1. Top of page
  2. Abstract
  3. Introduction
  4. A Review Of Relevant Literature
  5. Research Of Migration Motives
  6. Research Results
  7. Discussion
  8. Conclusion
  9. References

The main aim of this paper is to analyse the motives affecting the migration decisions of young people, particularly university students. Two scales were developed for measuring the perception of the importance of these motives. The data used in the research were collected via a survey of the opinions and attitudes of university students in Osijek, in June 2010. The paper also analyses psychometric properties of the scales – their dimensionality and reliability.

The results of a confirmatory factor analysis undoubtedly indicate that both scales are multidimensional constructs. A combination of the results of t-tests for an independent sample, factor analysis (exploratory and confirmatory) and reliability analysis suggest that emigration and stay motives are two sides of the same migration decision, and that they can be classified into several factors: the economic situation, social networks, insider advantages (that can be divided into inherited amenities and public-safety conditions) and the wealth of opportunities. Depending on the power of the initial and target destination, the factors can function as ‘push’ or ‘pull’ factors. The results of the study show social networks as being the only ‘pull’ factor for the city of Osijek, whereas the other factors, especially the economic ones, proved to demonstrate the ‘push’ effect. However, the effects of all factors were very mild.


  1. Top of page
  2. Abstract
  3. Introduction
  4. A Review Of Relevant Literature
  5. Research Of Migration Motives
  6. Research Results
  7. Discussion
  8. Conclusion
  9. References

Despite theoretical and empirical efforts, there is no unique and comprehensive migration theory (de Haas, 2007; Bodvarsson and den Berg, 2009). The reasons for that can also be found in the fact that a decision to stay in an area is not taken into account as a component of the decision-making process (Hammar and Tomas, 1997). There is a lack of knowledge as to why migration units, in spite of open opportunities and tendencies (as well as encouraging activities and measures taken by particular areas to draw people into their area), remain in their places of living and do not emigrate. This has been especially noticed in developing countries and Central and Eastern European countries (Hammar and Tomas, 1997; Fisher et al., 1997; Bodvarsson and van den Berg, 2009). This is also the case with Croatia where little research has been done to try to provide answers to those questions or to classify migration motives in a meaningful way.

The main aim of this paper is to analyse the motives influencing the migration decisions of young people, particularly university students. The identification of these motives, as well as their systematization and classification into factors, contributes to a better understanding of migrations. It can also become an important parameter for the management and development of an area. Based upon a re-assessment of previous theoretical and empirical findings, two scales have been developed for the purpose of measuring the perception of the importance of these motives, i.e. the list of emigration motives (LEF) and the list of stay motives (LSF). Young people belong to a category of those who are most prone to migration due to their generation habitude, as was noticed a long time ago by Ravenstein (1885, 1889) and Sjaastad (1962). This has been confirmed by numerous examples of empirical research studies throughout the world (Todaro, 1980; Franklin, 2003; Deshingkar and Grimm, 2005), as well as in Croatia, particularly when it comes to highly educated young people (Bozic and Buric, 2005; Sverko, 2005; Golub, 2006). The data used in this research were collected by means of a pilot study of opinions and attitudes of young people, i.e. students of the University of Osijek (Croatia), through a survey conducted in June 2010. Since the scales are relatively new in terms of their usage in empirical research, another aim of this paper is to analyse their reliability and dimensionality. Namely, only if the measurement instruments (such as scales) exhibited good psychometric characteristics, could conclusions drawn from them be considered usable and valid.

A Review Of Relevant Literature

  1. Top of page
  2. Abstract
  3. Introduction
  4. A Review Of Relevant Literature
  5. Research Of Migration Motives
  6. Research Results
  7. Discussion
  8. Conclusion
  9. References

The modern theory of migration has its roots in the second half of the 20th century. According to Bodvarsson and den Berg, 2009, it can be divided into three main streams, depending on the predominant motive for migration. The first stream of theories is concerned with a migrant offering his services (Sjastaad, 1962; Todaro, 1969; Haris and Todaro, 1970), who chooses a location enabling him to achieve the highest net income, or the location that is expected to do so. The initial approach of that stream of theories is a neo-classical approach that sees a migrant as an individual and a rational subject guided primarily by an estimation of economic factors (De Haas, 2007). The second stream of theories focuses on a migrant who is primarily a user of internal amenities and public goods offered by a certain location (e.g. Roback, 1982). Thus locations with various amenities not available in other locations (e.g. a pleasant climate, clean air and water) attract migrants. The third stream of theories focuses on a migrant as a producer of his own household goods and services (Becker, 1962). A location that can enable the production of the best combination of household goods and services will be the most attractive destination for a migrant. This stream is based on the so-called new household economics (Stark and Bloom, 1985; Stark, 1991).

There are also other streams that are likely to explain the decision to emigrate by other motives which are important for this paper. For example, social capital can play an important role in making a decision to emigrate and within it the existence of migrant or friend networks (e.g. Massey, 1990; Garip, 2008). This has made social capital an important migration resource. Still, its role must not be overestimated (De Haas, 2007) since it has to be considered in addition to other forms of capital (e.g. physical and human capital). The network theory is closely related to the migration systems theory, according to which migration is indirectly related to migrants' broader social environment, including social, cultural, economic and institutional conditions at both source and destination locations (Mabogunje, 1970).

It is necessary to point out that the economic migration theory (on a micro level) tries to identify migration units, reasons for migration as well as the consequences generated by such decisions for both source and destination locations (on a macro level), thereby neglecting the need for studying the decision-making process and motives in relation to the question as to why the majority of people stay in a certain area, and in this way choose not to emigrate (Hammar and Tomas, 1997). However, intensive research attention has been drawn to that aspect of migration consideration (e.g. Fischer et al., 1997). Within the framework of traditional theory, immobility is usually explained by various institutional obstacles, aversion to risk, control and discrimination. In recent years, the theory of Fischer et al. (1997), according to which the so-called insider advantages of some locations are the primary cause for staying in a certain place, has been especially popular. Insider advantages are linked to the accumulated results of work and leisure. The longer people stay in one place and the more insider advantages they accumulate, the less likely it is for them, according to the same authors, to emigrate. In this way social capital and local knowledge create new opportunities which do not exist to the same extent anywhere else.

Research Of Migration Motives

  1. Top of page
  2. Abstract
  3. Introduction
  4. A Review Of Relevant Literature
  5. Research Of Migration Motives
  6. Research Results
  7. Discussion
  8. Conclusion
  9. References

Research hypotheses

For the purpose of achieving the fundamental goal of our research, three research hypotheses were stated:

H1: a migration decision is a multidimensional construct dependent on a variety of motives that can be grouped into several factors, among which economic and social ones are of special importance;

H2: in the city of Osijek, economic factors are ‘push’ factors, whereas social factors are ‘pull’ factors;

H3: a decision to stay or to emigrate can be seen as two sides of the same migration decision affected by the same factors.

Measurement instrument, data and methods

A pilot study examining attitudes and opinions of students was conducted in June 2010 on a convenient sample consisting of 148 students of University of Osijek. The data were collected in lecture rooms in the presence of the authors and by means of an anonymous survey that had been used in 2008 with the same purpose (see Borozan et al., 2008). It was a so-called one-issue survey. To avoid bias of responses, the surveyed students neither took classes where migration had been discussed as a topic, nor did they have an opportunity to ask for any clarification regarding the survey questions. Although the survey's response rate was 100 per cent, 18 filled-in questionnaires were not included in the further analysis because of too many missing or irrelevant answers to survey questions.

The survey consists of three parts. The first part includes seven questions, mostly with closed answers, that refer to the demographic characteristics of respondents. In the second part, there follow open and closed questions in relation to the migration aspirations of respondents and assessment of the attractiveness of Osijek as a place to live, and finally a question (in the last part of the survey) which refers to the assessment of the importance of motives affecting a decision to emigrate or to stay in Osijek. In accordance with the defined goal, primarily data referring to emigration and stay motives were used in this paper. Two constructs were developed that were measured by two scales: a list of motives affecting decisions to emigrate (LEF) and a list of motives affecting decisions to stay (LSF). The structure of the scales LEF and LSF (see Table 1) was derived from theoretical and empirical findings presented briefly in the previous section (for a detailed explanation see Borozan and Barkovic Bojanic, 2011). Both scales consist of 24 variables (items) that are measured by means of a five-point Likert scale. Their perceived level of importance was assessed by all respondents (“Assess the level of importance the motives from the table below have on your decision to emigrate to a desirable place and their importance when deciding to stay in Osijek by grading the answers on the scale from 1 ‘completely unimportant’ to 5 ‘very important’”).

Table 1. Structure Of Measurement Scale: Researched Migration Motives
ItemsLegendItems LegendItems Legend
Housing opportunities (e.g. price per m2, availability)1Crime rate9Natural beauties17
Property (e.g. real estate)2Bribe and corruption (e.g. perception, spread)10Environmental cleanness18
Tax and local income tax3Vicinity of parents11Historical sights19
Income (e.g. regularity of payment, amount) Vicinity of friends12Cultural opportunities20
Availability of jobs 5Family roots13Recreational and sorting opportunities21
Cost of living 6Familiarity with people, places, etc.14Educational opportunities22
Size of place7People's life style15Entertainment opportunities 23
Quality of public service8Safety of place16Job characteristics (e.g. career advancement, challenges)24

When items make a scale, they should measure the same construct and the scale should be reliable - internally consistent. Both scales were therefore analysed with the goal of determining their reliability and their dimensionality.

Reliability is tested by means of the method based on inter-correlations among test items that test the internal consistency of the scale, as usual, by using the following criteria: item-to-total correlation, average item-to-total correlation and Cronbach's alpha coefficient (Nunnally, 1979; Bearden et al., 1989; Robinson et al., 1991). Their values range from 0 to 1, where higher values indicate higher reliability. An empirical suggestion for Cronbach's alpha coefficient is that the scale is acceptably reliable if the alpha value is equal to, or greater than 0.7 (Nunnally, 1979); for item-to-total correlations, items should have a correlation of 0.5 or greater in order not to be eliminated from the scale (Bearden et al., 1989); for average item-to-total correlation, they should be equal to, or greater than 0.3 (Robinson et al., 1991).

The dimensionality of both scales as well as the accuracy of the preset research hypotheses were examined by means of factor analysis (FA). As a generic term, FA, which can be exploratory and confirmatory, enables the analysis of interrelations which exist within the set of variables. When it comes to exploratory factor analysis (EFA), we applied the principal component method in which data are standardized by means of a correlation matrix. Normalized varimax rotation was used for factor rotation. The Kaiser criterion was applied as a factor selection criterion according to which all factors whose eigenvalues are greater than 1 are retained for further analysis. An EFA was applied for the purpose of reducing a set of variables, so-called items (manifest variables) to a smaller number of uncorrelated variables (so-called factors, latent variables) in order to classify them in a meaningful way without presupposed theoretical expectations incorporated in the model.

A confirmatory factor analysis (CFA) was conducted with the goal of testing hypotheses referring to pre-supposed relations among items. Therefore various models were tested, and an analysis was carried out by using the SEPATH module of Statistica 7. The same statistical software was used in the overall statistical analysis in this paper. Models were tested by the method of maximum likelihood. The CFA results were evaluated by following goodness-of-fit indices (GFI) that are usually used for that purpose: Joreski GFI, Joreski adjusted, AGFI, Bentler-Bonett Normed Fit Index (NFI), Bentler-Bonett Non-Normed Fit Index (NNFI), and Bentler Comparative Fit Index (CFI). Their values should be greater than 0.90 for an acceptable fit and greater than 0.95 for a good fit (Sun, 2005).

Research Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. A Review Of Relevant Literature
  5. Research Of Migration Motives
  6. Research Results
  7. Discussion
  8. Conclusion
  9. References

Influence of demographic variables on the segmentation of respondents with respect to migration

An average respondent is a 21-year-old female with a monthly income ranging from €137.25 to €274.49. The respondent's family members have a monthly income of between €686.23 and €1,372.45, the respondents have neither a flat nor a house, their housing problem is unsolved and they have one sibling.

A comparison of a sub-sample of respondents who would like to live permanently somewhere else (N = 69) with those who would not like to (N = 78) 1 shows that potential migrants: a) are more frequently female than male persons; b) have a lower monthly income at their disposal; c) have fewer siblings; and d) have unsolved housing problems. Namely, the independent sample t-test confirms that there exists a statistically significant difference between sub-samples when it comes to: gender (t-value: −2.78, p = 0.01), disposable family income (t-value: 3.38, p = 0.00), real estate (t-value: -2.89, p = 0.00) and a solution to the housing problem (t-value: 2.04; p = 0.04), at the 0.05 level of significance.

Many empirical research studies confirm the important role that demographic variables (e.g. age, gender, education, etc.) can play in the process of segmenting people according to migration aspirations (OECD, 2008, or in Central and Eastern European countries Szczepanikova et al. (eds.), 2006). The results of this research with respect to the comparison of demographic variables of both sub-samples confirm the world tendency, the feminization of the migration process and an increase in the tendency towards the migration of more educated people and people with an increasingly limited ownership of property. It is also in accordance with the developed profile of potential migrants from the new EU member states and acceding countries according to which these are young single people with better education, or third-level students (Hoffman, 2006).

Perception of the importance of motives affecting a decision to emigrate or to stay

Making a decision to stay is a process that is equally as complex as making a decision to emigrate. These two processes are actually mutually intertwined; but in real time, when it comes to the realization of the decision, they are also mutually exclusive. Hence the decision to stay is also determined by individual characteristics of a migration unit as well as contextual motives.

The paper focuses on the North-Eastern part of Croatia, particularly the Osijek-Baranja County and its centre – the city of Osijek. This is one of the counties lagging behind the Croatian average and its capital, Zagreb. This may be exemplified by the gross domestic product per capita of the Osijek-Baranja County, which was, according to the most recent data issued by the Croatian Bureau of Statistics (CBS, 2011), €8,871 for the year 2008, making it 83.0 per cent of the Croatian average.

Figure 1 graphically illustrates an average grade given by respondents for variables contained in both scales.

It is interesting that a statistically significant difference does not exist between grades referring to every motive of the LSF and LEF scale at the five per cent significance level, except where the question includes the possibility of finding a job (t = 2.29, df = 286, p = 0.02). That motive was perceived by respondents as a motive that is important for making the decision to emigrate rather than the decision to stay. Furthermore, concerning the question: “If you think about emigrating, would your decision include a cost-benefit analysis in relation to a) only a desirable place to live; b) the current and a desirable place to live; c) only the current place to live; d) I don't know?”, 95.89 per cent respondents (i.e. a total of 140) opted for b). This confirms hypothesis H3 in this paper.

In order to test the appropriateness of the sample for FA, we employed two tests: the Kaiser-Meyer-Olkin measures of sampling adequacy and the Bartlett's tests of sphericity. The results of the former test were 0.708 for the LSF and 0.788 for the LEF, exceeding the 0.50 required value for FA. Furthermore, the Bartlett's tests of sphericity (approx. chi-square = 4957.846, df = 1225, sig. = 0.000 for the LSF and approx. chi-square = 1575.982, df = 300, sig. = 0.000 for the LEF) were significant. This means that the strength of the relationship among variables is strong, and therefore our data are appropriate for FA.

Factor structure of the LSF measurement scale

The t-tests for the LSF scale showed that in the calculation of the mean, the importance of items differs from grade 3, i.e. a neutral attitude for all the items, except for the amount of taxes (t = 1.51, df = 141, p = 0.13) and the size of the area (t = 1.83, df = 141, p = 0.07) at the 0.05 significance level. In accordance with the aforementioned, respondents perceive all other factors as important in the process of making a decision as to whether to stay in the city of Osijek.

The existence of numerous items used for assessing the decision to stay also requires a reliability analysis. Statistically significant values of Cronbach's alpha coefficient (0.91) and the average inter-item correlation (0.30) show that all 24 items represent the scale well. The value of Cronbach's alpha remained unchanged (0.91) after the elimination of items not perceived by respondents as factors important for making the decision to stay (taxes and the size of the area), as well as property, whose correlation with the total scale was low (0.38). However, the value of the average inter-item correlation was slightly improved (0.32); hence in the procedure of further analysis we retained 21 items. The answer to the question as to whether the construct is one-dimensional or multidimensional can be given by a FA.

Theoretical considerations imply that the decision to stay in a certain place, as well as the decision to emigrate, is a multidimensional construct dependent on numerous factors. The CFA carried out on model 1, which includes 21 items and whose null hypothesis assumes a one-dimensional construct, confirmed that. Namely, the chi-square value, χ2 (χ2 = 908.95, df = 189, p < 0.001), as well as calculated goodness-of-fit indicator values (Table 2 for LSF, model 1), suggest a multidimensional construct.

Table 2. Results Of Confirmatory Factor Analysis Of The Lsf And Lef Scales
 ModelFactor structure, itemsaχ2; number of degrees of freedom Goodness-of-fit indices
Joreski GFIJoreski AGFIBentler Bonett NFIBentler Bonett NNFIBentler CFI
  1. “-“ implies that chi-square (χ2) and goodness-of-fit index values cannot be calculated since the corresponding number of degrees of freedom is missing.

  2. a

    Legend for items: see Table 1

LSF1Factor 1: all items908.95; 1890.5560.4570.4730.4720.525
2Factor 1:1, 4–6, 9–1088.34; 90.8320.6080.7780.6550.793
Factor 2: 11–13------
Factor 3: 16–2021.28; 50.9420.8250.9320.8920.946
Factor 4: 21–2422.85; 20.9380.6920.9040.7310.910
Factor 5: 8, 14–15------
Total Model 2459.42; 1790.7620.6930.7340.7820.815
3Factor 1: 1, 4–61.76; 20.9940.9700.9931.001.00
Factor 2: 11–1416.99; 20.9410.7060.9350.8250.942
Factor 3: 15, 17–193.43; 20.9870.9370.9830.9780.993
Factor 4: 20–2426.19; 50.9430.9550.9830.9870.996
Factor 5: 8–10, 167.38; 20.9770.8840.9590.9070.969
Total Model 3445.91; 1790.7710.7040.7420.7930.824
4Factor 1: 1, 4–6, 8–10103.5; 140.8270.6530.7680.6840.790
Factor 2: 11–1548.54; 50.8670.6000.8520.7260.863
Factor 3: 16–2021.28; 50.9420.8250.9320.8920.946
Factor 4: 21–2422.85; 20.9380.6920.9040.7310.910
Total Model 4479.53; 1830.7590.6960.7220.7750.804
LEF1Factor 1: all items1077.57; 2750.5730.4960.4000.4140.464
2Factor 1: 1, 4–63.47; 20.9880.9400.9790.9720.991
Factor 2: 11–1531.01; 50.9260.7790.9050.8360.918
Factor 3: 8–9, 16, 18–2034.22; 140.9380.8760.9140.9190.946
Factor 4: 21–2419.58; 20.9360.6780.9050.7360.912
Total Model 2390.54; 1640.7990.7420.7340.7940.823
3Factor 1: 1, 4–63.47; 20.9880.9400.9790.9720.991
Factor 2: 11–1411.69; 20.9580.7920.9560.8890.963
Factor 3: 15, 17–190.65; 20.9980.9890.9971.0211.000
Factor 4: 20–2433.99; 50.9090.7270.8570.7450.873
Factor 5: 8–10, 162.09; 20.9930.9640.9870.9980.999
Total Model 3462.90; 1790.7770.7120.9560.8890.793
4Factor 1: 1, 4–63.47; 20.9880.9400.9790.9720.991
Factor 2: 11–1411.69; 20.9580.7920.9560.8890.963
Factor 3: 8–9, 16, 18–2034.22; 140.9380.8760.9140.9190.946
Factor 4: 21–23------
Factor 5: 15, 24------
Total Model 4401.51; 1600.7940.7920.9560.8890.811

After conducting an EFA with normalized varimax rotation, five statistically significant factors were extracted, which leads to the conclusion that the list of stay motives is five-dimensional. With respect to gathering items (Table 2 for LSF, model 2), we deal with the following factors: the economic situation, social networks, inherited amenities, wealth of opportunities, and quality of public services. These five factors account for 68.72 per cent of the total variations in the model, and they also show an acceptable level of reliability (see Table 3 for LSF). Furthermore, items incorporated into factors mostly possess characteristics of convergent and discriminant validity. This means that most of the associated items have high factor loadings for corresponding factors and low factor loadings for other factors. However, there are items with a relatively small difference of loadings for two factors (quality of public services, crime rate, level of corruption and bribery), and there is an item with almost equal factor loadings for two factors (familiarity with people and place – factors 2 and 5) so that the factor it belongs to cannot be determined with certainty. Thus a dilemma occurred as to whether the model possesses better statistical properties with a different factor structure.

Table 3. Results Of Exploratory Factor Analysis And Reliability Analysis Of The Lef And Lsf Scales
FACTOR ANALYSISEigenvalue:7.422.741.841.381.04
% Total 35.3613.058.786.564.96
Cumulative (in %)35.3648.4157.1963.7568.72
Factor loading (FL)ItemsFLItemsFLItemsFLItemsFLItemsFL
90.65  200.69    
RELIABILITY ANALYSIS inline image 16.2214.7714.0815.0418.85
Std. Dv. 3.323.873.303.3053.76
Sample size140142141142142
Cronbach's alpha0.820.850.800.780.82
Standardized alpha0.830.850.800.780.82
Average item corr.0.560.600.510.480.49
Item-to-total correlation (ITC)ItemsITCItemsITCItemsITCItemsITCItemsITC
  1. Legend for items: see Table 1; Note: inline image and Std. Dv. represent mean and standard deviation, respectively.

FACTOR ANALYSISEigenvalue:6.182.522.231.41
% Total 30.8912.6211.177.07
Cumulative (in %)30.8943.4954.6761.74
Factor loading (FL)ItemsFLItemsFLItemsFLItemsFL
RELIABILITY ANALYSIS inline image 16.9418.3725.2315.93
Std. Dv. 2.754.495.173.12
Sample size145142143147
Cronbach's alpha0.760.850.850.80
Standardized alpha0.770.840.850.81
Average item corr.0.470.530.450.52
Item-to-total correlation (ITC)ItemsITCItemsITCItemsITCItemsITC

A CFA was also carried out for the purpose of resolving the dilemma that is testing statistical power showing not only the relation between latent variables and items at which the exploratory variable pointed as to belonging to them, but also other pre-supposed relations as well. Thereby four models were tested. Table 2 for LSF shows the structure of each of them. Under the assumption of the independence of measurement errors, model 1 assumes, as has already been mentioned, that the LSF scale is one-dimensional, i.e. that it contains only one latent variable. The structure of model 2 is a direct result of the conducted EFA. In model 3, items are classified on the basis of the results of EFA and their correction with other theoretical and empirical research findings. In model 4, four factors that account for 64 per cent of the total variance are assumed to describe the tested construct. Its structure is a result of the EFA. Results of the CFA for each of the models and associated factors are shown in Table 2 for LSF. When it comes to the model considered as a whole, goodness-of-fit indices were calculated under the assumption of the inter-correlation among factors.

Since the values of calculated goodness-of-fit indices show that model 3 corresponds best with empirical data, that model is explained in the sequel. It should indeed be taken into account that this did not completely achieve a perfect significance level either, which can be a consequence of the fact that the sample in this paper is not large, that it is convenient, and that the goodness-of-fit indices are sensitive to the sample size as well as the number of variables included in the model (Sun, 2005). However, taking the results pointing at a good level of internal consistency into consideration, the model definitely contributes to a better understanding of the factors influencing the decision to stay. According to model 3, stay motives can be classified into five factors: the economic situation, social networks, inherited amenities, public-safety conditions, and wealth of opportunities offered at the place of residence (Table 3 for LSF). This confirms hypothesis H1, specifically, the part that refers to the LSF scale.

Taking into account the values of the goodness-of-fit indices (which are greater than 0.90), each of the given factors shows a good fit of models to real data. Factors also express a high level of reliability. This may be exemplified by the criteria of internal consistency and Cronbach's alpha coefficients, which can be also seen in Table 3 for LSF.

Factor structure of the LEF measurement scale

Testing the hypothesis, which the mean of the assessment of the importance of every item making the LEF scale is equal to a neutral attitude, showed that respondents perceive every item as an important decision-making motive. The results of the CFA (χ2 = 1 077.57, df = 275, N = 25, p = 0.00) confirm the validity of hypothesis H1, i.e. they refer to a conclusion that the LEF is not a one-dimensional construct. Therefore, an EFA was carried out for the purpose of evaluating the internal consistency of extracted factors in a reliability analysis. Results pertaining to the application of the given methods are presented in Table 2 for LEF.

Four models underwent a CFA: model 1 checked out, as has already been mentioned, whether the construct in question is one-dimensional; model 2 tested the results of the EFA showed in Table 3 for LEF; model 3 tested the factor structure of the LSF construct which proved to be the best among respondents; and model 4 tested an alternative structure that was highlighted by the EFA as a potential one.

The results obtained by the CFA refer to the conclusion that model 2 best describes real data, whereas a reliability analysis (Table 3 for LEF) shows that factors are internally consistent. According to model 2, emigration motives can be classified into four factors: the economic situation (as the primary factor), social networks, insider advantages and wealth of opportunities. Thereby hypothesis H1 is completely accepted. The extracted factors partially correspond to results referring to brain-drain research conducted by Sverko (2005), according to which as reasons for emigration, the respondents (4th year students in Zagreb) list material reasons (earning money faster and easier, higher salaries, etc.), followed by reasons related to development (training, better working conditions, etc.), and then social reasons (uncertainty, bad economic situation, etc.). The domination of economic factors was also noticed in other researches conducted in Croatia (e.g. Golub, 2006).


  1. Top of page
  2. Abstract
  3. Introduction
  4. A Review Of Relevant Literature
  5. Research Of Migration Motives
  6. Research Results
  7. Discussion
  8. Conclusion
  9. References

The FA suggests that in both cases (stay or emigrate) the existence of several factors account for about two-thirds of the variance. These are the economic situation, social networks, insider advantages (inherited amenities and public-safety conditions) and a wealth of opportunities, by which it confirms the validity of hypothesis H3.

University students want to have a good quality of life and thus it is extremely important for them to be able to find jobs and to receive corresponding incomes. This need is especially emphasized by the economic recession in Croatia. Hence, the economic situation prevailing in the city of Osijek, as assumed by hypothesis H2, proved to be a key factor in affecting the decision to emigrate. This factor was extracted as a result of FA, and it accounts for about 1/3 of the total variance. It includes traditional economic variables related to efficiency and conditions for the functioning of the labour market, available jobs and incomes, the cost of living, and also to housing opportunities indirectly dependent on them. Respondents assessed the importance of these variables by means of allocating the highest grades (see Figure 1). If there are more new jobs available in a place, higher expected incomes, more favourable costs of living and more housing opportunities, it is more likely that a person will live there.

Empirical research results show that in developed countries motives linked to the labour market become less significant for making the decision to emigrate (Wikhall, 2002), although these motives still affect decision making (Gunderson et al., 2007). However, it may be expected that their importance has increased due to the consequences of the world economic crisis. These motives are still dominant in Central and Eastern Europe (Fassmann and Hintermann, 1998; Hoffmann, 2006), as well as in Croatia (Sverko, 2005; Borozan et al., 2008). With respect to the Osijek-Baranja County, taking the low level of economic development of that county, the high unemployment rate (28.1% in 2010) and average salaries lower than the Croatian average (€631.33 in the county compared to €708.32 in Croatia in 2008) 2 into account, the critical importance of the economic factor is expected. Based upon the higher average grades of the respondents (Figure 1), the economic factor functions slightly more as a ‘push’ factor.

The second factor that emerged from the FA is social networks, incorporating variables such as the vicinity of parents and friends, roots, familiarity with the area and people living there. It accounts for about 13 per cent of the total variance. The more important each of the given variables is, the greater the influence social networks will have on making a decision about the place to live. Numerous theoretical and empirical research studies confirm that social networks play a crucial role in decision-making by potential migrants as well as non-migrants (Massey, 1990; Massey and Aysa, 2005). University students are more mobile, single, less burdened by families and their own property. However, taking into account average grades implying that these are important decision-making motives (Figure 1), respondents feel that social networks are important factors that have a slightly higher ‘pull’ power towards staying in the city of Osijek. This confirms hypothesis H2.



Legend for items: see Table 1.

Download figure to PowerPoint

Insider advantages represent the factor with the most disperse structure and greatest variations between the decision to stay or emigrate. When it comes to staying, the FA showed that insider advantages can be divided into two factors: inherited amenities (including lifestyle, natural beauty, environmental cleanliness, etc.) and public-safety conditions (including the quality of public services, safety, crime rate, etc.). When it comes to the decision to emigrate, the given variables form one factor – insider advantages. Regardless of the factor being split into one or two factors, in both cases it accounts for about 12 per cent of the total variance. The term itself was coined by Fischer et al. (1997), who used it for the first time, and it includes location-specific factors that cannot be easily transferred to another location by an individual. On average, respondents perceive that the city of Osijek does not offer significantly different qualities compared with potential target locations.

The wealth of opportunities a location can offer to its citizens was also recognized as an important factor accounting for slightly less than 10 per cent of the total variance. It incorporates variables such as cultural opportunities, recreational and sporting opportunities, educational opportunities, entertainment opportunities in general, as well as job quality. The structure of that factor also implies its relation to the quality of life, just like the previous factor; however, it represents an upgrade that is more challenging and that enables people to gain new experience and self-actualize. Locations that are able to offer more of such opportunities will attract university students. Many authors have researched the importance of culture and cultural infrastructure as part of the regional infrastructure for highly educated people, detecting a positive correlation between the attractiveness of some locations and culture (Wikhall, 2002). When it comes to young and highly educated people throughout the world, then the wealth of opportunities represents the ultimate factor influencing the decision as to where to live (OECD, 2008), which also holds true for Croatia (e.g. Sverko, 2005; Golub, 2006).

An analysis of the results showed that factors affecting the decision to stay or to emigrate may be considered as two sides of the same final decision, but also as sides that both have to be taken into account. Factors that can attract university students to a certain location can also retain them in a location, depending on the power of each of them, as well as the mutual relationship between the sizes measured in the context of expected costs and benefits. The ‘push-pull’ coefficient (f) of 0.49 suggests that ‘push’ factors slightly overcome the ‘pull’ factors considering migration. This coefficient is calculated by using the following equation (1):

  • display math(1)

The f is the number between 0 and 1 where higher values indicate that people show more aspiration towards staying in their place of residence and lower values demonstrate emigrating tendencies. Numbers n1, n2 represent the number of ‘push’ and ‘pull’ items in the scales or factors. For the purpose of the equation we used the higher value. Ei and Sj are the mean values of ‘push’ and ‘pull’ items respectively. We subtracted the right expression from 0.5 to ensure that our result ranges from 0 to 1.

It is necessary to mention that the results of this research are indicative and that they cannot be unquestioningly generalized to the overall population of young people in Croatia. The primary cause for that is the sample size and structure. However, the classification of migration motives is important in order to provide a deeper insight into the migration aspirations of young people in the Osijek-Baranja County, particularly its university students. This is of importance for the regional government in its attempt to develop regional policies targeting young and highly educated people which would provide them with reasons and opportunities to stay. Furthermore, measurement scales developed for the purpose of this research have proved to be preliminarily reliable; yet they should be used repeatedly in future researches in order for all of the psychometric properties of the scale to be assessed.


  1. Top of page
  2. Abstract
  3. Introduction
  4. A Review Of Relevant Literature
  5. Research Of Migration Motives
  6. Research Results
  7. Discussion
  8. Conclusion
  9. References

Migrations represent a complex phenomenon and a process which has been in the focus of scientific interest for a long period of time. Still, there are many unknowns, especially the ones connected with not only a decision to migrate, but also a decision to stay in a particular area. Modern research studies show that migration considerations are also followed by the consideration of an option to stay somewhere. Thus, special attention in this paper was placed on the identification and classification of emigration motives and stay motives. Two measurement scales were formed, consisting of 24 variables each. The reliability analysis results implied that emigration factors and stay factors represent both constructs well. Namely, Cronbach's alpha has values 0.88 and 0.91 for the LEF scale and the LSF scale, respectively. Moreover, the CFA confirmed the expectations formed on the basis of relevant theoretical and empirical evidence showing that both measurement scales are multidimensional constructs.

The conducted EFA has pointed out that the primary factor with both scales is the economic situation. This means that when considering both options, potential migrants primarily consider these motives. It was proved that for the city of Osijek, the economic situation functions as a ‘push’ factor. On the other hand, social networks that were singled out as the second factor by the importance for variance explanation, function as a pull factor with regards to Osijek. The factor referring to insider advantage is the third factor in the variance explanation, and it can be considered either as a unit or as a factor divided into two parts – inherited amenities and public-safety conditions. With respect to the city of Osijek, respondents did not perceive these as especially attractive. The last factor singled out as a FA result is the wealth of opportunities. It is a factor which accounts for the smallest part of the variance. This factor enables the complete self-actualization of an individual, making the place offering it a real attraction for young, educated and creative people. Respondents objected that in the city of Osijek, there is a lack of such development opportunities.

The results showed that university students' decisions to stay or to emigrate are mutually related, that they are mutually conditional, mutually intertwined, and they are considered simultaneously in the context of costs and benefits. Finally, it is a strategic decision – to stay or to emigrate. A migration unit makes the decision at a certain moment of time, but it is re-assessed over the course of time, depending on the existing or expected future situation. Although only a slight majority of the respondents expressed their wish to stay in the city of Osijek (53.06%), the economic situation, lack of insider advantages, as well as the wealth of opportunities, lead towards a continuous re-assessment of the decision to stay in the city of Osijek and the County of Osijek-Baranja. Statistical data of the CBS for the Osijek-Baranja County (published in their first releases in the last ten years) show a negative net migration. Therefore, respondents' migration aspirations pertaining to emigration represent a serious threat to the future social and economical development of the County and the city of Osijek itself, as well as a serious warning and a call for the creation of high-quality regional/local policies. The problem is further emphasized by the fact that the people in question are young educated persons for whom empirical findings confirm that they are in favour of emigrating.

The reliability analysis confirmed a good internal consistency of factors, while the CFA confirmed that they describe real data well. Namely, Cronbach's alpha coefficients range from 0.78 to 0.85 for the LSF measurement scale, and from 0.76 to 0.85 for the LEF measurement scale, whereas average item correlations range from 0.48 to 0.60 for the LSF and from 0.45 to 0.53 for the LEF. Moreover, goodness-of-fit indices for individual factors have values greater than 0.9. Thereby, statistical measures lead to the conclusion that measurement scales could be reliable and that they could be used in other empirical studies. Indeed, the statistical procedure with respect to the psychometric properties of scales requires them to undergo repeated and further research of representative samples, which may be a new research topic. This research has also revealed other potential directions for future research. For example, it would be particularly interesting to compare the motives of university students with non-university educated young people, as well as with other age groups.

  1. 1

    Segmentation of the sample according to migration aspirations into two subsamples was done based upon the answer of examinees to the question: “Would you like to live permanently in some other place outside of Osijek?” with a binary answer offered: yes or no.

  2. 2

    Source of data referring to unemployment: CES, 2011; salaries: CBS, 2010, Table: 3.8.


  1. Top of page
  2. Abstract
  3. Introduction
  4. A Review Of Relevant Literature
  5. Research Of Migration Motives
  6. Research Results
  7. Discussion
  8. Conclusion
  9. References
  • Bearden, W.O., R.G. Netemeyer, and J.E. Teel 1989Measurement of Consumer Susceptibility Influence”, Journal of Consumer Research, 15, March: 473481.
  • Becker, G.S. 1962Investments in Human Capital: A Theoretical Analysis”, Journal of Political Economy, 70: 949.
  • Bodvarsson, O., and H. Van den Berg 2009 The Economics of Immigration: Theory and Policy, Springer Verlag, Berlin.
  • Borozan, Dj., I. Barkovic, and M. Dabic 2008The Attractiveness of the Place in a Region Lagging Behind”, International Journal of Business Research, 8(1): 915.
  • Borozan, Dj., and I. Barkovic Bojanic 2011Assessing the Motives Underlying the Decision to Migrate in a Region Lagging Behind”, Proceedings of the 3rd International Conference ECEE: Convergence, Opportunities and Challenges, 12–14 June 2011, Tallin, Estonia.
  • Bozic, S., and I. Buric 2005Croatia's Migration Potential – Micro-Analytic Aspects”, Migracijske i etnicke teme, 21(1-2): 933.
  • CBS (Croatian Bureau of Statistics) 2010 Statisticka izvjesca, No. 1419, Zagreb.
  • CBS (Croatian Bureau of Statistics) 2011 Priopcenje, No. 12.1.2, Zagreb.
  • Fassmann, H., and C. Hintermann 1998Potential East-West Migration: Demographic Structure, Motives and Intentions”, Czech Sociological Review, 6(1): 5972.
  • Fischer, P.A., R. Martin, and T. Straubhaar 1997Should I Stay or Should I Go?”, in T. Hammar, G. Brochmann and K. Tamas et al. (Eds), International Migration, Immobility and Development: Multidisciplinary Perspectives, Berg Publishers, Oxford: 4990.
  • Franklin, R.S. 2003 Migration of the Young, Single, and College Educated: 1995 to 2000, Census 2000 Special Reports, US Census Bureau, November, Washington.
  • Garip, F. 2008Social Capital and Migration: How Do Similar Resources Lead to Divergent Outcomes?Demography, 45(3): 591617.
  • Golub, B. 2006Socioprostorna pokretljivost hrvatskih znanstvenika”, Sociologija sela: casopis za istrazivanja prostornoga i sociokulturnog razvoja, 44(2/3): 261-288.
  • Gunderson, R.J., Pinto, J.V., and R.H. Williams 2007Assessing Domestic Migration at the County Level in the 4-Corners Region”, Northern Arizona University Working Paper series 07–06.
  • Hammar, T., and K. Tomas 1997Why Do People Go or Stay?”, in T. Hammar, G. Brochmann, K. Tamas and F. Faist (Eds.), International Migration, Immobility and Development: Multidisciplinary Perspectives, Berg Publishers, Oxford:120.
  • Harris, J.R., and M.P. Todaro 1970Migration, Unemployment and Development: A Two Sector Analysis”, American Economic Review, 60: 126142.
  • De Haas, H. 2007Migration and Development: A Theoretical Perspective”, COMCAD Arbeitspapiere – Working Papers, No. 29, Bielefeld.
  • Hoffmann, J. 2006 Working Migration from the Countries of Eastern and Central Europe: Present State and Perspectives. Report – Committee on Migration, Refugees and Population, Doc. 10842, 27 February
  • CES (Croatian Employment Service), Podrucna sluzba Osijek 2010 Godisnjak, Osijek.
  • Lee, E.S. 1966A Theory of Migration”, Demography, 3: 4757.
  • Mabogunje, A.L. 1970Systems Approach to a Theory of Rural-Urban Migration”, Geographical Analysis, 2: 118.
  • Massey, D.S. 1990Social Structure, Household Strategies, and the Cumulative Causation of Migration”, Population Index, 56(1): 326.
  • Massey, D.S., and M. Aysa 2005 Social Capital and International Migration from Latin America, UN/POP/EGM_MIG/2005/04.
  • Nunnally, J.C. 1979 Psychometric Theory, McGraw-Hill, New York.
  • OECD 2008 The Global Competition for Talent: Mobility of the Highly Skilled, July, Paris.
  • Ravenstein, E.G. 1885The Laws of Migration”, Journal of the Royal Statistical Society, 48: 167227.
  • Ravenstein, E.G. 1889The Laws of Migration”, Journal of the Royal Statistical Society, 52: 214301.
  • Roback, J. 1982Wages, Rents, and the Quality of Life”, Journal of Political Economy, 90, December:, 125: 71278.
  • Robinson, J.P., P.R. Shaver, and L.S. Wrigtsman 1991 Criteria for Scale Selection and Evaluation in Measure of Personality and Social Psychological Attitudes, Calif Academic Press, San Diego.
  • Sjaastad, A.H. 1962The Costs and Returns of Human Migration”, Journal of Political Economy, 70: 8093.
  • Stark, O. 1991 The Migration of Labour, Basil Blackwell, Cambridge.
  • Stark, O., and D. Bloom 1985The New Economics of Labour Migration”, American Economic Review, 75: 173178.
  • Sun, J. 2005Assessing Goodness of Fit in Confirmatory Factory Analysis”, Measurement and Evaluation in Counseling and Development, 37(4): 240256.
  • Sverko, I. 2005Studentske namjere odlaska u inozemstvo: velicina potencijalnog odljeva mozgova i njegove odrednice u 1995., 1997. i 2004. godini”, Drustvena istrazivanja, 14(6): 11401174.
  • Szczepanikova, A., Canek, M., and J. Grill (eds.) 2006 Migration Processes in Central and Eastern Europe: Unpacking the Diversity, Multikulturni centrum Praha, Praha.
  • Todaro, M.P. 1969A Model of Labor Migration and Urban Unemployment in Less-Developed Countries”, American Economic Review, 59: 13848.
  • Todaro, M.P. 1980Internal Migration in Developing Countries: A Survey”, in: R.A. Easterlin (Ed.), Population and Economic Change in Developing Countries, University of Chicago Press, London and Chicago, 361402.
  • Wikhall, M. 2002Culture as Regional Attraction: Migration Decisions of Highly Educated in a Swedish context”, European Regional Science Association, ERSA Conference papers, No. ersa02, Vienna: 318.