41
The Monitoring of Learning Outcomes in Sub-Saharan Africa: Senegal Languages of Instruction and Teachers’ Methods in Senegal Grade 3 Classrooms

The Monitoring of Learning Outcomes in Sub-Saharan … · La langue officielle de la république du Sénégal est le français. ... le malinké, le poular, le sérère, ... “Apprendre

Embed Size (px)

Citation preview

The Monitoring of Learning Outcomes in Sub-Saharan Africa: Senegal Languages of Instruction and Teachers’ Methods in Senegal Grade 3 Classrooms

The Monitoring of Learning Outcomes in Sub-Saharan Africa: Senegal Languages of Instruction and Teachers’ Methods in Senegal Grade 3 Classrooms The William and Flora Hewlett Foundation Grant # 2008-3367 RTI International Project Number: 0212014 Prepared for William and Flora Hewlett Foundation 2121 Sand Hill Road Menlo Park, CA 94025 Prepared by Pierre Varly, Consultant, on behalf of RTI International 3040 Cornwallis Road Post Office Box 12194 Research Triangle Park, NC 27709-2194 The author’s views expressed in this publication do not necessarily reflect the views of the William and Flora Hewlett Foundation or RTI International.

Table of Contents Page

List of Tables ......................................................................................................... 3 

List of Figures ........................................................................................................ 3 

About This Report ................................................................................................. 4 

Brief Overview of National Languages in Education in Senegal ............................ 4 

The Place and Nature of Wolof Language ............................................................ 7 

Is Arabic a National Language? ............................................................................ 7 

Scope, Objectives, and Methodology of the Study ................................................ 8 

What Languages Do Pupils Speak at Home? ..................................................... 10 

Matching Home Language and LOI to Improve Schooling, Equity, and Quality ............................................................................................................ 12 

What Do Teachers Say About Their Use of Language in the Classroom? .......... 13 

What Does Observation Tell Us About Languages Used in the Classroom? ....... 14 

What Activities Do Teachers Use to Teach Reading? .......................................... 16 Group Versus Individual Work ............................................................................. 16 

How Is Class Time Spent When Learning to Read? ............................................ 18 

How Much Time Do Pupils Spend Reading? ....................................................... 19 

Pupils’ Literacy Environment and Opportunity to Read (OTR) Index ................... 20 

How Well Do Grade 3 Senegalese Pupils Read? ................................................ 22 

Do Teachers’ Activities and Language Use Have an Effect on EGRA Test Scores? .......................................................................................................... 22 

What Variables Have an Effect on EGRA Test Scores? ...................................... 25 

A Proposal of Framework for Measurement and Action ...................................... 26 

Other Factors ...................................................................................................... 27 

Results from the EGRA Pilot Survey: Wolof Versus French ................................ 29 

Conclusion .......................................................................................................... 29 

References .......................................................................................................... 32 

Annex 1: Matching Official Versus Implemented Curricula .................................... 1 

Annex 2: Regression Models ................................................................................ 1 

Annex 3: Correlations Between Activities and Test Score (letters per minute) ............................................................................................................. 1 

Annex 4: EGRA Sample Map and Language Map ................................................ 1 

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 2

List of Tables Page

Table 1:  Average proportion of language speakers in the EGRA dataset ......... 12 

Table 2:  Proportion of pupils speaking the different languages home in the PASEC sample ........................................................................................ 13 

Table 3:  Percentage of language used declaring using national language in the classroom ............................................................................................. 14 

Table 4:  Use of French and national language by lessons subject ................... 14 

Table 5:  Use of French and national languages in bilingual classrooms per grade ....................................................................................................... 15 

Table 6:  Teachers’ activities .............................................................................. 18 

Table 7:  Pupils’ activities ................................................................................... 18 

Table 8:  Availability of reading materials (%) .................................................... 20 

List of Figures Figure 1:  Partial reproduction of the data collection tool ................................... 8 

Graph 1:  Proportion of language speakers in each classroom ....................... 11 

Graph 2:  Relationship between the proportion of Wolof speakers and use of Wolof in the classroom ........................................................................ 16 

Graph 3:  Distribution of teachers group work ................................................. 17 

Graph 4:  Distribution of pupils group work ...................................................... 17 

Graph 5:  Distribution of reading activities by pupils in reading lessons .......... 19 

Graph 6:  Relationship between letter recognition and OTR Index in the LOI dataset .................................................................................................... 21 

Graph 7:  Letters and words per minute distribution ........................................ 22 

Graph 8:  Share of letters per minute variance by categories of variables ...... 24 

Graph 9:  Share of comprehension variance by categories of variables .......... 25 

Figure 2:  A framework for measurement and action ....................................... 26 

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 3

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 4

About This Report This report is a product of a grant by the William and Flora Hewlett Foundation to RTI International, to carry out an assessment of early reading skills in Senegal, and to study the link between use of language of instruction, language of reading, and learning outcomes. The data for the study was collected in 2009 using the Early Grade Reading Assessment (EGRA) instrument and methodology1. This assessment was conducted by RTI and FocusAfrica with assistance from Associates in Research and Education for Development (ARED) and the National Institute of Research for Educational Development (INEADE) in Senegal. The evaluation took place from May to June in CE1 classes (third year of schooling) across 50 schools in 11 regions of Senegal. The results analysis for reading in French is based on a final sample of 687 students tested (51% boys and 49% girls). Additionally, interviews with 50 school directors and 70 teachers from the same classrooms provided additional data for interpreting students’ performance. A custom classroom observation instrument was also developed to determine how much instruction actually takes place in the language of the assessment, and how often and under what circumstances teachers use the mother tongue. More details about this instrument are provided in this report, which looks at the results of the reading diagnostic in terms of the methods and language(s) used to teach both reading and other subjects.

Brief Overview of National Languages in Education in Senegal

The first recorded experimentations of national languages in education were conducted by a French teacher in the beginning of the 20th century. This teacher noted great improvement among his pupils when Wolof was used as the language of instruction (LOI); however, upon this finding, he was sent back to France by the French authorities.

In 1971, Leopold Sedar Senghor, former president of Senegal and member of the Académie Française, stated2 that replacing French as the medium of instruction is neither desirable nor possible, and that two generations would be needed to turn one of Senegal’s national languages into an efficient learning tool. He also asked what the situation would be four decades later,

According to (Shiohata 2010), “Access to employment in the formal sector is often denied to those who are not literate in French,” and likely in many other African Francophone countries.3. However, national languages, and especially Wolof in Senegal, have been

1 See www.eddataglobal.org for more information. 2 “Remplacer le français, comme langue officielle et comme langue d’enseignement, n’est ni souhaitable, ni possible. Si du moins nous ne voulons pas être en retard au rendez-vous de l’An 2000. En effet, il nous faudrait au moins deux générations pour faire d’une de nos langues nationales, un instrument efficace pour l’enseignement des sciences et des techniques.” 3 Access to a civil servant position is often based on the single results of a French language test, diplomas and interview.

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 5

valued and broadly used in the media and by political authorities. Several sources estimate that between 80% and 90% of the Senegalese population can understand and speak Wolof.

French is the official language of Senegal, and 6 languages out of 38 have a national status in the 2001 Constitution, as well “as any other codified language”4.

According to the law5: “L'éducation nationale est sénégalaise et africaine: développant l'enseignement des langues nationales, instruments privilégiés pour donner aux enseignés un contact vivant avec leur culture et les enraciner dans leur histoire, elle forme un Sénégalais conscient de son appartenance et de son identité.” Thus, it is not clear if children have the right to learn systematically in any codified national language. Follow-up decrees in the Kolda (1993) and St. Louis (1995) colloquiums concluded that universal enrollment could only be achieved by the introduction of national language instruction in primary school.

4 “La langue officielle de la république du Sénégal est le français. Les langues nationales sont le diola, le malinké, le poular, le sérère, le soninké, le wolof et toute autre langue nationale qui sera codifiée.” Editor’s translation : “The official language of the republic of Senegal is French. National languages are Diola, Malinke, Pulaar, Serer, Soninke, Wolof, and any other national language that will be codified.” 5 Loi n° 91-22 du 16 février 1991, article 6. Editor’s translation : “Education is Senegalese and African : By developing the teaching of national languages--privileged instrument for giving learners direct contact with their culture and ground them in their history--it also creates a Senegalese population are aware of its heritage and identity.”

Languages diversity, status, categories and dataAménagement linguistique dans le monde from Jacques Leclerc (Université de Laval, Québec) and www.ethnologue.com maintained by SIL International are the two main references sources on language diversity. They are largely used by the United Nations Educational, Scientific and Cultural Organization (UNESCO), which keeps its own database on languages in danger of extinction. The African Academy of Languages (ACALAN), founded by the Mali president in 2001 and sponsored by the African union, has a project that maps languages and primarily focuses on the domain of harmonization of coding. Africa remains a continent of oral traditions, and the codification of languages is relatively recent. ACALAN (2009) provides three main categories, “widely spread languages” (also referred to as “dominant language” in the literature), “lesser spread languages” (or “nondominant language”). and “vehicular cross-border languages”. Ten vehicular cross-border languages are identified by the academy in West Africa, including Wolof; Pulaar; and Mandingue, and 41 for the entire African continent out of 2,000 languages spoken. African languages are divided into the following different linguistic groups: “Nilo-saharienne, nigéro-congolaise, bantoue,khoïsane, chamito-sémitique.” Language coding is referred as ISO 639-3. In many countries, including Senegal, national census data collect information on ethnic origin rather than languages spoken, and the proportion of the population speaking a given language is hardly estimated in a reliable manner. SIL provides estimated figures on number of speakers of each language, but this information is derived from different sources and years. It is not possible to calculate a ratio from a unique source for a given year. In Senegal, ethnic origin cannot be assimilated to a language. For instance, the Lebou ethnic group speaks Wolof, but a proportion of the Pulaar and Serere populations do not speak their own languages and speak Wolof instead (especially in Dakar), although this assumption cannot be based on reliable data.

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 6

National languages are actually more used in the informal sector (adult literacy and community schooling) than in formal sector, where the objective is: “Apprendre à l’enfant à lire et écrire, compter communiquer et résoudre ses problèmes de la vie quotidienne dans sa langue.”, (PDEF)6. Introduction of national languages will be drawn on experiences from informal sector: “Les acquis capitalisés dans le non formel seront valorisés dans ce cadre”.7

Several attempts were made to introduce national languages in primary education, producing very little results according to a recent declaration from the Directeur de l’Alphabétisation et des Langues Nationales, El Hadj Meissa Diop: ‘’La question de l’introduction des langues nationales dans le système éducatif sénégalais est un problème, l’expérimentation a été faite mais sans atteindre les résultats escomptés’’8 .

Several specific9 problems are reported in implementing national languages in Senegal (Dia 2009, IDEA 2009):

• There is a lack of school textbooks available in national languages.

• The revision of curricula has been pending for a decade.

• There is a lack of available, adequate teaching materials in national languages.

• There is no significant teacher training.

For primary education, no official statistics are available on national languages provision, but IDEA identified 465 “bilingual” classrooms in 2008. In pre-primary schooling, progress has been made in the use of national languages within the Case des Tout Petits initiative. On 4,983 educators, 86.9 % teach in French, 9.4 % teach in Arabic, and 3.7 % in national languages, according to official statistics.

As a result, only a small proportion of the population is literate in its own language, according to the 2002 National Census. Assuming the ethnic origin is a proxy for the main language spoken home, only 4.3% of Pulaar people are literate in their own language, 2.6 % of Mandingue, 2.7% of Wolof, and only 1.3 % of Soninke. A total of 37.8% are literate in French, but less than 10% of the population speaks that language daily, a proportion hardly estimated in a reliable manner. Significant differences are reported between males (between 15% and 20% speakers of French) and females (less than 2% speakers of French).

6 Editor’s translation: “To teach the child to read and write, count, communicate, and resolve problems of daily life in his language.” 7 Editor’s translation: “Lessons learned from the informal sector will be used in this case.” 8 http://www.en24heures.com/senegal/introduction-des-langues-nationales-a-lecole-le-probleme-demeure-entier-selon-un-officiel/2010/08/27/23122. Editor’s translation: “The question of the introduction of national languages in the Senegalese education system is a problem. Experimentation has taken place but without achieving the expected results.” 9 UNESCO (2010) lists commonly reported problems in implementing national languages in Africa.

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 7

The Place and Nature of Wolof Language In Senegal, Wolof is widely used as the communication tool among different ethnic groups and by the media. It is predominantly spoken in Dakar. According to census data, Wolof people are the largest ethnic group in Senegal, and their language has been established as the most widespread communication mode. However, languages spoken at home are quite diverse (according to pupils’ declarations), and some10 complain about the possible Wolof supremacy in education. The proportion of books and literacy materials written in Wolof is greater than in any other national languages, according to our data and Shiohata (2010).

Code switch (the practice of mixing Wolof and French syntax in a same sentence) is a very common feature in daily conversation, in the media, and on Internet forums. Moreover, Wolof has its own grammatical structure but has acquired many foreign or non-indigenous terms (Shiohata 2010). In modern urban Wolof: “Nangadef, loubes, comment ça va ?, nice,” is a typical series of salutations including Wolof, Arabic, French, and English words.

Is Arabic a National Language? In Senegal, 95% of the population is Muslim and learning Koran is part of the traditional Senegalese education. Arabic is learned through three types of education: 1) parental education or tutoring, 2) Koranic schools (daara), and 3) Franco-Arab schools. This language is not used as a communication language (Shiohata 2010) by the population, but passages of the Koran are transcribed11 in phonetic language, making everyone able to recite them.

If Franco-Arab schooling is officially included in the education statistics and supervised by the government (via formal education), Arabic language has no official status in Senegal. A significant proportion of pupils attend Koranic schools prior to entering formal education, but no statistics are available. The Government of Senegal wishes to develop Franco-Arab education and took recent measures to improve Koranic schools’ management, including forbidding street begging12 and child mistreatment. Pupils enrolled in Koranic schooling (to be distinguished from purely child begging) are more likely to enter formal primary education (André 2009).

In order to diversify the school supply and get more pupils enrolled into formal education, it is not clear whether the government policy is to develop Franco-Arab

10 Tous les pédagogues du monde sont unanimes pour affirmer qu’un enfant scolarisé dans la langue maternelle a beaucoup plus de chances de réussir. Or le wolof n’est pas la langue maternelle de tous les Sénégalais ,” Le Collectif des Sénégalais de la Diaspora pour l’émergence des langues nationals. “http://www.seneweb.com/news/article/26138.php. 11 Wolof used to be coded using Arabic script but is now coded in Latin script. 12 A common practice is for children to be sent to Koranic residential school to study Islam with a master teacher (a “marabout”). These children, known as “Talibé,” are sent by the Marabout to beg for money in the street.

schooling, introduce state supervision and financial support of Koranic schools, or foster national languages education.

If 25.9% of the population declared being Arabic literate, according to the 2002 Census, this ratio may likely overestimate the proportion of people really able to read, write, and speak Arabic. Rather it could be viewed as a measurement of the knowledge of the Koran than an indicator of actual; Arabic literacy (Shiohata 2010).

In the Programme d’Analyse des Systèmes Educatifs des Pays de la Confemen (PASEC) dataset (made up of 150 schools), only 2.5 % of the pupils declared speaking Arabic at home in grade 2 and 0.5 % in grade 5, a lower proportion than for French language. Arabic is an LOI for a significant proportion of children through informal and formal education (Franco-Arab schools). Its place in Senegalese society and competing with official national languages are an issue. If the majority of the population speaks Wolof, the diversity of languages used in Senegal is a key factor in determining whether to use national languages in education. French, the official administrative language, is the unique official LOI in most of the government sponsored schools but is seldom spoken at home by pupils.

Scope, Objectives, and Methodology of the Study In this complex landscape, this study will investigate what languages are really used as the LOI in public schools and whether introducing national languages in the education system is an avenue for improving learning outcomes in addition to better teacher pedagogical methods. The study relies on quantitative data, but results are contextualized using qualitative information, such as research papers and press review.

Data was collected using the EGRA protocol, and a custom LOI observation instrument (see Figure 1). Additionally, interviews with teachers and school directors elicited information about classroom practices and language use. The sample of classroom observations and interviews is from 50 schools in 11 regions of Senegal; EGRA data was collected from14 pupils from each observed classroom. Public schools were not selected with weights proportional to their size (number of pupils). No Franco-Arab school was included in the data. Data is collected through direct observation in the classroom.

Enumerators in a classroom note every two minutes what teachers and pupils are doing and what language is being used. Information on the classroom equipment is also collected.

Figure 1: Partial reproduction of the data collection tool

Method Activity Snapshot observations

1 2 3 4 5 6 Teachers work with a group of pupils

Teachers speaks FR FR WO Teachers writes FR

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 8

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 9

Method Activity Snapshot observations

1 2 3 4 5 6 …

Teachers work with an individual pupil

Teachers speaks FR FR Teachers writes

Pupils work in a group Pupils speak FR FR FR

Pupils read aloud …

Pupils work individually

Pupils speak Pupils read

silently FR FR FR …

FR: French language used; WO: Wolof language used.

The mock example instrument provided in Figure 1 shows us how the data collection works: • Snapshots 1 and 2: Two minutes and four minutes after the start of lesson, teacher

speaks in French; pupils speak in group in French. • Snapshot 3: Teacher writes in French; pupils speak in group in French.

The following research questions will be addressed in this report: • What are the languages spoken by the children at home? • How diverse is the ethno-linguistic situation in the classroom? • Do teachers and pupils use national languages at school? • What are the activities of pupils and teachers in the classroom? • What is the literacy environment of pupils? • How do pupils manage to read in grade 3? • Does the use of national languages or specific teaching methods have an impact on

test scores? • What further issues should be addressed with regards to national language in

education?

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 10

What Languages Do Pupils Speak at Home? In the Mali LOI data, the LOI in curriculum schools is spoken at home by the majority of classroom pupils, but choosing one language in multilingual areas is not straightforward (Traoré 2009). Graph 1 below shows the proportion of each language speakers in Senegal classrooms, which can be compared with the Malian context13.

13 Although we should strictly compare only bilingual classrooms of the two countries.

Quantitative versus Qualitative Data

The lesson sequence can be examined from the raw data, but, in this report, only aggregated data was used, giving the count of snapshots by type of activity and language, not the order of the activities. It is possible to calculate the percentage of group versus individual work, the share of each type of activity, and the percentage of use of each language by activity. It is crucial to keep in mind that data is collected every two minutes. Data does not state what teachers and pupils were doing during the last two minutes, but at a given point in time. The data collection method is a snapshot observation, and the data is a small sample of what really happens in the classroom. Comments from enumerators are a rich source of complementary qualitative information. Out of 176 lessons observed in Senegal, only 56 enumerators wrote comments (31% versus 78% in Mali). They provided feedback on the language use between snapshots but unlike Mali, few reported on teachers’ competencies, class climate, and equipment. In Senegal, qualitative data was therefore used to cross check quantitative data on language use vs. qualitative reporting between two snapshots. In Mali, comments were categorized and used as quantitative data. Cultural factors could be leading to such different responses and numbers. This is a common feature when using a common data collection tool in different countries. Following are a few comments showing how diverse language use among Senegal classrooms and within a lesson:

• “4ème et 6ème minutes: Maîtresse parle diola pour expliquer droits de l’enfant mais revient au français.” (Between the 4th and 6th minutes, teacher speaks Diola to explain the rights of the child, but then reverts back to French.)

• “Wolof souvent utilisé entre les clichés.” (Wolof is often used between snapshots). • “3ème minute: Maîtresse parle diola pour mieux expliquer et les élèves répondent en diola. 8ème

minute maîtresse parle en français associant le diola pour donner exemples.” (3rd minute: Teacher speaks Diola in order to explain better and the children respond in Diola. 8th minute: Teacher speaks in French, using Diola to give examples.)

• “Maîtresse parlait dehors avec une dame. Elle a employé une seule fois le wolof pour expliquer une question.” (Teacher speaks outside with a woman. She uses Wolof one time to explain a question.)

• “Maître demande aux élèves en français comment on appelle le feu en mandingue et peul. Elèves respondent.” (Teacher asks children in French how one writes “fire” in Mandingue and Peul. Children respond.)

• Deux élèves parlent en Wolof en se passant une éponge.” (Two children speak to each other in Wolof to pass the sponge.)

• “La maîtresse parle pulaar a un élève qui n’a pas d’ardoise.” (Teacher speaks Pulaar to a child who does not have a chalkboard.)

• “23ème minute un enfant parle pulaar pour donner le sens d’un mot et le maître lui interdit de parler pulaar.” (23rd minute: A child speaks Pulaar to give the meaning of a word, and the teacher forbids him to speak Pulaar.)

Graph 1: Proportion of language speakers in each classroom

* Sum of percentages can be more than 100 as pupils can speak several languages at home.

Large regional variations are observed in the use of Wolof at home, but the sample is not representative at the regional level to produce reliable estimates. See Annex 4 for the maps of Early Grade Reading Assessment (EGRA) sample and national languages.

It is important to make a distinction among the following: • Languages spoken at home (e.g., Wolof, Pulaar, Serere); • Language used at school to communicate with other pupils (typically Wolof); and • Language used in the classroom as official language of instruction (typically French).

Wolof is likely to be used in the school as a communication tool between pupils in Senegal, even if not spoken at home. In fact, the estimated proportion of the population that can speak Wolof is greater than the proportion of pupils whose native language is

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 11

Wolof , as estimated in our data. Language used by children to communicate and play is not anecdotic. In Mali, it is the key to propose an LOI to communities in multilingual areas (Traoré 2009). In Senegal, only four schools have 100% Wolof speakers at home, and 11 schools have no Wolof speakers. Clearly, then, identifying one national language as a medium of instruction in Senegal classrooms is not straightforward. Table 1 below estimates the average proportion of language speakers according to the pupils’ own declaration.

Table 1: Average proportion of language speakers in the EGRA dataset

Language spoken at home Wolof 39%

Pulaar 29%

Others 19%

Serere 11%

French 2%

Source : RTI/Focus Africa 2010, p.10.

Data from PASEC give us further insight, as all official national languages and Arabic were included in the questionnaire. Data are consistent with the grade 3 EGRA dataset, but there is a significant variation between grade 2 and grade 5 pupils (selected in the same schools). The proportion of language speakers varies among schools and also among grades in the same schools.

Matching Home Language and LOI to Improve Schooling, Equity, and Quality

The proportion of pupils not speaking the LOI at home is increasing over time, along with the progression of Universal Primary Education in Africa. In addition, education systems now reach a more diverse population (Varly 2008). If teachers’ methods and pupils’ materials are designed for pupils already speaking “colonial” languages at home, most African pupils remain marginalized in their own education systems.

When looking at PASEC 1995–2000 cohort study, with data on five countries, no significant variation of test scores according to a language group was found, although more recent studies (PASEC 2009, RTI/FOCUS AFRICA 2010, Michaelowa 2008) find that Wolof speakers (at home) tend to perform better. Table 2 shows the proportion of Wolof speakers is higher in grade 5 (58%) than in grade 2 (45.6%), and non-Wolof speakers seem likely to drop out more frequently. The proportion of French speakers is also higher in grade 5 (5.3%) than in grade 2 (2.7%).

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 12

Table 2: Proportion of pupils speaking the different languages home in the PASEC sample

Grade 2 Grade 5

Mean SE Mean SE Wolof 45.6% 3.2% 58.0% 3.9%

Pulaar 25.4% 2.9% 22.1% 2.6%

Serere 16.1% 2.7% 12.7% 2.1%

Mandingue 9.5% 1.8% 10.7% 2.4%

Soninke 6.5% 2.8% 6.6% 1.9%

Diola 5.3% 1.2% 7.7% 1.1%

French 2.7% 0.8% 5.3% 1.0%

Arabic 2.5% 0.7% 0.5% 0.2%

Nb. Pupils 2300 2089

Source: Author calculation on 2007 PASEC dataset.

Ethno-linguistic background is a key factor explaining school attendance and retention, along with poverty, for instance in Burkina Faso (Kobiane 2008). In Senegal, census data show different patterns of literacy according to ethnic origin but no data is produced crossing ethnic origin and formal schooling attendance rates.

Similarly, using national languages to match LOI and ethno-linguistic background is an avenue to get more kids into school—for better equity and possibly for improving learning outcomes. What national languages are really used in Senegal classrooms? How often and what for? Does their use really improve the learning outcomes?

What Do Teachers Say About Their Use of Language in the Classroom?

According to interviews, teachers do not receive training in using national languages as LOI, and only 1 teacher out of 50 declared having teaching materials in a national language. As indicated in Table 3, a total of 67.3% of the teachers reported using Wolof, but 80% of the teachers declared using any given national language less than 10% of a lesson time or less than 3 minutes. These ratios are consistent with the observations made in the classroom via snapshot instruments. Consequently, we can conclude that national languages are seldom used.

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 13

Table 3: Percentage of language used declaring using national language in the classroom

Wolof 67.3 %

Pulaar 34.7 %

Serere 8.2 %

Diola 8.2 %

Mandingue 6.0 %

Diakhanke* 2.0 %

Malinke 2.0 %

Saraxole* 2.0 %

Source: 2009 Senegal EGRA dataset.

*Not official national languages.

However, when they are used, it is as a way to improve the lesson (81.3% of teachers reported) and for discipline matters (29 % of teachers reported).

Direct observation is a better measure of the use of national language than teacher declarations. Data on teacher’s methods, pupil’s activities, and language used were collected through direct snapshot observations in the classroom during a lesson.

What Does Observation Tell Us About Languages Used in the Classroom?

The observations were conducted over the course of several lessons in different subject areas, and variation was found in the proportional use of French across subjects. Teachers use national languages more during math and other subjects such as “education civique.” In reading lessons, French is used quasi-exclusively by teachers. Pupils quasi- exclusively use French in reading and math lessons. The lowest incidence of use of French is 66% of observations for all subjects in all lessons for pupils and teachers.

Table 4: Use of French and national language by lessons subject

Subject # of Lessons

% Teachers use of French

% Pupils’ use of French

Math 23 96.9 100.0

Reading 47 97.9 99.5

Other 42 95.9 97.1

French* 64 99.1 99.4

Total 176 97.8 98.9

Source: 2009 Senegal EGRA dataset.

*Grammar, spelling, vocabulary.

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 14

Snapshot observations are a sample of what is actually happening in the classroom. Enumerators reported in their notes instances of teachers using Wolof or other national languages between snapshot periods, but this data is not reflected in the quantitative analyses presented in this report. If only 7.9% of teachers were observed using a national language during the snapshots, this proportion reaches 13% when looking at comments from enumerators reporting on national language use. Use of national language can be underestimated by the snapshot data collection method.

Table 5 shows us how French and national languages are expected to be used in “bilingual” experimental schools. French is used 50% of the time in grade 3 according to this model, versus 97.8% in the “regular” classes of our dataset, if we assume that the RTI snapshot observations is a proxy for time.

Table 5: Use of French and national languages in bilingual classrooms per grade

National language use (%)

French use (%)

CI 90% 10%1

CP 70% 30%2

CE1 50% 50%

CE2 30% 70%

CM1 20% 80%

CM2 10%3 90% 1 Orally only 2 Writing and as LOI for math 3 For reading only

Source: IDEA (2009).

According to the RTI LOI data, in regular classes, teachers use mainly group work with no significant differences across subjects, except for maths, where teachers work more with an individual pupil, as in Mali. Regarding pupils’ type of work, they work more on an individual basis in reading lessons than in other subjects.

The following analysis focuses only on reading lessons. Graph 2 clearly shows that teachers use Wolof in the classroom only when the proportion of speaker exceeds 50% except in two classrooms.

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 15

Graph 2: Relationship between the proportion of Wolof speakers and use of Wolof in the classroom

0.0

5.1

.15

.2%

use

of w

olof

by

teac

hers

0 .2 .4 .6 .8 1Proportion of wolof speakers in the classroom

Source: 2009 Senegal EGRA dataset.

Wolof is mainly used for group work and in little proportion (less than 20% of observations). The use of French is not related to the proportion of pupils speaking French at home or the diversity of the linguistic situation in the classroom. In 78% of the classrooms, no pupil speaks French at home, and the proportion of pupils speaking French at home does not exceed 15% in any classroom. According to the data, pupils tend to use their home language rather than the most used language (Wolof, which they are likely able to speak) in the classroom.

Finally, regarding the LOI, French is predominantly used. Linguistic variation within a classroom hinders the use of a single national language for teaching in the class. Moreover, national language can be used for discipline matters rather than for really improving teaching, although many teachers (and enumerators) reported using national languages to explain new words to pupils. There is not enough variation in the data to further analyze the LOI as a factor to improve learning. At the school level, there is no relationship between the use of national language and mean EGRA test scores.

What Activities Do Teachers Use to Teach Reading?

Group Versus Individual Work Consistent with Mali, there is a great variation in teachers’ method between classrooms in Senegal, but at least 50% of teachers’ observations are group work in all classrooms (see Graph 3).

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 16

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 17

Graph 3: Distribution of teachers group work

01

23

45

Den

sity

.5 .6 .7 .8 .9 1Distribution of teachers group work

Source: 2009 Senegal EGRA dataset.

What the data tells us about pupils’ work is different than Mali, as pupils tend to work more individually in Senegal (see Graph 4). This difference can be explained by the class sizes, which differ in the Mali and Senegal samples. In Mali, in classes over 60 pupils14, teachers tend to work in groups, whereas in Senegal no class exceeds 60 pupils in the sample, although no significant relationship was found between class size and teachers’ methods in Senegal.

Graph 4: Distribution of pupils group work

0.5

11.

52

Den

sity

0 .2 .4 .6 .8 1Distribution of pupils group work

Source: 2009 Senegal EGRA dataset. 14 Michaelowa identified the threshold of 62 pupils per class has having strong negative impact of pupils scores.

How Is Class Time Spent When Learning to Read? Abadzi estimated that in low income countries, 63% of class time is actually used in teaching activities, and 83% of that time was predominately lecture-based rather than interactive (Gillies 2008). Data on Mali and Senegal are consistent with these ratios if we consider observations classified as “Teacher speaks and reads aloud” as lecture-based.

Tables 6 and 7 below show the distribution by type of activities of the teachers and pupils, respectively, compared to Mali, during reading lessons.

Table 6: Teachers’ activities

Activity Senegal Mali With a group (79.4 %)

Teacher speaks 52.9 % 35.9 %

Teacher writes 18.6 % 13.1 %

Teacher reads aloud 6.3 % 13.4 %

Teacher does not speak 1.2 % 5.9 %

Teacher does not work 0.4 % 0.2 %

With an individual pupil (30.6 %)

Teacher speaks 10.1 % 18.5 %

Teacher writes 2.4 % 4.9 %

Teacher reads aloud 8.1 % 8.0 %

Source: 2009 Senegal EGRA dataset.

In lessons designed to teach reading, the teachers’ main activity is speaking (63% of observation time). Patterns are similar to those found in the Mali dataset, but teachers spend less time reading aloud in Senegal.

Table 7: Pupils’ activities

Activity Senegal Mali In a group (32.4%)

Pupil speaks 9.2% 18.5%

Pupil writes 16.9% 15.1%

Pupil reads aloud 6.4% 10.1%

Individually (67.6%)

Pupil speaks 13.5% 16.9%

Pupil writes 25.3% 20.3%

Pupil reads aloud 18.7% 14.5%

Pupil reads silently 10.1% 4.7%

Source: 2009 Senegal EGRA dataset.

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 18

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 19

As shown in Table 7, the pupils’ main activity is writing (42.2%), most likely copying sentences written on the blackboard onto their individual chalkboards. For the correlations between the different activities and letters per minute read, see Annex 3.

How Much Time Do Pupils Spend Reading? There is far more frequent pupil reading activity in Senegal than in Mali (35.1% versus 25%) (RTI 2011), with more pupils writing and fewer pupils speaking (see Graph 5). There is only one class where pupils do not read themselves in Senegal versus one third of the classrooms in Mali.

Graph 5: Distribution of reading activities by pupils in reading lessons

0.5

11.

52

2.5

Den

sity

0 .2 .4 .6RATIO_READING

Source: 2009 Senegal EGRA dataset.

The proportion of activities devoted to reading varies between lessons. Moreover, half of the grade 3 teachers declared spending 5 hours per week teaching reading and 27% spent 3 hours or less, according to our data. In a PASEC study of curriculum implementation15 in 30 grade 5 classrooms in Senegal, the required average of 3.5 hours spent teaching reading per week was found to have significant variation between classrooms (Lejong 2008). Moreover, if the official program requires 27% of comprehension activities at this level, pupils’ evaluation exercises in reading comprehension accounts only for 7% of what a typical teacher spends time on versus 33% for grammar (Lejong 2008).

Effective teaching time is a key indicator in explaining learning outcome, and therefore it is a useful measurement method to combine the data on teachers’ self-reporting (hours per week) and the share of reading activities observed using the research instrument. In this case, we find that a total of 50% of the pupils spend 1.25 hours or less actually

15 For more information on this study, see Blog on education in developing countries. http://varlyproject.wordpress.com/2010/05/31/curricula-analysis-in-francophone-africa/

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 20

reading themselves, which is insufficient (Gillies 2008). However, at the school level, there is no relationship between the estimated time spent reading per week and average school test scores. As pupils have also other opportunities to read outside school, we will try to capture their broader literacy environment, likely to influence their abilities and test scores.

Pupils’ Literacy Environment and Opportunity to Read (OTR) Index

Our study data give us a sense of a literacy environment among pupils, but with a limited scope considering the matrix designed by (Shiohata 2010). The data do not give information on the literacy environment outside the home and school, but there are still the places were kids are more likely to learn to read. Pupils could have opportunities to read themselves outside school, a key issue in Senegal, considering the small public schooling provision caused by the intensive teachers’ strikes in the recent years or use of double-shift in Dakar Region16. A total of 23.5% of pupils in our study report having no reading materials at home (textbooks or other books home). Table 8 provides the detailed distribution of textbooks and other reading materials.

Table 8: Availability of reading materials (%)

Have other books home?

No Yes Total Have school reading

textbook ? No 23.5 2.6 26.1 Yes 52.7 21.2 73.9 Total 76.2 23.8 100.0

Books available home are mainly in French (87.3%) and Wolof (6.4%) and only 1.9% in Pulaar. The other common language (7%) is Arabic, because the Koran is present in many Senegalese homes--sometimes it is the only book in a household. A total of 89% of pupils declared having someone literate in the home (mainly brothers and sisters), and 92% declared having homework, of which 77% have some help from members in the home.

16 Places to build public school in Dakar in its suburbs is also a problem, identified in a report of the Projet Education dans les Banlieux de Dakar, but not made public.

OTR Index

We can elaborate an OTR Index based on (Gillies 2008) proposal, (Shiohata 2010) as well as the pioneer Inclusive Education in Action (IEA) project work. Since there is little variation in availability of textbooks, or parents’ literacy in the sample, the index is strongly correlated with the estimated number of hours devoted to reading per week in the class (cor. 0.74). The idea is to capture the likelihood of pupils to read at school and at home altogether. Several variables can be introduced in the OTR Index:

• One person can read at home; • Pupil has reading textbook, and pupil has other at book home; • Pupil has homework; • Time spent reading per week at school (estimated at the class level): • Teacher has been absent; • Pupil has been absent; and • Literacy supports on the walls (e.g., posters, decorations, pupils’ writing).

This last component (literacy supports on walls), which was noted during classroom observations, can be considered as an indicator on how reading and writing are valued in the class. In the Mali data, there are many missing values on posters (18%), wall decorations (33%), and pupils’ writings (44%). Only poster data is used; when data was missing, we assumed there were no posters. Other missing data were imputed with the least frequent modality. Pre-schooling was not included in the indicators, because it is not known whether children are prepared or learned to read at this stage in the various forms of pre-schooling in Senegal. Moreover, IEA separates “early home literacy” and “home educational resources” in the PIRLS study.1 The index is a Z-score and was found to be correlated with socioeconomic status and pupils test scores (roughly 0.2).

Graph 6: Relationship between letter recognition and OTR Index in the LOI dataset

020

4060

80

-4 -2 0 2 4Standardized values of (OTR_INDEX)

clpm Fitted values

Source: 2009 Senegal EGRA dataset.

As the distribution of points along the trendline in Graph 6 indicates, the OTR Index is not a good predictor of pupils’ test scores but can be incorporated into regression models in order to compare teachers’ method size effects with pupils’ literacy background (including availability of reading materials and support). Given little use of national languages as LOI in the classrooms, small effects can be expected on test scores.

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 21

How Well Do Grade 3 Senegalese Pupils Read? EGRA Senegal test scores are low: “Among the students who could read at least one word, the results were still very low, with an average of 22.5 words per minute read correctly. Moreover, the comprehension score for this story was very poor among all readers, with only five students able to answer at least 5 out of 6 questions correctly.” (RTI 2010). The proportion of pupils demonstrating minimum French fluency (more than 45 words per minute) is only 11.2% for boys and 7.3% for girls. Graph 7 shows the distribution of scores for letters and words per minute, which show a high variance, but a tendency towards the lower scores.

Graph 7: Letters and words per minute distribution

0.0

1.0

2.0

3D

ensi

ty

0 20 40 60 80Correct letters per minute

0.0

2.0

4.0

6D

ensi

ty

0 50 100 150

Correct words per minute Source: 2009 Senegal EGRA dataset. Source: 2009 Senegal EGRA dataset.

Do Teachers’ Activities and Language Use Have an Effect on EGRA Test Scores?

There is no significant difference in test scores between pupils having a teacher using national languages and others. The global share of teachers or pupils’ group versus individual work is not strongly correlated with EGRA test scores (see Annex 3). When looking at disaggregated data by type of activities, in the EGRA core report (RTI 2010), teachers speaking Wolof with the group, and writing with pupils individually, were associated with significantly higher EGRA test scores. Moreover, some teachers’ self-reported methods were positively correlated with pupils learning outcomes. Indeed, seven teachers who declared that students can understand short stories from the first grade have better performing pupils than others. They are setting ambitious educational goals and are more able to move pupils from the level of graphemes to word recognition (see Annex 2). Similarly, teachers using the "Reciting a story they have already read" activity have pupils with lower results. Repetition techniques, very popular in Senegal, are not effective, as in Mali.

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 22

Using Regression Models to Link Teachers’ Methods and Test Scores

Data potentially allow us to try to identify the best teaching methods to improve learning outcomes, but little effect can be expected with use of the national language because there is little variation in the LOI used in the classrooms. As the use of national languages and type of classroom activity is possibly correlated with other variables (such as proportion of Wolof speakers in the classroom), regression models are suitable to estimate net effects of teachers methods on reading skills. Several models were tested. Initial models were made of the variables identified in the core EGRA report (pp. 39-40), adding teachers’ methods and LOI. Further control variables were introduced. For the comprehension, we considered the words per minute as initial score, and for words per minute, the number of graphemes. The idea is to measure the value added by the teachers’ method, given an initial level of competencies and background characteristics (Varly 2010). Therefore, the research questions we will try to answer through the regression model are as follows:

• What variables have an effect on number of letters read per minute? • What variables have an effect on number of words read per minute at given level of letter per

minutes? • What variables have an effect on the likelihood to understand at least one question at a given level

of words recognition? • Variables introduced to measure teachers’ expectations or commitments and practices are: Pupils

can read a little story right from CI, according to teacher declaration; Frequency of pupils’ writing evaluation; Teacher uses national language; Teacher’s share of group work (or teachers speak Wolof with the group and teachers write individually in French with pupils); Pupils’ share of group work (or Pupils write individually); OTR Index (or Number of reading hours per week or share of reading activities in a lesson);Teacher declared using national language for discipline.

Missing data were imputed with the most common modality, but that tends to reduce the variance of explanatory variables. This method should not impact the estimations when the response rate is over 95%.

To determine whether teachers’ activities and language use have an effect on EGRA test scores, first, we will try to determine how the teachers’ methods and LOI used contributed to the variance compared with the pupils’ OTR in and out of the school and other pupils, teachers, and schools head characteristics. Graph 8 below shows the different grapheme recognition variance contribution of variables. In this model, no initial test score was introduced.

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 23

Graph 8: Share of letters per minute variance by categories of variables

Source: 2009 Senegal EGRA dataset.

Teachers’ method and use of national language only explains 0.84% of grapheme recognition variance versus 2.21% for OTR Index. There is little variance in the use of LOI but a large variance in teachers’ methods, with little impact on test scores.

Pupils’ characteristics and prior schooling (pre schooling and repetition) explains a large amount of variance (13%). Class effect is calculated by including dummy variables for each school and gives a sense of the differences generated when a pupil is enrolled in School A versus School B (Varly 2006). A total of 68% of the variance remains unexplained.

To compare size effects of individual items, we standardize the response variables (CLPM) and consider the regression coefficient as a percent of standard deviation (SD). Except where the school received training in methods of teaching reading, no teacher or school variable is significant in the model. Speaking Wolof at home has a 41% of SD effect on test scores, graderepetition has a 34% effect on SD, and OTR has 19%.

Word recognition capacities are largely explained by the letter recognition abilities, but the relationship between comprehension and number of letters read is weak, as mentioned in the core EGRA report. They pertain to different categories of cognitive abilities. We will now produce models, including initial scores, e.g., letters per minute to explain words per minute and words per minute to explain comprehension. The regression models of CWPM on CLPM do not come up with any significant variables coefficient. The initial level of letter recognition explains itself 64.3% of the words per minute variance and the rest of the variables only 1.9% altogether. Therefore, these models are not presented here.

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 24

Graph 9: Share of comprehension variance by categories of variables

Source: 2009 Senegal EGRA dataset.

Graph 9 shows that words per minute performance explains 10.2% of comprehension variance, pupils’ characteristics 5.7%, OTR Index 1.4%, and teachers’ methods and LOI less than 1%. A total of 69.7% of the variance remains unexplained. These variance contributions are similar to those found in most of the PASEC studies using pre-test and post-test in a school year (Varly 2006). In Senegal, understanding a little story should be mastered at grade 3, according to teachers’ perception while word recognition should be mastered at grade 2 (see Annex 1). Since the study targets grade 3, it is assumed that word recognition is no longer taught at this level, and is reflected in an initial test score.

What Variables Have an Effect on EGRA Test Scores? Two regression models are presented in Annex 2. Several models were tested, and the OTR was disaggregated to check the effect of each component item such as textbooks and pupils absenteeism for instance. Globally, the teachers’ methods and use of national language explain very few variances of EGRA tests scores, when compared to pupils and teachers characteristics.

Pupils’ individual activities have a positive effect on comprehension, with similar results found in Mali. However, it is not significant in all the models tested. Pupils’ writing individually is positively correlated with better test scores (letter recognition and comprehension) and is significant in some of the regression models as well, while in Mali the correlation coefficient is negative. The hours of reading per week in school and the director training in reading methods have a positive effect on test scores for letter recognition and comprehension.

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 25

Wolof-speaking pupils also have significantly better tests scores on letter recognition but not in comprehension, and this result must be taken with caution. Using a logit model on PASEC data, the likelihood of speaking Wolof at home seems to be affected by the socioeconomic status. school equipment, and school location (urban/rural). Speaking Wolof is a proxy indicator for several living conditions and behaviors and is not only a language indicator. There is no causal relationship between speaking Wolof and reading abilities.

Pupils whose teachers think they should be able to understand a short story from first grade have better performance in comprehension. Teachers’ absenteeism has negative effects on comprehension and word reading. Repetition has a negative impact on test scores and possibly on further school attainment, as identified by (Glick 2009) and other similar research based on the Senegal cohort PASEC data.

No effect of using of national language was revealed but it could be slightly underestimated by the data collection “snapshot” method which does not capture every instance of national language use. In addition, the variance is too low to draw conclusions.

A Proposal of Framework for Measurement and Action From existing literature (Abadzi 2006, Verspoor 1991, Pasec 2009, Gillies 2008, Shiohata 2010), and looking at recent EGRA reports and findings (Gove 2010), as well as commonly reported problems affecting use of national languages as LOI (DIA 2009, UNESCO 2010), a framework can be produced to categorize factors influencing specifically reading skills and identify areas of action (see Figure 2).

Figure 2: A framework for measurement and action

Categories Items

Actual measurement

methods

Pupils

Individual characteristics Age, gender, rank in the sibling order Questionnaire

Prior schooling Nursery, repetition, disruption in schooling Questionnaire

Living conditions Water supply, number of

goods possessed, nutrition, sibling size

Questionnaire

Ethno-linguistic background

Ethnic origin, languages spoken home and outside

home Questionnaire

Literacy environment* Parents literacy, books home, classroom posters Questionnaire

Learning conditions* Textbook, class size, homework

Questionnaire and observations

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 26

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 27

Categories Items

Actual measurement

methods

Parents representations on national language

Do parents want their children to learn in the

home language? None

Teachers and

Schools

Teachers’ method Group versus individual work, activities, time on task Observations

Teachers’ and school characteristics

Status, age, gender, initial training, public school Questionnaire

Teachers language proficiency LOI and national languages None

Use of languages in the classroom Time spent using languages Observations

Teachers’ mental representations

Pedagogical objectives, best- perceived methods, LOI Questionnaire

*Also referred to as OTR.

Other Factors Parents’ representations on national languages in education and teacher’s language proficiency issues have not been measured and properly addressed thus far. A recent, large test of all primary and lower secondary teachers in Guinea Bissau (Benavente 2010) revealed problematic mastering of the official LOI (Portuguese) by teachers, even after several years of initial training. A significant proportion of community teachers do not have basic grade 4 competencies in Portuguese language and math.

Mali enumerators reported on several teachers not mastering the LOI, even when it was a national language. In Senegal, it is not clear whether the teachers have enough proficiency to teach in national languages, although they might speak it daily at home and in school. Wolof codification has been consolidated (even in advanced mathematics), but it is not clear whether teachers are able to properly write this language (or others), using acceptable standards17. It is worth noting that in the PASEC data, 20% of teachers declared not being able to adequately speak the pupils’ or local community language.

Another sensitive issue deals with the attitudes from parents and the education community on LOI. For example, “In Senegal, ambivalence about the issue of language in education is widely shared by government officials, teachers and parents” (Shiohata 2010).

Do parents really want their children to learn to read in national languages or in French or Arabic? The parents’ wishes might influence political decisions at a macro level, but it could also influence teachers’ practices at the local level. For Dia (2008), parents are

17 This article discusses the debate around the variance in Wolof language as used by the press, some arguing that Wolof is used improperly in the press and therefore diminishing the integrity of the language. http://fr.allafrica.com/stories/201005031364.html.

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 28

reluctant to support use of national languages in education18: “La réaction négative des parents apparaît presque dans tous les entretiens. Il faut dire qu’au début les parents avaient cédé sous l’insistance des directeurs et maîtres, mais ils ne voulaient pas que leurs enfants étudient en LN. Ce qui s’est traduit par une série de retrait de leurs enfants des classes expérimentales”19.

So far, few awareness campaigns have been made in favor of national languages in education and national languages-dedicated teachers are little valued socially and somewhat confused with “alphabétiseurs,” individuals who are in charge of adult literacy campaigns. The additional skills required to teach in bilingual classrooms might be underestimated by teachers, as illustrated by this teacher’s statement: “J’avoue que je suis très seul dans ma classe à la limite. Avec les collègues ça a été un peu difficile quand même parce qu’on nous considérait toujours comme une classe wolof. La classe wolof, on nous disait. Alors que ce n’était pas une classe wolof. C’était une classe bilingue,” (Dia 2008)20.

This might partially explain why national languages are so under-used in Senegal classrooms. They could be perceived by teachers’ colleagues as indicators of a lack of mastering of French language skills, which are still a key qualification to obtain higher responsibilities within the education system, access to a political positions, and mandatory to run for presidential office, according to the Constitution.

National languages as mediums of instruction should be better promoted among teachers, school heads and parents interviews, otherwise any pending reform for introducing national languages will be another pretext for strikes and a social movement in Senegal.

Introducing national languages in education is far from reaching a consensus in the education community, is still perceived as minimally sustainable in the upper grades, and viewed as an obstacle to access the formal sector market. However, there is substantial evidence worldwide and in Senegal that use of mother tongue as the LOI improves education quality. The 2008 EGRA pilot study and IDEA impact evaluation showed significant improvement of early grade skills when using mother tongue as the LOI.

18 Editor’s note: This finding was also supported by qualitative research and informal interviews conducted through the course of the EGRA implementation. (See Roggemann, K., 2009). 19 Editor’s translation: “The negative reaction of parents appears in almost all interviews. It seems that at first, parents gave in to insistence by directors and teachers, but they did not want their children to study in national languages. This resulted in parents finally removing their children from the experimental bilingual classrooms.” 20 Editor’s translation : I admit that I feel very isolated in my classroom. With colleagues it was a bit difficult because they always considered us a Wolof classroom. “The Wolof Classroom,” they would say. Although it wasn’t a Wolof classroom, it was a bilingual classroom.

Results from the EGRA Pilot Survey: Wolof Versus French

In the piloting of EGRA instruments, tests were administered in French and Wolof to 502 and 186 pupils, respectively, in 32 schools where teaching is provided in French or Wolof. Tests were administered to grade 1, 2, and 3 pupils. Although French and Wolof versions of the pilot tests are not strictly equivalent and the sample is small, Sprenger-Charolles (2008) report states:

Children learning to read in Wolof achieved better results than those learning to read in French for tasks involving spoken language. By contrast, the differences in language of instruction were significant in 4 of the 10 tests: children learning to read in Wolof scored higher for the two tasks that assessed oral language (listening comprehension and phoneme counting), children learning to read in French scored higher for the two tasks that assessed written language (orthographic skills).

The first difference may be explained by the fact that almost all children learning to read in Wolof speak this language at home, while few children learning to read in French speak French at home. The results obtained in tasks that assessed written language may be explained by the fact that written Wolof is not as developed as written French. Consequently, children learning to read in Wolof were probably less exposed to written materials in this language than those learning to read in French.

Wolof-learning pupils scored better in comprehension when responding to an oral stimulus, but differences in other activities score were relatively small. We can therefore recommend that in further studies, children should be tested in reading and math in order to compare the value added of using national languages. An impact evaluation conducted by IDEA (2009) revealed problems in grades 4 and 6 in math in the bilingual classrooms.

The pilot test report gave also guidance in terms of variables to be included in regression models explaining EGRA test scores variation. However, in our study, tests were only administered in French, and we cannot compare pupils that are learning in French versus Wolof.

Conclusion Schooling has improved quantitatively in Senegal in recent years but faces issues of social demands. Koranic and Franco-Arab schooling are alternatives modes of education that compete with the traditional, so-called “Ecole française.” National languages were introduced as experimentations in the classroom but with no significant evidence of an effect on quality thus far, given a questionable implementation (Dia 2009, IDEA 2009). The improved performance of pupils learning in their mother tongue in grade 2 was not

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 29

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 30

observed in grades 4 and 6, where pupils have even poorer performance in math than pupils learning in French.

It is not clear what the government’s commitment to education in national languages, and actions remain in the experimentation stage21. The repeated argument stating that actual results of the experiment are not an incentive to generalize national language in primary school may hinder further political reasons, including the sensitive issue of diversity of languages in Senegal.

Although Wolof is spoken by a majority of the population, it is not the LOI in Senegal classroom—that remains French. This political decision might be a solution to prevent potential ethnic conflicts arising when a few national languages are imposed but does not rely on a sufficient corpus of research on the pedagogical effectiveness of mother tongue use as LOI in Senegal. Mali reached consensus on what language to use as LOI, even in multilingual areas and communities and accounting for the fact that Bamanankan seems less widely spread than Wolof as the main national communication tool.

This study gives further insight on national languages in education, such as follows: • Teachers seldom use national languages as LOI and face a very diverse ethno-

linguistic situation in their class. Teachers use Wolof in their class as a communication tool when the proportion of pupils having this language as mother tongue reaches a certain level (50%).

• Teachers’ methods are largely based on group work, even in small size classrooms. Pupils’ speaking Wolof at home is associated with better performances but also with other contextual variables and no causal relationship can be established.

• No significant effect was found of use of LOI or a specific teaching method on test scores, except the pupils’ individual work, and especially writing.

• The effect size of teachers’ methods is small when compared with pupils’ OTR in school and at home. Insufficient time is given for children to read at school, with 50 % of the pupils spending less than 1 hour and 15 minutes actually reading per week.

• Teachers’ absenteeism and effective schooling time is a crucial issue in Senegal. In the actual configuration, the effective reading time does not match minimum standards, and teachers’ methods have only marginal effects. The introduction of national languages will not resolve this problem in Senegal and in Mali.

Further issues should be addressed such as parents’ actual desire for their children to be taught in national languages and teachers’ proficiency in national languages. A survey of opinion could be undertaken about national languages in education, but the coming 2012 elections are not a context conducive to important reforms. The Senegal education system already remains very instable.

21 http://www.education.gouv.sn/politique/langues-nationales.html.

If national languages were to be introduced in Senegal, teaching materials and textbooks should be improved and the curricula reform completed. Teachers having the minimum proficiency skills to teach in national language and in French should be clearly identified and more socially valued. There is clearly a mismatch between the social value attached to national languages in the media, daily life, literature, and their place in the education system.

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 31

References Abadzi (2006), Efficient learning for the poor: insights from the frontier of cognitive

neuroscience, World Bank, Washington.

African Academy of Languages (2009), National policies: The role of vehicular cross-border languages, the place of less disseminated languages in Africa, ACALAN.

Andre P., Demonsant J. (2009), Koranic schooling: an actual real barrier to formal education?, Job Market Paper.

ANSD (2008), Rapport national de présentation du RGPH, 3e Recensement de la population sénégalaise, Dakar.

Benavente B., Varly P. (2010), Balanço de competências de docentes em exercício na guiné bissau, UNESCO/BREDA.

Dia Y. & al (2009), L’introduction des langues nationales dans le système éducatif formel, ERNWACA/ROCARE.

Diallo I. (2010), The Politics of National Languages in Postcolonial Senegal, Cambria Press, forthcoming October 2010.

Fehrler S., Michelowa K. (2009), Education marginalization in Sub-Saharan Africa, Background Paper for the EFA Monitoring report, UNESCO.

Hewlett Foundation (2008), The Quality Education in Developing Countries Initiative Grantmaking Strategy.

Gillies J., Quijada J.. (2008), Opportunity to Learn: A high impact strategy for improving educational outcomes in developing countries, USAID/EQUIP2.

Glick P., Sahn D. E. (2008), Cognitive skills among children in Senegal: Disentangling the roles of schooling and family background, Economics of Education Review, Volume 28, Issue 2, April 2009, p 178-188.

Glick P., Sahn D. E. (2009), Early Academic Performance, Grade Repetition, and School Attainment in Senegal: A Panel Data Analysis, Rand Corporation.

Gove A. & Cvelich P. (2010), Early Reading: Igniting Education for All. A report by the Early Grade Learning Community of Practice. Research Triangle Park, NC: Research Triangle Institute.

Kobiane J.F, Pilon M. (2008 ?), Appartenance ethnique et scolarisation au Burkina Faso : la dimension culturelle en question.

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 32

Lejong M., Ineade (2008), Le curriculum officiel est-il implanté dans les classes sénégalaises en cinquième année ?, Document de travail, PASEC/CONFEMEN.

Lookheed M., Verspoor A. (1991), Improving Primary Education in Developing Countries, Oxford, Oxford University Press.

Michaelowa K. (2003), Les déterminants de la qualité de l'éducation primaire : Enseignements de la mise en œuvre du PASEC en Afrique subsaharienne francophone, Document de travail, ADEA.

Ministère de l’Education (2009), Rapport national sur la situation de l’éducation 2009, DPRE.

Ministère de l’Education (2003), Plan national de développement de l’éducation et de la formation (PDEF).

PASEC (2009), Evaluation diagnostique Sénégal, CONFEMEN.

Piper B. (2010), Kenya Early Grade Reading Assessment Findings Report, RTI International.

Piper B. (2010) Uganda Early Grade Reading Assessment Findings Report, RTI International.

Roggemann, K. (2009), Welcoming Knowledge: National Language as Language of Instruction in Senegal. Unpublished internship report submitted to RTI International.

RTI/FOCUS AFRICA(2010), Evaluation des compétences fondamentales en lecture au Sénégal, RTI International.

RTI (2011). The Monitoring of Learning Outcomes in Mali Language Of Instruction and Teachers’ Methods in Mali Grade 2 Curriculum Classrooms. RTI International.

Shiohata M. (2010), Exploring the literacy environment: a case study from urban Senegal, Comparative Education review, vol. 54, no.2.

Sprenger-Charolles L. (2008), EGRA-Results from Senegalese Primary School Students Learning to Read in French and in Wolof, Pilot tests, RTI International/WORLD BANK.

UNESCO (2010), Why and how Africa should invest in African languages and multilingual education, An evidence- and practice-based policy advocacy brief, UNESCO Institute for Lifelong Learning.

Varly P. (2006), Gestion scolaire et réussite éducative, Exposé de cadrage des Assises Francophones sur la gestion scolaire, Document de travail, CONFEMEN.

Varly P. (2008), Synthèse des résultats PASEC VII et VIII, Papier présenté à la 53ème session ministérielle CONFEMEN, Document de travail, CONFEMEN.

Varly P. (2010), Rapport de mission pour RTI EGRA-Sénégal & Mali, Appui au séminaire de communication et mobilisation sociale, Document de travail.

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 33

Senegal: Languages of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms 34

Varly P. (2010, Curricula analysis in francophone Africa, posted on a Blog on Education in the Developing countries.

Varly P. (2010), Language of instruction and teachers’ methods in Mali grade 2 curriculum classrooms, Draft paper, RTI International.

Annex 1: Matching Official Versus Implemented Curricula Les Curricula Souhaités, Implantés et Réalisés au Sénégal

Tâches

Niveau visé selon les enseignants (déclarations)

Niveau dans le curriculum souhaité

Résultats actuels des élèves de CE1 Objectifs

1. Lire à haute voix une petite texte avec peu de fautes CP

Oui, Mais Pas d’indication de niveau

26,6 % des garçons au seuil minimum, 20,1 % des filles ?

2. Ecrire son nom CP Pas d’indication precise

3. Comprendre les petites histoires qu'on a lues CE1

Indiqué au CP sans niveau

22,5 % de bonnes réponses ?

4. Identifier les lettres et dire leur nom CP

Pas d’indication precise

28 graphèmes par minute

5. Lire des mots inconnus CP Pas d’indication precise

9,7 mots par minute

6. Comprendre les petites histoires qu'on a entendues CE1 Indiqué au CP

7. Réciter l'alphabet CP Pas d’indication

8. Exécuter une consigne lue CE1 Indiqué au CP

Source : VARLY P. (2010)

Senegal: Languages Of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms A1-1

Annex 2: Regression Models Correct Letters Read Per Minute Regression Model

clpm_STD Coef. Std. Err. t P>t [95% Conf. Interval] gender_n -.0841537 .0654947 -1.28 0.205 -.2159121 .0476046

Age -.0038721 .0009895 -3.91 0.000 -.0058627 -.0018816

PSES -.0115051 .0248185 -0.46 0.645 -.0614335 .0384232

PLANG_WOLOF .4121457 .0845312 4.88 0.000 .2420909 .5822005

PREPEAT -.3407952 .1189994 -2.86 0.006 -.5801911 -.1013994

PPRESCHOOL .1311603 .0987749 1.33 0.191 -.0675491 .3298698

Z_OTR_INDEX .1872508 .0571631 3.28 0.002 .0722535 .3022481

TACADEMIC .0727592 .1121545 0.65 0.520 -.1528666 .2983849

TCTRAIN_FR .1178796 .1167946 1.01 0.318 -.1170809 .35284

DACADEMIC -.1094015 .1143243 -0.96 0.343 -.3393924 .1205893

DCTRAIN_READ .2374509 .1305891 1.82 0.075 -.0252605 .5001623

SELECTRIC -.2119777 .1247672 -1.70 0.096 -.4629768 .0390215

TPEDAG_OBJ~T .1044915 .1490084 0.70 0.487 -.1952748 .4042579

FREEVAL_READ -.0065064 .0423602 -0.15 0.879 -.0917242 .0787114

RATIO_MG .3996917 .4299099 0.93 0.357 -.4651755 1.264559

RATIO_EG -.0530656 .2401859 -0.22 0.826 -.5362573 .430126

TEACHER_US~N -.1548968 .1137346 -1.36 0.180 -.3837013 .0739077

DISCIPLINE -.0428262 .1229621 -0.35 0.729 -.2901939 .2045416

_cons -.2582199 .3834227 -0.67 0.504 -1.029567 .5131271

R² .1846. 650 observations.

Senegal: Languages Of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms A2-1

Senegal: Languages Of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms A2-2

Comprehension Regression Model

UNDERSTAND Coef. Std. Err. T P>t [95% Conf. Interval]cwpm_STD 6.878281 1.233319 5.58 0.000 4.397162 9.359399

gender_n -4.480424 2.510024 -1.79 0.081 -9.529941 .5690925

Age -.0543135 .0363969 -1.49 0.142 -.1275346 .0189075

PSES .5196357 .6969303 0.75 0.460 -.8824072 1.921679

PLANG_WOLOF -3.000766 2.18339 -1.37 0.176 -7.393181 1.391649

PREPEAT .0551806 3.291487 0.02 0.987 -6.566437 6.676798

PPRESCHOOL 9.069343 3.143465 2.89 0.006 2.745507 15.39318

Z_OTR_INDEX 3.919311 1.221808 3.21 0.002 1.461351 6.377271

TACADEMIC -6.215885 2.927754 -2.12 0.039 -12.10577 -.3260042

TCTRAIN_FR -1.028909 3.05158 -0.34 0.737 -7.167896 5.110078

DACADEMIC -1.309771 2.741835 -0.48 0.635 -6.825633 4.206091

DCTRAIN_READ 6.456842 2.808328 2.30 0.026 .8072161 12.10647

SELECTRIC -.2123716 2.866218 -0.07 0.941 -5.978459 5.553716

TPEDAG_OBJ~T 6.089483 3.263784 1.87 0.068 -.4764041 12.65537

FREEVAL_READ -.1631552 1.233644 -0.13 0.895 -2.644927 2.318616

RATIO_MG -4.690194 11.84364 -0.40 0.694 -28.51653 19.13615

RATIO_EG -6.797423 5.206507 -1.31 0.198 -17.27156 3.676719

TEACHER_US~N .8924202 2.918476 0.31 0.761 -4.978796 6.763636

DISCIPLINE -2.368614 3.312424 -0.72 0.478 -9.032353 4.295124

_cons 23.00868 12.25411 1.88 0.067 -1.643407 47.66077

R² 19.8. 532 observations.

Annex 3: Correlations Between Activities and Test Score (letters per minute)

TEACHERS’ ACTIVITIES

Group Individual

Clpm Speaks Writes Reads aloud

Don't speak

Don't work Speaks Writes Reads

aloud

Clpm 1.0000

With Speaks -0.0084 1.0000

a Group Writes 0.1973 -0.3143 1.0000

Reads aloud 0.0377 -0.3747 0.2805 1.0000

Don't speak -0.0137 -0.0675 0.0248 0.2588 1.0000

Don't work -0.0041 -0.1407 0.0245 0.0703 0.8570 1.0000

With and Speaks 0.0446 -0.2868 -0.1267 0.0204 -0.1629 -0.1423 1.0000

Individual Writes 0.0616 -0.1871 0.0930 0.1683 -0.0452 -0.0110 0.3869 1.0000

Reads aloud -0.1761 0.3418 -0.2981 -0.1302 -0.0744 -0.1135 -0.1496 0.0419 1.0000

PUPILS’ ACTIVITIES

Group Individually

Clpm Speaks Writes Reads aloud Speaks Writes Reads

aloud Reads silently

Group

Speaks -0.1886 1.0000

Writes 0.0298 -0.0670 1.0000

Reads aloud 0.0293 0.2317 0.1170 1.0000

Individually

Speaks -0.1062 -0.0557 -0.2370 -0.1335 1.0000

Writes 0.0414 -0.1474 -0.3636 0.0227 -0.0473 1.0000

Reads aloud -0.0092 -0.2953 -0.3131 -0.2699 0.1300 0.1726 1.0000 Reads silently 0.0422 -0.1304 0.0970 0.0182 0.2044 -0.1935 0.0251 1.0000

Senegal: Languages Of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms A3-1

Annex 4: EGRA Sample Map and Language Map Map of the 50 Sampled Schools in Senegal

Source : RTI 2009, EGRA report Senegal, p.9.

Language Mapping in Senegal

Source : Ethonologue.com, accessed October 2, 2010.

Senegal: Languages Of Instruction And Teachers’ Methods In Senegal Grade 3 Classrooms A4-1