The migration of highly skilled professionals is an important issue for many sectors, especially for IT. It has already been noted that highly skilled labor markets are crucial for a wide variety of economic sectors. The now famous examples of the San Francisco Bay Area and the Boston cluster are mobilized regularly by innovation consultants and by regional planners seeking to replicate the miracle of university/industry collaborations.
In this research we leverage a database that provides us with information on the location and academic production of Russian computer scientists. This dataset allows us to ask anew a series of questions: What is the pattern of mobility among Russian computer scientists who work in academia? Are those patterns related to scientific productivity? Russian computer scientists allow us to focus the larger questions raised by migration studies scholars onto this population in order to understand Russian highly skilled workers mobility. We explore Russian computer scientists’ mobility using bibliographic data as a source of information. We develop a combination of methods which allows us to go beyond a simple description of migration flow or individual trajectory analysis. The source of data is publications indexed in Web of Science, the scientific publications aggregator by Thomson Reuters.
Data and Method
We extracted our data set from the Web of Science using a query combining subject categories and publication years. We formally selected all categories related to computer science. We transformed the bibliographic data into a longitudinal form, a sequence of states, and then we explored it using Optimal Matching (OM) method for sequence analysis. After having clustered individual careers, we visualized and described the obtained clusters in order to understand resulting type patterns of individual career of researchers. Furthermore, we calculated and visualized the probability of transition between different states in the total population for each moment of time. That permitted us to seize how the flow of individuals from one type of publication activity to another changes in time.
Results and Discussion
Independently of affiliation location, the number of publications roughly increased from 1990 till 2009 with a temporary drop around 2006, and then we observe a permanent drop from 2009 to 2012. The fastest growth concerns publications with foreign affiliation, we supposing to observe a discipline expansion at international level from 1999 to 2007. The former USSR affiliated publications number seems to grow less dramatically. The fall in late years is due to objective fall in publication activity and by our database artifact: it was established in late 2012 and excludes some new publications. The number of states with a foreign affiliation solely is growing slower during early nineties and is even deposed by number of states with former USSR affiliation in 1994. It could be explained either by the fact that some of the journals published in the former USSR were included in the WOS database and/or by a more active publishing of research papers in English. In any case, these two explanations could have in common a supposition that this grow is based on research done before the collapse of the Soviet Union. Then we observe a decrease in 1995-1996, the moment of the lowest level of R&D employment due to fall in public support to science. Afterwards the growth of homeland affiliated publications is never as fast as of foreign affiliated ones – the effect of crisis of Russian science seems to be quite strong.
Computer scientists who originate from the former USSR and publish in Web of Science referred journals have, in most of the cases, foreign affiliation. Those few of them who have affiliation only in the former USSR and have homologous co-authors have poor citation level in most cases. So, international dimensions of scientific activity are an important factor of scientific success. But there is no best way to internationalize the publication activity. Even if the migration with stable publication pattern when researcher is working and publishing abroad seems to be related to a high citation level the difference with, firstly, foreign-affiliated scientists publishing with ex-compatriots, and secondly, with scientists who change publication patterns during their career is not so important. Moreover, for the overall population, growing publication activity is correlated with a growing rate of change of publication configuration and, after 2005, the most cited articles are made by scientists with turbulent careers.
The internationalization of former USSR computer scientists can be seen as a process when researchers located in an area characterized by relatively poor research funding and relatively small number of active colleagues try to accede, by collaborating and migrating, countries where the domain is strong and well founded. But this view is simplistic: scientists and institutions located in the former USSR seem to bring fruitful collaboration for researchers with foreign affiliation. We can observe complex patterns of relations between diaspora scientists and their homeland research universe and complex patterns of internationalization of foreign USSR based researchers. Those complex patterns are related to higher scientific success. Described dynamics concerns only some part of researchers; nevertheless, they show that the unidirectional “brain drain” vision of former USSR scientist’s migration should be nuanced. On the methodological level, our claim is that sequence analysis with optimal matching could be a good method to explore scientific career through bibliographic data.