Book Rcvitws 89 
overwhelmed my capacity to absorb it: it is a source to return to in the future when further 
information is required. However, I was recurrently irritated by textual references to places and 
people that appear on none of the seven maps provided in this book. Moreover, it took me a long 
time to formulate a (probably deficient) idea of the background against which these papers would 
be assessed. Surely some concession should be made to the reader whose knowledge of the 
country's issues is, by definition, less than that of the writers of these papers. 'Land reform* is, 
after all, of comparative interest even to those who know nothing of Botswana's specific 
circumstances. 
Re\ iewed bv A 1* Cheater. Department u! NIK ii>ln«4\. I'nherMiv i>i'Ziml>;ilme. 
Social Attitudes and Agricultural Productivity in Central Africa, Jean Bonvin, Development 
Centre of the Organisation of Economic Co-operation and Development, Paris, 1986 (pp 135, no 
price given). 
Actually this book presents the findings of research conducted in 1977 in one Central African 
country, Burundi, and its simultaneous publication in English as well as French is a very 
welcome addition to Anglophone knowledge of Francophone Africa. Perhaps its greatest use may 
lie not in any value its specific findings may have for policy purposes but rather in the areas of 
research methods, where it provides another textbook case of the problems of survey research on 
this continent. 
Bonvin provides, at the very beginning, a brief description of Burundi which, apart from 
saying nothing about the country's social characteristics, is useful, as is his detailed map of the 
country's communes. He then proceeds to attempt to steer a middle course between the 
theoretical rocks of marxism amd modernisation theory, in what he himself describes as more 
'applied economics' than the development of new theory. This application assumes that different 
districts of Burundi may be ranked on an evolutionary scale in their incorporation into the 
contemporary international political economy. Bonvin is therefore interested to investigate the 
differences between Burundi's more and less effectively incorporated areas, in respect of their 
economic modernisation. 
The idea is, in principle, interesting. Its execution was less satisfactory. While asserting that 
participatory observation would have been 'ideal' (p 84), Bonvin instead opted for a survey 
questionnaire, properly piloted, finally containing 395 questions, administered to 1 655 farmers 
in the two interior communes of Ruyigi and Ngozi, and four communes in the Rusizi plain area 
north of the capital, Bujumbura (where peasantisation had been established under the Belgian 
colonial regime and was in the process of collapse post-Independence). His reasons for selecting 
the survey process rested on the excessive time needed for anthropological investigation, and the 
problem of generalising its results. Bonvin is now wiser, having realised that the survey and its 
processing took him ten years anyway, and that his results were not only unreliable in many 
respects, but also had a limited capacity for generalisation! 
In the meantime, he experienced research that will undoubtedly have provided him with 
enormously funny stories with which to regale colleagues and students for years to come. He has 
realised that the statistical universe from which samples are drawn is, in third world countries at 
least, seriously defective at both central and local administrative levels, thus often necessitating 
the creation ot the universe before the sample may be drawn, that 'in spite of its size, the 
questionnaire used in the quantitative enquiry failed to uncover many significant elements of the 
process of change in the rural structures' (p 102); that university students may be unreliable 
interviewers and coders, no matter how well trained, especially when they are physically 
uncomfortable in the countryside; and that officials may mislead about computing capabilities in 
developing countries, thus leading to enormous difficulties in processing data. There are still a
90 Book Reviews 
few verities that Bonvin ha» yet to learn, however, not least that, in the role of research assistant, 
government officials would be not better than students (for structural rather than personal 
reasons); and that the mythology of religious behaviour often reflects outsider's prejudices rather 
than reality. 
Given the title of this book, it is extremely disconcerting to have no usable information at all on 
ethnic variables, household structure, modes of land tenure, or the history of the Rusizi plain 
peasant holdings. (Bonvin's concept of 'culture' is also somewhat strange: see below). One 
presumes that the normal tests of significance were applied to the variables dealt with, but the 
reader must remain content (after all of the defects of the survey) with imprecise results, whereby 
some variables 'seem to' suggest, or be correlated with, others. If one is going to use a particular 
method of investigation, one should at least stick by the formulae of its symbolic rectitude, even 
•while admitting that there are problems! 
Of more importance, perhaps, is the fact that Bonvin does not appear to have used much, if 
any, of the existing sociological works on the Rwanda-Burundi area, in formulating the areas he 
wished to investigate and the precise questions he wished to ask. It is all very well to pay pauing 
deference to the value of ethnographic work, while ignoring completely its content; and quite 
another to build generalisable research onto such earlier work. Bonvin appears to claim to have 
done the latter, but actually did the former, which may explain why he seems to regard 'social 
integration' as consisting of access to governmental authorities and the mass media, and 'culture' 
as compromising religious observance and food taboos. 
Finally, in this slim volume there is nonetheless a certain amount of unnecessary repetition of 
methodological issues. This space would better have been occupied by data tables with their tests 
of significance; or, better still, by a copy of the final questionnaire (available on request, from the 
Development Centre of the OECD, but not included in this book). 
Reviewed by A P Cheater, Department of Sociology, University of Zimbabwe. 
The Construction Industry in Developing Countries. Alternative Strategies for 
Development, Jill Wells, Croom Helm, London, 1986 (184 pp, £25.00) 
This book is one of the latest publications on the construction industry in developing countries. 
The subject has been looked at for a long time, but only by specialised Institutions such as 
UNIDO and ILO, and is not very widely known. However, for some years research and books 
(1) have investigated this subject with a technico-economic approach showing the key role of this 
sector and analysing States' Policies and their results. This book does not only that, but also gives 
us some proposals concerning alternative strategies for African developing countries. 
The author show the specific role of the construction sector in the development process and in 
economic growth (construction output represents often more than 50% of the gross capital 
formation of a country) and the clear relationship between construction output and economic 
growth indicated through statistical analysis. She describes the under-developed resources of the 
sector in some developing countries, namely, Kenya, Ghana and Nigeria. 
For the author, those situations are not only due to the poor level of development of a country 
but are principally a result of the structure of the construction industry and its organisation, 
characterised by the divorce of design (architects, engineers and quantity surveyors) and 
production (building and civil engineering contractors), and the system of competitive tendering. 
It has been inherited in most cases from the colonial period. The effects of such organisations on 
the cost and quality of construction are described in detail, and also how they could become a 
barrier to development. She gives several examples, one of which is the case of professional fees 
for the designers which are usually calculated as a fixed percentage of the final cost of the