Catálogo de publicaciones - libros

The quantitative appraisal, partly through bibliometrics, of science-technology connections has made great progress in the last decade. We investigate in this chapter the lexical linkage between articles and patents, an alternative method to the systematic exploitation of the citations of patents to scientific papers. We explore in particular the ability to establish correspondence tables between patent classification and scientific categories. After a reminder of the methodological background (S&T linkages, lexical methods, statistical measures) we report an exploratory study based on a subset of the Chemical Abstracts database (CA) that covers both articles and patents by a very precise indexing system. Connection measures have been established, first on controlled vocabulary, and secondly on some natural language fields. The comparison shows some robustness of the lexical approach, with clear limitations at the micro level: topic sharing between a particular article and a particular patent cannot be interpreted in the general case as the sharing of a research question. At the macro level, for example IPC sub-classes and ISI subject categories, the lexical approach is an appealing technique, complementary to usual citation based analysis built on very sparse matrices, because informetric performances of lexical methods can be tuned in a large scope of precision-recall features. The extension to databases specific either to articles or patents requires language processing which can be alleviated if macro level correspondence is solely sought.

Abstract: Despite the generally acknowledged importance of science in many high-tech areas of major economic relevance, there are no science-related statistics to be found in high-profile international benchmarking reports such as the European Innovation Scoreboard. Why? This chapter aims to provide an answer by advancing our understanding of the possibilities of indicators quantifying linkages between science and technology. Central are the concepts of ‘innovation capability’ and ‘science/technology interface’, which are used to assemble a wide range of empirical studies and quantitative indictors to summarise their possibilities and limitations for producing comparative statistics. The review focuses on indicators dealing with flows of written (‘codified’) information, and indicators of inventiveness that capture the non-codifiable ‘tacit knowledge’ dimension. General conclusions will be drawn with a view towards further developments in the foreseeable future, suggesting new avenues for the design and implementation of patent-based and inventor-based statistics to describe and assess the complex and dynamic web of relationships between scientific research and technical development within the context of regional or national systems of innovation.

Up to now the contribution of scientific institutions to technology is considered to be primarily indirect. However, an analysis of patent applications of European public research institutions shows that they contribute about one half of all patent applications in selected science-based technology fields in the life sciences and nanotechnology. This finding documents a high direct contribution of science institutions to the generation of technology. The share of public non-profit institutions proves to be important, in particular in early stages of the technology life cycle; scientific institutions obviously play the role of lead actors. A comparison between European and German data reveals a lower, but still quite high share of public research institutes in other areas of science-based technology, so that the general statement of a high direct technology contribution holds. With regard to scientific institutions, patent indicators do not replace, but rather complement publication indicators and reflect an additional dimension of performance.

This chapter analyses and extends existing studies of how to characterise, trace, and measure knowledge bases of firms, sectors, and countries. The chapter is structured in two main parts. First, we present the concepts of knowledge specialisation and knowledge integration as the relevant dimensions along which knowledge bases can be mapped. The concepts proposed build upon extensive qualitative research which has focused on a variety of of knowledge generation and use in a range of industrial sectors and organisations. The aim of this chapter is to demonstrate that these, largely qualitative, processes map into key characteristics of the knowledge bases they contributed to generating and shaping; and that these key characteristics can be measured relying on the innovative use of patents, citations and publications data. More specifically, the analysis of the evolution of knowledge specialisation over time provides information about the persistence of knowledge in firms and sectors. It hints at the cumulative, path dependent nature of learning processes. Integration is studied by analysing the evolution of specialisation across different typologies of research. It hints at the complex, non-linear inter-dependence that link the scientific and technological domains. The second part of the chapter will be devoted to the presentation of indicators of breadth and depth that capture the key characteristics of the concepts introduced in the first part.

Patent and scientific publication data offer useful information for the analysis of key features of national systems of innovation of less developed countries. However, the use of these statistics should be subjected to careful evaluation. This chapter uses data for 120 countries (patents, scientific papers, and GDP), investigating different patterns of interactions between science and technology according to different levels of development. Later this chapter focuses on the data for India and Brazil, investigating interactions between science and technology in three dimensions: inter-sectoral, inter-regional, and inter-temporal.