International mobility in the higher
education sector
A study of young academics employed at the University of Amsterdam
By Laurens Steehouder
Bachelor thesis Human Geography Course code: 734301380Y
Supervisor: dr. I.C. van der Welle Second reader: dr. M.A. Bontje Date: 20th of June 2014
Student code: 10192360
Email: laurens_steehouder_ik@hotmail.com Tell: +31(0)653559418
CONTENT
1. Introduction ... 4
2. Theoretical framework ... 6
2.1 Academic mobility in a globalizing world ... 6
2.2 Academics as Highly skilled workers ... 8
2.3 The shard context: Student mobility ... 9
2.4 Academic mobility ... 10
2.5 the geography of international mobility ... 11
2.6 Factors explaining the different levels of spatial mobility ... 12
3. Data and methods. ... 17
3.1 The research questions ... 17
3.2 Causal models ... 18
3.3 Cross-sectional research design ... 19
3.4 Research group and site ... 20
3.5 Research Method for collection of the data ... 22
3.6 Operationalization of the concepts ... 24
3.7 Sample & response ... 26
3.8 Responde and Representativeness of the sample ... 27
3.9 construction of the variables ... 29
3.10 Methods for analyzing the data ... 30
4. Describing the data ... 31
4.1 Mapping the short-term spatial mobility over 2013. ... 31
4.2 MAPPING the long-term mobility previous in career ... 33
4.3 Describing the momenarty long-term spatially mobile ... 36
4.4 Mapping the mobility ... 37
4.5 Why are the academics spatially mobile? ... 42
5. Explaining the different levels of cross-border spatial mobility ... 44
5.1 Personal factors ... 44
5.2 Previous international experience ... 49 2
5.3 international Network ... 50
5.4 financial factors ... 55
6. Conclusions and discussion ... 58
7. References ... 62
Attachment 1: Respondents per research institute ... 66
Attachment 2: construction of the variables ... 67
Attachment 3: The used questionnaire. ... 71
1. INTRODUCTION
In case you are reading this bachelor thesis, you are probably an academic yourself or connected to the university in some way. If this holds true, you might already have noticed that the university is changing, and increasingly becoming a globalized institution. Think for example of the increased amount of courses given in English, the increase of non-Dutch colleagues, and the expansion of international student exchange programs. Reputation and place in international university rankings are becoming important for a university that has to compete for top researchers and students on a global labour market (Stromquist, 2007: 82). In the 21st century, the universities are once again becoming the globalized institutions they were during the Middle Ages. The first universities were truly international with Latin as common language and international recruitment of professors (Avveduto, 2001; Altback & Teichler, 2001: 6). In the 21st century, English is taking over the role as universal scientific language which Latin once had. The Anglicisation in higher education eases communication between universities in different countries and opens the way for cooperation and competition on a global level.
Moreover, globalization has forced the universities to emphasise market forces and become liberalized institutions. Increased global competition has changed the nature of the university and now the decision making is more influenced by market rationale than purely scientific curiosity. Cooperation between universities in a national context is no longer enough to survive; they need an international dimension (Kim, 2009; Storme et al., 2013; Ackers, 2009; Altbach, 2013). Universities face the growth of a global labour market for highly skilled personnel where they have to compete for the best academics. Knowledge workers are increasingly willing to move across national borders for job opportunities, and universities have to make sure that they can attract the best researchers and lecturers (Jöns & Hoyler, 2013; Altback & Knight, 2007). Attracting highly qualified professors can have a positive influence on the university’s reputation and can help to attract new students and research funding. The changed communication and knowledge structures and increased mobility of academics lift the universities out of their local context and force them to play in a global framework (Altbach & Teichler, 2001: 5-6; Qiang, 2003: 249).
It is this increased mobility of scholars and researchers, as important aspect of internationalization in higher education, this bachelor thesis will focus on. Academics are
more than ever expected to travel internationally for their careers (Storme et al., 2013: 12, 13). However, the increased mobility of academics did not get a lot of academic attention. The exact scale of increased mobility and the factors explaining this international movement of academics has until now been a relatively less researched topic (Altbach & Teichler, 2001: 19).
By examining the international travel patterns of the young academics employed at the University of Amsterdam, this thesis examines to what extent the academic staff is cross-border spatially mobile. Also, it tries to get a deeper understanding of the underlying factors that can explain the differences within scholars’ level of international mobility. The central questions that are answered in this thesis are: How often do young academics at the University of Amsterdam travel abroad for reason related to their work at the university? Which locations do they visit? How often do they have contact with foreign scholars? Is a high level of international mobility a requirement if one desires a prestigious academic career? And what role do gender, academic networks, funding and having a partner play in the decision to go international or not?
This specific single case contributes to the existing literature on internationalization in higher education. The university is globalizing and changing mobility patterns of academics are an integral part of this process. Understanding the cross-border spatial mobility of academics and recognising the changes that are taking place within the university are important because future policy can adjust to the findings. Furthermore, the study on the movement patterns of young academics at the UvA contributes to the literature on mobility of highly-skilled in general. Academics can be seen as a specific group of highly skilled knowledge workers. Insights in the mobility of this group can help to understand the phenomenon and function as comparative data for studies on other kinds of highly skilled labour mobility.
The rest of the thesis is structured as follows. In the first part, the mobility of academics will be placed in a framework of globalization and highly skilled labour mobility. Additionally, in this theoretical chapter the most important concepts related to the level of mobility will be explained. Further, in chapter three the used methods in the thesis are explained and in chapter four the dataset is described. In the final chapter the different factors that explain the difference in spatial mobility are presented by using statistical analysis. The thesis will be concluded with an overview of the results and suggestions for further research.
2. THEORETICAL FRAMEWORK
Internationalization within higher education is not an isolated topic, but is part of a larger force that has been changing the world drastically over the last two centuries. To fully understand the growing spatial mobility of academics, the broader context in which the changes are taking place must first be examined. Many scholars have identified the economic, social and cultural changes connected to the development of communication and transportation technologies as a process of globalization (Castells, 2001; Omae, 1995; Eriksen, 2014; Altbach & Teichler, 2001); Castells, 2001 ). This ‘opening up of the world’ has led to increased pattern of movement, especially for the group highly skilled knowledge workers (Mahroun, 2001; Beaverstock, 2005). This theoretical chapter starts with describing the concepts globalization and mobility of the highly skilled in relation to the topic of academic’s spatial mobility.
2.1 ACADEMIC MOBILITY IN A GLOBALIZING WORLD
The concept of globalization has been a hotly debated topic in the social science literature. The term ‘globalization’ is often referred to in different ways and contradicting claims have been made about its consequences (Scholte, 2007). In this thesis the term globalization refers to a nonlinear process that makes different places on the world more interconnected. The development of transportation and communication technologies, among other things, has changed the role of geography. It is no longer the absolute distance that determines proximity, but the relative distance, the connection with other places, decides if a place is relatively remote or nearby (Eriksen, 2014, 15-16). Communication with internet and telephone make it possible to contact colleagues or friends located at the other side of the world in less than a second. This increased interconnectedness is illustrated by an increase in global flows of goods, services, capital and ideas (Altbach and Teichler, 2001). Moreover, the amount of international trade has never been as high as it currently is (Chase-Dunn et al., 2000). The increased interconnectedness is also evident in the higher education sector. Flows of academic who visit other universities and contact international colleagues through the internet illustrate this.
Changes in the role of distance come together with a change in the capitalist mode of production. Production is no longer concentrated in one intergraded industrial area like Ford’s car factory, but is produced in more flexible global networks of a multiplicity of firms. In the 6
big industrial factories all parts of the production process were concentrated in one single location (Scott, 2007: 1467-1469). The vertical disintegration of production created global production chains with mutually dependent actor in different locations. For this reason the production chain management as occupational sector emerged and requires highly educated personal (Gereffi et al., 2005)
There has been a change from an industrial society to an ‘information society’ or ‘network society’. In the information society, knowledge has become the most important factor of production in the economy. Not only the new economic sectors and production chain management, but the entire economy is depending on science and technology for innovation. The economic success of a country is far less dependent on the availability (or lack) of natural resources, but the present human capital is source of wealth accumulation (David & Foray, 2002: 9-11). In the network society, the networks accumulate knowledge and wealth and have become fundamental for social organization (Castells, 2011). In his book Castells (2000) describes the spatial dynamics in the network society as ‘spaces of flow’ instead of ‘spaces of places’. A place is defined by the connections it has with other places. In these locations the highly skilled elites occupy their management positions and are continuously connected through their networks to the rest of the world. In the knowledge society the university has become a highly important institution. As primary producer of the highly skilled knowledge workers, the university stands at the basis of the economy.
Due to the changed nature of space, the power of the nation-state is disappearing (Omae, 1995), or at least changing (Mann, 1997). The institutions of the nation-state lost their capacity to control the flows and the accumulation of wealth within the national territory (Castells, 2011) The TNCs are not bound within the borders of the state and can decide to outsource parts of their production process. The academics are also no longer contained within the state’s territory and can relocate themselves in other countries.
Despite development of the modern communication and transportation technologies and the importance of networks, empirical evidence suggests that ‘spatial proximity ’is still relevant. This is shown by looking at the geographical distribution of the leading offices in the banking sector. The vast majority of the headquarters is located in just a few ‘world cities’ (Baeverstock et al., 1999: 435,436). Moreover, in his famous article Porter (2000) argues that clusters have become the main units of analysis in the modern knowledge-based economy. 7
Despite the access to modern communication technologies, proximity is still important for a firm: being located in a cluster is important to get a competitive advantage. The crux here is that a specific kind of information, the so called ‘tacit knowledge’, can only be exchanged with face to face contacts. This tacit knowledge (contradictory to codified knowledge) has an informal character and is exchanged in unofficial places such as lunchrooms and bars. However, this form of informal knowledge is crucial to innovation and therefore proximity to firms in the same industry is crucial for a company (Gertler, 2002: 95-96).
The continued importance of a specific place and the fact that modern communication has not made proximity irrelevant is shown by the increased patterns of movement. Despite the spread of access to telephone and internet people are more physically mobile than ever. In 1800 US citizens travelled on average 50 metres a day, while they now travel 50 km on a daily basis. This significant level of spatial mobility is also shown in the tourism and travel industry, which constitutes for 10.3 percent of the global GDP (Urry, 2007: 3-5).
The increased mobility of academics has to be studied with these bigger societal changes in mind and can be viewed as a product of globalization. Universities are important institutions in a world where knowledge is becoming the major production factor. The tacit knowledge is bound to a geographical location and, therefore, spatial mobility is required. This required proximity for the exchange knowledge stands on the basis of the academics spatial mobility. Academics, students and universities have to look globally and can no longer limit their orientation to a national scale. The increased importance of mobility has been identified for all highly skilled labour and an extensive amount of literature has developed around this subject. The next subchapter will further elaborate the topic of highly skilled labour mobility.
2.2 ACADEMICS AS HIGHLY SKILLED WORKERS
The patterns of movement have increased specifically for the group of highly skilled knowledge workers. One of the most important explaining factors for this is the development of a global labour market for the highly skilled in the knowledge society (Altbach & Teichler, 2001: 8). This group is highly desired by transnational companies (TNC’s) and can search for jobs on a globalized labour market (Mahroun, 2000).
In his book Castells describes the development of a transnational managerial elite class (Castells, 2000). This group participates in regular non-permanent migration and thereby 8
embodies the international flows of knowledge. Highly skilled workers do not only bring their practical knowledge but more importantly bring their networks and connections to the workplace (Beaverstock, 2005: 246). TNCs are depending on this diffusion of tacit knowledge on a local scale for managing their global production networks (Mahroum, 2002). In the globalized economy, willingness to travel internationally for your job has proven to be an important asset for workers (Storme et al., 2013). Accepting a function at a foreign office of the company is seen to improve career prospects. The result is that the managerial elites often have international career paths, with short time migration to several countries (Beaverstock, 2005). The different motives for workers to choose for an international career can be divided in five categories: economic factors, political factors, family factors, cultural factors and career factors. The relative importance of these motives differs for specific personal preferences and migrant groups (Carr et al., 2005 388-391).
The existing literature on mobility of highly skilled workers is useful for the analysis of academic’s mobility. The academic are an part of the group of highly-skilled workers and share characteristics. However, it should be noted that academics are a specific group of knowledge workers and, therefore, should be handed uniquely. The motives for an academic to move somewhere might be his/her academic aspirations, while for a business manager it might purely reflect his/her employer’s preference (Mahroon, 2000). In addition, the scientific labour market is considered even more internationally integrated with a relatively high rate of mobility for academics in comparison with other professional areas. This makes the case of academics even more significant to study (Chompalov, 2000: 32; Ackers, 2009 : 102). Internationalization is considered an ongoing increasing process. Because the academic sector is in further stage of internationalization the results can be valuable for the future development of other sectors.
2.3 THE SHARD CONTEXT: STUDENT MOBILITY
The literature available on mobility in the higher education sector has mainly focussed on the mobility of students. The in-depth analyses of the ERASMUS program is one on the few examples of a good multifaceted study on mobility in higher education. The Erasmus program is a by the European Union funded exchange program and supports more than 100.000 students annually in Europe (Altbach & Teichler, 2005: 7,19). Student mobility is no new phenomenon, but its size has increased drastically in the last decennia. Between 2000 and 9
2005 the level of student mobility has risen more than ever with a 50 % increase in the amount of international students (Kim, 2009: 396).
A study among UK exchange students shows that most students expect that an international exchange program will increases their position on the international labour market. The gains from an international experience are not only the acquired human capital. Moreover, the biggest advantages result from the acquired cultural and symbolic capital. Thus experiencing another culture and building a network have more relevance than the acquired academic skills and knowledge (Balaz & Williams, 2004). In his article on student mobility and European identity Mol (2013:210) elaborates this by stating that European student mobility is mainly driven by experience, travel, educational, and leisure motivations .
The literature on student mobility is useful for the analysis of academics’ mobility because of the shared context. Both the academic and the student mobility can be seen in the light of internationalisation within higher education. However, the motives for mobility are expected to differ between the two groups. Academics are in a different life stage and a further point in their career. The future of the beginning academic will likely be within the university, while most of the students will end up working for companies or governments.
2.4 ACADEMIC MOBILITY
The mobility of university staff and post graduates has been a less studied topic in the immigration literature (Kyvik et al., 1999; Storme et al., 2012: 13). However, flows at a post-doctoral level or higher are of a higher societal significance. This group are the ‘best and brightest’ and contribute most in the global knowledge flows. Therefore this group has more economical value and a greater societal relevance than other groups (Ackers, 2005: 105). There is a need for more research on the motives and barriers behind the level of international mobility among academics (Thorn, 2009: 442; Vance, 2005). Academics are seen as a specific group and their motivations and level of mobility are expected to differ from the findings from studies on undergraduate students and highly skilled workers. This empirical research will contribute to the highly skilled labour mobility literature and by examining academics as a special category within the highly skilled labour force.The study will focus on both, the long term and short term mobility. Migration patterns of academics are less often permanent, but have become more temporary and multidirectional in nature (Albert & Hazen, 2005; Ackers, 2005: 114).
The expectations of academics to be willing to travel to foreign countries and engage in international cooperation have never been so high. This is true for both the short-term mobility and living in another country for a longer period of time. Expectations for academic to be internationally mobile majorly have to do with the place based nature of knowledge. Physical proximity has become an important factor in exchanging and creating scientific knowledge (Storme et al., 2012: 12). If one wants to pursue a successful academic career a high level of geographical mobility is a requirement (Leemann, 2010: 610; Ackers, 2005: 104). By traveling internationally the academic can acquire this locally bound knowledge. For this reason the following hypothesis can be adapted.
Hypothesis 1: The majority of the PhDs, post-docs and associate professors employed at the university of Amsterdam travel internationally for reasons related to their career.
2.5 THE GEOGRAPHY OF INTERNATIONAL MOBILITY
The study focusses not merely on frequency as indicator for geographical mobility, but will also adapt a geographical perspective. This geographical perspective gives more insight in where the academics go and analyses whether mobility is truly global or more regional phenomenon. Expectations on the direction of mobility can be drawn from the brain circulation literature.
The concepts of ‘brain gain’ and ‘brain drain’ are useful for explaining the spatial patterns of movement of highly skilled labour. A situation of ‘brain gain’ occurs when more science and technology labour is migrating in the country than is moving out the country (Mahroum, 2000: 2). Brain gain is favourable for the national economy and therefore governments try actively to attract the highly skilled workers. On the other hand ‘brain drain’, where more highly skilled labour is migrating out the country, can have a negative effect on the national economy (Straubhaar, 2000). In practise this often means that highly skilled personal is moving from lower developed countries to more advanced economies. High educated people from developing countries can earn more money for the same job if they more to wealthier countries (Massey et al., 1993: 433)
For the Amsterdam case this might implement either a brain drain to the US or a brain gain from developing economies like India and China. Because of its open economy and high quality educational services the US has been attracting foreign highly skilled workers. Almost
all of the PhD students in the US have a foreign background (Mahroum, 199: 19). The Netherlands is attractive for foreign student from low developed countries with a relatively good level of English proficiency, education and income.
However, the geography of mobility is multi-faceted and not a one directional process. Also, the migration patterns can vary on length of stay and per discipline (King, 2002: 94). During the 1990s the paradigm has shifted away from the focus of ‘brain drain’ to a more fluid perspective of ‘brain circulation’. This perspective has a world system standpoint and emphasis the temporary character of skilled migration and the interconnections between places (Alberts & Hazen, 2005: 133). The brain circulation viewpoint suggest that international mobility is no longer south – north, this is changing with the economic growth of Asia (Mahroun, 2000: 28,29). On the other hand, the connections, institutions and spatial proximity between European counties may suggest that mobility patterns are limited to the borders of Europe. The literature on student mobility shows that the geographical pattern is mainly concentrated in Europe. An explaining factor for this is the EU funded ERASMUS program that help students with organizing their trips abroad and financial support (Alberts & Hazen, 2005: 134). The ERASMUS program also exists for academics and this is expected to stimulate the intra-European mobility. The European Union has actively promoted the labour market mobility by removing barriers and reducing adjustment costs (Van Der Velde, 2004:101). The highly skilled workers are currently the most mobile group within the EU and this brain exchange is expected to grow even more in the future (De Gijsel & Janssen, 2000: 61-62).
Hypothesis 2: The geographical scope of the spatial mobility of the young academics employed at the University of Amsterdam is concentrated in the territory of the European states.
2.6 FACTORS EXPLAINING THE DIFFERENT LEVELS OF SPATIAL MOBILITY
Except for mapping the amount and the geography of international mobility, this thesis goes further and tries to understand the factors that can explain the differences in mobility. Not all the academics are spatially mobile to the same extent (Ackers, 2005: 113). This thesis tries to give more insight in these different levels of mobility by looking at international network,
previous international experience, personal factors, economic factors and the length of contracts.
A first explaining factor for spatial mobility is the change from lifelong employment to more short-term contracts of academics. Starting academics are no longer always getting a long term contract and do not stay at one university their whole career. This can be seen as a consequence of globalization: more liberalization of the universities has led to more academics are being offered a short-term contract. This means that they have less security and need to keep orientated on job openings on an international level. For this reason the academics move more frequently between states due to a new job (De Gijsel & Janssen, 2000: 67, 69) and they maintain their international networks (Kim, 2009: 399-341; King, 2002).
Hypothesis 3: Academics who have a contract for a shorter period of time have a higher level of spatial mobility than academics with a contract for a relative longer period of time.
Being involved in international networks has become a central element in the career of many academics. These networks are important for the production and diffusion of knowledge. For the academics it is important to be connected and be actively involved in international academic networks. The networks are important for staying up to date of the newest knowledge, getting published and cited frequently, and staying informed on job openings. Regular face to face contact with actors in these groups of academics is crucial. As noted earlier, this is because of the tacit knowledge that requires proximity for exchange (Storme et al., 2013: 12). The networks can be seen as transnational social capital and are, among other things, build up and sustained through personal contact, attending conferences and research cooperation. Actually meeting colleagues is a necessity to maintain and develop networks (Kim, 2009: 400-403; Altback, 2013).
Moreover, studies based on Swiss and UK students show that previous international experience is an important factor in explaining the level and geography of mobility (Ballaz & Williams, 2004; Leeman, 2010). An excessive percentage of scientists with a relatively high level of international mobility also participated in an international exchange while they still were on an undergraduate level. The network they established in this period is an important determent for their international mobility further on in their careers. Their mobility level can partly be explained by the networks they built up during this period they spend abroad (Ackers, 2005: 108) Thus, migration flows result in the development of networks consisting
of interconnected nodes that shape further international movements. Therefore, networks stimulate the international mobility and also partly determine the location of that mobility (Williams et al, 2004).
Hypothesis 4: Academics who have previous international experience are more involved in international networks and have a higher level of spatial cross-border mobility than academic who lived in the Netherlands their whole career.
Personal factors
A second group of factors that can help explain the difference in the level of international mobility among academics are the personal characteristics and personal situation. A first aspect here is the age of the academics, the expectations to acquire international experience is the highest among people who are in an early stage in their career (Ackers, 2005: 104; De Gijsel, 2000: 65). Also age is a determining factor in the relative importance of the other motives. The barriers and motivations differ in different age groups (Storme et al., 2013: 455). The pressure to be cross border spatially mobile is expected to be highest for academics at the start of their career. Therefore, the scope of the bachelor thesis will be limited to young academics only (PhD, postdoc, assistant professor).
Hypothesis 5: Academics at an earlier stage in their career are more internationally mobile than academics at a later stage in their career.
Other possible barriers for academics their international mobility are gender, having a partner and being a parent. Having a partner or a child are considered a factors that bind academics to a location and decreases the amount of international travel. Several studies show that both having a partner and having a relationship has a negative correlation with the amount international spatial mobility (Leeman, 2010; Storme etal., 2012).
Hypothesis 6: Academics who have a child are less internationally mobile than academics without a child.
Hypothesis 7: Academics who have a partner are less internationally mobile than academics who are single
This relationship is apparent for male and female respondents, but is stronger associated with female respondents. A study among Swiss academics shows that women are less often able to 14
combine an international career with having children in comparison to men. An explanation is that the male partner is less often willing to take over more responsibilities and invest time (Leemann, 2010). However there must be noted that this research under Swiss academics only considers stays in a foreign country with a period longer than five years as dependent variable, while this study also takes the short time stays into account. Long stays have a bigger impact on one’s life than short trips and the role of gender may be less prominent in the latter case.
Hypothesis 8: The negative effect of having a child on the level of international mobility is stronger for female academics than male academics.
Economic factors
A third factor that can influence the level of international mobility of young academics is the economic situation. The first theories on migration emphasized the importance of economic motives for migration. According to the neoclassical macroeconomic theory, migration is caused by geographical differences in supply and demand of labour. All people are assumed to act as rational actors; they always strive for utility maximization, and therefor migrate if wages are higher in another country (Massey et al., 1993: 433). The prospect of economic improvement is expected to be a motivating factor for academics to go abroad. In many cases this holds true not directly for higher wages, but indirectly for further career progression (Ackers, 2005: 104-105).
However, several studies point to the special value system of scientist and a priority of other aspects above financial gain. In general scientists attach greater importance to non-tangible factors and want ‘to do good science’. A motivational environment, freedom to choose research topics and more social factors, such as the position of science in society, can be important aspects (Ackers, 2005: 105-108).
A study on PHD students in Italy shows that the lack of financial funding also is an important barrier for mobility. International mobility is costly due to different living standards and travel costs, funding takes away a part of this financial barrier (Avveduto, 2001: 237-238).
Hypothesis 9: Academics who have more access to financial support travel more frequently internationally for reasons related to their career than academics without funding
Another factor highlighted in the literature is institutional arrangement and connections. Nevertheless, this study will not examine this aspect because this factor is expected to be the same for the whole population while they are all connected to the University of Amsterdam. Moreover, academics usually go on their own incentive and not through institutional agreements. Mobility is for academics usually shaped by networks, and not by organizational structures as is often the case with highly skilled personal in transnational corporations (Ackers, 2005: 103)
This bachelor thesis will now continue by examining the hypotheses deducted from the literature for a unique dataset of young academics employed at the University of Amsterdam. The data gathered from an online survey will show the cross-border spatial mobility that the academics undertake for their career, and explain the variety. But first, in the next chapter the methodological choices are explained and accounted for.
3. DATA AND METHODS.
The first goal of this bachelor thesis is to map the amount, geography and motivations of cross-border spatial mobility of young academics employed at the University of Amsterdam. Thereafter, it tries to understand the variety in the spatial mobility by looking at: the personal factors, previous mobility, extent of international network and financial factors. The first objective is more descriptive and, gives an overview of how often the group of young academics engages in international activities and. These activities may range from short trips that involve teaching, study, conferences or may be stays abroad for a longer period of time. Also, the locations of the activities are examined to get more insight in the geographic distribution of the international mobility.
The second aim is to get a deeper understanding of differences that exist in the volume and geography of cross-border spatial mobility. By using unique data gained from a survey among academics employed at the University of Amsterdam, this bachelor thesis tries to find patterns in the dataset. Statistical analyzing methods are used to test the in the theoretical framework formed hypothesis. It will be clear which factors characterize the relatively spatially mobile academics and which factors the non-mobile group.
3.1 THE RESEARCH QUESTIONS
The research question is: What are the volume, geographical distribution, and motives of job related cross-border spatial mobility of academics employed at the University of Amsterdam who are in an early stage of their careers? And to what extent can their level of spatial mobility be understood by looking at personal factors, previous international experience, network factors, and access to financial support?
The sub questions are:
- How often do the PhD-students, post-docs and assistant professors who are employed at the University of Amsterdam travel outside the Netherlands for reasons related to their work at the university?
- What is the geographic distribution of the short-term mobility of 2013 and the long-term mobility of academics in an early stage in their careers who are employed at the UVA?
- To what extent can differences in personal factors, previous international experience, access to financial funding, and international network account for differences in academics’ amount of cross-border spatial mobility?
- What are the most important reasons for academics to travel abroad?
- Do previous international experiences result in the establishment of an international network and a higher level of international mobility?
3.2 CAUSAL MODELS
Figure 1 shows the expected causal relationships that will be tested in the research. It visualizes the hypotheses previously explained in the theoretical framework paragraph of this thesis. In the model, the level of international mobility is the depended variable that can be explained by the independent variables: previous international experience, international network, personal factors, and economic factors.
Figure 1: Causal model of the independent variables that explain the dependent variable cross-border spatial mobility
Moreover, the effect of previous international experience on mobility may be an indirect causal relationship. It is expected that the correlation can be explained by international 18
network as intervening variable. If is held constant for international network, the effect of previous international experience on level of international mobility disappears. This relationship is indirect because the network build up during the exchange explains why academic who have previous international experience, have a higher level of international mobility. The effect of international network in the causal relationship is shown in the mediation model in figure 2.
Figure 2: Causal model of the expected intervening relationship of international network that explains the relationship between previouw international experience and international mobility.
The expected moderating effect of gender is shown in figure 3. The causal model visualizes hypothesis 7 in the theoretical framework. The expectation is that gender influences the relationship between having children and spatial mobility. The relationship between having a child and international mobility is expected to be stronger for female academics.
Figure 3: Causal model of the relationship between having a child and international mobility with gender as moderating variable
3.3 CROSS-SECTIONAL RESEARCH DESIGN
The study follows a survey design, or in Bryman’s typology a cross-sectional research design. He describes a cross-sectional design as follows:
A cross-sectional design entails the collection of data on more than one case and a
single point in time in order to collect a body of quantitative or quantifiable data in
connection with two or more variables, which are then examined to detect patterns of association (Bryman, 2008: 44)
This bachelor thesis research on academic’s cross-border spatial mobility fits perfectly in his description. The aim of the study is to look for variation between units of analysis, in this case the academics who filled in the online questionnaire. Also, the analysis does not involve a time component and the data is collected at a single point in time. The survey was active for three weeks from the 22st of April 2014 until the 13th of May 2014. This can be considered as one single point in time because the answers are not expected to differ between these dates. Questions in the survey apply to the spatial mobility over 2013 and, therefore, it does not matter when the questionnaire was filled in. Moreover, questions in the survey are standardized and the answers were pre-coded. This results in quantifiable data and allows for construction of a data- matrix. The data-matrix consists of quantifiable numbers for the different variables and therefor statistical analyses can be applied. The research looks for patterns, and aims to find causal relationships between the dependent and independent variables.
The decision for a more quantitative research strategy has some epistemological and ontological implications. This study aims to collect quantifiable and objective data to measure a social phenomenon. Different hypothesis are deducted from theory and the existing literature and tested for the dataset (Bryman, 2008: 13-23). However, the reader must be aware that this more positivist epistemological view and objective ontological assumption have its limitations. The thesis is able to demonstrate the factors that lead to different levels of mobility, but cannot explain why. The reason why these factors influence the reasons for mobility is unclear and is open for interpretation (Bryman, 2008: 159,160). To get a better understanding of how the processes work, more qualitative and in-depth information has to be collected.
3.4 RESEARCH GROUP AND SITE
The research group consists of academics who are at the beginning of their careers and currently employed at the University of Amsterdam. The decision to study academics in the beginning of their career is made, because internationalization in higher education is a relatively new phenomenon. The globalization of the universities and the increased spatial mobility of academics is a process of the last decennia. For this reason, the pressure to be 20
cross-border spatial mobile is strongest for the academics that are in an early stage of their career (Ackers, 2005: 104; De Gijsel, 2000: 65). The expectation is that a cohort effect exists; this means that academics that are born in a later period have a higher rate of international geographical mobility (Biemans, 1981: 72). However, looking at differences in age is not the goal of this research. Instead the scope of the study is limited to academics in the beginning of their careers only. It is assumed that academic in the beginning of their career have a higher level of international spatial mobility and are therefore a more interesting group to study. The process of globalization of the universities is unlikely to stop soon, and new academics will start within this international context. Research on the group with the highest amount of mobility is thus more relevant, because other groups are likely to receive the same pressure to be spatially mobile in the future. Thus, insight in the young academic can be relevant for the future development of other groups.
There is chosen to select the research group on function instead of age. The reason for this decision is that people can start at a high age with their PhD. By selection on function instead of age, people who start their careers at a higher age are still included in the research. The first group selected in the research are the PhD-students. Doing a PhD is a logical step after a person achieves his or her graduate degree and aspires an academic career. This does, however, not mean that all the PhD-candidates have a future career within the university. A big part of this group will end up at government or private institutions. Despite the fact that not all the PhD candidates will end up working within the higher education sector, this group is still relevant to study. The individuals who do not strive for an academic career are at the moment employed by the university and are doing research and are part of the higher education sector. This current research function makes them interesting to study. In the Dutch education system, a logical step after a PhD is ether becoming a Post-doc or an Assistant professor (UD). The difference between the two is that the former has more a research function, while the latter has a more teaching function. A post-doc position gives the opportunity to conduct individual research in a research group for another two years. Assistant professor and post-docs are likely to continue their careers within the higher education sector.
The University of Amsterdam is an interesting case because it is a big university with a great variety in disciplines. Unlike some universities the University of Amsterdam is not specialized in a specific faculty, but has a big social science, humanities and a science
department. This multiplicity of faculties makes the University of Amsterdam a representative case and provides an opportunity to study difference between faculties.
Further, compared to other European countries, such as the UK, France and Germany, the Netherlands has a relatively spatially mobile population. The geographic mobility in the Dutch university sector is relatively high. This fact indicates an internationalized university sector (Bekhradnia & Sastry, 2005). Also, English is widely spoken as a second language in the Netherlands. Especially persons enrolled in a university have a high level of English proficiency. In a language point of view, the Dutch universities are more internationalized than for example the German and French, where English is less often used.
Within the Netherlands, Amsterdam is the capital city and it has an international character. Because of the good accessibility of Amsterdam due to the presence of Schiphol airport, inability to travel is unlikely to act as a barrier.
The thesis also includes the academics with non-Dutch nationalities who are working at the University of Amsterdam. This group acts as a substitute for the staff formerly employed at the University of Amsterdam who are now abroad for a longer period of time. The staff that is in a foreign country for a longer period of time will not check their e-mail regularly and can therefore not be included in the research. Data from NUFFIC (2012) suggests that the ingoing mobility is about the same size as the outgoing mobility. So the amount of respondent with a high level of international mobility will not be overestimated if the foreign academics are included in the sample.
3.5 RESEARCH METHOD FOR COLLECTION OF THE DATA
A method of ‘self- completion questionnaire’ is used to collect the data for the research. The questionnaire was built in an online survey program named ‘thesis tools’. Thereafter, a link to the questionnaire with an introducing letter was sent to people in the sample. The respondents could activate the survey by clicking on the link and fill in the questionnaire directly.
As a consequence of the limited time and resources available for the research, the online questionnaire was the most suitable method. The use of the mail questionnaire made it possible to address a large sample (Bryman, 2008: 216-218). Due to cooperation with the different research institutes, a big part of the total population could be included in the sample. Also, the mail survey gives the responds the option to determine at which moment they want 22
to fill in the questionnaire. Because of the (in general) tight schedules of academics, this is considered as a big advantage. Completing the questionnaire took only about 10 minutes and the respondents could determine themselves the moment they wanted to start the survey. The used method of mail questionnaire also has some limitations. The survey is made up almost entirely of closed questions. Therefore, the data gathered is highly standardized and more explaining open answers could not be included. In addition, since the respondent filled in the questions alone, it was not possible to elaborate if specific questions were unclear. This might have biased the results, in case the respondents misinterpreted the questions and did not answer accordingly to the intention of the question. However, to limit the bias resulting from misinterpretation the questionnaire was piloted on two academics before it was distributed. These two respondents were asked to expound on the questions they did not fully understand the meaning of, or were not able to answer accordingly within the categories. Despite the survey was piloted, the amount of missing data is relatively high. The missing data is the result from respondent who were unable or unwilling to fill in one or more questions. Because of missing data not all the questionnaires are useful and not all respondents’ answers can be included in all the analyses.
Another disadvantage resulting from the decision for a mail questionnaire instead of the alternative of a structured interview is no control over the environment. The validity is threatened if other persons are present when the respondent fills in the questionnaire. Other person may read the questions and influence the respondent. Academics do often have shared offices and may discuss the questions. However, this bias is expected to be limited, since the questions are not very sensitive in nature and email addresses are private.
Also, the mail questionnaire typically results in lower response rates. It is easy for a respondent to ignore the request to fill in the survey in case he receives it by email. In case a person is addressed in person, it is harder for him to decline cooperation with the research. A low response rate is again risk to validity and representativeness of the results. Paragraph 3.9 will go further into detail on this topic.
Although, if instead a method of structured interviewing was used, the same sample size would not have been possible. With the limited time and resources the mail questionnaire was the best available method that allowed for the collection of data about 200 respondents.
3.6 OPERATIONALIZATION OF THE CONCEPTS
Before hypothesizes can be tested, the social reality had to be simplified, and standardized. In the theoretical framework different concepts, building blocks of theory have been introduced. However, to be able to effectively measure the social phenomenon these concepts have to be further operationalized (Bryman, 2008: 143-145). The variables in the survey are constructed by structurally operationalizing the concepts. The used questionnaire is included in attachment 3 and a schematic table of the operationalized concepts is shown in the table below.
Table 1: Operationalization of the concepts
Concept Dimension Indicator/variable
Cross-border
spatial mobility (short-term)
- Amount
- Geography
- Days in foreign country/ a year for work - Number of stays - Length of stay Previous international experience (long-term) - Amount - Geography - Study abroad - Living abroad - Non-Dutch nationality
Reasons for spatial mobility - Acquire international social capital - Acquire cultural capital - Acquire symbolic
- Develop research network - Meeting colleges
- Language
- Cultural perspective - Reputation
capital - Place specific - financial - Good on CV - Status of university - Tangible (infrastructure) - Intangible (good science) - Higher future wage Personal situation - Family
- Academic - Personal - Parenting - Partnering - Faculty - Function - Length of career
- Research and/or teaching function - Age - gender - Access to financial support - Funding - Capital - Costs - Amount of funding - Income level - Saved money - Travel cost - Cost of living
For the mobility concepts a multidimensional profile is formed and besides amount also geography is included as a dimension. This multidimensional profile is better able to measure the concepts of cross-border mobility (Bryman, 2008: 149). People can score relatively high on the amount of days they spend abroad, while their mobility is only concentrated between two countries. This is another kind of mobility, than when a person travels to more different places. For this reason, the amount of days a respondent spend abroad, but also the geography of the mobility is measured.
This thesis distinguishes between mobility for a short time period in the calendar year 2013 and long-time mobility for a period longer than 3 months. A distinction between the two is significant, because mainly the personal factors are expected to play a different role (Ackers, 25
2005: 114). Also, the previous mobility in the career is expected to influence the amount of mobility in 2013.
3.7 SAMPLE & RESPONSE
For selection of the sample a strategy of non-probability sampling had to be used. The main reason for this is that no sampling frame was available to randomly select cases from. A list with all the academics in the population does not exist and could not be constructed. Personnel at the University of Amsterdam are not organized by one central secretariat, but de-central by each research institute separately. In total there are a number of 20 research institutes connected to the University of Amsterdam, see attachment 1. This de-central organizational structure and privacy protection regulation made impossible to construct a sampling frame.
Instead, the units included in the sample are selected with a form of convenience sampling. Since random sampling is impossible, there is chosen to include as much units of the population as possible. The population size is relatively small and consists of approximately 1400 PhD’s, post-docs and Associate professors employed at the University of Amsterdam. The chosen method of online surveys makes it possible to conduct questionnaires with relatively small effort. For the research was tried to send the survey to the whole population. The academics at the University of Amsterdam are organized in twenty different research institutes with each its own secretariat. These research institutes vary in size from the ‘astronomical institute Anton Pannekoek’ with 7 members, to ‘Amsterdam institute for social science’ with over 200 connected researchers. The individual research institutes are again divided in different research groups. To select the units each research institute is approached with the question to forward an email with the survey to the PhDs, post-docs and assistant professors connected to the institute. Further, the secretariat was also asked to give information on the exact number of academics the email was sent to. In attachment 1 a list of the amount of respondents included in the sample per research institute is included.
Of the 20 research institutes, a total of 14 collaborated with the request and distributed the survey among connected academics. It is estimated that the survey is send to approximately 1000 Post-docs, PhDs and Assistant professors. The average age of the respondents in the survey is 31,98 years and the percentage women in the sample is 54,9%.
3.8 RESPONDE AND REPRESENTATIVENESS OF THE SAMPLE
The total size of the sample is 200 units of analysis. A limited amount of time and resources made it impossible to collect a larger sample. It was attempted to approach all the respondents in the sample, and 200 was the number of academics that could be reached and wanted to participate. To increase the number of respondents, there could have been sent a reminder email. Although, It was clear that not all the research institutes wanted to participate in sending a reminder email. For more consistency, there is chosen to approach all the respondents one time. Unfortunately a total unit of analysis of 200 is small for some statistical analysis. The limited amount of units gives problems with empty cells and cells with small number counts in the contingency table. Despite variables are recoded to have less categories, the cell counts are still too small for some three-dimensional contingency tables. For more detail on the recoding of the variables see attachment 2. Most analysis in the thesis are limited to bivariate analysis, where direct causal relationships are examined. Multivariate analysis were in many cases not possible.
The sample size of 200 respondents is not very large, and more response would have increased the reliability of the results. However, this sample size can be sufficient in getting an understand of academic’s mobility. Moreover, The sampling fraction is rather high. This means that a relative large part of the population is included in the survey. The reason for this is that no method of random sampling was used, but there was tried to include the whole population the sample. Although according to Bryman (2008: 179), this relative size does not say much about the representativeness of the sample. It can be the case that the part of the population included in the sample differs from the part that is excluded from the sample. Also, a small amount of extreme values can already have a big influence on the final results. This increases the chance that the findings are a result of coincidence and the found relationship does not really exist.
The response rate of the survey is somewhere between 15% and 20%. The calculation of the exact response rate is impossible because of uncertainty of the amount of respondents that received the questionnaire. This uncertainty is caused by the fact that not all the institutes sent the survey to the correct group. Three institutes have sent the survey to the whole research group while others sent an email to the PhD students only. So, if the research institute provided the exact amount of individuals that received the email, this does not mean that they are all included in the survey. In cases where the whole research institution was 27
addressed, individuals who are not allowed to participate in the study also received the questionnaire. Due to the lack of exact figures, the response rate had to be estimated based on the available data and information provided by the website of the University of Amsterdam (2012) and the research institutes.
Sending to the whole research institution does not result in direct problems. In the introduction letter of the questionnaire it was clearly stated that the survey was only meant for PhD’s, Post-docs and Assistant professors. Academics that do not belong to one of these groups just have to ignore the email. This does thus not result in specific groups being overrepresented or underrepresented in the survey. The only problem here is that it is hard to guess the amount of people in the population that received the survey and therefor it is difficult to calculate the exact response rate.
If Bryman’s criterion (2008:219) for acceptable response rate to postal questionnaires is used, the response rate of 20% would be unacceptable. A response rate lower than 50% already carries an unacceptable risk of an unrepresentative sample. However, in his article Mol (2013: 212) states that a response rate between 10% and 20% is common for web based studies among students. In his article he suggests that a student survey with 10% can still be considered trustworthy.
The biggest threat to the representativeness of the sample is that the group that did not participate in the study differs significantly from the respondents that are included. The figures in paragraph 3.7 show that the male to female ratio is equal to the mean at the University of Amsterdam. The percentage women that participated in the survey was 54,5%, compared to 51% women of all the PhDs.
However some figures suggest inconsistencies. First, Academics connected to the ‘Amsterdam Research Institute for Legal Studies’, ‘institute of biodiversity and ecosystem dynamics, ‘research institute AMC’ and ‘Research institute of child development and education’ are not included in the sample. Especially the exclusion of ‘research institute AMC’ and ‘Amsterdam Research Institute for Legal Studies’ may threat the representativeness of the sample. The two institutes are very large and have a highly specialized field. If the academic’s level of mobility in these field is higher than in other field, the results may give a wrong picture.
Also, the sending to PhDs only by some institutes may have led to an overrepresentation of PhD-students in the sample. The reason for this is that for some research institutes the post-docs and associate professor could not be included in the sample. The sample consists of 71, 3% PhD students to 28,7% post-docs and assistant professors. Based on figures from the University of Amsterdam (2012: personnel), a more equal ratio is expected. If the hypothesis of a higher level of mobility for academic in an early career stage is true, the mobility in the results may be overstated.
As final remark, there is a chance that selection bias has resulted in that the actual level of spatial mobility is overstated in the sample. Academics who travel internationally for their careers might be more willing to participate in the research than academics that do not travel. The spatially mobile group feels more connected to the subject and is therefore more willing to spend some time to answers the questions. Instead, it might be the case that the non-mobile academics do not see the relevance of participation in the research. If this is the case the results may suggest a higher level of international mobility of Academics employed at the University of Amsterdam than is actually the case.
3.9 CONSTRUCTION OF THE VARIABLES
The different questions from the questionnaire are used for statistical analysis in the computer program IBM SPSS statistics 20. However, the data matrix cannot directly be imported in the program, but modifications are necessary. To construct variables that are usable for the analysis most questions had to be recoded. A first reason for this is that the unusable values have to be excluded. If respondents misunderstood questions and values are unusable, they were excluded from the analysis. A second reason is that for some variables the values had to be grouped or regrouped in new classes. This is done because in this way the cell filling in the analysis is better. If the cell count in the contingency table is too low, the found relationship is not valued. A third reason is that in some cases new variables are constructed out of different questions in the survey. This is for example done for the construction of the network variable. Questions 19 and 20 are combined, so the variable presents the amount of people in the network and the amount of contact at the same time. The exact construction of the different independent and independent variables is highlighted in attachment 2.
3.10 METHODS FOR ANALYZING THE DATA
The first goal of this bachelor is very descriptive and aims to give an overview of the cross-border spatial mobility of the academics. For this reason, in chapter four several pie charts and histograms are used. Diagrams are able to present the data in a relatively easy understandable way. Further, the mean and standard deviation of the variables are used to give more information of the distribution and dispersion of the values. Finally, for representation of the geographical perspective, maps are made. The computer program GIS is used to visualize the data from the survey in a world map.
For the bivariate analysis in paragraph 5, contingency tables are used as method for finding causal relationships between the variables. The contingency table is an easy way of analysing relationships between the variables. The combinations of the percentages in the table and, Cramer’s V, gamma and test for statistical significance show whether correlation exists between the variables. Cramer’s V gives the strength of the correlation on a scale of 0 to 1 and is used for the nominal variables. Gamma is also able give the direction of the correlation and has a value from -1 to 1, where the former is a negative correlation. In case both the variables are of ordinal or dichotomous, the gamma is used a measure of association. Although gamma is a method of rank correlation and suitable to measure the strength of association between variables of an ordinal level, it can also be used for dichotomous variables. If the chance of non-existence of a relationship is very large, Cramer’s V instead of gamma is used. This is done because with a non-significant relationship, the direction of the relationship is of no value.
For acceptance of the hypothesis a maximum significant level of 5 percent is used. The significance level is the accepted level of risk that a found relationship is accepted, while there is in reality no relationship exists. If a found relations ship is significant at a lower significance level, this level of risk is used. This is done because in case a relations ship is valid at a significance level of 1 percent, the chance that the relationship does not exist is small.
4. DESCRIBING THE DATA
This chapter gives an overview of the amount of academic’s international travel, the reasons for their mobility, and its geographical dimension. In the analysis different dependent variables for level of international mobility are used. First, the short-term mobility variable is examined. This measures the amount of days a person has spent in a foreign country in the calendar year 2013. Second, the long-term spatial mobility previously in the career is examined. All academics that have been living abroad or are currently living abroad are labelled as ‘internationally mobile’. The third dependent variable measures only the number of respondents that are currently living abroad as ‘mobile’. In the survey this group consists of the academics that are not born in the Netherlands and are currently employed by the University of Amsterdam. As mentioned before, the third variable had to be constructed for purpose of analyzing the causal effect of having a child and having a partner.
After the different measures for cross-border spatial mobility are described, their geographical dimensions are examined and compared. In case of the long-term mobility the geographical dimension is examined for the outward mobility and inward mobility separately. This enables the examination of the difference between locations of Dutch academics that go abroad and non-Dutch(foreign) academics that come to the University of Amsterdam. Finally, the different reasons that respondents named for their international mobility are described.
4.1 MAPPING THE SHORT-TERM SPATIAL MOBILITY OVER 2013. In the questionnaire the respondents were first presented with questions concerning their spatial mobility over the calendar year 2013. The total number of days an individual has spent abroad during the year 2013 is presented in table 2. Only 20,7 % of the respondents did not travel internationally for reason related to employment at the UvA and can be labelled as internationally immobile for 2013. This means that 79,3% of the respondents can be considered cross-border spatially mobile during 2013. The figures in table 2 show that, 25,3% of the respondents spent seven days or less abroad; 23,7% spent between 7 and 14 days abroad; 14,6% spent between fifteen and 28 days abroad; and 15,7 % spent more than 25 days abroad.
Two open-ended questions on the amount of international travel and average duration of a trip were included in the questionnaire. The use of this second pair of questions on mobility
results in more consistency of measures and allows for the calculation of descriptive statistics. The respondents spent on average 16,5 days in a non-Dutch country for reasons related to their employment at the UvA in the year 2013. The standard deviation of the nominal distribution is 27,1 days. The respondents made on average 2,2 trips a year and the average duration of the trips was 7,6 days.
Table 2 Frequencies of the amount of days abroad for reason related to employment at the UvA in the year 2013.
Days abroad in 2013 Frequency Percentage
None 41 20,7 %
7 days or less 50 25,3%
Between 7 and 14 days 47 23,7%
Between 15 and 28 days 29 14,6%
More than 28 days 31 15,7%
Total 198 100%
The figures show that the majority of the academics (79,3%) were spatially mobile over the calendar year 2013. Of the total sample of 200 academics, 31 respondents (15,7%) are highly mobile. This highly mobile group spent more than 25 days in a foreign country in the calendar year 2013. Also, the figures in the table and the standard deviation suggest that there exists great variety in the amount of days academics spend abroad. A standard deviation of 27,1 days means that the 5% academics with the highest amount of international travel, travelled on average about 70,7 days abroad. In a nominal distribution approximately 95% of the values falls within range of the mean plus two times the standard deviation. The equal distribution of the values over the categories in the table suggests variety in the data. Diversity in the data makes it relevant to investigate the factors that cause these different levels of mobility. The next paragraph tries to explain this variety by looking at several personal attributes of the respondents.
Apart from the amount of spatial mobility, the specific purposes of the trips made over 2013 are of interest. Respondents were presented with a closed question concerning their reasons for international travel. Most academics made one trip a year and were therefore allowed to indicate multiple answers. A frequency table of their answers is presented in figure 4. The table shows that ‘visiting congresses’ was named most often as reason for international travel, three quarter of the respondents named this as a motivation. ‘Research’ was named second 32
most frequent with 42,7% of the respondents naming it as purpose of the trip; 33,1 percentage of the respondents named meeting colleagues as a reason; 2,8% teaching and 12,4% indicated other reasons. Examples of other reason named by respondents are: ‘doing fieldwork’, ‘participating in an exchange program’ and ‘summer school’. It is apparent that visiting congresses is the most important purpose for the trips made over 2013. However, it is also possible that a single trip has multiple purposes. The formal reason of international travel may be the visiting of a congress, but informal academics also value the meeting of colleagues. This importance of informal networking is highlighted in paragraph 5.3.
Figure 4: Percentage of the academics that named a specific purposes for the trips made abroad in the calendar year 2013 (N=193)
4.2 MAPPING THE LONG-TERM MOBILITY PREVIOUS IN CAREER Apart from the short-term spatial mobility of academics over the calendar year 2013, the long-term mobility previously in their career is also examined. The long-long-term mobility measures if a respondent has spent a period of three months or longer in a country, which is not his country of birth, since the start of his or her academic career. The variable is constructed by examining places where the academics were at different moments during their career since the start of their undergraduate education. For more details on the construction of the variables see attachment 2. 0% 10% 20% 30% 40% 50% 60% 70% 80% Teaching Studying Other Meeting colleagues Research Congress 33
Figure 5 shows the percentage of the sample that has lived abroad compared to the percentage that has not. According to the diagram 68,4% of the respondents have been long-term spatially mobile at a stage in their careers. Thus, the data would seem to suggest that the majority of the academics have been living in a foreign country compared to a small amount hat have not.
However, when only is looked at academics who are born in the Netherlands , the amount of long-term mobile academics decreases. It can be seen from figure 6 that 45,6% of the Dutch academics have lived in another country for longer than three months since the start of their careers. Does this mean that non-Dutch academics are more long-term spatially mobile than the Dutch respondents? The answer is no.
The difference is caused by the non-Dutch academics working at the University of Amsterdam that are included in the sample. It is expected that this group studied outside the Netherlands at undergraduate and/or graduate level prior to the start of their PhD or Post-doc in the Netherlands, and is therefore automatically considered ‘mobile’. This group of academics works as a substitute for the native Dutch academics who are currently doing a PhD or Post-doc at a non-Dutch university. As mentioned before, this was needed because the group of Dutch academics who are currently abroad could not be approached for the
46,6% 53,4% Non-mobile Mobile 25,4% 74,6% non-mobile mobile
Figure 6: percentage of the native Dutch academics that has been long-term spatially mobile during their career (N=103)
Figure 5: Percentage of the respondents that is long-term spatially mobile or has been long-term mobile during their career (source: survey questions 8,9,10,11, 12, 15, 16) (N=193)
