DOCUMENT RESUME

ED 343 940 TM 018 112

AUTHOR Kellaghan, Thomas; G,:eaney, Vincent

TITLE Using Examinations To Improve Education: A Study inFourteen African Cinntries. World Bank TechnicalPaper Number 165. Africa Technical DepartmentSeries.

INSTITUTION World Bank, Washington, D. C.REPORT NO ISBN-0-8213-2052-1; ISSN-0253-7494PUB DATE 92

NOTE 93p.

AVAILABLE FROM Distribution Unit, Office of the Publisher,Department F, The World Bank, 1818 H Street, NW,Washington, DC 20433 (free).

PUB TYPE Reports - Evaluative/Feasibility (142)

EARS PRICE MF01 Plus Postage. PC Not Available from EDRS.DESCRIPTORS *Academic Achievement; African Studies; *Developing

Nations; Educational Assessment; Educational Finance;*Educational Improvement; Educational Policy;Elementary Secondary Education; Foreign Countries;Government Role; *National Competency Tests; NationalPrograms; *Public Schools; Testing Problems; TestingPrograms; Test Results; *Test Use

IDENTIFIERS *Africa (Sub Sahara); Francophone Africa; TestingEffects

ABSTRACTA detailed description is presented of the types,

functions, performance levels, governance, administration, andfunding of public examinations in 14 Sub-Saharan African countrieswith different educational traditions, based on English, French, orother backgrounds. The countries are: (1) Kenya; (2) Lesotho; (3)Mauritius; (4) St7aziland; (5) Uganda; (6) Zambia; (7) Chad; (8)

Guinea; (9) Madagascar; (10) Mauritania; (11) Rwanda; (12) Togo; (13)

Cape Verde; and (14) Ethiopia. In virtually all of these countries,public examinations are offered at the end of primary (elementary)school, lower-secondary school, and upper-secondary school.Procedures for funding, constructing, administering, and scoring theexaminations are discussed. Issues discussed include passing rates,the effects of examinations on teaching and grade repetition, theroles of assessment, implications of national policies, and ouhercurrent international issues in education. Guidelines are offered forimproving the quality of examinations and their uses. This synthesisreport draws on a series of studies supported by the World Bank.Findings from the 14 studies confirm that public examinations mayhelp raise academic standards, but can also cause many problems in aneducational system. Nine table:- present data from the studies. A105-item list of references and three appendices with supplementalinformation are included. (SLD)

***********************************************************************

Reproductions supplied by EDRE are the best that can be madefrom the original document.

***********************************************************************

WORLD BANK TLCHNICAL PAITH NUMBLA 165

AFRICA TLCHNICAL Dri'AfILMEN 1 SERIES

c' Using Examinations to Improve Education

cez A Study in Fourteen African Countries

ceD

Thomas Kellaghan and Vincent Grep.neycx4

U.S. DEPARTMENT OF EDUCATION01 lice of Educational Research and Improvement

EDUCATIONAL RESOURCES INFORMATIONCENTER (ERIC/

tr/his document has been reproduced asreceived from the person or organizationoriginating it

0 Minor changes have been made to improvereproduction Quality

Points of view o opinions stated in this document do not necessarily represent officialOERI position or policy

"PERMISSION TO REPRODUCE THIS MATERIAL

IN OTHER THAN PAPER COPY HAS BEEN

GRANTED BY

T. FE-fir/1E

TO THE EDUCATIONAL RESOURCES

INFORMATION CENTER (ERIC)"

'&43

BEST COPY AllAILARE

A

RECENT WORLD BANK TECHNICAL PAPERS

No. 103 Banerjee, Shrubs in Tropical Forest Ecosystems: Examples from India

No. 104 Schware, The World Software Industry and Software Engineering: Opportunities and Constraintsfor Newly Industrialized Economies

No. 105 Pasha and McGarry, Rural Water Supply and Sanitation in Pakistan: Lessons from Experience

No, 106 Pinto and Besant-Jones, Demand and Netback Values for Gas in Electricity

No. 107 Electric Power Research Institute and MENA, The Current State of Atmospheric Fluidized-BedCombustion Technology

No. 108 Falloux, Land Information and Remote Sensing for Renewable Resource Mana,1,-.ment in Sub-SaharanAfrica: A Demand-Driven Approach (also in French, 108F)

No. 109 Carr, Technology for Small-Scale Farmers in Sub-Saharan Africa: Experience with Food Crop Productionin Five Major Ecological Zones

No. 110 Dixon, Talbot, and Le Moigne, Dams and the Environment: Considerations in World Bank Projects

No. 111 Jeffcoate and Pond, Large Water Meters: Guidelines for Selection, Testing, and Maintenance

No. 112 Cook and Grut, Agroforestry in Sub-Saharan Africa: A Farmer's Perspective

No. 113 Vergara and Babelon, The Petrochemical Industry in Developing Asia: A Review of the CurrentSituation and Prospects for Development in the 1990s

No. 114 McGuire and Popkins, Helping Women Improve Nutrition in the Developing World: Beating the ZeroSum Game

No. 115 Le Moigne, Plusquellec, and Barghouti, Dam Safety and the Environment

No. 116 Nelson, Dryland Management: 71w "Desertification" Problem

No. 117 Barghouti, Timmer, and Siegel, Rural Diversification: Lessons from East Asia

No. 118 Pritchard, Lending by the World Bank for Agricultural Research: A Review of the Years 1981through 1987

No. 119 Asia Region Technical Department, Flood Control in Bangladesh: A Plan for Action

No. 120 Plusquellec, The Gezira Irrigation Scheme in Sudan: Objectives, Design, and Performance

No. 121 Listorti, Environmental Health Components for Water Supply, Sanitation, and Urban Projects

No. 122 Dessing, Support for Microenterprises: Lessons for Sub-Saharan Africa

No. 123 Barghouti and Le Moigne, Irrigation in Sub-Saharan Africa: The Development of Publicand Private Systems

No. 124 Zymelman, Science, Education, and Development in Sub-Saharan Africa

No. 125 van de Walle and Foster, Fertility Decline in Africa: Assessment and Prospects

No. 126 Davis, MacKnight, IMO Staff, and Others, Environmental Considerations for Port and HarborDevelopments

No. 127 Doolette and Magrath, editors, Watershed Development in Asia: Strategies and Technologies

No. 128 Gastellu-Etchegorry, editor, Satellite Remote Sensing for Agricultural Projects

No. 129 Berkoff, Irrigation Management on the Indo-Gangetic Plain

No. 130 Agnes Kiss, editor, Living with Wildlife: Wildlife Resource Management with Local Participationin Africa

No, 131 Nair, The Prospects for Agroforestry in the Tropics

No. 132 Murphy, Casley, and Curry, Farmers' Estimations as a Source of Production Data: MethodologicalGuidelines for Cereals in Africa

No. 133 Agriculture and Rural Development Department, ACIAR, AIDAB, and ISNAR, AgriculturalBiotechnology: The Next "Green Revolution"?

(List continues on the inside back cover)

1-1

L.)

AFRICA TECHNICAL DEPARTMENT SERIES

Technical Paper Series

No. 122 Dessing, Support for Microenterprises: Lessons for Sub-Saharan Africa

No. 130 Kiss, editor, Living with Wildlife: Wildlife Resource Management with Local Participation in Afric0

No. 132 Murphy, Casley, and Curry, Farmers' Estimations as a Source of Production Data: MethodologicalGuidelines for Cereals in Africa

No. 135 Walshe, Grindle, Nell, and Bachmann, Dairy Development in Sub-Saharan Africa. A Study of Issuesand Options

No. 141 Riverson, Gaviria, and Thriscutt, Rural Roads in Sub-Saharan Africa: Lessons from World Bank Experience

No. 142 Kiss and Meerman, Integrated Pest Management and African Agriculture

No. 143 Grut, Gray, and Egli, Forest Pricing and Concession Policies: Managing the High Forests of Westand Central Africa

No. 157 Critchley, Reij, and Seznec, Water Harvesting for Plant Production, vol. II: Case Studiesand Conclusions for Sub-Saharan Africa

No. 161 Riverson and Carapetis, Intermediate Means of Transport in Sub-Saharan Africa: Its Potentialfor Improving Rural Travel and Transport

Discussion Paper Series

No. 82 Psacharopoulos, Why Educational Policies Can Fail: An Overview of Selected African Experiences

No. 83 Craig, Comparative African Experiences in Implementing Educ.7tiona1 Policies

No. 84 Kiros, Implementing Educational Policies in Ethiopia

No. 85 Eshiwani, Implementing Educational Policies in Kenya

No. 86 Galabawa, Implementing Educational Policies in Tanzania

No. 87 Thelejani, Implementing Educational Policies in Lesotho

No. 88 Magalula, Implementing Educational Policies in Swaziland

No. 89 Odaet, Implementing Educational Policies in Uganda

No. 90 Achola, Implementing Educational Policies in Zambia

No. 91 Maravanyika, Implementing Educational Policies in Zimbabwe

No. 101 Russell, Jacobsen, and Stanley, International Migration and Development in Sub-Saharan Africa,vol. I: Overview

No. 102 Russell, Jacobsen, and Stanley, International Migration and Development in Sub-Saharan Africa,vol. II: Country Analyses

No. 147 Jaeger, The Effects of Economic Policies on African Agriculture: From Past Harm to Future Hope

WORLD BANK TECHNICAL PAPER NUMBER 165

AFRICA TECHNICAL DEPARTMENT SERIES

Using Examinations to Improve Education

A Study in Fourteen African Countries

Thomas Kellaghan and Vincent Greaney

The World BankWashington, D.C.

Copyright © 1992The International Bank for Reconstructionand Development/THE WORLD BANK

1818 H Street, N.W.Washington, D.C. 20433, U.S.A.

All rights reservedManufactured in the United States of AmericaFirst printing February 1992

Technical Papers are published to communicate the results of the Bank's work to the development

community with the least possible delay. The typescript of this paper therefore has not been prepared in

accordance with the procedures appropriate to formal printed texts, and the World Bank accepts no

responsibility for errors.The findings, interpretations, and condusions expressed in this paper are entirely those of the author(s)

and should not be attributed in any manner to the World Bank, to its affiliated organizations, or to

members of its Board of Executive Directors or the countries they represent. The World Bank does not

guarantee the accuracy of the data induded in this publication and accepts no responsibility whatsoever

for any consequence of their use. Any maps that accompany the text have been prepared solely for the

convenience of readers; the designations and presentation of material in them do not imply the expression

of any opinion whatsoever on the part of the World Bank, its affiliates, or its Board or member countries

concerning the legal status of any country, territory, city, or area or of the authorities thereof or

concerning the delimitation of its boundaries or its national affiliation.The material in this publication is copyrighted. Requests for permission to reproduce portions of it should

be sent to Office of the Publisher at the address shown in the copyright notice above. The World Bank

encourages dissen ination of its work and will normally give permission promptly and, when the

reproduction is for noncommercial purposes, withoutasking a fee. Permission to photocopy portions for

classroom use is not required, though notification of such use having been made will be appreciated.

The complete backlist of publications from tne World Bank is shown in the annual Index ofPublications,

which contains an alphabetical title list (with full ordering information) and indexes ofsubjects, authors,

and countries and regions. The latest edition is available free of charge from the Distribution Unit, Office

of the Publisher, Department F, The World Bank, 1818 H Street, N.W., Washington, D.C. 20433, U.S.A., or

from Publications, The World Bank, 66, avenue d'Iéna, 75116 Paris, France.

ISSN: 0253-7494

Thomas Kellaghan is director of the Educational Research Centre, St. Patrick's College, Dublin.

Vincent Greaney is an education specialist at the World Bank.

Library of Congress Cataloging-in-Publication Data

Kellaghan, Thomas.Using examinations to improve education : a study in fourteen

African countries / Thomas Kellaghan and Vincent Greaney.p. cm. (World Bank technical paper, ISSN 0253-7494 ; no.

165. Africa Technical Department series)Includes bibliographical references (p. ).

ISBN 0-8213-2052-11. ExaminationsAfrica, Sub-Saharan. 2. Educational evaluation-

-Africa, Sub-Saharan. I. Greaney, Vincent. II. Title.III. Series: World Bank technical paper ; no. 165. IV. Series:

World Bank technical paper. Africa Technical Department series.

LB3058.A357K45 1992371.2.'6'0967dc20 92-164

CIP

Foreword.111=11=1110..

Public examinations tend to exert enor-mous influence on the nature of learning andteaching; they tend to dictate not only whatis taught but also how it is taugh. In devel-oping countries, while most examinationsserve a number of functions, including cer-tification and accountability, their main func-tion is to select students for the next highestlevel of the educational system. Their im-pact is most pronounced due to the short-age of places, particularly at the secondaryand tertiary levels of formal schooling.

This study presents for the arst time adetailed description of the types, functions,performance levels, governance, administra-tion, and funding of public examinations ina range of African countries with differenteducational traditions. The national public-examination systems of fourteen Sub-Saharan countries are reviewed. Six of thecountries Fr e Anglophone (Kenya, Lesotho,Mauritius, Swaziland, Uganda, and Zambia)and six Francophone (Chad, Guinea, Mada-gascar, Mauritania, Rwanda, and Togo). Thetwo remaining counties are Cape Verde andEthiopia. Public examinations are offeredin virtually all of the countries at the end ofprimary, lower-secondary, and upper-secondary school.

Procedures for funding examinations; forconstructing, administering, and scoringpapers; and for reporting results in each

country are outlined in the report. Issuesdiscussed include low pass rates, the back-wash effect that examinations have on teach-ing and on grade repetition, the roles ofschool-based and practical assessment, theimplications of national language policies,the use of quotas and compensation proce-dures, the publication of national schoolrankings based on examination performance,and the low level of technical support avail-able to examination authorities. Guidelinesare offered for improving the quality of ex-aminations and for using examinations toimprove education.

Many current international issues hi. thecontroversial topic of examinations andassessment are addressed. Thus, the studyshould be of interest to ministry of educa-tion officials, national examination bodies,development agencies, and national and in-ternational educational organizations. Byoffering concrete proposals for impro vingthe quality of public examinations, and byfocussing on the close interrelationshipbetween formal assessment, teaching andlearning, it helps pinpoint the way to rais-ing the level and the quality of education ofpupils in Sub-Saharan Africa.

Ismail SerageldinDirectorTechnical DepartmentAfrica Region

Preface

The studies synthesized in this reportoriginated from the emphasis on improve-ment in educational quality in the WorldBank's policy paper, Education in Sub-Sa-haran Africa, published in January 1988.This emphasis was highlighted by twelveAfrican Ministers of Education at the firstplenary meeting of the Donors to AfricanEducation (DAE), established in the samemonth to support priority educational de-velopments. When the DAE Task Force metin June 1988, a Working Group on SchoolExaminations (WGSE) was also established,in recognition of the important role exami-nations can play in quality improvement.

The World Bank Education and Train-ing Division of the Africa Technical Depart-ment (AFTED), in assuming responsibilityfor developing appropriate activities to fur-ther quality improvement, then preparedterms of reference to undertake studies onexaminations in primary and secondary edu-cation in Sub-Saharan Africa. The rnoin ob-jectives of the studies may be surn:iwrizedas follows:(1) to improve educational quality in a cost-effective way through adjustment of inputsrelative to examination systems(2) to help develop institutional capacity inSub-Saharan countries accordingly.

Five studies (funded by the Bank's Eco-nomic Development Institute (EDI) and theIrish Government), completed in Fall 1988by the Educational Research Centre (ERC),Dublin, were discussed at a seminar on Us-ing Examinations and Standardized Testingto Improve Educational Quality in Lusaka,Zambia in November 1988. At this semi-nar, AFTED presented its terms of reference

to deepen and extend the studies. Follow-ing consultation, fourteen countries agreedto participate in a set of studies under theterms of reference. Public examination sys-tems in six Anglophone countries (Kenya,Lesotho, Mauritius, Swaziland, Uganda, andZambia), six Francophone countries (Chad,Guinea, Madagascar, Mauritania, Rwanda,and Togo) and two other countries (CapeVerde and Ethiopia) were investigated

The studies were started in early 1989under AFTED management with Bank andIrish Government funding. Ireland agreedto act as lead donor, operating through theIrish Aid Agency, Higher Education for De-velopment Cooperation (HEDCO). An ini-tial model study was carried out in Swazi-land. Studies of examination systems in thesix Anglophone countries and in Ethiopiawere carried out under the direction of theERC, Dublin; examination systems in Fran-cophone countries and in Cape Verde werestudied by Management Planning and Re-search Consultants (MPRC), Bahrain. Animportant feature of these studies was theactive participation of counterparts nomi-nated by Ministries of Education (MOE). Inaddition, lead donor Ireland hosted periodicmeetings of the WGSE to advance the imple-mentation process. Progress reports werepresented at DAE Task Force meetings. Thefindings of this synthesis report were dis-cussed at a November 1990 World Bank-sponsored seminar in Killiney, Dublin, at-tended by examination officials from eachof the fourteen countries and representativesof donor agencies.

This synthesis report is based on: (1) ananalysis of existing examination systems;

vii

(2) diagnosis of qualitative problems; and(3) assessment of existing institutional ca-pacity. The report draws also on researchevidence from work undertaken in Westernas well as African countries to form an as-sessment of examination practice in Africaand to suggest guidelines for the construc-tion and administration of examinations thatimprove the quality of education.

1.2.3.4.5.6.7.8.9.

10.11.12.13.14.

Ethiopia:Kenya:Lesotho:Mauritius:Swaziland:Uganda:Zambia:Chad:Guinea:Madagascar:Mauritania:Rwanda:Togo:Cape Verde:

Preparation of the series of studies onUsi_tg Examinations to Improve Educationwas managed by James McCabe, PrincipalEducation Planner, Education and TrainingDivision, Africa Technical Department of theWorld Bank.

Individual country studies were pre-pared in collaboration with Mirtistry of Edu-cation authorities in each cotTravy, by thefollowing principal authors:

R. Reilly, M. O'Donoghue, V. GreaneyS. McGuinness, M. O'Donoghue, A. Yussufu, M. KithukeJ. McCabe, E. MacAogain, T. KellaghanN Baumgart, M. O'DonoghueT. Kellaghan, J. McCabe, S. SukatiJ . Sheehan, S. McGuinnessT. Kellaghan, M. Martin, J. SheehanA. Brimer, J. Hosman, D. Ouman HamidM Chaibderraine, I. SankhonN Constantine, J. Hosman, Mr. Rakoto, et al.M Chaibderraine, M. El Hafed OuldA. Brimer, J. Hosman, D. Kayinamura, E. MunyantwaliN. Constantine, A. SewaF. Oliveira, 0. Carvalho

ite authors are indebted to Stephenheyneman for his input into the de-sign stage of the studies; to MarlaineLockheed, George Madaus, Eoghan

MacAogain, and Ingvar Werdelin for com-ments on an earlier draft; and to ThomasPoole, Alicia Hetzner, and Michael Matovinafor preparing the report for publication.

viii

Acronyms for Public Examinationsand Examination Authorities

BEPC

CEECEPCEPD

CEPE

CEPEA

CFE/FP1

CFE/FP2

CFEPCES

CPEDECESLCEHSCKCPE

KCSE

KNECMESPCPLESCSECUACEUCEUCLESUNEB

Brevet d'etudes du premier cycle. External examination taken at end of first cycleF2condary (four years) in Togo, in Guinea, and in Chad.i-....ementary certificate, taken at the end of primary school, Mauritania.Certificat de fin d' etudes primaires. Primary-school leaving examination, Rwanda.Certificat de fin d'etudes de l'enseignement du premier degre External certificateexamination at end of primary; also admission examination to secondary school,TogaCertificat d'itudes primaires elementaires. External examination taken at end ofprimary school in Madagascar, Chad.Certificat d'études primaires elémentaires. External examination taken at end of

primary school in Chad, for Arabic-speaking students.Certificat de fin d'études de la formation professionnelle du niveau 1. Externalexamination taken at end of first two-year vocational training period (Level I) inMadagascar.Certificat de fin d'etudes de la formation professionnelle du niveau 2. Externalexamination taken at end of second two-year vocational training period (LevelII) in Madagascar.Certificat de fin d'etudes du premier cycle de l'enseignement secondaire.External examination taken at end of first cycle of secondary school in Madagascar.Certificate of Primary Education, Mauritius.Director of Examinations and Concours, Togo.Ethiopian School Leaving Certificate Examination.Higher School Cerfificate, Mauritius.Kenya Certificate of Primary Education, taken at the end of the eight-yearprimary cycle.Kenya Certificate of Secondary Education taken at the end of the four-yearsecondary cycle (ntroduced in 1989).Kenya National Examinations Council.Mauritius Examination Syndicate.Primarlr Certificate.Primary Leaving Certificate.School Certificate, Mauritius.Service of Examinations and Concours, Chad.Uganda Advanced Certificate of Education.Uganda Certificate of Education.University of Cambridge Local Examinations Syndicate.Uganda National Examinations Board.

ix

10

Contents

t SUMMARY

2. THE NEED FOR REFORM IN AFRICAN EDUCATIONGrowth in EducationConcern with QualityStrategies to Improve EducationExaminations and Reform

Effects of Examinations on CurriculaQuality of ExaminationsRole of Examinations in Reform

1

5567889

11

3. THE EDUCATIONAL SYSTEMS OF FOURTEEN COUNTRIES 15Countries in the Study 15Current Size and Participation Rates of Educational Systems 16Concern with Quality of Provision 16Scarcity of Places 18Repetition 18Dropout 19Teacher Quality 20

4. PUBLIC EXAMINATIONS 23Tradition 23Titles and Descriptions of Examinations 23Size of Examination Enterprise 24Setting of Examinations 24Use of Aptitude Tests 24School-based Components of Examinations 26Scoring 27Technoiv 28Functions ()1 Examinations 28

28Lerii.fication 29Accouradhility 30

Performance 30Examination Repetition 32Governance and Administration 33Localization of Examinations 33Funding 34

Fees 34

Security 35

Feedback 36

Comparisons among Schools/Districts 36

Analysis of Results 37

Timeliness 37

Challenges to Examination Results 37

Outside-School Tuition 38

5. ISSUES IN EXAMiNATIONS AND ASSESSMENT 39

Expanding Enrollments 39

Resources for Examinations 40

Effects of Examinations on Curriculum and Teaching 40

Reducing the Burden of Examinations andIntroducing School-based Assessnrent 42

Use of Multiple-Choice Tests and TheirAssociated Technology 44

Validity 45

Accountability 47

Language of Examination 48

Comparability of Performance 48

Quotas 49

6. USING EXAMINATIONS TO IMPROVE EDUCATION 51

Action Plans 52

Administration 52

Technical and Training Requirements 55

Equipment 56

Construction/Rehabilitation 56

Cross-National Cooperation 56Improving Assessment in Public Examinations 57Anticipating Problems and Undesirable Effects 60Conditions to Improve the Administration of Examinations 62

Examination Authority 62

National Cooperation 63

International Cooperation 63

Number and Format of Public Examinations 64

Formal Training 64

7. CONCLUSION 65

8. REFERENCES 67

9. APPENDICES 73

National Education Systems 73

Titles and Descriptions of Examinations 74

Kenya Certificate of Primazy Education: 80

Sample Essay and Marker's Comments 81

1 Summary

This Synthesis Report summarizes dataand information from a set of studies un-dertaken in:

(1) Kenya, Lesotho, Mauritius, Swazi-land, Uganda and Zambia;(2) Chad, Guinea, Madagascar, Mau-ritania, Rwanda, and Togo, and(3) Cape Verde and Ethiopia.

It outlines the findings, issues and :rec-ommended improvements from the four-teen individual studies undertaken on pri-mary and secondary school examinations.

Considerable reliance has been placedon public examinations in African educa-tion as a means of ensuring that teachersand students cover a common curriculum,and accordingly as a particularly effectiveinstrument for raising academic standards.However, it has also been argued that whilepublic examinations may help raise aca-demic standards, they may also give rise toproblems in the educational system. Thefindings of the fourteen studies confirm thispoint.

Findings

Public examinations have three mainfunctions: certification, selection and ac-countability. While most examinationsserve all three to some extent, selection isundoubtedly the main function. The para-mount importance of selection is reflectedin the large numbers each year who are notpromoted to the next highest level of theeducational system. There is also a grow-ing tendency to use a single examinationfor purposes of both selection and certifica-tion of level of attainment.

Students in virtually all of the systemstake public examinations at the end of pri-mary, first cycle secondary and second cyclesecondary school. Candidates for primary-school leaving certificate examinations inmost countries far exceed those taking sec-ondary school examinations. Countries inFrancophone areas tend to require their stu-dents to take more formal examinations thanthose in Anglophme areas; most often theyalso offer term, end-of-year and concours(competitive selection) examinations.

Setting of examinations seems a ratherhaphazard exercise, with some notable ex-ceptions. In some instances, a final selec-tion is simply made from a pool of ques-tions submitted by teachers. Most of theexaminations reviewed show serious weak-nesses. Very little use is made of school-based assessment.

Most scoring (or correction) of public ex-aminations is carried out by teachers, thougha relatively small number of systems use op-tical scanning and/or computer equipment.Two countries (Ethiopia and Swaziland)have had to abandon optical scanning.

The important issue of marker reliabilitydoes not appear to have been addressed inmost countries, though it should be, giventhe seriousness of the decisions made on thebasis of examination performance.

Pass rates tend to be low. In some in-stances the pass rate is determined by thenumber of places available at the next high-est level of schooling. Data on pass ratesshould be interpreted with caution. For ex-ample, quota systems are sometimes usedto attain national objectives such as equalityof male and female pass rates. Pass marksfrom region to region within countries may

1

10"

vary in an effort to ensure regional equalityof opportunity.

Public examinations undoubtedly exertenormous pressure on activities in schools.Teachers tend to gear teaching to the teststo be taken and to ignore material not fea-tured in such tests, even if it is mandated inthe official curriculum.

Most of the questions in public examina-tions tend to measure students' ability torecall facts. Relatively little attention is givento higher-order cognitive skills such as thoseinvolved in synthesis or evaluation. Exami-nation content tends also to be academic innature; 1; outside of school seldom featuresin examination questions. In most countries,particularly those of Francophone tradition,examinations are of the written essay or oraltype. Other forms of assessment (multiple-choice, project, practical, or aural) are lessfrequently used.

Whereas Ministries of Education admin-ister public examinations in Francophonecountries, independent or semi-independentexamination boards generally fulfil this func-tion in Anglophone countries. In a smallnumber of countries, universities have re-sponsibility for the secondary-school leav-ing examination.

For examinations run by Ministries ofEducation, fees are generally not charged atthe primary level. However, Anglophonecountries are more likely to charge fees.Most systems charge fees for secondaryschool leaving examinations; fees can be veryhigll in some cases.

Concern over security is a conspicuousfeature of most systems. Legal provision forbreaches of security vary.

Examination results are seldom used toprovide useful feedback to schools, admin-istrators or curriculum bodies. Thus, a goodopportunity to effect change is little ex-ploited.

2

Issues

To date, not enough attention appears tohave been focused on the educational con-sequences of existing examination systems.However, the likely impact of changingpresent systems needs to be studied care-fully, given the experience with public ex-amin ,tions in non-African settings.

It must be appreciated that however ex-cellent a test is in terms of validity, reliabil-ity, and efficiency, the pressure on studentsto perform will remain substantially un-changed while lack of places at the next high-est level persists. Where such bottlenecks ex-ist, it is naive to think that teachers will notattempt to gear teaching to testing, or thatstudents will not seek outside help to enablethem to advance to the next level.

In this context also, it must be empha-sized that the high repetition rate attribut-able to "failure" may represent a seriouswaste of scarce educational resources.

The need for definition and implemen-tation of language policies by governmentsis urgent, and in that regard, decisions arerequired on the most appropriate languagesto be used in public examinations. In mostcountries the present emphasis on Frenchand English has probably contributed to lowpass rates.

Clearly, practical subjects and school-based assessments will tend to receive littleemphasis until they are incorporated intopublic examinations. These types of assess-ment, however, are time-consuming, expen-sive, and difficult to moderate.

Efforts to localize (or nationalize) exami-nations, following long experience with sys-tems such as those offered by the Univer-sity of Cambridge Local Examination Syn-dicate, run the risk of undermining publicconfidence in the examination system.

likely impact of such tests on the vast ma-jority of students who may not proceed tothe next highest level must be considered.

Publication of school results and nationalrankings are designed to provide "incentiveinformation," a form of public accountabil-ity expected to motivate schools. Cogni-zance should also be taken, however, of themany factors (e.g., socio-economic) otherthan teaching effort and quality elementsthat affect examination performance.

Efforts to modernize the processing ofexamination data should take into accountthe level of technical support available andparticularly the extent to which foreign cur-rency is likely to be available for procure-ment of required equipment, materials, andexpertise.

Recommended Improvements

It is clear that a public-examination sys-tem can exert a highly positive influence notonly on the nature of assessment, but muchmore importantly on the nature of teachingand learning. While examinations cannotchange the educational system, they can bemade to reflect the objectives of curriculumdevelopers and educational planners. Thefollowing recommendations are offered:

1. Examinations should reflect the full cur-riculum, not merely a limited aspect of it.2. Higher-order cognitive skills should beassessed to ensure they are taught.3. Skills to be tested should not be lim-ited to academic areas but should also berelevant to out-of-school tasks.4. A variety of examination formatsshould be used, including written, oral, au-ral, and practical.5. In evaluating published examinationresults and national rankings, account

should be taken of factors other than teach-ing effort.6. The number of public examinationsshould be reduced to help diminish repeti-tion and dropout rates and the inevitablesense of failure experienced by students.7. The amount of time teachers spend ontesting and preparing for public examina-tions should be lessened to provide moretime for teaching.8. Detailed, timely feedback should beprovided to schools on levels of pupil per-formance and areas of difficulty in publicexaminations.9. Predictive validity studies of public ex-aminations should be conducted.10. The professional competence of exami-nation authorities needs to be developed, es-pecially in test construction.11. Each examination board should have aresearch capacity.12. Examination authorities should workclosely with curriculum organizations andwith educational administrators.13. Regional professional networks shouldbe developed to initiate exchange programsand share common interests and concerns.An example is the Association of Heads ofInstitutions Responsible for Examinationsand Education Assessment in East andSouthern Africa.14. A post-graduate degree course shouldbe established in an African country for ex-amination authority personnel.

Finally, while implementation of theabove can effect improvement, it should berecognized that other factors such as avail-ability of places, language of instruction andassessment, amount and quality of literacymaterials and quality of teaching, togetherwith examinations, will r:ay critical roles inimproving the quality of education.

2 The Need for Reform in African Education

After a period of massive expansion, edu-cation in many African countries has, dur-ing the last decade, encountered a series ofproblems. While demand remains high,progress towards universal provision at theprimary level and expansion at the second-ary level has slowed, mainly because of ashortage of money. More limited financialresources than had been available in the pastare being spread more thinly over increas-ing numbers of students (Fuller 1986; WorldBank 1988). Given this situation, the choicefor policymakers, for at least the remainderof this century, would appear to lie either inincreased efficiency in the use of existingresources or in the acceptance of decliningstandards of access, equity, and scholasticachievement (Windham 1986).

In this chapter, we outline some currentconcerns about education in Africa. Pro-posals to address these concerns relate to avariety of actionsthe provision of booksand equipment, teacher training, enlistingthe support of children's homes, and im-proving the nutritional status of children.Our concern is with the possible role ofpublic examinations, a well-established fea-ture of African educational systems, in im-proving educational achievement.

Growth in Education

Africa has the lowest primary; second-ary, and university enrollment rates of anyworld region (Nafziger 1988). This is sodespite the fact that educational systemsthroughout the continent expanded rapidlyin the 1960s and 1970s. For example, be-

tween 1960 and 1983, the number of pri-mary-school pupils in Sub-Saharan Africaincreased by well over 400 percent, from11.85 to 51.35 million. Gross enrollment ra-tio over the same period rose from 36 to 75percent (World Bank 1988). At secondarylevel, the increase, at 1,400 percent, was evengreater. Student numbers rose from .79 to11.12 million and gross enrollment ratio from3 to 20 percent (World Bank 1988).

Despite massive expansion, the discrep-ancy between provision at primary and sec-ondary levels remained great. The percent-age of students in Sub-Saharan countrieswho transferred from the last grade of pri-mary school to the first grade of secondarygeneral education in 1983 was 43 (WorldBank 1988). This figure has a particularrelevance to examinations because they areused to control the flow of students at thisjuncture, as well as at other points, in theeducational system. The success or failureof a student at any of the important selec-tion points in the system can have very se-rious consequences for his or her educationaland occupational future. It is precisely be-cause of their role as gatekeepers in educa-tional systems, in that the number of placesdiminishes as one ascends the educationalhierarchy, that examinations have acquiredthe importance they possess in African coun-tries.

Educational growth in Africa has slowed,and in some countries has stagnated or ac-tually declined, in recent years. This hasbeen attributed largely to poor economicconditions that were aggravated, in the first

5

1 G

instance, by the oil crisis in the 1970s and,more recently, by increasing levels of debt-repayment obligations (Windham 1986). To-tal public expenditure on education in Af-rica declined from $10 billion in 1980 to $8.9billion in 1983, while per-pupil expenditure,both in absolute terms and as a proportionof gross national product, declined mark-edly at both primary and secondary levels(World Bank 1988. \

There are a number of reasons why theeducational systems of Africa are particu-larly vulnerable to economic circumstancesand the conditions of government finances.First, the role of the public sector in financ-ing education is a major one. In 1983, forexample, 85.4 percent of primary studentsand 85.1 percent of secondary students werein public schools. The proportion for Fran-cophone countries was larger than for An-glophone countries (World Bank 1988).Second, state subsidies are large, as fees forpublic education recover only 5.7 percent ofthe cost of primary education, 11.4 percentof the cost of secondary education, and 1.9percent of the cost of higher education.Third, because of the low level of grossnational product (GNP) in many Africancountries, the cost of public education forone student as a percentage of per capitaGNP is higher in Africa than in any otherregion (Mingat and Psacharopoulos 1985;Nafziger 1988; Unesco 1990, Table 19).

Although resources are limited, and evendiminishing, school-age populations are in-creasing rapidly. During the past twentyyears, over 50 million new pupils enrolledin school throughout the continent, and it isexpected that the number that will becomeeligible in the next twenty years will be 110million (Windham 1986). In addition, thedemand for educated manpower to supporteconomic growth is likely to grow, puttingpressure on governments to expand educa-tional facilities. Moreover, many countriesmade commitments following independenceto provide education for social as well as

6

for economic reasons (see, for example,Swaziland, Ministry of Education 1986; Tan1985; Zambia, Ministry of Education 1977).

Attempts to meet these commitments canbe expected to contribute to the growth innumbers attending school.

Concern with Quality

Side by side with the problems of de-creased financial resources and stagnatingenrollments, commentators in a number ofAfrican countries, as in countries elsewherethroughout the world (see Heyneman andWhite 1986), have expressed concern abouta decline in quality of the education beingoffered in schools (see, for example, Evalu-ation Research of the General EducationSystem in Ethiopia ERGESE, 1986; Kelly1991; Zambia, Ministry of Education 1977).Although many countries now collect edu-cational statistics, for example on participa-tion rates, on a fairly rout:ne basis, evidencerelating to the quality of provision or ofoutput is more difficult to come by. Thereis some evidence of decline in the quality ofprovision. Supplies of key inputs, especiallybooks and other learning materials, are re-ported to be critically low in many coun-tries (World Bank 1988). Concern has alsobeen expressed about decline in qualif'output (as measured, for example,dent achievement). Such concern is I.for the most part on impressionistic evi-dence, evidence, of course, that should notbe ignored. The available empirical evi-dence on a decline in standards, however,is far from satisfactory. While the Interna-tional Association of the Evaluation ofEduactional Achievement (1EA) studies ofachievement in mathematics in Nigeria andSwaziland have been cited as evidence ofdeclining standards (World Bank 1988, p.33), these studies do not tell us anythingabout change as they were carried out ononly one occasion.

Examination performance over time is apossible source of evidence regarding chang-

1 7,I

ing standards (see Kahn 1990). However,the characteristics of students taking an ex-amination as well as the standard requiredto achieve a particular grade in an examina-tion, particularly if a predetermined propor-tion of students is assigned to each grade,may vary from year to year. Thus, knowl-edge of the numbers achieving particulargrades over time does not permit unambigu-ous inferences about changes in standards.It is worth noting, however, that in Swazi-land, pass rates increased on the CambridgeOverseas School Certificate (that was con-trolled by the Cambridge Syndicate) from36 percent in 1980 to 54 percent in 1988.There was an even greater increase in thepercentage of first and second-class passes.It is of interest, in the context of the presentreport, that the improvements have been at-tributed to strategies (that were implementedbetween 1984 and 1987) designed to increasethe competence of inspectors and teachersin the assessmc)nt of students (Lulsegged1988). It is also of interest that the nature ofstudent assessment in Swaziland, even be-fore the efforts to improve it, may have dif-fered from that in other countries. An analy-sis of IEA data, that were collected in Nige-ria and Swaziland in 1980-81, indicates thattime spent by teachers in monitoring andevaluating student performance was posi-tively associated with achievement in Swazi-land but not in Nigeria (Lockheed andKomenan 1989). The Swazi data, howeverone interprets them, do seem to contradictthe general perception of declining standardsof achievement in African schools, under-lining the need for more systematic empiri-cal evidence relating to standards of achieve-ment over time.

Strategies to Improve Education

Because of the perceived need in manycountries not just to maintain standards ofaccess, equity, and academic achievementbut tr, improve them, a number of strategieshave been considered for educational reform(see Fuller 1986; Fuller and Heyneman 1989;

7

Hallak 1990; Heyneman 1985, 1987; Mingatand Psacharopoulos 1985). The World Bank(1988) has shown a particular interest in thisproblem and has identified three major ar-eas to be considered in the formulation ofpolicy:

(1) Adjustment to current demographic andfiscal realities, that will require: (a) diversi-fication of sources of finance (in the form ofincreased cost sharing in public educationand encouragement of nongovernmentalsuppliers of educational services); and (b)unit-cost containment, that will be of greaterimportance than cost sharing in countries inthat the scope for cost sharing is negligibleor non-existent. One aspect of the organiza-tion of education that could contribute tounit-cost containment would be a reductionin the amount of grade repetition.

(2) Revitalization of the existing educationalinfrastructure, that will include: (a) a re-newed commitment to academic standards(principally by strengthening examinationsystems); (b) restoration of an efficient mixof inputs in education (especially increasingthe amount of textbooks and other learningmaterials); and (c) greater investment in theoperation and maintenance of physical plant.

(3) Selective expansion, that will be viableonly after measures of adjustment and re-vitalization have begun to take hold andwill involve: (a) renewed progress towardsthe long-term goal of universal primaryeducation; (b) the development of alterna-tive ways of delivering educational services(including distance education); (c) trainingfor those who have entered the work forcein job-related skills; and (d) research andpostgraduate education.

In a document adopted by the WorldConference on Education for All, that wasconvened jointly by the executive heads ofthe United Nation's Children's Fund(UNICEF), United Nations DevelopmentProgramme (UNDP), Unesco, and the World

Bank; and was held in Jomtien, Thailand,from March 5 to 9, 1990, the emphasis onbasic education and the achievements of stu-dents who receive that education is againevident. Article 3 of the World Declarationon Education for All (1990) states that "Basiceducation should be provided to all children,youth and adults" (p. 4). Recognizing thatthe provision of such education is onlymeaningful if people actually acquire usefulknowledge, reasoning ability, skills, andvalues, Article 4 of the Declaration states thatthe focus of basic education must be "onactual learning acquisition and outcome,rather than exclusively upon enrolment,co.ltinued participation in organized pro-grams and completion of certification re-quirements" (p. 5).

Examinations and Reform

The idea that examinations may have animportant role to play in effecting a reformin education in developing countries arisesfrom the belief that they exercise a stronginfluence on what is taught in schools andcan be used as instruments of accountabil-ity. It can also be argued that they providea relatively simple means of controlling asystem in that resources for other means ofcontrol, such as school inspection andteacher training, are limited and in that stu-dents attend private schools and non-formaleducational establishments as well as pub-lic schools.

Effects of Examinations on Curricula

What evidence do we have that exami-nations affect curricula and teaching inschools? Perhaps the most striking exampleof how an assessment procedure can affectthe content and skills covere ! in a curricu-lum is to be found in an anecdote of a per-son who carried out research on the selec-tion and training of U.S. naval personnelduring the second World War (Fredericksen1984). This person found that the best testsfor predicting grades given by instructors

8

were written verbal and reading comprehen-sion tests. This did not seem to make muchsense given the tasks that gunners were sup-posed to do, such as maintaining, adjusting,and repairing guns. On examining whatwent on in training, the researcher foundthat classes consisted of lecturing and dem-onstrating how to use equipment, ratherthan having students do the actual tasksthemselves. He then designed performancetests in that students were required to dem-onstrate competence by, for example, remov-ing and replacing the extractor plunger on a5 /38-inch anti-aircraft gun. Few of the stu-dents could perform the tasks.

New students soon found out what thenew performance iests were like and beganpracticing the assembly and disassembly ofguns. The instructors also moved from lec-turing to more practical work with guns andgun mounts. At the end of the course, stu-dents performed very much better on theperformance tests than students who hadexperienced the older instructional method.Further, the predictive validity of the verbaland reading tests dropped while the valid-ity of the mechanical aptitude and mechani-cal knowledge tests improved.

The interesting point of this anecdote isthat no attempt was made to change the cur-riculum or teacher behavior. The dramaticchanges in curriculum and achievement cameabout solely through a change in testing.

In the day-to-day workings of schools,the effects of assessment on curricula andon student achievement are not so dramatic.However, there can be little doubt that ex-aminations, to which high stakes are at-tached, exert considerable influence on whatgoes on in schools (Fredericksen 1984;Madaus and Kellaghan, in press). Thoseinvolved in an educational system that hassuch examinations will attest to the fact thatthe topics that teachers and students attendto in class and in study are the topics thatare likely to appear on examination papers.

One commentator has concluded that sinceexaminations represent "the ultimate goalof the educational career, they define whatare the important aspects of a school cur-riculum and they dictate to a large degreethe quality of the school experience for bothteacher and student alike" (Little 1982).

If the examinations are good, this mightbe a satisfactory situation. If the objectivesand skills to be measured are carefully cho-sen and if the tests truly measure them, thenthe goals of instruction will become explicitand well-defined targets for teachers andstudents on that they can focus their efforts.Furthermore, the examinations will providestudents and teachers with standards of ex-pected achievement. Given this situation,there should be no reason why studentsshould not work for marks, and good rea-sons why they should ( Fredericksen 1984).In practice, however, examinations may lack"construct validity" or, for other reasons,may not meet the high standards that wouldjustify teachers and students devoting theirefforts to performing well on them.

Quality of Examinations

Evidence from many countries through-out the world over the past century sug-gests that public examinations suffer from avariety of defects (Madaus and Kellaghan,in press). In African countries, defects ofexaminations have been pointed out on nu-merous occasions, in both official and unof-ficial reports. First, most examinations, atboth primary and secondary level, are lim-ited to pencil-and-paper tests and so ignorea variety of skills that cannot be measuredin this way. Second, examinations empha-size the achievement of scholastic skills (par-ticularly language and mathematics at theend of primary schooling) paying very littleattention to more practical skills. Third, inmost examination questions, the student isrequired to recall or recognize factual knowl-edge, rather than to synthesize material orapply principles to new situations. Fourth,

9

many examinations contain very little refer-ence to the everyday life of students outsidethe school, dealing with scholastic topics andapplications for the most part, rather than,for example, trying to find out if a studentcan use money in the market place. Fifth,the quality of actual items used in tests isoften poor (see Cambridge Educational Con-sultants 1988; ERGESE 1986; Kelly 1991:Lesotho 1982; Little 1982; Myeni 1985;Oxenharn 1983).

If schools gear their teaching to suchexaminations, then they are unlikely to bev. successful in developing in theirstudents the kind of knowledge and skillsthat most people would regard asdesirableskills of observation, problemidentification, problem-solving andreasoning, and particularly knowledge andskills that can be applied in the day-to-daylife of the many students who will have onlyminimal exposure to formal education (seeNational Commission on Testing and PublicPolicy NCTPP, 1990; Nigam 1982). Rather,the effect of such examinations, as wasobserved in Ghana, is likely to be "to supportand encourage rote-memorization, routinedrilling, bookishness" (Brooke andOxenham 1984, p. 158). Indeed, the viewwas expressed in a government report inLesotho (1982) that many problems withcurriculum and instruction seem to stemfrom

. . . .the inordinate emphasis givento the preparation for terminalexaminations which undermines theattainment of certain objectives thatare critical to the country's economicdevelopment. . .

The JC [Junior Certificate]examination heavily emphasizes theaccumulation of factual knowledgeand neglects general reasoning skillsand problem-solving activities" (p.94).

In Ghana, Brooke and Oxenham (1984)have noted that teachers tend to neglect non-

examination subjects even though the offi-cial timetable might require that allocationof time be adhered to strictly. Teachers evenpick topics within subjects according towhether they judge them as likely or not toappear in examinations. The same phenom-enon has been observed in Uganda, whereneglect of the teaching of practical skills, es-pecially at the primary level, and the conse-quences of this for students on leavingschool have been noted (Uganda, Ministryof Education 1989). In such situations, theeffect of examinations can be inhibiting, serv-ing to distort or prevent learning rather thanto promote or facilitate it (Heyneman andRansom 1990; Little 1990).

The use of examinations that were setand marked outside of African countriesprobably contributed to the perpetuation ofthe situation in that examinations took littlecognizance of the conditions in that studentslived and were likely ,) live in the future(Kellaghan 1991). As late as 1981, the school-leaving certificate examinations in sevenAfrican countries were set and marked bythe University of Cambridge Local Exami-nations Syndicate (Bray, Clarke, andStephens 1986), a situation that has nowchanged. Since preparation for examinationswas important in schools, we might expectthat the examinations contributed signifi-cantly towards the maintenance of a west-ern academic education. The examinationsalso, of course, contributed to the mainte-nance of standards and acceptance of certi-fication at the international level, importantconsiderations in developing countries.

The setting up of local examinationboards did not result in a sudden break withthe traditions of the colonial system in thecontent of curricula or examinations or inmodes of examining. For example, the WestAfrican Examinations Council was set up in1951 to serve the Gam lia, the Gold Coast,Sierra Leone, Nigeria, and later Liberia, as alocal independent body. However, in its1980 Advanced-level English Literature syl-

10

labus, only six of 39 authors named wereAfrican, and only three of these were WestAfrican. As well as studying Shakespeare,students had the option to study the poetryof Chaucer, John Donne, and GeorgeHerbert. The Senegalese Baccalaureat of-fered examinations in ten languages, noneof them African (Bray, Clarke, and Stephens1986).

While examinations at the primary levelare controlled locally in all countries, for-eign influences are to be found in curriculaand examinations at this level also. In astudy of the national primary-certificate ex-aminations of ten Anglophone countries,Hawes (1979) found that, in 1978, five usedonly the English language, a further fourused English in all papers except in the oneexamining the local language, and only oneused a local language as the main examin-ing language. Reliance on English in thesecases invariably reflects the fact that thereare several, in some cases a great many, locallanguages, none of that is spoken through-out the country. In linguistically heteroge-neous situations, learning and being exam-ined in a metropolitan language is perceivedto favor no particular group. Furthermore,the use of metropolitan languages, such asEnglish and French, is seen as conferringbenefits, particularly on those who are likelyto proceed to third-level education (Eisemon1990). However, the advisability of teach-ing and examining through English orFrench students who have poor proficiencyin the language and who rarely if ever usethe language outside school is somethingthat has been questioned many times. Stu-dents with limited knowledge of a languagewill inevitably be handicapped in the ac-quisition of knowledge and skills presentedin the language as well as in their ability todemonstrate in examinations the knowledgeand skills they have acquired (Eisemon1990).

A further problem with many external-examination systems in Africa is that their

f)

primary use is to control the flow of stu-dents through the educational system. Thisuse is understandable, and the selectionfunction of examinations will no doubt con-tinue to be important as long as there is ashortage of places at higher levels of educa-tional systems. However, there are a num-ber of daagers inherent in focussing on theuse of examinations for selection. First, theexaminations will tend to be geared to theneeds of pupils who are doing relatively wellin the system while the needs of lowerachieving students may not be adequatelymet. And, second, it is likely that the ex-aminations will focus on academic topics.In the early 1970s, the ILO (1972), in a com-ment that could have been applied to theeducational systems of many African coun-tries, noted that, in Kenya, most primary cur-ricula and examinations ignored the needsof terminating students.

To help avoid such undesirable effectsin constructing examinations, it is importantto keep clearly in mind the need for certify-ing the achievements of all students, as wellas the need for selection. Further, cognizanceshould be taken of the motivational effect ofexaminations on students and the influenceof examinations on what is taught and em-phasized in schools. Examinations that aredesigned for only the top 20, 30, or even 50percent of students and the curricula thatprepare students for such examinations arelikely to be seen as irrelevant to lower-achieving students.

Role of Examinations in Reform

Given the importance of public exami-nations in educational systems, to an extentth; t a school's success may be "judged

ictly by the performance of its students inthe examinations" (Fafunwa 1974, p. 193), itis not surprising that the role of examina-tions has received particular attention in thecontext of the problems facing education inAfrica today. For some observers, externalpublic examinations are perceived as con-

11

tributing to the problems, while others see arole for examinations as part of a possiblesolution. Those who perceive examinationsas part of the problem cite the evidenceconsidered above concerning the inadequacyof examinations. If examinations serve todistort or prevent desirable learning, it mightnot seem unreasonable to conclude thatpublic examinations should be abolished.Together with this view, it is usually pro-posed that a system of school-based assess-ment should be installed in place of exter-nal examinations.

An alternative view is that since there isa lack of the resources that would be neededto introduce alternative assessment proce-dures, for example school-based assessment,to the educational systems of most Africancountries and since the need for selectionwill continue in the foreseeable future, itwould appear to be unrealistic to talk of dis-pensing with present examination systems.This view also recognizes that examinationsystems in Africa are perceived by many tobe relatively fair and impartial and that theyalso serve to legitimate the allocation ofscarce educational benefits. Before an alter-native system could be introduced, it wouldbe necessary to demonstrate that it couldfulfil this task equally well. It is also ar-gued that examinations, if properly de-signed, could have a beneficial effect on thequality of education in schools. Because ofthe high stakes associated with examinationsin terms of student opportunities and teacheraccountability, changes in examinationswould most likely be reflected in changes ineducational practice in schools. If thechanges involve improving the quality andscope of examinations, these in turn shouldresult in improving the educational experi-ences of students in schools (Heyneman1987; Heyneman and White 1986;McNamara 1982). In Little's (1984) words,examination improvements could help turnthe educational system into one "which en-courages, rather than stultifies, desirable out-comes" (p. 228). Again, Little (1982) argues

that "the quality of the examination systemitself can have a considerable impact on thequality of skill formation encouraged by theeducation system, that skills in turn couldhave a considerable impact on the inputs tothe labor market" (p. 177).

The possible role of examinations in edu-cational reform has been considered in anumber of policy documents. In the WorldBank (1988) strategies outlined above, theneed to strengthen examination systems inthe interest of raising academic standards isspecifically mentioned in the context of re-vitalizing the existing educational infrastruc-ture. Further, examination and assessmentsystems could be rationalized and mademore efficient; this seems important at a timewhen numbers taking examinations areshowing lr.rge increases in many countries.Again, assessment procedures could be usedin a formative way to guide instructionaland learning processes in schools to reducedropout rates and grade repetition. In theirimportant role in educational selection, ex-aminations could contribute to greater effi-ciency in the educational system by identi-fying students most likely to benefit fromfurther education. Finally, in future expan-sion of educational systems, examinationand assessment procedures could be as-signed an important role in the provision offeedback information to schools on studentachievement levels. They could also havean important role in ensuring comparabilityof standards between school-based and non-school-based candidates, if non-school-basedways of delivering educational services aredeveloped further.

The fact that systems of examinations andassessment can be used for a variety of pur-poses should not he taken to imply that asingle system of examinations or assessmentcan readily serve all purposes equally well.For example, an examination system that ef-ficiently selects the pupils most likely to ben-efit from further education might contrib-ute to the identification of a technical elite

12

and over time might even have importanteffects on the economic performance of anation (Heyneman 1987). However, such asystem could have serious and damagingeffects on the educational experiences ofmany students if it ignores the fact that formany studentsin fact the majoritylearn-ing has to have utility beyond that of quali-fying individuals for the next level of edu-cation (World Bank 1988). Thus, a procedurethat most efficiently selects students may beinadequate for certification purposes. Simi-larly, a procedure that is adequate for certi-fication is unlikely to be the most appropri-ate one for monitoring the quality of perfor-mance of a school or of the educational sys-tem in general.

In considering examination reform, it isimportant to bear in mind the different func-tions of examinations and the many possibleeffects of examination systems on schools,teaching, and learning. It is also importantto realize that in selecting the specific inno-vations most likely to fulfill a country's re-quirements, insofar as is possible, effectsshould be anticipated and judged as desir-able or undesirable in the context of thegeneral goals and aims of each country'sown educational system (Eckstein and Noah1988; Heyneman 1987). We shall return tothese issues in Chapter 5.

Efforts to reform curricula and examina-tions are already in evidence in a number ofcountries. For example, a program of cur-riculum reform has been carried out in Le-sotho. Syllabi in the core subjects (Sesotho,English, Science, Mathematics, Social Stud-ies) have been designed and disseminated.Syllabi comprise units, specific objectives,suggested activities, skills and concepts tobe learned by students, and resource mate-rials. Side by side with these activities, skillschecklists for Sesotho, English, and Math-ematics (Standards 1-3) and sample testquestion booklets in English and Mathemat-ics (Standards 4-6) have been constructed toreflect the new curricula.

Initiatives in Ke9,57a have been more ob-viously directed towards the use of exami-nations to improve the quality of learningin schools. In these initiatives, that havebeen reported by Somerset (1987, 1988), thecontent of the public examinations at the endof primary schooling was changed and a sys-tem to provide feedback information to thepublic and to schools was introduced. Thereform in the content of examinations in-volved the inclusion of a much broader spec-trum of cognitive skills than had previouslybeen included in examinations, skills de-signed to measure comprehension and ap-plication, as well as skills that could be ap-plied in a wide range of contexts, in and outof school. This was done in recognition ofthe fact that examinations should not be con-fined to the measurement of students' abil-ity to memorize factual information butshould also promote the teaching and learn-ing of competencies that would be usefulnot only to those who stay in school butalso to the majority who would leave afterthe examination. Little (1984) has describedthe changes between 1971 and 1979 in thedistribution of items designed to test knowl-edge, comprehension, and application in theCertificate of Primary Education in Kenyaas "dramatic."

In the feedback system in Kenya, lists ofschools were published that reported theoverall mean of the performance of studentsin each school on the examination. Meanstandard scores for each district in the coun-try were also published. This was describedas "incentive information." "Guidance in-formation," based on an analysis of the per-formance of students nationally on indi-vidual questions, was also provided in anewsletter that was sent to schools. Thenewsletter explained changes in the contentand skills covered in examinations, identi-fied topics and skills causing problems, andsuggested ways of teaching these topics andskills. The incentive information was firstpublished following the 1976 primary-school

13

2 '4

certificate examination although the infor-mation did not receive wide publicity untilafter the 1978 examination. The first full-length newsletter was based on the 1978 ex-amination results.

Only limited information is available re-garding the impact of these procedures onschool practice or student achievement. Atthe time he prepared his report, Somerset(1987) indicated that relevant data on teach-ing methods and content had not been col-lected. He did, however, provide informa-tion on changes in the relative mean scoresof districts throughout the country between1976 and 1981. He attributed changes inthese scores between 1976 and 1979 mainlyto the impact of the incentive feedback sys-tem. If that is correct, then the impact wasnegative since mean performances betweendistricts, that it had been hoped would nar-row, substantially widened during the pe-riod. In the following two years, duringwhich guidance information was available,the trend was reversed and some districtsin which mean scores had been relativelylow improved their positions. It may bethat guidance information contributed to thisimprovement, but any firm conclusionsabout the effects of any of the changes inthe examinations system are not possible onthe basis of the available information.

Since students' test performances on theprimary-certificate examination each yearwere converted to standard scores derivedfrom a distribution with a mean of 50 and astandard deviation of 15, inferences cannotbe drawn about possible trends in studentachievement over time. Yussufu (1989), how-ever, has stated that the performance of can-didates on the Kenya Certificate of PrimaryEducation Examination (KCPE), first intro-duced in 1985, "has, in general terms, showna steady consistent improvement." This con-clusion is presumably based on changes instudents' raw scores. Information about thenature or extent of the changes, however, isnot provided.

Neither is information available on otherpossible effects of the examination and re-porting systems, though we might expecteffects on, for example, retention, repetition,and dropout rates in schools.

Eisemon, Patel, and Abagi (1987) observethat the introduction of new question itemson the primary-school leaving examinationin Kenya has not, in their opinion, changedprimary-school instruction "in ways inwhich greater emphasis on problem solv-ing, reasoning and explanation can be dis-cerned." They stress the need to build ex-aminations in accordance with cognitivetheories of the measurement of achievementif the examinations are to have a beneficialeffect on school practice. They studied theimpact of examinations that had been con-structed in accordance with such a theoryon instruction and learning in a Nairobi pri-mary school. The study is limited in itsscope in a number of respects: it is confinedto one school; it deals only with health edu-cation; and it was carried out under experi-mental conditions. Thus the findings mightnot apply in the normal circumstances u:der which examinations are administered.In some instances examination items assesspupils' abilities to integrate existing and newknowledge. These items are explicitly relatedto the competent performance of target be-haviors. Because of this deliberate structur-ing of items, teachers' explanatory behav-iors and their emphasis on procedures tofoster pupil understanding will improv e.On the basis of this study and of their con-sideration of reform efforts in Kenyan edu-cation, Eisemon and others (1987) concludethat:

Psychometric manipulation of exami-nation items in the absence of moreresearch into instructional and cogni-tive processes may produce bettertests and different teaching, but this

will not necessarily make schools anymore effective especially insofar asfostering outcomes like better healthare concerned.

How precisely examinations might con-tribute to the improvement of education inways that everyone regards as desirable isnot clear. Is it, for example, necessary tobuild tests in accordance with cognitive theo-ries of the measurement of achievement, asEisemon and others suggest? While such anapproach might be helpful, though its limi-tations must also be recognized in light ofour poor understanding of the process ofachievement measurement, it would hardlyseem to be necessary. Insofar as we know,no such theories guided the work of theperson involved in the measurement of theperformance of the young navy personnelin Fredericksen's study (1984), describedabove. Even if construction of a test alongcertain principles is more likely to assist instudent learning than a test that does notfollow those principles, we know very littleabout the mechanism that can translate ex-perience of examinations and knowledge ofresults (that might be in the heads of stu-dents, teachers, parents, or administrators)into higher student academic achievement.So iong as that is so, we should be circum-spect in our recommendations about the useof examinations in the interest of raisingstudent achievement. Actions based on de-cisions taken on an a priori basis, and with-out the beneit of conceptual analysis andempirical evidence, run the risk of being atbest ineffective and at worst damaging. Be-cause of this, we emphasize the need to con-sider possible problems and undesirable ef-fects that can arise when examinations areused to direct the activities of teachers andstudents. We consider these problems inChapter 5 after we set out guidelines relat-ing to the improvement of examinations.

3 The Educational Systems of Fourteen Countries

Cotuttries in Cte Study

In this report, we are concerned with theexamination systems of fourteen Africancountries. While many of these countriesexperienced a variety of colonizers (Arab,Belgian, British, French, German, Italian, andPortuguese), they are sometimes categorizedin terms of the European language that inaddition to local languages, is used in thecountry today. Using this criterion, six ofthe countries can be described as Anglo-phone (Kenya, Lesotho, Mauritius, Swazi-land, Uganda, and Zambia); six as Franco-phone (Chad, Guinea, Madagascar, Mauri-tania, Rwanda, and Togo); while two fallinto neither of these categories (Cape Verde,Ethiopia).

At the outset, it has to be recognized thatgreat variation exists among (and oftenwithin) the countries that were studied forthis report. They range in population fromless than a million (Cape Verde, Swaziland)to over 40 million (Ethiopia). They also varyconsiderably in size, from less than 2,000square kilometers (Mauritius) to 1,222,000square kilometers (Ethiopia). Eleven of thecountries fall into the World Bank (1990)low-income category and three into themiddle-income category. Economic differ-ences are reflected in annual per capita GNPthat ranges from $110 (in Ethiopia) to $790(in Swaziland) and $1,090 (in Mauritius)(Table 2.1). Even these figures may under-estimate the extent of poverty in countries(Swaziland, Ministry of Education 1985).Since they are based on the monetary as-sessment of activities related to the modernsector of the economy (that in many casesrelies heavily on foreign capital), they throw

15

2 G

Table 2.1POPULATION, AREA, AND PER CAPITA GNP

OF STUDY COUNTRIES, 19844=M11IMP

Countries/

linguistic

status

Population

(millions)

Area

(thousands of

square kms)

Per capita

GNP ($)

Anglophone

Kenya 19.5 583 310

Lesotho 1.5 30 530

Mauritius 1.0 2 1,090

Swaziland 0.7 17 790

Uganda 15.0 236 230

Zambia 6.4 753 470

Francophone

Chad 4.9 1,284

Guinea 5.9 246 330

Madagascar 9.9 587 260

Mauritania 1.7 1,031 450

Rwanda 5 26 280

Togo 2.9 57 250

Other

Cape Verde 0.3 4

Ethiopia 42.2 1,222 110

Source: World Bank (1988), except for Cape Verde, for which

Unesco (1988) data for 1935 are used.

little light on the standard of living of themajority of the population who in all thecountries live in rural areas and are engagedin the traditional sector of the economy, thatis mainly agriculture.

Some of the countries are ethnically andculturally fairly homogenous; others arevery diverse. In some countries, for ex-ample, a single language is spoken by mostof the population (Cape Verde, Lesotho,Madagascar, Swaziland). In others, and thisis the more common situation, a diversity oflanguages and dialects is in use (see WorldBank 1988).

Differences among countries, that we willagain advert to when considering the edu-cational and examination systems of thecountries, are outlined here to draw atten-tion to the fact that any general conclusionsreached in this report will have to beadapted and interpreted within the contextof the particular circumstances of each indi-vidual country.

Current Size and Participation Rates ofEducational Systems

Twelve of the fourteen countries in-cluded in our survey have either twelve orthirteen school grades altogether; one haseleven and one has fourteen. All the sys-tems are divided into primary and second-ary sectors. At the primary level, one coun-try has five grades, seven have six grades,four have seven grades, and two have eightgrades. At the secondary level, one countryhas four grades, four have five grades, fivehave six grades, and four have seven grades(Appendix 1).

Kenya has the largest number of pupilsin primary schools followed by Ethiopia,Uganda, Madagascar, and Zambia (Table2.2). In 1986, the gross primary enrollmentratio of the countries varied from 29 percent(Guinea) to 115 percent (Lesotho) (Table 2.2).The gross enrollment ratio for six of the

16

countries exceeded 100. Figures over 100percent can be attributed to the relativelyhigh number of over-age pupils in primaryschool. The percentage of primary-schoolstudents who were female ranged from alow of 28 percent (Chad) to a high of 56percent (Lesotho).

Relatively small numbers of students ad-vance to secondary school. In all but threecountries, the gross enrollment ratio at thesecondary level was less than 25 percent in1986. Ratios at the secondary level exceeded20 percent in only five of our study coun-tries (Lesotho, Madagascar, Mauritius,Swaziland, Togo). Female participation waslower at the secondary than at the primarylevel. In nine of the fourteen countries, fe-males constituted less than 40 percent of to-tal enrollment (Table 2.2). Lesotho is theonly country in which the number of femalesecondary-school students was greater thanthe number of males. The ratic of primaryto secondary pupils was particularly largein a number of countries, most notablyRwanda (43:1), Uganda (12:1), Kenya (11:1),Zambia (9:1), and Lesotho (8:1). Clearly, forthe vast majority of pupils, formal educa-tion can be equated with primary school-ing.

Concern with Quality of Provision

In recent decades, in each of ate fcur-teen countries reviewed, the challenge forMinistries of Education was to provideschools and teachers for a dramatically in-creasing population. To date, the emphasisappears to have been on quaniity rather thanon quality, with the result that concern hasbeen expressed with many aspects of thequality of educational provision. In manycountries, teacher morale, for instance, ap-pears to be low. In Ethiopia, one studyreported that 40 percent of primary and 76percent of secondary teachers would, ifgiven the opportunity, abandon the teach-ing profession (ERGESE 1986). In Lesothoin 1988, 805 teachers and almost 55,000 pu-

27

pils had no shelter. In four of the fourteencountries, class size has increased between1983 and 1986. The pupil-teacher ratio inMauritania has gone from 24:1 (1970-71) to50:1 (1986-87). Class sizes as large as 120have been reported for Togo. In many in-stances, attendance rates are poor. Childrenmay be absent to help their parents either inthe home or with planting or harvesting.

Table 2.2ENROLLMENT AND PARTICIPATION RATES

Primary Secondary

Countries/ Grosslinguistic Number enrollment

siatus (030s) %female ratio

GrossNumber enrollment(0000 %female ratio

Anglophone

Kenya 5124* 49* 96* 540* 40* 20

Lesotho 330* 56 115 39 60* 24

Mauritius 145 49 107 71 47 51

Swaziland 142 50 105 32 50 42

Uganda 2204 45 68 1 87 33 9

Zambia 1366 47 96 1 45 37 18

Francophone

Chad 341 28 43 44 16* 6

Guinea 270 31 29 77* 25* 15*

Madagascar 1492* 48 124 357* 44 42

Mauritania 157* 4-0* 51 37* 30* 16

Rwanda 904 48 65 21 34 3

Togo 511 38 102 78* 24 21

Other

Cape Verde 66* 49 108 7* 30* 14

Ethiopia 2884* 39 38 843* 39* 14

Source: World Bank (1990).

*Updated in national report.

17

In its policy study, Education in Sub-Saharan Africa, the World Bank (1988) con-cluded that "... the safest investment in edu-cational quality in most countries is to makesure that there are enough books and sup-plies" (p. 4). Other studies have shown that,in the case of developing countries, factorssuch as school facilities, textbook availabil-ity, and teacher training account for a largeportion of the variation in student achieve-ment (Eisemon 1988; Fuller 1987; Heynemanand Loxley 1983).

Resources that are lacking include suchbasic items as desks and chairs as in Togo.Many of the national reports commented onthe lack of basic materials, most notably text-books. In Madagascar, up to ten studentsshare a textbook. Over a sixteen-year pe-riod, education in Guinea was conductedpractically without any textbooks. Manytextbooks used in Cape Verde are producedin Portugal, primarily for the home market,and do not provide an adequate coverageof the official local syllabus. In Kenya andMadagascar, as wel! as in other countries,basic science equipment is either in shortsupply or, in some instances, is nonexistentin secondary schools.

Resource availability has implicationsboth for the format and functioning of pub-lic examinations. In particular, practical ex-aminations in subjects such as woodwork,metalwork, and home economics requirebasic materials, not alone at examinationtime but also during the school year. Lackof appropriate resources is one of the mainimpediments to the introduction of moreappropriate forms of assessment (essay,practical, oral, and aural) in Ethiopia. Prob-lems have been experienced in Chad andSwaziland, as well as in other countries, inthe acquisition of equipment for practicalsubjects.

Scarcity of Places

The percentage of students progressingfrom the final grade of primary to the firstgrade of general secondary education is in-fluenced greatly by the number of availablesecondary school places. Students complet-ing primary school compete against eachother for the relatively small number ofplaces at the higher level. At one extreme,80 percent of Ethiopian primary-schoolgraduates transfer to secondary school while,at the other, as few as 4 percent of Rwandanstudents with the same attainment level doso. However, in the former case, it shouldbe noted that the gross primary enrollmentratio is very low (38 percent). For eleven ofthe thirteen countries for which transfer datawere available (Table 2.3), less than 50 per-cent of students advanced from the lastgrade of primary to the first grade of sec-ondary school. In Uganda, for example,approximately one-third of the studentscompleting the primary-leaving examinationcan expect to be accommodated at second-ary level. In Kenya, of the 341,000 who satfor the primary-leaving examination (KCPE)in 1987, 51 percent were successful jn ob-taining places at the next level. In Malawi(examined in an earlier study), in 1986-87,the number of places in Form 1 of the sec-ondary system only amounted to 7.6 per-cent of the total Grade 8 enrollmer t(Kellaghan and Greaney 1989). The bottle-neck between primary and secondary schoolgives rise to great pressure on students toperform well on the primary scnool-leavingexamination, that normally serves as theselection test for secondary school.

Repetition

Traditionally, grade repetition has beenused as an indicator of educational ineffi-ciency. High repetition rates, especially in

18

educational systems that cannot accommo-date all children of school-going age, repre-sent a waste of scarce resources. While somerepetition is to be expected for reasons suchas illness or low achievement, the relativelyhigh levels reported for the countries in thisstudy (Table 2.4) is cause for concern. Rep-etition ranges from 2 percent of total enroll-ment in Zambia to 35 percent in Togo. On

Table 2.3PERCENTAGES OF STUDENTS

PROGRESSING FROMTHE LAST GRADE OF PRIMARY

TO FIRST GRADE OFSECONDARY SCHOOL, 1986

Countries/linguisticstatus Percentage

AnglophoneKenya 28*

Lesotho 26*

Mauritius 43

Swaziland 67

Ugand NAZambia 234-

FrancophoneChad 19

Guinea 27

Madagascar 28

Matnitania 29

Rwanda 12**

Togo 34

OtherCape Verde 45

Ethiopia 80

Source: World Bank (1990).

* 1970.Data for an adjacent year.

** 1989-90,

Table 2.4

REPEATERS AS A PERCENTAGE OF

TOTAL ENROLLMENT, 1986

Countries/linguisticstatus Primary Secondary

AnglophoneKenyaLesothoMauritiusSwazilandUgandaZambia

FrancophoneChadGuineaMadagascarMauritania (1987-88)

13

23

7

15

14

2

24

25

29*

19+

Rwanda (1987-88) 11+

Togo 35

OtherCape Verde (1988) 27

Ethiopia 10

NANA14

8

NA2**

11*

38

20**

First cycle 11+Second cycle 16+

6+32

15+

11

**

1970.

Adjacent year.

National Examinations Report.

the whole, repetition at secondary level,though still substantial, is lower than at pri-mary level. Data for Sub-Saharan Africancountries in general show that the rate ofrepetition for Francophone countries is ap-proximately four times that for Anglophonecountries. Among the countries in ourstudy, the highest repetition rates were re-corded for four Francophone countries:Togo, Guinea, Chad, and Madagascar.

Dropout

Dropping out of school represents awaste of human and financial resourcesunless students acquire basic skills duringtheir time at school. There is a considerablebody of evidence to indicate that large per-centages of students fail to reach the finalgrade of primary school in Africa. Some ofthis evidence has taken the form of report-ing the proportion of a cohort that reachesthe final grade.

Available data indicate that in almost halfof the countries in our study, fewer than 50percent of those entering primary school per-sist to the final grade of the primary school(Table 2.5). In Lesotho, it has been estimatedthat about one-quarter of those entering thelowest level complete primary school (Le-sotho, Ministry of Education 1987, p. 13).

Recent Ethiopian data highlight the ex-tent of early dropping out of school. In 1987,the enrollment figure in Grade 2 (N: 440,515)was less than half that in Grade 1 (N:942,541). In Grade 3 (N: 350,197), there wasa further drop of almost 100,000 pupils.While grade repetition would account forsome of these differences, early dropout ap-pears to be the main reason for the sharpdecline in enrollment figures.

In Lesotho and Swaziland, the highestdropout rates at the primary level were re-corded at the early and final grade levels.

Evidence from a number of countries sug-gests that the dropout rate is higher at thesecondary than at the primary level. In Togo,for instance, in 1986, the dropout rate at pri-mary level was 10.2 percent; it was almosttwice as high (19.4 percent) over the first four-year period of secondary schooling.

19

3op

Table 2.5

DROPOUT: PROPORTION OFCOHORT (N: 1000)

REACHING THE FINALPRIMARY-SCHOOL GRADE, 1986

Countries/linguistic

status

DropoutIndex

1111.1!

AnglophoneKenyaLesothoMauritiusSwazilandUgandaZambia

FrancophoneCh4dGuineaMadagascarMauritaniaRwandaTogo

OtherCape Verde

Ethiopia

NA517+

959+

615

NA801+

172

437+

NA923+

486+

592+

34G+

496+

Source: World Bank 1990.

Data for adjacent year.

In a Swazi study of reasons for droppingout at the primary level, students whodropped out and their parents cited finan-cial difficulty as the most important cause.Teachers, on the other hand, attributed drop-out to poor school performance and lack ofinterest in studies, though they accepted thatfinancial difficulty played a role. With oneexception, similar reasons were advanced fordropout at the secondary level; the excep-tion was the claim of dropouts that the main

20

reason for not continuing in school was thenon-availability of a higher grade in the areain that they lived (Swaziland, Ministry ofEducation 1986). We cannot, of course, as-sume that reasons for dropout would be thesame in other countries.

Teacher Quality

Most of the educational systems in ourstudy have large numbers of untrainedteachers. For example, the majority of pri-mary teachers in Cape Verde have no for-mal qualification. Frequently, those who failto gain admission to teacher training on thebasis of performance on a series of achieve-ment tests are offered teaching positions inareas where there is a shortage. In Lesotho,19 percent of primary and 16 percent of sec-ondary teachers have no formal teachingqualification. In Ethiopia, approximately onein four primary and secondary teachers havehad no professional training; in fact, a totalof 11 percent had less than twelve years offormal school experience. At the secondarylevel, one in five 7th and 8th Grade teachershad a college qualification, while fewer thanone in two of those teaching the final twoyears of secondary school had a universitydegree (ERGESE 1986). Madagascar authori-ties, faced with the difficulty of recruitingteachers for rural areas, have filled vacan-cies with men and women doing their na-tional service. While the percentage of quali-fied teachers has improved dramatically inrecent years in a number of countries (mostnotably Swaziland), the Kenya teaching forcehad a lower percentage of trained primaryteachers in 1988 than in 1963. Like most othercountries, Kenya has experienced a shortageof teachers of specialized subjects.

While entry qualifications to teacher-training colleges vary considerably, theygenerally tend to be low. Primary pre-ser-vice teachers in Mauritius embark on a two-

year teacher-education program after pass-ing the senior cycle examinations. In Mau-ritania, candidates for the one-year teacher-training program should have passed theBaccalauréat; those without this academicqualification may still opt for teacher train-ing by undergoing a three-year course. InMadagascar, on the other hand, entrants toteacher training for primary school are re-quired to have passed only the junior-cycleexamination (CFEPCES) of the secondaryschool. They are then required to undergoa two-year training period and a further yearof practical training.

Low salary levels are perceived to con-tribute to the low status of teachers thoughteachers' salaries in Africa are two to threetimes higher relative to income per capitathan in Asia or Latin America (Mingat andPsacharopoulos 1985). The median salaryaround 1986 for teachers in countries in thestudy for that we had data was close to$1,400 for primary teachers and $2,200 forsecondary teachers. But there wereexceptions to this. In Madagascar, teachersalaries generally range from $45 to $63 permonth. The value of this salary is broughtinto sharp focus when one considers thatrice costs an average family about $28 permonth, while rent for a simple apartment

21

amounts to around $26 per month. It isevident that in this instance a teacher's sal-ary cannot support a family. In Chad,unsalaried teaching assistants are often sup-ported by the villages. In Uganda, the verylow level of state salaries necessitates theraising of most of a teacher's income throughdirect parents' contributions and local fund-raising efforts.

It may be important to distinguishteacher quality from teaching quality. Thefindings of research relating to indices ofteacher quality, such as teacher educationand teacher experience, are not consistent.In some studies they have been found to berelated to student achievement, in others not.In general, however, it seems that formalteacher education is less effective in Africathan in other developing-country regions.Further, the effects of teaching experiencehave generally been found to be less posi-tive than the effects of formal teacher edu-cation. However, in analyses of IEA datacollected in Swaziland, while teacher qual-ity (education and experience) did not cor-relate with student achievement (when stu-dent background and school characteristicswere controlled), the effects of teaching qual-ity (expressed in terms of teaching processvariables) were manifest (Lockheed andKomenan 1989).

3..

4 Public Examinations

In this chapter, we provide a descriptionof the examination systems of the fourteencountries included in our study, based onthe reports prepared for the individual coun-tries.

Tradition

Examination syctems in African coun-tries, not surprisingly, have been stronglyinfluenced by traditions in the Europeancountries that colonized them. Followingtraditions that are both British and French,examinations are formal, terminal, and sub-ject-based. Countries in Francophone Af-rica use the French Baccalauréat system atthe end of secondary school, while Anglo-phone countries have had close ties withBritish examining boards. In fact, the finalschool-leaving examination (GCE) in two ofthese countries continues to be set andmarked in Britain. Anglophone countrieshave also tended to use features of Britishexaminations such as moderators. In recentyears, national examination systems have be-come more independent and have tendedto develop their own examinations. Theproblem of processing very substantial num-bers of examinations has led to an increas-ing reliance on multiple-choice tests, favoredby educational systems in the United States.

Titles and Descriptions of Examinations

A total of fifty-two public examinationsoffered at primary and secondary (general)levels of education was identified. In addi-tion to these, many examination bodies of-

111111111111=1EMMIIIIIIMIN

23

fered a range of other public examinationsin the areas of vocational education andteacher training. The numbers of candidatestaking these examinations are small com-pared to the numbers taking public exami-nations at the primary and secondary lev-els. Vocational and teacher-training exami-nations were considered to be beyond thescope of the present study. In general, pub-lic examinations are offered at the end ofprimary schooling, midway through thesecondary school, and at the end of second-ary schooling. Students in Francophonecountries tend to be subjected to more ex-aminations than those in Anglophone coun-tries. In particular, additional examinationsare administered during the primary schoolyears in Francophone countries. In a num-ber of Francophone countries, also, a com-petitive examination, termed the concours, isused to select pupils for the next highesteducational level.

Written essay examinations are offeredin most countries. In two countries, how-ever (Ethiopia and Zambia), the multiple-choice format predominates. The broad ar-ray of examination formats offered also in-cludes practicals, orals, and aurals. Ang]o-phone countries in general tend to use a va-riety of examination types. In each of thesecountries practical tests are used in publicexaminations. Virtually all use essay andmultiple-choice items in at least one publicexamination. Four Anglophone -xtuntries in-clude aural components in their public ex-aminations. Other formats (oral, projects,short-answer) are used less frequently. Fran-

0 UN

1.)

cophone countries tend to opt for the writ-ten and oral formats.

For each of the fourteen countries in ourstudy, details of examination titles, types,functions, and the stage of education atwhich they are administered are given inAppendix 2. The name of the agency thathas responsibility for each public examina-tion is also given.

Size of Examination Enterprise

The primary school-leaving certificate ex-amination is clearly the most significant innumerical terms (Table 3.1). Kenya(N:354,802), Ethiopia (N:281,734), Zambia(N:197,000), and Madagascar (N:192,405)have the largest numbers of candidates forthis examination. In all but two (Ethiopiaand Swaziland) of the fourteen countries thenumber of secondary junior-cycle candidatesin the system is less than half the number ofprimary candidates. Over 100,000 take thesecondary school-leaving examination inEthiopia (N:175,277) and Keny (N:131,000).In the case of all but these two countries,the senior-cycle school-leaving certificatecandidates total fewer than one-ouarter ofthe number of primary candidates. In ap-proximately half of the countries for whichdata were available, fewer than 5,000 candi-dates took the senior-cycle terminal exami-nation.

Setting of Examinations

Three distinctively different approachesare used for writing items and setting ex-amination questions in the countries re-viewed.

First, in a number of countries with de-veloped national examination centres, suchas Kenya, Swaziland and Mauritius, muchof the work is conducted by staff membersworking closely with others, usually teach-ers. In these countries, multiple-choice isthe preferred examination format.

24

Second, in Francophone countries, ingeneral, a less formal approach is used thatallows for much more direct input fromteachers. In Guinea, for instance, secondary-school teachers are asked to submit topicsfor examination. A commission adds its owntopics from a data bank. The Minister thenmakes the final choice of questions. Teach-ers who set the examinations may haveunrealistic goals and expectations. InRwanda and Mauritania, the Minister isclosely involved in the examination-settingprocess; concern over test security is the pri-mary reason for the adoption of this proce-dure. The ability of teachers to judge theappropriateness of items or to estimate thetime required to answer individual papershas attracted critical comment. In Rwanda,for example, an examination for which athree-hour time period was set aside wascompleted in a half-hour by the candidates.

Third, in some countries (Lesotho andZambia) formal courses for examiners areorganized by examination councils. In 1987,in Lesotho, Primary Certificate multiple-choice items were written by teachers andsubject specialists in specially organizedworkshops. Final examinOion items wereselected on the basis of their degree of dis-crimination and difficulty, and on the judg-ment of subject specialists.

In a small number of countries, the writ-ing of items for the terminal secondary-school examination is entrusted to univer-sity personnel. In Chad and Ethiopia, itemsare written by staff members from the Uni-versity of Chad and the University of AddisAbaba respectively.

Use of Aptitude Tests

Scholastic aptitude tests have not provedpopular in the context of public examina-tions in Africa. This is perhaps because theyhave not been very popular in Europe ei-ther, as compared, for example, with theUnited States. There are a number of other

3 I,

Table 3.1SIZE OF EXAMINATION ENTERPRISE

Countries/linguisticstatus Primary Secondary

AnglophoneKCPE KCE KSE

Kenya 0988) 354,802 (Discontinued, 131,000,1987)

PSLCE .JCE COSC

Lesotho (1987) 23,998 6,642 2,790

CPE SC HSCMauritius (1989) :53,083 11,141 3,977

PC JCE COSCSwaziland (1987) 13,943 7,500 (1988) 2,561

PLE UCE UACE

Uganda (1989) 134,669 42,507 13,248

G7 G9 SC

Zambia (1987) 197,000 59,933 15,343

Francophone

CEPEICEPEA Concours BEPC

Chad (1988) 17,840 20,550 7,985BacNA

7th yearEntrance BEPC Bac

Guinea (1988) 24,668 12,148 (i) 9290(ii) 4209

CEPE CFEPCES Bac

Madagascar (1988) 192,405 84,635 47,624

1st yearCEE Secondary BEPC Bac

Mauritania (1988) 23,680 23,680 3,463 7,361

Rwanda CEP Concours Examens de reclassement57,089 2-2,500

NA

CEPDTogo (1987) 46,364

Other

Cape Verde

Ethiopia (1988)

End of 3rdyear 3,500

BEPC18,190

(ii) 5610

NA

Bac(i) 5163

End of Basic First Cycle Second CycleElementary Complementary General Complementary

NA NA NA NA

G6281,734

G8237,011

ESLCE175,277

25

possible reasons for their lack of popularity.First, curricula in schools are more closelyprescribed in Africa (as they are in Europe)than in the United States. Second, in a num-ber of studies, scores on scholastic aptitudetests have been found to predict performancein third-level institutions less well than sub-ject-based examinations (see O'Rourke,Martin, and Hurley 1989). Third, the use ofan aptitude test might lead schools to em-phasize preparation for this kind of test atthe expense of covering material designedto achieve curricular objectives.

It can be argued that a scholastic apti-tude test may still have certain advantagesin certain situations. First, being neithercriterion-referenced nor based on a singleconcept of subject excellence, it could allowlocal school authorities to experiment withcurriculum. Second, it might have a par-ticular value in countries in which schoolinputs and facilities vary widely, insofar asit is less sensitive than achievement tests tothe differential opportunities of students tolearn (Heyneman 1987).

In Africa, two countries used aptitudetests for a number of years but discontinuedthe practice. An aptitude test was adminis-tered as part of the Ethiopian School-Leav-ing Certificate Examination between 1967and 1974. In Lesotho, verbal reasoning andnumerical ability tests were introduced inthe 1970s to be used in conjunction with theachievement tests of the primary-school cer-tificate examination to select students for sec-ondary school. Sebatane (1985) reports thatthe "predictive validity" of the tests was"professionally established" but does not re-port any details. The results of the abilityand achievement tests were combined, apractice described by Sebatane as "unten-able." The practice of administering the abil-ity tests has been discontinued.

The most consistent use of aptitude testshas been in Zambia. Verbal and nonverbal

26

reasoning tests are administered in conjunc-tion with four subject achievement tests inthe primary school-certificate examination.In a small-scale study in the 1960s, it wasfound that ability tests (measuring verbalreasoning, nonverbal L.easoning, and numeri-cal skills), that were administered in addi-tion to, but not as part of, the secondary-school selection examination of achievement,were better predictors than the selection ex-amination of performance on the JuniorCertificate national examination taken twoyears later (Irvine 1968). In further studiesof the predictive validity of the examina-tions, reported in a note prepared by the Psy-chological Service of the Ministry of Educa-V on, performance on the verbal-reasoningtest was the best predictor of success at sec-ondary school. Performances in the Englishcertification examination and on the nonver-bal reasoning test were the next best predic-tors. In particular, and not unexpectedly,the spatial component of the nonverbal testwas found to have "a very high correlation"with performance in mathematics and physi-cal sciences. Furthermore, performance onthe nonverbal reasoning test was found tocorrelate "very well with success at techni-cal courses at tertiary level." Unfortunately,we do not have more details of these stud-iesfor example, the magnitude of the cor-relations between performance on the apti-tude tests and later school performancebutfurther information about this researchwould obviously be of considerable interestin the context of the selective functions ofpublic examinations.

School-based Components ofExaminations

A few systems currently include assess-ments conducted by students' teachers aspart of the final examination grade. Thejunior-secondary cycle examination in Le-sotho has elements of school-based assess-ment, while in Ethiopia, a total of 30 per-cent of the total marks for the primary

3 G

school-leaving examination (G6) is awardedfor school-based assessments.

Despite the potential advantages of thisform of assessment, especially from the va-lidity perspective, a number of problemshave been observed in its use. A lack ofvariance in school-based marks has beennoted; marks tend to be inflated. Someschools fail to submit students' school-basedassessment marks in time. In such instances,candidates in Ethiopia are awarded the samepercentage marks for school-based work asthey achieved in the public examination.Since school-based marks are usually higherthan examination marks, these candidatestend to be penalized due to the failure toreturn school-based marks.

Marks derived from school-based assess-ments, at one time used to contribute to thetotal marks for the Baccalauréat in Guinea,are no longer used because serious discrep-ancies were found between the school-basedand non-school-based elements of the ex-amination. Discontinuation of the practicein Guinea has contributed to an observedincrease in discipline problems, less regularattendance, and a nonchalant attitude to-wards school examinations.

Scoring

In most systems, scoring or marking ofstate examinations is carried out by practic-ing teachers. In fact, this exercise is consid-ered a most useful form of in-service teachereducation for its participants since it high-lights the essential elements of the curricu-lum, standards expected, and marking pro-cedures. In Cape Verde, scoring is entrustedto the candidates' teachers. Responsibilityfor scoring some public examinations is as-signed to regional or local committees ofteachers in Togo and Madagascar. In Swazi-land, moderators are used to scrutinize asample (10 percent) of scripts marked byteachers and to recommend, if necessary,changes in the scoring.

27

In other countries, notably Kenya, Le-sotho, Mauritius, and Zambia, multiple-choice tests are machine-scored. Ethiopia'sESLCE has been machine-scored by the Uni-versity of Addis Ababa since 1977.Ethiopia's other national examinations arescored by large groups of university studentsor members of youth organizations.

Systems that require students to completeprojects and/or administer practical exami-nations tend to have assessments conductedby outside teachers. In this case, and alsowhere practicals are conducted by the stu-dents' own teachers, the process is oftenmonitored by the school inspectorate. Anumber of examination systems (Lesotho,Mauritius, Swaziland) have their senior-cyclesecondary school-leaving examinations setand scored by an agency outside the coun-try (the University of Cambridge Local Ex-amination Syndicate and London UniversitySchools' Examination Council).

No formal reliability studies of public ex-aminations appear to have been conducted.To ensure the reliability of the scoring pro-cedure of essay-type examinations, severaldifferent procedures have been used. InTogo, papers are corrected once except in"cases where the score distTibution shows alarge deviation from the norm alreadyknown by the committee." (Unfortunatelythe country report does not indicate thesource of this norm or elaborate on the term"score distribution"). In such instances, pa-pers are recorrected. In Mauritania andMadagascar, each paper is corrected twice;in Madagascar, the average mark is assignedexcept where the difference is "very large,"in which case the paper is scored a thirdtime. The scoring process adopted in Guineaclearly indicates that factors other than levelof performance on the set questions in theexamina don affects a candidate's final mark.Scorers are made aware of candidates' schoolgrades prior to correction. According to thenational report "these records allow the jury

to more fairly mark those pupils who havenot received a complete education coveringall of the questions posed at the examina-tions." It seems reasonable to assume alsothat a student who performed well on theexamination could have his or her mark re-duced in line with expectations derived fromschool grades.

Technology

The use of computers and optical-scan-ning machines allows for the processing oflarge amounts of examination data in a shortperiod. Furthermore, if multiple-choiceitems are used, computerized item analysisfacilitates the provision of feedback infor-mation to schools.

To date, the use of computers and opti-cal scanning devices appears to have beenconfined in the main to Ethiopia, Kenya, Le-sotho, Swaziland, and Zambia. Dependenceon modern technology is most pronouncedin Kenya. Over 360,000 primary-school leav-ing-certificate examinations are scanned andanalyzed in a year. Each year, through itsKCPE Newsletter, schools are informed of thepercentage of candidates in the country whoanswer each item correctly and in manyinstances the percentage opting for eachdistractor. Computers are also used to cal-culate national order of merit lists.

Some examination systems rely on out-side agencies for computer services. Pro-cessing of the Ethiopian ESLCE is done bythe University of Addis Ababa; in Zambia,machine-scoring )f multiple-choice tests isundertaken by the Department of Finance,while other processing, including the pro-duction of certificates, is doite at the Uni-versity of Zambia.

Swaziland and Ethiopia have encoun-tered serious problems that have thwartedtheir efforts to use modern equipment toscore large numbers of examinations effi-ciently. In the former, machine-scoring had

to be discontinued as the process producedtoo many errors. In Ethiopia, the inabilityto obtain appropriate scanner parts due to alack of appropriate arrangements for secur-ing foreign currency resulted in a return tohand scoring of the Grade 6 and Grade 8national examinations.

Ftmctions of Examinations

In theory, public examinations are ex-pected to serve a number of different func-tions. For teachers and pupils, they can actas an incentive for greater effort. They areregarded as a fair means of distributingscarce educational benefits. They can pro-vide feedback to teachers and students ontheir strengths and weaknesses. Public ex-aminations can help unify the teaching ef-fort throughout an entire country aroundcommon goals and themes. For students,they provide formal evidence of educationalachievement (certification). Public exami-nations may also serve an accountability roleby offering evidence of standards attainedby individual teachers and schools. Lastly,and most important in the present context,educational systems, governments, and otheremployers use examination results for se-lecting appropriate candidates for the nextlevel of schooling or for employment.

28

Selection

In practice since educational opportuni-ties form a pyramid with fewer places avail-able at each successive level, the primaryfunction of almost all public examinationsreviewed in the fourteen countries in ourstudy is selection. While most of the exami-nations also serve a certification function,some, most notably the concours in Franco-phone countries, are designed solely to se-lect students for the next highest level of theeducational system. The highly competitiveconcours is used, for instance, in Togo at theend of the secondary junior-cycle to selectstudents for the three-year senior-cycle pro-gram and again at the end of this cycle to

select candidates for colleges and universi-ties and for foreign study.

The capacity of educational systemsrather than the academic achievement ofstudents plays the crucial role in determin-ing the percentage of pupils who advanceto the next highest level. Lack of schools,especially secondary schools, inevitably re-sults in a severe bottleneck at the end ofprimary school. The number of selected orpromoted students depends ultimately onthe number of available places. Pressure fora place in a secondary school is most pro-nounced in Rwanda, Uganda, Chad, andZambia. In Ethiopia, on the other hand,little selection pressure at the end of pri-mary school is due to the ready availabilityof places; the vast majority of students cantransfer to secondary school. Here, as inthe other countries, however, the selectionfactor is emphasized at the end of seniorcycle. Evidence for this claim is providedby the large number who repeat the ESLCEin the hope of improving their grade-pointaverage to qualify for one of the relativelyfew available places in third-level educa-tional institutions.

In a number of Francophone countries,rigid marking systems are employed at end-of-year grade examinations, administered toselect or to promote students to the nexthighest grade. Among the most severe isthat employed in Chad secondary schoolsin which a mean mark of ten or more isrequired for promotion, greater than sevenand less than ten for repetition, while a markof seven or less results in expulsion. Givenwhat is known about the extent of markerunreliability in examinations, this procedureobviously presents problems. If, for ex-ample, a student obtained a score of 8.5 ona test with a standard error of measurementof 0.8 we could say that the individual's"true" score lies between 10.1 and 6.9 witha 95 percent level of confidence. This rangeencompasses each of the critical promotion,

29

retention, and expulsion cut-off points.Thus, it is reasonable to assert that a stu-dent who was expelled from school on thebasis of examination performance couldhave merited either retention or promodonhad an accurate measure of student achieve-ment been available.

Since the selection function of examina-tions relates to the identification of studentswho are most likely to succeed at the nexthighest level, the lack of studies of the pre-dictive validity of examinations is surpris-ing. Clearly it would be possible to increasethe efficiency of the selection process byusing shorter tests. Such tests would re-quire less supervision and less time in mark-ing and in analysis and so could result insubstantial financial savings. Overempha-sis on increasing the efficiency of the selec-tion function, however, would inevitably runthe risk of reducing the content coverage ofthe examinations (see Chapter 5). As a con-sequence, we would expect that over timesections of the curriculum not being exam-ined would be neglected in schools. Thecombination of selection efficiency and cur-ricular validity in examinations involves adelicate balance.

Certification

The emphasis on the selection functionof examinations is so pervasive that there isa danger of losing sight of their certificationfunction. With the exception of the concoursand those taken in some countries prior tothe primary school-leaving examinations,virtually all of the examinations have a cer-tification function. Formal certificates thatindicate candidates' academic achievementsmay be important credentials for gainingemployment or for qualifying for furthertraining. In Ethiopia, for example, posses-sion of the Grade 6 and Grade 8 certificatesmay be required as a minimum qualifica-tion for a range of occupations, for exampledrivers and filing clerks. Lower level cer-

3 LI

tificates, however, as the numbers possess-ing them increase, are losing their currencyin the labor market.

There has been a growing tendency inrecent years to use the same test to meetselection and certification requirements. InMauritius, for instance, prior to the intro-duction of the present Certificate of PrimaryEducation (CPE), a primary school-leavingcertificate served certification requirementswhile a different examination (Junior Schol-arship) administered around the same timesatisfied the selection requirement. A real-ization that the two formal examinationscontributed to excessive pressure on ten- andeleven-year olds led to the dropping of bothexaminations in favor of one (CPE) that nowserves both functions.

An examination may be an efficient se-lection device but may fail to serve the im-portant certification function. Certification,which is recognized as a function of exami-nations in all countries, is primarily con-cerned with providing a measure of curricu-lum coverage. Adequate curriculum cover-age is more likely to be provided in broad-based examinations that extensively samplethe prescribed syllabus than in examinationsthat are confined to narrow content domains,which cover material that is normally notdirectly prescribed in school curricula, suchas nonverbal reasoning ability, or that areconfined to one or two assessment proce-dures.

Accountability

Public-examination results are frequentlyperceived as providing evidence of schooleffectiveness. In a number of countries,schools and teachers may be held account-able for their students' achievements as re-flected in performance on public examina-tions. The use of examination results to holdteachers or schools accountable is magnifiedwhen results are published. Even when re-sults are not formally published, education

officials may congratulate officials of schoolsin which students have performed well orrequire less successful schools to explainpoor examination performance.

Performance

Attempting to summarize and providefigures comparable across countries for ex-amination performance (such as are pre-sented in Table 3.2) is problematic. Such asummary fails to take into account differ-ences in criteria for awarding grades or inthe structure of performance (that may in-volve compensation across subjects or reach-ing a particular level in specific subjects) thatcan merit a pass or a particular grade. Evenwithin the same examination, different cri-teria can be used to indicate that a studenthas "passed." For example, in the school-leaving certificate examinations of a num-ber of countries, such as Kenya and Lesotho,to obtain a school certificate, a candidatemust have passed in a specified number ofsubjects. However, candidates who do notobtain the specified number of passes mayobtain a certificate in the subjects which theydid pass. In our table, we provide figuresfor the students who pass the total certifi-cate examination and, when available, fig-ures for candidates who failed to get a gen-eral certificate but passed some subjects inthe examination. Thus, problems in compa-rability can arise from the fact that what ismeant by a pass can vary from examinationto examination, not only between but alsowithin countries. Further, in a number ofcountries (Kenya, Zambia), a distinction isnot made between pass and fail on the pri-mary school-leaving examination.

30

To complicate matters, success or failurein the eyes of students, parents, and teach-ers may not be determined so much by ex-amination-performance level as by whetheror not a student qualifies for a place at thenext highest level of the educational system.A mark of 80 percent on a primary school-leaving examination may gain admission in

LIG

Table 3.2PUBLIC-EXAMINATIONS PASS RATES

Country/linguisticstatus Primary Secondary

AnglophoneKCPE

Kenya (1987) Not applicable

PSLCELesotho (1987) 91.0%

Mauritius

Swaziland (1988)

Uganda (1987)

Zambia (1983)

Francophone

CPE58.4%

PC79.6%

PLE82.2% (1989)

G7Not applicable

KCE KACE78.0% 56.3%

CE COSC65.1-68.8% 19.4%-31.7%(1980-87) (1980-87)

62.2%-74.7%(GCE)

SC HSC58.1% 52.6%

JC COSC67.3% 54.0%

UCE UACE80.1% 66.4%

G9 SC35.8% 63.6%

+ 32.4(GCE)

CEPEICEPEA Concours BEPC BacChad (1988) 56.8% 35.3% 49.8% 12.9% (1987)

7th yearEntrance BEPC Bac

Guinea (1988) 56.6% 48.8% (i) 21.0%(ii) 30.6%

CEPE CFEPCES BacMadagascar (1988) 49.0% (1987) 26.7% (i) 20.1%

(ii)18.2%

CEE 1st year Sec. BEPC BacMauritania (1988) 22.0% 29.2% 37.6% 25.6%

ConcoursCEP for Sec. NA 93.2% (Public)

Rwanda (1988) NA 7-9% NA 70.0% (Private)

CEPD BEPC BacTogo NA NA NA

OtherEnd of Basic First Cycle Second Cycle

Elementary Complementary General Complementary

Cape Verde (1987) 67.3% 56.9% NA NA

Ethiopia G6 G8 ESLCE(1987) 82.5% 68.3% 97.7% (Amhric)

51.9% (Math)

3141

or country but may be insufficient to do soin another where there is a severe shortageof secondary-school places.

Most systems tend to use some form ofpreset achievement standard to awardmarks. While the percentage of studentsgetting various grades under this approachvaries relatively little from year to year,notable variation does occur. In 1986 inGuinea, for example, 27.4 percent of candi-dates passed the Grade 7 entrance examina-tion. Two years later, this figure had morethan doubled to 56.6 percent.

A number of systems transform exami-nation marks into normalized scores, for ex-ample Ethiopia and Lesotho. Marks aretransformed into standard scores so thatindividual students' reported scores reflecttheir performance relative to the perfor-mance of other students. Such an approachcan tell us little about what a pass means interms of a student's level of performance orknowledge. The standardization of scoresensures that, irrespective of level of achieve-ment, percentages passing, failing, or gain-ing distinctions are predetermined. Unduereliance on normalized scores can lead tothe adoption of grading practices that bearlittle resemblance to educational realities(Sebatane 1985). For example, an indepen-dent study of the secondary school-leavingexamination in Ethiopia reported that thecut-off raw score that distinguished betweenpass and failure on three of the 1986 sub-jects, English, history, and chemistry, ap-proximated the number of items that couldbe attributed to guessing the answer to eachitem (Cambridge Educational Consultants1988).

The summary data in Table 3.2 revealthat large numbers of candidates were un-successful in public examinations. In anumber of countries, frwer than one in fourpassed the primary school-leaving examina-tion. High failure rates (60 percent +) were

32

also recorded on the junior and senior-cyclecertificate examinations. In almost all coun-tries, pass rates were highest on the primaryschool-leaving examination and lowest onthe secondary senior-cycl. examination. Ingeneral, the secondary-school pass ratesseem to have been slightly lower in Franco-phone than in Anglophone countries.

Within individual countries, substantialvariation in performance rates has been re-corded between different regions. For ex-ample, the pass rate on the primary school-leaving examination (CEPE) in M4dagascarwas approximately 60 percent for candidatesliving in cities and as low as 20 percent forrural candidates.

Repeat students tend to lower the meanpass rates. In 1988, repeat students repre-sented 59.4 percent of all candidates in Ethio-pia for the ESLCE secondary leaving exami-nation. Their mean score (which was con-siderably below a D grade) placed themlower than other candidate categories. Onthe 1983 Zambian Junior Secondary-SchoolCertificate Examination, the success rate ofinternal candidates (52.5 percent) was morethan six times that of external candidates(8.4 percent).

Examination Repetition

Policies regarding examination repetitionvary. Students who fail public examinationsin Ethiopia may repeat the year in schoolprovided there is sufficient accommodation;otherwise they must study outside of thepublic-school system. Madagascar permitsmultiple repeats on its primary-leaving ex-amination. On the other hand, Guinea andMauritius limit the number of repeats whileZambia and Lesotho require the public-school candidates to pass some public ex-aminations on the first attempt. Both thesecountries, together with Kenya, Cape Verde,and Mauritius accept repeat pi ivate candi-dates.

Governance and Administration

Public examinations are organized andadministered either by the Ministry of Edu-cation (or special sections within the Minis-try) or by an independent body establishedspecifically for this purpose. In general, inFrancophone countries, responsibility isvested in the Ministry, while in a number ofAnglophone countries, it is vested in an in-dependent examination board or council. Ina relatively small number of countries, suchas Ethiopia and Togo, the Ministry has re-sponsibility for primary and junior-cycle sec-ondary examinations, but the senior-cycleschool-leaving examination (or part of it inthe case of Togo) is administered by univer-sity authorities. Appendix 2 identifies theagency or body that has been given respon-sibility for individual public examinationsor shares responsibility with external exami-nation organizations, for example the Uni-versity of Cambridge Local ExaminationSyndicate.

Examination systems in Kenya, Lesotho,Swaziland, Mauritius, Zambia, ahd Ugandaare run by independent boards. Member-ship of these boards tends to be drawn froma broad educational constituency. Ministriesof Education are represented by one or moremembers. The membership of the UgandaNational Examinations Board, for example,includes four ex officio members: the Vice-Chancellor of the University, the Chief Edu-cation Officer, the Chief Inspector of Schools,the Director of the National CurriculumDevelopment Cenixe; seven elected membersfrom professional bodies; and thirteen min-isterial appointees (mostly principals of dif-ferent types of schools and colleges).

In general, examination sections withineducation ministries and examination boardspublish syllabi and regulations for theirexaminations, appoint appropriate person-nel (markers, invigilators, etc.), and issue cer-tificates.

33

There is a trend in Anglophone coun-tries, though not in Francophone ones, to-wards the devolution of responsibility forthe administration of examinations to anauthority other than the Ministry of Educa-tion. However, even in countries in whichan examinations authority has responsibil-ity for examinations (for exartple, Ethiopia,Lesotho and Zambia), the authority stillmaintains close links with the Ministry andmay be dependent on other bodies for a va-riety of services, such as marking of testsand computer facilities.

Localization of Examinations

A trend in countries in which studentstake foreign examinations as part of theirnormal schooling is to localize all examina-tions, that is, to have examinations controlledwithin the country. There are only threecountries (Lesotho, Mauritius, and Swazi-land) in which the leaving-certificate exami-nation at the end of secondary schooling isdesigned and marked abroad (in Britain),and steps are being taken to change that situ-ation. Localization frees foreign currencyfor other activities, though whether themoney will be made available for improv-ing the resources available for examinationsis another matter. It also, however, raisesquestions about the maintenance of stan-dards and the acceptability of a country'seducational qualifications in other countries.The localization of examinations, particularlyat the end of secondary school, has beenviewed with apprehension in some coun-tries in case standards should fall as a con-sequence. Governments, examination bod-ies, and the general public are very keen tosee that this does not happen and that theirschool qualifications will have currency notonly nationally but internationally. Toachieve this, it will be necessary for govern-ments to invest more in evaluation and re-search activities relating to examinations.

ti r'o

Funding

Information from a limited number ofcountries that have made appropriate dataavailable suggests that levels of state fund-ing of public examinations vary consider-ably. At one extreme, Madagascar devotes10 percent of its total educational budget topublic examinations. Chad and Rwanda, onthe other hand, spend less than 0.5 percenton this exercise. Size of government expen-diture on public examinations expressed inmonetary terms or as a percentage of cur-rent government expenditure on educationmay represent only part of the total exami-nation expenditure. Where Ministries ofEducation meet all examination costs as theyarise, many, if not all of the costs, are ab-sorbed into the general operating budget ofthe Ministry.

The difficulty of estimating or conduct-ing a comparative analysis of examinationcosts is compounded by the fact that thenumber of cost-incurring items differsamong countries. In Cape Verde, for ex-ample, teachers who construct examinationpapers or who supervise at examinationsessions are not paid. Furthermore, at thebasic elementary level, markers are paid onlyfor correcting external candidates' scripts. Atthe secondary level, markers are paid by theschools.

In at least five of the countries reviewed(Kenya, Lesotho, Mauritius, Swaziland, andZambia), public-examination bodies allocateconsiderable portions of their budgets to thepurchase of foreign currencies. For securityreasons, Kenya devotes roughly 15 percentof its total budget to external printing of itsprimary and secondary examinations. InZambia, approximately 8,000 students takethe GCE 0-level examination each year, at acost of $782,500 to the Examinations Coun-dl. Since the amount that the Council ispermitted to charge candidates is not nearlysufficient to meet the GCE costs, it must sub-sidize the fee paid by each student. The

34

country's deteriorating financial position andthe difficulty of raising foreign currencyplace the continued operation of the exami-nation in jeopardy. Figures from Mauritiusfor 1986-87 indicate that foreign examina-tion syndicates acquired $1.16 million forfees for secondary-level examinations; Cam-bridge University Local Examination Syndi-cate received $970,000 of this total. Othercountries that pay substantial amounts offoreign currency, mainly for senior-cycleleaving-certificate examinations, are Lesothoand Swaziland.

While the inevitable increase in candi-date numbers will increase the cost of fund-ing public examinations, economies of scaleshould lessen unit costs. In Kenya and coun-tries dependent on externally produced andconrcted examinations, local printing, ex-amination production, and correction couldreduce costs.

Some examination authorities raise fundsthrough their publishing activities, notablyby publishing copies of previously adminis-tered public-examination papers.

Fees

In general, fees for end of primary-schoolexaminations are either not charged or arekept to a low level (see Table 3.3). This canbe attributed to government policies of en-couraging pupils to persist in school up tothe end of the first or compulsory cycle.Senior-cycle secondary-school examinationsare generally much more expensive for stu-dents than those offered at the end of juniorcycle. Candidates for examinations offeredby overseas examination boards or syndi-cates are required to pay high fees. The feefor the senior-cycle terminal examination inLesotho ($124.19) represents considerablymore than the mean monthly income percapita in the country.

Fees do not cover costs. In Kenya, forexample, although the National Examina-

Table 3.3CANDIDATE FEES FOR PUBLIC EXAMINATIONS

Countries/linguisticstatus

AnglophoneKenyaLesotho

MauritiusSwaziland

Uganda

Zambia

FrancophoneChadGuineaMadagascarMauritaniaRwanda

Togo

OtherCape Verde

Ethiopia

Scale

KCPE - $354; KCSE $42.49 (minimum)PSLCE - $1.86; ICE - c. $32.09; Senior Cert/COSCc. $124.19

CPE - none; Second level examinations - $67.00 averagePC - $1.25 per subject; ICE - Reg. fee - $3.33 + .

$2.92 per subject; COSC - Registration fee - $13.13 + $11.46per subjectPLE - $1.71; UCE - $13.14 (8 subjects); UACE- $13.43 (3 principal, 1 subsidiary subject)None G7 and G9 (except external); SC - $12.87 (8 subjects)

No feeNo feeCFEPCES $0.19; Bac.- $1.58NANone apart from Examen de recIassement andLe JunyCentral - $12.53

CEPD - $0.36; BEPC - $1.80; Bac.1 - $7.30;

Bac2 $12.90

Basic elementary - none; Basic complementary -$2.20; General secondary - $10.20; ComplementarySecondary - $2.00 per subjectGrade 6 none; Grade 8 none; ESLCE Gov, schoolsfree; Others: Registration fee - $2.50 + $2.00 per paper

tions Council effects considerable efficienciesthrough computer-scoring its primary leav-ing examination, its income from fees for theexamination represents less than 50 percentof the total cost. In Uganda, the ratios ofunit costs to unit fee for the three major pub-lic examinations are 1.34:1 for the PCE, 1.42:1for the UCE, and 1.44:1 for the UACE.

35

Security

A marked feature of public examinationsin Africa is the pronounced emphasis on se-curity. The security dimension is empha-sized throughout all stages of the examina-tion process: setting questions, printing,proofing, packaging, distribution, supervi-

sion, and correction. The security emphasissimply reflects the importance of examina-tion success in the lives of students and theirfamilies. Success offers some hope of secur-ing valued scarce employment or of advanc-ing to the next highest educational level. Fail-ure, on the other hand, can lead to a rela-tively frugal existence or at best a lost yeardevoted to preparing for the same examina-tion.

Commentators have noted "the ease withwhich malpractices can occue (Kelly 1986,p. 552). In Togo and Madagascar, concernhas been expressed that security is not wellmaintained. In Zambia in 1980, it was re-ported that candidates in three regions re-ceived help from invigilators during theGCE.Copying has been a problem in Ethiopia. Col-lusion between candidates and supervisors,though rare, has been reported in Swaziland.Premature opening of envelopes containingexamination papers has been noted inRwanda. In Uganda, types of examinationirregularities reported include leakage, im-personation, external assistance, smuggling,copying, collusion, and substitution (Ongom1990).

In general, however, examination authori-ties are confident their security precautionsare effective. Machine scoring appears tocontribute to test security because it reducesthe personnel requirement for scoring. Con-cern has been expressed, however, that dueto the development of sophisticated devices,the security of machine-scored data is by nomeans guaranteed. In Ethiopia, the typing ofGrade 6 and Grade 8 examinations is en-trusted to one typist. In many countries,people entering and leaving centres in whichexamination papers are printed are searched.Supervisors in Zambia must take an oath ofsecrecy. Similarly, employees of the KenyanNational Examination Council sign docu-ments binding them to the terms of the Offi-cial Secrets Act. Papers for examinations inKenya are printed in the United Kingdomdue to the absence of secure local printing

36

facilities. Elsewhere, papers are printed insecure premises, often as near as practical tothe date of the examination to avoid unnec-essary storage problems or security breaches.Distribution of papers is undertaken undersecure conditions.

Some systems use identification numbersto ensure that markers do not have access tocandidates' names. In the case of oral exami-nations, care is taken to ensure that candi-dates do not hear the questions beforehandor come in contact with others who haveheard the questions. When there is evidenceof misconduct at an examination centre inUganda, the entire examination may be can-celled. In Kenya, serious malpractice in rela-tion to the conduct of an examination canlead to a lengthy prison term.

Feedback

Public examinations can be used to pro-vide feedback to teachers, curriculum bod-ies, and educational policymakers. Severalof the countries in our review provide somekind of mechanism to make information onexaminations available to interested parties.

Past examination papers are importantsources of information in some countries andconsiderable use is made of them in schools.Such papers, however, are not available ev-erywhere, such as in Lesotho. The wide-spread involvement of teachers from variousparts of the country in item-writing, for ex-ample in Lesotho, has also been an importantsource of information to schools about theskills and knowledge tested in examinations.

Comparisons among Schools/Districts

Another source of feedback provided byexamination authorities is the publication ofthe examination results of students on aschool-by-school or district-by-district basis(Ethiopia, Kenya, Rwanda, and Swaziland).

4 61

Analysis of Results

Some examination systems have devotedresources to providing detailed analyses ofstudents' performance on public examina-tions. For example, in Lesotho, the test itemsfor the primary-school mathematics exami-nation, together with performance item sta-tistics for the population of examination can-didates, was sent to all schools. Item-analy-sis information has also been prepared inEthiopia, but the information was given onlyto Ministry of Education or ExaminationBoard personnel. Countries in which item-analysis data have been used for post-ex-amination reports are Kenya, Mauritius, andSwaziland. Chief Examiners may also pro-vide descriptive reports on the overall per-formance of students on individual exami-nations and even on individual examinationquestions. In some countries, such as Zam-bia, every school receives a copy of the ChiefExaminer's Report.

Information about examination perfor-mance can be helpful in identifying aspectsof the curriculum that appear to have beenpoorly taught or ignored by teachers. Wemight expect the information to contributeto improvement in the construction of ex-aminations and in future examination per-formance.

The KCPE Newsletter published by theKenyan National Examinations Council isan example of a systematic approach to-wards informing teachers about how stu-dents have performed on examinations, Thenewsletter, which is circulated to all schools,contains item-analysis data for each exami-nation paper. It also includes suggestionsas to why students may have selected par-ticular incorrect options. Examples of poor,average, and good answers to selected ques-tions are presented and teaching implica-tions are described. Teachers make frequentuse of the information provided in the News-

letter in their classes. For example, theyanalyze with their students the characteris-tics of "good" and "poor" essays presentedin the newsletter.

Timeliness

Prompt release of examination results isnecessary to allow for appropriate planningby schools and individuals. Results returnedafter the commencement of the new schoolyear represent an ineffidency, as they clearlydelay the assignment of pupils to appropri-ate classes. The lag between the time theexamination is given and the formal publi-cation of results depends on the efficiencyof collection procedures and the size of theexamination enterprise.

37

Even within countries, the tiLieliness offeedback varies from one examination to thenext. Evidence from Chad suggests that theresults of the CEPE and the Baccalaureat areavailable as soon as one week after the ex-amination while Grade 6 concours resultsmay take up to three months to be pub-lished.

A disadvantage associated with the Ke-nya KCPE Newsletter is that it does not ap-pear until the final term of the year follow-ing the examination, thereby lessening thepotential value of its feedback for teachers.Since, however, curriculum content tends toremain constant over the years, the overallutility of the newsletter is probably not se-riously diminished.

Challenges to Examination Results

A number of examination bodies permitappeal of examination results. In both Le-sotho and Swaziland, appeals are allowed,on payment of a fee, for a period of six weeksfollowing the publication of the primary andjunior-cycle results. In Lesotho, if the ap-peal is successful, the candidate is refunded

41 7

half the fee. Ethiopian authorities discour-age appeals, claiming that the process is bothexpensive and time-consuming, but theBoard does recheck results. Appeals in Zam-bia must be submitted through the princi-pal teacher; recheck fees are not charged.Similarly, in Swaziland, requests forrechecking the results of the CambridgeOverseas School Certificate Examinationmust be made through the principal. In thisinstance, the examination authorities do little

to encourage candidates to seek a recheck.It is stressed in an explanatory note that re-ports will be limited to one subject per school

on any one occasion of the examination andto the work of not more than six candidatesin that subject. The service is not madeavailable to private school candidates.

38

Outside-School Tuition

As happens elsewhere throughout theworld, students in examination classes fre-quently seek additional academic help out-side of school or in tuition offered in schoolsoutside of regular school hours. The empha-sis on private tuition is particularly pro-nounced in Mauritius. One study reportedthat 73 percent of Standard 6 pupils in theprimary school-leaving examination classtook private tuition. As many as 11 percentof Standard 1 pupils also took private tu-ition. Tuition frequently was provided ingroups of twenty-five or more. Approxi-mately 60 percent of secondary-school stud-ents also enrolled in tuition groups(Joywathsing and others 1988).

5 Issues in Examination and Assessment

In this chapter, we consider trends thatappear in the examination systems of at leastsome of the countries included in this studyand outline issues to which education au-thorities might direct their attention whenconsidering possible areas for reform in as-sessment and examination procedures.

In considering these issues, sight shouldnot be lost of the fact that considerable differ-ences exist between the examination systemsof the fourteen countries described in thisreport. In highlighting issues, one can easilygive the impression that contextual factorsthat operate in the individual countries areless important than they actually are (seeLittle 1990). Of obvious importance are dif-ferences in the educational systems of thecountries. For example, there are differencesin the numbers of students who are retainedat varying levels and differences in the use ofa vernacular language for instruction. Bothhave implications for the assessment of learn-ing. Among the differences in examinationsystems are differences in traditions, not onlyin the colonial period but also in the post-independence period, which reflect consid-erable diversity in the way in which nationalsystems of examinations have developed(Eisemon 1990). Examination systems alsodiffer greatly in the use of technology , forexample, in the use of computers and opti-cal-mark readers. Such differences have im-plications that go beyond the technology. Forexample, they affect the types of items usedin examinations, methods of marking, costs,and the ease with which information on ex-amination performance can be provided to avariety of publics.

39

Expanding Enrollments

1111.11=MINE111=1

Perhaps the most obvious trend in thedevelopment of examinations in recent yearsin many of the countries studied for thisreport, has been the enormous increase inthe numbers of students sitting for exami-nations. For example, between 1968 and1988, Swaziland registered an increase ofover 200 percent in numbers taking its pri-mary-certificate examination and an increaseof over 1,000 percent in the numbers takingthe junior-secondary school-certificate ex-amination. In Ethiopia, over a much shortertime span (1979-87), primary-certificate can-didates have shown a 240 percent increase,junior-secondary certificate candidates a 260percent increase, and senior-secondary cer-tificate candidates a 400 percent increase. Ingeneral, the higher the level of the examina-tion, the greater has been the rate of increase.

As we noted in Chapter 1, increases inthe numbers taking examinations, as wellas in enrollments generally, have been asso-ciated in people's minds with declining stan-dards in education. This has been particu-larly so when resources to deal with thisexpansion have not been available. Theincreased numbers have also forced a re-consideration of the content of curricula inschools. An academic type of curriculum,which might have been suitable when ac-cess to educatici was limited, is not likelyto be suitable when all or most of the popu-lation attend school. Curriculum change hasobvious implications for the nature of ex-aminations, implications that relate not justto content but to function. It is increasinglyrecognized that the dominant selective func-

4 Co'

tion of examinations must give ground tothe certilcation function. Just as curricu-lum change has implications for examina-tions, examinations may have implicationsfor curriculum. For this reason, changes arebeing introduced to the examinations ofseveral countries, including content andskills appropriate to the wider range ofachievements, needs, and interests repre-sented in the growing numbers of studentstaking examinations, to stimulate beneficialeffect on curricular practice in schools.

Expanding enrollments will conti-fiue to be apressing issue for many examination systems.For educational authorities they will lead to ademand for greater resources and to changes 'oothin the content and function of public examina-tions.

Resources for Examinations

Examination authorities appear to devotevirtually all of their time and resources tothe administration of the national examina-tions. Frequently they are understaffed.With a few notable exceptions, relativelylittle or no time is given to systematic analy-ses to improve examination quality. Indeed,most authorities do not have the researchcapability to undertake this task. Fre-quently, skilled personnel are not availablewithin the country. In a number of in-stances, however, people with formal train-ing in psychometric techniques have beenassigned to other sectors within educationministries. In Zambia, staff shortages withinthe Examinations Council have resulted inmembers of its Department of Research andTest Development having to devote most oftheir time to item-writing and pretesting.

In general those responsible for the con-struction of examinations have had little orno relevant formal training. Many, ofcourse, acquire skills through experience.However, weaknesses could be avoided ifadequate training facilities were available.

40

According to one national report (Ethiopia),...among the weaknesses observed insome papers were unclear instruc-tions, poor grammar, placing of ques-tions soir pages from the passageon which they are based, use of in-terrelated questions, (for example,the stem of a question prompts theanswer to an earlier question), andinadequate supporting art work" (p.40).

Governments and examination bodies willhave to address the issues raised by staff short-ages and lack of trained personnel.

Effects of Examinations on Curriculumand Teaching

We have already noted in Chapter 1 thatwhenever the results of external exarnina-tions are an important determinant of fu-ture education and life chances or are usedto evaluate teachers, the examinations willbe taken very seriously by teachers and stu-dents. In this situation, teachers pay par-ticular attention to the "tradition of pastexaminations." Having identified the con-tent and skills covered hi provious exami-nation questions, they coach students in theidentified content and skills. This :an havepositive effects on curriculum. For example,in Lesotho, the marks (25.0 percent) in theJunior-Certificate Examination awarded forpracticals help to ensure that this aspect ofthe curriculum is covered throughout theschool year. In Kenya, by including itemsthat relate to local conditions and that at-tempt to measure higher-order cognitiveskills in the KCPE, the Kenyan National Ex-aminations Council has drawn teachers' at-tention to a new range of knowledge andskills.

However, when people talk about the"tradition of past examinations" definingcurriculum in schools, very often they arereferring to negative effects of this tradition.

What they mean is that teachers are prob-ably spending inordinate time in simplyworking for the certifying tests, often at theexpense of subjects and skills not tested inthe examination.

The narrowing effects of the public ex-aminations on curriculum have been com-mented on in most of the national reportsand reinforce the findings of research in avariety of countries relating to the effects ofexaminations that we outlined in Chapter 1.In Ethiopia, curriculum areas that tend toreceive relatively little attention in the class-room include writing, vocational, oral andaural skills, use of reference material, andpractical work in science. In Madagascar,subjects such as music and physical educa-tion, which are not examined, tend not to betaught in primary schools. In Zambia, locallanguages, practical subjects, and productionactivities are not formally examined and con-sequently are undervalued in schools. InZambia as well as in Ethiopia, where theprimary school-leaving examinations arelimited to multiple-choice items, writingskills receive inadequate attention. In CapeVerde, the status of subjects such as visualeducation, handicraft, and physical educa-tion is diminished as these are not testedwhen ascertaining who is to be promoted.The Mauritius report observed that thehighly competitive Certificate of PrimaryEducation Examination has led to "a neglectof broader curriculum goals not included inthe CPE syllabus" (p. 34).

The importance of examination resultsfor the pupil and for the teacher is reflectedin the amount of time spent fostering test-taking strategies. A recent report fromUganda observed that "the system is soexam ridden that the entire teaching andlearning is geared to passing the exam andgetting the good marks needed for entry tohigher levels of education" (Uganda, Minis-try of Education 1989, p. 7). In Zambia,"even against their better pr( ressional judg-ment," teachers feel compelled to direct their

41

efforts to obtaining good examination re-sults, "out of a concern for their pupils, andto safeguard their own reputations" (Kelly,Nkwanga, Achola, Kaluba, and Nilsson1986, p. 551). Teachers in Cape Verde makeextensive use of the examination papersfrom previous years; virtually all of the timespent in school is devoted to preparing stu-dents for examinations. Practice in exami-nation-taking techniques is also a feature ofprimary education in Mauritius and Mauri-tania. In Ethiopia and Madagascar, routineclassroom examinations reflect closely theitem types used in the official examination.Teachers in Ethiopia and Kenya use the mul-tiple-choice format of the national ex,irnina-tion in their regular classroom tests as earlyas the junior primary grades. Indeed, teach-ers have been observed to use the multiple-choice format not only in examining but inteaching, as early as in the first grade.

The amount of time teachers in a num-ber of Francophune countries are requiredto devote to formal assessment must of ne-cessity limit teaching time. For example, inTogo, Madagascar, and Guinea, teachers areexpected to administer monthly, trimester,and end-of-year examinations.

Examinations are often regarded as animportant source of motivation and as en-couraging teachers and students to workhard. By targeting important elements ofthe curriculum, they can focus student learn-ing and may raise students' levels of achieve-ment. And success on a public examinationis an achievement of significance not alonefor students but also for parents and teach-ers. The significance of passing the KenyaCPE examination in the eyes of one youngcandidate is well portrayed in an essay writ-ten during the examination (Appendix 3).

Evidence from a number of countries,such as Mauritania and Zambia, however,indicaLs that examinations tend to promotea passive colcept of learning. Furthermore,since scoring high marks in examinations

becomes the dominatint, interest of pupils,"pupils who cannot secure places at the nexteducational level are stigmatized as failures.There is even some tendency to judge thepersonal worth of an individual on the ba-sis of his examination performance" (Kelly1986, p. 26).

Examination authorities must take into ac-count that while motivating students and teach-ers, important public examinations also help toensure that curriculum content likely to be testedwill be covered by teachers; that subjects thatare not tested tend to receive relatively littleteaching emphasis; that considerable class timeis devoted to fostering test-taking strategies; andthat passive concepts of learning are promoted.In addition, society at large and, in particular,pupils themselves frequently equate examinationperformance with personal worth.

Reducing the Burden of Examinations andIntroducing School-based Assessment

In a number of countries, efforts are be-ing made to reduce the burden of examina-tions. Three broad strategies to achieve thiscan be identified, although it is not usuallyclear in the case of individual countries whatstrategy will eventually be adopted.

One strategy involves the reduction ofthe number of public examinations throughthe abolition of examinations and certifica-tion at particular points in the educationalsystem. For example, in Kenya, public ex-aminations and certification no longer oc-cur at the end of the junior cycle of second-ary schools, while other countries (CapeVerde, Ethiopia, Guinea) plan to reduce thenumber of their examinations.

In the second strategy, procedures arealready in place or proposals exist to allo-cate a percentage of public-examinationmarks on the basis of school-based assess-ment. The school-based assessment couldbe used in areas in which a terminal exter-nal examination is inappropriate, for ex-

42

ample in the assessment of practical orproject work, or it could be more generaland overlap with elements of the externalassessment.

Examinations in Ethiopia (at the primarylevel) and in Lesotho (at the junior-second-ary level) already have elements of school-based assessment. There is also a definitecommitment in a number of countries (Le-sotho, Swaziland, Uganda, Zambia) to makeschool-based assessment an important fea-ture of public examinations. However, thisis being done with caution. For example, inZambia, the recommendation is that continu-ous assessment in the public-examinationsystem should be confined to performancein practical subjects (Zambia, Ministry ofEducation 1977). A recent report in Ugandacontained a strong recommendation thatschool-based continuous assessment shouldaccount for 20 percent of the marks allo-cated for each subject in the primary school-leaving examination (Uganda, Ministry ofEducation 1989). The assessment, whichwould be moderated by the schoolinspectorate and officials of the Uganda Na-tional Examinations Board, would includeassessment of performance in class, of par-ticipation in community projects, and of cre-ative activities. In Swaziland, it is hopedthat school-based assessment will eventuallyreceive a weighting of 50 percent of the to-tal mark in the primary-certificate examina-tion.

A third possible strategy is to base certi-fication entirely on school-based assessment.This is not envisaged in any counti: in ourstudy in the immediate future.

Arguments can be advanced in favor ofand against school-based assessment(Heyneman 1988; Pennycuick 1990). Argu-ments in favor go as follows:

Since assessment by teachers is a cru-cial component of good learning andteaching, every effort should be madeto improve teachers' competence in thisarea (Crooks 1988). If school-based as-

sessment becomes part of the certifica-tion process, it is likely that greater ef-fort will be invested in improvingteachers' general competence in assess-mem', and this should have beneficialeffects on teaching and learning.

School-based assessment providesimmediate feedback information toteachers on student achievement andteaching effectiveness.

Since school-based assessment is car-ried out over time and by a person whoknows students well, it is likely to pro-vide a more valid and reliable ap-praisal of a student's achievementsthan can a single external terminalexamination. In this context, one com-mentator in Zambia has observed thatthe school is "the only place wherethere is enough information to do rea-sonable justice to a pupil" (Kelly 1986,p. 20).

School-based assessment permits anextension of the range of curriculumtopics which are examined. The presentsystem of examinations limits the rangeof achievements that can be assessedand must narrow the curriculum inschools. Aspects of achievement thatcannot be satisfactorily assessed in aterminal examination include astudent's ability to plan and organizea project and persevere with it overtime. While the assessment of oral andpractical skills may be carried out in aterminal examination, inevitably it willbe limited, artificial, and expensive.

School-based assessment reduces theundesirable back-wash effects of exter-nal examinations.

School-based assessment, if spreadover the year, can increase the level ofpupil motivation and appiicationthroughout the year.

Some of the following arguments ad-vanced against the use of school-based as-sessment in the certification of students:

Its use can change the nature of the

43

relationship between teachers and stu-dents towards making the judicial as-pect of the teacher's role more promi-nent.

Marking standards in school-basedassessment are likely to vary bothwithin and amcng schools. Whilemoderation procedures can help, theytend to be expensive.

School-based assessment can subjectteachers to considerable parental pres-sure, especially during the periodsleading up to and immediately aftercritical public examinations.

School-based assessment would re-quire teachers to devote more time toassessment and recording.

School-based assessment gives riseto a variety of administrative problemsfor schools, such as what to do whenstudents are absent for tests or whenstudents transfer from one school toanother.

Teachers' assessments are subject toa variety of biases.

In many instances, it is difficult ifnot impossible to apply school-basedassessment to non-school-based candi-dates.

Such considerations led the participantsin a conference on school-based assessmentin Swaziland to conclude that if such as-sessment is to be used in combination withexternal-examination results in the finalnational summative assessment of students,a number of conditions should be fulfilledto at least mitigate, if not completely remove,the problems to which its use is likely togive rise. First, national item banks of testsshould be available to teachers. Second,clear guidelines for assessment should beprovided. Third, procedures for recordingand interpreting should be common to allschools. And fourth, standards betweenschools "should be controlled statistically"and should be moderated and controlled byschool inspection (Swaziland, Ministry ofEducation 1987).

Clearly advantages and disadvantages areassociated with both school-based assessmentand external examinations. Individual coun-tries should consider the appropriateness ofthese procedures for the conditions in whichtheir educational systems operatethe pres-sure for selection, the level of training ofteachers, and the availability of resources tosupport either a school-based or external sys-tem of assessment. When selection is an im-portant issue, any consideration of a systemin which teacher assessments are a basis forimportant decisions about students must rec-ognize that the assessments should not onlybe accurate but should be perceived to be so.To achieve this situation, teachers would re-quire the support of a high level of trainingand a range of services, including guidanceand secretarial support. In many countriesoutside Africa in which school-based assess-ment is used, teachers also have available arange of aptitude and achievement tests thatcan be used to set objectives, to monitor stan-dards, and to provide a basis for student guid-ance. The problems that can arise whenschool-based assessment is introduced into asystem in which classes are large and the nec-essary resources are not provided have beendocumented in Lesotho (Sebatane 1985) andin Nigeria (Ali and Akubue 1988).

The issue of school-based assessment is avery live one in many countries. As educa-tional systems develop and expand and as thepressure for selection eases, we may expect adecline in the need for external examinations.Countries will reach that point at differentdates. As they work towards investmentin the development of the required infrastruc-ture for an effective system of school-basedassessment (including teacher training)should have the short-term effect of improv-ing the quality of assessment and instructionin schools.

Consideration should be given to increasingthe role of school-based assessment. This will re-

quire extensive pre-service and in-service teacher-training programs as well as other resources.

Use of Multiple-Choice Tests and TheirAssociated Technology

Partly as a result of the growth in the num-bers of students taking examinations, therehas been an increase in many countries in theuse of objective multiple-choice tests. In ourstudy, half the countries used such tests intheir primary-school certificate examinations.Such tests are cheaper to score when largenumbers of candidates are involved. They arelikely to be more reliable and to providebroader curriculum coverage than essay tests.They also make the provision of feedback onstudent performance easier. Related to theuse of objective tests has been the growth inpretesting of items, in mechanical marking,and in the use of computers to process results.Despite the apparent advantages of these pro-cedures, the use of objective tests and the tech-nology usually associated with them, some ofwhich can be applied to more traditionalkinds of examinations, can cause a number ofproblems.

44

The first problem arises from the fact thatthe selective role of examinations may be re-inforced by the introduction of computertechnology and the ease with which statisti-cal procedures can be used to standardizemarks and determine cut-off points for gradeson the basis of the number of students obtain-ing certain scores. This can happen both withexaminations that use multiple-choice testsand ones that do not.

A second problem is that the new proce-dures, especially the selection of items thathave high indices of discrimination for mul-tiple-choice tests, tend to reinforce the viewthat the latent-trait model of classical psycho-metric theory adequately represents theschool achievements of students. The as-sumptions on which this approach is based

that when many discrete behaviors co-vary,they have a single underlying determinant, of

which individuals possess varyingamounts--can be challenged even in the caseof western students (Christie and Forrest1981). Their validity is more doubtful in thecase of students who live in two culturalworlds, with all that implies in terms of lan-guage use, modes of cognitive functioning,and methods of learning (Kellaghan 1961,

1968).

Rigid adherence to a policy of discardingitems with low item-test correlations runs therisk of eliminating items that prove to be ei-ther very easy or very difficult (Greaney1980). Thus, the routine discarding of itemsthat have low correlations with the total testscore also runs the risk of making the test lessrepresentative of the defined universe(Cronbach 1971; Kwasna 1974; Quansah1985).

A further problem arises from the fact thatthe introduction of the new technical devel-opments led in some countries to droppingessay examinations and to a heavy reliance onmultiple-choice tests. An examination sys-tem made up only of multiple-choice itemshas several disadvantages. For one thing, itcannot measure some of the outputs of edu-cation that most people would regard as im-portant, such as oral and practical skills, in-cluding the ability to write continuous prose.Further, muitiple-choice tests are most effi-cient in measuring recall and recognition offactual information; it is extremely difficult toconstruct such tests to measure higher-orderskills. Limiting an examination to multiple-choice items is likely to have undesirable ef-fects on what goes on in schools since, in pre-paring students for the examination, thetemptation will be to pay little attention toskills other than those which can be measuredby the test. If higher-order and writing skillsdo not form part of public examinations, theyare not likely to receive the attention theyshould in schools.

Examination authorities should be sensitive tothe fact that while multiple-cho:a tests offer thebenefits of efficiency, test reliability, and potentialfor feedback to teachers, they are subject to a num-ber of serious limitations. These include undulyarbitrary use of standardized marking schemesand cut-off scores attributable to the ready avail-ability of computers, an assumption of a unitaryachievement trait, psychometric procedures thatreduce the content validity of a test and the neglectof important educational skills not amenable tomultiple-choice test assessments. Since decisionsabout students are likely to be more valid and fairwhen based on multiple sources of information, in-sofar as it is feasible, candidates should be requiredto demonstrate their knowledge and skills in a va-

riety of contexts.

45

Validity

Validity is concerned with the appropri-ateness of descriptions, inferences, classifica-tions, or decisions that are made on the basisof examination results. As far as curriculumvalidity is concerned, the responsibility fordetermining I-hat questions in public exami-nations reflect adequately the offidal curricu-lum rests ultimately with the body entrustedwith the operation of the examination system.If an examination is to possess curriculum va-lidity, individual questions should representa balanced and adequate sampling of the cur-riculum course. While some systems, such asin Swaziland, take the precaution of usingmoderators to establish the extent to whichthe elements of the prescribed curriculum arecovered, others, for example Madagascar andTogo, simply select questions from lists sub-mitted by teachers.

Part of the reason for the lack of adequatecurriculum coverage in examinations is thatstudent selection is a major function of theexaminations. An examination can be anefficient selection instrument even if largesections of a curriculum are excluded fromit. Further, since an examination designedfor selection will have as a major interest

r, r-t./ ki

discrimination between students dose to thecut-off point for selection, questions in-cluded may be too difficult for the generalbody of candidates, something that is borneout by the poor level of student performanceon many public examinations in Africa.

Despite the emphasis on selection in pub-lic examinations, study of the predictive va-lidity of performance on the examinationshas received little attention. The limitedavailable evidence is not very encouraging.In one study in Ethiopia, students' perfor-mance on the ESLCE mathematics exami-nation did not predict success in first-yearmathematics in the university (Asmersonand others 1989). A somewhat similar studyconducted by the University of Zambia con-cluded that the low predictive validity ofexaminations rendered their use for selec-tion at all levels problematic (Kelly and oth-ers 1986).

Attempts have been made in a numberof countries to improve aspects of the valid-ity of examinations. One such attempt in-volves the pretesting of items used in mul-tiple-choice tests, a device that, it might beargued, should improve curricular validity.However, pretesting may not have this ef-fect (Madaus, Airasian, and Kellaghan 1980).Selection of items of intermediate levels ofdifficulty for the final test, like the use ofhigh discrimination indices, runs the risk ofmaking a test unrepresentative of the de-fined universe. Such a practice is more likelyto contribute to instructional validity (whatis taught in schools) than to curricular va-lidity (what is laid down in the curriculum),since only items that a given proportion ofpupils (about half) get right will be retainedin the test. Thus, only items pupils havehad the opportunity to learn (presumablyin school) will find their way into the test.If schools consistently ignore aspects of thecurriculum, items relating to those aspectswill not be included in the test, creating asituation in which certain curriculum areasare not tested (Kellaghan 1990).

46

However, discarding items that do notexhibit intermediate levels of difficulty willalso affect a test's instructional validity sincematerial that is well taught will be excludedfrom the test. It is thus important that de-cisions to include items in a test should notbe based on the common practice of select-ing only items of intermediate level of dif-ficulty, which will maximize a test's abilityto discriminate between student perfor-mances, but, at the same time, will reduceits curricular and instructional validities.

An alternative to norm-referenced stan-dardized tests was proposed in the Mauri-tius report. It suggested that a pass shouldbe redefined in terms of essential compe-tencies for primary-school graduates ratherthan in terms of standardized scores. Eachletter gmde would be defined in criterion-rather than in norm-referenced terms, andgrades would indicate the performance lev-els of students. The approach would alsohave the advantage of lending itself morethan norm-referenced procedures to themonitoring of changes in achievement stan-dards over time. Since a criterion-referencedapproach might not adequately serve the se-lection function of the examination, the Mau-ritius report goes on to recommend that stu-dents in the top grades, such as A and B, beranked for selection purposes. While theproposal has much to recommend it, the dif-ficulties of specifying essential competenciesshould not be underestimated.

A body responsible for the construction ofexaminations has a clear obligation to attend tovalidity issues. In this, the need to achieve bal-ance between the certification and selection func-tions of the examinatlons must be kept in mind.The examinations body should ensure that theprocedures it uses in the construction of exami-nations are more likely to contribute to, than tointerfere with, the examination's validity. If ex-aminations are used for selection, the examina-tion body should accumulate evidence to showhow well performance on examinations reflectsreal-life educational performance.

Accountability

Examination results are sometimes usedas a measure of school effectiveness. Whenthis occurs, teachers and others tend to beheld accountable for the performance levelsof students. Frequently, successful perfor-mance is equated with the percentage of can-didates who advance to the next highestlevel of the educational system.

In Kenya, national and provincialrankings of schools based on public exami-nations are published annually to providethe public with an opportunity of compar-ing schools. In districts that occupy thelower end of the merit order, teachers maybe accused of being lax, administrators ofnot performing their supervisory duties, andparents of not being "development-minded"(Oloo 1990). Zambian teachers also tend tobe blamed if too few of their pupils are se-lected on the basis of examination resultsfor further schooling. "Good" teachers inthe Prai region of Cape Verde are those whoattain a promotion rate of 50 percent at theend of the school year. Examination perfor-mance is used by the Ministry of Educationin Swaziland and by the general public as ameasure of school effectiveness. In Lesothoalso, examinations serve an accountabilityfunction; teachers and schools with highexamination marks are usually regarded as"good." In Uganda, where most of theschool fees are paid by parents, examina-tion results can have a fairly rapid impact;"good" results tend to increase the pressurefor school places.

The ranking of school districts or schoolsis more complicated than current practicewould seem to acknowledge. In fact, infer-ences made about the "effectiveness" of

schools or districts on the basis of rankingsmay be inaccurate and unfair. First, it hasbeen shown in the United States that rank-ings can vary greatly depending on theschool outcome that is used (Guskey andKifer 1990). Second, comparisons betweenschools on the basis of examination resultsgenerally fail to take into consideration thesocial or even physical conditions underwhich schools operate. However, if statisti-cal adjustments to school outcomes are madeon the basis of student characteristics andtheir backgrounds, level of teacher qualifi-cation, and availability of material resources,a very different order of merit might emerge.Third, whether or not adjustments are madeto school-output measures, errors in the mea-surements on which rankings are based arenot likely to be taken into account whenjudgments of merit are being made aboutschools. For example, will observers knowthat the difference between a rank, say, of35 and one of 38 may only be a matter ofchance? Finally, in cases in that parents orstudents have a choice of school, the publi-cation of results may lead to schools thatare perceived to be doing well to attract stu-dents of high levels of scholastic ability,aspiration, motivation, and parental supportwhile those that are perceived to be doingbadly, even though they may be more "ef-fective" than schools with better results, willbe avoided by such students. This situationin time could lead to low morale in indi-vidual schools and to a ghettoization of theschool system.

The effects on schools of using examinationresults as an accountability mechanism demandsimmediate attention from education authorities.In particular attention should be directed to theidentification of "successful" and "failing"schools and teachers on the basis of examinationresults and to the publication of school rankings.

47

rUI

Language of Examination

In thirteen of the fourteen countries inour study for which data were available,there was an average of 4.7 principal lan-guages (i.e., indigenous languages spokenby at least 10 percent of a country's popula-tion and any other language that serves asan official language, lingua franca, or me-dium of instruction) (World Bank 1988, Table542).

This situation creates serious problems re-lating to language of instruction and languageof examinations. Countries have adopted avariety of strategies to deal with the situationand, in some cases, have changed these strat-egies over time. The most common practiceis to use a local lanpage for instruction inthe early grades of primary school. In somecountries, the relevant international language(French or English) is also used at this stage.The medium of instruction in the highergrades of primary schooling in most countriesis officially either French or English. In allcountries, French or English is the mediumof instruction in secondary schools, except inCape Verde, where it is Portuguese. In somecountries, some schools also provide instruc-tion in an alternative language, for example,Arabic in Mauritania. In all cases, with theobvious exception of language subjects, thelanguage of examinations is the same as theofficial language of instruction.

For the majority of pupils in the countriesin our study, the language used in instruc-tion and in examinations (English or French)was not their mother tongue. We know rela-tively little about the difficulties that Africanstudents experience in the acquisition of a sec-ond language or about how educationthrough the medium of a second languageaffects the quality of that education. Not sur-prisingly, some commentators have attrib-uted low levels of achievement in public ex-aminations to the language difficulties of stu-dents (Eisemon 1990). In Mauritius, for ex-ample, where the lingua franca is Creole, some

48

candidates' poor examination performancesis due to a failure to understand questions setin English. The low pass rate (38.8 percent)recorded by Madagascar students in theTechnical Baccalauriat has been attributed tothe extensive use of French in the educationalsystem. In Ethiopia, where English is the me-dium of instruction in secondary schools, itis accepted that there will be a move towardsassessment in Amharic in the Grade 8 andESLCE public examinations.

The choice of language for examinations hasto be considered in the wider context of choice oflanguage of instruction. The decision is a diffi-cult one for many countries and requires contin-ued research, analysis, and review.

Comparability of Performance

Possession of a certificate, such as a pri-mary school-leaving certificate, would implythat students had attained an acceptable andcomparable level of performance in definedachievement areas. Further, one might as-sume that the certificated students had ad-dressed themselves in examinations to thesame tasks under similar conditions. How-ever, current practices in some countries re-lated to "compensation" raise a number ofissues that suggest that these assumptionsmay not be tenable. Basically, the practice ofcompensation involves allowing credits ormarks that a student has achieved in one cur-ricular area to make up for a relatively poorperformance in another area. Questions ob-viously arise about the comparability of stu-dents' achievements when different achieve-ments are involved or when the achievementsare measured in different ways.

In Guinea adjustments are made to exami-nation marks to take into account the fact thatsome teachers may not have covered certainelements of the curriculum, again raising theissue of comparability of results. Cape Verdestudents are exempted from an oral exami-nation if they achieve a certain level of per-formance in the written test. In fact students

may gain exemption from each of the threenational examinations on the strength of theirperformances on internal school-based ex-aminations. A Togo candidate who scores lessthan 10 but at least 9 out of 20 can compen-sate for this relatively low mark by taking anoral examination. In the Ethiopian Grade 6and Grade 8 examinations, it is possible tocompensate for poor performance levels incertain subjects. For example, candidateswho score poorly in Amharic, English, andmathematics, three important subjects at sub-sequent educational levels, can qualify for acertificate by moderate to good performancein political education, general science, and so-cial science.

In Lesotho a first-class pass is awarded toprimary certificate candidates who obtain agrade of at least 60 percent in English,Sesotho, and mathematics and at least 60 per-cent in the examination as a whole. To ob-tain a second-class pass, a grade of at least 50percent has to be obtained in these three sub-jects and at least 50 percent in the overall ex-amination. A third-class pass is awarded tocandidates receiving 40 percent in any threesubjects and also on aggregate. A studentwho obtains 90 percent in all subjects with theexception of, say, English, in which he or shegets 40 percent, would merit only a third-classpass. It could be argued that the overall per-forrnance of this candidate is superior to thatof another who obtained a first-class pass byscoring 60 percent in all subjects.

In these examples, it is often difficult todetermine what possession of a formal cer-tificate means in terms of levels of scholasticachievement since students' achievements areassessed in different ways. However, the useof different modes of assessment does notnecessarily give rise to this problem. Whileit may have this effect, multiple measures ofachievement can be useful in allowing a per-son to exhibit a particular skill, trait, orachievement.

49

5 '3

Examination bodies that use compensationprocedures should ensure that these do not giverise to ambiguities in the interpretation of exami-nation performance.

Quotas

To meet social or political objectives, quo-tas are used in some countries, along withexamination results, to control student flowto the next highest level. In effect, quota sys-tems guarantee that a certain number of can-didates from a particular category, who oth-erwise would not advance to the next level,do so. Where this happens, the success ofthese candidates occurs at the expense ofother candidates who would have advancedhad no quota system been in operation.

Following the correction of the concours atthe end of primary schooling, the RwandanMinistry of Education selects students for sec-ondary schools, bearing in mind the need tohave regional, ethnic, and gender balance.Mauritius also uses a gender quota; the top2,000 boys and 2,000 girls on the primaryschool-leaving examination are offered placesin secondary school. Both Mauritania andEthiopia set pass marks that vary from regionto region. In Ethiopia, the selected passingraw score is transformed into a percentilerank derived from the distribution of aggre-gate scores for all candidates. Regional offic-ers are instructed to determine the passingraw scores for their regions by identifying theaggregate score that corresponds to the na-tionally selected passing percentile rank. Theeffect of this on the Grade 6 National Exami-nation in 1987 was that while the passing rawscore in Eritrea was 42, in Gojam it was ashigh as 47. Thus, for many Ethiopian stu-dents, success or failure depended on placeof residence.

One effect of the operation of quota sys-tems is that students change their place ofresidence or their names to improve theirchances of selection.

The quota system enables examinationauthorities to examine the predictive validity

of selection procedures. To do this, the sec-ondary-school performance of students whowere selecWd on the basis of their examina-tion performance is compared with the per-formance of students who would not have

50

been selected on their examination perfor-mance but gained entry to secondary school

because of quota requirements.

The social, political, and educational implica-

tions of quotas merit serious consideration and

monitoring.

6 6

6 Using Examinations to Improve Education

In considering the role of examinationsin educational reform, one should not losesight of the range of problems faing educa-tion in Africa. The extent of these problemsvaries from country to country, but all coun-tries to a greater or lesser extent, face prob-lems relating to accommodation, the rel-evance of curricula, availability of books andmaterials, and teacher training. Further-more, these problems in the educational sys-tem co-exist with, and often are caused by,economic and social problems in the widersociety. Problems such as these are not go-ing to be solved by improvements in exami-nation systems alone. However, examina-tions could play a role in the reform of edu-cational systems and in improving the qual-ity of education in schools. At the very least,good assessment procedures must be betterthan bad ones. Furthermore, despite theproblems that will arise in implementingsome reforms, such as in measuring andteaching higher-order skills and in provid-ing feedback information to schools, at-tempts to do these things are preferable tocontinuing with examination systems that,in many countries are regarded as obviouslydefective in many ways. Besides, there aresome recommendations that can be made,which can hardly be regarded as controver-sial, such as the need to include in examina-tions assessments of students' ability to writeif writing is to receive the attention it meritsin schools.

We have seen that examinations occupya key role in the educational systems of thecountries considered in this report. This ispartly because of a resource problem, in par-

ticular the inability of educational systemsto accommodate all students who wish toremain at school. In this situation, exami-nations are used as the means of allocatingplaces to students. Whatever decisions aretaken about assessment and examinationsprocedures, it has to be recognized that theneed will remain, for some time at any rate,to operate a system of selection that is per-ceived to be fair and equitable. And if therehas to be selection, it nas been argued, for along time in European countries (Curricu-lum and Examinations Board - CEB 1985)and currently in developing countries(Heyneman 1985), that examinations are amore equitable way of doing the job thanother procedures that have been tried. Evenif examinations continue to be used for se-lection, this does not mean, of course, thatefforts should not be made to improve themas instruments of certification also.

In this chapter, for the fourteen coun-tries in our study, we present the proposalswhi,:h were set out in action plans forchanges in examination systems which, itwas believed, would lead to an improve-ment in the quality of examinations and ofeducation. Following that, we provide someguidelines which, if followed in the imple-mentation of action plans, should result inthe improvement of examinations. Afterthat, we consider problems and undesirableeffects that can arise from certain kinds ofexamination reform. Finally, we outlineconditions that should be present in theadministration of examination authorities ifthese authorities are going to be in a posi-tion to support examination reform.

51

6,

Action Plans

Each of the fourteen countries preparedan action plan designed to improve the qual-ity and functioning of its examination sys-tem. A tabular summary of the plans isprovided in Table 5.1, separately for Anglo-hone, Francophone, and other countries.

The objectives of the action plans can be cat-egorized as relating to Administration, Tech-nical and Training Requirements, Equip-rne.tt, and Construction/Rehabilitation. Itshould be pointed out that the objectiveswere not stated with similar clarity in allthe reports.

Administration

Expand Examination System. The prob-lems caused by the substantial increase inenrollment figures in general, and in thenumber of candidates taking public exami-nations in particular, are reflected in theplans to increase the administrative capa-bilities of publk. .7%amination authorities.Five of the six Anglophone countries andEthiopia indicated a need to provide for theexpansion of their examination systems tocope with anticipated increases in numbers.

Localize Examinations. Three countries(Lesotho, Mauritius, Swaziland) plan to lo-calize public examinations by developing al-ternative examinations to those prepared byoverseas agencies. The challenge for thesecountries is to develop within their exami-nation systems the ability to design, print,mark, and validate examinations whilemaintaining public confidence by ensuringthat the new examinations retain the statusof external examinations such as those pres-ently offered by the Cambridge OverseasSchool Certificate Examination Syndicate.

Phase Out Examinations. In 1987, theKenya National Examination Council dis-continued the Kenyan Certificate of Educa-tion. Guinea, Cape Verde, and Ethiopia planto reduce the number of public examina-

tions which they offer. Ethiopia, for ex-ample, has indicated its intention tc elimi-nate the Grade 6 National Examination whenthe shortage of junior-secondary schoolplaces is overcome.

52

Centralize Examination Authority.Three countries (Cape Verde, Ethiopia, Mau-ritania) have indicated a desire to centralizeresponsibility for public examinations withinone agency. In this way it is hoped to makethe system more efficient through a sharingof scarce resources. In Ethiopia one centralsystem is planned as a replacement for theEthiopia School-Leaving Certificate Exami-nation Office run by the University of AddisAbaba and the National Examination Boardof the Ministry of Education which has re-sponsibility for the Grade 6 ar -I Grade 8national examinations.

Centralize Operation of Examinations.Assigning printing and packaging of exami-nation papers to an external agency increasesthe likelihood of a breach of test security.Three large countries (Kenya, Uganda, Zam-bia) propose to assign responsibility for allaspects of public examinations to oneagency. At the moment, to cite but one ex-ample, the Kenya National ExaminationCouncil could achieve substantial savings ifit were in a position to transfer the printingof examination papers to an in-house facil-ity.

Create a Teclmical Bureau/Support Unit.The need for the creation of a technicalbureau or support unit to develop tests, toprovide training, and to evaluaLe and carryout research on aspects of the examinationsystem has been identified in five of theFrancophone countries (Chad, Guinea,Madagascar, Mauritania, Togo.) In Mada-gascar, for example, it is hoped that this newunit might take responsibility for improv-ing school-based assessment techniques, de-veloping standardized tests, and carryingout specialized studies designed to informthose responsible for decision making.

6 "

Administration

Table 51OBJECTIVES LN DRAFT PLANS

Kenya Lesotho Mautitius Swaziland Uganda 7..ambia Chad Guinea

1. Expand system to dealwith increasing numb

X X x X

2. Localize all exams X X X

3. Phase out exam(s)

4. C.,,ntralize responsibilityfor exams within one agency

5. Allow agency to assumeresponsibility for all aspectsof exams (e.g. printing) X

6. Create technical bureau/support unit to develop tests,provide training/evaluation X X X

7. Improve efficiency of examsby strengtheningmanagement structures X X X X X

Technical and Training

1. Improve quality of exams(improve exam-syllabus linkage,introduce criterion-referencedtests) X X X X X X X X

2. Improve assessment, researchand development capacityof systems by improving technicaltraining/resources (e.g. library) X X X X X X X X

3. Involve teachers more inassessment as a replacement for, orin conjunction with, public exams X X X X

4. Develop capacity to providefeedback information on public-exam performance to schools X X X X

5. Provide training for MOEofficials (school inspectors)/teachers/counsellors inassessment and exams X X X X X X

Equipment

1. Improve efficiency of examsthrough provision of equipment X X X X 14. X X X

Construction/Rehabilitation

1. Extend/improve accommodationfor administration of exams X X X X

53

Table 5.1 (continued)OBJECTIVES IN DRAFT PLANS

Administration

is& AMMMNIMiNM

Madagascar Mauritania Rwanda a Togo CapeVerde Ethiopia

1. Expand system to dealwith increasing numbers X

2. Localize all exams

3. Phase out exam(s)

,

X X

4. Centralize responsibiLlyfor exams within one agency X X

5. Allow agency to assumeresponsibility for all aspectsof exams (such as . printing)

6. Create technical bureau /supportunit to develop tests, providetraining/evaluation X X X

7. Improve efficiency of exams bystrengthening management structure5 X

Technical and Training

1. Improve quality of exams(improve exam-syllabus linkage,introduce criterion-referencedtests)b X X X

2. Improve assessment, researchand development capacity ofsystems by improving technicaltraining/resources (e.g. library) X X X X X

3. Involve teachers more inassessment as a replacement for,or in conjunction with, public exams X X X

4. Develop capacity to providefeedback information on public-exam performance to schools X

5. Provide training f9r MOEofficials (school inspectors)/teachers/counsellors inassessment and exams X X X X X

Equipment

1. Improve efficiency of examsthrough provision of equipment X X X X X

Construction/Rehabilitation

1. Extend /improve accorrunodationfor administration of exams X

a/ Being finalizedb/ Recommendations to produce tests of "creativity," "intelligence," "visualization," are nut included here.

54

6.41'

Improve Management Efficiency. A to-tal of ten of the fourteen countries specifiedthe need to improve the efficiency of publicexaminations by strengthening managementstructures. In many instances, the lack ofautonomy of the examination body was citedas a factor which restricted the efficiency ofthe examination system. There is also a needfor a substantial recruitment and trainingprogram to equip management personnelwith relevant skills and knowledge. Salaryscales should be sufficiently attractive to dis-courage skilled personnel, such as in the areaof computing, from seeking employmentelsewhere.

Technical and Training Requirements

Improve Examination Quality. Consid-erable agreement is to be found in the ac-tion plans regarding the need to improvethe quality of public examinations. Each ofth( fourteen countries specified that itneeded technical assistance in this area.Many recognized the need to improve thevalidity of their public examinations, inparticular to improve the link between thecontent of examination questions and thesyllabus. Psychometric support was alsocalled for to help, among other tasks, todevelop item banks and construct adequateassessment instruments. A number of ex-amination bodies recognized the need to in-troduce new modes of examining, involv-ing oral, aural, and practical tasks.

Improve Research and DevelopmentCapacity. The need to introduce or improvethe capacity of the examination system toconduct research directly related to publicexaminations was recognized in each of thefourteen countries. In Swaziland for ex-ample, it was proposed that this researchwould include validity studies, monitoringthe implementation of a proposed continu-ous-assessment system, and the effects ofchanges in the system on the internal andexternal efficiency of the educational system(dropout, repetition, levels of achievement.)

55

In most instances, it is envisaged that for-eign training will be required to enable per-sonnel to acquire the appropriate researchskills. In addition to training, research, anddevelopment, personnel will need to haveaccess to libraries equipped with appropri-ate technical books, journals, and sample as-sessment instruments.

Involve Teachers in Assessment.Slightly over half the countries identified theneed to increase teachers' involvement in theassessment process. In some instances it isenvisaged that assessment by teachers wouldeventually replace some external examina-tions. Most systems, however, favor the in-volvement of teachers in providing part ofthe overall assessment in a particular sub-ject examination. For Mauritius for example,it is suggested that, following the trainingof teachers in assessment techniques, theCertificate of Primary Education wouldshow the teacher's grade in addition to theexternal examination grade for each subject.

Develop Feedback Capacity. Four An-glophone countries (Kenya, Lesotho, Swazi-land, Uganda) and Ethiopia stressed theneed to develop the capacity of the public-examination system to provide feedback in-formation to schools on aspects of students'performance on examinations. It is believedthat such information could contribute to thequality of teaching and learning in theschools by identifying aspects of the curricu-lum that appear to be misunderstood or sim-ply ignored by examination candidates.Feedback information can also be useful toeducational authorities by identifying re-gional disparities in achievement and aspectsof the curriculum where candidates farepoorly, possibly due to lack of appropriateequipment, for example in physics andchemistry.

Provide Training for Inspectors andTeachers. Virtually all countries recognizethe need for the training of Ministry of Edu-cation officials, school inspectors, and teach-

ers in assessment terminology and proce-dures. In-service training seems particularlynecessary to familiarize teachers with thetechniques required to conduct school-basedassessment. Zambia's proposal envisagesthat as many as 10,000 secondary-schoolteachers would attend a ten-day in-servicetraining course over a four-year period. CapeVerde authorities propose that a foreign ex-pert should train a cadre of people who inturn would become trainers of others.

Equipment

Ail action plans indicate a need for equip-ment. Specified needs relate mostly to theproduction of examination papers, such asprinting equipment, and the processing ofexaminations. Data-processing equipment isrequired for scoring multiple-choice exami-nations, the preparation of results, and forstoring of assessment information. Main-frame and micro-computers have been re-quested. Books and journals on testing andmeasurement have also been listed. Otherequipment that is frequently requested in-cludes typewriters, photocopiers, and paper.A number of countries have asked for fur-niture, (Madagascar and Togo), and motorvehicles, (Lesotho and Mauritania).

Construction/Rehabilitation

Four Anglophone countries, three Fran-cophone countries, Cape Verde, and Ethio-pia include specific requests for constructionand/or rehabilitation of buildings. Kenyaand Uganda require the construction of largebuildings to house printing equipment. TheUgandan proposal outlines the need for abuilding to house, in addition to printingequipment, a computer facility as well asrooms for desktop publishing. Other coun-tries, for example Togo and Chad, requirethe construction of buildings to accommo-date all examination board personnel. Le-

56

sotho and Madagascar have indicated thattheir examinafion offices are badly in needof repair and improvement.

Cross-National Cooperation

A number of countries referred to thebenefits of cross-national cooperation in ex-aminations but did not actually include pro-posals for such cooperation in their actionplan. Two possible approaches to cross-na-tional cooperation are outlined here.

First, associations of examining bodiesmight be formed to enable those involved inexaminations to exchange information andexperiences and to provide short-term train-ing courses. Such an association (The Asso-ciation of Heads of Institutions Responsiblefor Examinations and Educational Assess-ment in East and Southern Africa) alreadyexists in the Anglophone areas of eastern andsouthern Africa. This association, however,is in need of financial support. There is alsoa need to cater for countries in non-Anglo-phone areas.

The second proposal for improving cross-national cooperation relates to ti e establish-ment in a university of research facilities andpost-graduate courses relating to measure-ment, assessment, research, and evaluation.It would probably be necessary to make thisprovision in both Anglophone and Franco-phone areas.

Before this step is taken, they will be aneed, as indicated in individual country ac-tion plans, to have individuals go to univer-sities in Europe or North America for post-graduate studies. Consideration should begiven to having individuals from severalcountries attend the same institution as thiscould help establish relationships whichwould form a useful basis for future co-op-eration when the individuals return to theirown countries.

Improving Assessment in PublicExaminations

A number of guidelines should be fol-lowed to improve assessment in the public-examination system. Before presenting themtwo general points about examination reformthat have already been made might be re-peated. First, such reform has to be consid-eree :n the general context of the educationalsystem, not as an isolated element in it.Changes in the assessment system are likelyto depend on, and to affect other aspects ofthe educational system. The dependence ofexamination reform on other aspects of aneducational system is perhaps most obviouswhen one considers the resources that willbe required to implement reform. On theother hand, the effects of reform are likely toexhibit themselves in the quality of studentsthat are selected for further education, as wellas in the character of curricula offered inschools. Second, in considering examinationreform, one has to be conscious of the factthat there is no perfect system of assessmentand that there is no point in looking for one.Any system will have advantages and dis-advantages. This will be particularly obvi-ous when an assessment system is used toserve more than one purpose. It is the taskof individual countries, in the light of theirtraditions, resources, values, and needs todecide what balance of options best suitsthem. Since it is not usually possible to an-ticipate precisely the advantages and disad-vantages of any system in advance, it is im-portant that the effects of examination re-form, particularly on curricula and teachingpractice, should be monitored. With thesepoints in mind, we may now turn to morespecific guidelines for the improvement ofassessment in the context of public examina-tions.

First, assessment should be based on theofficial curriculum. However, since noteverything included in a curriculum can beassessed, the areas which are selected willbe regarded by teachers and studeAts as very

important and will receive special attention.Hence, the areas selected for inclusion inexaminations should cover importantcurricular topics.

Second, insofar as it is possible, themodes of assessment (written, practical, andoral) which are used should be diverse andshould reflect the goals of curricula. Whatis tested is what is likely to be taught andwhat is not tested is likely to be ignored.Hence, if examinations do not require stu-dents to write continuous prose, the skillsinvolved in this activity will probably notreceive much attention in schools. Similarly,examinations that do Dot include oral orpractical components are likely to lead to aneglect of skills in these areas. However, costis a serious problem if external examinersare used, particularly in the case of oral andpractical skills. For this reason, the possibil-ity of using assessments by teachers withinthe school is worth exploring. If done prop-erly, such assessments are likely to be lessexpensive than external ones and are alsolikely to be more valid. However, some sys-tem of moderation may be required ifteacher assessments are incorporated into thepublic-examination system. In this case, itshould be borne in mind that moderationcan be cumbersome and expensive and, insome circumstances, quite impractical.

Third, assessment instruments shouldpay adequate attention to the measurementof higher-order thinking skills. Many tests,and these are the easier ones to construct,require little more than the recall or recog-nition of information. However, if it isagreed that "there is a need to make deeplearning a central goal of education, and tofoster development of this goal through theevaluation of students," then it will be nec-essary to "...place emphasis on umkrstand-ing, transfer of learning to untaught Prob-lems or situations, and other thinking skills,evaluating the development of these `; killsthrough tasks that clearly must involve morethan recognition or recall" (Crooks 1988). It

57

6",

is assumed in this view that the inclusionof tasks requiring higher-order skills in testswill lead to greater emphasis on teachingthese skills in classrooms. There is in factsome evidence from a study carried out byOgundare (1988) in Nigeria, in which pri-mary-school students were taught socialstudies using a problem-approach method.This suggests that when this emphasis ispresent, not only do students acquire morefacts, they also comprehend material betterand are better able to apply their knowl-edge to the solution of new problems andto evaluation activities.

Various taxonomies have been designedto help test constructors elicit higher-orderthinking skills in students. Some of thesecontain as many as six levels (such as Bloom1956, Perrott 1982). However the construc-tion of a test which distinguishes clearly be-tween that many levels might not be pos-sible. Problems arise particularly at higherlevels designed to measure synthesis andevaluation. For example, in Britain, thespecification of achievement levels withinsubject areas where higher-order or generalskills were involved has proved very diffi-cult in the context of developing a systemof assessment for the new national curricu-lum (Nuttall 1990). Because of this, it isperhaps better in test construciion to thinkin terms of a simpler taxonomy containingtwo levels rather than six. The first levelcould cover what might be called "minimal"objectives, which most students might beexpected to achieve, and would relate tothe acquisition of knowledge and basic skills(recall and comprehension in Bloom's tax-onomy). The second level would involvehigher-orc,er outcomes, defined in terms ofmore complex cognitive processes. Whiletests of minimal objectives would examinestudents' ability to deal with familiar con-cepts and rehearsed skills, the higher-orderones would examine students' ability to

58

apply learning to new material or situations(Gronlund 1973; Terwilliger 1989).

Even with a simplified taxonomy, it isimpossible to be sure that the taxonomiclevel of an item will evoke the targeted tax-onomic level in the candidate answering theitem. Whether or not it does will dependon the learning experiences of the candidate(Torrance 1986). Bloom (1956) was awareof this when he observed that if a problemdesigned to measure application is too sim-ilar to problems used in class, it will mea-sure comprehension rather than application.

A fourth guideline to be followed in theconstruction of examinations is that exami-nations should test students' ability to ap-ply their knowledge and skills not just inscholastic contexts but in situations outsidethe school as well. Thus the examinationsshould take into account the everyday lifeconditions of students and the kinds of prob-lem they are going to meet at home, on thefarm, in the shop, or in the market place.

The Kenya Primary School Certificate Ex-amination provides examples of items whichrelate to situations outside the school. Thefollowing are examples of such items:

Diarrhoea is a disease that kills manybabies in Kenya. When babies havediarrhoea they lose a lot of water andfoods. One correct way to treat ba-bies with diarrhoea is to:

A. keep them wrapped up and warmso that they sweat out the sickness

B. give them drinks of boiled coldwater containing some sugar anda little salt

C .give them solid foods containing plenty of carbohydrates

D.give them very little food orwater until the diarrhoea stops.

fi

Wangari lives in a village with manypeople. From which one of the fol-lowing sources can Wangari collectthe best drinking water?

A. a big river nearbyB. the rain from her mabati roofC. the dam near the villageD. a swamp on her farm.

A fifth guideline recognizes that exami-nation scores used for selection should dif-ferentiate between candidates on the basisof characteristics relevant to the opportuni-ties being offered. Thus, there is a need tocarry out studies of the predictive validityof selection instruments.

The fact that examination scores are im-perfect measures of students' achievementsis the basis of a sixth guideline. It is desir-able that the past performance and relevantexperience of students be taken into accountin making important decisions about stu-dents. This guideline points to the need todevelop the capacity of teachers and schoolsto carry out school-based assessment thatwill be acceptable outside the classroom.

A seventh guideline takes up the recom-mendation of several commentators that de-tailed feedback information on the perfor-mance of students in examinations shouldbe provided to schools and teachers. For ex-ample, Lewin (1984) has pointed out thattelling a teacher that 20 percent of a classhas failed does not communicate the causesof either the teacher's or the students' fail-ure. However, if teachers were told whichquestions and tasks students had completedsatisfactorily and which ones they had not,the information could be used as a basis forremedial action. The information would beeven more useful if accompanied by an in-terpretation of student difficulties as well as

by guidance on the design and implementa-tion of possible remedial actions.

The most serious efforts to provide feed-back information to schools have been madeby the Kenya National Examinations Coun-cil. While the Newsletter of the Council iscommended by Lewin (1984) as providing"excellent examples" of the kind of feedbackthat is likely to be helpful to teachers, it doesnot in fact provide teachers with the kind ofinformation recommended by Lewin. Sincethe information in the Newsletter is limitedto national statistics about performance,there is no way a teacher can ascertain howhis or her class performed on particularaspects of the examination. Thus, the basisfor remedial action is limited. While theprovision of information on examinationperformance on a school-by-school basiswould be necessary to meet Lewin's condi-tions, the logistical problems to which thiswould give rise are formidable.

59

The provision of detailed information onthe performance of students in the hope thatteachers will direct their efforts towards thetopics and skills covered in examinations andthereby raise students' test performanceraises a broader issue which merits attention.It is not surprising that research evidence in-dicates that the test performance of studentswill improve if they have been taught thespecific material covered by a test and if theyhave been frecrantly exposed to the test for-mat (Cooley and Leinhardt 1980; Le Mahieu1984). However, are we justified in conclud-ing on the basis of improved student test per-formance that students' actual achievement(in terms, for example, of problem-solvingskills), as di ctinct from their test performance,has also improved? The answer would seemto be no, at least until we have satisfied our-selves that some of the variance in test scoresis not attributable to the form of the test used(method of measurement) rather than to theindividual characteristics (achievement) that

the test is designed to measure (see Bolgeran°. Kellaghan 1990; Campbell and Fiske1959). Students are taught in such a waythat the match between irstructional pro-cesses and test items is very close, then in-ferences about students' actual skills andknowledge become very difficult to make.The problem is particularly acute if we wantto test whether a student can apply skills andknowledge to solve new problems, since inthis case, the problems must be new to thestudent and not ones that were taught inclass. Practice on a problem may improvestudents' test scores; however, not only willit eliminate the possibility of drawing un-ambiguous conclusions about students' abil-ity to solve problems, it may actually havethe effect of interfering with the developmentof understanding by students (Linn 1983a,b).

This situation creates a dilemma forwhich there is no obvious solution. To whatextent should one improve the match be-tween testing and instruction, not only topromote appropriate curriculum coveragebut also to ensure that what is tested is whatis taught, if, at the same time, the at,',ity tomeasure application and problem-solvingskills is going to be compromised? On theothlr hand, should one include new prob-lems and types of problems on examinationsw' , the objective of assessing higher-orderskiiis, knowing that this will create the riskof failing to measure what is taught? (Linn1983b).

An eighth guideline relates to the assess-ment of the impact of examination reformon student achievement. On the basis of theprevious discussion, which implies a distinc-tion between test performance and achieve-ment, it should be clear that the use of testinstruments, which are part of an instruc-tional treatment, to monitor the effects of thattreatment, as in Somerset's (1987) Kenyanstudy, cannot be regarded as satisfactory.The problem in this approach cuncerns con-

60

struct validity (LeMahieu 1984). If students'examination scores are used to assess the ef-fect of providing information to schoolsabout students' test performance, the sort oflearning that is revealed in the examinationscore may be made up largely of skills thatare more properly considered test-takingskills than achievement as indicated by stu-dents' ability to use higher-order skills. In-deed, as noted above, the more effective thetreatment is in terms of having teachers teachto the examination, the greater will be thematch between instructional processes andtest items, and, as a consequence, the morelimited will be our ability to draw inferencesabout students' actual skills and knowledgeas distinct from test-performance skills. Thisguideline then specifies the need to use al-ternative methods of measuring achievementwhen assessing the effects of an examinationreform which involves providing detailed in-formation on students' performance toschools.

The final guideline, concerns the need todevelop effective assessment strategies, dis-tinct from public examinations, to assess theperformance of the educational system andof individual institutions. One type of strat-egy will probably not meet the requirementsof these two purposes. A procedure involv-ing testing in a sample of schools may beadequate for the assessment of the perfor-mance of the educational system, while amore school-specific approach, which might,for example, be incorporated into the func-fions of the school inspectorate, will be re-quired in the assessment of individualschools.

Anticipating Problems andUndesirable Effects

In considering proposals to improve as-sessment procedures and, in particular, todesign examinations which will have a ben-eficial effect on curricula and studentachievement, it is important not to

70

underestimate the problems involved in thetask ',see Airasian 1988). It is also importantto guard against possible unintended conse-quences that may arise.

The first problem relates to the task ofspecifying desired outcomes of instructionon which curriculz and examinations mightbe based. The task involves important valuejudgments; hence it may not be easy to reachagreement on what the outcomes of school-ing should be. Besides, such outcomesvary, even within a single country, accord-ing to the different circumstances in whichschools operate and in which their studentslive.

Of interest in this context are proposalsto make education more relevant to theneeds of rural children (the majority of chil-dren in Africa). Such proposals have beenmade repeatedly in the history of Africaneducation (se- far example, Gifford andWeiskel 1971. Jones 1922, 1924). On thewhole, they failed to gain acceptance, prob-ably largely because of parents' antipathyto them. Since parents saw traditional aca-demic education as the route to occupationswith the greatest prestige and reward, afailure to provide children with such educa-tion was seen as limiting their occupationalopportunities (see Blaug 1973; Bray, Clarke,and Stephens 1986; Foster 1965; Keilaghan1991). It remains to be seen whether thecurrent emphasis being placed on local rel-evance will be more acceptable than pastefforts (see Lockheed and Verspoor 1990).It may be, as the number f children attend-ing school increases, that the prospects ofthe more desirable occupations to whichparents aspired for their children in the pastwill be seen to be less realistic. Much toomay depend on the degree of localizationthat is introduced into curricula. If curriculafor children in rural areas are limited toteaching basic principles relating to healthand agriculture and do not deviate from thetraditional curriculum to the extent that ruralchildren would be at a disadvantage in scho-

lastic competition, they may be more accept-able now than they were in the past. Fur-ther, the idea of relevance in curricula canbe defended without reference to the futurelives of children. Tasks that are related to achild's previous experience and are seen tohave relevance to everyday life are morelikely to motivate children to learn than aremore abstract tasks.

A second problem in the design of ex-aminations relates to the complexities in-volved in teaching stated objectives. Themere positing of goals, or even their incor-poration into examinations, does not guar-antee their achievement. Indeed, as we haveseen above, it may inhibit them. The prob-lem is likely to be most acute if the exami-nations, as test designers are increasinglystriving to do, place emphasis on measur-ing higher-order behaviors rather than, orin addition to, general minimum lower-levelbehaviors. The earlier discussion alluded tothe problem of measuring higher-order be-haviors. Evidence is also available to indi-cate that teaching such behaviors is differ-ent in many ways from teaching lower-levelrote behaviors. For example, higher-levelbehaviors take longer to teach, developgradually over time, are less amenable todirect instructional approaches, are often dif-ficult to locate in a curriculum, and may betoo diffuse to drive instruction (Airasian1988).

61

Faced with such difficulties, teachers mayresort to teaching students test-taking strat-egies which may undermine the objectivesof the examination (Madaus 1988; Madausand Kellaghan, forthcoming). Over time,teachers become familiar with the formal re-quirements of the examination and preparestudents to deal with these rather than withcurricular objectives. Strong evidence of thishas been observed in a number of Africancountries, and the evidence is not confinedto Africa. Selection-type questions, whenused in examinations, become a particularfocus of attention. Not only are students

taught how to cope with such questions, butnormal teaching may be presented in theform of statements accompanied by a rangeof options from which students are requiredto choose, (for example, The capital of Franceis: Brussels, Paris, Lyon, Toulouse). Thisform of teaching may be used, not just inthe examination class, but can filter rightdown the school, as has been observed inreception classes in African schools.

A further po3sible undesirable effect ofpublic examinations, particularly if the re-sults of schools are publicized, is that schoolsmay require pupils to repeat grades, eitherto take more time to preparn for the exami-nation or to avoid taking it at all (Madausand Greaney 1985). In the case of Kenya,Oloo (1990) states, though he does not pro-vide supporting data, that "it is fairly obvi-ous that the schools which perform best inKCPE are those which enter a few selectedcandidates." Further, pupils, as a result ofacademic pressures, may leave school at anearly stage. Grade repetition and drop-outare, in the United States at any rate, inter-related (Catterall 1989) and are probably theresult of a variety of iactors, which includefamily and student characteristics, low per-ceived benefits from education, and schoolineffectiveness (Lockheed and Verspoor1990; Swaziland, Ministry of Education1986). Care should be taken to ensure thatthe pressures to perform well on public ex-aminations do not lead to an exacerbationof problems of repetition and inefficiency inschools.

Finally, as has already pointed out inChapter 4, the ranking of schools in termsof examination success, especially when noaccount is taken of the entry characteristicsof pupils, their home circumstances, or thecharactefistics of schools, can be unfair andcan have disastrous effects on the morale ofschools.

Conditions to Improve the Administrationof Examinations

In this section, some conditions relatingto the administration of examinations an..1examination authorities are outlined. Theseare necessary if the efforts to develop publicexaminations which will have a positiveeffect on school curricula and studentachievement are to be effective.

Examination Authority

62

In the first place, the body entrusted withthe construction and administration of theexaminations must be adequately financedand staffed. The body should have sufficientresources and competencies to develop ex-amination instruments, to analyze examina-tion performance, and to carry out researchon examinations. Resources should includeaccess to foreign exchange to allow for pur-chase, maintenance and repair of ,-quipment.At present, most examination bodies in Af-rica have insufficient resources and are hardpressed just to administer the system, leav-ing no time for reflection or research.

Salary levels should be suffidently attrac-tive to ensure that skilled personnel are notforced to seek employment elsewhere. Fre-quently in the past, personnel with high-level skills have been attracted to other jobs,particularly in private industry, where sala-ries are higher.

In some countries, making examinationslargely self-financing through examinationfees has helped provide financial resourcesto enable the authority to operate u-ith somedegree of independence.

Whatever the location of an examinationsbody, it is crucial that it should be indepen-dent in the exercise of its functions. In some

countries the examination body is locatedin a Ministry of Education; in other coun-tries, it is a separate body, though workingin close co-operation with the Ministry ofEducation. Elsewhere, the examination bodyis dosely associated with a university; withinsome countries, in fact, universities and otherauthorities administer different public ex-aminations. There are factors in favor andagainst each arrangement. The decision thatis made in any individual country about thelocation and control of the examination bodywill probably be largely determined by thecountry's tradition and experience.

National Cooperation

Examinations represent one part of theeducational process. Structures should be de-veloped to allow the examinations author-ity to work in close liaison with bodies re-sponsible for educationai administration andcurricula. Cooperation can help ensure thatthe public examinations serve the needs ofall students in the system, improving thegeneral quality of education and satisfacto-rily serving the functions of certification andselection. It can also help to ensure thatexaminations adequately reflect curriculaand not only that, but that they are seen todo so. Public examinations that reflect na-tional curricular priorities can serve as apowerful positive change agent in the edu-cational system.

The examinations body should alsowork closely with teachers. Most examina-tion bodies make use of teachers to markexamination scripts. This can be a good wayof letting teachers know what is importantin examinations. Obviously, it is importantto have as many teachers as possible in-volved in this exercise so that first-hand in-formation about examinations is generallyavailable in the educational system.

63

Teachers can also be involved in con-structing examination questions. For ex-ample, teachers might contribute items to anational item bank, from which examina-tion questions might be selected. Apart frominvolving teachers directly in the examina-tion process, this practice might also serveto integrate public examinations with theinternal assessment procedures of schools.Teachers would come to know what thingsare assessed in public examinations andwould follow similar practices in their ownin-school assessments. At the same time,teachers would have an input into the ex-amination system ensuring that what isasked in the public examinations is relevantand appropriate to their students.

Communication between examinationboards and teachers could also be improvedconsiderably through the introduction ofnewsletters which would include detailedresults of pupil performance and how to dealwith problems.

Other forms of communication from ex-amination boards include providing infor-mation about standards through the schoolinspectorate and selected teachers, the pro-vision of in-service courses, and respondingto individual queries from teachers and otheragencies. Pre-service teacher training shouldinclude courses on assessment. Participantsin these courses should be introduced in apractical way to a variety of assessment pro-cedures.

International Cooperation

In considering the reform of examina-tions, the need for cooperation and sharingof scarce resources should he recognized. Ex-changes of item writers, markers, and othertechnical personnel, as 's commonly carriedout in a small number of countries, should

be fostered. Security threats and proceduresshould be reviewed. While countries are notlikely to administer common public exami-nations, international cooperative studiesmight be conducted on common curricularareas. Perhaps the main advantage of suchcomparative studies is the focus it couldbring to bear on existing curricular goals andobjectives in individual countries. Regionalorganizations such as the recently foundedAssociation of Heads of Institutions Respon-sible for Examinations and Educational As-sessment in East and Southern Africa shouldbe encouraged and supported. Such orga-nizations can provide a structure withinwhich regional and international coopera-tion can be promoted.

Number and Format of Public Examinations

Consideration should be given to reduc-ing the number of public examinations ad-ministered. Economies incurred should bedevoed to improving the quality of the pri-mary leaving and the senior-cycle second-

64

ary leaving-certificate examinations. Im-proving the quality of public examinationswill of necessity require examination au-thorities to change existing modes of assess-ment. In particular, examination officialsshould be requested to improve the mea-surement of higher-order cognitive skills,and skills related to everyday life, and todevelop techniques of oral, aural, and prac-tical assessment.

Formal Training

Well developed, advanced-degreecourses in measurement and assessmentshould be offered by at least one universityin an Anglophone area and one in a Franco-phone area. In addition to recruiting stu-dents in the normal manner, special leavearrangements should be made to allowmembers of national examination authori-ties to enrol. Courses offered should coveritem construction, psychometric theory, re-search design and methodology, curriculumtheory and design, computer applications,and management.

7 Conclusion

In considering the reform of public ex-aminations, it has to be recognized that someof the problems associated with examina-tions are not due to the format or adminis-tration of the examinations per se, but tostructural aspects of the educational systems.For example, the severe shortage of suffi-cient secondary-school places would, irre-spective of the quality of the examinationused for selection purposes, inevitably havea strong limiting influence on teaching styleand the nature of student learning towardsthe end of primary school. Thus, in the finalanalysis, it needs to be stressed that the roleof public examinations in contributing to im-provements in achievement levels will betempered by a range of factors over whichexamination authorities have little or no con-trol. Unless governments and ministries ofeducation tackle the problems created by alack of places in schools, poor quality of in-struction, inadequate supplies of books andequipment, and the inability of students tounderstand the language of instruction, re-forms in public examinations, no matter howfar-reaching, are unlikely to be very effec-tive.

While factors other than examinations af-fect school quality, they also place con-straints on the kind and extent of examina-tion reform that is possible. Thus, those con-cerned with proposing changes in examina-tions have to take into account the contextin which the changes will be implemented.In particular, the conditions which operatein schools, which vary considerably acrossthe countries considered in this report, andwhich include school facilities, the qualityand availability of text books, number ofunqualified teachers, class size, and provi-

65

sion for school supervision, will limit theoptions that are available to the examina-tion reformer.

Having said that, it should be also rec-ognize that unreformed examination systemsare likely to cause damage to the quality ofeducation offered in schools. It obviouslymakes sense to take steps to change suchsystems and to develop examinations whichare more in keeping with principles of goodassessment practice. Well designed and re-sponsibly used assessment can be an impor-tant source of information about what andhow well students are learning and whichones are most likely to make the most ofopportunities that cannot be provided forall (see NCTPP 1990.)

Most countries, to varying degrees, havein recent years actually implementedchanges in their examination systems. Thesechanges relate to the techniques of assess-ment, to a broadening of the scope of as-sessment and, in turn and related to this, ashift in responsibility for assessment.

In Anglophone countries, the mainchange in techniques relates to the introduc-tion of multiple-choice tests to replace free-response ones.

But there have been other changes. Insome countries, the scope of the examina-tions has been broadened considerably bycombining school-based assessments withthe result:, of a terminal examination. Thishas increased the involvement of teachersin the assessment process, though thesekinds of assessment are accompanied bystrict monitoring procedures.

At one level, changes represent an at-tempt to improve efficiency by applying newtechnology. For example, the use of ma-chine-scored multiple-choice items ratherthan person-scored essays helps to reducescoring costs, as non-multiple choice exami-nations are between two and five times moreexpensive to design and grade than mul-tiple-choice examinations (Heyneman 1986.)The changes may, however, represent morefundamental shifts in the control and role ofexaminations. The use of coursework as-sessment, for example, involves a shift inthe control of examinations and of the selec-tion process, which is based on it. Some ofthe other changes may herald a shift in em-phasis in the role of examinations from oneof selection to one of certification, as well asan attempt to replace assessment-led cur-ricula by curricula-led assessment. Attemptsto provide a more adequate sampling of cur-ricula in examinations suggest, in additionto a concern for curricular validity, a con-cern with the performance of all childrenrather than just the ones that are likely to beselected to remain in the educational sys-tem and an effort to restructure testing sothat talent is promoted rather than merelyscreened or classified.

These comments suggest that importantvalues may underlie changes in examina-tions. Consideration should be given tothese values and how they fit with acountry's aims and priorities for its educa-tional system. Insofar as any country canachieve consensus in its aims for education,it would seem important that, in consider-ing examination reform, an attempt shouldbe made to reach some level of agreementand establish some set of priorities in val-ues pertaining to education. For example,is the function of the educational system pri-marily to selea a small cohort of studentsfor higher levels of education or is it to pro-vide basic skills for the total population? Or

66

is it to do both, and if it is, what kind ofbalance should be struck between the dif-ferent functions? The rhetoric in most coun-tries would suggest that the provision ofbasic skills for the total population is moreimportant; present examination practicewould often seem to suggest that selectionis more important.

These considerations should not be taker,to imply that an ideal assessment systemwill neatly fall into place once resourceshave been assessed and the values and aimsof an educational system have been agreed,insofar as these can be agreed. As has beennoted, no examination system is without itsdefects. The design of a system will inevi-tably involve a series of trade-offs so thatfinal decisions will represent a set of provi-sional compromises among competing values (Ecl, .ein and Noah 1988; Noah andEckstein 1990.) Old and new values willmost likely continue to compete. For ex-ample, even if greater emphasis is placed inpolicy on the education of all children, it isunlikely that the use of examinations forpurposes of selection, even at a relativelyearly stage in the educational careers ofchildren, can be abandoned.

Compromise too will be forced on coun-tries by lack of resources. Thus, the guide-lines set out in this report to improve as-sessment in public examinations are meantto be just thatdirections in which to travelrather than objectives to be achieved in theimmediate future. It is our belief, that if re-forms are instituted which are pointed inthe directions indicated in the guidelines, ifthe necessary administrative steps, set outalso in the report, are taken, and if prob-lems and undesirable effects are anticipatedand steps taken to deal with them, then thequality of examinations and of educationalsystems in general will improve in Africa.

7,,

8 References

Airasian, P. W. (1988). Measurement driveninstruction: A closer look. EducationalMeasurement: Issues and Practice, 7, 6-11.

Ali, A., and Akubue, A. (1988). Nigerian pri.-mary schools' compliance with Nigerianational policy on education. An evalua-tion of continuous assessment practices.Evaluation Review, 12, 625-637.

Asmerson, K., and others (1989).Discontinuations of students from institu-tions of higher learning in Ethiopia. Unpub-lished document. Addis Ababa: Univer-sity of Addis Ababa.

Blaug, M. (1973). Education and the employ-ment problem in developing countries.Geneva: International Labour Office.

Bloom, B. S. (Ed.) (1956). Taxonomy of educa-tional objectives. The classification of educa-tional goals. Handbook1. Cognitive domain.London: Longman Green.

Bolger, N., and Kellaghan, T. (1990). Methodof measurement and gender differencesin scholastic achievement. Journal of Educa-ional Measurement, 27, 165-174.

Bray, M., Clarke, P. B., and Stephens, D. (1986).Education and society in Africa. London:Edward Arnold.

Brooke, N., and Oxenham, J. (1984). Theinfluence of certification and selection onteaching and learning. In J. Oxenham(Ed.), Education versus qualifications? Astudy of relationships between education,selection for employment and the productivityof labour. London: Allen and Unwin.

Cambridge Educational Consultants. (1988).The Ethiopian School Leaving Examination:A strategy for imp ovement. Report for theCommission for Higher Education, AddisAbaba. Cambridge,

67

7

Campbell, D. T., and Fiske, D. (1959). Conver-gent and discriminant validation by themultitrait-multimethod matrix. Psycho-logical Bulletin, 56, 81-105.

Catterall, J. S. (1989). Standards and schooldropouts: A national study of tests re-quired for high school graduation.American Journal of Education, 98, 1-34.

CEB (Curriculum and Examinations Board).(1985). Assessment and certification. Aconsultative document. Dublin.

Christie, T., and Forrest, G. M. (1981). Defin-ing public examination standards. London:Macmillan Education.

Cooley, W.W., and Leinhardt, G. (1980).The instructional dimension study.Educational Evaluation and Policy Analy-sis. 2, 7-25

Cronbach, L. J. (1971). Test validation. In R.L. Thorndike (Ed.), Educational measure-ment (2nd ed.). Washington DC: Ameri-can Council on Education.

Crooks, T. J. (1988). The impact of classroomevaluation practices on students. Reviewof Educational Research, 58, 438-481.

Eckstein, M. A., and Noah, H. J. (1988). Theform and functions of secondary schoolleaving examinations: Internationalvariations. Paper presented at 32nd An-nual Conference of Comparative and In-ternational Education Society, Atlanta,Georgia.

Eisemon, T. 0. (1988). Benefitting from basiceducation: School quality and functionalliteracy in Kenya. Comparative and Interna-tional Education Series, 2. Oxford:Pergamon.

Eisemon, T. O. (1990). Examination policiesto strengthen primary schooling in Afri-

can countries. International Journal of Edu-cational Development, 10, 69-82.

Eisemon, T. 0., Patel, V. L., and Abagi, J.(1987). Read these instructions carefully:Examination reform and improving healtheducation in Kenya. International Journalof Educational Development, 8, 55-66.

ERGESE (Evaluative Research of the GeneralEducation System in Ethiopia, Secretariate).(1986). A quality study: Summary reportpresented to the Executive Committee ofERGESE. Addis Ababa.

Fafunwa, A. B. (1974). History of education inNigeria. London: Allen and Unwin.

Foster, P. (1965). The vocational school fal-lacy in development planning. In C. A.Anderson and M. J. Bowman (Eds.), Educa-tion and economic development, Chicago:Aldine.

Fredericksen, N. (1984). The real test bias:Influence of testing on teaching and learn-ing. American Psychologist, 39, 19'3-202.

Fuller, B. (1986). Is primary school qualityeroding in the third world? ComparativeEducation Review, 30, 491-507.

Fuller, B. (1987). What school factors raiseachievement in the third world? Review ofEducational Research, 57, 255-92.

Fuller, B., and Heyneman, S. P. (1989). Thirdworld school quality. Current collapse,future potential. Educational Researcher,18, 12-19.

Gifford, P., and Weiskel, T. C. (1971). Africaneducation in a colonial context : Frenchand British styles. In P. Gifford and W. R.Louis (Eds.), France and Britain in Africa.New Haven, CN: Yale University Press.

Greaney, V. (1980). Developing nationalattainment measures: A study of Irishexperience. International Review of Educa-tion, 26, 3-16.

Gronlund, N. E. (1973). Preparing criterion-referenced tests for classroom instruction.New York: Macmillan.

Guskey, T. R., and Kifer, E. W. (1990). Rank-ing schonl districts on the basis of statewide test :esults: Is it meaningful or mis-leading? Educational Measurement Issuesand Practice, 9, 11-16.

Hallak, J. (1990). Investing in the future: Set-ting educational priorities in the developingworld. Oxford: Pergamon.

Hawes, H. (1979). Curriculum and reality inAfrican primary schools. Harlow, Essex:Longman.

Heyneman, S. P. (1985). Investing in educa-tion: A quarter century of World Bank expe-rience. Seminar Paper No.30. Washing-ton, DC: World Bank.

Heyneman, S. P. (1986). Testing to maximizenational economic performance. In S.Heyneman and I. Fagerlind (Eds.), Maxi-mizing national economic performance: Therole of educational testing. Washington, DC:World Bank.

Heyneman, S. P. (1987). Uses of examina-tions in developing countries: Selection,research, and education sector manage-ment. International Journal of EducationalDevelopment, 7, 251-263.

Heyneman, S. P. (1988). Improving univer-sity selection, educational research, andeducational management in developingcountries: The role of examinations andstandardized testing. In S. P. Heynemanand I. Fagerlind (Eds.), University examina-tions and standardized testing. Washington,DC: World Bank.

Heyneman, S. P., and Loxley, W. (1983). Theeffect of primary school quality on achieve-ment across twenty-nine high and low-income countries. American Journal ofSociology, 88, 1162-94.

Heyneman, S. P., and Ransom, A. W. (1990).Using examinations and testing toimprove educational quality. EducationalPolicy, 4, 177-192.

Heyneman, S. P., and White, D. S. (Eds.).(1986). The quality of education and economicdevelopment. Washington, DC: WorldBank.

ILO. International Labour Office. (1972).Employment, incomes, and equality: A strat-egy for increasing productive employment inKenya. Geneva.

Irvine, S. H. (1968). A five-year fellow-up ofsecondary school selection procedures incentral Africa, 1962-1967. British Journal of

68

Educational Psychology, 38, 208-211.Jones, T. J. (1922). Education in Africa. New

York: Phelps-Stokes Fund.Jones, T. J. (1924). Education in East Africa.

New York: Phelps-Stokes Fund.Joywathsing, M., and others. (1988). The

private costs of education in Mauritius.Reduit: University of Mauritius.

Kahn, M. J. (1990). Some questions concern-ing the standards of external examina-tions. Studies in Educational Evaluation,16,513-527.

Kellaghan, T. (1961). Some implications ofbilingualism for education in Nigeria.lbadan,11, 31-33.

Kellaghan, T. (1968). Abstraction and catego-rization in African children. InternationalJournal of Psychology, 3, 115-120.

Kellaghan, T. (1990). Examination systems inAfrica. In P. Broadfoot, R. Murphy, andH. Torrance (Eds.), Changing educationalassessment: International perspectives andtrends. London: Routledge.

Kellaghan, T. (in press). Examination sys-tems in Africa: Between internationaliza-tion and indigenization. In M. A. Ecksteinand H. J. Noah (Eds.), Examinations: Com-parative and international studies. New York:Pergamon.

Kellaghan, T., and Greaney, V. (1989). Astudy of examination systems of five Africancountries. Report of a study carried out forHigher Education for Development Co-operation. Dublin: Educational ResearchCentre.

Kelly, M. J. (1986). An overview of education inZambia. Lusaka: University of Zambia.

Kelly, M.J. (1991). Education in a decliningeconomy. The case of Zambia, 1975-1985.Washington : World Bank.

Kelly, M. J., Nkwanga, E.B., Achola, P.P.W.,Kaluba, L.H., and Nilsson, K. (1986).The provision of education for all. Towardsthe implementation of Zambia's educationalreforms under demographic and economic con-straints,1986-2000. Lusaka: University ofZambia.

Kenya National Examinations Council. (1986).K.C.P.E. Newsletter 1987. Nairobi.

69

7

Kwasna, K. B. (1974). Content validity andreliability of domain-referenced tests.African Journal of Educational Research, 1,73-79.

Le Mahieu, P. G. (1984). The effects onachievement and instructional content ofa program of student monitoring throughfrequent testing. Educational Evaluationand Policy Analysis, 6, 175-187.

Lesotho. (1982). The educa' .n sector survey.Report of the Task Force. Maseru: Govern-ment of Lesotho.

Lewin, K. (1984). Selection and curriculumreform. In J. Oxenham (Ed.), Educationversus qualifications? A study of relation-ships between education, selection for employ-ment and the productivity of labour. London:Allen and Unwin.

Linn, R. L. (1983a). Curriculum validity:Convincing the courts that it was taughtwithout precluding the possibility of mea-suring it. In G. F. Madaus (Ed.), The courts,validity and minimum competency testing.Boston: Kluwer-Nijhoff.

Linn, R. L. (1983b). Testing and instruction:Links and distinctions. Journal of Educa-tional Measurement, 20,179-189.

Little, A. (1982). The role of examinations inthe promotion of "the Paper QualificationSyndrome". In ILO. Jobs and SkillsProgramme for Africa. Paper Qualifica-tions Syndrome (PQS) and unemployment ofschool leavers. A comparative sub-regionalstudy. Addis Ababa: International LabourOffice.

Little, A. (1984). Combating the diplomadisease. In J. Oxenham (Ed.), Educationversus qualifications? A study of relation-ships between education, selection for employ-ment and the productivity of labour. London:Allen and Unwin.

Little, A. (1990). The role of assessment, re-examined in international context. In P.Broadfoot, R. Murphy and H. Torrance(Eds.), Changing educational assessment . In-ternational perspectiv,Is and trends. London:Routledge.

Lockheed, M. E., and Kcmenan, A. (1989).Teaching quality and student a chievement

in Africa: The case of Nigeria andSwaziland. Teaching and Teacher Educa-tion, 5, 93-113.

Lockheed, M. E., and Verspoor, A. M. (1990).Improving primary education in developingcountries. A review of policy options.Washington, DC: World Bank.

Lulsegged, A. (1988). Assessment of impact ofSIDA assistance to the Swaziland govern-ment. Unpublished manuscript.

Madaus, G. F. (1988). The influence of testingon the curriculum. In L. N. Tanner (Ed.),Critical issues in curriculum. Eighty-seventhyearbook of the National Society for the Studyof Education, Part 1. Chicago: NSSE.

Madaus, G. F., Airasian, P. W., and Kellaghan,T. (1980). School effectiveness. A reassess-ment of the evidence. New York: McGraw-Hill.

Madaus, G. F., and Greaney, V. (1985). TheIrish experience in competency testing:Implications for American education.American Journal of Education, 93, 268-294.

Madaus, G. F., and Kellaghan, T. (in press).Curriculum evaluation and assessment.In P. Jackson (Ed.), Handbook of research oncurriculum. New York: Macmillan.

McNamara, V. (1982). The long-term effectsof "carrots". School system structure andcurriculum, SSCEP and the problem ofmotivating learning that is functional.International Journal of Educational Devel-opment, 1, 49-60.

Mingat, A., and Psacharopoulos, G. (1985).Education costs and financing in Africa: Somefacts and possible lines of action. Discussionpaper. Report No. EDT13. Washington,DC: World Bank.

Myeni, A. D. (1985). Standardisation and theJunior Certificate programme. SIER Bul-letin, 6, 30-40.

Nafziger, E. W. (1989). Inequality in Africa.Political elites, proletariat, peasants and thepoor. Cambridge: Cambridge UniversityPress.

NCTPP (National Commission on Testingand Public Policy). (1990). From gatekeeperto gateway: Transforming testing in America.Report of the National Commission on Test-

ing and Public Policy. Boston College.Chestnut Hill, MA.

Nigam, S. B. L. (1982). Preface. In ILO. Jobsand Skills Programme for Africa. PaperQualifications Syndrome (PQS) and unem-ployment of school leavers. A comparativesub-regional study. Addis Ababa: Interna-tional Labour Office.

Noah, N. J., and Eckstein, M. A. (1990). Trade-offs in examination policies: An interna-tional comparative perspective. In P.Broadfoot, R. Murphy and H.Torrance (Eds.), Changing educational as-sessment: International perspectives andtrends. London: Routledge.

Nuttall, D. L. (1990). Proposals for a nationalsystem of assessment in England andWales. International Journal of EducationalResearch, 14, 373-381.

Ogundare, S. (1988). Cuiriculum develop-ment: A description of the developmentof the national curriculum for primarysocial studies in Nigeria. EducationalStudies, 14, 43-50.

Oloo, A. (1990). KCPE ranking is wrong.Daily Nation (Nairobi), February 3.

Ongom, D. L. (1990). Malpractices in ex-aminations. Paper presented to seminaron Malpractices in Examinations. Zomba,Malawi.

O'Rourke, B., Martin, M. 0.. and Hurley, J. J.(1989). The Scholastic Aptitude Test as apredictor of third-level academic perfor-mance. Irish Journal of Education, 23, 22-39.

Oxenham, J. (1983). What examinations testand emphasize: An analysis of the Le-sotho Primary and junior Secondary Ex-aminations. In Lesotho, The education sec-tor survey. Annexes to the Report of the TaskForce. Maseru: Government of Lesotho.

Pennycuick, D. (1990). The introduction ofcontinuous assessment system at second-ary level in developing countries. InP.Broadfoot, R. Murphy and H. Torrance'Eds.), Changing educational assessment.London: Routledge.

Perrott, E. (1982). Effective teaching: A practi-cal guide to improve your teaching. London:Longman.

70

Quansah, K. B. (1985). Current testing inAfrica: Selection or measure of real per-formance? Prospects, 15, 281-288.

Sebatane, E. M. (1985). Perspectives of edu-cational assessment in Lesotho. Studies inEducational Evaluation, 11, 269-278.

Somerset, H.C.A. (1987). Examination reformin Kenya. Discussion paper. Washington,DC: World Bank.

Somerset, H.C.A. (1988). Examinations as aninstrument to improve pedagogy. In S. P.Heyneman and I. Fagerlind (Eds.), Univer-sity examinations and standardized testing.Washington, DC: World Bank.

Swaziland, Ministry of Education. (1985).Report of the National Education ReviewCommission. Mbabane: Ministry of Edu-cation.

Swaziland, Ministry of Education. (1986).Wastage in the educational system. Survey onthe causes of repetition and drop-out in pri-mary and junior secondary schools inSwaziland.. Mbabane: Educational Re-search and Planning Department,Ministry of Education.

Swaziland, Ministry of Education. (1987).Final report on a management developmentconference for senior management. A strate-gic approach to continuous assessment andinstructional leadership. Mbabane: Educa-tional Research and Planning Section,Ministry of Education.

Tan, J-P. (1985). Private enrollments andexpenditure on education: Some macrotrends. International Review of Education,31, 103-117.

71

Terwilliger, J. S. (1989). Classroom standardsetting and grading practices. EducationalMeasurement Issues and Practice, 8(2),15-19.

Torrance, H. (1986). What can examinationscontribute to school evaluation? Educa-tional Review, 38, 31-43.

Uganda, Ministry of Education. (1989). Re-port of Education Policy Review Commission.Kampala.

Unesca (1988). Statistical yearbook, 1988.Paris.

Unesco (1990). Basic education and literacy.World statistical indicators. Paris.

Windham, D. M. (1986). Internal efficiency andthe African school. Discussion paper.Washington, DC: World Bank.

World Bank. (1988). Education in Sub-SaharanAfrica. Policies for adjustment, revitaliza-tion, and expansion. Washington, DC.

World Bank. (1990). Education in Sub-SaharanAfrica. Updated statistical tables. AMIDTechnical Note No.1. Washington DC.

World Declaration on Education for All.(1990). Adopted by the World Conferenceon Education for All, Meeting Basic Learn-ing Needs. New York: UNICEF House.

Yussufu, A. (1989). Promise and problemsassociated with the assessment of a re-vised curriculum based on behaviouralobjectives in the Republic of Kenya. Stud-ies in Educational Evaluation, 15, 279-284.

Zambia, Ministry of Education. (1977).Educational reform: Proposals and recom-mendations. Lusaka: Department of Tech-nical Education and Vocational Training.

9 Appendices

Appendix 1National Education Systems

Countries andlinguisticstatus

Compulsoryschoolage

Entrance age and duration of first andsecond-level (general) education

Anglophone

3 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Kenya PPPPPPPPSS SSLesotho 6-13 PPPPPPPS1 51 51 S2 S2

Mauritius PPPPPPS1 51 51 S2 S2 S2 S2

Swaziland PPPPPP PS1 S1 S1 S2 S2

Uganda PPPPPPPS1 S1 S1 S1 S2 S2

Zambia 7-14 PPPPPPPS1 51 S2 S2 S2

Francophone

Chad 6-12 PPPPPPS1 51 51 S2 S2 S2

Guinea 7-13 PPPPPP 51 51 S1 51 S2 S2 S2

Madagascar 6-13 P P P 13 P S1 S1 S1 S1 S2 S2 S2

Mauritania PP 13 P 13 PS1 51 51 S2 S2 S2

Rwanda 7-15 PPPPPPPPS1 51 51 S2 S2S2

Togo 6-12 PPPPPPS1 S1 51 51 S2 S2 S2

Other

Cape Verde 7-13 PPPPPPS1 51 51 S2 S2

Ethiopia 7-13 PPPPPPS1 S1 S2 S2 S2 S2

P=Primary; S1=Secondary Junior Cycle; 52=Secondary Senior Cycle

Sources: UNESCO (1988), and national reports

73

82

Countries/LinguisticStatus

When Given

Anglophone

KENYA

. Appendix 2TITLES AND DESCRIPTIONS OF EXAMINATI.ONS

Title Function Res onsibili

1. End of primary KCPEb(8)a

2. End of secondary KCSE(4)

Certification/Selection

Certification/Selection

Multiple- KNECchoice,Essay

Essay, KNECMultiple-choice,Practical,Oral, Aural

LESOTHO

3. End of primary Primary School Certification/ Ahnost all Ministry ofschool (7) Leaving Certificate

ExaminationSelection Multiple- Education

choice,Letter and Essay

4. End of junior junior Certificate Certification/ Essay, Examinationsg)cle secondary Examination Selection Short answer Council of

Practical Lesotho

5. End of senior Cambridge Overseas Certification/ Short UCLEScycle (2) Certificate Exam. Selection answer,

Essay

MAURITIUS

6. End of primary CPE Certification/ Multiple- MES(6) Primary Educ.

(CPE)Selection choice,

Essay

7. End of first SC Certification/ Multiple- MEScycle secondaryage 16+ (2)

Selection choice and UCLESEssay,Practical

8. End of second HSC Certification/ Multiple- MES incycle secondaryage 18+ (Z

Selection choice collaborationEssay, with UCLES andPractical University of

London GCE 0-and A- Levels

a. in parent eses in icate tb. Acronyms defined on page ix.

um Ts e num r o gra es at a particu ar level.

74

Appendix 2 (continued)

When Given Title Function Type

SWAZILAND

9. End of Primary(7)

10. End of firstcycle secondary(Form 3)

11. End of secondcycle secondary(Form 5)

UGANDA

12. End of primary(7)

13. End of firstcycle secondary(4)

Primary Certificate Certification/Selection

Jwtior Certificate Certification/Selection

Cambridge Overseas Certification/Certificate Selection

PLE

UCE

14. End of senior UACEcycle secondary(2)

ZAMBIA

15. End of primary(7)

16. End of juniorcycle secondary(2.)

17. End of seniorcycle secondary(3)

Certification/Selection

Certification/Selection

Certification/Selection

Responsibility of

Multiple-choice plusEssay,Practical,ListeningComprehension

Multiple-choice plusEssay,Practical,ListeningComprehension

ShortanswerEssay

Shortanswer

ShortanswerPractical,ProjectsOral, Aural

ShortanswerPractical,ProjectsOral, Aural

Grade 7 Composite Certification/ Multiple-Examination Selection choice

Grade 9Junior SecondarySchool Leavingk.2xarrtination

Joint SchoolCertificate andGeneral Certificateof Education Exam

Certification/Selection

Certification/Selection

Essay,Multiple-choice, oraland aural, practical

SwazilandExaminationCouncil

SwazilandExaminationCouncil

Swaziland Exam.Council and UCLES

IJNEB

UNEB

UNE13

ExaminationsCouncil ofZambia

ExaminationsCouncil ofZambia

Essay, Multiple Examinationschoice, Council ofPractical Zambia

75

84

Appendix 2 ( continued)

Whe.LE1 Type

Francophone

CHAD

18. End of primary CEPE(6) CEPEA

19. End of primary Primary(6) Concours

20. End of first BEPCcycle secondary(4)

21. End of firstcycle secondary(4)

Concours forsecond cycle

22. End of second Baccalauréatcycle secondary(3)

GUINEA

23. End of primary(6)

24. End of firstcycle secondary(4)

25. End of secondcycle secondary(3)

MADAGASCAR

Certification Written

Selection

Certification NA

Selection

Responsibili

SECMinistiy ofEducation

Dictation, SEC, Ministry ofarithmetic Education

Essay andOral

Selection / NACertification

Seventh year Entrance/ NASelection

BEPC NA NA

Baccalauriat Certifiction/ Written,Selection Oral

26. End of primary CEPE Certification/ Written(5) (Regional E..e.m) Selection

27. Annual firstcycle secondary(4)

School-based Selection/ Written,Promotion Oral

28. End of first CFEPCEScycle secondary(4)

Certification/ NASelection

SEC, Ministry ofEducation

SEC, Ministry ofEducation

NationalUniversity ofChad

NationalEducationMinistry

NationalEducationMinistry

NationalEducationMinistry

Ministry ofPrimary Ed. andProvincialDirector

School

Ministry ofPrimary andSecondaryEducation

76

.)

Appendix 2 (continued)

When given Title Function Type Responsibilityof

29. End of second Baccalauriat Certification/ Written Ministry ofcycle secondary Selection Higher

(3) Education

30. After two years CFE/FP1 Certification/ Written Ministry of

of post-primary Selection Secondary andPrimary Educatior

31. After two years CFE/FP2 Certification/ Written Ministry of

of senior cycleeducation

Selection Secondary andPrimary Education

32. End of second Baccalauriat Certification/ Written Ministry of

cycle secondary technique Selection Secondary and

(3) Primary Education

MAURITANIA

33. End of primary CEE Certification Written Ministry of

(6) Education

34. End of primary Secondary Selection Written Ministry of

(6) entrance Education

35. Fnd of first BEPC Certification/ NA Ministry of

cyde secondary (Not obligatory) Selection Education

(3)

36. End of second Baccalaur:fat Ceraication / NA Ministry of

cycle (3) Selection Education

RWANDA

37. End of primaryschool

Concours Selection Written Ministry ofPrimary andSecondaryEducation

38. Term and end oiyear

School-based Feedback/Promotion

Written,Oral,Practical

Teacher

39. Third, fourth Examens de Placement of NA Committee for

and fifth years reclassement students from Programs in

of secondary private orforeign schools

Secondary Schools

77

86

Appendix 2 (continued)

When given Title Function Type Responsibility of

40. NA Le Jury Central Certification/ Oral Committee forSelection forprivate andrepeatcandidates

Programs inSecondary Schools

41. End of secondary NA Certification Written Committee for(6) (Essay,

Multiple-choice, Oral)

Programs inSecondary Schools

TOGO

42. End of 6th grade CEPD Certification/ Written DECSelection

43. End of firstcyde secondary

DEPC Certification Written DEC

(4)

44. End of firstcycle secondary

Concours Selection NA DEC

(4)

45. End of second Baccalaureat Certification Written and DEC andcyde secondary oral Directorate of(3) Third Level

Education

46. End of secondary Concours Selection foruniversity

Written DEC

Other

CAPE VERDE

47. End of basiccomplementary(2)

48. End of generalsecondaryeducation (3)

Elementary cycle Certification,examination Selection(Exemption option)

,

Oral

General secondary Certification/ Written,examination Selection Oral(Exemption option)

78

e e

NationalExaminationCornniittee,Ministry ofEducation

NationalExaminationCommittee,Ministry ofEducation

Appendix 2 (continued)

When given Tide Function Type Responsibility of

49. End ofcomplementarysecondaryeducation (2)

Diploma of UpperSecondary Educ.(Exemption option)

Certification/Selection

Written,Oral,Practical

NationalExaminationCommittee,Ministry ofEducation

ETHIOPIA

50. End of primary Grade 6 National Certification/ Multiple- National(A) Examination Selection choice,

School-basedcomponent

ExaminationBoard, Ministryof Education

51. End of Grade 8: Grade 8 National Certification/ Multiple- NEB, Mimstry of(2 years - firstcycle)

Examination Selection choice Education

52. End of Grade 12: ESLCE Certification/ Multiple- University of(4 years - secondcycle)

Selection choice(Essayup to 1977)

Addis Ababa

79

S

Appendix 3

Kenya Certificate of Primary Education:Sample Essay awl Marker's Comments

That evening my father came home lookingunusually happy. He told me he had some-thing to tell us, but before he could do that, heasked me to go and call Heri. When we bothcame in we found that everyone else waswaiting for the two of us. My father as usualwas seated on his armchair with a cigarette inhis mouth. My mother was standing by himand my two other sisters were seated at thedining table. We all had an idea of what thegood news was but we were unsure.

When everybody had seated down, my fa-ther began talking. There was suddenly a cryof joy and everbody came rushing towardsme. It was so sudden that I almost fell off myseat. Everybody was hugging me and doingall sorts of funny things. I just could notbelieve my eyes. What I had been waiting forso anxiously was now revealed. It was toogood to be true. I do not know how I made it,but passing my examination was the bestnews I had ever heard. None of us couldcontrol ourselves. Some of us almost ju apedhigh enough to touch the ceiling with ourheads. All that was ringing in my head wasthat I had passed with flying colours and towhich school I would be chosen to go.

My father asked everyone to be silent since hehad not finished saying what he wanted.

80

#.1

Once resettled, he told us that I was amongthe top people in the schr)ol and that I wouldget an award for that. he also told us that hewas to go and check which school had chosenme to be one of their student,. I wasoverjoyedin such a way that I could hardly move fromwhere I was now seated. I could just remem-ber the way I was so nervous while doing theexams and all the words of good luck frommy friends and relatives. I was glad that allthe messages in the cards had been fulfilledand also the fact that I did not disappoint myparents. This had happened before with Heri,but I had challenged him and beaten him. Icould remember all the promises made to meby my uncles and aunts and I felt as thoughmy heart was smiling inside.

After having supper that n ,ght, I was called tomy parent's room and aF: T ed what I wanted tohave as a present. I named almost everythingone could think of and this made tliem laughsince they knew the situation I v. as in and inwhich they had also gone through. Since theycould not buy me everything, I wat.7 romisedthe most essential things. I would never for-get that day because it was the happiest day ofmy life and having that feeling inside memakes me feel that in future I will also do well

Ma' l< Awarded: 37

Marker's Comments:

Although this piece of writing was not thevery best, it did represent that top qualitywriting which is not very easy to come by.Candidates who reached this standard wereactually very few. Except for such errors asthe usage of 'seated' instead of the simple past'sat' and such flaws as 'after having supperthat ,Light'a candidate as this would evenhave scored higher marks had he been moreinteresting and had he used a greater varietyof structures in his writing.

From the brief analysis of errors in the com-positions included in this newsletter and inother compositions written by candidates in1985 KCPE we discovered that certain kindsof errors were being made repeatedly. Incor-rect verb tenses, wrong usage of words, spell-ing errors and errors in syntax were wide-spread.

81

90

Examples:

Incorrect verb tenses:

We all sit down and listen to him. (We all satdown and listened to him.)

My father did not agreed. (My father did notagree.)

They took the money and leaw him crying.(They took the money and left him crying.)

Numerous similar errors were made by can-didates.

Soul ce: Kenya National Examinations Coun-cil (1986)

ARGENT:NA4.1k0 itri'd4 SRL-

GA lera GWAV,0111

Poodo 145, 4j Sore.01.t. 451/4431333 Stara* Nona

AUST :SALM, PAPUA NEW cumw.SOLOMON ISLANDS,

VANUATU, AND WESIV EN SAMOAnooks &Jairods

Whioahom MedMadison 3132I/Kara

AUSTRIAGerad arid Co.Graben 31MOW Weer

BAHRAINnaltaa ileausrdi ond CreeoPcmy

Auga LtdErato/ Oaldlistt No 42Caplan Mk Alrte

Mom, Town 317

n1-4GI ADE MtMICK. Ind Natter% Otrvelopose

hiatetonor %day tALIDADHoax S, Kor 4114IN.MonwW1 RiAtwMelia 1%4)

Bootee offrac151, Nur Ahmed East.Otillogorec

7e, K.O.A. AvenueYJJAP

BELGIUMlarinov

Ay du Ita 2021060 Bawd,

CANADA011hoonur

C P. 19319 rve ActrakreSnuchervillo, VaNatberJOB SEE

fl IN AOdra Anancial & Eowittroo

Publiohlog HCRIbe41, Fo 5, Dams Lk&tang

CO Li:NM MA

Menace le itApanale Aorta 74170Rogrea E

CS:ME 0 ll-'0114.41CenTt d'Ultien et de ONPLeirri

&nee nos04 9 PAbodion 04 PLACA%

Cyr RUEMEM En Inhanoton SuatceoP 0 No. UV

osia

DENMARKSainheidoLiatrocurRownotion Alit. 11(Nk. Iwo Prat ecilobrez C

DOMINICAN RENOIR'P4A1 tor aToUte C por A.Remoras% e Wad la Crane,' 309Apartado Poeta 1190Santo Doodego

EGYPT, ARAB RE1/111.1C DEAl Moan

GrOw StreetCaro

The Middle Eat Ctraever41, Sherd StreetClue

EL SALVADOR;feriaAvenida Mandel Enrique Amp MGEd.6 coo SIM, ler PlanSki Salt oder

Distributors of World Bank PublicationsPiNIAKT)ktotearristeriKkolskauppa

C. nos 126Holamil 10SIN06101

FIANCEWald Buds Nthcononslot tram (Wmnil Porie

CEEMANTUNGVerLagPoppelad taw Melt 55D530) loan

KEW2t Ippodassoto Sawa Saes PIterfaaA 0511110 '1435

OJATEMA LAUtirearo Piedra Santa

C.410 7-5feZara 1Cuctornala City

HONG KONG, MACAOAais PM OA.4641 Wyndivea %netWhrreng Cane'ad SoarCadroi Hats Rang

Mad PubRehone Pratte Lul.2S1 Moat ReadMadras LOU 022

Beach reamISj N. ;it! idAir MtnBala?! EmNorBanbay OW 0216

13/14 Aal Ali Poe/New !NW 110(61

17 Mazer,* AvenueCalcutta ''//072

lry ed in a Hattie Ilairilny544 h4na Med aulltWociarSswgr.ore SLO

3.5 11 29 1:..whisu4a Crow Poi4Hyderabad So7.1 Ere,

Nrw)u Rain 7ro4 FloorNeer Thekce. Dewy idevangpatAhnotiatert AOC%

Pan411 HMO*14A Ashok Menle-7t.now 12.5 001

Ft. Indira Imutito&AM Riordeno 17

P 0 Box 1411aorta Pact

ISRAEL?woo Utr, wort ledP 0 Box .504'.I5I'd Aviv M SE0and

ITALYI :moo Casuniaaauria Senoral 5PA,

(AK. 1. Calahna, I /1Ciaello Prosaic 55230115 Florence

lAtANEanon Bove. Service74 Hato Samna, EriaNytnhu 113Inky's

KENYAAtrICA Book Servo* (EA./ Ltd)'3 fat 46115Nairobi

X) BEA, !JJIJIUC OrVan Keret NPAI CtrytitInanP 0 Box tal, Kw allgwitlinrilA501,41

onfointrM53orRE IMarcation %MowP O. Ern SFS

MALAYIERlerdvasity of Mal ora Co/medal

lidokshrip. Lira/KJv a 5,0, Itv, jlihA Psota NV .4Kuala LciapNe

mutasNrormApertedo Poet: 22-640IKKOTIodpon, Macaw D

MOROCCOS-14114 0' Zealot Matra% Maratha12 roe Mame. bd. d'AnleCamblance

IVITHERLANDSat UM tborm/InGoPabb.loatha

Soo 70274X) AB HaWabwrw

NEW 7.1t4L4NOHills Newry wel IsdorreaRott StarkeNaha BitNew Madat

NIC IMOIndeerety hest HanoiDora trawitendgdeitglendusPave% Mad nog S09SRada

NOWAYNervation, lisharnadem CantoSoak DoparairatF O. Era s115 EtteratedN.0443.o ado 4,

OMANWNW! litheemion %monoPC. Rao 1413, Saab AirportMulat

FAIOSTANhUrsep nook Agars,5. ShahrshinQuad.aAvass

P 0 Boo No 729Lahore 3

rr-kt,lido Swat DIA117011C SAApartado 3124LIVA

tizrtzrucsIntianacceAr Bonk CameSeth Aeon KIIiernse Ica BuildingAykia Av4111,14 MatatMitto Mande

rottmc ALLIVIASA PortugalRUN CA1,0470.741203 Lisbon

',AUDI *EWA, QATARlair nook SoWSP 0 Box 31%iyadh 11471

MDARB IniorxrwOol. WrvlowBrod' lir=AI Aim SurtoAl Diana Cadrerifle FIG-ZPO Boall%6.1y4d2o

Hop Abduolah Alma loW/dIngKing hludad StreetP 0 thar /349°motto

30, Mohamed 14/10118/1 Awed StreetP 0 Box 9776

$114:1APORIL TAIWAN,MYA'NMAILIIRLINEI1ittnV4SIM Publicadoia

Prise% Ltd0106 1.4 i. FeEllo Fedi:mid

Eleg.14 New radios/rid Road

RAG opeio 1953

SWIM AFRICA. BOTSWANAFor roe& WmOared Unweteity Praia

Souiltoon AfricaP 0 Boo 1141CapeTown 1030

row amarropiree CrawIncanuereal Subecapbon SavoP 0 pox MOCnoghallJohanntaburg 3324

SPAMMtadhPrensa Lebrea SA.Camila 37scot muted

Liberia Iniontasonsi AELTJGCarona de Care. 39109CD3 Emotion s

SPI LANKA AND TILL MALDIVESLaka HoulatiookshopP 0. Box 144100, Sir Chittampolun A

Gardiner MarathaColombo 2

SWEDENPer mulls fatsEmu* F. SoladoestestReeparlegogasi 17, Bar 1E354E. it* 27 Siva/taint

Fly serlsoratice orderWennergraN-WdEeina AllBat MOGI5-104 25%6E01m

5WETZERLANI3For toes fillorLlama Part4. rue CannaCole postale 3$1C/11111 a.ve 21

Far wbonalce IrianWear:: FartSu tr:ce des Abreact antsCat postale 3311

1CC2 lARI/Ahlte

TANZA NMOared Urovereity PeasP 0 Box 5299Dot es Salmon

THAILANDCenotl Dept-arum Sort306 Weal PlodI I **Lc*

TILLNIDAD & TOBAGO, ANTIGUAARI1L93 A, BARS A DOS,

DOM WIC& GEENADA, GUYANA,JA 14 MCA, MONTIEREAV, cr.MIS & NEVIS, ST. LUCIA.ST. vlsagqr B. GRENADINESSyne:nano, Sotedlea Untt9 Woo ServeCaveTrawled. Was India

L rrev ALA. 6 VAIRATEd14 1RnrIcoP 0 l'o WVF.Earah

UNTIED KINGDOMMiatrirda LitP 0 Boo 3AI ion. IL ourpchav G1114 2l4GEngland

'rENEZULIALianas dei EtaA oa oda 20337Caracas 1063rA

RECENT WORLD BANK TECHNICAL PAPERS (continued)

No. 134 de Haan and Bekure, Animal Health in Sub-Saharan Africa: Initial Experiences with AlternativeApproaches

No. 135 Walshe, Grind le, Nell, and Bachmann, Dainy Development in Sub-Saharan Africa: A Study of Issuesand Options

No. 136 Green, editor, Coconut Production: Present Status and Priorities for Research

No. 137 Constant and Sheldrick, An Outlook for Fertilizer Detnand, Supply, and Trade, 1988/89-1993/94

No. 138 Steel and Webster, Small Enterprises under Adjustment in Ghana

No. 139 Environment Department, Environmental Assessment Sourcebook, vol. 1: Policies, Procedures,and Cross-Sectoral Issues

No. 140 Environment Department, Environmental Assessment Sourcebook, vol. II: Sectoral Guidelines

No. 141 Riverson, Gaviria, and Thriscutt, Rural Roads in Sub-Saharan Africa: Lessons from World Bank Experience

No. 142 Kiss and Meermari, Integrated Pest Management and African Agriculture

No. 143 Grut, Gray, and Egli, Forest Pricing and Concession Policies: Managing the High Forest of Westand Central Africa

No. 144 The World Bank/FAO/UNIDO/Industry Fertilizer Working Group, World and Regional Supplyand Demand Balances for Nitrogen, Phosphate, and Potash,1989190-1995196

No. 145 Ivanek, Nulty, and Holcer, Manufacturing Telecommunications Equipment in Newly IndustrializingCountries: The Effect of Technological Progress

No. 146 Dejene and Olivares, Integrating Environmental Issues into a Strategy for Sustainable AgriculturalDevelopment: The Case of Mozambique

No. 147 he World Bank/UNDP/CEC/FAO, Fisheries and Aquaculture Research Capabilities and Needsin Asia: Studies of India, Thailand, Malaysia, Indonesia, the Philippines, and the ASEAN Region

No. 148 The World Bank/ UNDP/CEC/FAO, Fisheries and Aquaculture Research Capabilities and Needsin Latin America: Studies of Uruguay, Argentina, Chiie, Ecuador, ard Peru

No. 149 The World Bank/ UNDP/CEC/FAO, Fisheries and Aquaculture Research Capabilities and Needsin Africa: Studies of Kenya, IVIalawi, Mozamkque, Zimbabwe, Mauritania, Morocco, and Senegal

No. 150 The World Bank/UNDP/CEC/FAO, International Cooperation in Fisheries Research

No. 151 The World Bank/UNDP/CEC/FAO, Tropical Aquaculture Developmer : Research Needs

No. 152 The World Bank/UNDP/CEC/FAO, Small-Scale Fisheries: Research Needs

No. 153 The World Bank/UNDP/CEC/FAO, Small Pelagic Fish Utilization: Research Needs

No. 151 En vironment Department, EnOronmental Assessment Sourcebook, vol. III: Guidelinesfor Envircnmental Assessment of Energy and lndus try Projects

No. 155 Belot and Weigel, Programs in Industrial Countries to Promotc Foreign Direct Investment in DevelopingCountries

No. 156 De Geyndt, Managing Health Expenditures under National Health Insurance. 7'he Case of Korea

No. 157 Critchley, Reij, and Seznec, Water Harvesting for Plant Production, vol.11: Case Studiesam' Conclusions for Sub-Saharan Africa

No. 158 Hay and Paul, Regulation and Taxation of Commercial Banks during the International Debt Crisis

No. 159 Liese, Sachdeva, and Cochrane, Organizing and Managing Tropical Disease Control Programs:Lessons of Success

No. 160 Boner and Krueger, The Basics of Antitrust Policy: A Review of Ten Nations and tlw EuropeanCommunities

No. 161 Riverson and Carapetis, Intermediate Means of Transport in Sub-Saharan Africa: Its Potentialfor Improving Rural Travel and Transport

No, 162 Replogle, Non-Motorized Vehicles in Asian Cities

No. 163 Shilling, editor, Beyond Syndicated Loans: Sources of Credit for Developing Countries

No. 164 Schwartz and Kampen, Agricultural Extension in East Africa

The World BankHeadquarters1818 H Street, N.W.Washington, D.C. 20433, U.S.A.

Telephone: (202) 477-1234Facsimile: (202) 477-6391Telex: wui 64145 WORLDBANK

RCA 248423 WORLDBK

Cable Address: IN TBAFRAD

WASHINGTONDC

European Office66, avenue d'Iéna75116 Park', France

Telephone: (1) 40.69.30.00Facsimile: (1) 40.69.30.66Telex: 640651

okyo OfficeKokusai Building1-1 Marunouchi 3-chomeChiyoda-ku, Tokyo 100, japan

Telephone: (3) 3214-5001Facsimile: (3) 3214-3657Telex: 26838

ISBN 0-8213-2052-1