Back to Home Page of CD3WD Project or Back to list of CD3WD Publications


3. Methodological issues

3.1 Introduction
3.2 Methodology: Towards a more culturally appropriate approach
3.3 Methods: Using life histories to tell us more about real life in developing countries
3.4 The research project

3.1 Introduction

If Culture is indeed both knowledge and ideas that give meaning to beliefs and actions and the ideational tool that describes and evaluates that action, then it is appropriate for research itself to be viewed as a particularly significant area of human activity.

In developing a series of questions to frame the enquiry, attention was given to the wish to develop culturally-sensitive research methodologies and methods which would not only serve the research well in the gathering of good quality data but would advance our knowledge of how best to investigate educational and social problems, particularly in a setting different from the principal researchers.

This meant paying particular attention at the outset to the broad approach or methodology that was to be used and later to the range of techniques or methods that would be developed to gather the data. Finally, thought would be given to who would be involved in the research and to the ownership and dissemination of the results.

In trying to carry out research which is both of high quality and yet culturally-sensitive it is possible to identify four major questions:

· What, in terms of cultural factors need to be identified in the content and methodology of the research?

· Where, in terms of locus of control, will the research and publication be carried out?

· Why, in terms of linkage to development goals, will the research be done?

· Who, in terms of personnel will be involved and to what extent will the research be empowering and reflexive for the researcher and researched?

3.2 Methodology: Towards a more culturally appropriate approach

The writer's own extensive research in a number of African cultural settings (detailed elsewhere in Vulliamy, et al. 1990) and his interest in the promotion of qualitative research gives weight to the view that it is within qualitative research methodology we will find a more suitable way forward. In particular we can identify six common characteristics of this approach which also guided the Ghana study:

a) The focus on meanings and the attempts to understand the culture of those being studied predisposes researchers to work as far as possible in natural settings (Denzin, 1971). In terms of methods this suggests, for example, a preference for participant observation rather than experiments under artificial conditions, and preference for informal and less standardised interviews rather than for more standardised and formal ones. The use of life history as we shall see, also attracts as a particularly meaningful research method.

b) Rather than testing preconceived hypotheses culturally-sensitive research aims to generate hypotheses and theories from the data that emerge, in an attempt to avoid the imposition of a previous, and possibly culturally inappropriate, frames of reference (Glaser and Strauss, 1967). There are two important implications of this. First, it implies a greater degree of flexibility concerning research design and data collection over the duration of the research project; and secondly, it implies that the process of analysis occurs simultaneously with the process of data collection.

Though the Ghana study was guided by the six research questions reported in the introduction the continual process of analysis played a significant part in steering the on-going collection of data and in the generation of meaningful results.

c) In focusing on the processes of social interaction, qualitative research involves the ongoing collection of data rather than collection of material at discrete points in the research process. Culturally appropriate research in development is therefore more likely to be concerned with the process of implementation than with innovation outcomes. In the case of the Ghana study researching two field sites also meant that lessons and refinements learnt from one could be applied to the other.

d) Qualitative, culturally-appropriate research is holistic, in the sense that it attempts to provide a contextual understanding of the complex interrelationship of causes and consequences that affect human behaviour (Goetz and Le Compte, 1984). As stated earlier the focus needs to be more on context as research environment rather than context as background to the study. A consequence of this holistic emphasis is that qualitative research within the development field tends to incorporate a wide variety of specific research techniques, even within one research project. As we shall examine later a case can be made for increased use of research methods that relate specifically to patterns of local knowledge and the transmission of cultural meanings (Vulliamy, et al. 1990).

In Ghana this meant giving priority to the initial selection of research sites and the gathering of relevant contextual data. Becoming sensitive to and grounded in the contexts of the research sites involved an intensive initial period (of about six months) during which regular visits were made to both areas, previous researches conducted nationally and locally were consulted, and key individuals with an interest in the location were approached and listened to.

e) The validity or explanatory power of qualitative research depends on the researcher's ability to understand the relationship between macro- and micro-analytical levels of data collected, and to establish cross-cultural comparisons and contrasts. To a certain extent this resolves the earlier problem of objectification.

By choosing two field sites of very different characters - one in the South of the country in a more prosperous, largely Christian and matrilineal community; the other in the poorer, Northern community, Islamic and patrilineal it was intended that cross-cultural comparisons and contrasts could be drawn at a national level.

This micro-level case study data is also counter-pointed against macro background data drawn from international and national studies, secondary source literature, and donor community documentation.

f) The goals of educational researchers using qualitative methods are best served by using approaches which connect explicitly macro- and micro-structural levels of data collection and analysis from an interdisciplinary perspective which can assist policy makers and practitioners. (Trueba, et al. 1990)

In Ghana three domains of enquiry were established early on: the communal/home; the economic; and the educational/school. In order to describe those domains literature was reviewed pertaining to a number of relevant academic disciplines ranging from the classical anthropological (e.g. The Lions of Dagbon') to the economic (e.g. John Toye's 1991 study of the impact of structural adjustment on the Ghana Economy). Material was also drawn from newspapers, unpublished student theses, and the welter of documentation produced by NGO and bi-lateral aid organisations.

3.3 Methods: Using life histories to tell us more about real life in developing countries

3.3.1 Introduction
3.3.2 Nature of life history
3.3.3 Life history and the study of teachers' lives
3.3.4 Life history and issues of culture
3.3.5 Life history and educational research in developing countries: Problems and prospects

3.3.1 Introduction

The importance and value of life history and biography as a research method is now well established in educational research, particularly in the Western, so-called developed world. Life history and biography have been used to good effect for example in an understanding of individual-community relations at a time of change (Bertaux, 1981), in the organisation of individual life (Mandelbaum, 1973), and more specifically in the interplay between teachers, individual identities and the socio-historical context in which they work (Goodson 1980, Ball & Goodson, 1985; Woods 1981, 1984, 1987). Essentially life history research concerns the relationship between two inter-dependent worlds: that of the individual with their unique life story and that of the past, present and future contextual world through which the individual travels. Life story is, "the story we tell about our life" (Goodson, 1992) while life history is that life story, "located within its historical context" (ibid.). Given the potential utility of these methods to illuminate the realities of people's lives, particularly at times of great change, it is surprising to discover how little impact they have had on educational research in the developing world. There seems, however, to be the stirrings of an interest in exploring the use of these techniques in Third World settings (see Osler's (1997) recent paper for example).

An aim of this research project was to present a case for greater use of biography and life history in educational research and to describe and evaluate the use of these methods in a small scale research project in one national setting.

3.3.2 Nature of life history

What defines a life history has been a matter of some debate (see Hatch & Wisniewski, 1995) though it seems possible now to identify a broadly accepted set of characteristics:-

· It is a qualitative research method which like the closely allied method, narrative enquiry, focuses on, "the individual, the personal nature of the research process, a practical orientation and an emphasis on subjectivity" (Hatch & Wisniewski, 1995).

· It is a mix of 'life story' as told by individuals to the researcher and, what Goodson 1992, calls 'genealogies of context' which in turn become a 'life history'.

· It is essentially a personal type of research enquiry with priority for success being given to the establishment of rapport between researched and researcher. The dialogical, discoursive nature of life history and narrative work raises a number of questions, both ethical and ideological particularly when involving outside researchers investigating problems in the developing world.

· It is concerned with 'voice' and 'ownership'; emphasis given throughout from design to publication to what the individual researched has to say, how it is said, and the meaning made by the speaker to what has been said. As such it has great potential for imbuing the research process with a liberating, democratic ethos.

These characteristics, in turn, give rise to two parallel sets of tensions:

· first in the balance that needs to be struck between the individual and the contextually-situated nature of their individual experience.

Recognising that 'no man is an island' means that a major task in carrying out life history research is to present a view of larger, macro-issues through the lens of an individual's life experiences.

Bertaux (1981) suggests that:

"each individual does not totalise directly a whole society, he totalises it by way of the mediation of his immediate social context, the small groups of which he is a part" (p. 23)

If one, therefore, views individual life experiences as always in relation to the immediate social environment (which is particularly so in the developing world) and in relation to comparative experiences of those in similar situations it is possible to present an analysis which is both particular and universal. As we shall see when discussing the life histories of teachers and students in our case study, data can be presented in two ways: the individual teacher's life in relation to her immediate social environment and a composite analysis of groups of teachers and children's lives drawn from larger geographical areas. A particular strength of life history is, therefore, its potential to bestride the micro-macro interface than most other forms of qualitative research.

· Second in the balance the subjective and objective. In many ways life history and narrative methods reveal both the strengths and challenges of these forms of qualitative enquiry. These who use life history have no problems with extolling its strengths.

Ayers writing in Hatch & Wisniewski (1995) summarises the pro-views well:

"Life history and narrative approaches are person centred, unapologetically subjective. Far from a weakness, the voice of the person, the subject's own account represents a singular strength. Life history and narrative are ancient approaches to understanding human affairs - they are found in history, folklore, psychiatry, medicine, music, sociology, economics and of course, anthropology. Their relative newness to us is a reminder of how often we tail behind" (p 118)

As we shall see when analysing the data collected from the two case studies an attempt has been made to balance the contextual, background material with the 'voices' of the individuals interviewed; and to maintain equilibrium between 'objective' data - statistical information, survey reports, etc. - and the more personal 'subjective' accounts drawn from the respondents in the field.

3.3.3 Life history and the study of teachers' lives

Education is essentially concerned with what happens to people. Remembering this fact can guide us in making decisions about how to collect educational data and the purposes to which research should be put. There is a strong case to be made for research into policy and curriculum for example to take much greater cognizance of voices of teacher and pupils who daily experience the effects of decisions usually taken at a distance and by individuals at least once removed from the chalkface.

There seem to be at least four advantages in using life history in educational research:

Firstly by focusing on the lives of women teachers directly involved in the lives of their female charges, we are providing an antidote to the predominance of current research which tends to focus on the manager and administrator, notably those with power, access to decision making, and the prestigious 'voices' of the educational establishment (which often, though not always in Ghana, tend to be men).

Secondly, by understanding the individual teachers and pupils within a life history context we can identify significant moments and experiences which, in the eyes of the respondents, appear significant. Within that context and that level we can therefore suggest possible strategies for intervention.

Thirdly, we can give 'voice' to the marginalised, to the beginner female teacher at the chalkface, to the parent attempting to improve the lot of his daughter, and to the children themselves, bewildered by the problems of coping with factors often outside their control.

Finally, as Goodson, 1982, argues in his book 'Studying Teachers' Lives', a focus on the teacher and pupils in situ will generate much needed research into the relationship of 'school life' and 'whole life', a relationship that seems to be central to the problem of girls moving into and out of school.

These advantages take on greater weight when we consider the situation of many teachers and pupils in the developing world. As we shall see next when considering cultural issues and life history research, a noted feature of most, if not all, education systems in the Third (and increasingly) First World is their authoritarian nature. Despite calls for decentralisation and 'participatory planning' most decisions affecting teachers and pupils are taken by senior, usually male officials in ministries in the capitals of the world. Teachers, and more so pupils, are expected to be obedient and implement policy decided elsewhere. With low teacher salaries, few resources, and poor means of communication, a tradition is easily established in which the teacher is disempowered and disinclined to take any part in the educational decision making process. What teachers and pupils do though is develop support systems which mediate the demands of the above with the realities of classroom existence. Knowing about and making proper use of such knowledge can be brought about by life history research. Such data can also make some contribution to bridging the gulf between the macro, policy world of the capital city and the micro, classroom world where education is supposed to occur.

3.3.4 Life history and issues of culture

Reviewing the literature on life history and narrative research we are struck by the little that has been carried out in Third World settings. There are a number of reasons for this: the predominance of traditional empirical forms of research, the establishment of large teaching universities with little opportunities for research, and the recruitment of indigenous researchers by development agencies concerned only with macro, survey style evaluations seem to be three. Though some good, qualitative research is now being carried out by doctoral students around the world, little serious effort is being made to promote the incorporation of small scale, qualitative studies into the political and financial agendas of Ministries of education and donor organisations.

An exception is the recently published research by Robert Serpell into 'the significance of schooling' in one Zambian community (1993). Significantly, sub-titled, 'Life-journeys in an African Society' Serpell well illustrates the interface between life history research and cultural life.

In an important appendix to his book, Serpell sets out his "reflexive triangle" which bring together the respective "cultures of interpretation" of author, subject, and audience (see figure below):

Figure B2: Cultures of interpretation (perspectives): the reflexive triangle (Serpell, 1993)

As Serpell says:

"We can identify three different roles which feature in the communication situation: the subject whose behaviour is to be explained, the author who proposes the explanation and the audience to whom the explanation is addressed" (p. 281)

In his collection and analysis of the life histories or 'life journeys' of students from his Zambian community, Serpell sets out to explain the significance of schooling held by those represented in his reflexive triangle. He makes the important point that he is not only trying to explain how the author, say, views the educational values of his subjects i.e. school children but how the various parties to the explanation view their own values, perceptions, attitudes etc. This reflexive approach to the gathering of data means that 'culture' applies equally to the framing of the Zambian data as it does to experiences, knowledge and insights brought to the research from the outside.

As we shall see when analysing data collected through life histories in Ghana significant attention will be paid to the reflexive nature of the research process.

3.3.5 Life history and educational research in developing countries: Problems and prospects

Edwards, reflecting on his work at Oxfam and the Save the Children Fund, recalls that agencies like his rely upon, "detailed historical and anthropological monographs, often based on oral testimony and life-history techniques" (Edwards, 1989).

It seems to us that life history is a particular research method which is not only well suited to cultural conditions in developing countries but fits well into a methodological paradigm that is both qualitative and action-oriented.

Central to the question of improving the quality of education in developing countries is surely the work of the teacher - it is on the quality of her involvement at education's point of delivery that so much hangs. It is the teacher, and the teacher trainer, who remain central to achievement in our educational endeavour (Goodson (ed.) 1992) and it is therefore, somewhat surprising that so little attention is given to these practitioners by researchers.

To overcome any disinclination to make use of this, and other more 'people-centred' research methods, any problems inherent in the particular technique need to be addressed.

Two major problems need to be faced: first, that by focusing on the teacher or group of teachers at local level we may be gaining a great deal of culturally rich material, particularly of a local kind but will gain little of the national or broader picture. We must not ignore therefore the contextual parameters which so substantially impinge upon and constantly restrict the teacher's life (Goodson, 1992). Secondly, there is a danger of unintentional dis-empowerment, we must be wary in other words of all give and no take. Goodson and Walker (1990) suggest that by developing 'genealogies of context' we will produce research data that produces more of a 'complete picture' than an assortment of teacher biographies. Collaboration and an emphasis, as we have said earlier, on research environment, are therefore vital:

"Much of the work that is emerging on teachers' lives throws up structural insights which locate the teacher's life within the deeply structured and embedded environment of schooling. This provides a prime 'trading point' for the external researcher... Each see the world through a different prism of practice and thought. This valuable difference may provide the external researcher with a possibility to offer back goods in 'the trade'. The teacher/researcher offers data and insights, the external researcher, in pursuing glimpses of structure in different ways, may now also bring data and insights. The terms of trade, in short look favourable. In such conditions collaboration may at last begin" (Goodson and Walker, 1990)

In taking up this approach we are emphasising two dimensions of cultural importance: that the teachers' life histories be told in their own words and in their own terms; and that these stories or biographies be embedded in genealogies of context.

This leads us on to consideration of our second proposal, that life histories be generally accorded greater value in the research process.

Martin Cortazzi (1993) at the University of Leicester has recently drawn attention to the importance of storytelling by teachers as a research method. His work is innovative in two ways. First, it shows how narratives can be analysed from a variety of perspectives (and for our purpose his section dealing with 'anthropological models of narrative' is invaluable). Second, he shows that by focusing on teachers telling we get back to the knowing, which in turn provides us with a rich supply of experiential data necessary for the improvement of classroom life.

A distinctive characteristic of a life history is its relationship to culture. When we talk of culture here we are not just referring to an ethnic group but to the way teachers as a cultural group "tell their own story". In listening to teachers talk about what is important to them we can distinguish between what is recounted, how it is narrated and what teachers believe about their story to be important. Finally it will be important to relate these stories to cultural and professional environments, and the overall action-oriented nature of the research.

3.4 The research project

3.4.1 Introduction
3.4.2 Research purpose
3.4.3. Background and context

3.4.1 Introduction

The purpose of this section is five-fold, first the purpose of the research is stated and the research questions are clearly identified; second background is given on the problem of girls and Basic Education; thirdly the two case study contexts are briefly described; fourthly the way in which the research was carried out is outlined; and lastly principles and procedures relating to the analysis and presentation of the findings are discussed.

3.4.2 Research purpose

The purpose of the research is to gather data and develop solutions to the problem of why girls are not attending and dropping out of school in one African country.

The problem of non-enrolment and dropping out is investigated within the context of girls and women's life histories. These in turn are embedded in the cultural and institutional context of where the informants respectively live and teach/learn.

The research also explores the relationship that exists between what can be broadly be defined as culture and the research process. This is achieved by the overall methodological framing of the research and specific selection of research methods.

The research addresses the following questions.

i) Why do some children, particularly girls, fail to enrol and/or drop out of school during their basic education?

ii) What are the contextual and educational factors responsible for non-enrolment and drop-out?

iii) How far do the life histories of women and girls explain and provide solutions to the problem being studied?

iv) What intervention strategies and agents of change can we identify at school and community level to solve the problem?

v) To what extent is the problem essentially different within two contrasting areas of one country?

vi) To what extent can we develop culturally-sensitive research methodologies and methods in our work in educational development?

3.4.3. Background and context The International context The national context - Ghana Preliminary research conducted locally Case study contexts and research process Analysis and presentation of the data

The aim of this section is to briefly place the proposed research in its international and national contexts and to relate it also to some preliminary research that has been conducted locally. The International context

The general case that girls and women are relatively educationally deprived is easy to demonstrate. Female enrolments lag behind boys in most developing countries.

Figure C:

Male and Female Gross Enrolment Ratios by Level of Education 1990 and Adult Literacy (Percentages)


First Level

Second Level

Third Level

Adult Literacy









Developing Countries









Sub Saharan Africa









Arab States









Latin America










East Asia


1 14.9






66, 4


South Asia









Source: UNESCO 1991: 53 Table 3.2: 26, Table 2.2

Though disparities in enrolments have been reducing at the first level they remain high in Sub-Saharan Africa, the Arab States and South Asia. At the second and third levels progress towards more equal enrolment has been slower. Low enrolments are associated with lower GNP per capita and the association is stronger for countries with lower female enrolments (Behrman, 1991).

There is also a strong relationship between the incidence of low Gross Enrolment Ratios at primary level and relative under-enrolment of girls. Countries where female enrolment is low are also countries where overall gross enrolments are low (UNESCO, 1991). The bigger the disparity in enrolments between boys and girls the more likely it is that a smaller proportion of primary school children will be in school (Colclough with Lewin, 1993).

If access is a problem then drop out is equally worrying. There is evidence that drop out is higher for girls than for boys in the majority of African and Asian countries (UNESCO, 1991). Persistence rates to grade four for girls also pose serious problems for the educator concerned with issues of quality (Hawes and Stephens, 1990). The national context - Ghana

This picture is more or less reflected at a national level in Ghana, though, as expected, a closer look at enrolment and drop out rates as girls progress up the educational ladder reveal a number of complex and inter-related factors.

Figure C1: Enrolment by Sex and Grades in Basic and Secondary Education 1990/91 (approximate figures)






% Total


% Total

























































































































Figure C2: Girls in Primary School by region, 1991/92; Population, Enrolment and Enrolment Ratio

Gross enrolment rates for boys and girls attending primary school in the country were 77% in 1990 (80% in 1980). Disaggregating this by gender and by region reveals a more dramatic picture: 70% girls were enrolled in 1990 (71% in 1980) as opposed to 84% of boys (down from 89% in 1980) and regionally a large disparity in enrolment revealed 4% of girls dropping out in southern located Accra, the corresponding figure for the North being 20% (UNESCO, 1993).

Regional disparities and disparities within regions in respect of enrolment are also considerable. In the Greater Accra and Central Regions of the country, the gross enrolment rates is over 95% (including private schools) whereas it is only 43% in the Upper East Region. Gender disparities are greater in those regions too with low enrolment rates in the three northern regions. The girls' participation rate of under 40% there compares with the national average of 70% (Grieco et al, 1994).

Finally we can see a marked difference in what happens to a girl the further up the educational ladder she travels. The most recently available Government statistics reveal a significant reduction in the number of girls moving from the final year of primary six to Junior Secondary School. The following table is also notable for the disproportionate number of women teachers (particularly at secondary and teacher training levels) and the large number of untrained as opposed to trained teachers of both sexes, especially at kindergarten and primary levels.

Figure D:


No. of Schools


Trained Teachers

Untrained Teachers


Boys & Girls













Pre-Primary/ Kindergarten
























Jnr. Secondary








Snr. Secondary








Teacher Training








Technical/ Vocational








Special Education















Note: Two sets of figures are not additive. It may be noted that most of Middle Schools are having J.S.S. classes. Therefore, number of Middle Schools is included in Grand Total. Preliminary research conducted locally

In 1992 the Ghana Ministry of Education in collaboration with the local UNICEF office initiated a piece of "action research for equity improvement in primary schooling" (Research report-draft. Unpublished 1993). Utilising a largely quantitative approach (in spite of the inclusion of the term 'action research' in its title) the team of researchers collected a wide range of data from seven geographically representative areas of Ghana. The main purpose of the exercise was to answer a number of key questions e.g. 'why do some children drop out of school during their basic education?' 'Which of these reasons are the most powerful?' and 'Where are the strategic points of intervention, both locally and nationally?'

Teachers, children and community leaders were asked, through interview, questionnaires and the building up of 'family' and 'school' profiles, a range of questions concerning specifically non-enrolment and drop out1.

1 Data was collected from 7 districts, 148 teachers therein and from 22 schools. 49 community leaders and 79 children (including the non-enrolled and dropped out) were interviewed.

The research reports four main findings:

i that variation in reasons given for dropout and non-enrolment given by those interviewed relate to the characteristics of the local school e.g. use of harsh punishment.

ii that variation in the quality of schools both in terms of material provision and in terms of teacher commitment is important.

iii that "money-related causes" e.g. variation in occupation distributions of samples of families in each district is also important.

iv that family factors e.g. the impact of broken homes upon non-enrolment and drop-out is important in all types of school.

Overall, the authors of the draft report argue that, "money-related causes are cited as the most important reasons as compared to all types of school-related characteristics" (GES/UNICEF 1993). If economic and education reasons feature predominantly, the third factor of significance, "the impact of broken families upon non-enrolment and drop-out in all types of school", is viewed also as worthy of further consideration.

The report concludes with some possibilities for further research e.g. "a more focused study in a couple of districts which would be more ethnographic than the current one deploying qualitative methods to the full to try and understand the complex inter-relation between all the factors that have been proposed as causes of drop-out and the variety of educational provision being offered for children" (GES/UNICEF 1993).

It is with this background in mind that the current research was proposed. Case study contexts and research process

In practical terms the research was carried out in the following ways:

First two locations within Ghana were selected for the study: Winneba a coastal township (pop. 50, 000) 64 kilometres west of Accra, site of the developing University College of Education (where the writer worked whilst living in Ghana) and reasonably prosperous. A number of primary schools centred within the town were selected for examination in contrast to a number of outlying schools in the poorer surrounding rural areas. Tamale and Laribanga village in the Northern region (population 200, 000 and 5, 000 respectively), Islamic with the latter known for its problem with girl under-enrolment and drop-out.

Map of Ghana showing research sites: Tamale and Laribanga in the North; Winneba on the Coast

Laribanga Primary School

Esseukyr Primary School

Secondly the following research process was undertaken: initially a broadly -focused questionnaire was administered to teachers in both sites to establish some sense of the dimensions of the problem of girls under-enrolling and dropping out of school, then a team of indigenous researchers was briefed and trained in identifying sources of data and in particular the carrying out of life history interviews; data was then collected over a nine-month period (Easter 1995 to Christmas 1995) - comprising in all 89 in-depth interviews with teachers, parents, elders, girls in and out of school, and those responsible at district level with the administration of the educational system. Periodically focus group interviews were held with the research team. Individual research diaries were also kept by all the researchers.

To supplement the gradually growing pile of interview transcripts (823 pages of interview data finally produced) the research team collected enrolment and examination statistics from all the schools plus a vast array of government and non-governmental agency reports plus any secondary source material that would help provide background to what emerged from the field. The research team also observed lessons at the various primary schools and, where necessary, spent time in the homes and workplaces of the girl drop-outs. A large number of photographs were also taken.

An essential requirement of the researcher was that he or she was able to interview in the first language of the interviewee - interviews were in fact conducted in six Ghanaian languages - English, Twi, Hausa, Dagomba, Gonja and Kamara (the exclusive language of Laribanga) and then transcribed from tape and written out into English. All the research team - six in all - were qualified teachers, some having previous experience of research techniques.

By focusing on the lives of women and girls in the two contrasting areas (and within those the urban and rural) it was hoped that some sort of representative national picture would emerge illustrating similarities irrespective of location but also unique differences that in turn would highlight the necessity of not treating all national contexts as if in some way they were monolithic in character. Analysis and presentation of the data

Analysis of the data was both interesting and laborious, a feature well noted by those carrying out qualitative research of this kind. Essentially the data was analysed thematically: after an initial 'read through' of all the material three previously mentioned broad categories were identified and loosely described as 'family', 'poverty' and 'school'. This data was then re-read and roughly coded into more focused categories e.g. 'who pays my fees?', 'relations with parents', 'punishment at school' etc.

Gradually, utilising the idea of progressive focusing, three large 'banks' of categories and sub-categories were accumulated with, at each stage, analytic memos or aide memoires produced to 'capture' meaning and insights as they emerged from the data. In a number of instances reference was made back to a particular piece of data collected by one of the team members for clarification.

By analysing the data thematically i.e. through the worlds of home, school and the economy, it was hoped that a composite picture would emerge of lives lived past and present in those domains. The focus is therefore more on the collective experience of Ghanaian women and girls than on describing single lives in toto. It became apparent during the fieldwork process that an alternative approach would have been to have interviewed a much smaller group, say between six and twelve people, and to have presented their individual lives as exemplars of experiences common to many. Educational and cultural issues would then have been inferred from those life histories.

The decision, on the other hand, to analyse data by theme and to draw upon the life histories and interview data to illuminate those themes, was taken on the grounds that the resulting analysis would be of particular value to educationists and those working in Government and partner donor organisations interested in understanding the complexity of the issues and the experiences of Ghanaians struggling to improve the quality of their lives.

In an effort to retain a sense of the original interviews extended transcripts of three life history interviews are included in the appendix A of this report.

Finally attention throughout the research process has been on the cultural nature of the issues under examination and the culturally appropriate use of the various research methods and forms of enquiry. The question of girl drop-outs is a sensitive one, particularly to those who have dropped out, and as such the research term were required to act in a way that was both respectful to the communities involved and tactful towards the various individuals interviewed. With this in mind it was decided to use only the speakers' initials in reporting the findings.

The following sections present the findings. The three domains of home, economy and school provide the conceptual organisation of the presentation.

Each section begins with a broad description of the landscape, material drawn from a variety of sources: personal observation, national and international research studies, published and unpublished works. The purpose is to provide the backdrop within which the individuals interviewed live and work, to give a sense of the cultural fabric within which the two case studies are framed. Background and contextual material is provided therefore for the worlds of the home, the economy and the school.

Each section then presents the data drawn from the individual life histories. These "voices of experience" have been analysed and presented thematically; themes emerging from the life history interviews during translation, transcribing and through discussion within the research team.

A major aim of this study was to "allow" individual voices to be "heard" and for this reason, where it seems appropriate, individual testimony is accorded space. Occasionally sections of dialogue between researcher and researched are included to give some sense of the interchanges that occurred in the field.

Throughout an effort is made to link experience with theory and to provide insights for the parent, teacher and development worker.