The findings from the literature review formed a reference base for the field study, which was undertaken as a primary data source for this thesis. The main objective was to investigate the effect of purdah practices on the layout and organization of rural housing in an attempt to achieve an understanding of how this particular facet of the rural living pattern influences and is, in turn, influenced by its built environment.
The study was conducted during October/ November, 1991 for about a month. It consisted of:
1) Fieldwork in the rural study area for more than three weeks.
2) One week in the capital Dhaka, collecting relevant literature and publications, and discussions with professionals engaged in the field of rural housing and development for women.
4.2 Choice of locale
The field research necessitated a case study in an Islamic country. Bangladesh, a country of more that a 100 million Muslims was chosen as the location for the field study. Several factors helpful to the study were taken in consideration for the choice of locale:
(1) Population-wise, Bangladesh is one of the largest Muslim countries in the world. At the same time it is one of the poorest and least urbanized. In this country, I would be able to study the building traditions of a rural Muslim society with its own distinctive cultural identity. I would also be able to have an insight into the dynamics of poverty and religious regulations in the layout of a poor Muslim home - an important premise of my study.
(2) As a native of the country, I would have the assets of knowledge of the language, familiarity with the cultural background, contacts with professionals in the field and access to relevant Bengali literature.
(3) Last, but not least - as a woman, I would have the advantage of studying this segregated society from the closest possible range.
The main factor influencing the choice of the research site was that it should be in an area where the field work could be carried out in a safe manner and the time available could be utilized in the best possible way. In the uncertain political and law-and-order situation that presently prevails in Bangladesh, it was imperative to choose an area where I already had a secure base and precious research time would not be squandered in trying to "set up shop."
Shampur, a rural area in the Chapai Nawabganj district, with which I was already familiar, was chosen as the site for the field study. (Fig. 16)Thumbnail('fig16-sm.jpg','fig16.jpg'); ?>
To participate in village day-to-day affairs, to enter private homes and conduct interviews in a conservative society is not an easy task for a researcher without proper introductions and credentials. This particular aspect was made easy for me as I am related to some of the villagers by marriage. I had visited the village as a new bride and was already known well in a section of the village. As a kinswoman, I expected that access to the rest of the village and establishing a rapport should not prove difficult.
4.3 The research site
The village in Bangladesh has been called "a somewhat elusive unit." Usually villages do not have recorded boundaries, but for administrative purposes rural Bangladesh is divided into Unions with defined limits.
Each union is subdivided into a number of mauzas which consist of one or more villages. Shampur Union (pop 30,000) is divided into 10 mauzas, which in turn comprise 31 villages. Two neighbouring mauzas named Bajitpur and Shadashivpur were selected as the research site. These two mauzas, in structure, spatialorganization and function, were observed to be two complementary parts of a settlement surrounded by agricultural land.Thumbnail('fig17-sm.jpg','fig17.jpg'); ?>
Officially Bajitpur has seven villages and Shadashivpur has three villages. In fact, some of the settlements within the mauzas are known as para and others are called gram (village). Contrary to the official line, however, most villagers consider the two mauzas to be two adjoining villages with the individual paras scattered within. To avoid confusion, the villagers' delineation of the settlement was followed for the purpose of this study.
4.4 Research methodology
The format used in this study is known as exploratory research which has as its object "the exploration and clarification of some phenomena where accurate information is lacking and is intended to provide description as thorough as possible, often with a view to providing material and guidance for subsequent research."
Due to obvious constraints of labour resources and time, it was decided not to attempt a statistical survey with a large representative sample of the study area. Instead a smaller sample was relied upon to generate a pattern which would serve as a basis for a descriptive analysis of the research problem.
The sampling unit was a homestead with one or more households. According to the Union Office there are 425 homesteads in Bajitpur and 380 in Shadashivpur. The total sample size was 60 homesteads (7.5% of the total).
4.4.1 Selection of Sampling Units
Stratified sampling methods were used, in which the population was broken upinto several subgroups or strata (in this case different socio-economic groups) and separate samples were taken within each sub-group. These samples were taken in the same proportion as the proportion of the subgroups in the total population following the "proportional stratified" sampling system.
At first the different socio-economic groups of village society were identified. Because of the obvious constraints, it was not possible to attempt a census of the villages in order to determine the socio-economic make-up of the population. In the absence of reliable information as regards income (which in rural Bangladesh is often in kind rather than in cash), possession of landed wealth was taken as an indicator of social grouping. Land has traditionally been valued as the ultimate security in the subsistence agricultural economy of Bangladesh. Any surplus income is invested in land and it is disposed of only as a last resort. As ownership and use of land is the measure of wealth and status, this factor was taken as a basis of identifying the different groups in the social hierarchy.
After discussions with several long-time residents of the area, the following socio-economic levels were established:
Landlords - These are the descendants of the four original zamindar (feudal landlords) brothers, who came and settled in Bajitpur. Although the system of feudal land-ownership was abolished in 1950, these landlords still own the majority of agricultural land and mango groves in Shampur Union. They do not work on the land themselves but lease it out to tenants or sharecroppers. Supervision of tenants and management of their land and orchards is carried out by traditional caretakers known as the dafadars, who may be members of any of the following groups.
Surplus Farmers - They work on the fields themselves, but also hire labour or lease part of their land. Some of these rich farmers have stopped working on their fields as a sign of their improving wealth and status, but this group was not included in the "landlord" category as their status is not regarded as equal to the original landlord family.
Subsistence Farmers - These farmers live mainly by working on their own lands.Their produce is barely enough for their own consumption.
Marginal Peasants - This group owns some land, but not enough to support themselves. They earn their living mainly by working as sharecroppers and day labourers.
Landless labourers - They own no land except for their homesteads. They have to sell their labour on a daily basis in order to survive.
In the absence of census figures, the author relied on the estimation of various village heads regarding the size of the various groups. According to them, the percentage of landless is about 55-60%, landlords and surplus farmers about 8-10% and subsistence and marginal farmers constitute the rest of the population. Five village chiefs who were interviewed independently quoted these numbers. Consequently it was decided that the sample would roughly reflect these numbers. Taking into account the number of households that could be surveyed in a day - about 3 to 4 - it was decided to survey 60 homesteads in total. Of these 35 were landless, 9 marginal farmers, 8 subsistence farmers, 5 surplus farmers and 3 landlords. However, random sampling could not be carried out because, in many instances, personal introductions served as a basis for choice of households. At other times, homesteads were chosen at random, but the socio-economic grouping of the household was assessed before commencing study of the homestead.
4.4.2 Focus of the Study
The mutual interdependance and social dichotomy between these segments of village society has led to an interaction between the groups which is simultaneously function-wise closeknit and status-wise divergent. To study the built environment of these separate groups, an understanding of their complex social structure is important. To narrow the focus, the study concentrated on the status and roles of the women of this village, their activities within and outside the home and their interaction with each other. A study of how women use and perceive their local spaces would give insight into the physical form and cultural structure of their built environment. It would indicate the extent of the influence of societal norms of segregation in local housing and would give valuable indications about how the built environment affects living patterns of women.
4.4.3 Data Collection Tools
The method of data collection involved an ethnographic approach. Ethnographic research is a mixed data-collection process, consisting of observational techniques, participation, interviews (usually informal) and secondary information bases such as recorded history, oral traditions, physical artifacts and so on. Although this approach involves "an intensive study of some given society," it was modified to accommodate the architectural research that forms the main focus of this study. Instead of an exhaustive study of the society in question, data gathering was limited to those aspects of social life which were relevant to the physical environment of women. Also, even though there is no question that "participation" is an important tool for research of this kind, for the architectural researcher equipped with prior training regarding the physical environment, a less intensive and less prolonged participation may be sufficient to recognize what Sinha terms "environmental behaviour."
Instead of a structured questionnaire, an open-ended interview guide (see Appendix 1) was prepared and used in informal interviews. This was more appropriate given the limited literacy of the respondents and also because of the close interaction that could take place between researcher and respondent. In almost all cases, female family members were interviewed in groups or individually, except in a few instances when male household heads deemed their wives too "ignorant" to be interviewed. Some male members were, however, interviewed in addition to the females, to get an understanding of male perspectives of female space use and behaviour. The interviews were unstructured and informal to allow the female respondents to discuss their living patterns and felt needs freely and at ease. Conversations were taped rather than noted down as this seemed less inhibiting and also saved time.
Key informants were consulted about the history of the village and about village life in general. Photographs and measured plans and sketches of house plan and furniture layout formed an important part of the study. At least two to three hours were spent in each homestead to allow enough time for friendlydiscussions and for observing the family activities at close range. Space use and avoidance by women were also observed and noted, not only within the homestead, but also in the neighbourhood. Women's perception of spaces and their cognition of their larger environment was taken into account. Access of women to educational, commercial, religious, legal and medical institutions in the villages and their surroundings were researched and noted.
4.4.4 Constraints of the study
The field study was hampered to a great extent by several constraints. Time constraints proved to be a crucial factor. It became evident that a much longer period of field work is needed for ethnographic-environmental research of this kind to achieve a more complete and comprehensive understanding of the research topic.
Being a member of the landlords' family imposed additional constraints on my role as a researcher. In many ways, I was expected to conform to standards of behaviour in keeping with my status in the social hierarchy. For example, it was my intention to spend a full day and night with a landless family. This suggestion was met with horror by my hosts. I had to be accompanied at all times by a female and a young male companion (to carry my materials), as walking about unescorted would be unacceptable. As I stayed in the landlord's enclave, I had more opportunity to observe their living patterns over a twentyfour-hour cycle, rather than those of the poorest villagers, which was the focus of the study. My "known" status sometimes proved to be a disadvantage in directly acquiring information about sensitive issues, such as family assets, religious practices, compliance to purdah norms, etc. Villagers were understandibly reluctant to impart such personal information to an "insider." Also, many of my questions seemed self-evident to my informants. As a member of the same culture, I was expected to know about certain aspects of their daily life. At times, some of their observations seemed designed to project a favourable image, rather than reality. My role as a participant observer (and access to secondary sources) overcame such disadvantages to some extent.
As a woman and moreover as a daughter-in-law, I had to abide by the generalpurdah conventions and behavioural norms of the village society, a fact that I could not afford to ignore. A breach of these norms would have offended the villagers and brought disrepute to my affinal family and affected their standing in the community. It was impossible to do any work after dark, as it was totally against social norms to be out at that time. Gathering information in public spaces, such as mosques, markets or eating places was out of the question. Although my status as a researcher and as an educated professional afforded me flexibility, mobility and access to male respondents, however, strict rules for dress code and behaviour had to be followed.
Nevertheless, the constraints were more than offset by the advantage of easy access and introduction to the villagers. As a result the field work could commence on day one and a friendly rapport could be established with the majority of my respondents.
4.4.5 Methods for Analysis
A descriptive analysis was attempted rather than a statistical one. To achieve a systematic organization of the data, it was necessary to establish particular categories of research. During the course of the literature review, several aspects of space use of women emerged which were used as a basis for the interview guide. At the end of the field work some other factors were added and these along with the earlier ones were used to formulate the analytic categories to present and explain the study results. These results are contrasted with experiences of other researchers and the empirical generalizations thus produced helped to develop the concluding interpretations of the findings.