Universalization of primary education and literacy, phase II; …unesdoc.unesco.org/images/0006/000671/067149eo.pdf · 2014-09-25 · between formal and non-formal education and greater

Restricted Technical Report RP/1984-1985/11.2.3/1V.4.3

...-

ETHIOPIA Action to achieve better co-ordination between formal and non-formal education and greater continuity between the various parts of the education system

Universalization of Primary Education and Literacy: Phase I I - Research and Eva1 uation Studies

by Ko-Chih R. Tung Vinayagum Chinapah

Serial No. FMFUED/EPP/85/146

A United Nations Educational, Scientific and Cultural Organization UltSCO Paris, 1985

E T H I O P I A

UNIVERSALIZATION OF PRIMARY EDUCATION AND LITERACY:

Phase I1 - Research and Evaluatiron Studies 1

by Ko-Chih R. Tung Vinayagum Chinapah

V

Report prepared for the Government of Ethiopia by the United Nations Educational, Scientific and Cultural Organization ( Une s co )

U N E S C O

Technical Report ElP/1984-1985/11. 2.3/IVO4. 3 FT4R/ED/E?P/85/146 (Tung/Chinapah) 2 December 1985

I'

0 Unesco 1985 Printed in France

TABLE OF CONTENTS

Page

I. TNTRODUCTION . . . . . . . . . . . . . . . . . . . . 1 11. MISSION ACTIVITIES . . . . . . . . . . . . . . . . . 2 111. FINDINGS ON PHASE 1 OF THE "QUALITY OF

EDUCATION STUDY" . . . . . . . . . . . . . . . . . 3 ITS. RECOMMENDATIONS 8

A. PhaseI.. . . . . . . . . . . . . . . . . . . 8 B. Phase11 . . . . . . . . . . . . . . . . . . . 10 C. Further recommend'ations . . . . . . . . . . . ..I1

V. C O N C L U S I O N . . . . . . . . . . . . . . . . . . . . . 1 2

ANNEX I - UNIVERSALIZATION OF PRIMARY EDUCATION AND LITERACY (UPEL) PROPOSALS FOR PHASE 11: RESEARCH AND EVALUATION STUDIES . . . . . . . . . . . . . . . 16

ANNEX I1 - LIST OF NATIONAL OFFICIALS AND INTERNATIONAL EXPERTS MET DURING THE MISSION . . . . . . . . . . . 98

- 1 -

I. INTRODUCTION

1. The mission described in the present report was carried out from 11 November to 17 December 1984 at the request of the Government of Ethiopia, and was funded by Unesco under its Regular Programme for 1984-1985. The purpose of the mission was:

"to cooperate with the Ministry of Education in the Study on the Quality of Education, a) in the evaluation of Phase I b) in the implementation of Phase 11".

Y 2. The specialists selected for this mission were Dr; Ko-Chih Tung of Interscience Research Corporation, Solna,. and Professor of the Univer- sity of Stockholm, Sweden, and Dr. Vinayagum Chinapah, Vice-Director, Institute of International Education, University of Stockholm, Sweden.

3. Because of the technical work entrusted to the specialists, a careful effort was made to synchronize their missions, as had been requested by the Government.

4. Bearing in mind the need for close co-ordination, Dr. Ko-Chih Tung was asked to undertake the following tasks:

- examination of the appropriateness of the coding data utilized - recommendation of possible appropriate statistical tools and during Phase I of the National Quality Education Study;

methods necessary for further analysis of data; - assistance to the national staff in the development of computer programmes (encoding and software) appropriate to the nature of the data that need to be processed and tested, along with assistance in the use of various input media;,

- provision of a system approach framework for relating the data assembled in the four major areas of the project;

- and provision of general and specific advice to research teams, to assist in refining methods, methodologies, techniques and strategies to be used during the remainder of the project.

c

5;. Dr. Chinapah carried out the following duties: - analysis of data collected by the national authority on the Quality of Education Study, with particular emphasis on the curriculum and contents of education as well as general operation of the education system;

- evaluation of research requirements and analysis of needs for refining the data assembled in the four major areas of the Quality of Education Study and relating them with Phase'II of the afore-mentioned study;

- establishment of a framework and elaborate models for analysis and presentation of data to be processed, and hypotheses to be tested, prior to the implementation of Phase I1 of the Quality of Education Study;

- provision of advice on the launching of subsequent phases of the Quality of Education Study, particularly in the field of reinforcement of educational research and data analysis.

Dr. Ko-Chih Tungs' mission lasted from 11 November to 2 December, 1984; Dr. Chinapah's from 17 November to 17 December, 1984.

- 2 -

6; In genera , it may be said that both consuatants complied with the terms of reference, which were further specified and developed by the national authority in a meeting held on 30 November, 1984. The data collected by the national authority durhg Phase I was evaluated by both mnsultants in terms of their particular areas of competence.

7'. Suggestions for changes and improvements in methodologies were offered, along with advice on the launching of subsequent areas, phases amd models for analysis and presentation of the data. In addition, a- certain amount of training of national officials was done by the consultants. Finally, with specific reference to Dr. Tung's consultancy, advice on the development of a computer system was given. These aspects axe described in detail in the present report.

&. The main recommendations for Phase I are given in Part IV of this report. Part I11 provides some findings on Phase I of the Quality of Education Study, useful for implementing Phase I1 (Universalization of Primary Education and Literacy, UPEL). Part VI refers to further recommendations to be kept in mind in Phase 11, particularly in regard tD training in modern research and methodology in several areas. Part VI1 makes research proposals for UPEL studies, Phase 11.

9. In order to delineate relevant areas of research and integrate the research findings, this report has been focussed on the critical phases of the educational cycle, starting with access to education and ending with the impact of education on the community.

IO'.. Six areas of the educational cycle were earmarked for the introduction of UPEL in Phase I1 af the project. These areas are discussed in Annex I of this report.

4'

II.., MISSION ACTIVITIES

11. (i) The consultants met with and interviewed members of the Curriculum Department of the Ministry of Education, especially those connected with its research programme. They also interviewed Representatives of Unesco and UNICEF (the list of people met with during the mission appeared in Annex 11), and many of the consultants' recornendations are based on the results of discussions held with these officials.

(ii) A seminar was held on 27 November 1984 to consider the role of models in Phase I1 research and produce initial suggestions for research in the appropriate areas. The organizational structure of Phase I1 was also considered. MOE officials participated in this seminar. *

(iii)From 10-13 December 1984, the consultants carried ouk a training workshop on design and methodology for UPEL research and education. The target population was composed of: - all professional staff of the Research and Evaluation

' - representatives of divisions and departments involved in Division :

UPEL projects: - A.A.U computer programmers working with the task forces; - other concerned researchers.

- 3 -

(iv) The consultants examined and evaluated the work done by the national officials during Phase I of the project, and made recommendations for six studies to be conducted within the framework of the UPEL project. These studies are included in Annex I of the present report.

111. FINDINGS ON PHASE I OF THE QUALITY OF EDUCATION STUDY

l,2. In order to assess the work carried out on Phase I of the Quality of Education Study, agreements were made with the national authority to focus on the following problem areas:

c

(i) organizational framework of the Study,

(ii) objectives and research designs,

(iii) quality of instruments,

(iv) sampling frame and procedures,

(v) data capture, f ile-building and data analysis,

(vi) writing of reports.

In this regard, the consultants'findings were as follows:

(i) Organizational framework of the Study

13. The organizational framework of Phase I of the Quality of Education Study has been often associated with a research and evaluation model and structure, namely the ERGESE (Evaluative Research of the General Edu- I

cation System in Ethiopia). zational framework of the study with its various administrative and implementation bodies, task forces and sub-task forces.

This model clearly indicates the organi-

14. Originally, a system approach was used in the design of the ERGESE, but not in its actual operation. Moreover, the linkaqes built into the organizational framework were not applied to the other problem areas.

15. In the conduct of the Quality of Education Study, great emphasis owas laid on the development of an information base on the 'state of affairs' in education in socialist Ethiopia. The national commitment to the issue of the quality of education was very much in evidence throughout the organizational and administrative setting of the study. In terms of coverage (sampled units, research problem areas and personnel), the study may in many ways be regarded as unique. Nevertheless, from the perspective of research and evaluation, a complex endeavour of this type required trained and experienced personnel for the coordination and monitoring of the various phases of the study; particularly in terms of design and focus, choice of appropriate techniques and analysis of the data collected.

16. zational and administrative problems as to how the works of the respective task forces and sub+task forces were to be

&

In qualitative terms, Phase I of the Study faced severe organi-

a) b) co-ordinated in terms of research designs and objectives,

delimited in accordance with their areas of investigation;

Sam ling coverage, instrument of construction and data ana Y yses;

- 4 -

c) disseminated in the form of inputs and feed-backs;

d) incorporated into one plausible document for planning and policy-making, following the system approach that was designed at the very outset of the Study.

17. As a result, much of the work done on Phase I of the Study was hampered by the lack of a proper information network. Each different task force tended to adopt a sub-sectoral approach to the Study, with a high degree of overlapping, or duplicating, of tasks.

18. Consequently, the respective task forces were advised by the mission to focus on a few issues that were specifically related to the objectives set forth for their task force. will serve as input in the planning of the research and evaluation studies of Phase 11.

The results of this exercise 4

19. It was also suggested that a better notion of the system approach (its nature, role, and implications) be given to the teams assigned to Phase I1 of the Study during the training sessions.

(ii) Objectives and research designs

After a proper examination of the objectives and research designs 20. set forth for the respective task forces in the Quality of Education Study, it was concluded that Phase I of the Study was descriptive rather than inferential and/or empirically oriented. However, this distinction was not made clear, and as a result, some of the objectives and research designs seemed to call for a choice of instrumentsc data and analyses that were suitable for inferential purposes. This has led to a number of problems which are outlined below. It seems imperative in the planning of Phase 11, as well as in the short training sessions, to help national team members understand the nature, role, needs and implications of the various types of research design.

21. There was a general consensus that the objectives for the areas under investigation for each task force were too broad in nature and not properly defined. the implementation stages.:

This situation had the following results during

a) a vast amount of data was collected, b) cl the garbage in and garbage out process was employed in the

d) there were many invalid responses, as well as a quantity of

all possible methods of data collection were used,

sorting,screening and analysis of the available data and,

unreliable information and duplicated items. U

22. and research designs of the four task forces. The lack of feed-back generated here was clearly reflected in theplanning and organization of their investigations. different groups, which led to much waste of time.

23. No investigation in Phase I of the Study. defined the objectives, developed hypotheses, selected appropriate instruments, and chosen the right types of analysis. Therefore, it was recommended that the task forces should help develop such models in their analysis of the data from Phase I.

Problems also emerged in the simultaneous assessment of objectives

A number of objectives were duplicated by

real conceptual model was developed for the areas under A proper model would have

- 5 -

(iii) Quality of instruments

24. The types of data gathered in Phase I of the Study were conditioned by the factors mentioned above. The emphasis on both quantitative and qualitative methods of data collection was to have been accompanied by an appropriate choice of instruments and items that were suitable to these two distinct methods. This would also have facilitated the strategy for data analysis.

25. It was observed through some item-analyses that the instruments faced problems of:

a) clarity, b) format (depth and breadth)., and c) objectivity and reliability.

A major shortcoming of these instruments was that they were not tested on a pilot-study basis prior to full-scale investigatfon.

26. It was also realizalthat the amount of factual information/data (reports on facts) and attitudinal information/data (reports on opinions, beliefs, and feelings) required from the selected informants was too cumbersome, too detailed and too time-consuming. During the first dry-run analysis, many items had inconsistent and/or invalid responses.

27. In the absence of literature reviews, many results from studies previously carried out in Ethiopia were not used prior to the development of instruments. As a matter Qf fact, such reviews, particularly on Ethiopian studies, would have reduced the vast amount of unreliable items in the different sets of instruments. For example, information on annual costs and expenditures at school level were not available or were not properly recorded. This information was, however, available at higher administrative levels. Henceforth, questionnaires should be replaced by more qualitative data collected from school directors. The gathering of attitudinal information through open-ended questionnaires could also have been dispensed with, since unstructured interviews would have served just as well.

28. The duplication of items in the different sets of instruments was originally conceived as a control mechanism for cross-checkings and validity. However, the andytical design gave no opportunity for cross-checking. Informants! identification numbers were generally missing on the instruments and it was very difficult to relate item- responses from one set of informants to another. Furthermore, within a single given set of instruments, the analytical technique (simple calculation of items frequency distribution) did not provide an

recommended that in Phase 11, instrument constuction should be'closely linked to the data analysis phase.

, adequate test of validity and reliability. It has therefore been

(iv) Sampling frame and procedures

29. As mentioned earlier, there was a considerable national commitment to Phase I of the Study. This could be observed, for instance, in the amount of participation of Ethiopian professionals, researchers, educational planners and officers, along with school personnel at all administrative levels in the Study.

- 6 -

30, The principles and criteria for sampling were underestimated or ignored in the Quality of Education Study. On the other hand, much emphasis was laid on the numbers-game, 2.e. the selectfon of a sample with a large number of provinces (Awarajas) and schools, and with a wide category of informants (e.g. educational officers at all administrative levels, school directors, inspectors, teachers, students from primary and secondary education levels, heads of Awaraja pedagogical centres, community members and un2t leaders) at the expense of a proper sampling frame, representativtty, and use of less time and resources. During the discussion on the sampling frame and procedures, it was understood that very little attention was paid to sampling in preparation and design of Phase I of the Study. Special attention was therefore given to the teaching of sampling in the short training sessions: It was also advised that the consultancy team should help in develophg sampling designs for Phase I1 of the Study.

3 L - It was found that the majority of the task forces had selected over 1/3 of the 106 Awarajas in their sample. Was this necessary for making inferences about the population represented in the sample, i.e. the 106 provinces? Do the characteristics of these provinces vary that much across the country? Questions of this type were addressed during the meeting with the task force members. A different set of questions was raised with regard to the criteria used for sampling.

32. As the sampling of key informants for the respective task forces was based on:

(i) their functions and roles in the educational sector, (ii) their representativeness in terms of the educational .

hierarchy (central, district, community and school levels) , and

(iii) their common experiences in related spheres of activity, the sampling frame and procedures used had some simple technical and administrative weaknesses.

33. The way the instruments were designed ruled out any possibility of relating data from one category of informants to another, despite the fact that cross-checking of responses on a given topic from different categories of informants could have been an excellent test of validity.

34. The lack of identification numbers and an appropriate sampling frame did not allow this objective to be meet. Task force members were informed of this and were recommended to go back to their instruments and reorganize them in such a way that data from different sets of instruments and informants could be easily processed and analyzed.

(v) Data capture, file-building and data analysis

35. During the mission period, the task forces were still at an early stage of data processing, i.e. data capture, file-building and data analysis (with the exception of the task force on Education Adminis- tration, Structure, and Planning).

36. It was not very clear which types of data sets would require computer facilities and which would need simple techniques using

- 7 -

calculators. The linkage between tk staqescrf instrument construction and data analysis was not properly understood. Although great efforts were made to process some of the data sets at the Computer Centre of the Addis Ababa University, the local faci’lities (expertise, computer programmes-; and computer tfme) could not cope with the volume of data sets involved.

37. A number of items were instead tabulated on simple tally sheets. For example, the data collected by the task force on Education Admi- nistration, Structure and Planning was well tabulated with the use of simple calculators. The findings were then reported on the basis of univariates with frequency tables and percentage distribution.

38. In the process of sorting and screening out relevant and reliable data/information, it was recommended that dry-run analyses should start by:

(i> using simple descriptive tables, and (ii) developing constructs or composites from theoretically

relate& variables. In the former case, these tables would provide the necessary tools for discarding items with less that 80 per cent responses. In the latter case, the data reduction procedures would facilitate the grouping of variables into composites, and would provide opportunities for testing them empirically on the basis of factor analysis programmes.

39. It was also recommended that the common items from the different se& on instruments and categories of informants be grouped together for cross-checking purposes.

40. The above recoMendations were considered in the building of separate-and-merge files for data analysis in three stages:

(i>, univariates - one-way frequency tables with percentages, Cii) bivariates - two-way frequency tables with percentages

means and standard deviations,

in rows and columns, chi-square test results as well as simple correlation (Pearson r); and

if possible, multiple regression. (iii) multivariate analysis - analysis of Variance (ANOVA) and

This plan for data analysis was discussed in more detail by the other consultant. The information given to the task forces pertaining to this area was viewed as a useful exercise for planning the data analysis partof Phase 11, the UPEL research and evaluation studies.

41. In the data analysis stage, very little attention was given to the analysis of qualitative data. The pressure for quick computable results and the lack of competence and skills limited the scope for using qualitative methods in the data analysis (e.g. field reports from in-depth interviews and observations, content analysis of curriculum programmes, case study narratives and the like). The idea of laying more emphasis on qualitative methods of research in Phase I1 of the Study was much favoured.

I

(vi) Writing of reports

The reporting exercise could not be discussed in detail during the 42. period of the mission, as three of the four task forces were still at an early stage of data capture, file-buflding and data analysis. However, it was possible to discuss a tentative structure for the writing of their reports, A common outline was suggested which included the following:

introduction and summary, objectives, selected policy-questions, structure of the report, limitations, literature reviewed and conceptual framework, methodology - sampling and target-population, instruments and data collection techniques, findings, and conclusion and recommendations.

c

43. The task force on Education Administration, Structure and Planning had their reports nearly finished by the time of the mission.

44. It was found that the results, recommendations and conclusions from this task force were conditioned by the same factors (conceptual, methodological and analytical) confronting the other task forces.

IV - RECOFLXENDAT IONS A. Phase I

45. The following major recommendations on Phase Iof tkcruality of Education Study were presented to the task forces. These recommendations were also presented during the short intensive training workshopfor the sake of quality improvement in the research and evaluation

(i) studies of Phase 11. It is recommended that the system approach (ERGESE) originally designed for the Study be improved in the data analysis and reporting stages. A feasible climate for dissemination of inputs, feed-back, and outcomes from each task force and sub-task force should also be created. In addition, seminars and workshops should accommodate for all groups involved in the planning organization adm-inistra- tion, and implementation of the Study. At this late stage, there should be no more room for a sub-sectoral approach, and so a better notion of the system approach and model-building work ought to be emphasized during the short intensive-training workshop. It is strongly recommended that a more streamlined strategy be adopted in the examination of the broad study objectives. These objectives should serve as a point of departure in planning the analysis and in reporting results.

,

- 9 -

,

(iii)

(vi I

Consultations among respective task forces and sub-forces are essential. Qualitative rather than quantitative aspects of the data/information should be stressed. The absence of proper research designs ought to be reconsidered in the planning of Phase I1 studies. Clear-cut distinctions should be made between the different research designs and their purposes. These distinctions should be accompagnied by the choice of appropriate sampling, relevant instruments, target groups, and data analysis methods. It is recommended that literature reviews accompagnied by separate conceptual and statistical models be included in the reports of the task forces. loped for each task force with the assistance of the Unesco consultants during their missions period, or even eventually by the national specialists involved. Each task force should examine the output from the different instruments used, with a view to sorting and screening out those items that are not relevant to the sets of objectives outlined in their study. It was observed that a significant amount of data/information collected was irrelevant to the study. With regard to the quality of instruments, it is recommended that item-analysis be Performed and variables be organized along the lines of concepts/constructs for further analysis. A special distinction should be made between the analysis of data bearing upon factual information and data bearing upon attitudinal information. Likewise, emphasis should be laid upon the reporting of both structured (closed-ended) and unstructured (open-ended) items from the different sets of instruments. The strategy to convert all data/information into numerical forms for computerization should be limited to closed-ended items only.

Such models would be deve-

(vii)It is also recommended that the cross-checking of items from different categories of informants be given a low profile in the data analysis stage, as the necessary data capture prerequisites for such an exercise (e.g. recording and sorting of instruments by identification field -ID number) were missing in the preparation of the instruments.

areas (awarajas, districts, communities and schools) was given, the sampling principles were not observed. The sampling designs should have considered the population referred to, and its basic characteristics. Weighting of sampling strata should be carried out whenever over- representation and under-representation occur. The type of sampling procedures should have been made clear- (purpose, evaluation,- or possibilities proportional to size, etc.). It is recommended that sample weights be considered in the analysis where the sample as planned differs significantly from the sample as executed.

(ix) Instruments construction and data analysis should have been linked during the designs of Phase I of the Study. A trade-off is recommendedin the data capture, file-building and data analysis stage as to which types of information

(viii)Although the general frame for sampling of geographical

- 10 - should be analyzed through simple routine methods (tallying and computing with calculators) and which ones should be processed through computer facilit2es. In the former, it 2s recommended that selected tables with frequency distribution be presented; in the latter, a few bivariate and multivariate analyses ought to be carried through statistical computer Frogrammes. It is strongly recommended that for Phase I1 Study, statistical packages (e.g. SPPS available at the Central Statistical Office) be used in the processhg of the quantitative data. In their reports, the task forces are recommended to distinguish areas bearing on policy-making decisions from those which refer more generally to the status quo of education in Socialist Ethiopia. They shouldalso appraise the issues and supply recommendations.

(xi) It is recommended that frequent seminars and workshops be held, with a view to establishing close working relationships between the several research and evaluation studies in all fields.

(xii)It is recommended that funds from research endeavours be earmarked for the nationals (Ethiopian researchers and planners) in order to improve their skills at all levels.

B. Phase I1

STUDY No.l:Equality of educational opportunity in primary schooling

46. The consultants recommend that this study should identify the characteristics which differentiate school-age children who are unlikely to enter primary school from those who tend to enter the formal education pipeline at an earlier stage.

47. These characteristics should be examined under three levels:

- individual (e.g. sex, age, health) - family background (e.g. socio-cultural, language and economic) - social/geographi'cal environment (e.g. urban/rural; Amharic/ non-Amharic, literacy level, local school).

48. hinder access to education, the recommended research procedures for this study are as follows:

With a view to identifying the conditions which facilitate or

49- The study, together with study 2, should be launched as a priority measure.

50. The fiefdworkers of the study should be trained to carry out fieldwork. This training should be timed to coincide with testing of the f ieldvork instruments.

51. The computer should be used throughout the Project to cut down on time and cost as well as to increase precision and efficiency.

STUDY No.2:Linkage between formal and non-formal education

52. The consultants recommend that this study should examine two specific groups:

- 11 - (a) Those who are of school-age (7-16) but are not yet in school. (b) Neo-literate adults (16-491, with a view to identifying the

reasons for their non-participation in formal/non-formal education and their further educational needs.

53. This study should be a complement to Stud.v Fo. 1 and its design should be the same.

STUDY No. 3: Primary education performance and retention

54. The consultants recommend that this study should examine the factors which facilitate or hinder pupils' completion of the primary educational cycle and investigate different levels of proficiency in the-three R' s.

55. As before, the computer should be used throughout the Project.

STUDY No. 4: Quality improvement in primary education

56. The consultants recommend that this study should evaluate the ~

effects of human and material resources on the quality and results of the teaching-learning process at all levels, with a view to identifying the role of school inputs therein.

57. The consultants recommend that those who are carrying out this study should be shown how to conduct a cost-effectiveness study. fieldworkers should be trained, as above, at the same time that the field instruhents are being tested.

The

STUDY No. 5: From plans to actions. A study of planning, implementation and management processes. Obstacles and needs at the Awaraja and school level.

58. The consultants recommend that this study should examine the objectives set for Awaraja Pedagogical Centres and assess the relations between them from the standpoints of planning, management and implementation.

STUDY No. 6: Effects of primary schooling in local community development (A SIDA-supported research study)

59. The consultants recommend that this study should identify and examine the multiple roles that education can and should play in individual, family and community development, e.g. child-care, health and nutrition, small-scale production, sanitation and agriculture. I

t C. Further recommendations

60. The Ethiopian research staff should be given further training in modern research and methodology in many other areas. These areas would include survey and sampling methods, data processing and analysis, and measurement sciences such as intermediate statistics, psychometrics and econometrics. Some individuals could be trained abroad in a research environment. Others could be given on-the-job training by consultants.

61. The UPEL Project and the statistical survey unit must be e%ipped with hardware and software for statistical processing.

- 12 - 62. The library at the Research Centre and the statistical survey unit should be equipped wfth modern research literature.

63. Phase I1 must be provfded with consultancy at the initial stage and supplemented by a follow-up at the intermediate and final stages. Consultancy should definitely not be a salvage operation at the end stage when it is too late to change anything signlflcantly.

64- over the duration of the Project.

The Phase I1 consultancy period should be 3-4 months, spread

V - CONCLUSION

65. All of the six studies discussed above should be conducted within the framework of the Second Phase of research and evaluation, in connection with the UPEL. The subjects of these studies have been chosen after a documentary review of the UPEL documents and objectives, along with an examination of the major problem areas that may be encountered in attempts to reach these Objectives.

66. This approach shows how the studies are related in terns of educational policies and systems. purpose and desired outcome of each study and presents an outline of the overall research design.

It also highlights the specific

The rationale for the UPEL Research Programme

67. Obviously many areas of educational research deserve support, so the choice of a limited number of areas has to be fully justifiable. The six areas outlined above were chosen for the following reasons:

UPEL objectives in the Ethiopian context

58. As the context of this research programme is the Universalization of Primary Education and Literacy (UPEL) in Ethiopia, it is essential that its recommendations be consonant with the aims of the UPEL. A number of important gssues in Ethiopia and the UPEL were reviewed and discussed at the Workshop on the UPEL held at Ambo (29 March - 2 April 19841. The following documents were reviewed in order to identify areas in which research might contribute to attainment of the objectives:

(i)

(ii) The Ethiopian UPEL Workshop documents (UPEL/ETH/84/DOC/Ol) ,

(iii)Report on the Workshop on UPEL (MOE, Ambo, 29 March -

Unesco-UNICEF documents on the UPEL Co-operative Programme (UNICEF E/ICEF/Misc. 401, 11 April 19831,

and

2 April 1984).

firthermore, in connection with the mission's work, Phase I of the Research and Evaluation of the Quality of Education in Ethiopia was reviewed in order to maximize the transfer of useful experiences and findings and minimize duplication.

Policy issues and the educational cycle

69. 'Six areas of the educational cycle have been delineated in terms of systems parts, as illustrated in the following diagram:

- 13 -

Sy8t.m. Approach to UPEL Research Ptogtnme

1

r i

rial resource allocation

implanenta- tion

3

rmance L re-

f aclli tat i on- 'U- 2

I ob.rruction1 2 1

-> learning

70. For greater precision, policy relevance has been used as a criterion. The extent to which this research programme can make any contribution toward the attainment of the UPEL objectives is greatly dependent (il. on the relevance of its questions to the policy issued, and (ii) of its findings to the o~erational level of nolicy implementation. To ensure that the research findings will be useful in decision-making, the policy issues involved in each of the recommended research areas must be fully explained.

71. Here it must be noted that some subjects may overlap with those covered in the Phase I of the Research Programme. Where there is an overlap, Phase I should be used as a Pilot Study for the Phase I1 Research Programme.

Differential accessibilitv of education

72. The problem of the differential accessibility of primary education for children who are just reaching school-age is a key one in terms of UPEL objectives, and it must be taken into account in any plans to achieve universal access. There must be research in this area if we are to identify and understand what helps and what hinders equal access to primary education. Such knowledge is indispensable for specifying the target groups and the selection criteria for appropriate corrective measures.

- 14 - 73. The importance of this phase to the UPEL is strongly indicated by the following evidence (M. Selowsky, 1980, 97-112; as summarized by Simmons, 1980: 11);

(a) growing empirical evidence shows that pre-school children of families of low socio-economic levels score worst in most ability tests:

(b) to a large extent, these low scores are the result of poor diet and deprivatlon;

(c) future school enrolment will increasingly draw children from deprived socio-economic levels. This means that the future effectiveness of primary schooling will depend on present * policies aimed at boosting children's school performance by raising the quality of their early environment and nutrition.

Marginal groups and the linkage between formal and non-formal education

The educational needs of the many people who are outside the 74. formal system constttute another critical area for research. It is recommended that two particular groups be specially studied:

(a) people of school-age (7-161 who are not in school, and Cb) the increasing number of neoliterate adult (16-49) graduates

produced by the literacy campaign.

This area needs further examination for a coherent policy of support for special educational services aimed at providing continuity between formal and non-formal education.

Performance and retention within the primary school

75. At primary school, the attainment by pupils of basic skills within an efficient system is central to the attainment of UPEL objectives. A certain knowledge of the mainsprings of educational attainment, repetition and attrition is required if proper corrective measures for the reductionof wastage and for ensuring minimum learning continuum at primary level are to be formulated. Hence, the study should be clearly focussed on the progress of pupils between grades.

Cost-effectiveness of resource input in bringing about aualitative immovement in education

76. Any further expansion of the primary sector implies hard decisions on the allocation of resources, since resources are very scarce.

77. In this regard, the cost-effectiveness of human and material resources in bringing about qualitative improvement in both teaching and learning must be carefully considered. Thus the study should evaluate the effects of resources on the quality of teaching and pupils' learning outcome, within the context of Ethiopia's projected new curriculum for the years ahead.

- 15 - Curriculum implementation structure

78. The impact of any policy ultimately depends on its implementation at the target level. An important aspect of educational innovation in Ethiopia is the link between the National Curriculum Development Centre (NCDC), and the Awaraja Pedagogical Centre (APCJ and its counterpart at the school level (SPC). Thus, it is vital for the proper implementation of the UPEL that a thorough assessment be made of NCDC-APC-SPC links. A study should be made of APC in-service training and the development of teaching alds, with an evaluation of their effect on the SPC and, fn turn, on classroom environment and teaching practices. This study would assist policy-makers in identifying practical difficulties and help them make the necessary adjustments.

The imDact of education.on the communitv

79. Education should serve the heeds of individual, the family, community and society. Hence, the success of the UPEL can be judged by the effect of its work pupils after they have left school.

80. Thus it is recommended that a study be made to examine the effects of primary education,both on participants and on community life. The results of such a study would help policy-makers identify new ways in which primary education could benefit individuals, families and communities. Old age, childcare, health, nutrition, sanitation, agriculture and small-scale production are only a few of the areas in which progress could be made and to which progress must be made in Ethiopia.

- 16 - ANNEX I

UNIVERSALIZATION OF PRIMARY EDUCATION

RESEARCH AND EVALUATION STUDIES AND LITERACY (UPEL) PROPOSALS FOR PHASE II

Page

S m a r y . . . . . . . . . . . . . . . . . . . . . Overview of UPEL studies . . . . . . . . . . . . A presentation of the six research and evaluation

studies

STUDY 1:

STUDY 2:

STUDY 3:

STUDY 4:

STUDY 5:

Selected

of UPEL - Phase I1 . . Equality of educational primary schooling . . .

. . . . . . . . opportuntty in . . . . . . . .

Linkage between formal and non-formal education . . . . . . . . . . . . . . . Performance of primary education system

Quality improvement

From plan to action

list of references .

in primary education.

. . . . . . . . . . .

18

20

28

28

48

66

74

84

91

- 17 -

PREFACE

Universalization of Primary Education and Literacy (UPEL) is a top-priority goal for educational policy and planning in Ethiopia. To reach the UPEL objectlves, educational research has a strategic role to play. the yet unknown contours along the way to UPEL, monitoring the current progress, evaluation of past experiences, and planning the future. To this end, both Unesco and UNICEF have promoted educational research and evaluation in Ethfopia, with particular emphasis on the long-range objective of universalizatkon of primary education and literacy.

It can provide the vital information necessary for mapping

As part of this effort, one of the major tasks of the Unesco Consultancy Team (Drs. Ko-Chih R. Tung and V. Chinapah) has been to plan, prepare and submit proposals for research and evaluation studies to the Ethiopian Ministry of Education in the field of universalization of primary education and literacy. These UPEL studies constitute the second phase of a research programme known as the "Research and Evaluation of the Quality of Education in Ethiopia".

The report contains six research and evaluation study proposals. Five of them are presented in detail, while the sixth one is a joint SIDA-Curriculum Department proposal, details of which are, therefore, not included in this report. As for the authorship of this report, the two consultants have been assigned special tasks. Dr. Chinapah was generally responsible for the study description sections (i.e., from "Purpose" to "5. Short Literature Review"), all of Study 5 proposal and section 7 of Study 1. Dr. Tung was responsible for the research design sections (i.e. from "6. Parameters of research design" onwards, except section17 of SfudyJ), all of Study 4 proposal, and the introductory overview chapter. The summary was written jointly.

- 18 - SUMMARY

The following outline provfdes a summary of each of the research and evaluation studies proposed under the UPEL umbrella.

STUDY 1:

The main purpose of this study is to identify at an early stage of life among those who are just reaching school-age, the characteristics which differentiate those children who have low probability of entering primary school from those who are more likely to enter the formal educational pfpeline. Three levels of characteristics will be examined: individual (e.g. sex, age, health), their family background (e.g. socio-cultural, language and economic) , and their community (e.g. urban/rural, Amharic/non-Amharic, literacy level).

The desired outcome of this study is a better understanding of the conditions that facilitate, and those which hinder, equal access to primary education. The knowledge gained in this study will be useful for the identification of the target groups and selection criteria for corrective measures, such as positive discrimination and compensatory/remedial education, strategies for the reduction of home and neighbourhood handicaps, etc.

STUDY 2: Linkage between formal and non-formal education - An in-depth study of the margfnal groups' educational needs.

This is an in-depth study of marginal gfoups. Two specific categories will be examined: (a) those who are of school-age (7-16) and yet are not in school, and (b) nediterate adults (16-49). The main purpose is to identify (iJ the reasons for their discontinuance and (ii) their futther educational needs.

The desired outcome of this study is that its findings will be used for the identification of critTcal areas where appropriate - educational services could be supported,. extended and integrated with services of other sectors in the community in order to serve the needs of the marginal groups, and to suggest ways of strenghtening the linkage and continuity between formal and non-formal education.

STUDY 3: Performance of the primary education system - A study of the nature, extent and causes of educational wastage and the attainment of basic skills in primary schools.

The overriding aim of this study is to identify and to examine factors which facilitate of hinder pupils' completion of the primary educational cycle and to investigate the different levels of proficiency in the basic skills of reading, writing and calculation.

The desired outcome of this study is to provide empirical evidence for an identification of possible corrective measures for reducing educational wastage, and for ensuring minimum learning continuum through the primary education cycle.

- 19 - STUDY 4 : Quality improvemen't in pr'imary. edu'c'ation. '-' 'The e.ffe.ct of

'human a.n'd .m'ate.ri'al 're:s.ooil'r%'e:s. an' 'the 'cru'a:li'tv 'of 'teachina and learning outcomes.

The purpose of this study 2s to evaluate the effects of human and material resources on the quality. of the teaching-learning process, and learning outcome in the children's cognitive, affective and psychomotor development in different classroom settfngs.

The desired outcome of this study is an identification of major quality control inputs related to the teaching-learning process that significantly affect primary pupils'learning outcome, and identification of the most cost-effective set of policies in regard to acquisition and allocatlon of resources and designing of teacher training in relation to different target groups. An important indirect outcome will be the strengthening of scientiffc evaluation research as an important tool in the development.of the Eew curriculiim.

The purpose of this study is to examine the relationship between the planning, implementatlon and management processes, obstacles and needs for primary schooling,. especially in regard to the relationship between the APCs and the SPCs.

The desired outcome is to provide information on the implementation of the objectives set for these two institutions; i.e., the input- output relationshipbetween APCs and SPCs; and to assess the importance and relevance of in-service trafning anddevelopment of appropriate technology to classroom implementation through the application of knowledge and skills of APC trainees.

STUDY 6:

The purpose of this study is to identify and to examine the multiple roles that education can and should play in individual, family and community development (e.g. childcare, health and nutrition, small-scale production, sanitation and agriculture).

The desired outcome is to provide some insiqhts into the differential effects primary edication has on its participants and community life. To provide particular information on re-organization of the content and method of the primary curriculum with respect to health and nutrition in order to better fitcommunity needs and improve practices';

- 20 - OVERVIEW OF UPEL STUDIES

1. INTRODUCTION

The Consultancy Team has been working in joint collaboration with the Curriculum Department of the Ministry of Education in Ethiopia in preparing the research and evaluation studies under the UPEL (Univer- salization of Primary Education and Literacy) umbrella. In all, - six studies, covering subjects ranging from primary educational opportunity to the impact of primary schooling on the local community, are herewith recommended to be conducted wtthin the framework of the Second Phase of research and evaluation in connection with the UPEL. These particular subjects have been chosen after a documentary review Of the UPEL documents, especially in regard to the UPEL objectives and the major problem areas that may be encountered in attempts to reach these objectives. It is hoped that the findings of these studies will provide vital information for formulating policies toward reaching the UPEL objectives.

the viewpoint of policy fssues along the educational cycle from a system perspective; provides informatfon on the specific purpose and desired outcome of each study; and presents an outline of the overall research design.

This overview explains the relationship between the studies from

2. THE RATIONALE FOR THE UPEL RESEARCH PROGRAMME

Therg are, of course, mary areas of educational research which deserve support. Hence, a selection of particular subject areas,. while neglecting many others, requires plausible justification. It is the intention of this section to provide the rationale behind the selection of the six areas to be recommended.

2.1. The UPEL objectives in Ethiopian context

of Primary Education and Literacy (UPEL) in Ethiopia, it is essential . that the recommendations be consonant with the aims of the UPEL. A number of important issues related to the educational objectives of Ethiopia and the UPEL have been reviewed and discussed at the Workshop on the UPEL held at Ambo (29 March - 2 April 1984).

As the context of this research programme is the Universalization

2.2. In preparing these recommendations, therefore, the following documents have been reviewed in order to identify relevant areas in which research may contribute toward the attainment of the objectives:

- UNICEF: Document of UPEL co-operative programme (UNICEF E/ICEF/ Misc.401, 11 April 1983),

- The Ethiopian UPEL Workshop Documents (UPEL/ETH/84/DOC/Ol), - Report on the Workshop on UPEL (MOE, Ambo, 29 March -

2 April 1984).

2.3. The original terms of reference for the Consultancy Team were revised in order to make maximal use of thefr time and experiences for (a) Phase I of Research and Evaluation of the Quality of Education

- 21 - Project in Ethiopia which was the major assi-ent of Dr. Ko-Chih Tung; and (b) the identification and planning of research and evaluation studies under the UPEL's umbrella, which was the major assignment of Dr, Vinayagum Chfnapah. In the report-writing, both consultants were instructed to give priori'ty to the Research Proposals for the UPEL studies. The leading staff of the MOE'S Curriculum Department, especially those connectedwith fts research programme, and the representathes of Unesco and UNTCEF, have been interviewed. Mr. A. Idris of Unesco, Mr. Alan Court of UNICEF, Ato Tesfaye Dubale (Head of the Curriculum Department), Ato Tekle Ayano (Head of Research and Evaluation Division of the Curriculum Departmentj and Dr. Barbara Junge (UNICEF project officer] have been particularly helpful in this preparatory work.

2.4. The recommendations presented here are a distillation of the contributions of many people, including those who have participated in the UPEL Workshop and those who have been interviewed specifically for this purpose.

3. SYSTEM APPROACH TO THE EDUCATIONAL CYCLE

3.1. To delineate the subject areas to be studied within the research programme., it is useful to employ a conceptual framework, one which is global enough to capture the entfre field but specific enough to indicate concrete sub-areas. Among the many models for conceptual organization of educational research,, the one presented by John Simmons (1980: 19-66) is particularly interesting, because it is a model employed to generate a discussion on "an Overview of the Policy issues in the 1980's'' in an anthology entitled "The Education Dilemma: Policy issues for Developing Countries in the 198G's'l (Simmons, Ed., 1980).

3.2; The influential factors in Simmons' model, which encompasses formal, non-formal and informal education, are orqanized in chronoloqical order. environment of pre-school chfldren, including such factors as health

Hence, the starting point of this learning system is the early

(infectious disease, parasitic disease, malnutrition), personality (skills, intelligence) as well as the usual family (income, occupation, child rearing behaviouf) and non-family (production mode, rural/urban, peer group, legal and social sanctions) environmental factors. The intervening phases are influenced by "later factors", namely schooling, non-formaland informal learning, which produce behavioural change in areas pertaining to employment socialization and politicization. At the'butcome'' level are the assumed benefits, such as productivity, participation, etc.

3.3. Such a chronology is also implied in the systems approach employed here (see Figure l), which, however, differs in some important aspects. One is that Simmons' model, being more concerned with the economic aspects, puts more emphasis on input and output rather that the in-between processes. The approach employed here is amenable to an examination of the content and process of schooling as well as the input-output factors. The other is an extension of the educational impact. As the ultimate objective of education is not only a modification of the individual, but also to improve the environmental factors, the outcome factor has been extended to include the impact of education upon the community.

c

i

- 22 - FIGURE 1: SYSTEMS APPROACH TO UPEL RESEACH PROGRAMME

Human and Material

Envl ronmental /community faci 1 i tation- 0 bstructi on

Formal Performance I and Retention

I v

Study 2 1 I ,I Nonformal and I

Study 6

3.4. In order to delineate systematically the areas of research and eventually to integrate the research findings, the approach employed here is to focus on the critical phases of the educational cycle, starting from differentfal access to education as the initial point, and the impact of education on the community as the end phase of each cycle, between which are the htervening phases.

4. POLICIES ISSUES AND THE EDUCATIONAL CYCLE

4.1. Six areas of the educational cycle have been delineated in terms of systems parts, as illustrated in the diagram above. To achieve sharper focus, --policy-relevance has been employed as a criterion. The extent to which this research programme can make any contiibution toward the attainment of the UPEL objectivesis greatly dependent on the relevance of its questions to the policy issues and its findings to the operational level of policy implementation. the research findfngs will be relevant to decision-making, incumbent upon those making the recommendations to define the policy issues involved in each of the recommended research areas.

To ensure that it is

b

Note. Some of the subjects may seem to overlap with those covered in Phase T of the research programme. and, where there is an overlap, Phase I is to be used as a feedback for Phase I1 research Frograme.

However, the emphasis is different

- 23 - 5. TH3 SIX XESEALICH AND STUDY AYEAS

5.1. Differential accessibility of education

At the gateway to the formal educational system, the first question to be asked pertains to the entry phase of primary education. Of immediate relevance to the UPEL objectives is the problem of differential accessibility of primary education for children who are just reaching school-age. For the policy-makers, this aspect constitutes a major complex problem which must be taken into account i’n order to formulate appropriate measures for fac2litation of universal access. Research in this area is required in order to identify and better understand the conditions that hinder and those which facilitate equal access to primary education. Such knowledge is indispensable for specifying the target groups and the selection criteria for appropriate corrective measures.

The importance of this phase to the UPEL is strongly indicated by the following set of evidence (M. Selowsky, 1980: 97-112; as summarized by Simmons, 198O:ll) :

(a) growing empirical evidence shows that pre-school age children of families of low socio-economic levels tend to score less in most ability tests; to a large extent, these low scores are a product of a deficit after birth in environmental stimulation as well as nutritional intake, not genetic deficiency; and future school enrolment will increasingly draw children from these socio-economic levels. This means that the future effectiveness of primary schooling will depend on the present policies aimed at boosting - via a better quality of early environment and nutrition - the ability scores of those children.

(bl

(c)

5.2. Marginal groups and the linkage between formal and non-formal education

Outside the formal educational system, the educational needs ofthe vast number of people who are not in school constitute another critical area for research. specially studied: (a) the school-age (7-161 cohort who are not in school and (b) the increasing number of neo-literate adult (16-49) graduates of the literacy campaign. Further knowledge in this area is required for the formulation of policy in regard to the provision of support for special educational services and the integration of the latter with other service sectors and to strenghten the linkage and continuity between formal and non-formal education.

5.3. Performance and retention within the primary school

Within the primary school setting itself, pupils’ attainment

It is recommended that two particular groups be

of basic skills and further reduction of educational wastage tition and attrition) are essential to the attainment of UPEL objectives. attainment, repetition and attri’tion is required in order to formulate policy in regard to corrective measures for the reduction of wastage, and for ensuring minimim learning continuum through the primary education cycle. Hence, it is recommended that a study focus upon .

aspects related to the between-grades progression of pupils;

(repe-

Knowledge on the nature, extent and causes of differential

- 24 - 5.4. Cost-effectiveness of resource input in bringing about

qualitative improvement in education

In planning for further expansion of primary education while constrained by resource shortage, policy-makers are inevitably confronted with the difficult problem of making decisions on the acquisition and allocation of resources that would, hopefully, have optimal effect as intended. It is vitally essential for informed decision-making to acquire more knowledge on the cost-effectiveness of various human and material resources in bringing about qualitative improvement of teaching and learntng outcomes fn different educational environments. It is therefore recommended that a study focus upon the evaluation of the effects of resources on the quality of teaching and pupils' learning outcome in the context of the planned extension of the new curriculum, which has been adopted as the model to be implemented throughout Ethiopia, in steps, over the coming years.

5.5. Planning, management and implementation

implementation at the target level. An important transmission line for the dissemination and implementation of educational innovation in Ethiopia, as conceTved when it was laundhdin 1977, is the link between the National Curriculum Development Centre (NCDC) and the Awaraja Pedagogical Centre (APC) and its counterpart at the school level (SPC). To the extent that implementation is dependent upon this transmission line, it is all that more important for the implementation of the UPEL to get an assessment of the APC-SPC links.

The impact of any policy is ultimately dependent upon its actual

It is recommended that a study focus upon the evaluation of the areas in which, and the extent to which, APC in-service training and its development of teaching aids have had impact on the SPC=and, in turn, on the actual classroom environment and teaching practices. Addressed to the policy-makers, such a study should identify the bottlenecks along the implementation chain. This information should lead to the recommandation of policy relevant to the particular nature and location of the respective bottlenecks in order to ensure more effective implementation of educational innovation.

5.6. The impact of education on the coinmunity

The intention of schooling is, needless to say, not education for education'sake, but to serve the needs and to improve the conditions of livelihood and enlightenment at the level of the individual, the family, the community and the society. Viewed from this angle, a measure of progress toward the objectives of the UPEL is the impact of schooling on life beyond the school. It is hence recommended that a study examine the effects of primary education on its participants and on community life. Addressed b the Folic$-nakers, such a study should produce recommendations on the ways in which primary education can further the development of the individual, the family and the community in matters pertaining to e.g., childcare, health, nutrition, sanitation, agriculture and small-scale production.

To the extent that education has any impact on the quality of life, it has immediate relevance also for the differential accessibility of education at the entry point, where this review of policy-

-

- 25 - relevant research areas started. One of the few generalizations in educational research that Is almost universally valid is that. children of poverty-stricken environment, and suffering from maldevelopment due to nutritional and other defIclenc2es, are less l5kely than other children to enter school and, among. those who are in school, may remain long enough to complete the pr2mary education. complete.

The cycle is now

6. OVERVIEW OF TKE RESEARCH DESIGN

As explained above, the complex processes involved in an educational system have been dfssected fnto six identifiable phases in an educational cycle. These phases represent different settings and different processes. Hence, different populations C"actors"l and different phenomena are involved in the resnectiveprocesses. They

measurement, etc. Yet, the systems perspective provides a framework for a systematic research design. In specifying the design, it is important to bear in mind that even as the complexity has been decom- posed into separate parts, it must enable a reconstruction of the whole from its parts. In order to ensure such a possibility, there must be an overall design within which the sub-studies must fit. It is this overall design that will be presented in this section.

I require different approaches on matters pertaining to sampling,

6.1. Parameters of the research design

All of the studies to be carried out within this research programme are cross-sectional, representing a cross-section of the respective population in Ethiopia. To infer about phenomena over time, such as trends and changes over time, there must be observations at different time points. Such observations are absent in these studies. This does not mean, however, that they do not allow infer- ~

ences about processes. Tf It is observed that those who have been exposed to X-stimulus behave differently from those who have not been thus exposed, and that other possible causes for this difference can be ruled out, then it is possible to attach greater confidence to the attribution of the "cause" of the difference to X. It is to maximize the validity of such inference that the competing sources of variance must be controlled.

6.2. Basic model

The overall model has been presented above (Figure 1). Each of the sub-studies will focus upon a particular part of this model. Hence, Study 1, for example, will investigate the two different paths eminating from environmental facilitation and obstruction. This represents a global model specffying the major processes; In the following presentation of the sub-studies, the model will be further elaborated in respect to the pertinent phases in the educational cycle.

6.3. Populations and sampling

to the settings, there are two groups of studies: those which are based upon only the school-based population (Studies 3-5) and those based on the community/household populat2on (Studies 1, 2, 6). The latter category includes those who are wlthin as well as outside the school.

Different elements are involved at different phases. In respect

As the two groups of studies pertain, at the individual level, to different populations, they require different samples and, hence, different data collection procedures. Fieldwork must be organized

- 26 - accordingly. Some studies will involve only household survey, while others will involve within-school data-collection.

Inasmuch as the major processes are related to each other, the sampling at community level must be co-ordinated. In other words, samples of households and schools should be drawn from the same communities.

Multi-stage sampling will be employed. The first stage consists of sampling at the Awaraja-level. The sampling-frame for the first stage is a list of all Awarajas stratified by the following variables:

(1) Residential ecology: urban/rural (2) Literacy level: ahove/below national median (3) Dominant lingo-ethnic group: Amharic/non-Amharic (4) School environment: has none/at least one "pilot" school.

These 4 dichotomous variables define sfxteen strata. From each stratum, a sample of Awarajas should be drawn according to probability, proportional to size. From these Awarajas, the household and school samples will be drawn according to instructions as stated in the respective studies below.

Note. - corresponding computer-based files should be maintained for sampling as well as for adjustment for sampling desing and analysis of the relationships between the ecological variables with other variables.

All sampling-frame data should be systematically collected and

6.4. Data-collection instruments and variables

In each of the proposed studies, there will be references made to types of data-collection instruments and variables. All of these are only recommendations indicating the types of instruments and variables that may be used to carry out the recommended studies. Within the time limit of this consultancy, it is not possible to actually construct the instruments. In all cases, adaptation of existing and well-tested instruments is preferable to using something yet untried.

In order to enable integration of the data as well as the findings from each of these studies, it is essential that the instruments contain identification system and codes that are standardized for all studies. An Awaraja will be identified by one particular code that will be used throughout all the studies and will not vary from study to study.

7. THE SIX RESEARCH STUDIES' PROPOSAL

For each study with the exception of Study 6, a detailed'proposal is made. The proposals consist generally of the following aspects:

Cl) Title of study (2) Purpose (31 Desired outcome (4) Background of study (5) Aims and object2ves. (6) Structure of study (71 Scope and limitations

- 27 - Short literature review Parameters of the research design Basic model Variables and instruments Strategy of analysis Sampling procedure Operational overview Note on computer usage

- 28 -

A PRESENTATION OF THE SIX RESEARCH AND EVALUATION STUDIES OF UPEL PHASE IT

Purpose

To examine equality of educational opportunity from the standpaint of school participation and scholastic achievement of children from different age, sex, family, lingurstic, cultural, socio-economic, regional, and school backgrounds.

Desired outcome

To provide a better understanding of the underlying factors that facilitate or hinder equal access and success in primary education. The knowledge gained throughout this study will be useful for (a) the identification of school-age children belonging to the "difficult-to- reach populations"; (bl the design of possible positive discrimi- natory measures to remedy inequalities among the educationally disadvantaged groups of school-age children; and (c) the examination of specific home, neighbourhoodand school factors that are statistically significant in determining variations in scholastic achievement of primary school children.

Note. - The theoretical and methodological approach for a research study on equality of educational opportunity emphasizes on both dlfferential access to schooling and on.determinants of scholastic performance. (See section 5.) However, for operational purposes, the other studies under the UPEL's umbrella have to be taken into consideration. We therefore suggest that the part dealing with determinants of scholastic performance in Study 1 will be based on the research design, sampling, instrument construction and data collection from Study 3.

1. BACKGROUND OF THE STUDY

1.1. In its Ten-Year Perspective Plan (TTYP), the Government of Ethiopia attaches significant importance to both formal and non-formal education for the development of the country's great human resource potential. Within a broader framework for democratization of education in Ethiopia, equality of educational oppbrtunity in primary education is considered as one of the priority areas. The Ethiopian TTYP aims to attain equity fn the provision of educational opportunity through (a) the provision of primary education for rural areas not having other means for the provfsion of education, and Cbl the provision of adequate primary school buildings.

- 29 - 1.2. Despite the enormous educational efforts made after the revolution in 1974, there fs still a lot to be done in order to redress educational imbalances amoung the disadvantaged groups of rural dwellers, females, certqtn nationalities and nomads iin remote rural areas. The system of education (both formal and non-formal] was no more than a reflection of the di'smal socto-economic conditions and political realities. characteristic of the system.

Discrepencles and inequalities were

"The revolution of 1974 revealed a system of education wholly inadequate for the needs of a nation. In order to brtng the action into the modern world, rt was necessary to bring masses previously excluded. In particular, fmbalances had to be redressed in favour of the disadvantaged groups or rural dwellers, females, certain nationalities and nomads in rural areas. The formal and non-formal education both had to be rapidly expanded to provide a basic, functional education and to begin to develop the great human resource potential of the nation." (MOE-Education Team Report, 1984, p. 10.1

1.3. The present study will.attempt to account for the degree of educational inequalities amoung the primary school-age population in Ethiopia, and will attach particular importance to inter and intra regional differences, age and sex differences, linguistic and cultural differences, and socfo-economic differences in school participation and scholastic achievement of primary school children. UPEL Study 2 will complement the present study by looking at the groups of children and adults who are not participating in the formal and non-formal education sectors.

2. AIMS AND OBJECTIVES

2.1. The overall aim and objectfve of this study is to identify the degree of school participation and to examine the scholastic achievement profile of primary school-age children (7-12+) in Ethiopia. Equality of educational opportunity is studied both in terms of equality of access and equality of success in primary education.

The specific objectives of the study are:

2.2. To make a diagnosis of the major trends of educational development across the country for disthct groups of primary school-age children, and to identify the underlying factors facilitating or hindering the Government's policy for greater equality of educational opportunity.

2.3. To study the patterns of disparity in children's school participation between and within Awarajas,Weredas: socio-economic, linguistic and cultural groups: and between the two sexes.

2.4. To study differences in the home and school environments of learning as measured through sets of home and school background characteristics and process variables.

2.5. To evaluate the importance of home and school factors on children's scholastic achievement in primary schoolfng, and to identify certain pollcy-related measures for ensuring a better quality of primary education. data collected in Study 3.1

(This aspect is treated from the

- 30 - 2.6. To develop an information-base and research competence at the Curriculum Department for planni'nq, manAtoring, and processfng of school survey research data.

3. STRUCTURE OF THE STUDY

The present study is d2vSded into three interrelated stages.

3.1. Stage I is a baseline study of school partic2pation along age, sex, regional (Awarajas and Weredas) li'nes. Measures of internal efficiency of the system (e.g. gross enrolment rates, net enrolment rates, age-specific enrolment rates, promotion rates, pass rates at Grade 6, drop-out and repeater rates) are chosen to assess the performance of primary schooling for the respective age, sex and regional characteristics of the primary school-golng population. Data from Phase I of the Quality of Education Project as well the official educational statistics can be re-evaluated to suit this stage of the study. In addition, UPEL Study 3 treats in more detail the wastage problems in primary schoolhg.

3.2. Stage I1 of the study deals with a sample or schools and students (see Section on Sampling procedure) selected for anidepth examination of educational access and educational success in primary schooling. Children's environments of learning (home, school and neighbourhood) is the focus of this stage of the study. Information about the home, school and neighbourhood factors Cbackgroudcharacteristics and process variables) will be collected in order to describe and to account for variations in the different learning settings of primary school children (see also the Section on Instruments).

3.3.- Stage I11 is the analytical part of the study where inferential statistical analyses are used to account for (a) the differential access to primary schooling for children of different characteristics and backgrounds; and (b) significantdeterminants (home, school and neighbourhood factors) of scholastic achievement at the end of the primary school cycle. As mentionned earlier, determinants of scholastic performance will be treated from the data collected in Study 4. The findings from the study will be used to indicate which policy- related inputs can be manipulated in order to attain higher equality of educational access and better ways for the distribution of educational benefits to primary school-age population in the country.

4. SCOPE AND LIMITATIONS

4.1. The scope of this study has to be seen within the context of the other five UPEL studies. The study's coverage is limited to the equality of educational opportunity issue for the primary school-age population (7-12+) who are in school. HoP7e-$er, Study 2 which looks at the linkage between the formal and non-formal education sectors will focus on the group of children (7-161 and neo-literates (16-49) who for certain reasons did not continue their education.

4.2. The study is not a longitudinal one, and it does not provide the opportunity to follow a g'iven cohort of children in order to account for the changes in educational access and educational henefits over a certafn period of time for dfstfnct groups of primary school population. However, the trend-analysfs may provfde to a large extent the information pertaining to equality of: educational opportunfty over a certain period of time. The present study can be used later as a first step for a research design accommodating longitudinal studies.

- 31 - 5. A SHORT LITERATURE REVIEW

5.1. In Ethiopia as in many other develop-ng countries, research studies on access and equ2ty. in education are of rather recent date. However, a major educational policy of these countries fs to ensure social equity (or at least to reduce fnequity) fn the prov3sion of educational opportunfties and benefits (Weiler, 1984). In the affluent and industrialized countrfes, a bourgeoning literature is available on research studies dealimg wi'th equality of educational opportunity (Chinapah, 1983).

5.2. Since the 1950s the concept of equality of educational opportunity has been exposed to an intensive philosophical, methodological, and empirical debate. Often the nature-heredity/nurture-environment controversy was the focus of the research endeavours. The primary question has been whether learning and cognitive growth of children is more conditioned by their inherited genes that by their environmental factors. Husen (1974) outlined a series of antitheses to that effect. Among others, he found that the hereditarian school are supporters of a "status quo" where each individual child should take advantage of the equal educational opportunity provided on a competi- tive and selective basis. Contrary to this view, the environmentalist school demands important changes in the society in order to correct for certain differences in the starting chances of children in school, namely individual, linguistic, cultural, socio-economic and regional differences. These differences call upon policies ready to distribute educational opportunities in a compensatory way. In the past few years, the environmentalist school has influenced the research studies on equality-of educational opportunity in both developed and developing countries, with particular emphasis on the differences between home, school and neighbourhood environmental factors on childrens' access to school and their learning outcomes.

5.3. Among the few studies carried out in the developing countries, there is no consensus as to whether an expansion of educational opportunities can compensate for the home and neighbourhood handicaps certain groups of children are usually exposed to. Marked differences in school participation and scholastic achievement have been observed along age, sex, linguistic, cultural, socio-economic, and inter and intra-regional lines. Anderson (1983) and Chinapah (1983) used the selectivity index to measure whether and to what extent increases in enrolment rates have ensured greater equality of educational opportunity for children differing in age, sex, cultural (ethnic or racial), socio-economic and regional (inter and intra) backgrounds.

5.4. The controversial nature of the research findings on equality of education in the developing countries does suggest, however, that across-nation comparisons can be misleading, and research on this issue has more relevance to policy-makers, educational planners, curriculum developers, and educators if a country-specific perspective approach is being used. to map the distinct features and underlying factors facilitating or hindering school participatfon and scholastic achievement of primary children in Ethiopia.

6. RESEARCH DESIGN FOR THE STUDY OF DIFFERENTIAL ACCESS

It is therefore imperative for this study

Under the theme of "Equality. of Educational Opportuntty fn Primary Education", this study incorporates two different areas of investigation: (a) differential access to primary educatton, and

- 32 - (b) determinants of scholastic achievement. As the former pertains to the entry process [see the overall model for the research programme), it is a study about the selection-recruitment process among children who are just reachfng school-age (6-72. The latter, on the other hand, is a study about scholqst2c achievement among children in the school. They refer to dilfferent processes, settings, and populations and, therefore, require dtfferent theorfes, models, samples, measures, instruments, fieldwork, etc. Tn short, they require two different research designs. access will be presented first and that of scholastic achievementin sectfon 7.

The reseamh design for the study of differential

6.il. Parameters of the research design

6.1.1. The focal point of the study on differential access 2s the selection process at the entry phase of the educational cycle. Its general strategy is to ident5fy the condftfons that facflitate or hinder access to education throuqh the study of the characteristics which differentiate those children who have low probability of, enter.iing primary school from those who are more ltkely to enter the formal educational pipeline.

6.1.2. As the study pertains to the entry phase of primary education, its target group must consit of those who are just reaching the entry age. Tn Ethiopia, thrs group consfsts formally of cheldrenin the 6-7 age-bracket, although iln practice for some, it may take some years after this age.

6.1.3. In order to identify the characteristics which distinguish children who are likely to enter the formal educational pipeline from those who are likely to Be excluded from it, observations for this study must be based upon a suitably varied population.

6.1.4. The proper target population for this study, therefore, consists of children in the 6-7 age-bracket, who may be or may not be already in school. constitute an important target group for reaching the UPEL objective. Naturally, the study cannot bebased solely upon the within-school population.

Those who are lfkely to be excluded at an early age

6.1.5. As elaborated below, the conditions which are assumed to affect the probability of entry/no-entrLJ characteristics: individual, family and the community (extra-familial environment), including the characteristics associated with the school. The setting for this study, therefore, nust contain variation in the characteristics at these three levels.

refer to three levels of

6,1,6.,Taking the above points into account, it is reconmended thata controlled variation in these Characteristics be obtained through multi-stage stratified sampltng (as described.below1.

6.1.7. Inasmuch as all the studies within this research programme are to be systematically integrated wkthin the overall systems model and the ecological data from this study is relevant to the other studies, it is furtkrrecommended that the same communities be used for all the studies and that thks study together with Study 2 (study on the marginal groups),be launched first. 6.1.8. As the fieldwork for this study is to be launched first, recommended that it be carried out as quickly and'snootly as possible. The research design should be kept very simple.

it is

It should, never-

- 33 - theless, be rigourous enough to produce a reliable set of ecological data for present and future use; i.e., a description of. the conditions which facilitate or hinder children's entry into primary school.

6.2. Basic model

In hypothesizing the importgnce of early environment of child development, the basic conceptual model suggested here is similar to that of Simmons (1980:23);. It is also in accord with Selowsky's assessment (1980:97-112) of research findings in this field. One well-known example of findings on the primacy of the pre-school environment is that reported by Jencks and his colleagues (Jencks, et al., 1972:256) :

"Our research suggests ... that the character of a school's outnutdepends largely on a single input, namely the characteristics of the entering children. Everything else - the school budget, its policies, the characteristics of the teachers .- is either secondary or completely irrelevant."

As the living conditions for children in Ethiopia vary greatly, it is important that their early environment be explicitly included in the conceptual model to guide the research.programme. To the extent that such factors as health, nutrition and socialization are related to early child development, there is a corresponding limit as to how much a school-centred strategy for UPEL can achieve by itself.

The first model presented here will enable an evaluation of the degree to which pre-school environment plays a role in determining the probability of entering formal school; As explained in the introduction of this research programme, this is only part of

: a more comprehensive concept.

In some ways, the model presented here is different from that of Simmons, in that it encompasses more dimensions within the general concept of the environment (for specifics, see section 6.3 on variables). However, the model will be presentedfirst at a very elementaty level. Its elaboration will take place upon later specification of the variables involved in each of the dimensions-.'

Simply stated, the parts of the model are the following:

6.2.1. The outcome variable is the probability of a child entering or not enteEing the primary school.

6.2.2. The entry-probability is assumed to be dependent upon characteristics pertaining to (1) the child, (2) the parents, and thek interaction within (3) the family structure and process, which are conditioned by the characteristics of (4) the community and (5) the school. This model hypothesizes that the impact of the characteristics in blocks 1 through 3 will depend on the characteristics of the community and the school. This postulate has direct inplications for the choice of a corresponding statistical model.

6.2.3. The variables that are encompassed within each of the above five groups of characteristics Will be further specified in section 6.3. They cover biological, social, cultural and economic dimensions.

6.2.4. As to the question of an appropriate statistical model, this study is to be launched first, and one of its parameters - - . is simpli-

- 34 - city (see section 6.1.8) and, therefore, it is recommended that this also be kept simple and quick. Hence, the method of crossbreak of means (percentage likely to enter) and standard deviations into the respective categories of the characteristics variables should be employed in the first step. It will generate much useful information which will reveal how the likelihoods of entry/no-entry vary across various conditions. Qualitative observations should supplement the description of these conditions associated with high or low probabilities.

Later om,. there is always the possibility of applying more advanced multivariate methods of analysis. The model stated above hypothesizes that the impact of the characteristics of the child, the parents, and the family structure and process is conditional on the characteristics of the communtty and the school. In statist-ical terms, this means that linear additive models are insufficient to test the model. Interaction terms must be included in order to achieve an isomorphism between the stated model and the statistical model.

6.3. Variables and instruments

6.3.1. In connection with the measurement of the variables, several types of instruments will be recommended. The guiding principles for the recommendation are approprtateness to the phenomena to be measured and accuracy and simplicity in administration, processing and analysis.

All variables which are to be measured through interview should be based upon the structured Interview-questionnaire format. The non-interview method of data-collection should be based upon a standardized observation schedule and checklist, which can be supplemented with an impressionistic written account.

Complete reliance upon unstructured interview and observations are to be avoided, because the time and resources required to translate the unformatted verbal accounts to summarizable information are more that the worth of the information thus obtained. However, when verbal accounts are directly tied to structured questions, these remarks can be very informative, since they tend to be more to the point. Such information which is expected from the interviewer must be explicitly solicited and, therefore, built into the interview- questionnaire. Otherwise, the information is even less useful that that obtained from open-ended interviews.

The main variables that are encompassed within each of the blocks in the basic model will be presented below. The variables listed below are only suggestions, which may be modified and supplemented.

6.3.2. The dependent variable is the probability of entryho-entry in respect to the primary school. This is to be measured by (1) whether or not the child is already enrolled in the school and (2) in case the child is not enrolled, whether or not the child is expected to start school in the very near future (say, within a year). This information is to be obtained through structured interview with the parents.

6.3.3. The variables included under the term ''child characteristics" include, inter alia, the following:

- 35 -

The

biological: sex, age, birth-order among siblings health, nutrition, physical handicap personality development: motor skills, verbal ability in Amharic assessment of own ability attitude toward schooling expectation of going to school time spent by child on household chores (e.g.,looking after younger siblings, livestock herds, agrlcultural work, collecting fuel, water) attitude toward learning Amharic.

biological data can be obtained through a structured interview. As to the measurement of the health and nutrition dimension, a health-and-nutrition expert should recommend some indicators that are not too difficult to administer. An example might be a check-list of symptoms of malnutrition, physical and mental handicap.

The method of observation in regard to " personality development I' should be kept at a simple level. An example of a fairly simple method of observing motor skills would be to give a child a paper and pen, ask him/her to draw a picture, and observe how the child goes about holding the pen, drawing a line, etc. An example of measuring verbal ability would be a simple word-knowledge quiz, adninisteredby letting a child name and explain in Amharic some objects in a picture.

6.3.4. Parental characteristics which may affect the probability of a child attending school include thefollowing variables:

(1) biological : sex, age (2) level of formal education (3) currently attending school (4) attending of having attended literacy course (5 1 II other courses or training

(6) indicators of intellectual-cultural level (e.g. possession

(7) occupational status and economic sector of employment (8) income and other indicators of wealth and standard of

11 11

programmes.

and frequency of reading of newspapers, books and journalsf;

living (e.g. size of livestock, availability of tapwater, toilet, electricity, radio, number of rooms, articles of furniture, etc.)

(9) mother-tongue and language use at home (10) fluency in Amharic (11) socio-cultural values (e.g. religion as an indicator) (12) attitude toward education and formal schooling (13) assessment of utility of ability to use Amharic.

The concept of "parents" may appear to be restrictive in an environment where the notion of family includes more that a nuclear family consisting only of parents and their children. Since we are dealing with an early childhood socialization environment, it is those who are primarily responsible for the "parental role" in child-rearing who are of prime relevance. Hence, the term "parents" as employed here applies to the social parents, not necessarily the biological parents.

- 36 - Even where parental role is diffusely shared among family

members as in an extended family, the choice of social parents may be regarded as a sample from thfs set of primary socialization agents within the family. The relevant 'social parents'' are those who spent more time in actually taking careofthe child during his/her infancy up to the time of the interview. Because of the importance of adults as social models Ce.g.,sex-role), it is important that, where present-,. both the male and the female social parents be included in the data collection.

6.3.5. The data for variables pertaining to garent-child interaction within the home processes are to be collected through structured interview with EACH of the social parents and through direct observations. These variables include the following:

Concerning first the parent-child relationship:

(1) child-rearing behaviour toward the child (2) attitude toward childxearing in respect to sons and

(3) conception of desirable attributes of girls and boys (4) conception of proper occupation for boys and girls (5) language of communTcation between the child and the parent.

daughters

Concerning child's education and future occupation:

(6) assessment of child's intellectual ability (7) assessment of the utility of education for the child (8) assessment of the utility of the child learning Amharic (9) attitude toward language of instruction in school

and occupation

and occupation.

(10) aspirations for the child in regard to his/her education

(11) expectations for the child in regard to his/her education

Concerning the role of the child in thehouseholdeconomy:

(12) assessment of the current economic situation for the family (13) expectations and aspirations for the family's economic

(14) indispensability/substitutability of the child's labour situation in the coming year

. for the household economy.

Concerning other household members:

(15) number of siblings and other members in the household.

For each member, obtain the following information:

(16) age, sex, health and physical handicap (17) level of formal education (18) attending school (19) attending or having attended literacy course (20) other courses or training

(21) language used at home and outside (22) fluency in Amharic.

It II I1

programmes

- 37 - 6.3.6. The extra-fam lial environmental var-ables per-ain to the community and school characterlstics. Community variables which may be particularly relevant to the probability of entry/no-entry are the following:

(1) general level of literacy of community members (2) urban/rural setting (3) language of the major2ty members (4) cultural values,e.g.,religion as an indicator (5) distance between child's home and the nearest town/village.

In respect to the nearest school to the child's residence, the school-related variables are the following:

(6) distance between the child's residence and the nearest

(7) number of years since the school was established (8) number of available places relative to the school-age

primary school

population size; enrolment relative to school-age population size

(9) sex ratio of the pupils in the school (10) internal efficiency, repetition and attrition rates (11) number of qualified teachers relative to the enrolment size.

6.4. Strategy for analysis

6.4.1. Based upon Awaraja-level database (see section on operational overview), the first step is an analysis of the stratification variables (see section on sampling procedure). The purpose of this analysis is to identify those variables, among the stratification variables, which are the most independent from others [i.e., least correlated with others). As for the hfghly correlated variables - since highly correlated variables are tapping the same variance - it would be redundant to retain all of them. Theoretically, the least important among the correlated variables can be dropped without loss from a sampling point of view. This step must be carried out as soon as possible-, since no study can commence without the information required for sampling.

6.4.2. The analysis of target community is the next step. Based upon the Awaraja-level database, analysis at the aggregate level can be carried out rather speedily. Tt will produce much useful information which can be employed to describe the environment, as related with high and low participation rates. This, at the first-grade level, is equivalent to the probability of entry/no-entry.

6.4.3. The next step is to analyze the "qualitative" observations, based upon interviewers' written text in the interview-forms, in order to get an impression of the fieldwork and garner hints -for further analysis (for caveats on the method of collection of such information, see introductory paragraphs in section 6.3). This information, along with simple descriptive statistics, can be used to describe the variations in children's environments.

6.4.4. The relationship between variables based upon data collection from fieldwork, together with the aggregate data, should be analyzed next. Simple percentage figures can be employed to discover the relationship betweenthe variables: It is important that such analysis be carried out beyond simple bivariate tables, since the

- 38 - conditional effects, as hypothesized in the basic model, can be analyzed only if the conditional variables, such as those of the community and the school, are explfcitly included in the analysis. as control variables. The impact of parental aspiration for the child's future occupation, for example, may be much greater in an urban than a rural setting.

6.4.5. Since no analysis is useful until its message has been communicated to and understood by the relevant audience,numerical. results should be converted to a medium that can easily communicate the message. One of the most effective means of communfcating the patterns embodied in the numerical results is a graphical display. Interpretation of a graphical display requires neither ability in interpreting tabulated numbers nor knowledge in statistics. There are many ways to summarize numerical information graphically: scatterplot,. frequency histogram, pie chart, curve plot, etc.

6.4.6. Later on,. especially in order to untangle complex relationships, multivariate statistical analysis may be employed. Even with a high-powered multivariate analysis, it would be necessary first to summarize the data through various techniques of data- reduction, such as removal of ftems with low variance of low discriminating power, index and latent-varlable construction, etc.

Once this is done, the variables can be sumitted to an analysis with a multivariate statistical model. With the probability of entry/no-entry as the dependent variable (i.e. a dichotomous variable), there are many such methods (see Cooley & Lohnes; Fergusson; Kerlinger) . As stated earlrer, however, linear additative models are not appropriate for testing the interactimeffects. The choice of which particular statistical model to apply depends on the information which the analyst is seeking as well as his/her knowledge and the availability of the necessary statistical programme and computer facilities.

6.5. Sampling design

6.5.1. From the point of view of maximizing precision of parameter estimates for a given simple size, a simple random sampling of households,as for example,. from the Census data files, would be the most appropriate. However, the fieldwork costs involved in collecting date from a simple random sample of householdswould be prohibitive. For this reason, and because the overall design of this research programme requires a controlled environmental variation, a multi-stage stratified cluster sampling design (Kish, 1965:359-383; Moser & Kalton, 1971:106-111) will be recommended here. The first stage is the sampling of Awaraja. The second stage is the sampling of Wereda within the sampled Awaraja. The third stage is the sampling of households. This section will be detailed, since it will lay the common ground for all the subsequent studies.

For administrative convenience-,. it is recommended that Awaraja be selected as the Primary Sampling Unit (PSU). Furthermore, in order to minimize the necessity of post-weighting the data for the estimation of population parameters, it is recommended that the sampling of the Awaraja with Probability Proportional to Size, PPS, (Kish, 1965:217-253; Moser & Kalton, 1971: 111-116) be employed at the first stage. Simply stated, the number of Awarajas to be selected from the respective strata should be proportional to the population size in the corresponding strata.

- 39 - As for the sample size, there are a number of points to take

into consideration. Reliable estimates of "intra-class correlation" are required for calculating the optimal sample size in multi-stage cluster sampling method (Kish, 1976:161-173; Moser & Kalton, 1971: 100-106). In the absence of any such estimates, it is virtually impossible to calculate the optimal sample size on a rational ground. Using a rule of thumb derived from the experiences of many national sample surveys, one is on the safe side with a sample size of about 1,800 households. Problems of ffeldwork and statistical computation can be reduced considerably by a good sampling design that takes such aspects into consideration.

data should be collected and corresponding data files should be maintained for sampling as well as for adjustment for sampling design and analysis of the relationships between the ecological variables with other varfables.

At each step of the multi-stage sampling, all sampling-frame

Studies 1, 3 and 4 are especfally dependentupon statistical analysis and, therefore, the sampling design is particularly Gertinent to these studies. The other studies (2, 5 and 6) are less dependent upon statistical estimation, since these will rely more upon non-quantified verbal accounts of Vobservatlons". This does not mean that sampling design is irrelevant for these latter studies, since the criteria of representativeness should also apply to such "observations

6.5.2. First stage stratification

The purpose of stratifying is to improve the representativity of the sample. The variables most appropriate for stratification purposes are those whfch are closely related to the main variables under study,. not closely related to each other,. and for which reliable data are available prfor to drawing the sample. the different purposes of the proposed studies and using the Awaraja-level database as the sampling-frame, the following variables are proposed for stratification of Awarajas:

Based upon

(1) Degree of urbanizatibn (percent urban population) according to the lates census (if this is found difficult to implement, then population density may be employed). Awarajas should be grouped into two categories: above or below the median.

(2) Linguistic classification according to nationality statistics. Based upon the predominant nationality within the Awaraja, classified into two categories: (a) Amharic,

I (b) Others.

(3) Literacy-level classification according to the latest Literacy Campaign statistics. Awarajas should be grouped into two groups: below or above the national median.

(4) Number of pilot schools within Awaraja, based upon information available at the MoE. The Awarajas should be grouped into two categories of density: above or below the median.

Note. The inclusion of the fourth variable (number of pilot schools) in the sampling design is requfred for Study 4.

- 40 - If one were to use all of the above variables with their given

categories for strat'Tflcation, there would be a total of 16 strata (. 2 x 2 x 2 x 2 ), whfch mfght be organfzed as Tn Table 1. Some of the strata will probably contailn very few or no Awaraja(s).

Table 1: Cell-numbers in an example of Awaraja stratification by 4 variables

URBAN NON-URBAN LITERACY -LEVEL PTLOT-SCHOOL Axnhar2c Non-A Amharic Non-A

High Many 1 2 3 4

Low Many 9 10 11 12 High Few 5 6 7 8

Low Few 13 14 15 16

It is recommended that, before the finalization of the stratification design, an empirical analysrs of the relationship between the stratification variables be carrfed out. If it is decided to reduce the number of strata, a practtcal criterfon is to delete the variable that is theoretfcally less important among the variables that are highly correlated.

If two variables are highly correlated, it means that both are covering the same area of variance. Highly correlated variables will create zero-cells. Hence, ft is redundant to include both of them in the stratification desfgn and, therefore, the variable that is least important (among the correlated variab1es)from. the theoretical or representativeness viewpoint can be deleted. If a variable is thus deleted, then there would be only 8 strata (2 x 2 x 21.

Another method to select the optimal set of stratification variables is to conduct a multiple regression analysis, with, for example, the enrolment ratio as the dependent variable and the above-mentioned stratification variables as the independent variables. Choose those variables which have the highest predictive power and drop those which have the least. (See section on strategy for analysis. )

The above-mentioned stratification variables are only suggestions, since these variables appear to be particularly relevant to the proposed studies. There are other variables that may be considered as relevant, not necessarily from a theoretical, but from a "representativeness" point of view; as, for example, predominant religion of the Awaraja population, geographical location of the Awaraja (e.g. coastal region, mountain region, pastoral region, North-South, East-West), etc. Such variables should also be included in the Awaraja database and be included in the above-mentioned analysis and selection of stratification variables.

6.5.3. Probability proportional to silze (PPS)

As for the total number of Awarajas to be selected, it depends on the resources one 2s able to invest for samplfng precfsion. The greater the number of Awarajas, the greater the precision, but also,

- 41 - the greater the cost of spreadkng fieldwork over more areas. Financial and personnel resources will set the maximum total number of Awarajas that can be sampled.

As for determining the number of Awarajas to be sampled within respective strata, it is recommended that the principle of PPS be applied. Suppose that a total of 20 Awarajas are to be sampled. For a stratum in which there is 1/10 of the population, one should select 1/10 of 20 Awarajas, i.e. 2 Awarajas. Since it would not be convenient to sample, for example, 2.3 Awarajas, fractions should be rounded to the nearest whole number (e.g. 2 Awarajas) .

In some strata in which the population is small, the number of Awarajas to be selected may be less that 0.5 in which case the categories which are producing such distribution may be collapsed. Since the variables contain only two categorfes each, collapsing is tantamount to the removal of such variables from the stratification design for a given set of cells. On the basis of the collapsed matrix, recalculate the samplkng fractions. Consequently, the actual number of strata will most likely be fewer that the number with which one started.

A clear advantage of this sampling method is that it simplifies the later stages (to be described below) and statistical computation in that it minimizes the necessity of weight-correction for non-proportional sampling.

6.5.4. Second stage stratification

As for the second stage sampling unit, one would have to choose wreda or Kebele/peasant association, wh2cheven has more reliable sampling frame data and Is more administratively convenient. As the Wereda is a more establishd administrative unit with a special office and an officer, it seems to be a more appropriate unit for the second stage sampling. Ultimately, the decision on this point should be based upon the experience of the fieldwork fo the Phase I of the ERGESE and the data-collection organization enployed by the Planning Services' Annual Educational Statistical Survey. The following discussion is based upon the assumption that bereda has been selected as the second-stage sampling unit.

For each of the selected Awaraja,obtain a complete list of Weredas. For each of the Weredas, obtain the following information:

(1) name and location (2) urban/town/village/isolated settlements (3) number of individual members (4) number of households (5) percent of membership "literaten by the standards of the

Literacy Campaign (6) number and addresses of primary schools within its boundary (7) of which, number and addresses of "pilot" schools.

This is a list of minimal Informatlon required for the creation of data file for sampling-frame and "community" level characteristics. More variables may be added, if they are deemed relevant for the studies. Variables 6 and 7 will be used specif2cally for the second-stage stratification. Accordingly, all Weredas wTthin the-

- 42 - selected Awarajas should be classified into the following categories in respect to the existence of primary school in their respective boundary area:

(1) no school (2) at least one normal school, but no "pilot" school (3) at least one"pi1ot" school.

For each Awaraja, select randomly one Wereda located in a geographical area where there fs no school at all; one Wereda where there is at least one school, but no pilot school; and one Wereda where there is at least one pilot school. Thus, in principle, three Weredas are supposed to be selected in each Awaraja. However, in practice only two can be selected in Awarajas where there are no pilot schools at all. Hence, assuming 20 Awarajas in the first stage, there should logically be 60 Weredas (20 x 3) or, assuming 10 Awarajas in the first stage, 30 Weredas (10 x 3). In case of sample dropout, it should be replaced by an adjacent Wereda.

6.5.5 .. Third stage sampling

from each of the selected Wereda a list of all households. The number of households to be selected depends on the following calculation. The total number of households to be sampled in all Ethiopia is about 1,800. Distributed equally over 20 Awarajas, it means that about 90 households must be selected in each Awaraja. Assuming an expected 20 percent sample dropout, it should be raised to about 113 households per Awaraja.

However, further adjustment is required in order to correct for the round-off that was done at the first stage. The target size (e.g. N = 113) must be multiplied by a factor of theoretical Awaraja sampling fraction (Ft) divided by actual Awaraja sampling fraction (Fa) :

The third stage of samplfng fs at the household Level. Obtain

*

N = (N x Ft/Fa). Suppose that the calculation in the first stage resulted in

2.3 Awarajas for a particular first-stage stratum. Assuming that 2 Awarajas have been actually selected from this stratum, the number of Awarajas sampled in this stratum was actually less than theoretically demanded. If one were to subsequently take a sample of only 113 households from each of the 2 Awarajas, the total nuqber of households in this stratum would be less that the proportion in the population.

To correct for such round-off, one would have to employ a correction factor, which in this example is (2.3/2) = 1.15. The actual number of households per Awaraja to sample within thiskstratum is (113 x 1.15) = 130, not 113. Had it been an upward round-off from 2.6 to 3 Awarajas, the number of households to sample per Awaraja would be (113 x 2.6/3) = 98.

Having adjusted for the round-off, the number of households should be distributed across the three school environment strata in proportion to the household-population (data collected in the second- stage stratification). Suppose that the target for the Awaraja, after correction for round-off, is 120. If the household population is equally distributed across the three school environment strata, then 120/3 = 40 households would be selected from each of the three Wereda. However, if 60 percent of the households live in Weredas where there

- 43 - is no school at all, then 0.6 x 120 = 72 households should be selected for that particular Wereda whfch represents the no-school area.

For the selection of households, systematic sampling (Kish, 1965:113-23; Moser & Kalton, 1971:83-4, 116, 1951, rather than random sampling, is recommended in order to ensure a good spread. Households can be listed by, for example, thefr membership size or, if available and up-bdate, household number as used in the Census. If the Census household number is avaflable, regardless of whether

included, for it would unlock the Census information and make it available to the study.

, it is used in the systematic sampling, th2s information must be

As the calculation fnvolved may seem complicated, a special calculation sheet should be developed fn order to simplify the procedure. All these adjustments are recommended to be done beforehand, because this would elimfnate the problem of later weightiqof data at the time of analysfs.

6.5.6. Households and the target groups for Studies 1 and 2

Finally, the selected household can be approached in the actual fieldwork. In case of dropout, select the household immediately below of above the household lfst used for the systematic sampling. Households in which there is at least one child within the 6-7 age-bracket constitute the target group for Study 1. Households in which there is at least one member in the 7-46 age-bracket who is not going to shoo1 are the target group for Study 2. Therefore, sampling and fieldwork for these two studies must be conducted together.

6.5.7. Caveat

Section 6.5.6. above leaves uncertainty about the final individual-level sample size for the respective studies. The reason for this is related to the question of availability of reliable data on the actual age-structure of all the households covered within the Wereda area. If sampling is conducted as in section 6.5.6., it would provide information necessary for the estimation of the age-structure of the area covered by the Wereda. However, the fieldworkers will not have the information beforehand as to how many households will be subjected to Study 1 or Study 2 interviews and observations.

Quota-sampling for the target age-groups would eliminate this uncertainty. However, without reliable data on the actual age-

correct for the design effect. Hence, it would lead to another sort of uncertainty of greater consequence: inability to make inferences about the actual population in Ethiopia, since the observed distribution is an artifact of the quota. If this problem can be overcome (i.e., if reliable data on the age-structure of the households is obtainable, prior to door-knocking], quota sampling of the target groups is preferable.

” structure, it would not be possible to adjust the observations to

6.6. Parameters of generalization

6.6.1. The observations fn thTs study are based upon a cross-sectional sample of households. The analysis that is possible’within this study does not span over time. Tt enables inferences about the relationship

- 44 - between early childhood envfronment and the probability of entering or not entering the primary school.

6.6.2. To the extent that sample dropout is correlated with the theoretically important variables, such as remoteness from the school or the nearest town/village, there w2ll be a systematic bias toward underestimation of the impact of these variables. Hence, all efforts must be made to avoid such bias, both Tn the fieldwork and in the completeness of data in the sampling-frames at each stage of sampling.

6.6.3. If the above steps have been followed, there is no need for weighting of the data later for the statistical analysis, since the result is based upon proportional sampling. This sampling design will produce an unbiased estfmate of the mean. To the extent that the actual sampling in the ffeld deviates from the above outline, the data must be adjusted accordfngly.

6.6.4. However, straightforward calculation of variance -.without taking cluster sampling design into consideration - will tend to underestimate the variance. Computing vareance requires separation of the clusters. Hence, it is very fmportant that cluster identification be explicitly fncluded in the data collectfon instruments and in the data files, whereby. each person/household can be uniquely identified with one, and only one, sampling cluster (i.e., specific Wereda) . 6.7. Operational overview

6.7.1. The first step is an establishment of an Awaraja-level database. Data for all Awarajas in Ethiopia is to be collected for the variables mentioned herein from various sources, such as the MoE's Planning Division (the Annual Educational Statistics), the Adult Education Division (Literacy Campaign Statistics] ana the Central Statistical Office (demographic and economic data). Obviously, data that can be obtained only through fieldwork (e.g., distance between child's domicile and the nearest school) cannot be included in this step.

6.7.2. The dependent variable can be measured already at an aggregate level (enrolment size relative to the school-age population for each of the sexes) and analyzed in relation to the ecological variables. This information should be regarded as only rough appro- ximation, because the data are unreliable. Actual age-range among pupils is greater than "school-age" range. However, the latest data collection instruments employed in the annual educational statistical survey do have age-sex breadkdowns. The problem is on the side of the census data from which the "school-age" population, which is reported to be unreliable. With this reservatim, it is possible to engage in analysis work at an early stage on this study. A written report should be produced as soon as possible so that' its information can be used for later steps.

6.7.3. Once the Awaraja database is established, the samplinc; work begins. Here the stratification variables must be selected on the basis of analysis. One possibility is a correlation analysis and the other is a regression analysis (for more details, see section on strategy for analysfs).

- 45 - 6.7.4. Employing the selected stratification variables and applying the procedure as outlined above in the sectfon on samplfng,select the Awarajas. For each of the selected Awaraja, collect data on the variables for the Weredas, build up a database and merge with Awaraja data.

6.7.5. Stratify the Weredas on the 'school variables' and select a subset of the Weredas as fnstructed in the sampling design. The geographical area for ffeldwork is now identified and planning of the logistics may begin.

6.7.6. Simultaneous or even prfor to the creation of the Awaraja database (the two are not mutually dependent), the instruments for interview and observations for the measurement of the variables as outlined in the section above on "variables and instruments'' must be developed. As stated earlfer, effort must be made to construct such instruments which are quick and easy to administer under greatly varying fieldwork condltlons.

6.7.7. In conjunction with the training of the fieldworkers, the instruments may be subjected to "tryouts" in order to identify the remaining defects in the fnstruments and to improve them as well as to train the fieldworkers. The hstruments should be tested on a small scale sample of the fntended target age groups. This is very important especially in regard to the young children. Encou- rage the fieldworkers to report any difficulty encountered in this tryout-administration of the instruments. Once the instruments are finalized and the fieldwork has begun, it is too late to make any changes.

6.7.8. When the instruments have been finalized, the data-capture routines, data management system (DBMS) and the programme for fieldwork administration should be immediately implemented. The data-capture programme should routinely screen for wild codes and store directly into the DBMS. Be especially careful to design the DBMS in such a way that the files can be linked through standardized case-identification fields. Data for each child, for example, must be linked his/her parents'data. It is mandatory that the data capture routine, the DBMS and the statistical analysis programme package are mutually comFatible and designed to enable easy statistical output directly after data entry. As part of this preparatory work, the data-collection instruments should be checked again for their compatibility from the viewpoint of data-processing.

6.7.9.. Logistics of fieldwork must be organized in accordance with the geographic distribution of the target communities-. Fieldwork assignments and packaging of instruments, arrangement for transport, etc. must be carefully planned. As for the timing of the fieldwork, it should be carried out just prior to the start of an academic year, in that the object of study pertains to the probability of entering school. Not only must the fieldworkers conduct data collection for both Study 1 and 2, as stated earlier, they must also initiate Study 3 (see Study 3 for more details) on these visits to the sampled areas.

6.7.10. Aside from the instruments, the fieldworkers should be provided with a mandatory check-off list and report forms for review and examination of the work at the end of each data collection session. These forms must be such that the worker must review the

contents of the filled-im Tnstruments and account for any peculiari- ties and missing cases. If any ambkgufty should arise, the worker is then able to follow it up Tmediately while he/she fs still at the data collection site.

6.7.11. The tompletdfnstruments and the reporting forms should be delivered to the fieldwork co-ordinator, who will check for their completeness and correctness. These will then be sent to entry of data into the DBMS.

6.7.12. Finally, analysis can be carried out according to the recommended "strategy for analysis" as outlined above. It is important that written reporting be carried out continuously at each step and not wait until all analysis is done. The final report should contain not only the ffndings in regard to the characteristic features of the high and low entry-probability groups, but also policy-implications in regard to possible remedial and compensatory measures to rectify the situation and evaluate the plausibility of the UPEL target-schedule.

6.8. Note on comDuter' 'us'acre

Direct amess to interactive computer usage on daily basis throughout the duration of the project fs absolutely essential. To cut down time and cost and increase precision and efficiency, access to a computer is required from the very start of the project: from instrument construction, database constructton, sampling, training, data-entry, coding, statistical analysis to project administration and report-writing. (See above section on operational overview; and for hardware and software requirements, see Tung's separate report. )

7. A MODEL FOR THE STUDY OF EQUALITY OF EDUCATIONAL OPPORTUNTTY

As mentionned in the ahoe section (61, the reseach design and the model for the study of equality of educational opportunity would have to consider two areas: (a) differential access to primary education, and (b) determinants of scholastic achievement. The former was discussed in section 6.

7.1. In this section, we shall present a conceptual model for the study of equality of educational opportunity., The conceptual model per se. is an integration of the models proposed for Study 1 and Study 3 (see Figure below).

7.2. The conceptual model

where the relationship (causal and non-causal) and the effects '(direct, indirect and total) of environmental factors (characteristics and processes) on (i) access to primary schooling, and (ii) pupils' scholastic achievement in primary schooling, were examined from the data collected in a national school survey. 7.3. Parameters of the model

This conceptual model is a modification of Chinapah's Model (1983)

The,rnodel consists of several blocks of variables and the hypothesized relationships between them. DTfferential access to primary schooling (Study I] can be examined through the hypothesized relationships between the variables in Block 1 (Home and Pupil's

- 47 - characteristics) and Block 3 (School Characteristlcs). In other words, it is possible to gather through the research hstruments, the main characteristics of the target groups (those reachfng the 6-7 age-bracket) and the probabilTty of entry or no-entry to schools. In addition, access to schools of di'fferent types and qualities for given characteristics of the target group (age, sex, ethnic, regional, socio-economic, etc.) can be cross-examhed.

7.4. Determinants of scholastic achievement (Study 3) can be studied through the hypothesized relatlonshfps among all blocs of variablesin the model. Direct, indirect and the total causal effects of the predictor variables on scholast2c achievement can be estimated through different multivariate analyses using for example, Educational Production Function (EPF), Commonality, PLS, and/or LISREL analytical strategies. An example of the application of these analytical strategies can be found in Cooley and Lohnes (1976); Munck (19791, Jbreskog and Wold (1982a and 1982b); and Chinapah (1983). (Refer to Study 3 for detailed presentation of the research design, parameters, sampling, variables, and strategy of analysis for study of determinants of scholastic achievement.)

FIGURE: A CONCEPTUAL MODEL FOR .STUDY OF EQUALIIY I)F DUCATIONAL OPPORTUNITY. ENVIRONMENTAL CHARACTERISTICS

Homc School: Parental Socloeconomlc Status School Type Home POSseSClOnl School RllplOn Home Utsraey Wnlclty School md Clas8 Slu Pmmnal Uumctirirtlu Teachrr Salules

schoo~ md Teacher aumy

- 48 - STUDY 2

Linkage between formal and non-formal education An in-depth study of the marginal groups!

educational needs

Purpose

The main purpose of the present study is to examine among 7-16 year-old children and neo-lfterate adults between 16 and 46 years of age the nature of causes of non-partlcipation and discontinuance in the two main delivery systems of educatton - the formal education (primary level) and the non-formal education Clfteracy and post- literacy). The study of non-partic2pation ard'discontinuance will enable us first, to identify marginal groups among these children and neo-literate adults and second, to examine their future educational needs in both delivery systems as well as in other systems where such educational programmes exfst.

Desired outcome

To set-up an information base that may be useful for the identification of critical areas where appropriate educational services could be supported, extended, and integrated with other community development servfces marginal groups and to suggest ways of strenghtenlng the linkage and continuity between the formal and the non-formal education sectors.

fn order to serve the needs of these


1.1. The linkage and continuity between the formal and the non-formal education are important components of the strategy towards mass involvement in education and mass participationin development in contemporary Ethiopia. Although viewed as a stepta-step process in the TYPP, the linkage and continuity in the two delivery systems can guarantee some form of educational opportunity for the broad masses in Ethiopia. Accordingly, mass educational provision calls upon the identification of the new target groups.

"Education in Socialist Ethiopiahas a dual function: to provide the adult population with knowledge and skills to carry out the work of constructing a new society and economy, and to bring up the new generation in the best possible way to continue this work.!' (MOE, 1984a, p. 16.)

1.2. The increasing demand and needs for the formal and the non-formal education make the two delivery systems mutually supportive. The linkage and continuity between-them may therefore reinforce the role of education as multiplying agents of developmentin the country. However, the mutually dependent relationsh9between formal and non-formal education requires proper coordination in a number of key domains (e.g. organizagional, administrative, operational, pedagogical, and attitudinal).

1.3. Besides primary educational programmes and the National Literacy Campaign, a network has Been established for the diversificatfon of education and training in Ethiopfa. There are several supporting

- 49 - agencies to serve this purpose, among others, the Ministry of Health for the training of health agents and tradfti'onal mtdwhes, the Ministry of Agriculture usfng farmer trainfng centres and the Agarfa Multi-purpose Farmer Trainfng Centre for the training of peasants for agricultural and community development purposes, and the Handicrafts and Small-scale industries Development Agency whose functions are to improve and to dissemhate new technologfes for handicrafts and small-scale industries.

1.4. The linkage and continuity between the formal and the non-formal education sectors emphasfzes a reconsideration of the following three key problem areas: {a] fncreased social demand and pressures for both delivery systems; (bl scarcrty of educational resources; and (c) educational wastage.

1.5. Despite the rapid growth of primary educatfon since the 1974 revolution, a substantial proportion of certain groups of 7-16 year-old children (girls, those from rural and poor family background, and nomads) are either unschooled or dropped out at an earlier stage from the primary education sector. Similarly, a significant proportion of neo-literate adults did not pursue their education any further, and the chances for them to relapse into literacy are high. Non- participation and discontinuance become therefore key areas of consideration for the overriding objective of the universalization of primary education and literacy in the country. The mutually supportive character of the two dellvery systems is expressed as follows:

"Floating between the two systems could well be the destiny of children and youth of school-going age, many of whom, for various reasons may grow up to be illiterate, uncatered for by the literacy programme on the one hand and the formal education system on the other... It is clear that universal primary education and literacy will have to go side by side for perhaps a longer period than expected.'' (Messele Mitiku, 1984, p. 47-48.)

2.

2.1.

AIMS AND OBJECTTVES

The main purpose of this study is to examine among the 7-16 year- old children and the neo-literate adults between 16 and 49 years of age, the nature and causes of non-participation and discontinuance in the formal and the non-formal education sectors (primary education and literacy programmes). Access, equity and efficiency are the three major policy domains under consideration in this study.

1 More specifically, the study attempts: __ 2.2. to identify and examine the background characteristics of the 7-16 year-old children who (a) are never enrolled in primary education or in the literacy classes; and(b) dropped out withinathe primary education cycle; and (c) neo-literate adults (16-49) who after certification did not conthue in post-literacy educational programmes;

2.3. to study the reasons for non-particfpation and discontinuance between and within thesetwo delivery systems and to relate them to some selected variable such as age, sex, linguistic, cultural, socio- economic and community background Characteristics;

- 50 - 2.4. to investigate the educational needs and aspisatkons of the marginal groups of children and adults interms of (a) the rnknimal amount of formal and non-formal educatEon required: (b) the type and content of that education; and (cJ the relevance of such education for satisfying their basTc individual and community development needs

2.5. to examine the institutional/educational linkage at the community level for such programmes run by different bodies (e.g. Ministries of Health and Agriculture).


The present study consists of three main parts:

3.1. The first part of the study fs an inventory based on school records, Kebele/Peasant-Association records to identify the groups of children and neo-literate adults who did not continue their education (drop-out from primary. education and post-literacy classes) and those children who are never enrolled in any type of formal and non-formal education.

3.2. The second part will consist of an in-depth follow-up study of some selected marginal groups through household interviews. Information will be gathered on the reasons for non-partici,pation or discontinuance In the two delivery systems.

3.3. The third part of the study wfll confine itself to an analysis of the information gathered from households, school records and Kebele/Peasant-Association records to (a) identify the major reasons for non-participation and discontinuance in the two delivery systems; (b) examine the possibilities for educational programmes and services that could be supported, extended and integrated with other community development services.


4.1. The nature of the present study calls for a limited sample of informants. As an in-depth follow-up study, much of the resources will be concentrated on qualitative type of data collected through detailed interviews. However, as a complement to Study 1, the information on the general characteristics of the home, school and community environments of learning will be used to describe the learning settings of the groups sampled for the present study.

4.2. The study will be to a large extent a descriptive one. suggested that several studies of a similar nature should be performed in the future.

It is

5. A SHORT LITERATURE REVIEW

5.1. The literature on research studies dealing with formal education and non-formal education has indfcated a certain trend in terms of the definition and application of the two concepts in different situational contexts. However , education as such whether formal, non-formal or informal is generally understood as a process where an individual acquires a certain amount of knowledge, values, skills and attitudes. The concept is sensitive as to the place and the time it is referred to (Chinapah and Fagerlind, 1979).

5.2. Although the most common use of the education concept is in association with formal schooling, we find the definition given by Coombs and Ahmed (1974) more appropr2ate to the present situationin

- 51 - Ethiopia, namely that education is broadly associated wit,, learning - regardless of where and when the learning occurs'. words, the role attached to education in Ethiopia as multiplying agents of develqnentemphasizes how formal and non-formal education interact with each other, reinforce one another through various media and become mutually supportive. It is imperative to look at the Linkage and continuity between and withinthese two delivery systems for the simple reason that the primary education sector and the national literacy campaign are vital agents for reaching the broad masses of the country.

In other

5.3. Some countries have in the past broadly educated its masses through the non-formal educational sector before education became formalized (e.g. Norway and Sweden). The situation is rather different in many developing countrfes where the rapid growth of formal schooling has caused an under-estimation of the potential role of non-formal education in terms of social prestige, access to employment facilities and generating earnings.

5.4. However, the social demand and pressures for education (formal and non-formal) have intensified the linkage and continuity aspects of the two delivery systems. The outflows from the non-formal educational sector (neo-llterates) would relapse into illiteracy if possibilitiesfor post-lrteracy educational programmes are not available; Ethiopia is among the few developing countries where the linkage and continuity between formal and the non-formal education maybe fully realized. For example, neo-literate adults are given the opportunity to be enrolled as third graders of primary

* education level. However, as Messe Mitiku (1984) rightly pointed out, the success would depend also on an articulated strategy for the provision of mass education through the coordinated knctioning the of two delivery systems at grassroot level.

5.5. We shall conclude this short literature review by referring to two questions that came up from'a body of researchers and prac- titioners meeting in Paris (October 1983) on "the diversificationof the educational field". These two questions were found to be of utmost importance forlesearch on linkage adcontinuity between the formal and non-formal education sectors. They are:

-what are the possible combinations between formal and non-formal education, and which-social groups, sexes, ages, urban/rural groups benefit most from the drfferent opportunities?

and social mobility of people concerned? - how do the various combinations impact on geographic, professional, 6. PARAMETERS OF THE RESEARCH DESIGN

6.1. The focus of the study

From an examination of the main title, sub-title and description of this study preceding this section, It is evident that this study contains two sub-studies: (a) linkage between the formal and the non-formal educational systems and (b) educational needs of the out-of-school population. However, these two subjects require different research designs, inasmuch as they refer to different units of analysis, structures and processes. Design will differ depending on the selection of the primary subject. Therefore, the main subject of this study must be first determined.

- 52 - 6.1.1... The main subject of study

As it has been agredpreviously that this study is to be an adjunct of Study 1 and that the two studies will share data, it is assumed that the primary unit of analysis of this study is at the same level as in Study 1. In other words, the main object of study pertains to attributes, behaviour and conditions of individuals or groups of individuals, not those of institutions or programmes. Hence, it is assumed that the subject of the educational needs of the out-of-school population is to be the main focus of this study. Consequently, the subject of institutional linkage will be treated only as secondary, as potentially instrumental to meeting the educational needs of the out-of-school population. This interpretation is further supported by an implicit assumption in the study description that 'linkage" will somehow promote better possibility of meeting such needs.

Focus upon the educational needs of the marginal groups will enable study to make a significant contribution to the international body of knowledge in a subject area which has been heretofore much neglected. Clifford W. Gilpin of the World Bank (which has financed the first 100 Community Skills Training Centres in Ethiopia and involved sirnilarily in Tanzania, Sudan, Somalia, Papua New Guinea and several other developing countries) in a work entitled "Issues in Non-Formal Education and Training for Rural Development" points out that the translation of the broad objectives of rural education and training into programme content and delivery approach should take into consideration three principal factors (Gilpin, 1979:3-4):

(a) the availability of resources (b) the development possibilities and constraints of the area (c) the characteristics and prioritiesof the target group.

"The third factor", Gilpin (1979:4) points out, "is the one most often neglected but is as critical as the others".

Furthermore, more concretely, this study has a potential for providing key information for the improvement of the curriculum in non-formal sector. G.D. Green, Chief Technical Adviser, Skill Development for Self-Reliance (ILO project financed by SIDA), states (1981:422) "Regrettably, one of the common failings with non-formal training lies in curriculum development." It is the neglect of acquiring research knowledge of the "learning needs of the intended clienteles" prior to curriculum development that is dften the cause for the failure of the non-formal training programmes ..."

"The approach recommended is to d m w q social profiles of selected communities and therafter to analyze the findings in order to identify appropriate training activities for various social groups. 'I

6.1.2. Four sub-topics

According to section 2 "Aims and objectives", four specific aspects are to be investigated. The first is a descriptive study of the background characteristics of the marginal groups: The second concerns the reasons for their non-participation or discontinuation and the third, their educational needs and aspirations. Finally, the institutional linkage between the formal and the non-formal educational system run by different bodies is to be investigated.

- 53 - Hence, this is primarily a study about the potential clientele

for the non-formal as well as formal educational programmes. It is not a study of the non-formal educatlonal programme per se.

It is important to point out, however, that the subject of the non-formal educational programme itself should not be neglected. There is a general agreement in the literature on education that a non-formal educational programme can play a very important strategic role in economic development, especially in the rural sector. Such a study is strongly recommended. However, because of its special requirements, it should be conducted outside the present framework.

U' 6.2. Definition of "non-formal" education

Before one can even conduct a discussion about the relationship between formal and non-formal educational systems, let alone locate the setting of the research target groups, it is first necessary to define these concepts. Wfthout a clear definition of the boundaries, it would be not possible to locate the setting of the "non-formal" educational system in such a situation as described by Cole Brembeck:

"Non-formal education is sometimes called the sflent educational system. It does not require elaborate campuses and buildings. Its 'curricula' will not often be found spelled out in cata- logues or brochures. Yet non-formal education occurs almost everywhere in fields, factories, offices, cooperatives, associations. I'

6.2.1. Coombs-Ahmed co'nti'nuum *

The distinction between formal, non-formal and informal educational systems is typically based upon Philip Coombs and Manzoor Ahmed who equate education to learning on a continuum from informal to formal. Along such a conthuum, they employ the following set of definitions to distinguish between the three educational systems (Coombs and Ahmed, 1974:81 :

Formal education: It... institutionalized, chronologically graded and hierarchically structured educational system, spanning lower primary school and the upper reaches of the university."

Non-formal education: ... any organized, systematic, educational activity carried on outstde the framework of the formal system to provide'selected types of learning to particular sub-groups in the population, adults as well as children."

Informal education: 'I... the lifelong process by which every person acquires and accumulates knowledge, skills, attitudes and insights from daily experiences and exposure to the environmenk. . .

6.2.2 Contextual speciftcity of "non-formal" education

In the present context, the formal education refers to that conducted within the framework of the primary school curriculum. More problematic is the definition of the "non-formal" educational system. In one of the most recent and comprehensive reviews in this area, the SIDA-commissioned study entitl&i"Non-Formal Training Programmes for Rural Skill-DeveloFment: A State of the Arts Study", the author describes the situation thus (Gorham, 1983:16-17):

. .... . .

- 54 - "Although there is general agreement that the definition of non-formal education fiust take as its starting point the out-of-school connotation, the latter by itself identifies a huge and amorphous field and does little to provide a means of discriminating among the components which make up that field. Attempts to achieve conceptual clarity have, therefore, largely given up the search for characteristics which would support a global definition of non-formal education and sought instead to allocate the problem to speciffc contexts. Defining non-formal education has thus become a functional issue which emplasizes the grounds on which 'formal' fs distinguished fran'non-formal.' in a particular instance, or on which degrees of 'non-formality' are alloted to various types of non-formal activity."

6.2.3. Governmerkconducted trainina moarammes

Although formal deffnition of "non-formal" education is absent in the study description, there are some references to empirical objects which are explfcftly of implicitly equated with the concept. Aside from the literacy and post-literacy programmes, there are references to the training programmes conducted by the Ministry of Health for health agents and mfdwives, the Ministry of Agriculture for training peasants and farmers and the Handicrafts and Small-Scale Industries Development Agency. Apparently, "non-formal" education is understood to mean educational (leteracyj or skfll training programmes conducted by Government agencies outside the formal school curriculum.

6.2.4. An alternative for non-educated adults

According to the Ethiopia/SIDA education sector review of 1979, non-formal education is equated wfth "adult education", presumably under the supervision and co-ordination of the MOE's Dept. of Adult Educatiom,- and encompasses five matn components (1380:18):

- basic education (functional literacy training) - rural adult education - urban adult education - women's education, within the framework of women's associations. While it is apparent that non-formal education is an alternative

for the non-educated adults, it is not clear what age groups are included in the term "adult". It appears as if there is alternative to formal education for school-aged children who are no yet in school.

6.3. The- setting of non-formal education

As Brembeck pointed out in an earlier cited work, "non-formal education occurs almost everywhere - fn fields, factories, sefices, cooperatives, associations...". This may be so, but it cannot be left so open-ended, if serious research is to be carried out on this subject area. It is necessary that the setting on non-formal education in Ethiopia be further specified in order for the research to find its direction of enquiry.

6.3.1. Organizational location of non-formal activities

The 1984 "Education in Socfalist Ethiopia" (MOE, 1984:44) makes a reference to "Post Literacy and Adult Education", where activities outside the formal primary school curriculum is presented. These

..

- - 5 5 - activities include hose of the Community Skill Tra-ning Centres (CSTC's), Basic Education Centres, and Community Education Centres. It concludes :

"In general, adult educqti.on, whfch embraces the work of all extension and development agents and the development work Of urban kebeles and rural peasant associations, can be brought to a higher level on the assumption of basic lfteracy."

6.3.2. The structure of non-formal education

Examination of the documents pertaining to the Ethiopian I. educational system reveals that a comprehensive system of non-formal

education has been developing through the establishment of facilities at the village, Wereda, Awaraja and national levels4 In this development, the CSTCs play a central role with an emphasis on skill training and developing non-formal educational activities, including basic education and literacy( The Centres are supervised by the MOE, but used also by development agencies of parttcularly the Ministries of Agriculture and Health. An inter-ministerial committee coordinates training programmes in agricultural extension, health, community development, etc. Additionally, within the MOE, there are the Departments of Adult Education, Mass Media (educational radio) and the Correspondence Educatfon Unit, all of whfch are engaged in educational programmes outsfde the formal school currfculum.

6.3.3. Research settins

In our review of the definition of non-formal education and the empirical objects associated with non-formal education in the Ethiopian context, it has become evident that, "in principle", virtually all educational and skfll-trainfng programmes conducted by the MOE, by other ministries and by develophentagencies outside the formal school curriculum in cooperation with various local organizations, are considered to be part of the non-formal educational system.

These many variations pose a great problem in trying to narrow down the research setting to a speciffc institutional setting. There are two alternative solutions to this problem. One alternative is to adopt an exploratory approach and take first an inventory of the types of non-formal educational activities conducted by those institutions and organizations mentioned in the above sections 6.3.1. and 6.3.2. Then procede to conduct research, taking into consideration this variation in non-formal educational activities. Cost in terms of extended time and resources and lack of sharp focus are likely to limit the immediate benefits of this research, although such an exploratory approach may bring long-term benefits.

programme which can be selected on the grounds of its strategic role in the attainment of the UPEL. As this research programme is to play a supporting role in the UPEL drive, such a criterion can bring a sharper focus to a more immediately relevant point. Of all the non-formal educational programmes mentioned in the documents, one in particular is not only by definition directly related to the core of the UPEL objective, but also plays a central role in the link between the marginal groups, the formal and the non-formal sectors. This is the literacy programme, For those people not in school, literacy is not only a bridge to formal education,.. but also, as we have seen,

Another approach is to focus on a specific non-formal educational

- 56 - (section 6.3.1.), considered by educational planners in Ethiopia as a pre-requisite for the improvement of the standards of all other non-formal education.

6.4. The research target groups

There are four specific aspects to be investigated. The first is a descriptive study of the background characteristics of the marginal groups. The second concerns the reasons for their non- participation of discontlnuation and the third, their educational needs and aspirations. Finally, the hstitutional linkage between the formal and the non-formal educational systems run by different bodies is to be investigated. Tn sum, from the perspective of data-collection design, this study concerns two levels of analysis: individuals belonging to marghal groups and the institutions/ programmes' which constitute the formal and the non-formal educational systems. The research target groups are, first, the potential or actual clients of the non-formal educational programmes and, second, the informers of the institutional aspect of these programmes.

6.4.1. Research population

According to the descriptfon of this study, two particular groups are to be specially studied: school-age children (7-16 year olds) who are no in school, and neo-literate adult (16-49) graduates of the literacy campaign. However, research questions should identify attributes, perceptions, motivations, behaviour, etc. which distinguish between participants and non-participants in respect to formal and non-formal educational programmes. Therefore, inasmuch as the identification of unique background characteristics which distinguish non-participants from participants requires comparative reference groups, the population for this study fn practice will encompass all individuals between the ages of 7 and 49 years old. Thereby, it will be possible to study the correlates of probability of participation and non-particfpation.

6.4.2. In-depth exploratory interviews

As "causes", "reasons" and "needs" cover a very wide field and many different levels, and assuming that there 'is no prior systematic research on these aspects in Ethiopia., it is recommended that in-depth exploratory interviews with a relatively small sample (see section on sampling) be employed to tap such basic but complex information. This approach implies that many questions will be posed to a few people, rather than a few questions to many people as En sample survey to estimate population parameters. Thks recommendqtton applies to the interview portion dealing only with "causes, reasons, needs". As the sample size is reduced, tts samplimg error can be expected to increase. The results of these intervTews can provide much information on types of "causes", "reasons", "needs", but the reliability of the data for the estimation of their distribution in the population is questionable. As for the estimation of the population belonging to the "marginal groups'', there will be sufficient data from the Study 1 household survey.

6.4.3. Perceptions from different perspectives

Inasmuch as data-collection is dependent upon the subjective response of those intervfewed, it is fmportant to bear in mind that one is dealing with perception data. As such, differences fn the perceived "needs" of the people in the way of education and the

- 57 -- reasons attributed to non-participation are most likely to be related to the role occupied by different groups involved in education. Perceived needs and problems may be examined franboth the clients' as well as from the agents' points of view. Aside from differences related to socio-economi'c, linguistic, regional and other such variations within the Ethfopfan population, there are likely to be important differences in percefved "needs" and "causes" between such groups as potential/actual students themselves, community leaders, teachers/instructors, adminAstrators, educational planners, development agencies, etc.

In order to capture the perceptual di'fferences related to the roles involved in formal and non-formal education, the target groups for data-collection should include both those who belong to the research population (as defined in secti'on 6.4.1.) and the personnel responsible for formal and non-formal education. Also to be included are those who are formally entrusted with the task of articulating the "needs", i.e., local peasant and urban dwellers associations. Wereda-level Revolutionary and Development Committees, and other development agencies.

6.4.4. Sub-set of Study 1 household members

In summary, the sample of observations for the study of the marginal groups is a sub-set of the household-members sample for Study 1, supplemented by a sample of the local personnel responsible for formal and non-formal education as well as those who are formally entrusted with the task of articulating educational needs at various levels of the organizational structure. For a discussion on the theoretical justification for the inclusion of these supplementary groups, see sections 7 and 8 below.

7. BASIC MODEL

"Years of research have gone into the shaping of formal education. On the other hand, research history on non-formal education is all too brief and inadequate to give much guidance." (Cole Brembeck, 1978:13).

This is certainly applicable to the present situation in regard to research on the out-of-school population as potential clients of non-formal education. To be sure there are now many descriptive studies of particular $on-formal educational programmes, about their successes and failures . However, there are hardly any studies about the marginal groups who might benefit from such a programme. There are no commonly accepted research paradigms and associated models that can be employed for the desigh of such a study. Hence, we are proceeding with an exploratory approach, not a confirmatory approach, as in the evaluation of the plausibility of a given theory ormodel.

Inasmuch as this study is recommended to be only exploratory and conducted parallel with and as a by-productof Study 1, it will not be necessary to develop an elaborate assumed causal model for analysis.

* See, for example, conference and workshop report on Non-Formal Education and Rural Poor, Ed. by Richard 0. Niehoff; and case studies in Simkins, Non-Formal Education and Development, 1977.

- 58 - However, if it is to serve as a guide for the data-collection work, it will require some a priori: conceptual organizatlon. It is only at this level that a model will be consfder,ed here. As for the overall relationships, see the Overview section.

As we embark on a discussion on research model, let us recall that. ithad been clartfied in the sectfon on the parameters of the research design that the main subject of this stidy is "the educational..needsof the out-of-school population". The secondary subject of institutional linkage enters the-dkscussion in the context of-its potential for meeting such educational needs. Hence, the relevant model should reflect this particular perspective.

7.1. Basic approaches

According to the SIDA-commissioned "State of the Arts Study" of the "Non-Formal Training Programmes for Rural Skill-Development" (Gorham, 1983:25-28), approaches in the non-formal sector are conditioned, as with formal education, by perceptions of education's role in a particular context. In the context of improved rural development, two schools of thought predominate. The first of these is the "functionalist" approach and the other is the Eonflict" approach.

7.1.1. Functionalist school

The first approach views the problem in "functional" terms and systematically sets out to design an effective strategy to overcome existing dysfunctionalit2es (Coombs and Ahmed, 1973). It is based upon the view that non-formal education and training are potentially ppwerful forces for rapid socio-economic improvement; and the belief that, properly applied, they can eliminate or significantly reduce many disparities in income, productivity and living standards.

7.1.2. Conflict scho'ol

The second approach regards functional strategy for non-formal education as a means whereby existing inequality can be maintained and legitimkzed (see Paulston & LeRoy, 1975). This approach is based upon a class-conflict/political power structure model. According to this school of thought, those aspects which promote the interests of the power-holders (i.e., those who have the means to. impose their will) are more likely to be implemented that those which contravene their interests. Whether or no the outcome will benefit the potential clients or promote economic development is of secondary importance.

7.1.3. Approach in this study

These two approaches refer to the causal agents assumedLto be operating upon the effectiveness of the educational system. If the purpose of this study is to evaluate these causes, such models should of course, be considered.

However, in the limited context of the present study, it is unnecessary to go too deeply into the reasons behind such processes. Our starting point is an assumption that all human beings have certain basic needs, some of which can be fulfilled of facilitated through education or training programmes. The mapping of such needs, and participation in the programmes intended to meet them, are the only levels at which an appropriate model ls required to organfze the study concept.

- 59 - 7.2. Xicro-macro perspective

Although the primary focus 2s on the educational needs of the marginal groups, this study also refers to the relatlonship between these needs and particlpation in the educational programmes made available by resource-allocating and mobilizing authorities. Hence, a model limited to individual-level phenomena or institutional-level phenomena would be inadequate. The two levels should be related. It is this perspective that Bock and Papagiannis advocate in their paper entitled "The demystification of non-formal education" (1976).

They (Bock and Papagiannis, 1976:19] view education as

"that specialized-set of social arrangements which specifies and delineates the socialization process and which subsumes all learning that is purposive and direct... At the same the social arrangements themselves represent patterned experiences which convey 'indirect' lessons to those who experience them."

In this view, non-formal education Is seen (ibid., p.21) as

''a socialization sub-system shaped by society or some aspect of it (class, sub-culture, political movement, etc.) which it is embedded.

Hence, in this model, differential participation is (ibid., 24-29) partly a function of the "educative agencies'" differential selection, recruitment as well as accessibility and partly a function

of non-formal education in relation to social values; for example, better employment opportunitles.

-. of the individuals ' expected "exchange value" or anticipated "payoff

In its basic outline-, their approach can be encompassed within the model presented in the "Overview" of the proposed research programme, to which the reader is referred. As for the relashionsh2p between individual-level attributes, perceptions and behaviour pattern and the environmental context, the model presented in Study 1 can provide a conceptual organization of this study with some modifi- cations. First, references to primary school must be supplemented with corresponding references to non-formal education. Second, the concept of "needs" and its relation to participation must be further specified. Third, the model is inappropriate for generating descriptive categories for the vareous aspects of the non-formal educational programme. Hence, there has to be a framework appropriate for

< describing such programmes, not hdividual/group behaviour.

7.3. A needs-participation model

A model that incorporates the above discussion on the ma'rginal group's learning needs and differential participation in educational programmes pertains to individual/gPoup behaviour under various environmental conditidns. In constructing such a model, one can begin with the assumption that "participation" (or, inversely, non- participation or even discontinuation) is partly a function, following the Bock-Papagiannis postulate, of the expected "exchange value" or "payoff", which is sought in order to satisfy some basic "needs". These basic relationships are condi€ional upon educational agencies' differential selection, recruitment, avallabil2ty and accessibility vis-a-vis the potential clfent&les.

- 60 - 7.3.1. The concept of "needs": objective and subjective

"real needs". This is an important distinction inasmuch as Gilpin (1979: 8-9) makes a distinction between "felt needs" and

"One of the principal problems fn introducing effective non- formal education ks the marriage of !felt needs' and 'real needs' ... Where the gap 2s too wide, the failure of the population to respond may wreck the programme."

This distinction will be referred to hereafter in terms of Objective and subjective needs.

7.3.2. The concept of "facilitation": objective and subjective

In order to gather under one term all conditional factors such as selectiom;, recruitment, availability and accessibility, the concept of "facilitation" (or, inversely, hindrance) is used here. As in the case with "needs" even facilltation can be "real" or "felt". Conse- quently, this distinction will be made in terms of objective and subjective facilitatkon.

7.3.3. Conceptual model

The discussion on the mode.1 so far can be summarized in the form of a flow diagram (see figure below).

CAUSAL FLOW DIAGRAM OF PARTICIPATION IN NONFORhW PROGRAMME

c SUBJECTIVE

NEEDS

NEEDS

FAC I L I TAT I ON

SUBJECTIVE FACILITATION U

PAT1 ON

- 61 - The model assumes that variations exist in objective, "real"

needs,and facilitative conditions. These account for variation in subjective, "felt", needs and facilitation. All of these variables in turn are assumed to influence the variation in the expected value of payoff of participation. Finally, these variables will converge to determine the Variation in actual particlpation rates.

One practical advantage of this model is that, having located the sources of variation through this study, it will be possible to identify the necessary improvements in terms of specific "objective" conditions and "subjective" aspects of the target groups. Non- participation due to some subjective aspects of the target groups, for example, can be modified through information and motivational changing actions.

7.4. A descriptive framework: the systems approach - TRUGA If the study is to be maimly exploratory and descr5ptzve as

recommended, then a conceptual organizat2on that well be very useful for generating descriptive categorEes is that of ILO's "Trainfng for Rural Gainful Activities (TRUGAJ". Thfs represents a recent attempt to apply the systems approach to the design and delivery of non-formal skill-development programmes in the rural areas of LDCs. As summarized by Gorham (1983:43-461, Truga consists of eight major components, each representing a step in the process of designing and delivering an effective rural skills-development programme. The functional relationship between the Truga system components is presented in a diagram (page 62).

Further details Qf thi's framework and its utility will be taken up in section 8 in connection with the varkables and in sectfon 9 in connection with the strategy for analysis.

As the TRUGA model considers only the non-formal system in isolation from the formal educational system and from other sub- systems, it must itself be consfdered only as a sub-system of a more coqxehensive system as outlfned fn the Overview, including the reiationships between (1) formal and (2) non-formal learning systems and their causal precedents, (31 environmental facilitation/obstruction, (4) human and material resource allocation and (5) curriculum implementation. Such a comprehensive perspective will be necessary, for example, in order to study the linkage aspect.

8. VARIABLES AND INSTRUMENTS

The general comments on the lnstruments as mentioned in section 6.3.1. of Study 1 apply in this study as well. As for the variables, on the whole the indivfdual-level attrebutes and perceptions will be measured by variables contafned fn Study 1. Most of the variables pertaining to non-formal educational programmes will be those connected with the eight components of the TRUGA systems framework.

8.1. Individual-level attributes, behaviour and perception

I

Generally, the variables and the instruments for measuring at the individual-level are identical to those employed in Study 1, with the following modification:

- 62 -

F \

Training for Rural Gainful Activities (TBUCA)

~~

IETUT definition of existing skills knowledge etc. in target group

I FuNcrIOIo identification of skilled manpower

FUNCTION I idantif ication of potential eco- fnomic ac iviries in rural areas , 1 requirements to meet activity needs I

OurpUI definition of learnkg obi ec- tives to meet defined needs 1

I y< 4 INSTBUCTIONAL~CONTENT 'training content required to U

1 close output-input &p *

I I OPEBATING CONDITIONS I I identify training constraints, facilities, personnel, resour- 1 ces, budget, time, equipment etc. [

I 1 1 I COMPONENT DESIGlJ and PRODUCTION design appropriate methods, de-

.~ 1 livery sy stem, support services ] I I I I EflECUTION execute programme and observe pro- grcss I

I evaluate the results and provide

- 63 - 8.1.1. The dependent variable

The dependent variable is the probability of participation in non-formal educational programmes. This is to be measured by (1) whether or not the indivfdual has ever participated in any non-formal educational programmes, and, ff not, (2) whether or not the individual is expected to participate in such a programme fn the near future (say, withfn a year].

8.1.2. Objective and subjecthe needs

Variables pertafning to perceived needs which the respondents associate with skill-development programmes, as well as corresponding variables to be measured by some external "objective" criteria, are to be introduced. Interview questions on educational needs should be posed to the potential/actual clients: to those responsible for the courses; and to those who are entrusted with the task of articulating such needs on behalf of the urban and peasant associations. It Is interesting that Gflpin (1979:9) makes a distinction between subjective and objective needs in terms of the informer: "'felt' needs as defined by the target groups and 'real' needs as defined by the planners." In the context of the present study, the "needs" as identified by the target groups can be classified as subjective and those identified by the supplementary groups as objective.

As we have stated earlier, even the response of the supplementary groups is subjective. Hence, an external "objective" measurement will be required. Even Gilpin suggests (1979:8) that "Real needs may be determined objectively by reference to the economic conditions and standards of living". ILO-SIDA project "Skill Development for Self- Reliance" (SDSR) has developed such indicators.

As a fulllisting of the variables and description of the instruments is not possible within the context of this paper, the reader is referred to the 1LO's"Community Profile Programme: Community Surveys, The Survey Instrument". A careful selective approach is advisable, since the paper is very comprehensive and voluminous.

8.1.3. Objective and subjective facilitation

In measuring differential selection, recruitment, availability and accessibility, sub-sumed under the concept of "facilitation", both the objective and the subjective facts must be included. As in the case with "needs", these questions should be posed to the groups mentioned in section 8.1.2.

L Indicators of objective facilitation are such variables as geographical distance to the nearest relevant educational programme, qualification requirements, language of instruction, real costs, opportunity costs, etc. The indicators of subjective facilitation are the individual's perceptions of those variables which may hinder or facilitate his/her participation in the educational programmes.

8.2. Linkage variables

The instftutlons/programes which are encompassed in the concept of "non-formal" educatton in the Ethiopian context have been discussed in section 6 above. Here,linkage variables will be discussed,in particular various linkage relationships presently under way in Ethiopia.

- 64 - 8.2.1. Bridging

According to a special art3cle, "Linking Formal and Non-Formal Education" in The NFE Exchange (1982, No.23, pp. 1-21, the linking has often taken the form of "br2dghg". Students participated in non-formal education programmes to gasn access to either the formal educational system or to enter the job market. The educators may see it from the perspective of learnfng as a lifelong process. may see it as a cost-effective way to make the best use of a nation's limited resources to meet the ever-increasing demand for education and to promote national development goals.

Planners

8.2.2. Sharing

According to the same artfcle, linkages may be forged at the local community and regional levels by the sharfng of (a) facilities, (b) human resources, and (CJ methods and content. Typically, the linkage is found where there is an overlap fn the form of vocational/ employment orientation.

This type of sharing already exfsts between the formal and the non-formal systems in Ethiopra, as stated in the 1979 Ethiopia/SIDA sector review (1980:19), "It 2s assumed that the target Dopulation will participate in actfvfties organized in regular school buildings and also in the Community Skfll TrainEng Centres". inst.itutiona1 Ekage at the communfty and regional levels remain to be investigated.

Other forms of

8.2.3. Policy coordination

As for linkages at the policy-making level, the NFE article

(a) an institutionalization of an fnter-ministerial advisory body to coordinate educational policy and resource allocation, to facilftate support and sufficient funding for the efforts of each ministry. "This could help a country meet its manpower needs inore adequately and make better use of its available resources for both formal and non-formal education".

(1982:5) makes two recommendations:

(b) "Centralization and linkage should be conducted in a way that leaves room for small-scale, locally-initiated projects." It is important to stimulate local participation and self- reliance, "... now usually considered essential if development is to benefit the poor" and all villagers, especially disavantaged groups such as women.

1

Linkages at the national level, as we have already mentioned earlier, exist both within the MOE and between Ministries; they are even formalized in an inter-ministerial coordinating committee. Institutional linkages at the local level are biefly discussed in the above-cited sector review (1980:19). The Community Skill Training Centres :

"are meant to represent the master key to the development of non-formal education... The CSTC programmes should reflect the needs of the local communities, as identffied by local peasant and urban dwellers' associations in cooperation with the Wereda-level Revolutionary. and Development Committees and other

- 65 - development agencies. The commfttees are also responsible for the management at the local level. The Irlini’stry of Education assigns a coordinator to each CSTC.”

Linkages, therefore, do exist - at leat on the drawing board. The question is whether or not they extst in actual practice. extent and form between the local level and the fnter-ministerial level remain unknown.

Their

8.2.4. Three dimensions of l2nkage

Investigation of linkages, will require documentary as well as interview sources of informatAon, from the National Government to the local associations and persons respons2ble for the training programmes. Linkage variables, in summary, must span the following three types of relationship:

(a) policy coordination mong decision-making bodies, (b) sharing of faciliti’es, human resources, methods and content, (c) flow of individual participants between the two systems.

The extent to which linkages are conceptualized must be compared with their extent at the grassroots level.

These linkage dlmensions should be applied in conjuction with the eight functions of the Truga system components (see section 7.4 above) . 9. OTHER DESIGN CONSIDERATIONS

For other considerations in respect to this study, see Sections 9 to 14 in Study 1 proposal, as these two studies will be conducted parallel to each other.

- 66 -

STUDY 3

Performance of Primary Education System

A study of the nature, extent and causes of educati'onal wastage and the attainment of basic skflls in primary education

Purpose

The overriding purpose of this study fs to fdentffy and to examine factors which facilitate or hinder pupilstcompletion of the primary educational cycle and to tnvestigate the defferent levels of proficiency in the basic skills as reading, writing and calculation.

Desired outcome

To provide necessary emptrtcal evidence for an identification of possible corrective measures for reducing educational wastage and for ensuring a minimum learning contfnuwn through the primary education cycle.


1.1. Mass education is a very recent phenomenon in socialist Ethiopia. Any attempt at assessing the prospects of mass education (in quantitative or qualitative terms) since the 1974 Ethiopian revolution should account for (i) the educational backwardness of the country prior to the revolution, and Cif) the progress achieved in the past 10 years. Within this background, one of the UPEL's objectivesof "keeping the child 2n school for the' 'full 'cyc.le' o'f 'primary' 'e'duca'tbn and enforcing a m.in.imum. 's.tan'da'r.a 'of '1e'a'r.n'fn.g ac'hf'e.Ve.me'nt 'I i: s examined in the present study.

1.2. In the 1970's, Ethiopia had an educational backlog that could hardly be compared to any other nation on earth. However, the progress achieved in the past ten years puts Ethiopia among the most educationally progressive countries of the past decade. In a period of ten years, primary education participation rate had reached 43.6 percent as compared to.some 20 percent in 1974. Enrolments in rural schools increased nearly three-fold from 25 percent to some 60 percent. Girls' enrolments also rapidly increased (around 37 percent today) as a result of the new roles attributed to girls and women in the building of socialist Ethiopia.

1.3. It is often argued that a trade-off between the quantitative growth and the qualitative change in educational development is imperative, irrespective of a country's historical, political', socio-economic, demographic or cultural conditions. Such is the case in Ethiopia today. The explosive nature of educational demand and supply has resulted in what the nation as a whole most needed, i.e. "TIRAT" or quality.

1.4. An assessment of educational wastage and minimum learning standards at primary education level will require both quantitative and qualitative performance indicators. In this context, data that are already available from Phase I of the "Quality of Education Project" will have to be re-analyzed and cross-examined; findings from studies of a similar nature that have been carried out in

- 67 - Ethiopia or elsewhere w. 1 have to be properly s-udied. In other words; it seems to be approprTate and useful to do a meta-analysis (research on existing research studfes) before embarking on any new research endeavour.

1.5. The achievement of a full-cycle of primary education im Ethiopia requires a set of strategfes or scenarios to achieve an acceptable degree of internal efffcfency. Thfs requfres research and evaluation, mainly identification, cost-effectiveness analyses, and the feasi- bility of applying correctfve measures in order to improve retention and to better ensure a mfnimum learnfng continuum throughout the primary education cycle.


2.1. The overall aim of thfs study is outlined above. More specifically, the study attempts:

2.2. To present a profrle of educational wastage (drop-outs, repeaters, failures, and the lfke) at drfferent grade-levels of primary education and along age, sex, geographical, socio-economic and cultural lines.

2.3. To identify major causes of educational wastage from a sample of participants (family, households, students, teachers, and school directors).

2.4. To assess the relative Tmportance of school and out-of-school related factors on retention and attainment of basic skills in prixfcary. schools.

2.5. To identify possible correctfve measures for reduchg educational wastage and for ensuring a minimum learning continuum through the primary education cycle.

2.6. To create an information base and the necessary infrastructure for the collection, compilation, dissemination, and analysis of relevant data for educational research and evaluation.

2.7. To ensure a competence building programme for the conduct of present and future research and evaluation studies in the country.

It should be noted that the objectives 2.6 and 2.7 are also strongly emphasized in all the other studies covered by UPEL.


The present study consists of four parts.

3.1. The first part of the study deals with an in-depth literature review on performance of primary education systems in developed and developing countries. It also includes a desk-study of Phase I of the "Quality of Education Project" where some relevant data are avaiLable on the question of educational wastage and minimum learning standards.

3.2. The second part of the study is a pilot, in whtch a limited number of local school communittes will be sampled for a diagnosis of the problems related to educational wastage and attainment of

.' .

- 68 - basic skills among primary sch.ool children exposed to different environmental settings - urban/rural; high-medium-low levels of school attainment; cultural-ltnqulstlc homogenefty/hetereogeneity; and economically diversified areas.

3.4. The third part of the study consi'sts of (e) pre-test of research instruments - their wordhg, format, and sequence - and evaluationof alternative methods of data collection (e.g. questionnaires, interviews, and participant observatlon); and (ii) analysis of the results from the pilot study in terms of clarity ardformat of the instruments used, variance in responses, fnternal validation of ftems, constructing tables, indices, scales and factors.

3.5. The last part of the study will focus on a broader survey of educational wastage and determinants of scholastic performance (minimum learning continuum). Tt should be noted that the sample for the survey should be selected in the same fashicnas the pilot- study sample.


4.1. The scope of this study has ta be seen within the strategic framework set for all the research and evaluation studies under the UPEL's umbrella. In other words, there is a two-way traffic between the present study and the remafning five studies in terms and inputs, outputs or feedbacks. The resources available and the competence required would not permit a nation-wide study of all aspects of educational wastage and scholastic performance in the country. However, the competence for such a nation-wide study should be built up during the planning and execution of the present study.

5. A SHORT LITERATURE -REVIEW

5.1. It is generally assumed that a literature review serves the dual purpose of "state of the art exercise" prior to the study in question, i.e. through an increase of the available stock of knowledge and competence required to perform the study; and of "delimiting" scope and coverage; i.e. from the general to the specific needs of the study.

5.2. The burgeoning literature and available empirical evidence from research on educational wastage and scholastic performance can already be used as potential sources (mainly research findings) for planning the present study. Apart from the research results from Ethiopian studies, the following major findings on educational wastage and determinants of scholastic performance may be used as inputs for the design of the study. They are associated with:

(i) the relevance, utility or remoteness of the teaching- learning matter (textbooks, syllabi, teacher-guides, etc.) to the-daily needs of children in-and-out of school, e.g. the regional discrepancies, socio-economic realities, the cultural and linguistic differences, and the disproportio- nately academic-oriented, as opposed to practical-need oriented, type of curricula;

(ii) the examination-selection-evaluation processes of qualification for further schooling; in other words, a system whereby the teaching and learnfng activities focus on examination instead of child-developmental needs;

- 69 - (iiildifferences in the home learning environments (e.g. the

cultural and educational capital of Farents and siblings, attitudes, belfefs, and expectations from schooling, parental reinforcements, opportunity costs for schooling, treatments at home), marked differences in school learning environments (:e.g. classroom climate, peer group characteristics, attitudes, Behavfour, and expectations, disci- pline of teachers and children, teacher characteristics, morale, behaviour, and expectations, neTghbourhood characteristics, geographical disparities in educational services and other servfces, and indivfdual differences - age, sex, health and nutrttfon conditions, time allotted for homework, distance of school from home, and so forth;

(iv)

5. 3. This

the corrective measures for improving retention and scholastic performance: - the need for a minimum learning continuum approach where diagnoses, feedbacks, and remedial teaching are carrfed out at varfous crftical moments in the ?rimary cycle for slow-learners and educationally handicaFned groups, automatic, promotion + teacher apprenticeship + compensatory education formulae; ava2lability of free textbooks and learning mater5als; increased access to pre-schooling; boarding facilfties, use of the mother- tongue as a medium of instruction in the early grades, increased participation of women in the teaching-force, in-service trahing for unqualified teachers, introduction of feeding-programme, etc.

short literature review provides some directives as to progress elsewhere on the causes of educational wastage and low level of school attainment/retention. A more comprehensive and systematic literature review is needed. The results from the Phase I Study should be properly studied, along with research results.Once the major sources, courses, and. outcomes are known (including results from the second stage of the study] final sets of instruments, as well as data collection and data analysis techniques can be selected for the final survey.

6. PARAMETERS OF THE RESEARCH DESIGN

6.1. The focus of this study is on the investigation of factors which facilitate or hinder pupils' cycle and to investigate the level of proficiency in the basic skills -

the target-groups are to be identified in terms of their probability of promotion, repetition and attrition respectively.

completion of the primary educational

. of reading, writing and calculation. Through this investigation,

6.2. The design is such that research can be conducted within a much. shorter time that Phase 1 research and within the manpower cabacity of the division.

6.3. It is assumed that hiqhest dropout rates occur in the first three years of the primary. education'- and that repetition rates are higher in grades 1 and 6, the fnTt2al and the terminal stages in the primary educational cycle.

6.4. It is assumed that by the end of each year of the primary education cycle, pupils should have attained certain literacy and numeracy skills.

- 70 - 6.5. With the above-mentianed points as parameters, the most efficient and the most relevant desfgn would be to focus upon grades one, two, four and six and to set the tirming of iinterview and observation at the end of term, looking hack to the Beginning of the term and looking forward to the next academi'c year.

7. BASIC MODEL

7.1. An outcome variable in thisinvestigation is the transition behaviour. This cons2sts of three modes: promotion, repetltion and attrition (dropout].

7.2. The probability of the respective behavioural modes is affected by (1) individual's performance of the screening/~atekeeping tests, (2) attainment of basic skflls, (3) teacher-learner interaction, (4) curriculum relevance, (5) attendance record, (6) socio-economic conditions and socializatton process wlthin the pupil's home environment and (7) community-level ecological contexts such as the community's economic base of social organlzation (urban/rural) , literacy level (above/below median] and cultural (llnguistic and religious) compositfon which are assumed to produce differential impact on the importance of education and opportunity costs of keeping children in school.

7.3. The statistical model corresponding to point 7.2. is the Discriminant Analysis Model, fn which the linear combination of the variable-blocs (1) through (71 named above define the discriminant functions and predict the probability of pupil's behavioural modes.


8.1. The dependent variable is the between-level transftfon behaviour (promotion, repetition and dropout) to be collected through a special follow-up report from the director's office.

The intervening variablesconsist of the followfng in-school vartables:

8.2. Pupil's attendance hfstory since reaching school-age as measured by attendance record, self-report and parental report in interview.

8.3. The proficiency level attained in reading, writing and arithmetic as measured by achievement tests. Performance on the screening/ gatekeeping tests as measured by the end-of-year tests used for passing or failing pupils.

8.4. The cognitive, affective and behavioural dimensions of teacher- pupil interaction as measured by interview and observation schedule.

8.S. Curriculum relevance as measured by interview and observation in regard to pupil's perception;. affection and behaviour toward the school subjects in both the school and home environment.

8.6. The exogenous variables conskt of the following:

(i) Teacher's sex, age, education, language, attitudes, perception and expectat5ons in regard to teacher-pupil relationship in general and his/her own pup2ls in particular as measured by questionnaire.

- 71 - - (ii) Pupil's home conditions, language, and processes, especi

parental expectatfons, attitudes and behaviour in regard to their children's schooling and socio-economic role as measured by household interview and observation.

(iii)Community-level ecological factors (urban-rural/literacy/ above/below/median/Amharic/non-Amharic, Christian/Moslem) as measured by Awaraja classification based upon census and geographfc data.

9. STRATEGY OF ANALYSIS

9.1. Data reduction is the first step, in which the items which lack variance should be deleted rom further analysis and the remaining items should be recoded, where necessary, and ordered for scaling and construction of composity variables (indices, factor scores, etc.) The reliability and validity of these measurements is then evaluated.

9.2. Each of the behavioural modes (promoted, repeated, dropped out) should be analyzed in relation to the "cleaned" variables in order to identify those factors which are empirically correlated with the different outcomes. The correlates identified can be used to describe those conditions which facflitate or hinder completion of education.

9.3. The target groups (high probability of repetition or attrition) for remedial action can be identrffed fn terms of distinguishing characteristics.

9.4. Through the analysis of multivariate contingency tables * ("cross-tabulation" or "cross-breaks" of mean-scores and standard deviation), some elementary evaluation of the "causal" mechanism and estimation of the direct and indirect effects is possible.

package, the following are examples of possible types ot analy,

9.5. The combination of ANOVA-MCA or GLM can be used to estimate independent as well as interaction effects of the predictor-variables, which can be mixed scales (nominal, ordinal, interval), producing results that can be easily interpreted and presented.

9.6. The Discriminant Analysis Model can be used to map probability areas of the respective khavioural outcome on a multidimensional space defined by the linear functions of the correlated predictor variables. Analysis of the discriminant functions in terms of the loadings (correlations) will enable identification of the major dimensions which predict the different behavioural outcome. It may reveal that, for example, predictors of promotion are not th# same as those which predict drop out.

9.7. Through Commonality Analysis, the relative impact of the respective variable-blocks in terms of the proportion of variation in behaviour statistically "explained" by the respective variable- blocks, "uniquely" as well as "jointly". Here, one must bear in mind the "Coleman-Report controversy", i.e. the causal sequence of the respective variables.

- 72 - 9.8. Complex causal model, including "unobserved" variables (based upon "manifest" var2ables or observed indicators) can be analyzed by such methods as PLS (Noonan & Wold , fn econometric tradition), LISREL (Joreskog, in psychometric tradition) and other modelling techniques.

9.9. There are various other methods which can be used to analyze different types of relationships and types of data (see Chinapah, 1983).

10. SAMPLING PROCEDURE

10.1. The population consists of all the pupils ingrades 1, 2, 4 and 6 in Awaraja where there is at leat one "experimental".school. However, it is assumed that these Awarajas are representative of all Awarajas in Ethiopia. The existence of the experimental school is required for Study 4 design.

10.2. The sampling-frame for the first stage of the procedure is a list of all Awarajas where there is at leat one experimental school. One Awaraja should be selected randomly from each of the 16 possible strata defined by classiflcation of Awaraja according to the latest census and geographic data on the following variables and their respective classification categorres:

(a) Dominant econlmoc base for social organization: urban/

(b) Literacy level: above/below median. (c) Dominant linguitic-ethnic group: Amharic/non-Amharic. (d) Density of ''pilot'' school" above/below median.

agrarian/pastoral.

See Sampling section of Study 1 for more details.

10.3. The sampling-frame for the second stage is the list of all primary schools in the selected Awarajas. Two types of schools should be selected randomly:

(a) One non-exper2mental school +. One reserve. (b) One experimental school (Study 4) + One reserve.

10.4. The sampling-frame for. the third stage is the list of all classes in grgdes 1, 2, 4 and 6 in the selected schools. This list should be obtained from the schools and it should contain the names and addresses of all first, second, fourth and sixth graders who had registered at the beginning of the academic year. Special reply- forms should be sent to these schools as soon as they have been selected for the study.

One cl+s should be randomly chosen from each of the grade levels.

A copy of the selected class-lists should be sent back to the school with a request for attendance record-keeping during the acadeifiic year.

10.5. The interview and observation ffeldwork will be dfrected to the members of the selected classes, both the pupils; their teachers and families.

- 73 - 10.6. Data collection is completed when the school directors have returned the special follow-up forms for reporting on each sample of individual in respect of promotion, repetition, and dropout.

variables.

11. PARAMETERS OF GENERALIZATIONS

11.1. The sample is based upon those whose names appeared on the re- gistration list at the start of the academic year in the year of observation. Hence, it does include all those who subsequently dropped out, repeated or were promoted. However, it includes neither those who dropped out before the year of observation and did not return to school, nor those who never entered the school system. Hence, the results should be interpreted in terms of the probable fate of those who registered for school admission and the phenomena associated with their behavioural outcomes. The study does not enablegeneralizations about individuals who have not yet entered the school system (see Study 2).

11.2. The design of this study is such that it enables analysis of relationships between phenomena which vary cross-sectionally across population sub-groups. Tt is n0t.a longitudinal study and does not enable analysis of variation across time beyond the span covered directly by the data collection instruments and observation.

11.3. Estimations based upon data unadjusted for the sampling design should either be reported for the respective strata (in which case generalizations should be confined to them accordingly) or else the stratification variables should be explicitly included in the computational procedure as in multivariate analysis.

11.4. To compute population parameters, the data must be first adjusted for sampling design through weighting of the cases belonging to the respective sampling strata.

12. NOTE ON COMPUTER USAGE

Direct access to interactive computer usage on a daily basis throughout the project duration is absolutely essential. time and cost and increase precision and efficiency, the computer should be used from the very start of the project: from instrument construction, sampling,data entry, coding, file-building in DBMS, and analysis to report writing. Computer hardware & software specification.)

To cut down

(See separate report on Tung's

- 74 - STUDY 4

Oualitv Improvement in Primarv Education The effect of human and material resources on the quality

of teachfng and learning outcomes.

Purpose

To evaluate the effects of human and material resources on the quality of teaching-learning process and the learning outcome in the children's cognitive, affective and psychomotor development in different school-class settfngs.

Desired outcome

Identification of major quality-control inputs related to teaching-learning process that significantly affect primary pupils' learning outcome. Identification of the most cost-effective set of policies in regard to acquisftion and allocation of resources and designing of teacher training fn relationto the different target groups. An important indirect outcome here will Be the strengthening of scientific evaluation research as an important tool in the development of the new curriculum.


1.1. In order to meet the UPEL objectives, there are many problems which must be solved simultaneously. One of the most important of these is the scarcity of human and material resources that can be .devoted to the school sector. From the viewpoint of planning further expansion of the primary education system, it is imperative to acquire more knowledge about the effects and cost-effectiveness of human and material inputs in the school system in bringing about qualitative improvements in teaching and learning outcomes.

1.2. A study of this aspect has become not only possible, but. also policy-relevant, due to the introduction of the "experimental" or "pilot" school-classes in some parts of Ethiopia and the plans for its gradual expansion to the whole educational system.

1.3. The new curriculum, according to "Education in Socialist Ethiopia" (MOE, Oct. 1984), has been issued to selected schools as a guide to programme implementation. A process of controlled implementation and evaluation was initiated in 1980 in 70 pilot schools. These schools were equipped and staffed by specially trained teachers in order to implement the new curriculum.

1.4. Schools possessing the following characteristics were selected : (i) urban and rural areas; (ii) community of the main nationality, where there are more than one nationality in the Awaraja; (iii) availability of a teacher to be assigned for the pilot class and the availability of classroom with desks and blackboard sufficient for 50 students; and (iv) community of high population density.

1.5. The educational objectives embodied in the new curriculum focus upon education for (il production, (ii) scientific consciousness and (iii) socialist consciousness. The pilot school programme has a grade-by-grade intensive study of the viability of the new curriculum.

- 75 - It begins with a three-year experimental phase in which use of the new curriculum is observed and evaluated by NCDC (National Curriculum Developmentcentre) experts at the Curriculum Department of the Ministry. The draft curriculum is then revised and adjusted and .

finally published as the tested and approved programme of studies for the grade.

1.6. Consequently, the introduction of these "pilot" classes has created three different types of environment in terms of human and material inputs: (i) the "pilot" classes themselves, (ii) the "adjacent" classes, i.e., clgsses within the same schools where there are pilot classes, and (iii) the ordfnary classes in other schools, which are now more frequently referred as "transitional" classes.


2.1. Whereas the new curriculum is carefully followed, its qualitative outcome depends on the specially trained staff and specially allocated material resources in various contexts; and it remains yet to be scientifically evaluated, especially in terms of cost-effectiveness. It is urgent that this aspect be assessed properly, in the light of the plans for expansion, as well as for the transformation of other schools according to this new model. Such an evaluation will enable educational policymakers, in decisions pertaining to acquisition and allocation, to rank-order the required resources in terms of cost- effectiveness in bringing about qualitative improvement in teaching and learning outcome and, hence to increase the probability of reaching the goals of the UPEL. This is the overall aim and objective of this study.

The specific objectives of the study are:

2.2. To evaluate the content of the respective domains of learning, the cognitive (knowledge) , affective (attitude] and psychomotor (skills), inthe light of the educational objectives embodied in the new curriculum. Learning outcome in this study refers to the qualitative aspect of these domains.

2.3. To evaluate the impact of the respective school-input variables on learning outcomes and, thereby, to identify the input variables that can serve as cost-effective instruments of policy in regard to acquisition and allocation of resources.

2.4. To evaluate the impact of the respective teaching-learning (T-L) process variables upon learning outcomes and+. thereby,. to identify the T-L variables that can serve as effective instruments of policy in regard to the recruitment and training of teachers as well as the preparation of the teaching materials and guidelines.

2.5. To evaluate the impact of ecological factors upon the efficiency of the input and T-L variables and thereby to identify the most cost-effective set of policles aimed at dffferent target groups.

2.6. To develop the necessary competence for conducting scientific evaluation research of the educational programme and its use as an integral part of the development of more effective educational planning, implementation and management.

- 76 - 3. STRUCTURE OF THE STUDY

The structure of the study consists of four sequentially dependent parts:

3.1. The first partof this study consists of a review of previous studies on the effects and cost-effectiveness of school inputs on the teaching-learning process and the quality of learning outcomes. In conjunction with this review, the research staff will require a training course in the specialized research methodology employed in this subject area.

3.2. The second part is devoted to the study of the different conditions pertaining to the school settings - the "pilot" and the "transitional" - in order to identify those factors which may be particularly relevant for the further refinement of the study design.

3.3. The third part is related to instrument development, in terms of construction, testing and refinement. Tn connection with this task, a rapid small-scale pilot study should be carried out. This pilot study would simultaneously provide opportunities for training of the staff.

3.4. Finally, the main study will consist of a major survey (to be conducted in coordination with Study 31, data analysis,. interpretation of the results and drawing conclusions about policy options in regard to resource allocation, teacher training programme, curriculum, textbooks, facilities, etc. that may be required for the further development of the new educational system.

For more details, see section 11 on operational overview below.


4.1. Internationally, there are many previous studies on the relative effects of school inputs. tendency related to the level of development, although this tendency has stillto be confirmed by other studies in different contexts.

The results show on the whole a general

4.2. One tendency which is of particular relevance in the present context is that the importance of school-inputs is positively correlated with the scarcity of resources: i.e., school-inputs play a more significant role in an environment where such input resources are relatively scarce; inversely, such inputs play much less role where a minimum level of input requirements has already been achieved.

4.3. Thus, research studies in Western industrialized societies have come to a general consensus that scholastic performance of children is due less to the influence of the school, but more to that bf the home environment (Colsman et. al., 1966; Plowden, 1967; Peaker, 1971; Jencks, 1972; Bowles and Gintis, 1976; Coleman, 1981; and Fagerlind and Leal, 1982).

4.4. In contrast, earlier studies in the developing countries have not found such a general pattern. It appears that scholastic performance is determined dffferently at different levels of education (primary, secondary and tertfaryl fn dffferent settings (rural and urban); from one school subject to another (Heyneman, 1976; Schiefelbein and Simmons, 1979; Husen et.al., 1978, Chinapah, 1983).

the findings also vary from study to study and

- 77 - 4.5.'However, having considered the ambiguity of previous studies, along with research findings from 29 countries, Heyneman and Loxley (1981: summary), report thus:

"The lower the income (hence, less resources) of the country the greater the power of school and teacher quality to determine a student's academic achievement. Generalizing to the world's population of school children, it is possible to conclude that the predominant inlfluence on their learning is the quality of the schools and the school teachers to which they have an opportunity to be exposed."

4.6. Just why this is so may be construed from the fact that the standard of the human and materfal resources for primary education in the developing countries has not yet reached even the minimal requirement, as aptly described by Heyneman (1983:18) .

"In spite of the significant advances made in providing places in primary schools, these schools in the developing world generally have poor resources and therefore pass on less effectively than they could the increasingly complex skills required of school leavers today. The education of teachers and the availability of furniture, equipment, and materials are normally well below the standards considered minimal for schools in industrial societies."

4.7. A concise statement of the situation in the developing countries in regard to the role of such inputs as trainedtegchers, textbooks, furnitures, etcs may be found in Stephen P. Heyneman's "Improvingthe quality of Education'' (Finance and Development, March 1983:18-21) and in Heyneman and Loxley's "Influences on Academic Achievement Across High and Low Income CountrTes: A Re-analysis of IEA Data" (Sociology of Education, 1982, 55:13-21).

4.8. Asthe proposed study incorporates many variables and complex relationships combined with the ambiguity of many of the previous stud*, it is advisable that a systematic review of the relevant literature be conducted in order to refine the design and instruments for this study. One area to be reviewed concerns the general models that relate school inputs to T-L process and learning outcomes, (e.g.s., Carrol, 1963; Bloom, 1964; Lundgren, 1972; Harnischfeger and Wiley, 1976; Cooley & Lohnes, 1976; Munck, 1979; Anania, 1983; Noonan & Wold, 1983; Chinapah, 1983; Walberg & Haertel, 1984; Ryan & Anderson, 1984).

4.9. mother aspect is that of cost-effectiveness of school inputs in bringing about improved educational outcomes. This is usually sub-sumed under the subject of economics of education. A good starting point for a general review is Blaug's Economics of Education and a collection of works in this area found in Thomas A; Goldman (Ed.), Cost-Effectiveness Analysis, 1967) as well as the World Bank publication series in this area. In this latter series, there are also works on closely related subjects such as CostBenefit Analysis by Psacharopoulos (1980) and Internal Efficiency by Gorham and Tung (1983).

4.10. Most of the above-mentioned studies focus on the cost-

proposed study focusses instead upon the cost-effectiveness of the inputs in bringing about qualitative improvements in the teaching and learning outcome.

' effectiveness of graduate output in quantitative terms. The

I

- 78 - 5. PARAMETERS OF THE RESEARCH DESIGN

5.1. The focal point of this study is the role of school inputs in the teaching-learning process and its outcome. In order to measure whether or not school inputs make any difference, the setting for this study must contain input-variation. Tn other words, there must be differences between schools in terms of curriculum content, teacher background, availability and utilization of facilities and materials.

5.2. Among the many possible sources of input-variation, those that can be and are affected dfrectly and significantly by educational policy, especially in line with the projected policy direction, are of prime relevance to this study. In other words, the variables of interest are those which are associated with the different stages of the implementation of the new curriculum. By focussing of this aspect:,. this study will be able to contribute toward the development and facilitation of the implementation of the new curriculum.

5.3. It is assumed that the cognitive, affective and psychomotor development of children are affected by factors within as well as outside the realm of the school. Hence, in order to identify within-school factors, those variations which are outside have to be isolated. This calls for a desfgn that would control for the outside- school environments.

5.4. It is further assumed that even within the school environment, there are many different types of quality-control inputs which interact with the different dimensions of T-L processes (see Levy, 1977). Hence, in order to enable identification of particular inputs in combination with particular dimensions of T-L processes, the designsbuld be such that variations in the input and process variables can be measured. This calls for a design which is focussed on settings where school-related factors vary maximally and, simultaneously, where outside-school factors (for reasons stated in 5.3.) vary minimally.

5.5. It is assumed also that factors related to classroom or factors that operate upon pupil-groups set the conditions for differential impact of the input-process interactions upon the- outcome. This assumption is based upon the fact that teaching in school environment is conducted largely in a group setting; such as one set of classroom facilities and equipment and one set of teacher-attributes and behaviour for one group of pupils, and another set for another group of pupils. Hence, it calls for setting where there are variations between classes within the same school and grade level.

5.6. Whereas major factors related to the different settings (between-classes, between-schools and outside school) have been taken into account, there remain at least two major factors which may threaten the validity of this study; They pertain to differences in the composition due to selectivity at the entry-level and attrition (dropout) through the grade levels. It is therefore recommended that this study select its setting and conduct its data-collection with Study 3 in order to control for both selection and attrition.

5.7. With the above-mentioned points as parameters, the most efficient and the most relevant design is to focus upon the chofce of settings where the outside-school factors are sfmilar (to satisfy condition 5.3), school-related factors (condition 5.4) and between-classes factors (condition 5.5) vary maximally.

- 79 - 5.8. Considering the above points, it is recommended that the following settings be included in this study: (11 the pilot classes where the new curriculum has been implemented; (21 the adjacent classes, where the new curriculum has not been fully implemented, but have geen affected partially by virtue of the fact that bhese classes are in the same "pilot" schools; and, finally, (3) other schools, sometimes referred in Ethiopia as "transitional" schools, which have not yet been affected by the new curriculum. It is very important that the sampling desegn be such that extraneous factors can be controlled. Thfs would ensure that the three stages of the implementation of the new currfculum are included as well as it would maximize the possibility of isolating the effects of factors related specifically to lmplernentation stage, class, school and outside-school respectively.

6. BASIC MODEL

6.1. The outcome variable in thfs investigation is the qualitative level of learning outcome pertaining to the cognitive, affective and psychomotor domalns of learnfng.

6.2. The qualitative level of learning outcome is seen to be dependent upon (1) pupil characteristics, (2) teacher characteristics and the interplay between the pupil and the teacher in (31 teaching-learnina processes, which are conditioned by (4) teaching-learning material availability and (5) ecological factors (see Conceptual Model in Study 1).

6.3. There are several possible statistical models. One statistical I

model which corresponds to 6.2. is the General Linear Model (GLM), including interaction terms. A simple linear additative model, e.g., path analytic model, can also be employed, in which case, the interaction terms will be disregarded. After having identified the input variables which have shown to have significarkimpact on the teaching and learning outcome, the cost-effectiveness analysis may be conducted by multiple regression analysis.


7.1. The dependent variable is the qualitative level of learning outcome, which is measured by three sub-variables pertaining to the three domains of learning:

Domains Content Examples

cognitive : thinking stages and skills, knowledge in the sciences affective : attitude of scientific enquiry, attitude toward

psychomotor : productive skills, utility of skills to home and work and the community

work life.

7.2. The cognitive component can be measured through written tests, combined with laboratory demonstration, and/or interview; the affective, through questionnaire/interview; and the psychomotor, through task assignment and observation.

7.3. The areas sub-sumed under the term "input-var2ables" include curriculum content, teacher-trainimg, human recruitment (staff and

- 80 - students), teaching and learning materials, and school and classroom facilities. Measurement of these variables wfll require various instruments, ranging from content analysis to inventory. 12sts.

7.4. The T-L areas to be investigated pertain to teaching behaviour, learning behaviour, type and frequency of re-enforcements/sanctions, language of communication, type and frequency of one-to-one or group interaction among teachers, among students and between teacher and students, classroom implemented curr2culum, type and frequency of different learning activites, use of teachfng/learning materials and facilities, etc. Here, the most appropriate means of measurement is the observation technique.

7.5. The ecological factors to be investigated pertain to the community characteristics such as the communfty's economic basis of social organization (urban/rural), literacy level (above/below the national median) and cultural composition (linguistic and religion). of the communities according to the available Awaraja statistics.

Measurement in this context will involve classification

8. STRATEGY OF ANALYSTS

8.1. Data reduction is the first step, in which the items which lack variance should be deleted from further analysis and the remaining items should be recoded, where necessary, and ordered for scaling and constructton of composite variables (indices, factor scores, etc. ) . 8.2. Each of the sub-variables of the domains of lgarning should be analyzed in relation to the "cleaned" variables in order to identify those factors which are empirically correlated with the different outcomes.

8.3. The correlates identified in 8.2. can be used to describe those conditions which are associated with high and low qualTtativeoutcomes.

8.4. Selected variables can be now entered into a statistical model (e.g.s., GLM, ANOVA, Path Analysis) corresponding to the expected pattern of relashionships for hypothesis -testing and estimation of the effect parameters. The model should be tested on the whole sample as well as separately for each of the sample strata, in order to arrive at general and specific tests in respect to different target environments (i.e., urban-rural, high-low literacy area, etc.).

8.5. The input-variables which have been shown to have significant effect on the T-L process variables and learning outcomes should now be costed, i.e., cost estimated, for cost-effectiveness analysis. Multiple regression may be employed for this analysis.

9. SAMPLING PROCEDURE

(See Study 3, Sampling design.)

10. PARAMETERS OF GENERALIZATIONS

10.1. The observations in this study are based upon cross-sectional sample. In drawing inferences about processes, one important factor to be considered is the effect of attrition (dropouts). To the extent that there are differences between the school-classes, the results are correspondingly biased. Hence, the sample must be corrected for differential attrftion probabilities (estimated in Study 3).

- 81 -

10.2. Being a cross-sectional study, it does not enable analysis of variation across time beyond the span covered directly by the data collection instruments and observation,

10.3. Estimations based upon data unadjusted for the sampling design should either be reported for the respective strata and generalizations be confined accordingly to them or the stratification variables be explicitly included in the computational procedure as in multivariate analysis.

10.4. To compute population qarameters, the data must be first

to the respective sampling strata. .adjusted for sampling design through wefghting of the cases belonging

11. OPERATIONAL OVERVIEW

11.1. The present study will begin with a literature review of the major studies on the effects and cost-efsctiveness of school inputs on the T-L process and the quality of learning outcomes and areview of the relevant findings of the Phase I of the Research andevaluatin of the Quality of Education in Ethiopia. To the extent required, the staff should recefve training fn the conduct of a cost-effectiveness study. This preparatory phase should result in written reports on the major hypotheses, operationalization (identification of the relevant variables, measurement scales, questionnaire/interview/ observation items, etc.) and specification of the strategy. for analysis (statistical models) . 11.2. In the second phase of this study; there are three tasks to be conducted simultaneously. One task is the collection of data on the school and community settings which will be used for a study of the settings themselves; both of which (i.e., the data and the setting-study) will be used for samplingpurposes. The product of this task should be a data file on the characteristics of the community and.the schools and a written report on the statistical distribution of the characteristics of the communities and the schools.

11.3. The second task is the development of the instruments (in.line with the results of the above-mentionned review reports), in the course of which a "pilot" study, with a small sample, must be conducted in order to test the instruments and the fieldwork. The analysis of the results should be directed toward the improvement of the instruments and the fieldwork. This task will be completed upon production of the final set of instruments.

Note. Some items used in the Phase I of the Research and Evaluation on the Quality of Education may be considered for inclusion in this present study. In that case Phase I may be considered as the pre-test. Item analysis can then be carried out on the basis of the data from Phase I. This will no only shorten the required time, but also cut down the cost immensely, because of the opportunity for selection of the "best" items.

-

11.4. The third task within bhe second phase of the study is the training of the fieldworkers. This should take place within the context of the pilot-study setting where the instruments are to be tested so that the f2eldworkers will get experience with both the

- 82 - instruments and the real setting. Their criticisms and suggestions should be solicited and collected systematically fn regard to both the instruments design, content and wording as well as the method of administration. Their comments will be a valuable source for the improvement of both the instruments and ffeldwork. This task should produce a corps of trained field workers and an instruction manual for them. Care should be taken to produce a standardlzed set of instructions so that all field workers will be followfng the same procedure.

11.5. The third past consists of three tasks: sampli'ng, record- keeping and preparation of a data processing system.

11.6. The first task 2s samplfng, which should be conducted well before the start of a new academic year, and communication with the schools that have been sampled. The selected schools must be notified just prior to the start of the academic year, because of the requirement for systematic record-keeplng (see next paragraph) starting from the beginning of the academic year. The sampling task is completed when the requfred number of schools have responded that they have agreed to partfcipate and have started with the record- keeping. A written report on the sampling procedure and the initial response rates for the respective sample strata should be issued.

11.7. The collection of data starts immediately on the part of the school staff designated for the purpose of record-keeping. With the first communication to the schools, record-keeping forms and an accompanying letter of instruction will be sent. It is important that individuals at each school be designated as coordinators to provide continuity in the cooperation between the school and the research unit.

11.8. Another task is the preparation 'of the data-capture routines, data management system (DBMS) and the programme for fieldwork administration (see section belowl. The data-capture Programme should routinely screen for wild codes and store directly into the DBMS. Special care should be taken to design the DBMS in such a way that the the files can be linked through standardized case-identification fields. Data for each pupil, for example, must be linked to his/her own teacher's data. It is mandatory that the DBMS are compatible with the statistical analysis programme package. The data-capture routines, the DBMS and the statistical analysis system should be designed to enable easy statistical output directly after data entry. As part of this preparatory work, the data-collection instruments should be checked again for their compatibility from the viewpoint of data-processing.

11.9. The fourth phase is centred around the main data collection. 1

11.10. The final version of all the instruments must be printed. By use of the sampling report, instruments should be packaged appro- priately in accordance with the sample size for each sampled group (school and class).

11.11. The fieldwork administrative computer programme should serve to keep track of progress fn the fieldwork. One of the first tasks here will be the assignment of the fieldworkers. The record of each workers's progress should be updated so that the current status can be reported at any time. Here, records on target groups should also be updated as the data collection makes progress so that appropriate action can be taken immediately.

- 83 - 11.12. The actual ffeldwork out in the schools and communities should start with a conference with the school coordinator assigned at the beginning of the term fn order to get acquainted with the conditions that impinge uponthe data collectfon work; Subsequently, the fieldwork team can proceed with observation, interviews and administration of the questionnaires. Asfde from the instruments, the fieldworkers should be provided with a mandatory check-off list and report forms for reviey and examhation of the work at the end of each data collection session. worker must review the contents of the filled-in instruments account for any pezuliarities and mfssing cases. should arise, the worker is while he/she is still at the collection site.

These forms must be such that the and a

If any ambiguity then able to follow it up immediately

11.13. The completed instruments and reporting forms should be delivered to the fieldwork director, who will check their completeness and correcteness. data processing and file-building in a DBMS environment (see section 11.8.) .

These will then be sent to the data entry unit for

11.14. The fifth phase will be data analysts, which consfsts of data reduction, scaling, index construction, statist2cal analysis and graphic display. In order that the operationalizatfon rement and statistical carried out. Each task

11.15. The sixth phase the input variables on

the work of this phase be efficient and useful, of the variables and specifications of the measu- models (see section 11.1.) must have already been carried out must be meticuously documented.

is the interpretative analysis of the effects of the'teaching and learning outcomes. The iden-

tified-"significant" variables will be used in the cost-effectiveness analysis. The final report should contain not only the findings of the cost-effectiveness analysis, but also policy-implications in regard to acquisition and allocation of the school resources in order 'to ensure their optimal impact on the implementation of the new curriculum throughout the country.

12. NOTE ON COMPUTER USAGE

Direct access to interactive computer usage on a daily basis throughout the duration of the project is absolutely essential. cut down time and cost and increase precision and efficiency, a computer is required from the very start of the project: instrument construction, sampling training, data-entry, coding, file-building in DBMS, statistical analysis to project adninistration and report-writing. (See above section on operational overview; and, for hardware and software requirements, see Tungs' separate report. )

To

from

. .

- 84 - STUDY 5

From Plan to Action A study of planning, management and imp lenent a t ion

processes, needs and oSstacles at Awaraja and school level.

Purpose

for Awaraja Pedagogical Centres (APCs) and School Pedagogical Centres (SPCs) and to assess the relatronship between them from the standpoints of planning, management and Implementation.

The main purpose of thfs study is to examine the objectives set

Desired outcome

To provide an understanding of the relationship between the APCs and the SPCs and to identify the needs and obstacles of these two types of institutions 2n ensuring quality improvement at the primary school level. To study the importance and relevance of in-service training and innovative appropriate technology to the classroom teaching-learning processes.


1.1. After the 1974 revolution, the national call for mass education was partly satisfied by a n?id quantitative increase in the educational system. This led to a serious qualitative deterioration of education in the country. Among other things, the immediate need for radical educational reform was self-evident.

"A drastic reform of the education system became necessary. Some of the reforms undertaken by the Ministry were the setting educational goals, the bringing of schools under the control of the people and reorganization and decentralization of the administrative structure. APCs were thus created to serve as supporting arms in the implementation of the decentralized scheme, primarily for the purpose of facilitating the operation of the teaching-learning process." (M0E:Working Document, UPEL Workshop, 1984d.)

1.2. The social demand and pressures for mass education gave rise to very heavy school-enrolment figures. Many schools had to be created in rural areas. Human, physical and material inputs for educational development were either inadequate or of an inferior quality. For example, a large number of untrained teachers were employed.

1.3. The implementation of the educational reform emphasized 'the need for in-service training of teachers, school-head teachers and super- visors. It was also felt that community participation and leadership was a vital instrument for implementing educational reform. While community participation and leadership can contribute to making education responsive to social needs at the local level, equally important was the role of the community in mapping out human and material resources effectively and to lay the foundation for self-reliance in educational provision at the local level.

I

- 85 - 1.4. Most of the 106 Awarajas today have their own APC. However, many schools still do not have their own SPC. APCs are seen as having a "multiplier effect" on the qualitative improvement of education in Ethiopia. As centres of change and innovation, they carry out a series of activitiss (e.g. in-service training of educational personnel, mapping and usage of human and material community services, development and dissemination of innovative appropriate technology and instructional aids, and adaptation of the national curricu'kum to local needs).

1.5. The APCs have four major functions. First, they serve as a platform for coordination - to promote the active participation and contribution of the community to the development of the education system. Second, APCs serve as centres for research and planning. Third, they play a vital role in upgrading educational personnel through workshops, seminars, study visits, etc. Fourth, APCs serve as a forum for innovation and development of appropriate technology - instructional materials and teaching aids' prototypes. 1.6. Implementation of the objectives set for APCs can be better examined through the works of the SPCs. SPCs are regarded as a forum for the application and dissemination of the knowledge, innovative experiences, and skills acquired by the APCs' trainees at school level.

"It is an educational support component where teachers take a pioneer role in organizing and leading the instructional activity. It is a centre where the beneficiaries of the education system get togeaher to discuss common educational problems and issues to seek possible solutions; it is a research and innovation centre for teachers, students and others; it is a centre for pookgthe knowledge, experiences and talents of teachers, students and community resources (human and material)

(MOE, 1982, p. 28.) - for improving and enriching tkquality of education."

1.7. Studying the structural liaison between APCs and SPCs and as se s s ing the imp l.ement a t ion process e s at the SPC 1 eije l--may fuX- in understanding the needs and obstacles that are encountered during the implementation procedures. These needs and obstacles may be conditioned by institutional, managerial, communicational, attitudinal, and resource factors. It is to be hoped that the present s.tudy (although limited in scope) will provide the necessary source of information for a larger study on planning and implementation.


2,l. The overall aims of this study is examine the planning, management and implementation processes, needs and obstacles'at Awaraja and school level. The objectives set for APCs and SPCs are studied and the relationship between them assessed from the standpoints of planning, management and implementation.

The specific aims of the study are:

2.2, To assess at the SPC level the impact of in-service training of the educational personnel (APC's trainees) on the teaching-learning processes.

2.3. To study t

- 86 - .e dksseminat-m of informat-on w,,,, regard to

innovative appropriate technology and teaching practice from the APC to the school level.

2.4. To study the managerial and coordinating functions between APCs and SPCs and the needs and obstacles met at the stage of implementation.

2.5. To examine what records (information-base) are kept about APC trainees in respect to teaching-learning methods and use of innovative appropriate instructional aids and materials in classrooms.


The present study is divided into two main parts.

3.1. Part I of the study deals with an evaluation of the objectives and the functions of APCs and SPCs in general. Field visits accompanied with interviews and observation check-lists will be carried out out in some selected ARCS and SPCs to record the various activities and managerial and operational constraints of these institutions. The informants will be APC coordinators and trafners and SPC coordinators.

3.2. Part I1 of the study will focus on the actual inplementation procedures at the SPC level. Records of the participants (former APC trainees) will be examined. Participant observation will be carried out in the classrooms of former APC trainees and non-APC trainees to examine the degree of variation in teaching-learning methods, usage of instructional aids and materials, and the teaching-learning processes in general.


4.1. The present study should be viewed as a pilot for testing possibilities in developing appropriate evaluation and research methods/ techniques of evaluation for examining the mutually dependent relationship between planning and implementation. The wide coverage of the study of educational planning in Phase I of the Quality of Education Project made it practically impossible to stress the implementation aspects of educational planning.

4.2. There are important areas where the dynamics between educational planning and implementation may be studied. It is hoped that the approach in choosing to study APCs and SPCs can only be served as an example. Being pilot in nature, the results from the present study need to be properly considered for the launching of further studies (of a wider scope) on educational planning and implementation in Ethiopia.


5.1. It is generally assumed that there is a watertight relationship between planning and implementation. However, the amount of lite- rat= available on educational planning as compared to that on educational implementation can be taken as an example to.illustrate that such a relationship makes more sense for theoreticians and decision-makers than for practioners of implementors. Planning may be regarded as a political exercise in optimism while implementation is a monitoring exercise in pessimism.

- 37 - 5.2. The links between planning and implementation are highlycomplex. They involve an interactive as well as a negociating process involving many participants, e.g. policy-makers, administrators, curriculum planners, educators, pedagogues and students.

"It is (therefore) impossible to study a process without giving consideration to the actors and structures, or without delimiting the scope of what is negot2able, without taking into account the nature of the arguments put forward by the different actors involved, or without communication procedures." (Unesco-IIEP, 1984, p. 57.)

5.3. The decentralization or deconcentration processes of Planning and implementation in the APC-SPC model in Ethiopia may be only realized if the conditions for training the local personnel in administration and management are met and the information and communication system improved. Weiler (1984) argued that competent, flexible and informed planning of education at the local level required trained personnel in administration and magagement as well as complete, up-to-date and accurate information on resource alloca- tions, staff assignments, material supplies,building maintenance and the like. Among other things, the condition of schools, the backgrounds and distribution of teachers, the demographic characteristics of students,the supply of teaching materials and other instructional aids, and the options in financial allocation are factors that are decisive for planning, implementation and management of the educational system at all levels.

5.4. In their survey of educational planners and administrators from some 25 developing countries trained at the International Institute for Educational Planning - Unesco, Paris, Chinapah and Lbfsted (1982) observed that despite the considerableattention, at least in theory, given to information and communication needs in educational planning, management and implementation, very little is known of the actual problems and constraints in the three phases of planning, management and implementation. Those involved have not been seriously consulted.

5.S. The survey also indicates that there is a huge difference between ideal planning and real-. planning. is generally underestimated. Planners are looked upon as "technical implementors or fact-finders". Besides economic constraints in the implementation of educational plans, it was observed that the major actors did not in principle possess the necessary skills and competence.

5.6. In the UNICEF document "Moving towards Universal Primary Education and Literacy" (April 19831, several indicators were given for effecting planning, management and implementation of the UPEL strategy at the local level. The building-up of the institutional capacity to manage and monitor the system of UPEL and developing an organizational structure that allows a high degree of local responsibility and community participation were regarded as important features of the strategy for support through cooperative programme.

The professional role of planners

5.7. The main features of the Ethiopian APC-SPC Model provide the necessary inputs for a research and evaluation of planning, management and implementation of education at the local levels. Several domains of needs and obstacles can be studies. In other words, the study may shed light on the needs and obstacles in planning, management and implementation in the following domains: (a) structural

- 88 - and organizational, (b) training and compe-ence, (c) information and communication, (d) particfpation, and (e) monitoring and evaluation.

6. PARRMETERS OF THE RESEARCH DESTGN

6.1. This study contains three unit-levels of observation: APCs, SPCs and classroom. However, classroom observations are conducted in Study 3, from which the results will be obtained of the purposes of this study as well. The desfgn of this study will, therefore, be limited to the other two units of observation - the APCs and the SPCs. At the APC level, the design should provide a comparison of the sampled APCs. This comparison should focus on similiarities and differences between APCs in terms of their objectives and operational means to promote teacher-training services, to develop innovative appropriate technology, and to ensure a linkage in the implementation process with their respective SPCs.

6.2. The design at the SPC level would serve a dual purpose. First, it muld focus on the managerial and implementation aspects. SPC coordinators and APC trainees will be considered as units of observation. Second, the research design will incorporate non-APC's trainees in order to allow at a classroom level, the observation of variance between the group of APC trainees and the group of non- APC trainees in terms of their behaviour in classroom management, usage of instructional materials, pupil-teacher interactions, and creative skills and innovations in teaching-learning activities.

6.3.,It is assumed that APCs are operating, to a great extent, under a decentralized framework and, hence, constrained within the existing conditions and means of their respective provinces. Variations in implementing the common set of objectives for APCs would therefore be conditioned and/or constrained by the socio-cultural, political, and economis settings across.the provinces of the country. The conditions of the teaching-force, the need for upgrading the teachers, the attitudes and commitment of the community, the resource-base (human and financial), the managerial and administrative competence and skills of the coordinating organs, and the infomation- communication network, would altogether determine the extent to which the common set of objectives for APCs are met and the possi- lities or opportunities given to account for specific needs of the respective provinces.

6.4. The points raised above (6.3.) have direct implications on the activities of the SPCs in the respective Awarajas.. Whereas the general socio-cultural, political, and economic conditions in the Awarajas would have a direct bearing on their SPCs, variations or similarities from the standDointof management and implementation at the SPC level can nonetheless be observed. Are managerial issues centred around the conditional factors mentioned above, or are they centred around the individual skills, competence, attitudes and control mechanisms (records and supervision) of the individual SPC coordinator? Do APC trainees share a common set of skills, competence and attitudes towards the teaching-learning processes? Are the differences more marked among APC trainees than between APC trainees and non-APC trainees?

6.5. In view of these questions, it is suggested that (a) variations at provincial-level (socio-cultural, political, and economic); (b) variations at the school and/or classroom-level (APC trainees and non-APC trainees) be accounted for in sampling procedures, instrument construction, data collection and data analyses.

- 89 - 7. BASIC MODEL

The basic model for this study is based on the official channel of policy implementation and dissemination of educational innovation along the chain from APC to SPC to classroom. It is assumed that, along this channel, the APC 2n-service training and its development of teaching aids have some impact on the SPC, which in turn, influence the actual classroom teaching-learning environment and practice. Furthermore, the outcome at the SPC and classroom levels are also affected by local environmental factors.


8.1. The operationalization of the variables in the model includes (a) a set of environmental indicators common to Study 1 and Study 3 for the APC unit of observation; (b) a set of frame factors examined through observation checklists at APC and SPC levels; (c) a set of attitudinal items for APC and SPC Coordinators as well as APC trainees constructed with the help of structured interviews; and (d) a set of items devoted to participant observation at the classroom level to account for actual implementation processes.

8.2. The outcome of the dependent variables in the model refers to school-level observations in respect to (i) administrative linkage/ or barriers; (ii) human and financial resource constraints; (iii) lack or presence of managerial skills, competence and positive attitudes; (iv) lack or presence of necessary skills, competence and attitudes among APC trainees for innovative teaching-learning practices; and (v) lack or presence of the required control mechanisms (record-keeping, monitoring and supervision) for ensuring implementation at the SPC and classroom levels.

8.3. It is clear from the above that the nature of the present study does not allow for an elaborated statistical approach to the quanti- fication and analysis of the information collected. This is partly due to the qualitative type of information searched for and partly to the qualitative methods chosen to gather such information. However, information pertaining to the independent variables that describe the general environmental settings (from Study 1 and Study 3) and those that are gathered from the check-lists at APC and SPC levels can be easily compiled for statistical analytical purposes.

8.4. Data pertaining to process and outcome variables in this study are more of a qualitative nature. It is therefore suggested that data collected through structured interviews and classroom participant observation be reported in the forms of field notes, impressions and views.

9. STRATEGY OF ANALYSIS

9.1. The first part of the quantitative analysis pertaining to environmental indicators (point (a) in 8.1. above) should be just a secondary or meta-analysis of the data from Study 1 and Study 3. The second part of the quantitative analysis based on data from the observation check-list would consist of descriptive summary statistics (univariates with frequency tables - percentages, modes, means and standard deviations) for items on the check-list and their breakdowns by specific categories of Awarajas and the SPCs they are associated with.

- 90 - 9.2. The major analytical, part of this study will be based on detailed summaries from structured interviews with APC and SPC coordinators and APC traznees. Tn addition, the field notes from participant observation fn APC trainees and non-APC trainees classrooms will be organized systematically according to the areas under investigation (see 8.2.1 and the two groups will be compared.

10. OTHER DESIGN CONSTDERATTONS

The fieldwork related to SPC and classroom observations for this study is to be conducted parallel to Study 3, which is based upon school sample. Hence, the sampling procedure in respect to the school and classroom observations is that of Study 3. Awaraja level data as well as its descriptlve statistics will be derived from Study 1. The APCs are to be sampled from the Awarajas selected for Study 1, which in any case should be the same even for Study 3. Consequently, readers are referred to the relevant sections in these two study-proposals.

...

- 91 -

SELECTED LIST OF REFERENCES

Alexander, L., and Simmons, J. 1975 The Determinants of School Achievement in Developing

Countries: The Education Production Function Model. World Bank Staif Working Paper 201. Washington, D.C: The World Bank.

Anania, J. 1983 The Influence of Instructional Conditions on Student

Learning and Achievement. Evaluation in Education, Vol. 7 No. 1

Anderson, C.A. 1983 Social Selection in Education and Economic Develop-

ment. Paper presented for the Education Department of -World Bank. Washington, D.C: The World -Bank.

Babble, E.R. 1973 Survey Research Methods. California: Wadsworth Inc.

Beeby, C.E. 1969 Qualitative Aspects of Educational Planninq. Paris.

Blalock, H.M. 1971 Causal Models in the Social Sciences. Chicago: Aldine

Press.

Blaus, 14. 1967- Economics of Education: A Selected Annotated Bibliog-

raDhv. Third Edition. Vol 3, Perqamon Press, England. A - - -

Bloom, B.S. 1964 Stability and Change in Human Characteristics.

.New York: John Wiley & Sons.

. Bloom, B.S. 1976 Human Characteristics and School Learning. New York:

McGraw-Hill.

1. Bowles, S., and Gintis, H. 1976 Schooling in Capitalist America: Educational Reform

and the Contradictions of Economic Life, New York: Basic Books.

Brembeck, C. 1978 Formal Education, Non-Formal Education and Expanded

Conceptions of Development, Occasional Papers No. 1, Non-Formal Education Information Center, Michigan State University.

Bock and Papagiannis. 1976 The Demystification of Nonformal Education. Universi-

ty of Massachussets.

- 92 -

P Carroll 1963 A Model of School Learning. Teachers College Record,

64, 723-733

Carron, G., and Chau, T.N. (Eds) 1980a Reduction of regional disparities: The Role of educa-

tional planning: Diagnosis and Policies for reduction. Paris: Unesco, IIEP.

Carron, G., and Chau, T.N. (Eds) 1980b Reqional Disparities in Educational Development: A

Controversial Issue. Paris, Unesco, IIEP.

Chinapah, V. , and Fagerlind, I. - 1979 The Role of Education in the Basic Human Needs

Strategy. Research Report No. 39. Stockho1m:Institute of International Education, University of Stockholm.

Chinapah, V., and Lijfstedt, J.I. 1982 Educational Planning in the Developing Countries:

Problems and Research Needs. Oslo Seminar Paper on Education in Developing countries, Oslo, Norway.

Chinapah, v. - 1983 Participation and Performance in Primary Schooling:

A Study of Equality of Educational Opportunity in Mauritius. Studies in Comparative and International mucation No.8, University of Stockholm, Sweden.

Coombs, P., and Ahmed, M. 1974 Attacking Rural Poverty: How Nonformal Education can

Help. IBRD, Washington D.C: The World Bank.

Coleman, J.S. et al. 1966 The Equality of Educational Opportunity. Office of

Education. Washington, D.C: U.S. Department of Health, Education, and Welfare.

Coleman, J.S. et al. 1981 High School and Beyond: A national longitudinal study

for the 1980s. Chicago,Illinois: The National Opinion Research Centre.

Cooley, W.W., and Lohnes, P.R. 1976 Evaluation Research in Education: Theory and Princip-

les,and Practice. New York: John Wiley & Sons.

Dave, R.H. 1963 The Identification and Measurement of Environmental I

Process Variables that are Related to Educational Achievement. University of Chicago: Unpublished Ph.D dissertation.

Davis, R.G. 1980 Planning Education for Development: Volume I, Issues

and Problems in the Planning of Education in Develo- ping Countries. Cambridge, Massachussets.

- 93 -

Ferquson, G.A. 1981 Statistical Analysis in Psychology and Education.

Fitth Edition, IlcGraw-Hill Kogakusha, Ltd. Tokyo, Japan.

Faigerlind, I. 8 and Leal, L.C. 1982

Gilpin , 1979

Gorham, 1983

Gorham, 1983

P4athematics in Portugal: The influence of School and Family Environment on Results in Grades 7,8 and 9. Research Report No. 55. Stockholm: Institute of Inter national Education, University of Stockholm.

C.W. Issues in Non-Formal Education and Training for Rural Development. World Bank, Department of Education, Washiniton D. C.

A. Non-Formal Training Programmes for Rural Skill-Deve- 'loprnent: A State of the Art Study. .SIDA, Stockholm.

A., and Tung, KC. Internal Efficiency of Higher Education in Developing Countries. Staff Working Paper (revised edition torthcoming ) , Washington D. C : The World Bank.

Goldman, T.A. (Ed.) 1967 Cost-Effectiveness Analysis. 1JY:Praeger

Green, G.D. 1981 Training for self-reliance in rural areas. Interna-

tional Labour Review, Vol 120. No.4

Harnishfeger, A. , and Wiley, D.E. 1976 The Teaching Learning Process in Elementary Schools: A

Synoptic View. Curriculum Inquiry, 6: 5-43.

Heyneman I S. P. 1976 A Brief Note on the Relationship between Socioeconomic

Status and Test Performance among Ugandan Primary School Children. Comparative Education Review, 20: 42-47

Heyneman S. P. 1983 Improving the quality of Education. Finance and

Development.

Heyneman, S. P. , and Loxley, W. 1981 The Impact of Primary School Quality and Academic

Achievement Across Twenty-Nine High and Low Income Countries. Paper presented at the Annual Meeting of the American Sociological Association, Toronto.

Heyneman,S.P. 8 and Loxley, W. 1982 Influences and Academic Achievement Across High and

Low Income Countries: A Re-Analysis of IEA Data. Sociology of Education, 55: 13-21.

- 94 -

House, E.R. 1980 Evaluating with Validity. Sage Publications, London.

Husen, T. 1974 Talent, Equality and Meritocracy. The Hague: Flartinus

Nijhoff.

Husen,T., et al. 1978 Teacher Training and Student Achievement in Less

Developing Countries. World Bank Staff Working Paper No. 310. Washington, D.C: The World Bank.

Jencks, C., et al. 1972 Inequality: A Re-assessment of Family and Schooling

in Anerica. New York: Basic Books.

Johnstone, J.N. 1981 Indicators of Education Systems. International

Institute for Educational Planning, Unesco, Paris.

Jljreskog, K.G., and SiSrbom, D. 1981 LISREL IV- Analysis opf Linear Structural Relations-

hips by the Method of Maximum Likelihood. Chicago: National Educational Resources.

JBreskog, K.G., and Wold, H. (Eds) 1982a Systems' under Indirect Observation. Part I.

Amsterdam: North-Holland Publishing Company.

JGreskog, K.G., and Wold, H. (Eds) 1982a Systems under Indirect Observation. Part I1

Amsterdam: North-Xolland Publishing Company.

Kerl.inger, F.N. 1981 Foundations of Behaviorial Research. Second Edition,

Holt,Rinehart and Winston, Inc. New York.

Kish, L. 1965 Survey Sampling. New York: John Wiley and Sons.

Lewy, A. (Ed. ) 1977 Handbook of Curriculum Evaluation. International

Institute for Educational Planning, Unesco, Paris.

Lundgren , U. P. 1972 Model Analysis of Pedagogical Processes. Lund :

Gleerup.

Malpica, C. N. 1980 Educational Administration and its Relations with

Planning and Research. IIEP, Paris.

Mar joribanks, K. 1974 Environments for Learning.. Slough, England: :PER

Publishing Company Ltd.

- 95 -

Messele , M. 1984 Universizing Primary Education and Literacy in Ethio-

pia. Paper prepared for the Advanced Training Prog- ramme in Educational Planning and Administration, IIEP, UnescO, Paris.

-

Ministry of Education- MOE (Ethiopia)

1977

1981

1982a

1984a

1984b

1984c

1984d

1984e

MINEDAF 1982

MOE: Basic Information on Education in Ethiopia.

MOE: Some Explanatory Notes on Educational Development in Ethiopia.

MOE: Awraja Pedagogical Centre: What it is and How it Operates.

MOE: The Case of Ethiopia MINEDAF 5. The Conference of Ministers of Education and those responsible for Economic Planning of African member States on Educa- tional Policy and Cooperatian.

MOE: Education in Socialist Ethiopia

MOE: Introduction to Research and Evaluation Project on the Quality of Education in Ethiopia.(Draft)

MOE: UPEL -Proposal for Supplementary Funding

MOE : Working Documents - UPEL Yorkshop MOE: Ethiopian Education Team Mission to Sweden.

Development of Education in Africa: A Statistical Review. Conference of Education and those Responsible for Economic Planning in African Member States. Paris: Unesco.

Moser and Kalton 1971 Survey Methods in Social Investigation. London :

Heineman.

1979 Model Building in Comparative Education. Applications of the LISREL Method to Cross-National Survey Data. IEa Monograph Studies No. 10. Stockholm: Almqvist & Wiksell International.

Nie, H.H. et al. 1375 SPSS Statistical Package for the Social Sciences.

Second Edition. New York : McGraw-Hill.

NFE 1982

Exchange Linking Formal and Nonformal Education. The W E Exchange. No. 238 Michigan State University.

- 96 -

Niehoff, R.O. (Ed.) 1977 Non-Formal Education and the Rural Poor. Michigan

State University.

Noonan, R. 8 and Wo1d.H 1983 Evaluating School System Using Partial Least Squares.

Evaluation in Education, Vo1.7, No. 3.

Paulston and Leroy 1975

Patton 8 1980

Peaker , 1975

Plowden 1967

Strategies for Non-Formal Education. Teachers Record. Vol. 76, No.4.

College

M.Q. Qualitative Evaluation Methods. Sage Tublications, London.

G.F. International Studies in Evaluation VII: An Empirical Study of Education in Twenty-one Countries: A Techni- cal Report. Stockholm: Almqvist & Wiksell Internatio- nal/LJew York: John Wiley & Sons.

Report Children and Their Primary Schools. A Report of the Central Advisory Council for Education (England) 11: Research and Surveys. London: Her Majesty’s Stationery Office.

Psacharopoulos, G. 1980 Higher Education in Developing Countries: A Cost--

%enekit Analysis. Staff Working Paper No. 440. Washington, D. C : The World Bank.

Ryan, D., and Anderson, L. (Eds.) 1984 Rethinking Research on Teaching: Lessons Learned from

an International Study. Evaluation in Education, Vol 8. NO. 2.

Schiefelbein, E., and Simmons, J. 1979 The Determinants of School Achievement: A Review of

the Research for Developing Countries. Canada: IDRC-MR9

Schiefelbein, E. et al. 1983 The Influences of School Resources in Chile: Their

Eftect on Educational Achievement and Occupational Attainment. World Bank Staff Working Papers No. 530. Nashington, D.C: The World Bank.

Selltiz,. C. et al. 1976 Research Methods in Social Relations. Third Edition.

New York: Holt, Rinehart and Winston.

1980 Pre-School Age Investment in Human Capital. In Educa- tional Dilemma, (Simmons, J. Ed.) p. 97-111

- 97 -

Shaeffer, S., and Nkinyangi, J.A. 1983 Educational Reseakh Environments in the Developin2

World. IDRC, Canada.

SIDA 1980 Primary education and non-formal education in Ethio-

pia. SIDA, Stockholm. - Silver, P. 1983 Educational Administration: Theoretical Perspectives

on Practice and Research. Harper & Row, Publishers, Inc. New York.

Simkins, T. 1977 Nonformal Education and DevelopInent . Manchester

I-lonograph, No. 8.

Simmons, J. (Ed.) 1980 The Education Dilemma. Oxford: Pergamon Press Ltd.

Simmons, J., and Alexander, L. 1978 The Determinants of School Achievement : A Review of

the Research. Economic DeveloDment and Cultural - Change.

UNESCO * 1982 From Planning to Plan Implementation. In Seven

Training Modules. Unesco Regional Office For Education in Asia and the Pacific, Bangkok, Thailand.

UNESCO - I IE P 1984 Educational Planning in the context of current deve-

lopment problems: An IIEP seminar. Unesco-IIEP, Paris.

UNICEF 1983 Moving Towards Universal Primary Education and Lite-

racy: A co-operative programme to foster universal primary education and literacy. UNICEF-E/ICEF/ 401.

Walberg, H., and Haertel, 6. 1984

Weiler, 1984

Weiler, 1980

Community Influences on Learning. Evaluation in Edu- cation, Vo1.8 No.1.

H. Nm Bank Group Education Sector Policy Paper. Final Draft African Development Bank.

H.N. (Ed.) Educational Planning and Social Change. Unesco/IIEP, Paris.

Wold, H. 1975 Path Models with latent variables: The. NIPALS app-

roach. In H.M. Blalock et al. (Eds.), Quantitative Sociology. New York: Academic Press.

- 98 - ANNEX r1:

LIST OF NATIONAL OFFICIALS AND INTERNATIONAL EXPERTS MET DURING THE MISSION

H. E. Mr. Bili-llign Manfred0 Mfnister of Education

(i) National Off fcials

Mr. Tekle Ayano

Mr. Dender

Mr. Asefa

Mr. Tesfaye Dubale Mr. Dendir Dansamo Mr. Mazengia Makommen Mr. Asefa Sileshi Mr. Altaye Gizaw Mr. Assefa Beyene Mr. Derese Mengistu Mr. Bernanu Malore Mr. Berhanu Dbaba

(ii) International Experts

Dr. Barbara Junge

Mr. Alan Court Mr. Awad Idris

Head, Research and Evaluation Division, MOE Research and Evaluation Division, MOE Research and Evaluation Division, MOE Head, Curriculum Department, MOE Curriculum Department,

11 I1

11 I1

II 11

11

11

11 11

11 11

MOE I1

~ ,UNICEF Project in Ethiopia

'UNICEF Programme Officer Unesco Representative to Ethiopia