بهره برداری از ازمکمل ها در فرایند تولید علمی در سطح آزمایشگاهی

The paper analyses the scientific research production of more than 80 laboratories belonging to Louis Pasteur University, a large and well-ranked European research university. We study research organisation of the labs focusing on the structure of research personnel and outcomes. The paper proposes a typology of laboratories, which enables us to stress different design for research organisation. The main results show how appropriate combinations of research personnel may strongly influence the publication and patent productivity.

This paper argues that the laboratory is the locus of many complementarities between researchers that should be taken into account to understand academic research organisation and production. In that respect, we depart from the literature in economics, which usually focuses on the individual level of analysis (Diamond, 1986, Levin and Stephan, 1991 and Stephan, 1998). In the meantime, the literature recognises that complementarities are important to understand scientific productivity and scholars also repeatedly argue on the necessity to take into account the collective level of organisation and especially the laboratory level (Dasgupta and David, 1994 and Stephan, 1996). Nevertheless, there are only few economic empirical contributions devoted to the laboratory level of academic organisation. Joly and Mangematin (1996) build a typology of public laboratories based on three categories of variables: scientific production, type of funding and the research themes. They analyse the type of relationships each category of laboratory establishes with private partners. Laredo and Mustar (2000) develop a model for characterising the ‘activity profiles’ of labs based on their relative involvement in five different activities: production of certified knowledge, embodied knowledge, participation to competitive advantages, to public debates and involvement in the construction of public goods. Bonaccorsi and Daraio (2003) stress that the distributions of the average age of scientists in the labs and the size of the labs are correlated. In three out of six domains, they found that the size of the labs is negatively correlated with productivity. The originality of our study resides mainly in the attention devoted to the structure and characteristics of both laboratory personnel (status, age, full-time research or teach-and-research position, non-researchers, sub-disciplines, etc.) and laboratory outcomes (publication counts and distribution, co-publication behaviour and patent counts). Our work is based on an original and unique database concerning the research activity of more than 80 scientific labs belonging to Louis Pasteur University (ULP) of Strasbourg and, covering more than a decade. We show that permanent and non-permanent researchers are complements: Professors tend to attract Ph.D. students and post-docs choose labs with highly recognised full-time researchers. Moreover, we observe that the labs that patent more are also those that publish with industry and with international co-authors. Even if some permanent researchers (university professors and un-promoted permanents) are less productive than others (full-time scientists), complementarities between them exist: the shares of the categories of personnel affect their productivity. For instance, an equal share of professors and full-time scientists stimulates productivity. The paper is organised as follows. In Section 2, we underline the added value of the laboratory level of analysis. Section 3 offers information on the data and some descriptive statistics. In Section 4, we develop a correlation study: between the variables characterising the labour force of the research laboratories, between patents and publications and finally between input and outcome indicators. Section 5 presents the typology, which identifies five ‘styles of research production process’ at the lab level. The last section discusses our main results.

This paper offers a first empirical investigation of an original set of data describing the research activity of a large European university. Such dataset allows us to analyse the organisation of research at laboratory level. We examine the labour force composition of labs and its influence on various outcomes. We also study the output structure of labs in order to stress possible complementarities or crowding out between outputs. Finally, the typology of labs highlights different types of research production processes or design. As expected, we find collective effects on research production at the laboratory level. There are complementarities between different types of research personnel: appropriate combination of different types of research personnel has significant impact on the research productivity of the lab. An equal share of full-time researchers and professors maintains incentives for the latter to perform research. Moreover, there are specific links between permanent and non-permanent researchers. Full-time researchers increase the research performance of the lab and attract post-docs. Even if university professors decrease all the scores in average outcomes, they increase the number of Ph.D. candidates thanks to previous contacts. Moreover, our results highlight the often ignored impact of non-permanent researchers. Especially, national post-docs increase significantly the average number of patents invented by permanent researchers. This result could indicate that such post-docs are dedicated to invention activities which may be explained either by the their weak autonomy in research agenda selection or by their early involvement in a career path turned toward research in industry. Another added value of our study comes from analysing the outcome structure of the labs. We find quite surprisingly that the share of international collaborations is not associated with a higher average publication performance: those who collaborate more intensively with international co-authors are not necessarily those who publish more. On the contrary, the average performance in international collaborations goes along with a high performance in terms of collaborations with industrial partners. One important result is that the intensity of patenting activity is correlated with all publication intensity measures: strongly with the intensity of publications with industrial partners and weakly with international co-authors. These results are summed up and combined in the typology which highlights the various designs for academic research organisation and production. We find that one class of labs exhibits performance scores which contrasts with other comparable ones: while their permanent researchers publish even more than the standard research-intensive labs, they also patent over eight times more. Such high performance is due to an interesting combination of personnel: The presence of younger and promoted permanent researchers, equally allocated between full-time researchers and university professors, focused on various sub-fields, allows to attract both many Ph.D. students and post-docs.