an actual subject, and not just part of “language,” devote more time to it, make sure teachers have very specific support in how to teach reading, make sure that the reading materials for the children line up with the teachers’ lessons, etc.). Second, if possible, make sure that the intervention’s recurrent costs (that is, taking away the initial costs of design, research, evaluation, etc.) are lower or at least not much higher than what the country’s budget could absorb, at least assuming that the government takes education seriously and budgets reasonably. Third, make sure that the evaluations are as well-designed as possible, because, in spite of the role of politics and political economy, good evaluations can help. As the saying goes, having good data does not necessarily drive good policy, but having no data can often ensure bad policy.

Despite hundreds of workshops on reading, some key issues are not really discussed. In France, UK, the whole word method of teaching reading is not used anymore, after years of debate. Neurosciences experts like Stanislas Dehaene clearly said that this method is not the right one from a cognitive point of view. In the developing world, there is little debate on this. Most early grade reading interventions rely on phonics approach and syllabic method. However, no organization (donors or NGOs) clearly took position against the whole word method. Don’t you think international community does lack on courage on this topic in particular and in general when the time comes to take position?

Well, maybe.  Or maybe it is just good tactics.  But it varies a lot. In some countries the phonics (or some blend) method is actually the preferred method in the curriculum. But the approach is pretty theoretical and teachers do not know how to use it, and they do not know how to assess their children in the framework of phonics, namely, in terms of what in fact is in their curriculum. In these cases the donor or NGO intervention is just showing the counterparts how their own curriculum can be implemented better. In other countries, the whole language method is so entrenched that anything related to phonics cannot even be mentioned, though this is also decreasing in developing countries. This is all fairly ironic, because in languages that are more transparent than English or French, phonics is a no-brainer: the orthography of the languages is almost made for phonics!  So, it varies a lot. And maybe it is wise, good tactics, not to take this issue head on.

On languages, as you mention, there is a wide acceptance that teaching in national languages fosters learning how to read; and the use of national languages have been promoted as medium of instruction. That probably make sense from a pedagogical point of view but they are several aspects to it : diversity of languages and ethnic groups, political aspects and so on. For instance, a country where there is a large number of endangered languages is the USA (the natives languages), probably since the policy is assimilation. What do you think of national languages as medium of instruction in the developing countries?

This is a pretty « fraught » issue.  I don’t think it is only (or even mostly) a technocratic issue. From a purely pedagogical or technocratic point of view, and all other things being equal, my read of the evidence is pretty clear, especially in situations where the teachers are good at the home language: work to teach reading (and maybe other fundamental skills) in the home language for at least the first few years. But there are all sorts of practical difficulties in urban areas where there might be a big mix of home languages in the classroom. And, political issues (both in the « good » and not so good senses, e.g., in terms of nation-building and creating a unifying ethos, or in the « bad » sense of imposing one language on others) are also important. So I think one has to temper the pedagogical with the practical and the political. There is also the issue that sometimes what parents value about schooling is precisely the acquisition of skill in the language of commerce and power, and that might often be a Western language, or a regionally-dominant language such as Arabic. And, they might think that starting right away in that language is best. And, if a program of education that starts in the home language is poorly designed and implemented, the parents may well be right. What would « good » mean? A program that simultaneously ensures that the child is a really good reader in the home language, and has started to have a really good oral dominance of the Western language. Then the transition is pretty effortless, and the reading skills in the Western language would be easier to develop. But it seems to me that those circumstances are not often found in many countries. I can think of my own case. Spanish is my mother tongue and I was a good reader in it by age 8 or so. At the same time I learned oral English by being immersed in English with children of my own age, who did not speak Spanish, for hours and hours over many months. My transition to reading English, then, happened relatively effortlessly, with just a little help from my mother, who used a simple English reader to help me.  But, kids in rural Africa will typically not have these privileges. So, this is a difficult issue. But I think the « technical » case for simultaneous use of the home language in reading, oral acquisition of the 2nd language at the same time, and then transitioning to reading in the 2nd language, is pretty clear.

If there are things that might not work, there are others that could work. Accountability and reporting to the population on education results is often cited as a good way to improve the quality of learning. Do you think it is key? What do you think of the population led assessments such as UWEZO in eastern Africa?

I think that, in general, the notion that “pure” accountability could do it, is very attractive, because in some ways it is the least expensive thing you could do. I myself (as an economist) used to think that way: that if you have enough public accountability (and given the fact that some schools seem to know what they are doing even without NGO or government help), schools would copy good methods from each other, or teachers would seek out the methods and copy from other teachers in other schools. Much as say, barber shops copy popular styles from each other. But schooling is more complex.  And, therefore, I think that the experimentation of the last few years is showing us that this is not as easy as one had hoped. I think that in most cases it is necessary not just to pressurize schools, or have league tables, or parental activism, and so on, but also to support them with very specific, practical advice. Coaching, for example, has emerged as a pretty respected practice.

What other things (let’s say that would not cost much) can be done?

Well, one thing is to start thinking not in terms of cost per student but cost per completer. In some simulations I have been doing recently, I find that if you can improve reading in the early grades, and that leads to a reduction in repetition and dropout of reasonable magnitude, the expense per child may go up (coaching is not cheap!), but the expense per completer may go down, because the internal efficiency of the schooling system has gone up. If you make some very efficient interventions in the first couple of grades, then repetition should decrease, and since it is repetition that is partly responsible for dropping out, then the cost per completer will go down. Maybe not hugely, but likely by 10% – 15%, even if costs per child have gone up by 10% to 20%. That said, there are some aspects of the interventions that are not that expensive, but “simply” require some policy changes and political will. An example is making the decision to allocate more time to reading, so that children can master this fundamental skill.  That’s costless in some ways, it just means reallocating time around.

Do you think that the election of Donald Trump (that claims to reduce federal budget of education and very likely development aid) is going to jeopardize all the efforts and money invested in the early grades reading in the developing countries ?

It is much too early to tell!  Any policy stance that reduces spending on education assistance is likely to have negative consequences for the agenda. That said, we all know that donor expenditures are not always efficient. So, my argument is that efficient donor expenditures are defensible, and necessary, if the international community is going to continue to assist in making sure children are actually learning.

Interview by email by Pierre Varly

Luis Crouch, PhD, is a recognized international leader in providing high-level advice to governments involved in complex educational systems change. From 2011 to 2013, Dr. Crouch served with the World Bank’s Global Partnership for Education Secretariat as head of the Global Good Practices Team.

He currently leads work addressing important challenges in education, workforce and youth, and « Data Revolution for Development. » He provides input and oversight to key areas of work in all of the International Development Group’s themes.

Dr. Crouch is also researching fundamental issues at the leading edge of applied scientific work on education while continuing to pursue his policy advisory work with specific countries in areas such as school funding and educational decentralization.

This interview does not engage the responsibility of any organisation.

Propulsé par WordPress.com.