will vary with the desired clusters' characteristics. The median method--the algorithm we used--determines:
d(ci,cj,ck) = 0,5d(ci,ck)+0,5d(cj,ck)-0,25d(ci,cj),
where d(ci, cj,ck) is the distance between cluster ck and the cluster obtained by merging clusters ci cj.
There are two main characteristics of this method. First, the two combined clusters are equally weighted in the centroid computation, regardless of the number of cases in each; i.e., this method is robust; second, it should not suffer from 'chain effects', i.e. progressive agglomerations of objects that are not homogeneous with regard to one principal component.
An important problem then was deciding how many clusters to choose, since hierarchical methods (as the median one) provide many alternatives (from n to 0 groups); this was done by cutting the dendrogram (the 'tree' that describes the grouping dynamics) above the low aggregations (which bring together the elements that are very close to each other), and under the high aggregations, which lump together all the groups in the population.
Atkinson A. B. and Stiglitz J. ( 1985), "'A new view of technological change'", Economic Journal, 00: 000-000.
Baah-Nuakoh A., and Teal F. ( 1993), 'Economic reform and the manufacturing sector in Ghana', mimeo, Africa Technical Department, Regional Program on Enterprise Development, World Bank.
Barba G. Navaretti ( 1994a), 'The acquisition of technological capabilities: Ghana, Kenya and Zimbabwe compared', mimeo, Africa Technical Department, Regional Programme on Enterprise Development, World Bank.
----- ( 1994b), "'What determines intra-industry gaps in technology? A simple theoretical framework for the analysis of technological capabilities in developing countries'" Centro Studi Luca d'Agliano-Queen Elizabeth House Development Studies Working Papers, no. 71, Turin and Oxford, November.
----- ( 1995), "'Trade policy and foreign investment in textile and clothing: an analytical framework'", in G. Barba Navaretti, R. Faini, and A. Silberston (eds.), Beyond the Multifibre Arrangement: Third World Competition and the Restructuring of the European Textile Industry, Paris, OECD Development Centre.
Dasgupta P. and Stiglitz J. ( 1988), "'Learning by doing, market structure and industrial and trade policies'", Oxford Economic Papers, 40: 246-68.
David P. ( 1993), "'Knowledge, property and the system dynamics of technological change'", in Proceedings of the World Bank Annual Conference on Development Economics 1992, Washington: The World Bank.
Dosi G. ( 1988), "'Sources, procedures and microeconomic effects of innovation'", Journal of Economic Literature, 26: 1120-71.
-148-