Attachment in cultural context

Differences in Attachment between Eastern and Western Europeans and the Role of Attachment Styles in Eastern European Migrants Adjustment

ISBN-978-90-367-3306-9

2007, Elbieta Polek

No part of this publication may be reproduced in any form, by print, photoprint, microfilm, or any other means without written permission from the author, except in the case of quotations in reviews, articles, books etc.

The present research was founded by the Ubbo-Emmius Foundation of the University Groningen

Printed by Grafex The author can be reached through email: elapolek@go2.pl

RIJKSUNIVERSITEIT GRONINGEN

Attachment in Cultural Context

Differences in Attachment between Eastern and Western Europeans and the Role of Attachment Styles in Eastern European Migrants Adjustment

Proefschrift

ter verkrijging van het doctoraat in de Gedrags- en Maatschappijwetenschappen aan de Rijksuniversiteit Groningen op gezag van de Rector Magnificus, dr. F. Zwarts, in het openbaar te verdedigen op donderdag 24 januari 2008 om 14.45 uur door Elbieta Polek geboren op 3 juli 1974 te Olsztyn, Polen

Promotores: Beoordelingscommissie:

Prof. dr. J.P.L.M. van Oudenhoven Prof. dr. J.M.F. ten Berge Prof. dr. B. de Raad Prof. dr. R. Goodwin Prof. dr. P.L.C. van Geert

Contents
Preface and acknowledgements Chapter 1: Introduction 1 5

Chapter 2: Evaluating the Replicability of the Bartholomew and Horowitz Model of Attachment in one Native Dutch and Four Immigrant Samples 15 Chapter 3: Attachment Styles and Demographic Factors as Predictors of Sociocultural and Psychological Adjustment of Eastern European Immigrants in The Netherlands 35 Chapter 4: Evidence for a Migrant Personality: Attachment Styles of Poles in Poland and Poles in The Netherlands 49 Chapter 5: A Cultural Look at Parenting and Attachment Styles in Eastern and Western Europe Chapter 6: Discussion and Implications Summary References Appendix

63 83 94 100 121

Preface and Acknowledgments The present research was conducted with the help of many people. I would like to thank some of them for their support. First and foremost, I am very grateful to my supervisors Professors Jan Pieter van Oudenhoven and Jos ten Berge for their great contribution to this dissertation and to my development as a researcher. When being supervised by them, I also had an opportunity to co-supervise the work of Joachim Whrle, whom I would like to thank for an inspiring exchange of ideas, his help in data-entering and his sense of humor. Our cooperation truly established a new era in PolishGerman relations. I wish to thank all the respondents who participated in my surveys, my colleague, Jacomijn Hofstra, for access to the data from the Dutch sample used in chapters 2 and 4, Veerle Brems, for collecting the data from the Dutch parents used in chapter 5, Naima El-Boundati, for collecting data from the Moroccan sample used in chapter 2, and Anna Szewczyk, for helping to collect the data from Polish parents used in chapter 5. Special thanks go to my Hungarian friends - Tams Br and Andrea Szentgyrgyi, for translating questionnaires and their help in the survey in Budapest. Katia Kalinina thank you for Russian translations and helping me in St. Petersburg. Also I would like to thank Pr. Sawomir Trypu and Bartomijej Mays, and Mrs. Magorzata Bos-Karczewska for their help in data collection from Polish immigrants in the Netherlands. Many thanks to my roommates Jacomijn, Ester, and Irene and other colleges from DPMG, especially Jorge, Rink, Arnaud, and Ilse; Hanny Baan, the secretary of our department who always swiftly handled my requests, and the ever kind and competent librarians, portieres and staff of the Repro, who printed the hundreds of questionnaires used in this project. I would like to thank my dear friends and colleagues who made my stay in Groningen memorable: Koen van Breackel, Yorgos Vleioras, Carmen Carmona, Huadong Yang, Anna Lichtwark, Laura Ballato, Simon Dalley, Istvn Back and my fantastic housemate Paco Guzman-Muoz. Thanks to Magda Goachowska and Kacper Wojtal, I didnt forget Polish language; Dimitry Balshoje spasiba za razgavory! Hartelijk dank to Pieter van Hoek for our wonderful talks about life and philosophy and for debriefing me about Dutch politics. Many thanks to Max Nuijens for supporting me when I really needed it.

I am thankful to Anne Boomsma, for being not only a knowledgeable professor, but also a wise person, and AIO-advisor Dr. Jose Heesink, for her help, as well as many other people for their support. Please forgive me for not mentioning all of your names individually. I am also obliged to Ela & Romek Mizera for their generous support, Leszek Woejko for a technical support, Luc Wasungu and Joanna Tadzik for all the good things I experienced from them. Finally, I would like to say thank you to Benjamin Cope, who did a proofreading of this thesis. Credit and thanks for the successful completion of my Ph.D. project I happily share with all these people. I dedicate this dissertation to my mom and dad, and Krzysztof, my wonderful brother.

Chapter 1 Introduction

When I arrived in the Netherlands to start my doctoral studies four years ago, emigration from Eastern Europe was a popular topic in public debate. Every day headlines in Dutch newspapers informed about the expected influx of immigrants after Eastern European countries joined the European Union. These articles expressed concerns related to socio-economic issues, but also, in an indirect way, a salient anxiety evoked by the approaching fusion with the unknown East. It was apparent that the topic of our research project was of great social importance. Later, experiencing myself the challenges of immigration and Eastern-Western differences, as well as hearing the stories told by my immigrant respondents, my conviction about the practical importance of our studies became even stronger. The migration research carried out hitherto was almost entirely focused on non-European immigrants, implicitly assuming that adjustment problems occur only when an immigrants culture of origin and the host culture are relatively distant. In the same vein, many cross-cultural studies focused on remote cultures, e.g., Western and Non-Western (African or Oriental), implicitly assuming that differences within European national cultures are negligible. Meanwhile, the statistics indicate that most of the cultural contacts in Europe occur between European national cultures (Statistics Netherlands, 2006). The present dissertation focused on Western and Eastern European cultures. What characteristics help Eastern Europeans to adjust to Western European society, and what are the similarities and differences in the individual characteristics of members of three Eastern European cultures and one Western European culture was the main theme of the present dissertation. Since all but the first and last chapter of this thesis are submitted journal papers, some overlap was unavoidable. Immigration from Eastern to Western Europe Until the late 80s social mobility between East and West Europe was very restricted due to the political division of Europe. The beginning of the 90s brought political changes that facilitated cultural and economic exchange, and increased population movements between East and West Europe. Throughout the communist era the reasons for emigration often were political persecution and emigrants desire to improve their economic status. Currently, it seems that it is economic motivations for emigration that dominate. Social and economic transitions in Eastern Europe created

INTRODUCTION

unemployment and difficulties with adaptation to market requirements among the working class, the rural population, and the poorly educated. These developments stimulated unskilled labourers to find work abroad. It is thus they who form the biggest group of immigrants from Eastern to Western Europe (Iglicka, 2000). In addition to economic motivations, a number of people migrate also with the aim of starting a new life with a partner of a foreign nationality. Recent statistics indicate a growing number of marriage immigrants from Eastern Europe in the Netherlands (Harmsen, 1999). During the 1990s Dutch immigration policy aimed at attracting knowledge immigrants (Doomernik, Penninx, & Van Amersfoort, 1997). This policy resulted in the immigration of well-educated professionals mostly medical personnel and computer specialists. Furthermore, European programs of student and university staff exchange also stimulated the emigration of this group from Eastern to Western Europe. Whether for work, love or study more and more people migrate within Europe. It has been estimated, for example, that about one and a half million Polish people left Poland for short or long-term emigration after Poland joined the European Union in May 2004 (Lipiski, 2006). In 2006, the estimated number of immigrants from Poland, Russia and Hungary living in the Netherlands amounts, respectively, to 34831, 35962, and 5736 (Statistics Netherlands, 2007). Freedom to travel, common economy, currency, legislation, and governmental institutions within the EU increased the population movement and the process of merging of national cultures. Statistics show that in last year about seven million people, which is 2% migrated within the EU, and this number will continue to increase in the future (Eurostat, 2007). It seems that we witness, nowadays, the formation of a European multinational society, consisting of individuals with diverse national backgrounds. Conceivably, cultural differences between European nations will diminish in the future. However, at present they still exist. Studying them may help us to better understand the nature of these differences and to develop acceptance for diversity. Differences between the four national cultures under study Although the four national cultures under study belong to the broader European culture, noticeable differences can nonetheless be found between theses cultures. Aside from the most apparent such as language differences (Dutch is a Germanic language; Polish is a Slavic language with a Latin alphabet; Russian is a Slavic language with a Cyrillic alphabet; and Hungarian is a Finno-Ugric language), there are differences forged by recent history. In the past the Netherlands was a part of the so called Western Bloc, where people could enjoy democracy and freedom of speech, whereas Poland, Hungary and Russia were parts of the Eastern (or Communist) Bloc, where democracy and freedom of speech were very limited. This

CHAPTER 1

division resulted in big economic differences; the Netherlands is listed as the 9th most affluent country in the world as measured by GDP per capita, whereas Hungary, Poland and The Russian Federation, although now economically blossoming, are much less wealthy: these countries are ranked respectively at 39th, 48th and 59th position in terms of GDP per capita (HDR, 2006). Accordingly, the welfare system in the Netherlands is far more developed giving Dutch people more existential security than the welfare systems in Eastern European countries. A better economy and a better welfare system contribute to a higher satisfaction with life and a higher general health reported in the Netherlands, as compared to Central and Eastern European countries. In the Netherlands Satisfaction with Life ranked 15th in the world, while in Poland, Hungary, and Russia 99th, 107th, and 167th, respectively (UN, 2006). Peoples perceptions of their health and level of control over their life was also found to be better in the Netherlands than in the three eastern European countries (Carlson, 1998). Also the percentage of the population with higher (university or college) education differs in the four countries: 18% in Dutch society; as compared to 7 % in Poland; 11 % in Russia and 9% in Hungary (CBS, 2006; HCSO, 2001; PCSO, 1996; CIA, 2007). Religiosity is lower in the Netherlands than in the three Eastern European countries: 41 % of Dutch people are not affiliated to any religion, 31 % are Roman Catholic, 20% are Protestant (Dutch Reformed or Calvinist), and 5.5 % are Muslim; in Poland, 90 % are Roman Catholics, 1.3 % Eastern Orthodox, and 0.3 % Protestants; in Hungary, 52% are Roman Catholic, 19% are Protestant, 3% - 14 % declare themselves non-believers; in Russia there are 15-20 % Russian Orthodox, 10-15 % Muslims, and 3040% are non-believers (CIA, 2007). Differences in religiosity and traditionalism are probably related to attitudes towards the family and children in the four countries. Statistics show, for example, that the average age of marriage is the eldest in the Netherlands (28 for women and 31 for men), a bit younger in Hungary (26 for women and 28 for men) and in Poland (for women 24 and for men 27), and the youngest in Russia (21.8 for women and 24.8 for men) (UNICEF, 2002; UNFPA, 1990). Also the average age for women to give birth to a first child is the eldest in the Netherlands 28.6 years old, followed by Hungary -25.1, Poland - 24.5, and in Russia 22.7 (NationMaster, 2003). The ratio of dissolved marriages differs in the four societies as well; in 2004 the Russian Federation had the highest divorce rate of the four countries with 4.42 divorces per 1000 residents, while Hungary, the Netherlands and Poland had 2.44, 1.91 and 1.48, respectively (United Nations Statistics Division, 2004). Differences are also found in policies promoting gender equality: the Netherlands, Poland, Hungary, and Russia rank respectively 7th, 30th, 41st and 62nd in the world on the Gender Empowerment Measure (HDR, 2006). Differences between the cultures

INTRODUCTION

under study were found in cross-cultural psychological studies. For example, Bond and colleagues found that Russian respondents scored higher (M=3.34, SD = 0.30), than Hungarian (M=2.98, SD = 0.41), and Dutch (M = 2.89, SD = 0.37) on dynamic externality, the social axiom referring to the general belief that effort, knowledge and careful planning will lead to positive results (Bond, Leung, Au, Tong, & Chemonges-Nielson, 2004). In the same study Russian respondents scored higher (M= 2.98, SD = 0.38) than Hungarian (M= 2.90 SD=0.47), and Dutch (M=2.59, SD=0.47) on cynicism the dimension referring to a negative view of human nature (a Polish sample was not included in this study). Hofstede (1980) found that respondents from the Netherlands, Poland, Hungary and Russia score differently on cultural dimensions: Power distance, Individualism-collectivism, Uncertainty avoidance, Masculinity-femininity, Long vs short-term orientation. Power distance was the lowest in the Netherlands (the calibrated position was 38), as compared to Hungary (46), Poland (68) and Russia (93). Individualism was equal in the Netherlands and Hungary (80), lower in Poland (60) and the lowest in Russia (39). Masculinity was the highest in Hungary (88), lower in Poland (64), Russia (36) and the Netherlands (14). Uncertainty avoidance was the highest in Russia (95), lower in Poland (93), and Hungary (82) and the lowest in the Netherlands (53). Long term orientation was the highest in Hungary, lower in the Netherlands and lower again in Poland (32) (no data available for Russia) (http:/www.geert-hofstede.com). In sum, the three Eastern European cultures seem to be more similar to each other, than each of them are to Dutch culture. Acculturation: Definition and Historical Perspective When culturally disparate people come into continuous direct contact with each other, the cultural differences between them tend to become salient and their original cultural patterns may undergo changes under the influence of the other culture (Bochner, 1982). This process is called acculturation. The term acculturation was introduced by American anthropologists as early as the 1880s (Jasinskaja-Lahti, 2000). Numerous definitions of acculturation have been presented in the literature, most of them adaptations of the definition proposed by Redfield, Linton, and Herskovits, 1936 (quoted from Arends-Tth, 2003): Acculturation comprehends those phenomena which result when groups of individuals having different cultures come into continuous firsthand contact, with subsequent changes in the original culture patterns of either or both groups. (p. 149)

CHAPTER 1

Although changes in both cultural groups are implied in this definition, in fact most changes occur in the non-dominant group as a result of influence from the dominant group. Consequently, migration research focuses on changes which migrants undergo, more than on changes which they induce in the host society. Acculturation was originally defined as a group-level phenomenon; later the term psychological acculturation (Graves, 1967) was proposed, to be replaced afterwards by the term adjustment. Facets of Immigrants Adjustment Leaving familiar surroundings and starting a new life in a foreign country is a major life-transition for an immigrant. It is easy to imagine that such a lifetransition might be problematic. Language barriers, cultural differences, the lack of a social network are all a challenge to immigrants adjustment. The term adjustment (or adaptation) is used in the present dissertation with reference to the outcome of the process, which occurs as a response to the change of environment or life situation. In recent migration literature a distinction has been drawn between sociocultural and psychological adjustment (e.g. Searle & Ward, 1990). Sociocultural adjustment is the process by which immigrants become a part of the new society, both as individuals and as a group. This process is related to changes in their identity, acquiring fluency in the host country language, increasing frequency of the use of the host country language, an increasing amount of contact with members of the host society and the satisfaction given by social interactions with the host society (Berry, 1992). It depends upon successful sociocultural adjustment, whether an immigrant is able to establish new social networks in the immigration country. Psychological adjustment refers to overall psychological well-being (or psychological health) and depends on strategies for coping with the stress associated with immigration (e.g. Ataca & Berry, 2002; Holmes & Rahe, 1967). Consistent with findings in general populations (Berkman, 1977), studies on immigrants show that social relationships are of major importance to psychological health and that psychological and sociocultural adjustment are correlated (e.g. Ataca & Berry, 2002). John Berrys Framework and its Critique Early research on sociocultural adjustment focused on immigrants orientation towards the host culture. It was believed that when immigrants acquire features and characteristics of the host culture they relinquish characteristics of their culture of origin (see Figure 1-1).

INTRODUCTION

10

Maintaining native culture

Adapting new culture

Figure 1-1. The unidimensional model of acculturation Recently, this unidimensional approach was replaced by a bidimensional approach in which maintaining the original and acquiring the new culture are perceived as independent, non-excluding processes. The most wildly used bidimensional framework in immigrant studies has been John Berrys (1992) framework. He proposed that immigrants have to deal with two major issues in their daily encounters: the extent to which they wish to maintain their own culture and have contacts with members of their native cultural group; and the extent to which they wish to have contacts with members of the host society. Treated as dichotomous dimensions, they can be used to define four acculturation orientations: integration, assimilation, separation and marginalization (Berry & Kim, 1988) (see Figure 1-2). When immigrants regard contact with members of the host culture as important, and at the same time wish to maintain their native culture, they are said to prefer the integration strategy. Assimilating immigrants are those who prefer contact with the new culture and give up their native culture, while separating immigrants maintain bonds with their native culture and reject the new culture. Finally, marginalization is characteristic of those who have no relation with either culture. A number of studies indicate that the majority of immigrants prefer integration (e.g. Berry, Kim, Power, & Young, 1989, Van Oudenhoven, Prins & Buunk, 1998). It is known that the integration strategy is related to the most adaptive forms of adjustment (Berry & Kim, 1988), whereas marginalization, when both cultures are rejected, is related to poor adjustment (Berry et al., 1989).

CHAPTER 1

11

Cultural adaptation
low high

Cultural maintenance

Separation
high

Integration

low

Marginalization

Assimilation

Figure 1-2. The bidimensional acculturation model of Berry (Arends-Tth, 2003) The four acculturation attitudes described above (e.g. Berry et al., 1989) are the basis of the fourfold measurement of acculturation attitudes. Although the fourfold measurement was used in numerous studies, it has met a lot of criticism, for poor psychometric qualities amongst others. For instance, double-barrelled items (e. g. I like both Dutch and Russian culture) used in a fourfold measurement are ambiguous, because we do not know if the respondents answer pertains to only the first part (e.g. Dutch culture), the second part (e.g. Russian culture) or both parts of the question (Dutch and Russian cultures) (Rudmin & Ahmadzadeh, 2001). Moreover, previous studies indicate that the acculturation strategy used by immigrants may vary in different domains of their lives. Some immigrants, for example, choose assimilation in their professional life, but prefer a separation strategy in private life (Arends-Tth & Van de Vijver, 2003). Therefore, it is often impossible to gauge the general (not domain-of-life specific) acculturation strategy of an individual. In Chapter 3, instead of assessing four acculturation strategies, we used a bidimensional measurement of sociocultural adjustment based on the two issues presented in Berrys model.

INTRODUCTION

12

Correlates of Immigrants Adjustment Research on acculturation focuses on two types of variables: variables related to immigrants individual characteristics, such as psychological and demographic characteristics and acculturation attitudes, and variables related to the acculturation context, such as characteristics of the host society and of the migrant group, as well as the relation between the migrant and majority group, and in particular the cultural distance between the host and migrants native cultures. In the present research we investigate the role of the first category of variables in immigrants adjustment. The relation between demographic characteristics and immigrants adjustment has been relatively well researched (for an overview see: Ward, Bochner, & Furnham, 2001). It has been found that the better the educational attainment of immigrants, the better is their sociocultural and psychological adjustment (Jayasuriya, Sang, & Fielding, 1992) and fluency in a host countrys language (Scott & Scott, 1989). Young age at emigration also relates positively to linguistic adjustment (Stevens, 1999; Jancz, 2000) and negatively to acculturation stress (Ha Kristic, 2000; Padilla, 1986). Length of residence proved to be related to a positive attitude towards the host culture (Cortes, Rogler, & Malgady, 1994) and to positive mental health (Ouarasse & Van de Vijver, 2005). While the role of demographic variables for immigrants adjustment has frequently been studied, far less attention has been given to personality factors, and to attachment styles in particular. In this research project we focused on attachment styles, as attachment is a promising, and up-till now rarely used, framework in immigration research (for exceptions see: Bakker, Van Oudenhoven, & Van Der Zee, 2004; Van Ecke, Chope, & Emmelkamp, 2005). The Bakker et al. study (2004), for instance, demonstrated that attachment styles were better predictors of immigrants psychological and social adjustment than Big Five personality traits. Moreover, a number of studies showed that attachment style is related to psychological adjustment and coping with problems (e.g. Cooper, Shaver, & Collins, 1998; Lapsley & Edgerton, 2002; Lopez, Mauricio, & Gormley, 2001) and affects regulation and problem coping (Lopez at al., 2001), attitudes towards the out-group (Mikulincer & Shaver, 2001) and own group members (Smith, Murphy, & Coats, 1999). Last, but not least attachment is a concept that attributes individual differences to environmental influences, in particular, parental practices, which, on the one hand, acknowledges the meaning of culture in individual development, and on the other, gives an optimistic view of the human condition: if cared for properly, we grow up to become open-minded and well-balanced human beings.

CHAPTER 1

13

The Foundations of Attachment Theory Attachment theory was developed by John Bowlby who described its foundations in a three-volume series, Attachment and Loss (1969/1982, 1973, 1980). In the first volume, Attachment, Bowlby argued that human beings evolved biologically to fear being alone, especially in potentially dangerous situations. This inborn tendency demonstrates that a person, especially during infancy and early childhood needs the protection and support of others, preferably others who have a special interest in the persons survival and well-being. An organized system of emotions and behaviours that enables the establishment of a close relationship with a protective caregiver an attachment relationship was according to Bowlby developed as an outcome of evolutionary selection; attachment to a protective caregiver increased the likelihood of survival. An attachment relationship serves as a safe base from which to explore the environment and to which to come back to when the exploration brings unsettlement. The first such relationship in most peoples life is the one with the mother, because in most societies mothers take care of their newborns. But an infants primary attachment figure might also be the father, a grandparent, an older sibling anyone who regularly plays the role of caregiver and comforter. Attachment theory uses the term primary caregiver to suggest that there is a hierarchy of attachment figures, with one being preferred in times of trouble. In his second volume, Separation: Anxiety and Anger, Bowlby explained the significance of the quality of early attachment relationships. First, he argued that if a person is confident that the attachment figure will be available whenever needed, the person will be much less prone to chronic fear, than will a person who for whatever reason does not have such confidence. In other words, self-confidence and freedom of anxiety are based on the quality of a persons present and past attachments. Second, Bowlby proposed that confidence in availability of attachment figures is built up during infancy and early childhood; whatever expectations of an attachment figure are developed during those years tend to persist relatively unchanged throughout the rest of life. Thus, attachment relationships affect personality development, and insecure and abusive attachments can cause lasting psychological problems. Early attachment relationships are conserved as mental representations which Bowlby and later attachment researchers call internal working models. When attempting to understand a persons feelings and behaviour in later love relationships, it is helpful to know what happened in earlier ones. In his third volume, Loss: sadness and Depression, Bowlby stated that the mind naturally recoils from the idea that there is no attachment figure to rely on. When a primary attachment figure departs or dies, a person often grasps at an alternative attachment figure, which helps to recover from the loss.

INTRODUCTION

14

Attachment theory assumes that all of us come into the world with a capacity to monitor attachment figures, to become anxious when left alone, and to make a fuss in order to assure proper care. Drawing on Bowlbys theory, Ainsworth and colleagues (1978) demonstrated how these universal abilities are expressed in different caregiving environments to produce important personality differences. In her landmark study - The Strange Situation experiment, she observed mother-infant dyads when infants encountered novelty toys, an unfamiliar adult experimenter, and a temporary separation from mothers. Ainsworth noted the differences between three major kinds of mother-infant dyads: secure, anxious-ambivalent, and avoidant. Infants form secure dyads kept up their exploration activities even when mother was temporarily out of sight and sought for proximity when mother returned. Infants from anxious-ambivalent dyads showed less exploration activities, and were not easily soothed on the return of the mother; finally, infants from avoidant dyads kept up exploration in the absence of mothers and seemed to actively avoid the mother after her return. In one of the earliest studies on attachment patterns continuity during infancy, Waters (1978) found that 96 percent of fifty infants from middle class homes exhibited the same attachment when assessed twice in the Strange Situation, once at twelve months of age and again at eighteen months. Studies of this kind have since been conducted with children of a wide variety of ages and with different methods of assessment. Most of these studies suggested that around 80 percent of infants whose environment is fairly stable show the same attachment pattern or style across time (Shaver & Hazan, 1994). This continuity supports Bowlbys idea that early attachment experiences shape our personalities. A number of attachment styles and measures have been proposed in the recent literature (see Cassidy & Shaver, 1999 for an overview). The present research was based on the Bartholomew and Horowitz (1991) model of attachment. The following chapter will give a more detailed description of this model and will examine its validity. Overview of the present dissertation In total, the data from 4089 respondents were analysed (743 immigrants in the Netherlands, and 3346 non-emigrant respondents) in the present dissertation. Immigrant respondents were approached in the places where they usually meet (clubs, parishes, etc.) and via immigrant organisations. We used the snowball method, and we also placed questionnaires on the internet (via internet we received in total 196 responses). Non-emigrant respondents were students of secondary schools and their parents in three countries. We

CHAPTER 1

15

approached them at schools. In Chapter 4, non-emigrant respondents were recruited in a survey in public places and via the snowball method. The main methods for data analysis used in this dissertation were factor analysis, multiple regression analysis, and multivariate analysis of covariance. Factor analysis was used to explore and identify the dimensions of individual differences (Chapter 2 and 5). Multiple regression analysis was used to test the relation between variables and to examine the exploratory power of certain variables with reference to other variables (Chapter 3). Multivariate analysis of covariance was used to examine differences in the mean scores on certain variables when controlling for the effect of other variables (Chapter 4 and 5) and to test the main and interaction effects of predicting variables on dependent variables (Chapter 5). This dissertation comprises four empirical studies described in Chapters 2-5. In Chapter 2 we address questions about the validity of the model of attachment styles and its measurements. In Chapter 3 we compare the predictive value of attachment styles and demographic factors for immigrants adjustment. In Chapter 4 we examine if individuals display predispositions for migration in terms of attachment styles and what is the role of attachment styles in the well-being of migrant and non-migrant respondents. Finally, in Chapter 5, we examine differences in attachment styles between the respondents from Eastern and Western Europe and the effect culture and parental characteristics have on attachment styles.

Chapter 2 Evaluating the Replicability of the Bartholomew and Horowitz Model of Attachment in one Native Dutch and Four Immigrant Samples

Introduction Attachment theory and measurement Attachment theory has spawned a large number of theoretical and empirical studies in developmental and social psychology. Originally formulated by Bowlby (1973, 1969/1982), attachment theory conceptualizes the universal human need to form affectional bonds with others. Within this theory, attachment is described as a behavioral system which is based on hereditary motivation (Bowlby, 1969/1982). It has emerged as a result of genetic selection, which favored attachment behavior because it increased the likelihood of child-mother proximity and, consequently, of offspring survival. The function of the attachment system is to provide children with a sense of security. Through interactions with caregivers the child develops internal working models, which include expectations and beliefs about the care and responsiveness of the caregivers, and also beliefs about whether the self is worthy of care and attention (Ainsworth, Blehar, Waters, & Wall, 1978). Thus, early in life an individual acquires a certain attachment style, or characteristic manner of developing bonds with others (Becker, Billings, Eveleth, & Gilbert, 1997). Attachment theory underlines that attachment is a lifespan phenomenon (Hazan & Zeifman, 1999), existing from the cradle to the grave (Bowlby, 1969/1982, 1973, 1980; Hazan & Zeifman, 1999, p. 336). Bowlby (1969/1982) argued that an infants relation with a caregiver is internalized as a working model and therefore generalized to new relationships. Bowlby also suggested that working models are the main source of continuity between the infants relationship with a caregiver and later relationships (Mikulincer & Nachshon, 1991). A number of adult attachment styles have been proposed. For example, Hazan and Shaver (1987), drawing on the work of Ainsworth et al. (1978), specified three attachment styles: a secure, an avoidant and an anxious/ambivalent style. Individuals with secure attachment are able to form close bonds with others and are comfortable with interdependent relationships, while individuals with an avoidant attachment are distrustful of others and afraid of intimate relationships. Anxious/ambivalent individuals

CHAPTER 2

17

want close relationships, but they suspect that other people do not truly care about them. The instrument used by Hazan and Shaver (1987) to assess attachment styles consists of three vignettes descriptions of the three attachment styles. Respondents indicate which of the three characterizes them best. Although the Hazan and Shaver instrument has been widely adopted by researchers (e.g., Baldwin, Fehr, Keedian, Seidel, & Thomson, 1993; Feeney & Noller, 1990; Mayseless, 1993; Mikulincer, Florian, & Tolmacz, 1990), it has also met abundant criticism. First, the Hazan and Shaver measure assumes that all individuals can be classified into one and only one category of attachment. As a result, the extent to which a chosen category characterizes respondents is ignored, as are potential individual differences in attachment profiles (Becker et al., 1997). These psychometric limitations of Hazan and Shavers instrument have led to the development of other measures. Based on Bowlbys claim that early attachment experiences are internalized as working models of self and others, Bartholomew and Horowitz (1991) proposed that the models of self and models of others can be either positive or negative. The combination of the model of self and others thus yields a four-category model of attachment styles (Figure 2-1). These are the secure style with positive models of both self and others, the fearful style with negative models of both self and others, the preoccupied style with a positive model of others and a negative model of self and the dismissing style with a positive model of self and a negative model of others. The structure of the Bartholomew and Horowitz model implies that the secure and fearful styles, as well as the preoccupied and dismissing are each others conceptual opposites (Figure 2-1). Drawing on the four-category model, Bartholomew and Horowitz (1991) measured attachment styles by means of vignettes, including descriptions of each attachment style (Table 2-1). Other researchers have separated the Bartholomew and Horowitz vignettes into individual items and derived scales to measure attachment styles with factor analysis methods (Collins, & Read, 1990; Simpson, Rholes, Nelligan, 1992). In some factor analysis studies, evidence was found for a four-factor structure corresponding to a four-category model (Kobak, Cole, Fleming, FerenzGillies, & Gamble, 1993; Hofstra, Van Oudenhoven, Buunk, 2005). Other studies, however, did not support a four-category model (Ng, Trusty, & Crawford, 2005; Brennan, Clark & Shaver, 1998; Kurdek, 2002).

THE REPLICABILITY OF THE BARTHOLOMEW AND HOROWITZ MODEL

18

positive model of others

PREOCCUPIED
negative model of self

SECURE
positive model of self

FEARFUL
negative model of others

DISMISSING

Figure 2-1. The four-category model of attachment styles proposed by Bartholomew and Horowitz (1991).

Table 2-1. Attachment Style Prototypes (Bartholomew & Horowitz, 1991). Secure It is relatively easy for me to become emotionally close to others. I am comfortable depending on others and having others depend on me. I dont worry about being alone or having others not accept me. I am comfortable without close emotional relationships. It is very important to me to feel independent and self-sufficient, and I prefer not to depend on others or have others depend on me. I want to be completely emotionally intimate with others, but I often find that others are reluctant to get as close as I would like. I am uncomfortable being without close relationships, but I sometimes worry that others dont value me as much as I value them. I am somewhat uncomfortable getting close to others. I want emotionally close relationships, but I find it difficult to trust others completely, or to depend on them. I sometimes worry that I will be hurt if I allow myself to become too close to others.

Dismissing

Preoccupied

Fearful

CHAPTER 2

19

Most studies on attachment measures report the structure obtained in one sample only, and do not examine the replicability of factors in a number of samples (e.g., Hofstra et al., 2005, Sibley, Fischer, Liu, 2005). Despite the advances in attachment measurement, the question as to whether or not the four-category model proposed by Bartholomew and Horowitz is supported by factor analysis is still open. The present study on the cross-sample replicability of factors obtained in a multi-item measure of attachment could provide an answer to this question. The Attachment Style Questionnaire (ASQ, see Appendix 1) by Van Oudenhoven, Hofstra and Bakker (2003) was originally developed in Dutch and used in samples of Dutch students and in a midlife sample. Authors of the ASQ started with an initial set of 35 items which were generated on the basis of the Bartholomew and Horowitz model, and probed both a specific type of attachment (quadrant) (e.g., I feel at ease in emotional relationships) and an underlying dimension (e.g., I presume that others are trustworthy). Based on the Principal Component Analysis in the Dutch sample they chose 23 items, which best represented four factors, and left out 12 items. In the present study we removed one more item, and used the 22item Attachment Style Questionnaire, see Table 2.3. The solution found in the Dutch sample corresponded to the Bartholomew and Horowitz model. Four oblique factors matching the four theoretical attachment types were found in the original version of the ASQ (Hofstra et al., 2005). The ASQ assesses attachment through multiple scores (on each dimension separately), allowing the description of individuals along each of the four attachment dimensions: secure, fearful, preoccupied and dismissing. Most attachment measures assess only relationship specific attachment. The ASQ is unique in that it measures general attachment to other people, thus allowing assessment of the general sociability of a respondent. Respondents give answers on a 5-point scale, ranging from 1 (strongly disagree) to 5 (strongly agree). First we examine whether or not the four factors obtained in the two Dutch samples are replicated in samples of Polish, Russian, Moroccan and Hungarian immigrants. If the four factors are replicated across the studied samples, it will provide support for the four attachment types postulated by the Bartholomew and Horowitz model. Second, we are interested if the correlations between factors support the structure of the model; this would imply that preoccupied attachment is the conceptual opposite of dismissive attachment and that secure attachment is the conceptual opposite of fearful attachment.

THE REPLICABILITY OF THE BARTHOLOMEW AND HOROWITZ MODEL

20

Method

Procedure A translation of the ASQ from the Dutch language into the Polish, Russian, and Hungarian target languages was made, and these were then translated back into Dutch. We compared the return translations with the original version in order to check the accuracy of the translations. We consulted Russian and Hungarian-born psychologists who have spent more than 15 years in the Netherlands and are fluent in the Dutch language, in order to eliminate the few discrepancies that emerged between the return translations and the Dutch version of the questionnaire. The final version of the questionnaires in Polish, Russian and Hungarian respectively, as well as the Dutch version, were presented once again to other translators who were asked to evaluate the equivalence of each translated item with the original version. We also asked translators to tell us if they would give the same responses to the Dutch items as to the corresponding translated items. The translators rated the equivalence of the original and translated items as high.

Samples We conducted the study on six samples: first year Dutch students and a Dutch midlife cohort (Van Oudenhoven et al., 2003) and four samples of immigrants living in the Netherlands: from Russia, Morocco, Poland and Hungary. The ASQ was administered to Polish, Russian and Hungarian first generation immigrants in their native languages. Moroccan first and secondgeneration immigrants received the ASQ in Dutch. Immigrants from Poland, Russia and Hungary were approached in places where they usually meet: communities, clubs, Sunday schools and parishes. Apart from the questionnaires and pre-paid envelopes, they were provided with a letter containing brief instructions and an explanation of the aim of the study. We asked participants to complete their questionnaires and also to distribute them among friends of the same nationality living in the Netherlands. Additionally, we sent questionnaires by e-mail to some of the participants through organizations in the Netherlands. Through the e-mail survey we obtained a total of 196 respondents, which accounted for 20, 5, and 18 percent of our Polish, Russian and Hungarian respondents respectively. The Moroccan participants were obtained using the snowball procedure. The Dutch first-year psychology students filled in the questionnaire as a course requirement. The Dutch midlife participants were obtained via a mail survey. Because there were no major differences in factor

CHAPTER 2

21

structure between the samples of the Dutch students and the midlife cohort, we merged them. Through the rest of the paper we will use the name Dutch sample to refer to the merged Dutch samples. Missing data were treated using variable mean replacement, and in this way a total of 84 missing data were replaced. Table 2-2 gives a brief description of the samples. Participants from immigrant samples and the Dutch midlife sample filled in the questionnaires voluntarily and without any monetary compensation. Table 2-2. Description of the Samples under study.
Male Female Age Years of residence in the Netherlands

SD

SD 7.58 5.12 6.77 15.52

Dutch 923 27 73 21.56 6.87 students Dutch midlife 175 30 70 50.10 4.70 cohort Polish 409 32 68 33.50 10.93 6.54 sample Russian 100 21 72* 36.71 11,93 6.20 sample Moroccan 111 42 58 26.50 9.32 20.25 sample Hungarian 116 76 24 36.51 14.23 13.79 sample *Seven respondents from Russian sample did not specify their gender

Analyses

Confirmatory analysis Multiple Group Method To assess the replicability of four factors across samples, a confirmatory method termed Multiple Group Method (MGM) was employed (Gorsuch, 1983; Nunnally, 1978; Nunnally & Bernstein, 1994). MGM is a method to test the researchers assumption that a given item represents a given factor. When a theory states that certain items will impact on the same factor and others on another factor, MGM is particularly appropriate. Items defining a particular factor are given weights of 1 (or -1 if the item is reversely keyed) and items defining other factors are given weights of zero to define the factor. The same is done for each of the other factors. In other words, in

THE REPLICABILITY OF THE BARTHOLOMEW AND HOROWITZ MODEL

22

MGM the factor is defined as the sum of the scores from a group of variables accepted as defining the factor with each variable given equal weight. For example, the standard scores from all the defining variables may be simply added together and this total score is considered the factor (Gorsuch, 1983, p.80). The correlation of an item with a factor which the item represents indicates how much the item contributes to this factor (Gorsuch, 1983, p. 88). However, the correlation of an item with the particular factor it is assigned to is spuriously high because the item correlates perfectly (r=1) with itself, and this self-correlation inflates the correlation between the item and any factor the item is assigned to. One way of countering self-correlation artifacts in MGM is to have large numbers of items per factor, so that the impact of self-correlation is negligible. In the ASQ the number of items per factor is small, so correction for self-correlation is imperative. The correlation between item and factor, corrected for self-correlation, is the correlation of the item with the score calculated by summing up the remaining variables assigned to the factor. As a requirement for correct assignment to a factor, one may demand that each variable correlates higher with the sum of the remaining variables in the factor than with any other factor (Gorsuch, 1983; Nunnally, 1978). Since an identical, binary weight matrix is used to define factors in all five samples, MGM can be seen as a form of cross-validation of component weights (Ten Berge, 1986). MGM does not rely on a formal statistical test, such as the chi-square test used in Maximum-Likelihood factor analysis. However, recent simulation studies (Stuive, Kiers, Timmerman, ten Berge, 2005) showed that in some cases MGM was more accurate than Confirmatory Factor Analysis (CFA) in indicating whether or not the suggested assignment of items to subscales is supported by the data. Hence, MGM although rarely used, is still an interesting alternative method to CFA in Lisrel.

Exploratory stage: Simultaneous Component Analysis In order to further examine the structure of the ASQ, we performed Simultaneous Component Analysis (SCA) (Kiers & ten Berge, 1989). In SCA, common weights are used to define factors in all the samples subjected to analysis and loadings are computed for each sample separately. The program used to perform SCA (Kiers, 1990) allows us to find optimal weights, such that the sum of the squared elements of residual matrices over samples is minimal. It is noteworthy that common weights are computed in this program based on correlation matrixes that are calculated for each sample separately. Thus, the problem of response tendencies in cultural samples is accounted for, as the calculations are performed on the data

CHAPTER 2

23

standardized within samples (Fischer, 2004; Van de Vijver & Leung, 1997). After computing the common weight matrix it was rotated by Varimax following the oblique procedure of the SCA program. This option tends to optimize simple structure for the weights matrix (Kiers, 1990). Imposing the constraints of common components is reasonable if the amount of variance accounted for by SCA factors does not differ much from the amount accounted for by the factors found in separate PCA. Regarding the questionnaire used in our study, we wanted to see if items show a consistent pattern of primary and secondary loadings across all five samples. SCA enables us to see how a given item correlates with a given factor in each sample, as factor loadings are calculated for each sample separately. This is not possible in PCA performed on the merged samples, because factor loadings are then calculated for the merged samples, not for each sample separately. Thus in PCA performed on merged samples we cannot see if a given item switches factors across samples, so we lose information about how the relation between an item and a factor differs across samples. In SCA a given item is assigned to a given factor across all samples, due to common weights, and thanks to this, factors have equal interpretations. This gives an advantage of SCA over Principal Component Analysis (PCA) performed on each sample separately, because the equal interpretation of the factors in SCA is assured across all samples included in the analysis.

Results

Confirmatory stage We performed computations for the MGM with a special option of the computer program Simultaneous Component Analysis (SCA) (Kiers, 1990). Corrected item-total-correlations were obtained in reliability analysis performed with SPSS. Items were assigned to the factors based on the original version of the ASQ. Primary loadings indicate that the secure, fearful and preoccupied factors proved to be relatively replicable across the five samples, while the dismissing factor was less replicable (Table 2-3). The four-category model of Bartholomew and Horowitz defines attachment style based on two underlying dimensions: the model of self and the model of others. Secure attachment is characterized by a positive model of self and others, while fearful attachment is characterized by a negative model of self and others. Thus, secure and fearful attachments are, to a large extent, opposite concepts, as are preoccupied and dismissing attachment

THE REPLICABILITY OF THE BARTHOLOMEW AND HOROWITZ MODEL

24

(Griffin & Bartholomew, 1994). Therefore, one can expect negative intercorrelations between the secure and fearful, and between the preoccupied and dismissing scales. One can also expect negative secondary loadings (i.e. second highest factor loadings) of items representing the secure scale on the fearful factor (and vice versa) and negative secondary loadings of items representing the preoccupied scale on the dismissing (and vice versa). As we can see in Table 2-3, the pattern of secondary loadings of items from the secure and fearful scales are consistent with the model; items representing the secure scale had negative secondary loading on the fearful scale, and items from the fearful scale had negative secondary loadings on the secure scale. However, we did not observe the expected pattern of secondary loadings for items from the preoccupied scale. Instead, most items from the preoccupied scale had their secondary loadings on the fearful rather than on the dismissing factor. The possible explanation for this may be that both fearful and preoccupied attachment styles are defined by an underlying negative model of self. Items from the dismissing scale, in their turn, did not show any clear pattern of secondary loadings in any of the three samples. Moreover, in the Russian sample the dismissing factor failed to emerge. Medium negative correlations between the secure and fearful scales (Table 2-4) in the five samples indicated that secure and fearful attachment styles are opposites as predicted by the model. The correlations between preoccupied and dismissing were negative and close to zero. These correlations and patterns of secondary loadings did not support the structure of the Bartholomew and Horowitz model that implies that preoccupied attachment is the conceptual opposite to dismissing attachment. Relatively high correlations between the secure and fearful factors entail secondary cross-loadings of items representing these scales. Although relatively high correlations between fearful and secure factors are conceptually legitimate, they become methodologically troublesome, because in some samples items assigned to the secure scale load higher on the fearful factor, and vice versa. Exploratory stage First, we compare variances explained in SCA and separate PCA for each sample, to see if imposing the constraint of common component does not cause a big decrease in the explained variance. The central and right-hand panels of Table 2-5 display the amounts of variance explained by the four factors obtained in SCA and PCA, respectively. The amount of variance explained by each factor separately is presented in the bottom row of Table 2-6. We remind the reader that the variances explained by oblique factors cannot be added up. The explained variances from Table 2-6 merely convey

CHAPTER 2

25

how much variance each factor would explain if it were the only factor. Table 2-5 shows that the explanatory power of the common components of SCA is only slightly inferior to that of separate PCA for all data sets. The largest loss of explanatory power was found in the Russian and Moroccan samples, 2.25% and 2.47%, respectively. Common components describe all the samples fairly well. Table 2-6 shows factor loadings, defined as correlations between variables and factors, in the five samples. The dismissing factor now proved to be more replicable as compared to the results of the MGM; only item 17 from the dismissing scale loaded on different factors in four samples.

Table 2-3. Factor Loadings for the ASQ items and Explanatory Power of Factors Across Five Samples (MGM)

S
Secure (S)
1. I feel at ease in emotional relationships 3. I feel uncomfortable when relationships with other people become close 7. I avoid close ties 9. I trust other people and I like it when other people can rely on me 12. I find it easy to get engaged in close relationships with other people 13. I feel at ease in intimate relationships 16. I think it is important that people can rely on each other 20. I trust that others will be there for me when I need them

Dutch sample F P D
-.27 .46 .39 -.37 -.38 -.31 -.11 -.31

S
.47
-.38

Polish sample F P D
-.27 .25 .22

Russian sample S F P D
.50
-.32 -.38 -.20

Moroccan sample S F P D
.18

Hungarian sample S F P D
.41
-.23 .30 .43 -.41 -.21 -.30 .20 .20

.54 -.57 -.61 .49 .51 .65 .36 .45

-.27 -.36 -.43 -.47

-.21 .25 .42

-.34 -.15 -.24 -.02 -.23 .02 .22 .20

.44
.44 -.39 -.20 -.31 -.09 -.37

.49
.33 -.24 .17 -.24 -.26 -.42 -.24 -.33 -.24

-.25
-.38

-.40 -.55 .46

.34

-.54 .52 .42 .52 .33 .42

-.37 .24 .28 .29 -.35 -.33 -.51

-.40 .47 .26 .44 .43 .35

-.36

.46 .43 .51 .30 .34

-.32 -.34 -.24

-.43
-.36 -.15

.66 .37
.43

-.50 .68 .60 .61 .60

.22 .36 .34 .29

Fearful (F)
2. I would like to be open to others but I feel that I can't trust other people 4. I would like to have close relationships with other people but I find it difficult to fully trust them 18. I am afraid that I will be deceived when I get too close with others 21. I am wary to get engaged in close relationships because I am afraid to get hurt

.50 .63 .59 .60

.25

.55 .60 .51 .51

.23 .29 .31 .25

.64 .53 .37

.50

.35 .32 .23 .38

.20 .21

.48 .60 .40 .28

.31 .27 .25 .29 .24

-.51 -.38 -.53 -.45

-.39 -.29

-.53
-.40

-.30

Preoccupied (P)
6. I often wonder whether people like me 8. I have the impression that usually I like others better than they like me 10. I am often afraid that other people don't like me 15. I don't worry whether people like me or not 19. I usually find other people more interesting than myself 22. It is important to me to know if others like me

.71 .51 .73 -.55

.18 -.22 -.26

.51
.41

.53

.35 .44 .51 .24 .45 .30

-.25

.41 .46 .60 .23 .30 .45 .19 -.24

.21 -.09 -.08 -.20 -.27 -.41 -.32

.62
.37

-.21 -.21

.33 .34

.46
.30

.47
-.23 .15 .39 .17 .34 -.15 .26

.44
.40

.55 -.31

.67 -.49 .37

.20

.23
.24 .15

.17 .14

.38 .52
-.14

.22

.36 .44

.19

.63 .34
-.14 -.14 -.22

Dismissing (D)
5. I prefer that others are independent of me and I am independent of them 11. It is important to me to be independent 14. I like to be self-sufficient 17. I don't worry about being alone: I don't need other people that strongly -.22 -.09 .10 -.23 -.10 -.26

.39 .45 .36 .33

2.06 3.78

.18 -.14 .10 .08

.35 .40 .36

.08

-.27

.32
-.16

.20 .42
.14

-.19
.09

.17

.22

-.18 -.21
3.81
.07

.25
1.65 4.11 4.11

.55 .47 .20

2.18

-.11
2.85

-.26
2.88 1.89 3.00 2.92

Explained variance

4.31

3.69

3.49

3.64

1.96

3.96

2.70

3.55

Table 2-3. Note. The loadings on the factors the items were assigned to were corrected for self-correlation. For each variable (the highest) primary loading in a sample is in bold type and the secondary (second highest) in bold italics. Only primary, secondary, and loadings equal or above .20 or below -.20 are displayed. Table 2-4. Intercorrelations of Scales in the Five Samples. obtained from MGM Dutch sample ASQ scales:
Secure (S) Fearful (F) Preoccupied (P) Dismissing (D) Secure (S) Fearful (F) Preoccupied (P) Dismissing (D)

Polish sample S
-.52 -.16
-.10

Russian sample S
-.53 -.15
-.16

Moroccan sample S
-.43 -.08
-.08

Hungarian sample S
-.59 -.23
-.12

S
-.49 -.13
-.14

F
.31
.10

P
-.18

F
.34
.18

P
-.14

F
.44
.22

P
-.10

F
.27
.16

P
-.23

F
.37
.10

P
-.16

obtained from SCA

-.52 -.04
-.05

.32
.09

-.14

-.47 -.04
-.01

.30
.16

-.16

-.47 -.04
.01

.30
.17

-.18

-.34 -.02
-.12

.24
.06

-.20

-.60 -.09
-.03

.36
.08

-.15

THE REPLICABILITY OF THE BARTHOLOMEW AND HOROWITZ MODEL

28

Intercorrelations between the secure and fearful factor ranged between .34 and -.60, and between the preoccupied and dismissing factor ranged between -.20 and -.14 (Table 2-4). Table 2-5. Variances Explained by Four Factors Jointly in MGM, SCA and PCA in the Dutch, Polish, Russian, Moroccan and Hungarian samples.
MGM Dutch sample Polish sample Russian sample Moroccan sample Hungarian sample Total Variance Accounted for 11.15 9.82 9.61 8.84 10.74 50.19 % 50.67 44.63 43.66 40.17 48.99 45.63 11.17 10.08 10.29 9.30 10.83 51.68 SCA % 50.78 45.83 46.78 42.29 49.22 46.98 11.32 10.25 10.79 9.85 11.20 53.41 PCA % 51.46 46.59 49.03 44.76 50.93 48.55

Note. The number of items in the questionnaire is 22. Similarly to MGM, intercorrelations between the secure and fearful scales, as well as the pattern of secondary loadings of items from the secure and fearful scales, were consistent with the Bartholomew and Horowitz model. Most items from the a priori preoccupied scale had their secondary loadings on the fearful factor. Low intercorrelation between the preoccupied and dismissing factor (Table 2-4) and the lack of secondary loadings of the items from the preoccupied scale on the dismissing factor (Table 2-6) indicate that, contrary to the assumption of the model, preoccupied attachment is not the conceptual opposite of dismissing attachment. Correlations between the preoccupied and fearful factor, which ranged between .24 and .36, and medium and positive secondary loadings of items from the preoccupied factor on the fearful factor, suggest that the preoccupied attachment and fearful attachment are positively rather than negatively related. Internal consistency We used SPSS-reliability procedure to compute the internal consistency of the scales. Items were assigned to the scales based on the SCA results. Alpha coefficients are displayed in Table 2-7. Item 17 was removed from the questionnaire.

Table 2-6. Factor Loadings for the ASQ Items and the Explanatory Power of Factors Across Five Samples (SCA Solution Obtained after Varimax Rotation of the Weight Matrix).

S
Secure (S)
1. I feel at ease in emotional relationships 3. I feel uncomfortable when relationships with other people become close 7. I avoid close ties 9. I trust other people and I like it when other people can rely on me 12. I find it easy to get engaged in close relationships with other people 13. I feel at ease in intimate relationships 16. I think it is important that people can rely on each other 20. I trust that others will be there for me when I need them

Dutch sample F P D
-.32 .54 .52 -.40 -.40 -.35 -.11 -.38

S
.62
-.44

Polish sample F P D
-.29

Russian sample S F P D
.68
-.38 -.36

Moroccan sample S F P D
.38
-.33 -.47 -.19

Hungarian sample S F P D
.51
-.45 -.24

.67 -.62 -.67 .64 .64 .78 .60 .59

-.24 -.32 -.40 -.46

.54
.52 -.40 -.22 -.29 .14 -.36 .20

.59
.46 -.23 .19 -.24 -.21 -.44 -.33 -.21

.40 .53
-.31 -.11 -.21 .22 -.26 .28

.46
.54 -.44 -.43 -.38 -.14 -.50 -.28

-.62 .66 .57 .70 .56 .57

-.32 -.28 .23 .25 -.29 -.51

-.49 .63 .48 .68 .62 .54

-.27 -.31

-.59 .62 .53 .79 .56 .60

-.49 -.33 .23 .21 -.51 -.48

.63 .62 .73 .58 .50

-.23 -.24

Fearful (F)
2. I would like to be open to others but I feel that I can't trust other people 4. I would like to have close relationships with other people but I find it difficult to fully trust them 18. I am afraid that I will be deceived when I get too close with others 21. I am wary to get engaged in close relationships because I am afraid to get hurt

.67 .77 .75 .73

.36 .51

.72 .77 .70 .66

.35 .23 .22

.78 .75 .56 .66 .67 .60

.20

.66 .79 .65

.79 .75 .75 .69 .29 .59

.49

.26 .27 .23 .20

-.26 -.53

.33

-.31

.50
.43 .45 .37

Preoccupied (P)
6. I often wonder whether people like me 8. I have the impression that usually I like others better than they like me 10. I am often afraid that other people don't like me 15. I don't worry whether people like me or not 19. I usually find other people more interesting than myself 22. It is important to me to know if others like me

.82 .57 .82 -.75

.11 -.28

.68 .49 .70 -.62

-.20 .08

-.35

.46 .51

.58 .53 .76 -.49 .40 .71 -.57

-.20 -.37

. -.38 -.30

.78
.47

.61
.41

-.21

.46

-.25

.47 .29 .45

.68 -.54 .53 .65

.23 -.27 -.20

.79 -.71
.15

.33 .37 .24

.30 .16

.46 .74 .65

-.12

.33

.47 .70
.27

.28 .27

.45 .81 .64

-.16 -.16

.18
.22

Dismissing (D)
5. I prefer that others are independent of me and I am independent of them 11. It is important to me to be independent 14. I like to be self-sufficient 17. I don't worry about being alone: I don't need other people that strongly -.23

.70
-.18 -.10

-.25

.41 -.18

.57 .87 .63

1.90
.20 .15

.23

.25

.15

.82 .72 .62

-.29

.81 .74
1.97

.20

.76 -.68 .51

-.24

.84 .79
2.07

.67
4.19

-.32

.68
2.58 3.60

-.36

.38
2.38 2.85

-.19

Explained variance

4.17

3.34

1.89

3.59

3.97

4.48

3.32

2.50

1.72

3.94

4.50

3.33

THE REPLICABILITY OF THE BARTHOLOMEW AND HOROWITZ MODEL

30

Table 2-6. Note. For each variable (the highest) primary loading in a sample is in bold type and the secondary (second highest) in bold italics. Only primary, secondary, and loadings equal or above .20 or below -.20 are displayed.

Table 2-7. Alpha Coefficients for SCA-derived Scales of the ASQ in the Dutch, Polish, Russian, Moroccan and Hungarian samples.
Dutch sample scale: Secure attachment (items: 1, 7, 9, 12, 13, 16, 20 ) scale: Fearful attachment (items: 2, 3, 4, 18, 21,) scale: Preoccupied attachment (items: 6, 8, 10, 15, 19, 22) scale: Dismissing attachment (items: 5, 11, 14) 0.78 0.79 0.81 0.58 Polish sample 0.74 0.76 0.70 0.63 Russian sample 0.68 0.75 0.62 0.49 Moroccan sample 0.65 0.64 0.64 0.41 Hungarian sample 0.71 0.77 0.78 0.65

Note: A minus sign preceding an item number means that the item has been keyed reversely Discussion Let us now address two central issues we put forward at the end of the introduction. 1. Is the factor structure obtained in the original version of the ASQ replicable in the four immigrant samples? The answer is: to some extent. We found four factors similar to the original version of the ASQ across the five samples. However, a number of items did not load on the hypothesized factor. Thus, the replicability was at best fair. In general, the SCA factor structure obtained in the five samples was more stable than that of the MGM. This is not surprising, since SCA is designed to find a more stable structure across analyzed samples. In both MGM and SCA we observed that items switched factors. In particular, items from the a priori secure scale switched to the fearful scale and vice versa. This was due to intercorrelations of these scales. In the next three studies described in this book we used the scales obtained in the SCA. Bartholomew and Horowitz (1991) proposed that denial of the need of others, characteristic of dismissing individuals, is expressed by an excessive self-sufficiency and an avoidance of relationships with others. Accordingly, two types of items measuring excessive self-sufficiency and measuring avoidance of relationships were included in the dismissing scale. However, the items of the two types yield one factor only in the Dutch sample. Remarkably, in none of the immigrants samples items probing the avoidance of relationships loaded on the dismissing factor. This indicates

CHAPTER 2

31

that dismissing attachments of immigrants surface mainly through selfsufficient behavior. A first possible explanation is that the lack of a familiar social network imposes a high degree of self-sufficiency on immigrants, which is not really connected to an avoidance of closeness, as is the case in the native Dutch sample. In another study (Polek, Van Oudenhoven, Ten Berge, 2007) a positive relation was found between length of residence and the dismissing attachment of immigrants, which means that the circumstances of immigrant life may strengthen dismissing behavior. A second possible explanation is that self-sufficiency is unrelated to immigrants avoidance of closeness. Immigrants from collectivistic cultures may adapt to the more individualistic Dutch culture and become more selfsufficient, but not more closeness-avoidant. Third, it is possible that immigrants are simply a pre-selected sample of individuals with high selfsufficiency and an independence that caused them to immigrate in the first place. The fourth option is that avoidance of closeness and self-sufficiency might be related, but still empirically clearly distinguishable concepts. Such an explanation is also suggested by the results of Duggan and Brennan (1994) who found two distinguishable factors: fearful avoidance related to shyness (social inhibitions, anxiety) and dismissing avoidance (low sociability and affiliation needs). As the dismissing factor comprising of items expressing excessive self-sufficiency in the present study was negatively correlated to secure, and almost uncorrelated with a fearful attachment, it seems that self-sufficiency is more related to dismissing avoidance and low affiliation needs, than to fearful avoidance. 2. Do we find support for the four attachment types proposed by the model of Bartholomew and Horowitz and the structure of the model? In the present study, we found four factors, which clearly corresponded to the attachment types (quadrants) postulated by the model. However, we found only partial support for the structure of the model. The structure of the Bartholomew and Horowitz model implies that secure and fearful - as well as preoccupied and dismissing - attachment styles should be considered as conceptual opposites. The correlations between factors obtained in SCA, as well as the patterns of secondary loadings, supported this assumption only for the secure and fearful attachment styles. Similarly to the findings of Schmitt et al. (2004), in the present study the correlations between preoccupied and dismissing factors did not support the opposed relation between preoccupied and dismissing attachments styles presumed by the model. Rather, the low correlations suggest that the preoccupied and dismissing attachment should be considered as independent. Our results are consistent with recent models of attachment system dynamics which operationalize anxiety (related to preoccupied attachment) and avoidance

THE REPLICABILITY OF THE BARTHOLOMEW AND HOROWITZ MODEL

32

(related to dismissing attachment) as independent dimensions (Mikulincer & Shaver, 2003). The model also assumes that preoccupied attachment has one underlying dimension a positive model of others in common with secure attachment. Therefore, one should expect that preoccupied attachment is positively correlated with secure attachment and, consequently, that items from the preoccupied factor would have positive secondary loadings on the secure factor, especially when the secure scale consists of items testing the model of others (as in the ASQ). We did not observe this. Correlations between preoccupied and secure attachment were close to zero in SCA, whereas between preoccupied and fearful they were positive and medium. Subsequently, items from the preoccupied scale had positive secondary loadings on the fearful factor, and not, as the model would predict, on the secure factor. This suggests that preoccupied attachment, similarly to fearful attachment, is characterized by a negative model of self and also that both concepts relate to behavior earlier classified by Hazan and Shaver (1987) as anxious-ambivalent. Our findings are consistent with earlier studies in which a positive correlation was found between preoccupied attachment measured with ASQ and distrust (Hofstra et al., 2005). However, we are not in favor of merging the fearful and preoccupied factors. Differentiating between fearful and preoccupied attachment allows for a wide range of correlations with third variables, e.g., fearful attachment is more strongly negatively related to the psychological and sociocultural adjustment of immigrants than preoccupied attachment (Polek et al., 2007). There are several limitations to this study. First and foremost, we conducted the study in a sample of native Dutch and samples of immigrants. Therefore, similarities in factor structure may be due to contact with Dutch culture. On the other hand, the differences we found cannot be unequivocally attributed to cultural diversity. Possibly, differences in factor structure are due to a situation specific to immigrants, as we found in the dismissing scale. To clarify this it is necessary to conduct a follow-up study on samples of respondents living in their country of origin. Second, the initial set of items in the original version of the ASQ was not large enough to guarantee that items represent attachment prototypes exhaustively. In most studies on attachment assessment, the initial set of items was much larger (e.g., Sibley, Fischer, Lin, 2005). Still another problem is the presence of double-barreled items. It is not certain what a respondent will answer if he or she agrees with only one part of the item and not with the other. In addition, several items in the initial set of items contained statements with negations, which may have confused respondents and therefore produce conceptually irrelevant variance caused by different

CHAPTER 2

33

language abilities of respondents (Holden & Fekken, 1990) or the ambiguity of an item (Angleitner, John, & Lohr, 1986). Third, the snowball recruitment approach may have meant that individuals with certain (insecure) attachment styles may have been underrepresented. In particular, those with a dismissing style may be less interested in participating or less connected to other people, and therefore less likely to be recruited. The above limitations notwithstanding, the methods used in this study represent a more rigorous test of psychometric properties than has been employed in most previous studies of attachment instruments (Griffin & Bartholomew, 1994; Hazan & Shaver, 1987). The ASQ appears to be original in so far as it combines theory driven attachment prototypes with an empirically based dimensional approach and that it allows the assessment of general, non relationship-specific attachment.

Chapter 3 Attachment Styles and Demographic Factors as Predictors of Sociocultural and Psychological Adjustment of Eastern European Immigrants in The Netherlands

Introduction The admission of Eastern European countries into the European Union (EU) in 2004 ignited much debate about increased immigration from Eastern to Western Europe. Observers expected a massive influx of so-called economic migrants from former Eastern bloc countries. This expectation was confirmed by recent data, which show that Polish people became the largest group of immigrants currently arriving in the Netherlands (Statistics Netherlands, 2006). In view of another possible immigration wave from Bulgaria and Rumania, countries recently admitted to the EU, it becomes vitally important to examine factors influencing psychological and sociocultural adjustment of Eastern European immigrants in order to plan effective immigration polices. Many studies have investigated the role of demographic factors such as age at immigration, length of residence, and education for immigrants adjustment. Age has been reported to be related to acculturation outcomes (Stevens, 1999); length of residence to a positive attitude towards the host culture (Cortes, Rogler, & Malgady, 1994) and mental health (Ouarasse & Van De Vijver, 2005); and educational level to sociocultural and psychological adaptation (Jayasuriya, Sang, & Fielding, 1992), and acquisition of a host countrys language (Scott & Scott, 1989). While the role of demographic factors for immigrants adjustment has been relatively well researched (see Ward, Bochner, & Furnham, 2001, for an overview), little attention has been given to personality factors, and even less to attachment styles, in particular. In the present study we focused on attachment styles, as attachment is a promising, and up to now rarely used framework in immigration research (see, for exceptions, Bakker, Van Oudenhoven, & Van Der Zee, 2004; Van Ecke, Chope, & Emmelkamp, 2005). Attachment styles are stable tendencies in the way people relate to others (Cassidy & Shaver, 1999), which influence aspects of social functioning such as quality of relationships (Hazan & Shaver, 1987), attitude towards out-group (Mikulincer & Shaver, 2001) and in-group members

ATTACHMENT AND ADJUSTMENT OF IMMIGRANTS

36

(Smith, Murphy, & Coats, 1999), social competencies (Mallinckrodt, 2000), the way of approaching unfamiliar others (Ainsworth, Blehar, Waters, & Wall, 1978), psychological adjustment and problem coping (Cooper, Shaver, & Collins, 1998), and attitudes towards acculturation (Hofstra, Van Oudenhoven, & Buunk, 2005; Van Oudenhoven & Hofstra, 2006). Our review of the literature led us to expect that attachment styles are likely to influence immigrants capacity to deal with the challenges of immigrant life and, consequently, to have an effect on their psychological and sociocultural adjustment. A number of attachment styles have been proposed in the literature (see Cassidy & Shaver, 1999 for an overview). The present study was based on the Bartholomew and Horowitz (1991) model, in which they distinguished four attachment styles: the secure style, characterized by trusting oneself and others; the fearful style, characterized by lack of trust in oneself and in others; the preoccupied style, characterized by a desire for close relationships in order to gain acceptance of others and, at the same time, fear of being rejected; and the dismissing style, characterized by trust in oneself, avoidance of relationships, and excessive self-sufficiency. The model was empirically validated and used as a framework in adult attachment research (Griffin & Bartholomew, 1994). Using measures of attachment styles proposed by Bartholomew and Horowitz (1991), Handojo (2000) demonstrated that attachment styles are related to acculturation attitudes. Van Ecke et al. (2005) found evidence for a relation between insecure attachment and distress among immigrants. Our first goal was to clarify the relation further between immigrants psychological and sociocultural adjustment and attachment styles. In addition, we examined the relation between demographic factors and psychological and sociocultural adjustment. The second goal of this study was to evaluate the usefulness of the attachment styles versus the demographic measures as predictors of immigrants adjustment. We now turn to our predictions concerning the relation between attachment and adjustment. Secure attachment was found to be associated positively with social competences and the belief that others are trustworthy (Bartholomew & Horowitz, 1991). These characteristics help to establish positive relationships with members of the host and native cultures. Therefore, we expect secure attachment to be positively related to psychological adjustment as well as identification and contact with the Dutch and native cultures [Hypothesis 1]. By contrast, fearful attachment is marked by interpersonal problems, social inhibition, and lack of assertiveness (Bartholomew & Horowitz, 1991). Social inhibition will be least strong among members of ones own cultural group. Consequently, we expect fearful attachment to be negatively related to

CHAPTER 3

37

psychological adjustment of immigrants and identification and contact with the Dutch culture, but positively related to identification and contact with the native culture [Hypothesis 2]. Preoccupied attachment is characterized by the desire to gain acceptance of others and fear of being rejected by others. This ambivalence may lead to rejection of other people, as a strategy of prevention of being ultimately rejected by others. Van Oudenhoven and Hofstra (2006) found, for example, that preoccupied attachment was positively associated with immigrants approval of separation from the mainstream society. Thus, we expect preoccupied attachment to be negatively related to psychological adjustment and to identification and contact with the Dutch culture [Hypothesis 3]. Dismissing attachment is characterized by a lack of trust in others (Bartholomew & Horowitz, 1991) and the use of a defensive strategy for affect regulation, such as denial of experiencing emotions (Onishi, Gjerde, & Block, 2001). Accordingly, we expect dismissing attachment to be unrelated to psychological adjustment and to be negatively related to sociocultural adjustment [Hypothesis 4]. As we saw above, education, young age at arrival, and length of residence were found to be associated with better psychological adjustment. Hence, we expect that age at immigration is negatively correlated with psychological and sociocultural adjustment, whereas education and length of residence are positively related to psychological and sociocultural adjustment [Hypothesis 5].
Method

Participants Four hundred and eight immigrants from Poland (68% female), 100 from Russia, (72% female), and 123 from Hungary (76 % female) participated in our survey. The mean age of respondents in years was 33.50 (SD = 10.93) in the Polish sample, 36.71 (SD = 11.93) in the Russian sample, and 36.51 (SD = 14.23) in the Hungarian sample. From the total number of respondents, 196 participated in our survey via the Internet, which accounted for 20%, 5%, and 18% of respondents in the Polish, Russian, and Hungarian samples, respectively. The average length of stay in the Netherlands, education, and age at immigration for all samples are reported in Table 3-1. Procedure A translation of the questionnaires from Dutch into Polish, Russian, and Hungarian was checked and revised by native Polish, Russian, and Hungarian psychologists who had spent more than 15 years in the Netherlands and were fluent in Dutch. This version of the questionnaires, as well as the Dutch version, was presented once again to other Polish, Russian, and Hungarian

ATTACHMENT AND ADJUSTMENT OF IMMIGRANTS

38

translators, who were asked to evaluate the equivalence of each translated item with the original version. The equivalence was rated as high. Polish, Russian and Hungarian immigrants in the Netherlands were approached in places where they meet: communities, clubs, Sunday schools, and parishes. Respondents were given the questionnaire, a letter with short instructions, and a pre-paid reply envelope. They were asked to complete the questionnaire and to distribute additional questionnaires among friends of the same nationality living in the Netherlands. We also placed questionnaires on the Internet, and asked immigrants organizations in the Netherlands to email our request for participation to their members. All letters, e-mails, and questionnaires were in the respondents native language. Since, in the preliminary check, the data obtained in Internet survey and ordinary mail did not differ on mean scores and respondents characteristics, we pooled the data from traditional and Internet surveys and analysed them jointly for each national sample. Respondents filled in questionnaires voluntarily and without monetary compensation. Missing data were treated using variable mean replacement. Missing data accounted for 1%, 7%, and 1% of the data in the Polish, Russian, and Hungarian samples, respectively. Instruments The questionnaire started with demographic questions about age, age at emigration, education, and length of stay in the Netherlands. Psychological adjustment was measured with Psychological Health, a 9-item scale from the RAND 36-item Health Survey (RAND Health Sciences Program, 1992; How much of the time during the past four weeks have you been a very nervous person?) and a 5-item Satisfaction With Life Scale (SWLS) by Diener, Emmons, Larsen, and Griffin (1985; sample item: I am satisfied with my life). Both measures had a 5-point answering scale ranging from strongly disagree (1) to strongly agree (5). Alpha coefficients of the scale of Psychological Health were .87, .91, and .87 and of Satisfaction With Life Scale were .80, .83, and .78 in the Polish, Russian, and Hungarian samples, respectively. Sociocultural adjustment was measured with a shortened version of the Social Support List Interactions (SSL-I) by Van Sonderen (1993). The 9-item SSL-I used in the present study was derived from a factor analysis on the original list of 64 items (Van Oudenhoven & Van der Zee, 2002). A sample item was: Does it ever happen to you that people are affectionate towards you?. Respondents gave answers on a 4-point scale ranging from seldom or never (1) to very often (4). Alpha coefficients were .84, .88, and .83 in the Polish, Russian, and Hungarian samples, respectively.

CHAPTER 3

39

Two measures of immigrants orientation towards the host as well as towards the native culture were used. Initially, on the basis of a number of unstructured interviews, 37 items concerning identification and contact with the native culture and 42 items concerning identification and contact with the Dutch culture were generated. Respondents gave answers on a 5-point scale: from strongly disagree (or never) (1) to strongly agree (or very often) (5). The answers from the three samples were subjected to Simultaneous Component Analysis with an oblique rotation (SCA; Kiers & Ten Berge, 1989). After excluding items loading on both factors, and items with loadings below .20, 16 items remained and these were subjected to SCA once again. We obtained two factors with 8 items loading on the Dutch, and 8 items loading on the native factor. Correlations between these two factors were .05, .14, and .08 in the Polish, Russian, and Hungarian samples, respectively. Alpha coefficients of the scale Identification and Contact with the Dutch Culture were .78, .74, and .67 in the Polish, Russian, and Hungarian samples, respectively, and the scale consisted of items like: I feel a member of the Dutch society. Alpha coefficients of the scale Identification and Contact with the Native Culture were .80, .80, and .74 in the Polish, Russian, and Hungarian samples, respectively, and the scale consisted of items like: I am proud of being Polish (Russian / Hungarian). Attachment was measured with the Attachment Style Questionnaire (ASQ; Van Oudenhoven, Hofstra, & Bakker, 2003). Four oblique factors matching the attachment styles postulated by the Bartholomew and Horowitz (1991) model were found in the original Dutch version of the ASQ (Hofstra et al., 2005). In another study the replicability of the ASQ scales across the Polish, Russian, Hungarian, and Moroccan immigrant samples, and two native Dutch samples, proved satisfactory (Polek, Ten Berge, & Van Oudenhoven, 2006). In the study of Hofstra et al. the construct validity and the stability of the ASQ, measured after one year with Pearson correlation, also proved to be satisfactory: .63 for the secure style, .60 for the fearful style, .69 for the preoccupied style, and .63 for the dismissing style. In contrast to many existing attachment measures that focus only on relationship-specific attachment, the ASQ measures general attachment, which makes it possible to assess the general sociability of a respondent. The ASQ assesses attachment through multiple scores (on each dimension separately). Thus, participants were not classified into one attachment category, but received scores on all four attachment scales. For all ASQ scales, a 5-point scale was used, ranging from strongly disagree (1) to strongly agree (5). An example of an item from the 8-item secure scale was: I feel at ease in emotional relationships; from the 4-item fearful scale: I am afraid that I will be deceived when I get too close with others; from the 6-item preoccupied scale: I often wonder whether people like me; from

ATTACHMENT AND ADJUSTMENT OF IMMIGRANTS

40

the 3-item dismissing scale: It is important to me to be independent. The four attachment styles are not independent; as theory would predict, secure style was correlated negatively with three other styles (see Table 1). Alpha coefficients of the secure scale were .71, .67, and .73; of the fearful scale .74, .71, and .80; of the preoccupied scales .70, .66, and .78; and of the dismissing scale .63, .49, and .65 in the Polish, Russian, and Hungarian samples, respectively. Results Attachment and Psychological and Sociocultural Adjustment The first four hypotheses focused on the relation between attachment styles and psychological and sociocultural adjustment. Accordingly, we calculated Pearson correlations between attachment styles and indicators of psychological and sociocultural adjustment. Secure attachment correlated positively with satisfaction with life, psychological health, social support, and identification and contact with the Dutch culture across all three samples (see Table 3-1). Identification and Contact with the native culture correlated with secure attachment in the Polish and Russian samples. Thus, our first hypothesis was almost entirely supported. Fearful attachment proved to be negatively correlated with psychological health, satisfaction with life, social support, and identification and contact with the Dutch culture across all three samples. Our expectation of a positive relation between fearful attachment and identification and contact with the native culture was confirmed in the Hungarian sample only (see Table 3-1). Hence, the second hypothesis was supported to a large extent.

Table 3-1. Means, Standard Deviations, and Pearson Correlations between the ASQ and Sociocultural and Psychological Adjustment.
1. Secure attachment
Sample: Polish Russian Hungarian Polish Russian Hungarian Polish Russian Hungarian Polish Russian Hungarian Polish Russian Hungarian Polish Russian Hungarian Polish Russian Hungarian Polish Russian Hungarian Polish Russian Hungarian M 3.94 3.85 4.09 2.63 2.49 2.23 2.73 2.80 2.89 4.12 3.94 3.75 3.47 3.41 3.62 4.13 3.78 4.08 3.16 3.24 3.64 3.31 3.18 3.34 2.60 2.65 2.55 SD 0.54 0..57 0.60 0.82 0.78 0.96 0.68 0.96 0.82 0.70 0.66 0.65 0.70 0.63 0.63 0.53 0.58 0.49 0.50 0.63 0.53 0.84 0.89 0.75 0.51 0.55 0.50 2 -.52* -.53* -.59* 3 -.16* -.15* -.23* -.34* -.44* -.37* 4 -.12* -.11 .34* -.17* -.26* -.13* -.11* -.01* -.32** 5 -.25* -.45* -.25* -.26* -.30* -.23* -.09* -.05 -.15* -.06 -.05 .17 6 -.22* -.21* -.10* -.03* -.02* -.22* -.05* -.05* -.21* -.05* -.25** .10 -.03* -.08* -.07 7 -.32* -.32* -.43* -.34* -.49* -.40* -.28* -.41* -.32* -.06* -.15 -.03 .19* .33* .28* -.01* -.05* -.12 8 .28* .35* .25* -.40* -.51* -.34* -.21* -.41* -.31* -.05* -.22** -.01 .26* .38* .30* -.08 .15* -.08 .45* .62* .48* 9 .45* .53* .45* -.29* -.41* -.43* -.06 -.18 -.17 -.13* -.19 -.14 .17* .34* .09 .20* .17* .02 .18* .17* .37* .28* .37* .25* 10 .11* .14 .02 -.22* -.25* .02 -.12* -.01 .10 -.15* -.01* -.10 .19* -.01* -.12 -.09* .02* -.15 .15* .03* .04 .20* .01* .03 .09* -.09* .08 11 .08* -.01* -.08* -.07* -.05* -.05* -.07* .00* -.29* -.05* .01* .13 .41* .05* .32* -.07* .10* -.02 .04* -.26* .08 .07* -.08* .26* .01* -.14* -.21* 12 -.05* -.11* -.05* .01* .16* .02* -.03* .10* .02* .03 .10* -.02 -.02* -.24* -.19* .03* .09* .01 .02* -.19* -.05 -.07* -.26* -.03 -.08* -.01* -.05

2. Fearful attachment

3. Preoccupied attachment

4. Dismissing attachment 5. Identification and contact with the Dutch culture 6. Identification and contact with the native culture 7. Psychological health

8. Satisfaction with life

9. Social support

10. Education

Polish Russian

4.35 4.44

0.83 1.03

.12* .15*

.18* .26*

Hungarian

4.61 6.54 6.20 13.79 26.57 30.19 24.08

0.80 7.58 5.12 15.52 8.13 10.47 6.17

-.36*

.05 .03* .09* -.20*

11. Length of residence

Polish Russian Hungarian Polish Russian Hungarian

12. Age at immigration

Note. For education 1 means primary school, 2 -some secondary, 3-vocational, 4-secondary, 5-BS, BA or higher degree. *p < .01.

CHAPTER 3

43

Preoccupied attachment indeed correlated negatively with psychological health and satisfaction with life across the three samples, which was in accordance with the third hypothesis. However, contrary to our expectations, preoccupied attachment was unrelated to identification and contact with the Dutch culture and other indicators of sociocultural adjustment. The exception was the Hungarian sample, in which preoccupied attachment was positively correlated with identification and contact with the native culture. Dismissing attachment was unrelated to psychological health in all three samples and negatively related to satisfaction with life in the Russian sample. There was a negative and weak relation between dismissing attachment and social support in the Polish sample, and identification and contact with a native culture in the Russian sample. The fourth hypothesis was supported by the data to some extent. Psychological and Sociocultural Adjustment and Demographic Factors Education proved to be positively correlated with identification and contact with the Dutch culture and with psychological health and satisfaction with life in the Polish sample only. In none of the three samples did we find a relation between education and identification and contact with the native culture. As expected, length of residence was positively related to identification and contact with the Dutch culture in the Polish and Hungarian samples. Contrary to predictions, it was unrelated to psychological health in the Polish and Hungarian samples, and negatively related in the Russian sample (Table 3-1). Length of residence was positively related to satisfaction with life only in the Hungarian sample and, surprisingly, it was negatively correlated with social support in that sample. Age at immigration was, as predicted, negatively correlated with psychological health, satisfaction with life, and identification and contact with the Dutch culture, which means that arriving at a young age is favourable for identification and contact with the Dutch culture, and consequently beneficial for psychological adjustment. Altogether, the predictions concerning the relation between demographic factors and adjustment were partially supported. The Predictive Value of Attachment and Demographic Factors In addition we examined the predictive values of attachment styles and demographic factors and if attachment styles explained more variance in psychological and sociocultural adjustment beyond that explained by demographic factors. For that purpose we carried out two regression analyses. In first regression we entered attachment styles alone as predictors (see Table 3-2). Next, we carried out hierarchical regression analysis with demographic factors alone as predictors entered in step 1, and attachment styles and demographic factors jointly in step 2 (see Table 3-3). When entered alone,

ATTACHMENT AND ADJUSTMENT OF IMMIGRANTS

44

attachment styles (see Table 3-2) explained more variance than demographic factors entered alone in step 1 in hierarchical regression (see Table 3-3) in every dependent variable except for Identification and Contact with the Dutch culture. In step 2 (see Table 3-3) attachment styles made a significant improvement of the prediction as compared to the prediction made by the demographic factors in step 1. Table 3-3 shows that the increase of R2 in predicting Psychological Health and Satisfaction With Life was evident, ranging between .14 and .34 in the three samples. For Social Support and Identification and Contact with the Dutch culture the increase in R2 was between .22 and .36 and .12 and.24, respectively. For Identification and Contact with native culture, however, the increase of explained variance was smaller .07-.17 and significant only in the Polish and Russian samples.

Table 3-2. Summary of Regression Analysis for Attachment Variables Predicting Psychological and Sociocultural Adjustment of Polish (N = 408), Russian (N = 100) and Hungarian (N = 123) immigrants. Sample Identification and Identification and Psychological health Satisfaction with life Social support contact with contact with the Dutch culture a native culture B SE B B SE B B SE B B SE B B SE B Secure Polish .22 .08 .18** .28 .06 .29** .19 .05 .22** .17 .09 .11 .40 .05 .44** Russian .43 .12 .40** .31 .12 .32** .11 .14 .09 .17 .17 .11 .43 .10 .44** attachment Hungarian .22 .14 .23 .01 .11 .01** .35 .10 .38** .00 .16 .00 .18 .10 .22 Fearful Polish -.13 .05 -.15* .05 .04 .08 -.09 .04 -.16* -.32 .06 -.31** -.04 .04 -.06 Russian .00 .10 .00 .07 .10 .10 -.29 .11 -.33** -.39 .13 -.34** .09 .09 -.13 attachment Hungarian .01 .09 -.01 .07 .07 .13 -.04 .06 -.07 -.24 .10 -.29* -.19 .06 -.36* Preoccupied Polish .01 .06 .01 .03 .04 .04 -.15 .04 -.21** -.11 .06 -.09 -.01 .04 -.01 Russian -.12 .07 -.19 -.02 .06 -.04 -.19 .07 -.27** -.22 .07 -.28** .03 .06 -.05 attachment Hungarian -.12 .09 -.15 .01 .07 .15 -.14 .06 -.21* -.22 .09 -.24* .02 .06 -.03 Dismissing Polish .01 .05 -.01 .04 .04 .05 -.02 .04 -.05 .01 .06 .01 -.02 .04 -.03 Russian -.02 .10 -.03 .25 .09 .30* -.05 .11 -.05 -.16 .13 -.09 -.10 .08 -.12 attachment Hungarian .14 .11 .15 .08 .08 .11 -.10 .08 -.12 .01 .12 .01 .02 .07 .03 Polish .08 .07 .19 .17 .23 R2 Russian .23 .16 .31 .36 .31 Hungarian .13 .07 .28 .19 .29 Note. *p < .05; **p < .01

Table 3-3. Summary of Hierarchical Regression Analysis for Demographic Factors and Attachment Styles Predicting Psychological and Sociocultural Adjustment of Polish (N = 408), Russian (N = 100) and Hungarian (N = 123) immigrants. Sample Identification and contact Identification and contact Psychological health Satisfaction with life Social support with with the Dutch culture a native culture B SE B B SE B B SE B B SE B B SE B Step1 Polish .16 .04 -.15** -.05 .03 -.10 .08 .03 .16** .19 .05 .19** .05 .03 .09 Education Russian -.03 .07 -.05 -.01 .07 -.01 .04 .07 .06 .04 .11 .05 -.05 .06 -.09 Hungarian .01 .08 -.01 -.11 .07 -.17 .03 .08 .04 .09 .11 .08 .04 .08 .06 Polish .04 .00 .38** .00 .00 .01 .00 .00 .01 .00 .00 .01 .00 .00 -.03 Length of Russian .01 .01 .06 .01 .01 .13 .03 .01 -.24* -.01 .02 -.04 -.01 .01 -.12 residence Hungarian .01 .00 .24* .00 .00 -.08 .00 .00 .09 .02 .01 .35** -.01 .00 -.15 Polish 00 .00 -.05 .00 .00 .04 .00 .00 .00 -.01 .00 -.10 -.01 .00 -.10 Age at Russian -.02 .01 -.25* .01 .00 .16 .02 .01 -.25* -.02 .01 -.25* .00 .01 -.01 immigration Hungarian -.03 .01 -.24* .01 .01 .08 .01 .01 -.06 .01 .02 .07 .00 .01 .01 Polish .18 .09 .02 .04 .02 R2 Russian .07 .05 .12 .06 .03 Hungarian .15 .03 .01 .11 .03 Step 2 Polish .10 .04 .12* -.06 .03 -.10 .04 .03 .08 .11 .05 .11* .01 .03 .02 Education Russian -.07 .07 -.12 -.01 .07 -.01 .02 .07 -.03 -.06 .09 -.07 -.13 .05 -.23 Hungarian .01 .08 -.01 -.12 .07 -.20 .05 .07 .08 .13 .11 .12 .07 .06 .10 Polish .03 .00 .37** .00 .00 -.01 .00 .00 -.03 -.01 .01 -.01 .00 .00 -.07 Length of Russian .01 .01 .07* .01 .01 .14 .03 .01 -.25** .00 .02 -.02 -.01 .01 -.11 residence Hungarian .01 .01 .27* .00 .00 -.01 .00 .00 .09 .02 .00 .31** -.01 .00 -.16 Polish .00 .00 -.03 .00 .00 .05 .00 .00 .02 -.01 .00 -.08 .00 .00 -.07 Age at Russian -.01 .01 -.16 .01 .01 .17 .01 .01 -.14 -.01 .01 -.11 .01 .01 .15 immigration -.20* .06 .01 .12 .10 Hungarian -.24 .01 .01 .01 .00 .01 .02 .01 .01 .01 Polish .19 .07 .15** .27 .06 .28** .21 .05 .23** .15 .09 .10 .39 .05 .42** Secure Russian .42 .12 .39** .34 .12 .35** .13 .14 .10 .25 .18 .16 .48 .10 .48** attachment Hungarian .27 .13 .27* .03 .11 .04 .33 .11 .37** .16 .16 .13 .20 .09 .25* Polish -.11 .05 -.13* .04 .04 .06 -.08 .04 -.14* -.30 .06 -.29** -.04 .04 -.07 Fearful Russian .02 .11 -.02 .06 .11 .08 -.27 .11 -.30* -.37 .15 -.32* -12 .09 -.17 attachment Hungarian .02 .08 .04 .09 .07 .16 -.09 .07 -.16 -.13 .10 -.16 -.20 .06 -.39** Polish .04 .05 .04 .04 .04 .05 -.16 .04 -.22** -.09 .06 -.07 .00 .04 .00 Preoccupied Russian -.10 .07 -.16 -.02 .06 -.03 -.18 .07 -.27** -.23 .08 -.27** -.01 .06 -.02 attachment Hungarian -.05 .09 -.07 .11 .07 .19 -.11 .07 -.16 -.20 -.10 -.22 -.03 .06 -.05 Polish .01 .05 .01 .04 .04 .04 -.02 .04 -.04 .03 .06 .03 -.03 .04 -.04 Dismissing Russian .-.01 .10 -.01 .26 .10 .31** .05 .11 -.05 -.10 .14 -.07 .09 .09 -.11 attachment Hungarian .14 .10 .15 .12 .08 .16 -.15 .09 -.18 .00 -.12 -.01 .02 .07 .03 Polish .24 (.24**) .27 (.07**) .20 (.18**) .18 (.14**) .23 (.22**) R2 (R2) Russian .28 (.21**) .22 (.17**) .28 (.27**) .40 (.34**) .39 (.36**) Hungarian .27 (.12) .15 (.11) .32 (.30**) .25 (.15**) .39 (.36**) Note. *p < .05; **p < .01

CHAPTER 3

47

Discussion The main purpose of this study was to examine the relation between attachment styles and adjustment of immigrants. The predictions with respect to secure and fearful attachment were largely supported. The findings clearly indicate a positive relation between secure attachment and psychological and sociocultural adjustment. Fearful attachment was, as expected, negatively associated with psychological and sociocultural adjustment. With respect to the other two attachment styles, the findings are partially in accordance with the hypotheses. Preoccupied attachment was, as predicted, negatively associated with psychological adjustment, but appeared unrelated to sociocultural adjustment. Dismissing attachment was, as expected, unrelated to psychological health but, contrary to expectations, showed no relation to sociocultural adjustment. To sum up, the results are largely consistent with our hypotheses. Moreover, they are consistent with previous studies in which immigrants attachment styles, particularly secure and fearful styles, were related to both psychological and sociocultural adjustment (Bakker et al., 2004) and acculturation attitudes (Van Oudenhoven & Hofstra, 2006). Next there was some, but not consistent, evidence for the relation between demographic factors and adjustment. Education was found to be positively associated with psychological adjustment and identification and contact with the host culture. Length of residence appeared to be positively related to identification and contact with the host culture. Age at immigration appeared to be negatively related to psychological adjustment and identification and contact with the host culture, which is in line with other findings (Stevens, 1999). In general, demographic factors appeared to be more related to cultural than to psychological adjustment. Attachment styles, in turn, were more strongly related to psychological adjustment than demographic factors, and clearly appeared to be better predictors of psychological and sociocultural adjustment of immigrants than demographic factors. To our knowledge, this is the first study in which the predictive power of attachment styles and demographic factors has been compared. The findings of the present study indicate that problems of immigrants psychological and sociocultural adjustment are intertwined with problems of attachment more than with demographic factors. Since the attachment framework proved to be useful in immigration studies, future studies concerned with immigrants adjustment could benefit from taking an attachment perspective. Due to the cross-sectional design of the present study, causal relationship between variables may not be inferred. Future studies incorporating a longitudinal design should clarify the causal relationship between attachment and psychological and sociocultural adjustment. Another

ATTACHMENT AND ADJUSTMENT OF IMMIGRANTS

48

limitation of the study was the sampling procedure; data were collected from members of immigrants organizations, thus immigrants detached from their culture of origin might be under-represented. Also, using a snowball method may have caused undesirable homogeneity of the sample. We studied immigrants from three Eastern European countries that, despite many similarities, are characterized by some real cultural differences. Further studies will clarify if differences between the samples with respect to psychological and sociocultural adjustment may be attributed to cultural background of immigrants.

Chapter 4 Evidence for a Migrant Personality: Attachment Styles of Poles in Poland and Poles in The Netherlands

Introduction Migrant Personality The beginning of the 21st century is marked by globalization and mass international migration. Currently there are an estimated 191 million emigrants worldwide (United Nations, 2005). What is the driving force stimulating people to leave their lands of origin? Modern theories of migration point to economical, demographic and network factors in attempts to answer this question (see Massey, Arango, Hugo, Kouaouchi, Pellegrino, & Taylor, 1998, for an overview). These factors, however, do not explain why some individuals, living in similar socio-economic conditions, become emigrants whereas others do not. Therefore, psychological factors favoring and impeding emigration may play a role in this selection process. Conceivably, emigrants possess characteristics, which predispose them to emigrate and help them to endure the challenges of emigrant life. In the literature the term pioneering personality was used to refer to psychological characteristics of individuals prone to emigration (Morrison & Wheeler, 1976). More recent research has indicated that emigrants are less prone to anxiety and insecurity than non-emigrants (Ray, 1986). Many cases of successful emigrants Marie Curie, Madeleine Albright or Salman Rushdie, to mention just a few, seem to support the hypothesis of a favorable migrant personality. While the outflow of well-educated individuals from donor societies, referred to as a brain drain, has been broadly discussed, personality drain induced by the outflow of individuals with favorable personality traits has largely been overlooked. Hence, the present study endeavors to address this issue. Linking Attachment and Emigration As we saw above, several studies have supported the idea of a migrant personality and even specified characteristics associated with this kind of personality (Boneva & Frieze, 2001; Ray, 1986). Remarkably, however, these studies have not paid attention to attachment styles, whereas these characteristics refer to the nature of affectional bonds between individuals

EVIDENCE FOR A MIGRANT PERSONALITY

50

(Bowlby, 1973), and groups (Smith, Murphy, & Coats, 1999), and more interestingly -the way of dealing with new situations and people (Ainsworth, Blehar, Waters, & Wall, 1978). As formulated by Bowlby, (1973, p.147) attachment ties induce a common tendency for humans to remain in a familiar locale and among familiar people. Emigrants, leaving their familiar milieu, deviate from this common tendency. Therefore, it is particularly interesting to investigate whether their attachment styles predispose them to do so. Furthermore, establishing bonds with new people and coping with new situations are important elements in the daily life of emigrants. As a result, attachment styles might be crucial to the successful adjustment of emigrants. Previous research has shown, for example, that attachment is a better predictor of emigrants adjustment than the Big Five personality traits (Bakker, Van Oudenhoven, & Van der Zee, 2004) and demographic variables (Polek, Van Oudenhoven, & Ten Berge, 2007). Accordingly, the first goal of the present study is to investigate, if attachment styles of emigrants differ from attachment styles of their non-emigrant fellow nationals. In addition, we tested if these differences, should they exist, may be assumed to exist prior to emigration (as predispositions for emigration), or rather emerge after emigration (as a result of emigration), induced by contact with the host culture, or, alternatively, by the mere fact of emigrating. Third, we investigated if attachment is a better predictor of the psychological wellbeing of emigrants as compared to non-emigrants. Attachment vs. Personality Originally formulated by Bowlby (1973), attachment theory conceptualizes the universal human need to form affectional bonds with others. Within this theory, attachment is described as a behavioral regulatory system which provides infants with the capacity to use one or a few primary figures as a secure base from which to explore, and to which to return to when seeking safety and proximity, in cases when exploration brings distress. Every individual constructs mental representations of their own secure base, thus conserving their attachment experience and translating it into other relationships. A secure attachment style evolves when early attachment experience brings the belief that the attachment figure is responsive and caring. Conversely, an insecure style evolves when the primary attachment figure is unresponsive and uncaring. Like the attachment framework, the personality framework also emphasizes the early formation of individual characteristics, but views personality traits as the effect of the interplay between genetic predispositions and environmental (e.g. family) influence. Up till now literature on personality and attachment focused mainly on divergent aspects of these two frameworks (Kobak, 1994; Waters, Crowell, Elliott, Corcoran, & Treboux, 2002), failing to notice apparently convergent

CHAPTER 4

51

aspects such as 1) the crucial role of primary caregivers in both developing personality traits (Collins, Maccoby, Steinberg, Hetherington, & Bornstein, 2000; Maccoby, 2000) and attachment styles (Bowlby, 1973; Diehl, Elnick, Bourbeau, & Labouvie-Vief, 1998;); 2) the evidence that personality traits (Larsen & Ketelaar, 1989; Zelenski & Larsen, 1999) as well as attachment styles (Barry, Lakey, & Orehek, 2007) reflect dispositional susceptibilities to affective states; 3) the stability of personality traits (McCrae & Costa, 2003) and attachment styles (Benoit & Parker, 1994; Hofstra, et al., 2005); 4) substantial correlations between attachment styles and personality traits (Bakker et al., 2004; Chotai, Jonasson, Hgglf, & Adolfsson, 2005; Diehl et al., 1998); 5) finally, the relation between attachment styles and personality disorders (Aaronson, Bender, Skodol, & Gunderson, 2006). In the present study we propose a cautious integrative approach, which acknowledges a distinction between both frameworks, yet affirms that both theories explain individual differences and that therefore it is reasonable to talk about a migrant personality with respect to individual characteristics such as attachment styles. Hypotheses A number of attachment styles have been proposed in the literature (see Cassidy & Shaver, 1999 for an overview). In this study, we will focus exclusively on the secure and dismissing styles (Bartholomew & Horowitz, 1991), as secure and dismissing attachment styles especially seem to distinguish emigrants from non-emigrants. Both secure and dismissing attachment styles have been found to be positively related to a readiness to explore new environments and to approach unfamiliar others (Ainsworth, Blehar, Waters, & Wall, 1978). A secure attachment appeared also to be positively related to positive attitudes towards out-group members (Mikulincer & Shaver, 2001), self-efficacy and self-esteem (Hofstra, Van Oudenoven & Buunk, 2005). Since becoming an emigrant is probably preceded by a readiness to explore a new environment, and to encounter unfamiliar others, as well as a sense of self-efficacy when approaching new people and dealing with new situations, we expect emigrants to score higher on the secure style than non-emigrants [Hypothesis 1]. Contrary to the secure attachment style, the dismissive attachment style is marked by avoiding close ties and dependence on other individuals (Bartholomew & Horowitz, 1991) and groups (Smith, Murphy, & Coats, 1999), and appears to be negatively related to the perception of the importance of contact with ones native culture (Bakker et al., 2004). Recent results indicate that social avoidance may be caused by two different mechanisms fear of intimacy (fearful avoidance) and an absence of attachment needs (dismissing avoidance) (Duggan & Brennan, 1994). In an earlier study, we also found that dismissing

EVIDENCE FOR A MIGRANT PERSONALITY

52

avoidance and fearful avoidance form two separate and replicable factors (Polek, Ten Berge, & Van Oudenhoven, 2006). Accordingly, in the present study we focused on the dismissing avoidance of the respondents. Boneva and Frieze (2001) reported a higher power motivation of emigrants, but lower affiliation, motivation and family centrality. It seems that these characteristics may correspond to the dismissing avoidance of emigrants. Given that emigrants voluntarily choose to leave familiar others and their culture of origin, we expect that they score higher on dismissing attachment than non-emigrants [Hypothesis 2]. Since a dismissing attachment was only slightly negatively related to a secure attachment in the present study, we deemed it reasonable to assume that an individual may score high on both a secure and dismissing attachment style. Ideally, a comparison between emigrants and non-emigrants would be done at the moment emigrants leave the country. Since approaching emigrants at the very moment of emigration is technically difficult, we administered questionnaires to emigrants who have already lived in their host country for some time. Because attachment styles are relatively stable characteristics of individuals (Ainsworth, Blehar, Waters, & Wall, 1978), we assume that the differences in attachment styles between emigrants and non-emigrants posited above exist prior to emigration and are stable over time [Assumption 1]. However, to control for a possible influence of the host culture, and a possible effect of the action of emigrating on the attachment of emigrants, we compare mean scores in subsequent cohorts of emigrants (each cohort differentiated from the next by two more years of residence). In addition, we compare the cohorts of emigrants with their non-emigrant fellow countrymen and a host sample. If the mean scores and the magnitude of differences in the mean scores between emigrants and non-emigrants, and in addition between emigrants and the host Dutch sample, remain similar across cohorts, it would suggest that the attachment styles of emigrants remain relatively stable over time. In that case, we can conclude that the differences between emigrants and non-emigrants were not produced by the influence of the host country, nor the mere action of emigrating, but were present prior to emigration. With respect to the possible influence of the host culture on the attachment of emigrants, we will furthermore test if emigrants who identify themselves strongly with Dutch culture resemble the Dutch native people more on secure and dismissing attachment styles than those emigrants who show a weak identification with the Dutch. The absence of an effect of the host culture on attachment styles of emigrants would support our assumption 1. It has been demonstrated that secure attachment is positively associated with social competencies (Mallinckrodt, 2000), psychological well-being and coping with problems (Cooper, Shaver, & Collins, 1998), as well as with the

CHAPTER 4

53

psychological and socio-cultural adjustment of emigrants (Polek et al., 2007; Van Ecke, et al., 2005). Therefore, we suppose that secure attachment is a predictor of psychological well-being for both emigrants and non-emigrants. However, since emigrants are exposed to unfamiliar others and new situations far more than non-emigrants, we expect that secure attachment is a better predictor of psychological well-being for emigrants than for nonemigrants [Hypothesis 3]. We do not expect that dismissing attachment is a relevant predictor of psychological well-being, for either emigrants or native respondents. We do not expect a difference in the predictive power of a dismissing attachment style between these two samples.

EVIDENCE FOR A MIGRANT PERSONALITY

54

Method Participants Four hundred and eight emigrants from Poland (68% female) living in the Netherlands, 587 Polish respondents (59 % female) living in Masuria, a region of Poland where the unemployment rate is 30% making it a typical emigration region, and 181 Dutch respondents living in the Netherlands (67 % female) participated in our survey. The average age of respondents was 33.50 (SD= 10.93) in the sample of Polish emigrants living in the Netherlands, 35.10 (SD=13.23) in the Polish native sample living in Poland, and 49.10 (SD=4.18) in the Dutch sample. Procedure A translation of the questionnaires from the Dutch language into Polish was made, and this translation was then translated back into Dutch. We compared the return-translation with the original version in order to check the accuracy of the translations. The final version of the questionnaires in Polish, as well as the original Dutch version, was presented once again to another translator who was asked to evaluate the equivalence of each translated item with the original version. The equivalence of the original and the translated questionnaires was rated by the translator as high. Polish emigrants were approached in places where they meet: communities, clubs, Sunday schools and parishes. Respondents were given the questionnaire, a letter with brief instructions and an explanation of the aim of the study and a pre-paid envelope. They were asked to complete the questionnaire and to distribute additional questionnaires among friends of the same nationality living in the Netherlands. Polish respondents living in Poland were approached in schools, universities and companies. They received questionnaires, instructions and envelopes, and they also were asked to distribute questionnaires among their friends and acquaintances. Questionnaires were collected back after a few days from the respondents. Data from the Dutch respondents were obtained via a mail survey. All respondents filled in questionnaires voluntarily and without any monetary compensation. Missing data accounted for 7%, 2%, and less than 1 % of the data in the Polish emigrant, Polish native and Dutch samples, respectively; we used pair-wise exclusion of the missing data from the analyses. Instruments We asked respondents about their age, gender, education, and marital status. Emigrant respondents were also asked about their age at emigration, and length of residence in the Netherlands.

CHAPTER 4

55

Two scales from the Attachment Style Questionnaire (ASQ; Van Oudenhoven, Hofstra, & Bakker, 2003, see Appendix 1) were used to assess secure and dismissing attachment of all the respondents. Inter-correlation between the secure and the dismissing scale was significant, but low in the Polish emigrant sample (r = -.12, p < .05) and insignificant in the Polish incountry and the Dutch native sample (r = -.04 and r = -.06, respectively). These two scales showed satisfactory factor replicability (Polek et al., 2006), validity (Hofstra et al., 2005), and stability measured after one year with Pearson correlation (.63 for the secure style, and .63 for the dismissing style) in the study on the Dutch sample. In contrast to many existing attachment measures that focus only on relationship specific attachment, the ASQ assesses general attachment, which is an individuals predisposition to build up social relationships. The ASQ assesses attachment through multiple scores (on each dimension separately). As we already mentioned in the introduction, participants were not classified into one attachment category, but received scores on each attachment scale. A 5-point answering scale was used for this questionnaire, ranging from strongly disagree (1) to strongly agree (5). An example of an item from the 8-item secure scale was: I find it easy to get engaged in close relationships with other people; from the 5-item dismissing scale: I like to be self-sufficient. Alpha coefficients of the secure scale were .71, .76, and .77; and of the dismissing scale .63, .56, and .54 in the Polish emigrant, Polish non-emigrant and Dutch samples, respectively. As a measure of psychological well-being, we used a 9-item scale, the Psychological Health scale obtained from the RAND 36-item Health Survey (RAND Health Sciences Program, 1992). A sample item from this scale is: How much of the time during the past four weeks have you been a very nervous person? Respondents gave answers on a 5-point scale ranging from all the time (1) to none of the time (5). Alpha coefficients of the scale were .87 and .83 in the Polish emigrant and native Polish samples, respectively. An 8-item scale assessing Dutch identity and contact with Dutch culture were used in the sample of Polish emigrants living in the Netherlands. A sample item from this scale is: I like to chat with the Dutch. Respondents gave answers on a 5-point scale: from strongly disagree (or never) (1) to strongly agree (or very often) (5). Alpha coefficients of the scale reached .78. Results After we checked whether the data met the requirement of multivariate normality, we carried out a multivariate analysis of covariance (MANCOVA) comparing Polish emigrants with Polish respondents living in Poland with respect to their secure and dismissing attachment styles, when controlling for age, gender, and education in both groups. Polish emigrants scored significantly higher on secure and dismissing attachment than Polish

EVIDENCE FOR A MIGRANT PERSONALITY

56

respondents living in Poland. A univariate test revealed a significant effect of the group (Polish emigrant vs. Polish non-emigrant) for secure attachment F(1, 924) = 175.06, p < .001, 2 = .70, and for dismissing attachment F(14, 1052) = 152.90, p < .001, 2 = .67. To test the assumption that the influence of the host culture and the mere act of emigration is minimal on emigrants attachment, we compared mean scores on secure and dismissing attachment in consecutive cohorts of emigrants (each cohort differentiated from the other cohort by a further two years of residence). We also compared cohorts of emigrants with Polish respondents living in Poland, as well as with native Dutch respondents. In these comparisons we controlled for age, gender and the education of respondents (Table 4-1). We reasoned that our assumption about the lack of influence of the host culture on emigrants attachment would be proven if the mean scores in consecutive cohorts of emigrants remained similar and the differences in attachment styles between Poles living in Poland and cohorts of Polish emigrants remained similar across cohorts. In the same vein, if the differences between the Dutch respondents and consecutive cohorts of emigrants remained similar, we could conclude that there is no evidence for the influence of the host culture on attachment styles of emigrants. Table 4-1 and Figure 1 show that means and estimated marginal means remain similar across cohorts. Also the differences between consecutive cohorts of emigrants and the native Polish sample, and the differences between consecutive cohorts of Polish emigrants and the native Dutch sample remain, in most cases, similar and significant over cohorts. A multivariate test revealed a significant main effect of the group (Polish non-emigrants / cohorts of Polish emigrants / the Dutch native sample) F(14, 1052) = 43.34, p < .001, 2 = .37, gender F(14, 1052) = 2.59, p < .05, 2 = .01, and age F(14, 1052) = 3.67, p < .05, 2 = .01. A univariate test revealed a significant effect of the group (Polish non-emigrants / cohorts of Polish emigrants / the Dutch native sample) for secure attachment F(14, 1052) = 175.06, p < .001, 2 = .70, and for dismissing attachment F(14, 1052) = .152.90, p < .001, 2 = .67. The differences remained relatively similar across cohorts, thus supporting our assumption. We may therefore conclude that attachment styles are relatively stable characteristics and that differences between emigrants and non-emigrants in attachment styles already existed prior to emigration. The difference in mean scores on dismissing attachment between cohorts of emigrants and the Dutch sample became insignificant in cohorts living in the Netherlands over 16 years. However, this effect may be due to the small number of respondents in these cohorts as the mean scores remain very similar across all cohorts, including those with a length of residence longer then 16 years (see Table 4-1). To further test the possible influence of the host culture on attachment styles of emigrants, we checked whether

CHAPTER 4

57

emigrants who scored high on Dutch identity had more similar attachment patterns to Dutch people than those who scored low on Dutch identity. For that purpose we calculated the absolute discrepancy scores between the mean scores provided from the Dutch sample and the responses of the Polish emigrants. These scores estimated the magnitude (but not direction) of the difference between the scores of the Polish emigrants and the Dutch respondents. Next, we subdivided the sample of Polish emigrants into a subsample with a high Dutch identity (above the average 3.47) and a sub-sample of emigrants with a low Dutch identity (below the average) and compared two sub-samples in terms of absolute discrepancy scores. We reasoned that if emigrants scoring higher on Dutch identity had lower discrepancy scores on attachment styles, it would suggest that the host culture can influence the attachment of emigrants. MANCOVA with Dutch identity (low versus high) as a factor, discrepancy scores between the Polish emigrants and the Dutch respondents on secure and dismissive attachment as dependent variables, and age, gender and education of respondents as controlled variables, revealed no significant effect of the level of Dutch identity for discrepancy scores in secure attachment F(1, 371) = 1.73, p < .19 (mean difference -.04), or for discrepancy scores in dismissing attachment F(1, 371) = .14, p < .71 (mean difference .02). Once more, we found evidence for the assumption that the influence of the host culture on emigrants attachment is absent and, consequently, that differences in attachment styles between emigrants and non-emigrants exist prior to emigration. To examine if secure attachment is a better predictor of psychological wellbeing of emigrants as compared to non-emigrants [Hypothesis 3], we conducted regression-by-group analyses, in which psychological health was entered as a dependent variable, and secure attachment as predictor in the Polish emigrant and Polish native samples, respectively (Table4-2). Next, a Chow test of equality between coefficients in linear regressions (Chow, 1960) was performed in order to compare the Beta slopes obtained in regression analyses in the emigrant and native Polish samples. This test revealed that the secure attachment was a significantly better predictor of psychological health for emigrants than for native respondents F(1, 904) = 5.08, p < .01. Thus, we found support for hypothesis 3. The same calculation was done for dismissing attachment. As expected, we did not find evidence that dismissing attachment is a predictor of psychological well-being in either sample (Table 4-2).

EVIDENCE FOR A MIGRANT PERSONALITY

58

Table 4-1. MANCOVA of Secure and Dismissing Attachment of Polish Emigrants and Polish Native, and Dutch Native Sample, Controlling for Gender, Age and Education. Secure attachment style Samples N M (SD) Estimate d marginal means Estimated differences in mean scores A Polish native Dutch native Polish emigrants (total)
Cohorts of Polish emigrants (length of residence):

Dismissing attachment style M (SD)

Estimated marginal means

Estimated differences in mean scores C D -

B 2.07 0.70 3.50 0.61 4.12 (0.70) 2.07 3.48

524 171

2.19 (0.57) 3.68 (0.47) 3.93 (0.54)

2.19 3.63

3.95

-1.77**

4.15

-2.10**

1 (0 - 2 years)

3.85 4.19 3.88 -1.69* -.25 (0.58) (0.63) 2 (2.1 - 4 3.95 4.08 73 3.98 -1.79* -.35* years) (0.52) (0.69) 3 (4.1 - 6 3.90 4.06 57 3.92 -1.73* -.29 (0.53) (0.63) years) 4 (6.1 - 8 3.85 4.21 24 3.87 -1.68* -.23 (0.57) (0.63) years) 5 (8.1 -10 4.18 4.39 12 4.19 -2.01* -.56 years) (0.51) (0.66) 6 (10.1 - 12 4.12 4.30 18 4.12 -1.93* -.49* years) (0.46) (0.58) 7 (12.1 - 14 3.94 4.06 24 3.94 -1.75* -.31 (0.58) (0.85) years) 8 (14.1 - 16 4.11 4.44 9 4.1 -1.91* -.47 years) (0.52) (0.55) 9 (16.1 -18 3.99 3.44 9 3.97 -1.79* -.34 years) (0.30) (1.10) 10 (18.1 - 20 3.98 4.29 7 3.96 -1.78* -.34 (0.52) (0.71) years) 11 (20.1 - 22 3.72 4.21 8 3.70 -1.51* .00 (0.89) (0.75) years) 12 (22.1 24 3.92 4.22 9 3.89 -1.71* -.26 years) (0.26) (0.55) 13 (above 24 4.18 4.06 11 4.14 -1.95* -.51 (0.36) (0.71) year) Levels of significance of the differences in means *p < .05, ** p < .001 A - mean scores of Polish natives minus mean scores of Polish emigrants 114

4.20 4.09 4.08 4.24 4.44 4.33 4.08 4.45 3.50 4.35 4.29 4.29 4.13

-2.14** -2.03** -2.02** -2.17** -2.37** -2.27** -2.02** -2.39** -1.44** -2.28** -2.23** -2.22** -2.07**

-.72** -.61** -.59** -.76** -.96** -.85** -.60** -.97** .00 -.87 -.81 -.80 -.65

B - mean scores of Dutch natives minus mean scores of Polish emigrants C - mean scores of Polish natives minus mean scores of Polish emigrants D - mean scores of Dutch natives minus mean scores of Polish emigrants

CHAPTER 4

59

Estimated 5 means 4.5

4 3.5 3 2.5 2 1.5 1 0.5 0 P 1 2 3 4 5 6 7 8 9 10 11 12 13 D

Secure attachment

Dismissing attachment

Samples

Figure 4-1. Estimated means on secure and dismissing attachment styles of Polish native (P), cohorts of Polish emigrants (1-13), and Dutch native sample (D).

Table 4-2. Summary of Regression Analysis for Secure and Dismissing Attachment, Entered one by one, Predicting Psychological Health in the Polish Native (N = 524) and Polish Emigrant (N = 380) Samples Polish native sample Polish emigrant sample 2 B SE B B SE B R R2 Secure .17 .03 .23* .05 .29 .04 .32* .10 attachment Dismissing .00 .03 -.03 .00 .00 .04 .06 .00 attachment Levels of significance * p < .001

EVIDENCE FOR A MIGRANT PERSONALITY

60

Discussion In the present study we examined whether there was empirical evidence for a migrant personality. We compared Polish emigrants and Polish respondents living in Poland in terms of their secure and dismissing attachment, when controlling for respondents age, gender and education. Emigrants were found to have more secure and dismissing attachment styles than non-emigrants. We did not find evidence that the attachment styles of emigrants change under the influence of the host culture or as a consequence of emigration. Therefore, we assume that the observed differences in attachment styles between emigrants and non-emigrants existed, most likely, prior to emigration. Moreover, we have seen that secure attachment is a better predictor of psychological well-being for emigrants than for non-emigrants. In sum, we found evidence for a migrant personality. As the results suggest, this personality is functional, helping emigrants to better adjust to the new environment. The results also indicate, in line with other findings (Hazan & Shaver, 1987; Kobak & Sceery, 1988), that adult attachment styles may be considered as stable characteristics of an individual, similar to personality traits. It may appear paradoxical that the same group of respondents scored high on secure attachment style which is characterized by comfort with closeness and intimacy, and dismissing attachment, characterized by the absence of a need for close ties. A review of other findings indicate, however, that such a result is very plausible. For example, in the classical studies on infants reaction to strange situations, both secure and avoidant attachment were found to be positively associated with a high exploration activity in the absence of the primary caregiver (Ainsworth, Blehar, Waters, & Wall, 1978; Thompson, 1988). Previous results showed that social avoidance might be caused by two different mechanisms: fear of intimacy (fearful avoidance) and denial or lack of attachment needs (dismissing avoidance) (Duggan and Brennan, 1994; Kafetsios & Nezlek, 2002). The latter type of avoidance was reported to be positively correlated with self-directedness (Chotai, Jonasson, Hgglf, & Adolfsson, 2005), a personality trait (Cloninger, Svrakic, & Przybeck, 1993) that might be high in individuals prone to emigration. Moreover, both secure and dismissing attachment styles are presumed to be underpinned by a positive model of self (Bartholomew & Horowitz, 1991). It is possible that a positive model of self may cause the propensity for emigration in individuals scoring high on a secure and a dismissing attachment. It is noteworthy that although both dismissing and secure attachment predispose individuals to emigration, only secure attachment turned out to be related to the psychological adjustment of emigrants. Possibly, a dismissing

CHAPTER 4

61

attachment makes individuals more detached from their social surroundings and thus prone to emigration, but it does not help them to become psychologically well-adjusted in the new social environment. Present findings, in line with the study of Bakker et al., 2004, suggest that emigrants who score high on dismissing attachment, may show little need to socialize with members of their native or host cultures. Possibly, emigrants who score high on both secure and dismissing attachment have the capacity to build up social relationships, but do not have the need to do so, either because of their low affiliation motivation, or the priority of work and achievement over personal ties (Boneva & Frieze, 2001). Some results suggest, for example, that a high level of affiliation motivation could be predictive of a desire to stay in the country of origin (Scott & Scott, 1989). In accordance with our results, the study of Van Vianen, Feij, Krausz, & Taris (2003) showed that both secure attachment and sensation seeking predicted voluntary job mobility. Correlational studies also demonstrated a positive association between secure style and novelty seeking (Chotai et al., 2005). Conceivably, novelty seeking (Cloninger, Svrakic, & Przybeck, 1993) underscores the high exploratory activity of secure and dismissing individuals. Yet, the differential predictive power of secure and dismissing attachment with respect to psychological well-being suggests that the psycho-social adjustment of individuals might be more related to attachment styles then to novelty seeking. Further studies are needed to clarify the relation between personality traits, such as novelty seeking, and attachment styles, as well as their predictive value for individuals psychological and social well-being. The present study has some limitations. First, we administered questionnaires to emigrants who had already emigrated, and not to emigrants to-be. Thus, we had to take into account the possibility that the differences in attachment styles between emigrants and non-emigrants that we observed may have evolved due to the influence of the host culture or the mere fact of emigration. To control for the influence of the host culture a longitudinal design would have been ideal. As a substitute to this, we carried out comparisons of consecutive cohorts of emigrants. We also have to keep in mind that studies on emigrants are carried out on individuals who have been able to cope efficiently enough with cultural shock to stay in the emigration country. Those who did not possess effective coping strategies might have returned to their country of origin, and therefore may not be included in the present study. Thus, a higher secure attachment style of emigrants might be caused, to some extent, by this selection mechanism. In the introduction we hypothesized about personality drain, a phenomenon which, similarly to brain drain, may occur in donor societies due to the outflow of individuals with favorable personality traits. As we have seen, emigrants seem to possess a personality setup making them prone to

EVIDENCE FOR A MIGRANT PERSONALITY

62

exploration and helping them to endure the unsettlement it may bring. We may conclude from the present study that personality drain may indeed happen in donor societies.

Chapter 5 A Cultural Look at Parenting and Attachment Styles in Eastern and Western Europe
Introduction For almost half a century Eastern and Western Europe had been separated from each other by different ideologies, and different social and economic systems. Political changes after 1989 ceased this long-lasting separation and increased economic, social and cultural contacts between countries of the previous Eastern and Western Blocs. As a consequence of these contacts, and more generally through the globalization process, Eastern and Western European cultures have arguably become more similar. Owing to the fact that national borders are more permeable, culture has a much more international and unified character than several years ago. Take television for instance: many serials and programs are broadcasted internationally contributing to the unification of cultural patterns. Or the Internet, to take another example, also internationalizes social contacts and unifies cultures. Furthermore, the freedom to travel increases the frequency of direct social encounters between Eastern and Western Europeans. It is tempting to ask if, in the case of the generation raised in post Cold War Europe, Eastern and Western Europeans have more similar psychological characteristics when compared to the generation of their parents, who grew up when differences between the East and the West were very large and pervasive. Studies stemming from temperament and personality frameworks suggest that culture plays a marginal role in the emergence of individual differences (e.g., Fulker, Eysenck, & Zuckerman, 1980). But there again, temperament is conceptualised as an inborn feature; consequently, one should not expect a significant cultural component in it. The present study will attempt to elucidate if culture may influence attachment, which is conceptualised as a construct related more closely to the environment than to genes. The Attachment Framework Attachment theory integrates biological (hereditary) and environmental approaches, proposing that the hereditary mechanism underlines the development of the attachment system, but that the environment, and parental practices in particular, is the main source of individual differences in attachment styles. The author of attachment theory, Bowlby (1984), argued that children have a tendency to seek for proximity to, and contact with, a specific caregiver in times of distress, due to a genetic self-protection

ATTACHMENT AND PARENTING STYLES IN EASTERN AND WESTERN EUROPE

64

mechanism which urges them to look for a powerful attachment figure who could provide protection in potentially dangerous situations. The presence of a caring and responsive attachment figure would induce a development of a secure attachment style in the child; inconsistent and unresponsive care would induce insecure attachment. The early attachment experience is preserved in working models as shaping future patterns of relationships with other people in adulthood. A number of attachment styles have been proposed in the literature (see Cassidy & Shaver, 1999 for an overview). The present study draws on the Bartholomew and Horowitz (1991) model, in which four attachment styles are postulated: the secure style characterised by trusting oneself and others; the fearful style characterised by lack of trust in oneself and in others; the preoccupied style characterised by a desire for close relationships with others and, at the same time, a fear of being rejected; and the dismissing style characterised by trust in oneself, avoidance of relationships and excessive self-sufficiency. The model was empirically validated and used as a framework in adolescent and adult attachment research (Bartholomew & Horowitz, 1991; Griffin & Bartholomew, 1994). Goals of the Study Parent-child relationships are embedded in a broader social system, such as the extended family, school and neighbourhood. This broad social system affects parenting styles and patterns of parent-child interactions. Due to the fact that all societies expect certain characteristics and behaviours from people in order to function adequately as members of their society, functional parenting, apart from universal biological practices such as caring for and protecting children, often includes specific practices related to the values and standards prevailing in the particular society (see Rubin & Chung, 2006 for an overview; Minturn, & Lambert, 1964; Hanono, 1999; Keller, 2003). Given that cultures differ with respect to ideas about parenting, it is reasonable to assume that cultural differences can be found also with respect to attachment styles. Some evidence indeed was found that individuals from diverse cultures may endorse different romantic attachment styles (Schmitt, 2003; Schmitt et al., 2004). Previous studies focused on cultural variation in romantic attachment, without paying attention to general, non relationspecific attachment. Moreover, comparative studies on parenting hitherto have focused on relatively remote cultures, e.g., Occidental versus Oriental (see Rubin & Chung, 2006 for an overview; Minturn & Lambert, 1964; Hanono, 1999), whereas parenting practices within Europe have received scant attention. The question as to whether the long-lasting separation of Eastern and Western Europe exerted any influence on parenting and attachment styles, to our knowledge, has not been addressed. Previous research has shown, for example, differences in personality traits of

CHAPTER 5

65

respondents from West and East Germany, with the latter scoring lower on openness (Angleitner & Ostendorf, 2000) and self esteem (Fischer, Meas, & Schmitt, 2006). Moreover, Schmitt, Allik, McCrae and Benet-Martnez (2007) found that Western Europeans scored higher than Eastern Europeans on Neuroticism, the dimension of the Big Five personality theory. Accordingly, the first goal of the present study was to compare Eastern and Western European adolescents and their parents with respect to attachment styles. In addition, we were interested if the unification of Eastern and Western European cultures may have caused a bigger similarity of attachment styles in adolescents from Eastern and Western Europe as compared to their parents. The second goal was to examine what is the role of parental characteristics and culture in the development of the attachment styles of adolescents. Differences in Attachment in Eastern and Western Europe Previous studies have demonstrated that people in collective cultures evaluate the self in terms of interconnectedness and the value they provide to others, more than do people from individualistic cultures (Markus & Kitayama, 1991; Kitayama, Markus, Matsumoto, & Norasakkunkit, 1997). In other studies, insecure romantic attachments were found to be associated with harsh environment and economic hardships (Schmitt, 2003), and preoccupied attachment co-occurred with high rates of collectivism (Schmitt et al., 2004). Western European societies are more individualistic (see Chapter 1), and more affluent than Eastern European ones, so we should expect that Eastern Europeans will score higher on a preoccupied attachment than Western Europeans do [Hypothesis 1]. Although differences between Eastern and Western European cultures are still noticeable, the process of globalization and unification of cultures may increase the similarities in these societies and in people from the East and West (Mott, 2004). Therefore, it is valid to speculate that the differences between Eastern and Western European respondents with respect to attachment styles will be smaller in the adolescent respondents, who were more exposed to cultural unification and cultural exchange, than in their parents [Hypothesis 2]. As clarified above, culture may influence attachment styles. The ecocultural theory of development postulates that cultural influence can be direct e.g., when a child interacts with peers or is taught norms and values promoted by institutions (schools, youth organisations), and indirect, e.g., when values and standards prevailing in the culture influence parenting practices. A number of studies have demonstrated the influence of culture on parenting practices (Dobbins, 2006; Anhalt, 2001; Lamb & Lewis, 2005; Nair & Murray, 2005). Still, the direct influence of culture on attachment styles is less documented. When we find that Eastern Europeans score higher on preoccupied

ATTACHMENT AND PARENTING STYLES IN EASTERN AND WESTERN EUROPE

66

attachment, we will further investigate whether this difference exists due to the direct effect of culture on preoccupied attachment, or, perhaps due to differences in parenting characteristics and parental practices. We expect to find a bigger direct effect of culture (when controlling for parental characteristics) on preoccupied attachment than on secure, fearful and dismissing attachment styles [Hypothesis 3]. Linking Attachment Styles of Children and their Parents The transmittance of attachment styles from parents to children (e.g., Ricks, 1985), and the relation between parents attachment styles and quality of parenting is very well-documented. Secure attachment was associated, for instance, with a higher motivation to parenting than fearful and dismissing styles; and a preoccupied attachment style was associated with a parents desire that the child fulfils the parents dependency needs (Wright, 2000; Rholes, Simpson, & Blakely, 1995). Results also suggest that individuals with a dismissing or preoccupied attachment style may experience problems in developing a strong helping relationship (Ettema, 1999). Dismissing parents experienced greater stress related to parenting, perceived parenting as less satisfying and as less personally meaningful (Rohles, Simpson, & Friedman, 2006), were less helpful and responsive, and displayed generally poor parenting skills (Ettema, 1999). We expect, therefore, that attachment styles of parents will positively predict an analogous attachment style of the child. In other words, we hypothesised that a secure attachment style of a parent is a positive predictor of a secure attachment style of the child; a fearful attachment of a parent is a positive predictor of a fearful attachment style of the child, a preoccupied attachment of a parent is a positive predictor of a preoccupied attachment style of the child; and a dismissing attachment style of a parent is a positive predictor of a dismissing attachment of the child [Hypothesis 4]. Linking Parenting Styles and Attachment Styles of Children There is ample evidence about the relation between parenting practices and childrens attachment styles (e.g., Nair & Murray, 2005; Demo & Cox, 2000; Bowlby, 1984). For instance, sensitivity, acceptance and the emotional accessibility of parents are positively associated with the secure attachment of a child (Ainsworth, Blehar, Waters, & Wall, 1978; Karen, 1998), whereas inconsistent care-taking, rejection and punitive parental beliefs are associated with the insecure attachment of a child (Bridges & Connell, 1991; Egeland & Faber, 1984). Gamble and Roberts (2005) demonstrated that adolescents who perceive their parents as critical and perfectionist tend to report insecure attachment, characterised by difficulty of getting close to others, fear of abandonment, and a low self-esteem. A low level of maternal warmth was

CHAPTER 5

67

associated with an insecure attachment of a child (Martens, 2006), whereas less punitive parenting was positively associated with a supportive and nurturing home environment (Reis, Barbera-Stein, & Bennett, 1986). Many categorisations of parenting styles have been proposed in the literature. In the present study we draw on the Baumrind (1966) classification that proposes three parenting styles: authoritative, authoritarian, and permissive. These three parenting styles seem to parallel the behaviour of parents of children with different attachment styles (Neal & Frick-Horbury, 2001). For example, similar to the parent of a securely attached child, the authoritative parent is sensitive to a childs needs, emotionally responsive, does not use punitive discipline and is moderately controlling (Baumrind, 1966). The authoritarian parent, like the fearful and dismissing parent, is unresponsive to the needs of the child and tends to reprimand and uses physical enforcement, harsh punishment and high control (Kochanska, Kuczymski, & Radke, 1989). Children of authoritarian parents are described as anxious and withdrawn, and as having difficulties with interactions with peers and low self-esteem (Baumrind, 1967; 1971; Elicker, Englund, & Sroufe, 1992). The preoccupied parent, like the permissive parent, is described as inconsistent, being either too lenient or too controlling and punitive, and using withdrawal of love as a punishment. Baumrind (1967) reported that children of permissive parents are anxious, immature and low on self-control and self-reliance. Our review of the literature led us to expect that authoritative parenting is a positive predictor of a secure attachment style in children, while authoritarian and permissive parenting styles - of insecure attachment styles (fearful, preoccupied and dismissing) [Hypothesis 5]. Parental Psychological Health and the Attachment Styles of Children A number of studies clearly demonstrate that parents psychological health is related to that of their children (e.g., Gerlsma, Snijders, Van Duijn, & Emmelkamp, 1997). For instance, Umberson (1989) found that parental psychological health was positively associated with the quality of a parentchild relation. Ettema (1999) showed that the psychological health of a mother was related to her sense of competence as a parent, and psychological health problems in her were found to be associated with a high chance of child abuse. Studies also indicate that parental depression or angermanagement problems are associated with insecure attachment and aggressive behaviours in children (Hutto, 1998). Therefore, we hypothesised that parental psychological health is a positive predictor of a secure attachment style in a child and a negative predictor of insecure ones (fearful, preoccupied and dismissing) of a child [Hypothesis 6]. The Issue of Measurement Equivalence

ATTACHMENT AND PARENTING STYLES IN EASTERN AND WESTERN EUROPE

68

Several authors addressed the problem of bias and instrument equivalence in cross-cultural studies (e.g., Van de Vijver & Leung, 1997; Poortinga, 1995). Equivalence of measures appears particularly critical in cross-cultural comparisons (Van de Vijver & Tanzer, 2004). Sources and types of bias have been discussed at length in the literature, as well as have procedures to minimize them (see Van de Vijver & Leung, 1997 for an overview). It was a focus of attention to minimise the source of bias and to strive for full score comparability equivalence in the present study. Therefore, aside from utilizing equivalent sampling, translation (Van de Vijver & Hambleton, 1996), and application procedures (Van de Vijver & Tanzer, 2004), we tested the construct equivalence of attachment and parenting styles by checking the factorial structure of the questionnaires across the four cultures (Van de Vijver & Leung, 1997), using a pattern of high correlations with related measures (convergent validity) and low correlations with measures of opposite constructs (discriminant validity) (Van de Vijver & Tanzer, 2004). In addition, we controlled for possible response and item bias in measures of attachment and parenting styles before testing the hypotheses (e.g., Poortinga, 1989; Van de Vijver & Poortinga, 1982).

CHAPTER 5

69

Method

Participants Students of secondary schools (between 15-18 years old) from the Netherlands (71% female, mean age 16.08, SD = 1.09), from Poland (75% female, mean age 16.17, SD = 1.34), from Russia, (65% female, mean age 15.30, SD = 1.21), and from Hungary (79% female, mean age 16.27, SD = 1.03), as well as their parents participated in our survey. The numbers of adolescents, mothers and fathers from each country are shown in Table 6-1. The mean age of parents was 47.52 (SD = 3.83), 43.48 (SD = 5.10), 42.29 (SD = 6.49), and 44.40 (SD = 5.13), in the Dutch, Polish, Russian and Hungarian sample. Responses only from mothers accounted for 26, 34, 34, and 33%, and responses only from fathers accounted for 9%, 2%, 4%, and 5 %, and responses from both parents accounted for 65%, 64%, 62%, and 62% in the Dutch, Polish, Russian and Hungarian samples respectively. Procedure Translations of the Dutch questionnaire into Polish, Russian, and Hungarian were checked and revised by native Polish, Russian, and Hungarian psychologists who had spent more than 15 years in the Netherlands and were fluent in Dutch. These versions of the questionnaires, as well as the Dutch version, were presented once again to other Polish, Russian, and Hungarian translators, who were asked to evaluate the equivalence of each translated item with the original version. The equivalence was rated as high. Students of secondary schools in the Randstad (the largest urban area in the Netherlands), Warsaw (the capital city of Poland), Saint Petersburg (the second biggest city in Russia) and Budapest (the capital city of Hungary) were approached at schools during classes, with the consent of school principals and teachers. Students were told in their native language about the aim of the study, and each student received a set of questionnaires (one for a student and two questionnaires for parents), a letter with short instructions for students and parents as well as three small envelopes and one big one. Students were instructed to fill out the questionnaires at home and give questionnaires to their parents. We explained that after completion, questionnaires should be placed in the supplied small envelopes and be sealed, and subsequently placed into one big, common envelope and sealed. This procedure was used to ensure the total privacy of students and parents, and to avoid the mutual influence of family members on each other while responding to our survey. Putting the set of three small envelopes into the big envelope guaranteed that sets of questionnaires from each family were returned individually. All instructions, as well as the questionnaires, were in the respondents native language. The response rates were 25%, 33%,

ATTACHMENT AND PARENTING STYLES IN EASTERN AND WESTERN EUROPE

70

34%, and 75% in the Netherlands, Poland, Russia and Hungary respectively. Missing data were replaced by variable means, separately for children, mothers, and fathers in each of the four samples. Missing data accounted for 2%, 2%, 3%, and 1% of the data in the students samples, 3%, 3%, 7%, and 5% in the mothers samples, and 4%, 5%, 9%, and 6% in the fathers samples from the Netherlands, Poland, Russia and Hungary respectively. Instruments The questionnaire for children contained questions about age, gender, and attachment styles. The questionnaire for parents started with demographic questions about age, gender and the kind of relation to the child participating in our survey (parent, step-parent, adoptive-parent, other). Further, the questionnaire contained questions concerning attachment styles, parenting styles, psychological health, satisfaction with life and social support. Attachment styles of children and parents were measured with the Attachment Style Questionnaire (ASQ, see Appendix 1) by Van Oudenhoven, Hofstra and Bakker (2003) consisting of four scales secure, fearful, preoccupied and dismissing, matching the four attachment styles postulated by the Bartholomew and Horowitzs (1991). In previous studies, the replicability and reliability of scales (Hofstra, Van Oudenhoven, & Buunk, 2005; Polek, Ten Berge, & Van Oudenhoven, 2006), and the construct validity and stability (Hofstra, et al., 2005) had proved to be satisfactory. In contrast to many existing attachment measures that only focus on relationship specific attachment, the ASQ measures general attachment, which makes it possible to assess the general sociability of a respondent. The ASQ assesses attachment through multiple scores (on each dimension separately). Thus, participants were not classified into one attachment category, but received scores on all four attachment scales. For all ASQ scales, a 5-point scale was used, ranging from strongly disagree (1) to strongly agree (5). An example of an item from the 8-item secure scale was: I feel at ease in emotional relationships; from the 4-item fearful scale: I am afraid that I will be deceived when I get too close with others; from the 6-item preoccupied scale: I often wonder whether people like me; from the 3-item dismissing scale: It is important to me to be independent. The four attachment styles are not independent; as theory would predict secure style was correlated negatively with three other styles. Alpha coefficients for the secure scale were .70, .69, .67, .72; for the fearful scale were .84, .71, .79, .78; for the preoccupied scale were .83, .74, .71, .82; for the dismissing scale were .53, .60, .61, .59, .57 in the Dutch, Polish, Russian, and Hungarian samples, respectively. Parenting styles were assessed with the Parental Authority Questionnaire Revised (Reitman, Rhode, Hupp, & Altobello, 2002), based on an earlier

CHAPTER 5

71

version of PAQ (Buri, 1991) consisting of scales representing the three parenting styles proposed by Baumrind (1991): authoritarian, authoritative and permissive. Previous studies indicated a satisfactory validity of this measure (Buri, Louiselle, Misukanis, & Mueller, 1988; Buri, 1991). For all scales, a 5-point answering scale was used, ranging from strongly disagree (1) to strongly agree (5). Similarly to the ASQ, respondents were not classified into one parenting category, but received scores on all three parenting scales. Confirmatory factor analysis was performed on the original 30-item PAQ-R in the four national samples of mothers and in the four national samples of fathers. Fourteen items that formed three robust factors across all samples (see Table 5-1) were retained in the final version of the questionnaire. An example of an item from a 4-item authoritarian scale is When I ask my children to do something, I expect it to be done immediately and without questions, from a 5-item authoritative scale is Once family rules have been made, I discuss the reasons for the rules with my children, from a 5-item permissive scale is Children need to be free to make their own decisions about activities, even if this disagrees with what a parent might want to do. Alpha coefficients for the authoritarian scale were .63, .72, .67, .76; for the authoritative scale were .69, .73, .76, .75; for the permissive scale were .78, .65, .64, .60 in the Dutch, Polish, Russian and Hungarian samples respectively. The psychological health of parents was measured with a 9-item scale from the RAND 36-item Health Survey (RAND - Health Sciences Program, 1992; sample item: How much of the time during the past four weeks have you been a very nervous person?) and Satisfaction With Life with the 5-item Scale (SWLS) by Diener, Emmons, Larsen and Griffin (1985; sample item: I am satisfied with my life). Both measures had a 5-point answering scale ranging from strongly disagree (1) to strongly agree (5). The social functioning of parents was measured with a shortened version of the Social Support List Interactions (SSL-I) by Van Sonderen (1993). The 9-item SSL-I used in the present study was derived from a factor analysis on the original list of 64 items (Van Oudenhoven & Van der Zee, 2002). A sample item was: Does it ever happen to you that people are affectionate towards you? Alpha coefficients for these scales ranged from .73 to .90 in the samples under study.

ATTACHMENT AND PARENTING STYLES IN EASTERN AND WESTERN EUROPE

72

Results Preliminary analyses Visual inspection of mean plots of scores on items used to measure attachment and parenting styles indicated that, regardless of the content, in 53% of the items the mean scores of Polish respondents were higher than those of respondents from the other three samples, whereas the mean scores of Dutch respondents were lower than those of the respondents from the other three samples in the case of 57% of the items. The mean scores of respondents from the Hungarian and Russian samples were between those of the Polish and Dutch respondents in 54% and 58% of the cases. In addition, ANOVA performed on items from the ASQ revealed significant differences between national samples in: 50% of items in the sample of children, 45% in the sample of mothers and 46% in the sample of fathers. ANOVA on items from the PAQ showed significant differences in scores between national samples in 35% of the items in the sample of mothers, and 28% in the sample of fathers. This observation suggests response tendencies: regardless of the contents of items Polish respondents used higher rates on the answering scale, Dutch respondents used lower rates on the answering scales, and Russian and Hungarian respondents - middle rates on the answering scale. As the analysis in the next section was carried out on the two groups Western Europeans (the Dutch sample) and Eastern Europeans (the Polish, Russian and Hungarian samples, jointly), we further checked for response tendencies in these two groups by comparing correlations obtained in these two groups between the two opposite concepts - secure and fearful attachment styles (see: Chapter 6). The correlation between secure and fearful attachment was .34, p< .001 in Eastern European group and -.51, p< .001 in Western European group. The difference between correlations was significant statistically (p< .01), which, again, suggested that Eastern Europeans used more acquiescent response styles. In order to account for these response tendencies we carried out analysis on the data standardized within samples (Fischer, 2004), and on the data standardized across all samples. However, in both cases, the results were similar and statistically significant; therefore, for the sake of brevity, we report below only the results on the data standardized across samples. Confirmatory Factor Analysis (CFA) was performed on responses to the Attachment Styles Questionnaire of children, mothers and fathers from each country separately, and on the responses of mothers and fathers from each country to the Parental Authority Questionnaire. In addition, CFA was performed on the merged samples of children, mothers and fathers from the four countries (see Table1). The results of CFA indicated satisfactory replicability of factors in the ASQ and the PAQ across all the samples under

CHAPTER 5

73

study, thus showing satisfactory structural equivalence of the measures of attachment and parenting styles. In order to further examine the construct equivalence, we checked if intercorrelations of the scales of the ASQ remain similar across studied samples (lack of similarities in correlation patterns would suggest low construct equivalence). In general, we found consistent patterns of internal correlations between secure - fearful (ranging from -.26 to -.59, p <.001), suggesting satisfactory cross-cultural equivalence of these constructs. Less consistent internal correlations were found for preoccupied and dismissing, which were probably due to the generally lower validity of the scale measuring a dismissing attachment. We also examined the patterns of positive correlations with related measures and negative correlations with measures of the opposite constructs, to check for convergent and discriminant validity respectively (Van de Vijver & Tanzer, 2004). Positive correlations ranging from .10 to .44, p<.01 between secure attachment and Social Support, Psychological health and Satisfaction with Life were found across all samples. Consistently across all samples, fearful and preoccupied attachment styles were correlated negatively (in the range of -.01 to -.44) with Social support, Psychological health and Satisfaction with life. A less consistent pattern of correlation was found for dismissing attachment. In sum, these results suggest satisfactory construct equivalence and validity for measures of secure, fearful and preoccupied attachment styles and a less satisfactory one for dismissing attachment. Differences in Attachment in Eastern and Western Europe In order to test the hypothesis that Eastern Europeans will score higher on preoccupied attachment than Western Europeans [Hypothesis 1], and that the magnitude of the differences between Eastern and Western European respondents with respect to attachment styles will be smaller in the adolescent respondents as compared to middle-life respondents [Hypothesis 2] we carried out Univariate Analysis of Variance (Toothaker, 1993), in which four groups were compared: Eastern European parents (N = 852), Eastern European adolescents (N = 553), Western European parents (N = 151) and Western European adolescents (N = 99). The analysis was performed for each attachment style separately, and post hoc Scheffe test was used as a robust procedure for the comparison of groups with different sample sizes (Keppel, 1991). The results revealed significant overall differences in preoccupied attachment between Eastern and Western European parents, and between Eastern and Western European adolescents F(3, 1470) = 4.65, p < .01, partial 2= .01. In the post hoc pair-wise comparison the mean difference in preoccupied attachment estimated based on marginal means was .14 between Eastern and Western parents, and .18 between Eastern and Western adolescents. In both cases, as predicted,

ATTACHMENT AND PARENTING STYLES IN EASTERN AND WESTERN EUROPE

74

Eastern Europeans scored higher, but contrary to our prediction, the magnitude of the differences was higher in the adolescent sample. In a post hoc pair-wise comparison of Eastern and Western European adolescents on secure attachment, we found that the differences between them were not significant, while the differences between Eastern and Western parents were significant (Eastern European parents scoring .23 higher) F(3, 1443) = 16.54, p < .001, partial 2= .03. Neither overall, nor post hoc differences in fearful attachment were significant between either sample F(3, 1484) = 2.28, p =.08, partial 2= .00. In a post hoc pair-wise comparison of Eastern and Western European adolescents on dismissing attachment, we again found that the differences between them are not significant, while the differences between Eastern and Western European parents again were significant (Eastern European parents scoring .06 lower) F(3, 1489) = 3.16, p < .05, partial 2= .01. To further test the hypothesis about Western and Eastern parents differing more with respect to attachment styles than Western and Eastern adolescents [Hypothesis 2] we carried out two ANOVAs, first comparing parents and next comparing adolescents from the East and West. In each comparison we additionally tested statistical power to check if the differences in group size might have influenced the results. In the first ANOVA we found that Eastern parents scored 0.24 higher on secure attachment than Western parents F(1,892) = 7.63, p < .01 (statistical power 0.94). Eastern parents scored also 0.12 higher on preoccupied attachment than Western parents F(1,907) = 5.63, p < .05 (statistical power 0.70). In the second ANOVA we found that Eastern adolescents scored 0.14 higher on secure attachment than their Western peers F(1,552) = 14.76, p < .001 (statistical power 1.0). Eastern adolescents scored 0.16 higher on preoccupied attachment than Western adolescents F(1,573) = 5.96, p < .05 (statistical power 0.79). Thus, again we found that parents differed more on secure, while adolescents on preoccupied attachment, and the sample sizes did not affect the results, as indicators of statistical power were similar when comparing parents and adolescents.

Table 5-1 Summary of Confirmatory Factor Analyses on the Attachment Styles Questionnaire and Parenting Authority Questionnaire in Samples of Children, Mothers and Fathers from the Four Countries Separately, and for a Merged Sample.
Attachment styles questionnaire (children) Attachment styles questionnaire (mothers) Attachment styles questionnaire (fathers) Parental Authority Questionnaire (mothers) Parental Authority Questionnaire (fathers) Sample: Dutch Polish Russian Hungarian Merged Dutch Polish Russian Hungarian Merged Dutch Polish Russian Hungarian Merged Dutch Polish Russian Hungarian Merged Dutch Polish Russian Hungarian Merged N 99 186 146 221 652 89 172 135 198 594 62 116 92 139 409 89 172 135 198 594 62 116 92 139 409 2 222.18 387.83 306.75 346.33 768.42 371.26 413.03 398.46 488.87 858.96 365.20 372.10 384.20 590.70 988.47 116.48 150.75 99.15 144.40 237.43 116.76 208.95 132.40 207.58 269.76 df 183 183 183 183 183 183 183 183 183 183 183 183 183 183 183 74 74 74 74 74 74 74 74 74 74 p .03 .04 .06 .12 .05 .01 .04 .06 .00 .14 .10 .08 .50 .20 .08 01 .07 .03 .06 .10 .01 .08 .04 .20 .09 GFI .82 .83 .77 .87 .89 .84 .82 .72 .83 .88 .73 .75 .84 .80 .86 .85 .89 .87 .91 .95 .85 .86 .84 .88 .94 AGFI .77 .78 .71 .84 .86 .76 .77 .65 .78 .85 .66 .75 .71 .74 .83 .79 .85 .82 .88 .92 .79 .80 .77 .83 .91 RMSEA .05 .08 .08 .06 .07 .08 .08 .10 .08 .07 .09 .09 .09 .09 .08 .07 .07 .05 .06 .06 .07 .08 .08 .07 .06 NFI .69 .63 .50 .68 .73 .58 .55 .48 .57 .69 .56 .58 .53 .49 .62 .66 .63 .68 .74 .83 .52 .50 .49 .61 .72 CFI .89 .76 .67 .81 .79 .65 .68 .61 .67 .74 .69 .67 .65 .57 .66 .82 .74 .86 .84 .87 .72 .58 .64 .70 .78 IFI .89 .77 .69 .82 .79 .65 .69 .62 .68 .74 .70 .71 .75 .58 .67 .83 .75 .87 .85 .88 .75 .60 .67 .71 .78

Note. GFI = goodness-of-fit index; AGFI = adjusted goodness-of-fit index; RMSEA = root-mean-square error of approximation; NFI = normed fit index; CFI = comparative fit index; IFI = incremental fit index.

ATTACHMENT AND PARENTING STYLES IN EASTERN AND WESTERN EUROPE

76

In sum, support was found for Hypothesis 1 that Eastern Europeans score higher on preoccupied attachment than Western Europeans, whereas little evidence was found supporting Hypothesis 2 that the magnitude of the differences between Eastern and Western European respondents with respect to attachment styles will be smaller in the adolescent respondents as compared to middle-life respondents. The results of the Univariate Analysis of Variance suggested that Eastern and Western parents differed more than adolescents with respect to secure and dismissing attachment styles. Adolescents from the West and East differed more than their parents with respect to preoccupied attachment. No differences were found with respect to fearful attachment in either sample. The magnitude of the main effect for secure and dismissing attachment, as indexed by partial eta-square (2), although significant, was rather small. Additional ANOVA confirmed the findings with respect to secure and preoccupied attachment styles, but not with respect to dismissing and fearful attachment styles. Attachment Styles, Culture and Parental Characteristics In the introduction we hypothesised that the direct effect of culture on the preoccupied attachment of adolescents will be stronger than on secure, fearful and dismissing attachment styles [Hypothesis 3]. Further, we hypothesised that a secure parental attachment style is a positive predictor of a secure attachment style of the child; a fearful attachment of a parent is a positive predictor of a fearful attachment style of the child, a preoccupied attachment of a parent is a positive predictor of a preoccupied attachment style of the child; and a dismissing attachment style of a parent is a positive predictor of a dismissing attachment of the child [Hypothesis 4]. We also expected that authoritative parenting is a positive predictor of a secure attachment style of children, while authoritarian and permissive parenting styles are predictors of insecure attachment styles (fearful, preoccupied and dismissing) [Hypothesis 5] and that parental psychological health is a positive predictor of a secure and a negative one of insecure attachment styles (fearful, preoccupied and dismissing) of the child [Hypothesis 6]. To test Hypotheses 3-6 we used an Analysis of Covariance (ANCOVA) in which culture (coded as a dummy variable) was entered as a random factor (Wildt & Ahtola, 1978), with characteristics of parents as predictors and attachment styles of adolescents one by one, individually, as dependent variables. This analysis is an extension of the analysis of variance carried out in the previous section. Since the correlations between responses of mothers and fathers were significant for each variable, apart from dismissing attachment and ranged from .16 to .50, in order to avoid a collinearity problem we separately carried out an analysis when

CHAPTER 5

77

characteristics of mothers were entered as predictors, and when characteristics of fathers were entered as predictors. In sum, we report below eight analyses: two for each attachment style: one analysis utilizes responses of mothers on the ASQ, PAQ, RAND, SWLS, and SSL-I, the other one utilizes responses of fathers on these scales as predictors of adolescents responses on the scales of the ASQ. The interaction effects of culture and each parental characteristic were also tested in these analyses. Contrary to expectations we did not find a significant main effect of culture for preoccupied attachment when using responses of mothers nor when using responses of fathers. Neither have we found a significant interaction effect between culture and parental characteristics. This suggests that the difference between Eastern and Western Europeans we found in the analysis of variance in the previous section exists due to the effect of parental characteristics. When controlling for parental characteristics, the differences between Eastern and Western Europeans were not significant. The analysis showed that preoccupied attachments of adolescents were negatively predicted by their mothers satisfaction with life B = -.14, p = .04, and positively predicted by preoccupied the attachment of the mother B = .20, p = .01, mothers permissive parenting B = .16, p = .02, and the preoccupied attachment of the father B = .27, p = .01. The B values indicate that, keeping all variables equal, if the predicting variable increases by one unit, the predicted variable increases with B value of the unit. We found a significant main effect of culture when predicting secure attachment with mothers F(1, 395) = 7.60, p < .01, partial 2= .02 and fathers F(1, 290) = 5.26, p < .05, partial 2= .02 characteristics. The significant difference between marginal means indicated that Eastern European adolescents scored higher than their Western peers in the range of .31, when fathers characteristics are controlled and .41 when mothers characteristics are controlled. The above result indicates that beyond the influence parental characteristics exert on the secure attachment of their adolescent children, Eastern European culture seems to enhance the secure attachment of adolescents. Secure attachment amongst adolescents were negatively predicted by a fearful attachment of fathers B = -.20, p = .05, and positively predicted by the authoritative parenting of mothers B = .15 p = .003, mothers psychological heath B = .11, p = .02, and how mothers perceived the social support they receive from other people B = .18, p = .007. A significant interaction effect between culture and authoritative parenting was found when predicting the fearful attachment of adolescents with fathers characteristics F(1, 301) = 5.56, p < .05, partial 2= .02. Authoritative parenting of fathers from East Europe (interaction: culture x authoritative parenting) predicted negatively a fearful attachment of children B = -.22, p = .02. When fearful attachment was predicted by mothers characteristics,

ATTACHMENT AND PARENTING STYLES IN EASTERN AND WESTERN EUROPE

78

fearful attachment of the mother F(1, 411) = 7.70, p < .01, partial 2= .02 has an effect on the fearful attachment of the child. Neither significant main, nor interaction effects were found with respect to the dismissive attachment of the adolescent by mother and fathers characteristics respectively. The estimated parameters showed that a preoccupied attachment of the mother was a negative predictor of the dismissive attachment of a child B = -.06, p = .02. In the analyses, in which mothers characteristics were entered as predictors, R squared reached: .13, .08, .06, .06 and adjusted R squared reached: .08, .03, .02, .01, for secure, preoccupied, fearful and dismissing attachment style respectively. In the analyses in which fathers characteristics were entered as predictors, R squared was: .11, .07, .08, .05 and adjusted R squared was: .04, .01, .02, .02, for secure, preoccupied, fearful and dismissing attachment style respectively. To sum up, the results are not in accordance with Hypothesis 3 we did not find a stronger direct effect of culture on preoccupied attachment as compared to other attachment styles. On the contrary: the results suggest that the difference between Eastern and Western European adolescents with respect of their preoccupied attachment style exists due to differential parental practices. Surprisingly, we found that culture may exert a direct influence on the secure attachment styles of adolescents, beyond the influence of parental characteristics. Further, we found support that secure, fearful and preoccupied attachment styles, and to some extent - dismissing attachment styles of parents are predictors of an analogous attachment style of their children [Hypothesis 4]. Evidence was found that authoritative parenting has a positive effect on a secure attachment style of children, while permissive parenting styles - on insecure attachment styles, however no evidence was found that authoritarian parenting would predict insecure attachment [Hypothesis 5]. The results supported our expectations that parental psychological health is a positive predictor of a secure and a negative one of insecure attachment styles (fearful, preoccupied and dismissing) in a child [Hypothesis 6]. Discussion In the present study we were interested in, whether the long-lasting separation between Eastern and Western Europe may have resulted in differences in the psychological characteristics of people from these regions. We focused on attachment styles, because attachment is defined as a construct shaped by the immediate (parental practices) and broader environment (culture). Therefore it is plausible that differences in culture will be reflected in differences in attachment styles.

CHAPTER 5

79

First, the present results indicate that Eastern Europeans have more preoccupied attachment than Western Europeans. This difference, as the findings suggest, exists mainly due to different parenting practices, more than due to the direct influence of the culture on an individual. The present results show that parental practices are of greater importance for child development than the broad social network. This challenges the assumption of the ecocultural theory (Weisner, 2005) proposing that the environment and parental practices are equally important for a development of the child. Second, the study addressed the question of whether the generation of Eastern and Western Europeans, which grow up in a more unified and internationalized world show more similarities in their attachment styles, than the generation of their parents. The findings suggest a negative answer to this question: there were no differences in either sample on the fearful and dismissing attachment, and the differences between adolescents from Eastern and Western Europe were smaller than the differences between their parents only with respect to secure attachment styles, but slightly bigger with respect to preoccupied attachment. A possible explanation of the latter might be that in puberty the worries as to whether one is liked and accepted by others, related to preoccupied attachment, are particularly strong. This tendency may be more salient in collectivistic, thus, Eastern European societies, than in more individualistic Western European societies, as Eastern cultures were found to induce self-evaluation, which is based on interconnectedness more than do Western cultures (Markus & Kitayama, 1991). In addition, we found that the attachment styles of parents predicted the attachment styles of their adolescent children, and the relation between them was, generally speaking, as we expected, and as described in the literature (e.g., Nair & Murray, 2005; Demo & Cox, 2000; Bowlby, 1984, Ainsworth, 1967): secure attachment of parents positively predicted a secure attachment style of the child and negatively insecure attachment styles of the child. Conversely, the insecure attachment of parents predicted negatively the secure attachment of the child and positively insecure attachment styles of the child. The results also suggest that an authoritative parenting style increases the chance that the child will develop a secure attachment style, which is in line with previous findings (Ainsworth et al., 1978; Karen, 1998). Surprisingly, the study failed to support the predicted effect of authoritarian parenting on the attachment of children. Possibly, children possess some kind of a recovery mechanism, which helps them to develop a secure attachment style even in the case of dysfunctional parenting. With respect to psychological health, the results were consistent with earlier findings (Gerlsma et al., 1997) that psychological health, regardless of cultural context, is an important factor for the child to develop a secure attachment.

ATTACHMENT AND PARENTING STYLES IN EASTERN AND WESTERN EUROPE

80

When controlling for parental characteristics, culture turned out to have the strongest effect on secure attachment, and not as expected, on preoccupied attachment. This result gives us a hint that Eastern European culture is more conducive for adolescents to develop secure attachment than Western European culture. Interestingly, in the comparisons of parents, Eastern Europeans also scored higher then their Western peers. It might be attributed to rapid economic growth and social progress, and the higher optimism of Eastern Europeans resulting from this. In accordance with earlier studies (Van IJzendoorn & Sagi, 1999; Schmitt et al., 2004) showing that a secure type of attachment is normative across diverse cultures, we also found that adolescents and their parents from all four samples scored higher on a secure than on insecure (fearful, preoccupied and dismissing) attachment styles. The present study has some limitations. First, there was a trace of differential response strategies among respondents from different countries; Polish respondents had a tendency to use higher, Dutch - lower, and Russian and Hungarian - middle categories of the answering scale. Although we found the same results in comparisons on the standardized data within samples, still we have to keep in mind the possibility that the outcome of the comparisons might have been influenced, to same extent, by the response styles. Our observation with regard to response strategies are in line with other studies which indicate that respondents from more collectivistic cultures show more acquiescent (positive) response styles (Lamm & Keller, 2007). Another limitation is related to the fact that in our fourfold comparison we had, on the one hand, adolescent respondents and, on the other, mid-life respondents. It must be noted that in adolescents attachment styles are not completely stable. This restrains the validity of the present findings, as it might well be that attachment styles measured in the same respondents will change in a few years. One should not forget, however, that since the political changes started 18 years ago, adolescent participants represent the first generation of Eastern Europeans born in a post Cold War Europe. Finally, it should be noted that respondents in the present study were recruited among the inhabitants of the biggest cities in the four countries; thus generalization to other non-metropolitan samples must be done cautiously.

CHAPTER 5

81

Chapter 6 Discussion and Implications

Summary of the main findings

Attachment: Conceptualisation and Measurement In a large number of studies, attachment styles are defined and measured as relation-specific phenomena (e.g., Schmitt et al. 2003, 2004). In this dissertation, attachment styles are defined as a trait describing the general way in which people relate to others. In Chapter 2, we found satisfactory reliability of three of the four dimensions of attachment styles proposed by the model of Bartholomew and Horowitz (1991) secure, fearful and preoccupied attachment. However, we did not find evidence for the postulated structure of the model which implies that the preoccupied attachment is defined by a positive model of others and a negative model of self. The data rather suggest that preoccupied individuals, like fearful individuals, have a negative model of others. The results in Chapter 2 also suggest that the fourth dimension proposed by the model dismissing attachment yielded two sub-scales: avoidance of close emotional ties and excessive self-sufficiency. These findings are in accordance with previous studies that suggested two distinguishable constructs: fearful and dismissive avoidance (Duggan & Brennan,1994). Attachment and Migration Numerous studies have demonstrated the role of demographic variables in immigrants adjustment (Ward, Bochner, & Furnham, 2001). The present research showed that attachment styles are better predictors of immigrants adjustment than demographic variables such as education, age or age at emigration (Chapter 3), and that the combination of the two attachment styles high secure and dismissing attachment may predispose people to migration (Chapter 4). Attachment styles seem to be relatively stable characteristics no evidence was found that they change under the influence of the host culture (Chapter 4). We also found that a secure attachment style is a better predictor of the well-being of immigrants than of non-immigrants (Chapter 4). Cultural Differences in Parenting and Attachment Styles In Chapter 5 we found that the wider culture of a society may exert an influence on the attachment styles of people. In accordance with previous

DISCUSSION AND CONCLUSIONS

84

studies (Schmitt et al., 2003, 2004), we found that respondents from Eastern Europe scored higher on preoccupied attachment than respondents from Western Europe. However, we did not find evidence that attachment styles in the generation of Western and Eastern Europeans growing up in a more unified and internationalized world show more similarities than attachment styles in the generation of their parents from Eastern and Western Europe. Moreover, we demonstrated that the attachment styles of parents, parenting styles and parental psychological health predicted the attachment styles of their adolescent children.

Implications and Conclusions

Implications for the Conceptualization of Attachment Styles In the literature on psychological development, attachment styles typically refer to how infants organize their attachment behaviour with respect to a particular caregiver (e.g., Ainsworth, Blehar, Waters & Wall, 1978). Accordingly, research stemming from a classical developmental attachment framework conceptualizes attachment as a relationship construct. In contrast, research stemming from individual differences defines attachment as a personality trait rather than as characteristics of relationships. Kobak (1994) argued that attachment theory is first and foremost a normative theory and that the conceptualisation of attachment as an internal personality structure rather than a quality of relationships is unfortunate (p. 42). His criticism is based on evidence that attachment styles show, even in the relatively brief period between 12 and 18 months, considerable instability and openness to environmental change. Furthermore, he points out that conceptualising and measuring attachment as a personality construct causes the attachment system to be abstracted from a relationship context, whereas classical infant research indicated that attachment quality may depend on the attachment figure. Reducing a persons internal models to a single personality style may, according to Kobak, ignore the complexity and relationship specific character of attachment ties. The present findings contradict this argumentation. First, we have seen that such a dramatic change in life as migration and the process of acculturation did not influence the attachment style of consecutive cohorts of immigrants (Chapter 3). In Chapter 5, we observed high and positive correlations between the general attachment styles of parents and their children. We also found correlations between the general attachment styles of children and parenting characteristics. These results indicate that parental characteristics may influence not only attachment in the dyad parent-child, but also the general attachment of

CHAPTER 6

85

children. Aside from the arguments brought forth by the present findings, it should be noted that patterns of interpersonal relationships show considerable consistency within individuals; people tend to use similar tactics in their interpersonal interactions with different interaction partners. Definitions of secure attachment (when conceptualised as a relation construct) include perceived support as a key component (Collins & Feeney, 2000, 2004). The positive correlations between secure attachment and perceived social support found in Chapter 3 indicate that secure attachment, defined as a general trait, is probably underscored by the same belief in a supporting and responsive other, as it is in the case of relationship-specific secure attachment. Nevertheless, the present author would not opt for abandoning the measurement of relation-specific attachment. Barry, Lakey and Orehek (2007) found, for example, that when measured independently, relation-specific and general attachments were differentially related to different concepts: relationship-specific attachment accounted for more variance in measures of affect, whereas general attachment style accounted for more variance in self-esteem. Livesley, Jackson and Schroeder (1992) in their factor-analytic study on a hundred personality inventories found independent dimensions referring to generalized insecure attachment and intimacy problems in close relationships. The present findings suggest that attachment, conceptualised and measured as a general trait, constitutes an addition to the personality framework, valid for describing individual differences in general sociability. It seems that people interact with their social environment while relying on a few close attachment relations, and that relation-specific attachment and general attachment are associated. Some evidence suggests, for example, that differences in interactions with close and not close friends are more pronounced for fearful individuals, as compared to those who have secure, dismissing or preoccupied attachment styles (Kafetsios & Nezlek, 2002). Further studies could elucidate the relation between relation specific and general attachment. Future studies could also bring more understanding of how working models are organised and how they shape later attachment (whether treated as a relationship concept or a general trait). Drawing Conclusions about Cultural Differences The average scores on attachment styles in the Polish sample of nonemigrant respondents in Chapter 4 and in the samples of Polish parents in Chapter 5 differed considerably. These differences can be attributed to demographic variables: respondents in Chapter 4 come from a region of Poland with a high (30% in 2005) unemployment rate, which entails a much lower socio-economic status of those respondents when compared to the respondents living in the capital city who participated in our survey in

DISCUSSION AND CONCLUSIONS

86

Chapter 5. These results are in accordance with previous studies, which showed that people with lower socio-economic status had an anxiousambivalent attachment rather than a secure or avoidant attachment (Shaver & Hazan, 1994). Apart from the differences in attachment styles which might have been induced by socioeconomic status, age-related differences also played a role. For example, Polish teenagers in Chapter 5 had significantly higher secure attachment styles than their parents. The foregoing example shows that differences in mean scores often occur due to demographic characteristics and not culture per se. Thus, in cross-cultural research which aspires to inspect differences between nations, big and representative samples with respect to age, country regions, and social strata should be used. As the present research did not use such samples, prudence is especially warranted before generalizing from the findings. Some Methodological Considerations The model of Bartholomew and Horowitz (1991) posits four attachment types: secure, fearful, preoccupied and dismissing (see introduction to Chapter 1). Secure and fearful types are conceptual opposites, so are preoccupied and dismissing. The measurement developed based on this model the Attachment Style Questionnaire developed by Van Oudenhoven et al. (2003), used in the present dissertation combines theory driven attachment prototypes with an empirically based dimensional approach. While in the typological approach it is necessary that categories in opposition, such as secure and fearful, encompass all the individuals subjected to the categorization, in the dimensional approach, where scales are derived from factor analysis, having categories in opposition may become methodologically troublesome. Items representing the secure factor also loaded highly on the negatively correlated fearful factor. High secondary loadings cause items to switch factors, which decreases factor replicability when a study is conducted in small samples. Assessment of opposite, negatively correlated constructs, such as secure and fearful attachment, may however also be useful in cross-cultural studies. Below we give an example of its usefulness. It is common practice that, when accounting for cross-cultural differences in acquiescent response style, researchers use positively and negatively-formulated items and subtract the scores respondents obtain on positively and negatively formulated items (Schmitt & Allik, 2005). If the number obtained in such a way, is close to zero, it indicates that the response style was non-acquiescent. It is known, however, that statements with negations may entail conceptually irrelevant variance caused by the different language abilities of respondents (Holden & Fekken, 1990) or the ambiguity of an item (Angleitner, John, & Lohr, 1986). Moreover, it has been found

CHAPTER 6

87

that respondents are generally less likely to endorse negatively phrased items (Schmitt & Allik, 2005). Another problem with negative formulation is related to syntax equivalence in different languages. For example, an English sentence with one negation I never go to the cinema is translated into Polish by Nigdy nie chodz do kina, a sentence with two negations. In addition, a Likert-type answering scale with categories agree-disagree will introduce another negation. When English respondents are faced with two negations (one comprised in the item and one in the answering scale), Polish respondents will have to process three negations. Such a lack of syntax equivalence may evoke differential item functioning and add culture-specific variance that is not relevant to the concept under assessment. Consequently, accounting for response bias using negatively formulated items may not be a reliable method. Negative items could be omitted in favour of opposite categories such as fearful and secure attachment, assessed with simple and positively formulated items, and acquiescent response tendencies could be appraised in diverse samples based on differences in correlations between two opposite concepts. For example, in Chapter 5 we appraised response tendencies comparing correlations between secure and fearful attachment computed in the sample of Eastern and Western Europeans. We found that Eastern Europeans have more acquiescent response tendencies than Western Europeans, which was congruent with the results of item analysis (see: Chapter 5).

Practical Implications The present findings have a number of practical implications; I will start from implications for psychological counselling for immigrants. Maladaptive behaviours resulting from a lack of secure attachment may be a source of suffering that can be soothed in psychotherapy. Horowitz, Rosenberg and Bartholomew (1993) reported that patients with attachment-related problems such as excessive social dependency, vulnerability to depression and interpersonal loss showed satisfactory improvement in a brief dynamic psychotherapy. Therefore, in clinical practice, it would seem helpful to treat some immigrants problems from an attachment perspective. Some conclusions from the present studies can also be applied to migration policies. In Chapter 3 we demonstrated that secure attachment was more strongly and positively associated with immigrants adjustment than any of the demographic variables. This suggests that the adjustment of immigrants is less related to immigrants education, or other demographic characteristics, but depends more on an immigrants attachment style. Paraphrased differently the maladjustment of immigrants results from the lack of a secure attachment. Thus, policy-makers, who try to prevent the

DISCUSSION AND CONCLUSIONS

88

maladjustment of immigrants, should, apart from developing policies that focus on immigrants education, develop policies that focus on interventions reducing factors causing insecure attachment. An example of a primary factor is dysfunctions in immigrants families which entail insecure attachment in children. Another example of secondary factors is social pressure at schools on children of foreign descent to give up their native identity. Perhaps due to isolation from mainstream society, immigrant parents are less likely than Dutch parents to seek external help for their problems. Psychological counselling provided at schools, targeted especially at immigrant parents, could reduce their isolation and help them to create conducive conditions for their children to develop secure attachment and be better adjusted in the future. The results presented in Chapter 4 showed that certain attachment patterns (the combination of high secure and dismissing attachment) predispose people for migration, and that secure attachment helps them to adjust after migration. One can easily imagine that, when a company sends an employee abroad, it is important to know the capacities of that person to cope with cultural differences and the stress resulting from being away from home. In such cases, assessment of attachment styles could help to select the best candidate. The findings described in Chapter 5 indicate that the globalisation of culture can lead to less cultural diversity in terms of psychological traits. It might be perceived as a positive phenomenon because less cultural differences will make us more similar and enhance communication. Some would argue, however, that less cultural diversity is bound to impoverish global culture; by analogy to theories of biodiversity which claim that the more species, the better for the planet, one could posit that the more cultures, the better for our civilization. Cultural diversity helps us to see problems from different angles, thus facilitating progress in society. For example, if members of one culture perceive a problem as unsolvable, someone with other cultural experience may cope with it adequately. But, there again, in order to benefit from cultural diversity, an open-mindedness coming from secure attachment is necessary. Strengths and Limitations of the Present Research The strengths and weaknesses of the present research come side-by-side, and therefore will be discussed together. First, in the present research we used data from diverse cultural samples, which strengthens the validity and generalizability of our findings. However, the culture of a respondent was treated solely as a selection criterion. No culture-related or contextual variables were included in the analyses. This limits the inferences we could make about the role of some

CHAPTER 6

89

aspects of the native cultures in the adjustment of immigrants, and the inferences about the nature of cultural differences. Van de Vijver and Leung (1997) suggest that including contextual and cultural dimensions could improve exploration of cross-cultural differences. Adding contextual variables in future studies including a bigger number of cultural samples and multilevel design could clarify which characteristics of national cultures explain differences in attachment styles. Second, our measures of the host and native identity of immigrants were empirically derived in factor analysis based on the data from the three samples of immigrants. That represents more methodologically rigorous assessment than one-item assessment or scenario-approach used in many other studies. At the same time, a few limitations follow from such an assessment. For example, we included items which loaded highly on the native and host factors across all samples under study, and emic (culturally specific) items were omitted. Also such an assessment entailed the assumption that the process of developing host identity and maintaining native identity are independent, whereas recent studies (Arends-Tth & Van de Vijver, 2004; 2007) suggest that they are related. In the next section, we propose possible measures in addition to independent dimensions related to native and host cultures. Third, we collected data from anonymous, voluntary respondents; therefore, we can assume they gave honest answers to our questions. On the other hand, such a sampling procedure may have caused an underrepresentation of certain attachment styles, mainly fearful and dismissing. Also, using a snowball method may have caused an undesirable homogeneity of our samples in terms of demographic characteristics, such as age, education and socio-economic status. Finally, two other limitations relate to the cross-sectional design of the study and to measurement that relied only on self-report. The crosssectional design of the study makes it impossible to say if, for example, cultural adjustment is the cause or rather the effect of psychological adjustment. Future immigration studies incorporating longitudinal design may examine the causal relationship between cultural and psychological adjustment, and could be more suitable for testing if the host cultures or the mere fact of migration have influence on the attachment styles of immigrants (in Chapter 3).

DISCUSSION AND CONCLUSIONS

90

Future directions

Psychology and Cuisine The majority of recent studies on acculturation have employed a bidimensional model of acculturation. In this model, maintaining the original and acquiring a new culture are perceived as independent, non-exclusive processes (see Chapter 1). Such an approach implies independent assessment of the expressions of the native and host identity of immigrants. We used such an independent assessment in Chapter 3. It can be inferred, however, from studies on identity conflict in immigrants (Ward & Searle, 1991) and cultural distance and adjustment (Ward, Bochner, & Furnham, 2001) that perceived discrepancies between host and native culture may play a more important role in socio-cultural adjustment than identification with either culture. The study of Arends-Tth and Van de Vijver (2004) on implicit acculturation theories also suggests that the process of maintaining the native and developing the host identity are not completely independent. In a similar vein, studies on bilingualism indicate the loss of the first language by a minority group, when a second (dominant) language is mastered (Lambert, 1978). Possibly, the relation between the two processes depends on whether identity is defined through the differences or similarities between the host and native cultures, through the perceived differences and similarities between cultures, or through the need for identity coherence (Berzonsky, Macek, & Nurmi, 2003). When an immigrant perceives the differences between his or her native culture and the host culture as big, it may motivate him or her to renounce a host identity in order to maintain identity cohesion. Put differently: apart from items such as I like Russian cuisine, which we used to measure native identity, in the future we ought to include items such as, To my taste, Russian and Dutch cuisine are very different. More Questions than Answers The present research gave answers to a number of questions, and at the same time gave rise to many more. Given that a secure attachment style was more important for the adjustment of emigrants than non-emigrants, it seems plausible to suppose that it is more important for the adjustment of immigrants whose culture of origin is very different from the host culture, than for immigrants coming from a similar culture. Further studies could explain the role of the interaction of attachment styles and cultural distance in immigrants adjustment. Mikulincer and Shaver (2001) demonstrated that priming secure attachment evokes a positive reaction towards out-group members. This result suggests that secure individuals might be less prone to experience an

CHAPTER 6

91

identity threat in a situation of contact with out-group members, and that they may define their identity as based more on similarities than on differences with other people. This issue also needs further research which would address the question of whether there is any relation between attachment styles and a susceptibility to identity conflict, or if attachment styles may predict reactions to identity threat. It would be also interesting to investigate a more general matter: if there is any relation between attachment styles and the organization of identity? Bartholomew and Horowitzs (1991) model of attachment posits that secure attachment is defined by a positive model of others and of the self. This, in turn, raises a question: what is, if any, the relation between attachment styles and stereotypes and metastereotypes? Individuals high in secure attachment, due to having a positive model of the self and others, might be more prone to activate positive metastereotypes than non-secure individuals with negative model of self and others. All these questions lead to the conclusion that migration studies would benefit from integrating two research traditions: individual differences and social cognition. Future studies could elucidate the role of the interplay between attachment, identity and meta-stereotypes in immigrants adjustment. Attachment and Bungee Jumping At more or less the same time when Ainsworth and her colleagues (1978) were conducting their classical experiments on attachment behaviour in mother-infant dyads (see Chapter 1), Fulker, Eysenck and Zuckerman (1980) found in their famous twin studies a high, 58%, heritability of sensation seeking. From that point onwards, research conducted within the attachment and temperamental frameworks grew apart, only occasionally acknowledging the existence of the other approach. Attachment researchers attributed infants behaviour almost entirely to the quality of care-giving, whereas researchers focused on temperament and personality attributed it to genetics. More recent studies, which eventually started to examine the connection between personality and attachment, examined the predictive value of personality and attachment with regard to other variables (Bakker, Van Oudenhoven, & Van der Zee, 2004; Kurdek, 2002; Shaver & Brennan, 1992). These studies seemed to focus on the issue of whether measures of attachment are distinct from the dimensions that personality researchers measured over years. The question as to whether and how attachment experience and temperamental traits interact to influence behaviour, dynamics of relationships and well-being has been rarely addressed. In the study of Thornqust, Zuckerman and Exline (1991), high sensation seeking was found to be related to a strong independence tendency and low satisfaction with a relationship tendency. Remarkably, a similar effect was

DISCUSSION AND CONCLUSIONS

92

found for a dismissing attachment (Shaver & Brennan, 1992). It is plausible that high sensation seeking itself may not be detrimental for close relations, but that the interaction of it with a dismissive attachment could be. Conversely, the interaction of a secure attachment and sensation seeking may annul the detrimental influence of the latter. In Chapter 4 we found that immigrants scored higher on secure and dismissing attachment than nonemigrants. Drawing on these results and on the findings of the classical experiment of Ainsworth, in which secure and dismissing individuals demonstrated high readiness to explore new environments and to approach unfamiliar others (Ainsworth, Blehar, Waters, & Wall, 1978), we speculated in that chapter that novelty seeking may underscore the high exploratory activity of secure and dismissing individuals. Conceivably, the combination of secure attachment style and sensation seeking on the one hand, and dismissing attachment style and sensation seeking on the other, may differentially influence psychological well-being. It could be inferred from the results of Chapter 4, in which secure attachment, unlike dismissive attachment, was found to be a positive predictor of well-being, that the combination of secure attachment and high novelty seeking is especially favourable for psychological well-being. In order to test the relation between dismissive (avoidant) and secure attachment and novelty seeking and their influence for well-being, a study should be designed to separate variance into temperamental and attachment components. Without being overly serious, one could imagine a study on the well-being of bungee jumpers: surely, they would score high on sensation seeking and endorse secure attachment! In the end, it is hard to escape the conclusion that the blossoming of research on attachment created an opportunity to move from overly deterministic models of personality to study the interaction of genetic predispositions and the environment. Combining the two perspectives seems to be a particularly promising avenue for future work.

Summary Political changes after 1989 and resulting from the European Union enlargement increased emigration from Eastern to Western Europe. In the introductory Chapter 1 the data about migration from the Eastern to Western Europe are presented, as well as the data about cultural, social and economic differences between Russia, Hungary, the Netherlands and Poland. In this chapter we also introduce A theoretical basis of acculturation and attachment theory. In Chapter 2 we focused on the replicability of factors in the Attachment Style Questionnaire (ASQ). Although the Bartholomew and Horowitz model has been used as a framework in many studies, there is a lack of clear evidence whether or not the model is supported by factor analysis. In the original Dutch version of the ASQ, four oblique factors were found, which corresponded to the four attachment types postulated by the model. We reasoned that if factors found in the ASQ are replicable in other samples, this would constitute support for the model. We applied the confirmatory Multiple Group Method, and the exploratory procedure of Simultaneous Component Analysis (SCA) to examine the replicability of factor structure. In both MGM and SCA we found three clear factors, similar to secure, fearful and preoccupied in the original version of the ASQ in all studied samples. However, we did not find evidence for the postulated structure of the model, which implies that the preoccupied attachment is a conceptual opposite of the dismissive attachment. The data rather suggest that preoccupied individuals, like fearful individuals, have a negative model of others. Moreover, the dismissing factor yielded to sub-dimensions: avoidance of close ties and excessive self-sufficiency. In Chapter 3 we examined the relationship between adult attachment styles and psychological and sociocultural adjustment of Polish, Russian, and Hungarian immigrants to the Dutch society. In addition, we also examined the relationship between demographic factors and adjustment and we compared the predictive value of attachment styles and demographic factors for immigrants adjustment. The Attachment Style Questionnaire was used to assess respondents attachment. Psychological adjustment was measured with the Psychological Health Scale and the Satisfaction With Life Scale. Sociocultural adjustment was measured with the Social Support List Interactions scale. Two scales for measuring identification and contact with the native and with the Dutch culture were developed and used as indicators of cultural adjustment. We found relations between attachment styles and psychological and sociocultural adjustment. Secure attachment was positively related to psychological and sociocultural adjustment, fearful attachment was negatively associated with psychological adjustment, and

SUMMARY

95

more negatively with identification with the Dutch culture than with identification with the native culture. Preoccupied attachment was negatively related to psychological adjustment and to identification with the Dutch culture. Dismissing attachment was weakly negatively related to sociocultural adjustment. Correlation patterns across the three immigrants samples indicate that dismissing individuals remain relatively indifferent towards their native and the Dutch culture. Regarding demographic factors we found that education and age at immigration were positively associated with psychological and sociocultural adjustment, and length of residence appeared to be positively related to sociocultural adjustment. In general, demographic factors showed a stronger association with sociocultural than with psychological adjustment. Regression analysis revealed that attachment styles were better predictors of immigrants' psychological and sociocultural adjustment than demographic factors education, age at immigration and length of residence. In Chapter 4 we examined empirical evidence for a concept of a migrant personality using the attachment framework. We compared Polish emigrants in the Netherlands to the Poles living in Poland measuring their secure and dismissing attachment styles, while controlling for age, gender, and education in both samples. The results showed that emigrants are more secure and more dismissing than their fellow countrymen. Furthermore, we examined, whether attachment styles might be influenced by the host culture. For that purpose, we subdivided the sample of emigrants into eleven cohorts (each cohort differentiated from the other by two more years of residence) and we compared the mean scores on secure and dismissing attachment styles in consecutive cohorts of emigrants. We observed that the mean scores on secure and dismissing attachment remained similar over cohorts of emigrants. Moreover, the differences between emigrants and non-emigrants were of similar magnitude across cohorts. This suggests that the attachment styles of emigrants are not influenced by the host culture, and that the observed differences in attachment styles between emigrants and nonemigrants last over time, and presumably, existed prior to emigration. Finally, we found that secure attachment was a stronger predictor of psychological health in the emigrant sample than in the in-country sample. In sum, evidence for a migrant personality has been found. This personality, as the results suggest, is functional for emigrants. In Chapter 5 we investigated the differences in attachment styles between Eastern and Western Europeans. As predicted based on earlier studies, we found that Eastern Europeans have more preoccupied attachment than Western Europeans. This difference, like the results suggest, exist mainly due to the differential parenting practices, more than due to the direct influence of the culture on an individual. Also we addressed the question of

96

weather the generation of Eastern and Western Europeans, which grow up in more unified and internationalized world, show more similarities in their attachment styles, than the generation of their parents. The results imply a negative answer to this question. In addition, we found that the secure attachment of parents positively predicted a secure attachment style of the child and negatively insecure attachment styles of the child. Conversely, the insecure attachment of parents predicted negatively the secure attachment of the child and positively insecure attachment styles of the child. The results also showed that an authoritative parenting style increases the chance that the child will develop a secure attachment style. With respect to psychological health, the results were consistent with earlier findings that psychological health, regardless of cultural context, is an important factor for the child to develop a secure attachment. In Chapter 6, overviewing the findings, we highlighted conceptual, methodological and practical implications of the present findings. We propose that the findings may be useful in policy-making and counselling targeted to immigrants. After discussing the major weaknesses and strengths of this research, the direction of the future studies have been proposed, which would combine the research on individual differences and identity of immigrants and which would investigate the relation between personality, temperament and attachment styles.

SUMMARY

97

Samenvatting Politieke omwentelingen na 1989 en de daarop volgende uitbreiding van de Europese Gemeenschap hebben tot een toename geleid van emigratie van Oost- naar West-Europa. In het inleidende hoofdstuk 1 worden gegevens gepresenteerd over migratie van Oost- naar West-Europa, alsook over culturele, sociale en economische verschillen tussen Rusland, Hongarije, Nederland en Polen. In dit hoofdstuk geven we voorts een inleiding in de theorie van acculturatie vanuit het perspectief van de hechtingstheorie. In hoofdstuk 2 richten we ons op de repliceerbaarheid van factoren van de Attachment Style Questionnaire (ASQ). Hoewel het model van Bartholomew en Horowitz al in veel onderzoek als uitgangspunt is gebruikt, is onvoldoende duidelijk geworden of het model ook steun vindt in factoranalyse. Voor de oorspronkelijke Nederlandse versie van de ASQ zijn destijds vier gecorreleerde factoren gevonden, die overeenkwamen met de vier soorten hechting die in het model gepostuleerd worden. Als die factoren ook in andere populaties aangetroffen zouden worden, dan zouden we dat als steun voor het model kunnen zien. Om een en ander na te gaan pasten we de (confirmatieve) Multiple Groep-Methode toe, en de (exploratieve) methode van Simultane Componentenanalyse (SCA). Met beide methoden vonden we, gebruikmakend van de oorspronkelijke versie van de ASQ, in alle onderzochte groepen drie duidelijke factoren, overeenkomend met zekere, angstige en gepreoccupeerde hechting. Maar de gepostuleerde structuur van het model, onder meer implicerend dat gepreoccupeerde hechting het tegenovergestelde van vermijdende hechting zou zijn, werd niet teruggevonden. De gegevens wijzen veeleer uit dat gepreoccupeerde personen, net als angstige personen, een negatief model van anderen hebben. Bovendien gaf de vermijdende factor twee gescheiden deelaspecten te zien: vermijden van nauwe banden en zelfgenoegzaamheid. In hoofdstuk 3 onderzochten we de relatie tussen hechtingsstijlen bij volwassenen en psychologische en sociaal-culturele aanpassing van Poolse, Russische en Hongaarse immigranten aan de Nederlandse samenleving. Ook onderzochten we de relatie tussen demografische factoren en aanpassing en vergeleken we de predictieve waarde van hechtingsstijlen en demografische factoren voor de aanpassing van immigranten. De Attachment Style Questionnaire werd gebruikt om hechting in kaart te brengen. Psychologische aanpassing werd gemeten met de Psychological Health Scale en de Satisfaction With Life Scale. Sociaal-culturele aanpassing werd gemeten met de Social Support List Interactions scale. Twee door ons zelf ontwikkelde schalen voor het meten van identificatie en contact met resp. de eigen en de Nederlandse cultuur werden gebruikt als indicator van culturele aanpassing. Zekere hechting bleek positief samen te hangen met

98

psychologische en sociaal-culturele aanpassing, angstige hechting hing negatief samen met psychologische aanpassing, en sterker negatief met identificatie met de Nederlandse cultuur dan met identificatie met de eigen cultuur. Gepreoccupeerde hechting hing negatief samen met psychologische aanpassing en met identificatie met de Nederlandse cultuur. Vermijdende hechting vertoonde een zwak negatief verband met sociaal-culturele aanpassing. Correlatiepatronen over de drie immigrantengroepen geven aan dat vermijdende personen relatief onverschillig staan tegenover zowel de eigen als de Nederlandse cultuur. Inzake demografische factoren vonden we voor opleiding en leeftijd ten tijde van de emigratie een positief verband met psychologische en sociaal-culturele aanpassing, en lengte van verblijf hing positief samen met sociaal-culturele aanpassing. In het algemeen hingen demografische factoren sterker samen met sociaal-culturele dan met psychologische aanpassing. Regressie-analyse liet zien dat hechtingsstijlen betere voorspellers zijn van psychologische en sociaal-culturele aanpassing van immigranten dan demografische factoren als opleiding, leeftijd bij immigratie en lengte van verblijf. In hoofdstuk 4 zochten we binnen het kader van de hechtingstheorie empirische steun voor een begrip als emigranten-persoonlijkheid. We vergeleken Poolse emigranten in Nederland met in Polen woonachtige Polen, en maten hun zekere en vermijdende hechtingsstijlen, controlerend voor leeftijd, geslacht en opleiding in beide groepen. Emigranten bleken zekerder en meer vermijdend dan hun niet-gemigreerde landgenoten. Verder onderzochten we of hechtingsstijlen benvloed zouden kunnen worden door de cultuur van het gastland. Daarvoor deelden we de emigrantengroep op in elf cohorten, van elkaar gescheiden door telkens twee jaar langer verblijf in het gastland, en we vergeleken hun gemiddelde scores op zekere en vermijdende hechting. We zagen dat die scores stabiel bleven over de cohorten. Bovendien bleven de verschillen tussen emigranten en nietemigranten stabiel over cohorten. Dit vormt een aanwijzing dat hechtingsstijlen van emigranten niet worden benvloed door de cultuur van het gastland, en dat de gevonden verschillen in hechtingsstijlen tussen emigranten en niet-emigranten duurzaam zijn, en vermoedelijk al ten tijde van de emigratie bestonden. Tenslotte vonden we dat zekere hechting een betere voorspeller was van psychologische gezondheid in de emigrantengroep dan in de niet-emigrantengroep. We concluderen dat er zoiets als een emigranten-persoonlijkheid bestaat, een type persoonlijkheid waar emigranten baat bij hebben. In hoofdstuk 5 onderzochten we verschillen in hechtingsstijlen tussen Oost- en West-Europeanen. Zoals te verwachten op grond van eerder onderzoek, vertoonden Oost-Europeanen meer gepreoccupeerde hechting dan West-Europeanen. Dit verschil is vooral toe te schrijven aan verschil in

SUMMARY

99

opvoedingsgewoonten, in plaats van aan directe invloed van cultuur op de persoon. Ook gingen we na of de huidige generatie van Oost- en WestEuropeanen, die opgegroeid is in een meer open, genternationaliseerde wereld, meer op elkaar lijkt qua de generatie van hun ouders. Hier werd geen aanwijzing voor gevonden. Wel vonden we dat zekere hechting van ouders positief samenhangt met zekere hechting van het kind. Ook bleek dat een gezag-afdwingende opvoedingsstijl de kans vergroot dat het kind een zekere hechtingsstijl zal ontwikkelen. Inzake psychologische gezondheid waren de resultaten consistent met eerdere bevindingen. Psychologische gezondheid is, ongeacht de culturele context, een belangrijke factor voor het ontstaan van zekere hechting bij kinderen. In hoofdstuk 6 geven we een terugblik op de belangrijkste bevindingen, en bespreken we de begripsmatige, methodologische, en praktische implicaties ervan. We stellen dat de resultaten van nut kunnen zijn voor beleid inzake begeleiding van immigranten. De belangrijkste beperkingen van ons onderzoek worden besproken, en de richting voor verder onderzoek wordt uitgestippeld. Met name wordt gepleit voor gecombineerd onderzoek naar individuele verschillen en identiteit van immigranten, met aandacht voor de relatie tussen persoonlijkheid, temperament en hechtingsstijlen.

REFERENCES

100

References

Aaronson, C., Bender, D., Skodol, A., & Gunderson, J. (2006). Comparison of attachment styles in borderline personality disorder and obsessive-compulsive personality disorder. Psychiatric Quarterly, 77, 69-80. Ainsworth, M. (1967). Infancy in Uganda. Baltimore: Johns Hopkins. Ainsworth, M. D. S., Blehar, M., Waters, E., & Wall, S. (1978). Patterns of attachment: A psychological study of the strange situation. Hillsdale, NJ: Erlbaum. Angleitner, A., & Ostendorf, F. (2000, July). The FFM: A comparison of German-speaking countries (Austria, Former East and West Germany, and Switzerland). Paper presented at the 27th International Congress of Psychology, Stockholm, Sweden. Angleitner, A., John, O. P. & Lohr, F. J. (1986). It's what you ask and how you ask it: An itemmetric analysis of personality questionnaires. In A. Angleitner & J. S. Wiggins (Eds.), Personality assessment via questionnaires (pp. 61107). New York: Springer-Verlag. Anhalt, K. (2001). The relation between parenting factors and social anxiety: A retrospective study. Dissertation Abstracts International: Section B: The Sciences and Engineering, Vol 62(1-B). West Virginia University, US. Arends-Tth, J.V., & Van de Vijver, F. (2004). Domains and dimensions in acculturation: Implicit theories of Turkish-Dutch. International Journal of Intercultural Relations, 28, 19-35. Arends-Tth, J.V., & Van de Vijver, F. (2007). Acculturation attitudes: A comparison of measurement methods. Journal of Applied Social Psychology, 37(7), 1462-1488. Arends-Tth, J.V. (2003) Psychological Acculturation of Turkish Migrants in the Netherlands. (Doctoral dissertation, Universiteit van Tilburg). Ataca, B. & Berry, J.W. (2002). Psychological, sociocultural, and marital adaptation of Turkish immigrant couples in Canada. International Journal of Psychology, 37, 13-26. Bakker, W., Van Oudenhoven, J.P., & Van der Zee, K. (2004). Attachment styles, personality, and Dutch emigrants' intercultural adjustment. European Journal of Personality, 18, 387-404. Baldwin, M. W., Fehr, B., Keedian, E., Seidel, M., & Thomson, D. W. (1993). An exploration of the relational schema underlying

REFERENCES

101

attachment styles: Self-report and lexical decision approaches. Personality and Social psychology Bulletin, 19, 746-754. Barry, R., Lakey, B., & Orehek, E. (2007). Links Among Attachment Dimensions, Affect, the Self, and Perceived Support for Broadly Generalized Attachment Styles and Specific Bonds. Personality and Social Psychology Bulletin, 33, 340-353. Bartholomew, K., & Horowitz, L. (1991). Attachment styles among young adults: A test of a four-category model. Journal of Personality and Social Psychology, 61, 226-244. Baumrind, D. (1966). Effect of authoritative control on child behavior. Child Development, 37, 887-907. Baumrind, D. (1967). Child care practices anteceding three patterns of preschool behaviors. Genetic Psychology Monographs, 75, 43-88. Baumrind, D. (1971). Current patterns of parental authority. Developmental Psychology, 4, 1-103. Baumrind, D. (1991). The influence of parenting style on adolescent competence and substance use. Journal of Early Adolescence, 11, 56-95. Becker, T. E., Billings, R. S., Eveleth, D. M., & Gilbert, N. W. (1997). Validity of scores on three attachment style scales: exploratory and confirmatory evidence. Educational and Psychological Measurement, 57, 477-493. Benoit, D., & Parker, K. (1994). Stability and transmission of attachment across three generations. Child Development, 65, 1444-1456. Berkman, L.F. (1977). Psychological resources, health behaviour, and mortality: A nine-year follow-up study. Paper presented at the American Public health Association Annual Meeting. Washington, D.C. Berry, J., & Kim, U. (1988). Acculturation and mental health. Thousand Oaks, CA, US: Sage Publications, Inc. Berry, J., Kim, U., Power, S., & Young, M. (1989). Acculturation attitudes in plural societies. Applied Psychology: An International Review, 38, 185-206. Berry, J. (1992). Acculturation and adaptation in a new society. International Migration, 30, 69-86. Berzonsky, M. D., Macek P., & Nurmi J.E., (2003) Interrelationships Among Identity Process, Content, and Structure: A Cross-Cultural Investigation. Journal of Adolescent Research, 18, 112-129. Bochner, S. (1982). Cultures in Contact: Studies in Cross-Cultural Interaction. Pergamon Press: New York.

REFERENCES

102

Bond, M., Leung, K., Au, A., Tong, K., & Chemonges-Nielson, Z. (2004). Combining social axioms with values in predicting social behaviours. European Journal of Personality, 18(3), 177-191. Boneva, B., & Frieze, I. (2001). Toward a concept of a migrant personality. Journal of Social Issues, 57, 477-491. Bowlby, J. (1969/1982). Attachment and loss: Vol.1 Attachment. New York: Basic Books. Bowlby, J. (1973). Attachment and loss: Vol.2. Separation, anxiety and anger New York: Basic Books. Bowlby, J. (1980). Attachment and Loss: Vol.3. Loss, sadness and depression. The international psycho-analytical library. London: Hogath Press. Bowlby, J. (1984). Violence in the family as a disorder of the attachment and caregiving systems. The American Journal of Psychoanalysis, 44, 9-27. Brennan, K. A., Clark, C. L., & Shaver, P. R. (1998). Self-report measurement of adult romantic attachment: An integrative overview. In J. A. Simpson & W. S. Rholes (Eds.), Attachment theory and close relationships (pp. 46-76). New York: Guilford Press. Bridges, L., & Connell, J. (1991). Consistency and inconsistency in infant emotional and social interactive behavior across contexts and caregivers. Infant Behavior & Development, 14, 471-487. Buri, J. (1991). Parental Authority Questionnaire. Journal of Personality Assessment, 57, 110-119. Buri, J., Louiselle, P., Misukanis, T., & Mueller, R. (1988). Effects of parental authoritarianism and authoritativeness of self-esteem. Personality and Social Psychology Bulletin, 14(2), 271-282. Cassidy, J., & Shaver, P. (Eds.). (1999). Handbook of attachment: Theory, research, and clinical applications. New York: Guilford. Chotai, J., Jonasson, M., Hgglf, B., & Adolfsson, R. (2005). Adolescent attachment styles and their relation to the temperament and character traits of personality in a general population. European Psychiatry, 20, 251-259. CIA: The World Factbook (2007). Retrieved June 22, 2007 from the web site of the American Central Intelligence Agency (CIA): https://www.cia.gov/library/publications/the-world-factbook/ Cloninger, C., Svrakic, D., & Przybeck, T. (1993). A psychobiological model of temperament and character. Archives of General Psychiatry, 50, 975-990. Collins, N. L., & Feeney, B. C. (2000). A safe haven: An attachment theory perspective on support seeking and caregiving in intimate

REFERENCES

103

relationships. Journal of Personality and Social Psychology, 78, 10531073. Collins, N. L., & Feeney, B. C. (2004). Working models of attachment shape perceptions of social support: Evidence from experimental and observational studies. Journal of Personality andSocial Psychology, 87, 363-383. Collins, N. L., & Read, S. J. (1990). Adult attachment, working models, and relationship quality in dating couples. Journal of Personality and Social Psychology, 58, 644-663. Collins, W., Maccoby, E., Steinberg, L., Hetherington, E., & Bornstein, M. (2000). Contemporary research on parenting: The case for nature and nurture. American Psychologist, 55, 218-232. Cooper, M. L., Shaver, P. R., & Collins, N. L. (1998). Attachment styles, emotion regulation and adjustment in adolescence. Journal of Personality and Social Psychology, 74, 1380-1397. Cortes, D. E., Rogler, L. H., & Malgady, R. G. (1994). Biculturality among Puerto Rican adults in the United States. American Journal of Community Psychology, 22, 707-721. Demo, D., & Cox, M. (2000). Families with young children: A review of research in the 1990s. Journal of Marriage & the Family, 62, 876895. Diehl, M., Elnick, A., Bourbeau, L., & Labouvie-Vief, G. (1998). Adult attachment styles: Their relations to family context and personality. Journal of Personality and Social Psychology, 74, 1656-1669. Diener, E., Emmons, R. A., Larsen, R. J., & Griffin, S. (1985). The satisfaction with life scale. Journal of Personality Assessment, 49, 71-75. Dobbins, T. (2006). Attributions: Relations to attachment and caregiving representations. Dissertation Abstracts International: Section B: The Sciences and Engineering, Vol 67(6-B). University of North Carolina, Greensboro, US Doomernik, J., Penninx, R., & Van Amersfoort, H. (1997). A migration policy for the future. Possibilities and limitations retrieved from: http://de.scientificcommons.org/8060608 Duggan, E. S. & Brennan, K. A. (1994). Social avoidance and its relation to Bartholomews adult attachment typology. Journal of Social and Personal Relationships, 11, 147-153. Egeland, B., & Farber, E. (1984). Infant-mother attachment: Factors related to its development and changes over time. Child Development, 55, 753-771. Elicker, J., Englund, M, & Sroufe, L.A. (1992). Predicting peer competence and peer relationships in childhood from early parent-child

REFERENCES

104

relationships. In R. Parke and G Ladd (Eds.), Family-Peer relationships: Models of Linkage. Hillsdale, NJ: Erlbaum. Ettema, J. (1999). Adult attachment: Its relationship to parenting, infant health and development, and helping relationships. Dissertation Abstracts International: Section B: The Sciences and Engineering, Vol 59(10-B). Michigan State University, US. Eurostat (2007). Data source: Internal migration within the EU. Retrieved September 10, 2007, from: http://epp.eurostat.ec.europa.eu. Feeney, J. A., & Noller, P. (1990). Attachment styles as a predictor of adult romantic relationships. Journal of Personality and Social Psychology, 58, 281-291. Fischer R, Maes J, & Schmitt M. (2007). Tearing down the 'Wall in the head'? Culture contact between Germans. International Journal of Intercultural Relations, 31, 163-179. Fischer, R. (2004). Standardization to Account for Cross-Cultural Response Bias: A Classification of Score Adjustment Procedures and Review of Research in JCCP. Journal of Cross-Cultural Psychology, 35, 263-282. Fulker, D. W., Eysenck, S.B.G., & Zuckerman, M. (1980). The genetics of sensation seeking. Journal of Personality Research, 14, 261-281. Gamble, S., & Roberts, J. (2005). Adolescents' Perceptions of Primary Caregivers and Cognitive Style: The Roles of Attachment Security and Gender. Cognitive Therapy and Research, 29, 123-141. Gerlsma, C., Snijders, T., van Duijn, M., & Emmelkamp, P. (1997). Parenting and psychopathology: Differences in family members' perceptions of parental rearing styles. Personality and Individual Differences, 23, 271-282. Gorsuch, R. L. (1970). A comparison of Biquartin, Maxplane, Promax, and Varimax. Educational and Psychological Measurement, 30, 861872. Gorsuch, R. L. (1983). Factor analysis. Second edition. Hillsdale. New Jersey. Graves, T. 1967. Psychological acculturation in a tri-ethnic community. South-Western Journal of Anthropology, 23, 33750. Griffin, D. W., & Bartholomew, K. (1994). Models of the self and other: Fundamental dimensions underlying measures of adult attachment. Journal of Personality and Social Psychology, 67, 430-445. Ha Kristic, C. (2000) Ascertaining the relationship of acculturation and the perception of psychological stress, self-concept, and locus of control, among Vietnamese refugees and immigrants in the United States of America. Dissertation Abstracts International, 60 (7-B). Kent State University, Kent, USA.

REFERENCES

105

Handojo, V. (2000). Attachment styles, acculturation attitudes/behaviors, and stress among Chinese Indonesian immigrants in the United States. Dissertation Abstracts International, 61 (4-B). Fuller Theological Seminary, Pasadena, USA. Hanono, J. (1999). Cross-cultural differences in adult attachment styles. Dissertation Abstracts International: Section B: The Sciences and Engineering, Vol 59(7-B). Fairleigh Dickinson University, US. Harmsen, C. N. (1999). Cross-cultural marriages. Monthly Bulletin of Population Statistics, 10. Hazan, C., & Zeifman, D. (1999). Pair Bonds as Attachments. Evaluating the Evidence. In Cassidy, J. & Shaver, P. (Eds.). Handbook of attachment: Theory, research, and clinical applications (pp.3-20). New York: Guilford. Hazan, C., & Shaver, P. (1987) Romantic love conceptualized as an attachment process; Journal of Personality and Social Psychology, 52, 511-524. HDR: Human Development Report. (2006). Retrieved June 22, 2007 from HDR web site: http://hdr.undp.org/hdr2006/statistics/ Hofstede, G. (1980). Culture's Consequences International Differences in Work-related Values. Culture's Consequences International Differences in Work-related Values. Sage Publications: Beverly Hills, CA, US. Hofstra, J., Van Oudenhoven, J. P., & Buunk, B. P. (2005). Attachment styles and majority members attitudes towards adaptation strategies of immigrants. International Journal of Intercultural Relations, 29, 601-619. Holden, R. R., & Fekken, G. C. (1990). Structured Psychopathological Test Item Characteristics and Validity. Psychological Assessment, 2, 3540. Holmes, T.H., & Rahe, R.H. (1967). The social readjustment rating scale. Journal of Psychosomatic Research, 11, 213-218. Horowitz, L. M., Rosenberg, S.E., & Bartholomew, K. (1993). Interpersonal Problems, Attachment Styles, and Outcome in Brief Dynamic Psychotherapy. Journal of Consulting and Clinical Psychology, 61, 549-560. Hungarian Central Statistical Office (HCSO). (2001) Retrieved June 22, 2007 from HSCO http://www.euromonitor.com/factfile.aspx?country=HU Hutto, C. (1998). The relationship of selected family variables to college student adjustment. Dissertation Abstracts International: Section B: The Sciences and Engineering, Vol 59(5-B), pp. 2455.

REFERENCES

106

Iglicka, K. (2000). Mechanisms of migration from Poland before and during the transition period. Journal of Ethnic & Migration Studies, 26, 61-73. Jancz, M. W. (2000). Social and psychological adjustments of first generation Polish immigrants to Australia. (Doctoral dissertation, University of Sydney, Australia). Jasinskaja-Lahti, I. (2000). Psychological Acculturation and Adaptation among Russian-Speaking Immigrant Adolescents in Finland. Helsinki: Helsingin. Jayasuriya, L., Sang, D., & Fielding, A. (1992). Ethnicity, immigration and mental illness: A critical review of Australian research. South Carlton, Australia: Bureau of Immigration Research. Kafetsios, K., & Nezlek, J. (2002). Attachment styles in everyday social interaction. European Journal of Social Psychology, 32, 719-735. Karen, R. (1998). Becoming attached: First relationships and how they shape our capacity to love. New Your: Oxford University Press. Keller, H. (2003). Socialization for competence: Cultural models of infancy. Human Development, 46, 288-311. Keppel, G. (1991). Design and analysis: A researcher's handbook (3rd ed.). Englewood Cliffs, NJ, US: Prentice-Hall, Inc. Kiers, H. A., & Ten Berge, J.M.F. (1989). Alternating least squares algorithms for simultaneous component analysis with equal component weight matrices in two or more populations. Psychometrika, 54, 467-473. Kiers, H.A. (1990). SCA: A program for simultaneous component analysis. Groningen: IEC ProGamma. Kitayama, S., Markus, H., Matsumoto, H., & Norasakkunkit, V. (1997). Individual and collective processes in the construction of the self: Self-enhancement in the United States and self-criticism in Japan. Journal of Personality and Social Psychology, 72, 1245-1267. Kitayama, S., Markus, H., Matsumoto, H., & Norasakkunkit, V. (1997). Individual and collective processes in the construction of the self: Self-enhancement in the United States and self-criticism in Japan. Journal of Personality and Social Psychology, 72, 1245-1267. Kobak, R. (1994). Adult attachment: A personality or relationship construct? Psychological Inquiry, 5, 42-44. Kobak, R., & Sceery, A. (1988). Attachment in late adolescence: Working models, affect regulation, and representations of self and others. Child Development, 59, 135-146. Kobak, R., Cole, H., Fleming, W., Ferenz-Gillies, R., & Gamble, W. (1993). Attachment and emotion regulation during mother-teen problem-

REFERENCES

107

solving: A control theory analysis. Child Development, 64, 231245. Kochanska, G., Kuczymski, L., & Radke, Y. (1989). Correspondence between mothers self-report and observed child rearing practices. Child Development, 60, 56-63. Kurdek, L.A. (2002). On being insecure about the assessment of attachment styles. Journal of Social and Personal Relationships, 19, 811-834. Lamb, M., & Lewis, C. (2005). The Role of Parent-Child Relationships in Child Development. Developmental science: An advanced textbook (5th ed.) (pp. 429-468). Lawrence Erlbaum Associates Publishers. Lamm, B., & Keller, H. (2007). Understanding Cultural Models of Parenting: The Role of Intracultural Variation and Response Style. Journal of Cross-Cultural Psychology, 38, 50-57. Retrieved Monday, April 30, 2007 from the PsycINFO database. Lapsley, D. K., & Edgerton, J. (2002). Separetion-individuation, adult attachment style, and college adjustment. Journal of Counseling and Development, 80, 484-492. Larsen, R., & Ketelaar, T. (1989). Extraversion, neuroticism and susceptibility to positive and negative mood induction procedures. Personality and Individual Differences, 10, 1221-1228. Lipiski, . (2006). Polska si nam rozjechaa. [Poland is in movement]. Gazeta Wyborcza. Retrieved June 11, 2006, from Portal Gazeta Wyborcza site: http://www.gazeta.pl. Livesley W.J., Jackson, D.N., & Schroeder, M.L. (1992) Factorial Structure of Traits Delineating Personality Disorders in Clinical and General Population Samples. Journal of Abnormal Psychology, 101, 432440. Lopez, F. G., Mauricio, A. M., & Gormley, B. (2001). Adult attachment orientations and college student distress: The mediating role of problem coping styles. Journal of Counseling & Development, 79, 459-464. Maccoby, E. (2000). Parenting and its effects on children: On reading and misreading behavior genetics. Annual Review of Psychology, 51, 127. Mallinckrodt, B. (2000). Attachment, social competencies, social support and interpersonal process in psychotherapy. Psychotherapy Research, 10, 239-266. Markus, H., & Kitayama, S. (1991). Culture and the self: Implications for cognition, emotion, and motivation. Psychological Review, 98, 224253.

REFERENCES

108

Martens, P. (2006). Parenting and internal working models in preschool-age children. Dissertation Abstracts International: Section B: The Sciences and Engineering, Vol 67(2-B). Northern Illinois U., US. Massey, D.S., Arango, J., Hugo, G., Kouaouchi, A., Pellegrino, A., & Taylor, J.E. (1998). Worlds in motion: Understanding international migration at the end of the millennium. Oxford: Clarendon Press. Mayseless, O. (1993). Gifted adolescents and intimacy in close same-sex friendships. Journal of Youth and Adolescence, 22, 135-146. McCrae, R., & Costa, P. (2003). Personality in adulthood: A five-factor theory perspective (2nd ed.). New York: Guilford Press. Mikulincer, M., & Nachshon, O. (1991). Attachment styles and patterns of self-disclosure. Journal of Personality and Social Psychology, 61, 321-331. Mikulincer, M., & Shaver, P. R.(2001). Attachment theory and intergroup bias: Evidence that priming the secure base schema attenuates negative reactions to out-groups. Journal of Personality and Social Psychology, 81, 97-115. Mikulincer, M., & Shaver, P. R. (2003). The attachment behavioral system in adulthood: Activation, psychodynamics, and interpersonal processes. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 35, pp. 53-152). New York: Academic Press. Mikulincer, M., Florian, V., & Tolmacz, R. (1990). Attachment styles and fear of personal death: A case study of affect regulation. Journal of Personality and Social Psychology, 58, 273-280. Minturn, L., & Lambert, W. (1964). Mothers of six cultures. Oxford, England: Wiley. Morrison, P. A., & Wheeler, J. P. (1976). The image of elsewhere in the American tradition of migration. Santa Monica, CA, US: RAND Corporation. Mott, W. H. (2004). Globalization: People Perspectives, and Progress. Westport, CT, US: Praeger Publishers. Nair, H., & Murray, A. (2005). Predictors of Attachment Security in Preschool Children From Intact and Divorced Families. Journal of Genetic Psychology, 166, 245-263. NationMaster (2003) Retrieved June 22, 2007 from the NationMaster web site: www.nationmaster.com Neal, J., & Frick-Horbury, D. (2001). The effects of parenting styles and childhood attachment patterns on intimate relationships. Journal of Instructional Psychology, 28, 178-183. Ng, K., Trusty, M., & J. Crawford, R. (2005). A Cross-Cultural Validation of the Attachment Style Questionnaire: A Malaysian Pilot Study.

REFERENCES

109

Family Journal: Counselling and Therapy for Couples and Families, 13, 416-426. Nunnally J. C. & Bernstein I. H. (1994) Psychometric theory. McGraw-Hill. New York. Nunnally, J. C. (1978). Psychometric theory. McGraw-Hill. New York. Onishi, M., Gjerde, P. F., & Block, J. (2001). Personality implications of romantic attachment patterns in young adults: a multi-method, multi-informant study. Personality and Social Psychology Bulletin, 27, 1097-1110. Ouarasse, A. O., & Van de Vijver, F. J. R (2005). The role of demographic variables and acculturation attitudes in predicting sociocultural and psychological adaptation in Moroccans in the Netherlands. International Journal of Intercultural Relations, 29, 251-272. Padilla, A. M. (1986). Acculturation and stress among immigrants and latergeneration individuals. In D. Frick (Ed.) Urban quality of life: Social, psychological and physical conditions (pp. 41-60). Berlin, Germany: Walter de Gruyter. PCSO: Polish Central Statistical Office. (1996). Data source about population in Poland. Retrieved on September 10, 2007, from: http://www.stat.gov.pl/gus/index_ENG_HTML.htm. Polek, E., Ten Berge, J. M. F., & Van Oudenhoven J. P. (2006). Evaluating the replicability of the Bartholomew and Horowitz model of attachment in one native Dutch and four immigrant samples. Manuscript submitted for publication. Polek, E., Van Oudenhoven J. P., & Ten Berge, J. M. F. (2007). Attachment styles and demographic factors as predictors of sociocultural and psychological adjustment of Eastern European emigrants in the Netherlands. International Journal of Psychology (in print). Poortinga, Y. (1989). Equivalence of cross-cultural data: An overview of basic issues. International Journal of Psychology, 24, 737-756. Poortinga, Y. (1995). Cultural bias in assessment: Historical and thematic issues. European Journal of Psychological Assessment, 11, 140146. RAND: Research and Development, Health Sciences Program (1992). RAND 36-Item Health Survey 1.0. Santa Monica, CA: RAND. Ray, J. (1986). The traits of emigrants: A case study of the Sydney Parsees. Journal of Comparative Family Studies, 17, 127-130. Redfield R., Linton R., Herskovits M. J., (1936) Memorandum for the Study of Acculturation. American Anthropologist, 38, 149152. Reis, J., Barbera-Stein, L., & Bennett, S. (1986). Ecological determinants of parenting. Family Relations, 35, 547-554.

REFERENCES

110

Reitman, D., Rhode, P. C., Hupp, S. D. A., & Altobello, C. (2002). Development and validation of the Parental Authority Questionnaire-Revised. Journal of Psychopathology and Behavioral Assessment, 24 , 119-127. Rholes, W., Simpson, J., & Blakely, B. (1995). Adult attachment styles and mothers' relationships with their young children. Personal Relationships, 2, 35-54. Rholes, W., Simpson, J., & Friedman, M. (2006). Avoidant Attachment and the Experience of Parenting. Personality and Social Psychology Bulletin, 32, 275-285. Ricks, M. (1985). The social transmission of parental behavior: Attachment across generations. Monographs of the Society for Research in Child Development, 50, 211-227. Rubin, K., & Chung, O. (2006). Parenting beliefs, behaviors, and parentchild relations: A cross-cultural perspective. New York, NY, US: Psychology Press. Rudmin, F., & Ahmadzadeh, V. (2001). Psychometric critique of acculturation psychology: The case of Iranian migrants in Norway. Scandinavian Journal of Psychology, 42, 41-56. Schmitt, D. (2003). Are men universally more dismissing than women? Gender differences in romantic attachment across 62 cultural regions. Personal Relationships, 10, 307-331. Schmitt, D., & Allik, J. (2005). Simultaneous Administration of the Rosenberg Self-Esteem Scale in 53 Nations: Exploring the Universal and Culture-Specific Features of Global Self-Esteem. Journal of Personality and Social Psychology, 89, 623-642. Schmitt, D., Alcalay, L., Allensworth, M., Allik, J., Ault, L., Austers, I., et al. (2004). Patterns and Universals of Adult Romantic Attachment Across 62 Cultural Regions: Are Models of Self and of Other Pancultural Constructs?. Journal of Cross-Cultural Psychology, 35, 367-402. Schmitt, D., Allik, J., McCrae, R., & Benet-Martnez, V. (2007). The Geographic Distribution of Big Five Personality Traits: Patterns and Profiles of Human Self-Description Across 56 Nations. Journal of Cross-Cultural Psychology, 38, 173-212. Scott, A. W., & Scott R. (1989). Adaptation of Immigrants. Individual differences and determinants. Toronto: Pergamon Press. Searle, W., & Ward, C. (1990). The prediction of psychological and sociocultural adjustment during cross-cultural transitions. International Journal of Intercultural Relations, 14, 449-464.

REFERENCES

111

Shaver P.R. & Hazan, C. (1994). Attachment. In A. L. Weber & J.H. Harvey (Ed.) Perspectives on close relationships (pp.110-131). Boston, US: Allyn and Bacon. Shaver, P., & Brennan, K. (1992). Attachment styles and the 'Big Five' personality traits: Their connections with each other and with romantic relationship outcomes. Personality and Social Psychology Bulletin, 18, 536-545. Sibley, C. G., Fischer, R. & Liu, J. H. (2005). Reliability and Validity of the Revised Experiences in Close Relationships (ECR-R) Self-Report Measure of Adult Romantic Attachment. Personality and Social Psychology Bulletin, 31, 1524-1536. Simpson, J. A.; Rholes, W. S., & Nelligan, J. S. (1992). Support seeking and support giving within couples in an anxiety-provoking situation: The role of attachment styles. Journal of Personality and Social Psychology, 62, 434-446. Smith, E. R., Murphy, J., & Coats, S. (1999). Attachment to groups: Theory and measurement. Journal of Personality and Social Psychology, 77, 84-110. Statistics Netherlands (2006). Data source: Population by origin, generation, sex, age and marital status. Retrieved September 10, 2007, from Statistics Netherlands Web site: http://www.cbs.nl. Statistics Netherlands (2006). Press release 10-05-2006. Retrieved June 10, 2006, from Statistics Netherlands Web site: http://www.cbs.nl Statistics Netherlands (2007). Data source: Population by origin, generation, sex, age and marital status. Retrieved September 10, 2007, from Statistics Netherlands Web site: http://www.cbs.nl. Stevens, G. (1999). Age at immigration and second language proficiency among foreign-born adults. Language in Society, 28, 555-578. Stuive I., Kiers H.A.L., Timmerman M.E. & ten Berge J.M.F. (2006). Comparison of Methods for adjusting incorrect assignments of items to subtests: Oblique Multiple Group method versus Confirmatory Common Factor method. Manuscript submitted for publication. Ten Berge, J.M.F. (1986). Rotation to perfect congruence and the crossvalidation of component weights across populations. Multivariate Behavioral Research, 21, 41-64; 262-266. Thompson, R. (1988). The effects of infant day care through the prism of attachment theory:A critical appraisal. Early Childhood Research Quarterly, 3, 273-282. Thornquist, M. H., Zuckerman, M., & Exline, R. V. (1991). Loving, Liking, Looking and Secsation Seeking in Unmarried Collage Couples. Personality and Individual Differences, 12, 1283-1292.

REFERENCES

112

Toothaker, L.E. (1993). Multiple comparison procedures. Sage university paper series on quantitative applications in the social sciences, 07089. Newbury Park, CA: Sage. Umberson, D. (1989). Relationships with children: Explaining parents' psychological well-being. Journal of Marriage & the Family, 51, 999-1012. UN: United Nation (2006) Human Development Report, Retrieved June 22, 2007 from web site: http://en.wikipedia.org/wiki/Satisfaction_with_Life_Index UN: United Nations (2005). Trends in total migration stock: The 2005 revision. Department of Economic and Social Affairs, Population Division, Retrieved June 22, 2007 from web site: http://www.un.org/esa/population/publications/migration/UN_Migr ant_Stock_Documentation_2005.pdf UN: United Nations Statistics Division (2004). Demographic yearbook 2004. Retrieved June 22, 2007 from United Nations Statistics Division web site: http://unstats.un.org/unsd/demographic/products/dyb/dyb2004.htm UNFPA: United Nations Population Found (1990). Retrieved June 22, 2007 from http://www.unfpa.org/worldwide/ UNICEF: United Nations Childrens Found (2002). TransMONEE database 2002. UNICEF Innocenti Research Centre (IRC), Florence. Retrieved June 22, 2007 from European Foundation for the Improvement of Living and Working Conditions web site: http://www.aegis.com/files/unicef/socialmonitor2002.pdf Van de Vijver, F., & Leung, K. (1997). Methods and data analysis for crosscultural research. Thousand Oaks, CA, US: Sage Publications, Inc. Van de Vijver, F., & Poortinga, Y. (1982). Cross-cultural generalization and universality. Journal of Cross-Cultural Psychology, 13, 387-408. Van de Vijver, F., & Tanzer, N. (2004). Bias and equivalence in crosscultural assessment: An overview. European Review of Applied Psychology/Revue Europenne de Psychologie Applique, 54, 119135. Van Ecke, Y., Chope, R. C., & Emmelkamp, P. M. G. (2005). Immigrants and attachment status: Research findings with Dutch and Belgian immigrants in California. Social Behavior and Personality, 33, 657673. Van Ijzendoorn, M., & Sagi, A. (1999). Cross-cultural patterns of attachment: Universal and contextual dimensions. New York, NY, US: Guilford Press. Van Oudenhoven J. P. & Hofstra J. (2006). Personal reactions to strange situations: attachment styles and acculturation attitudes of

REFERENCES

113

immigrants and majority members. International Journal of Intercultural Relations, 30, 783-798. Van Oudenhoven, J. P., & Van der Zee, K. I. (2002). Predicting multicultural effectiveness of international students: The Multicultural Personality Questionnaire. International Journal of Intercultural Relations, 26, 679-694. Van Oudenhoven, J. P., Hofstra, J., & Bakker, W. (2003). Ontwikkeling en evaluatie van de Hechtingstijlvragenlijst (HSL) [Development and evaluation of the Attachment Styles Questionnaire]. Nederlands Tijdschrift voor de Psychologie, 58, 95-102. Van Oudenhoven, J., Prins, K., & Buunk, B. (1998). Attitudes of minority and majority members towards adaptation of immigrants. European Journal of Social Psychology, 28, 995-1013. Van Sonderen, E. (1993). Het meten van sociale steun met de Sociale Steun Lijst-Interacties (SSL-I) en Sociale Steun Lijst-Dicrepanties (SSLD): een handeleiding [Measuring social support with the Social Support List-Interactions (SSL-I) and Social Support ListDiscrepancies (SSL-D): A manual]. NCG Reeks Meetinstrument-2. Groningen: Northern Centre for Health Care. Van Vianen, A., Feij, J., Krausz, M., & Taris, R. (2003). Personality Factors and Adult Attachment Affecting Job Mobility. International Journal of Selection and Assessment, 11, 253-264. Ward, C. Bochner, S., & Furnham, A. (Eds.). (2001).The psychology of culture shock. New York, NY: Routledge Ward, C., & Searle, W. (1991). The impact of value discrepancies and cultural identity on psychological and sociocultural adjustment of sojourners. International Journal of Intercultural Relations, 15, 209-225. Waters, E. (1978). The reliability and stability of individual differences in infant-mother attachment. Child Development, 49, 483-494. Waters, E., Crowell, J., Elliott, M., Corcoran, D., & Treboux, D. (2002). Bowlby's secure base theory and the social/personality psychology of attachment styles: Work(s) in progress. Attachment & Human Development, 4, 230-242. Weisner, T. (2005). Attachment as a Cultural and Ecological Problem with Pluralistic Solutions. Human Development, 48, 89-94. Wildt, A. R. & Ahtola, O. (1978). Analysis of covariance. Sage university paper series on quantitative applications in the social sciences, 07012. Newbury Park, CA: Sage. Wright, D.E. (2000). The effect of parental attachment on the college adjustment of urban and rural students. Dissertation Abstracts

REFERENCES

114

International: Section B: The Sciences and Engineering, Vol 61(5B), pp. 2788. Zelenski, J., & Larsen, R. (1999). Susceptibility to affect: A comparison of three personality taxonomies. Journal of Personality, 67, 761-791. .

APPENDIX

Appendix 1 The Attachment Style Questionnaire (Van Oudenhoven, Hofstra, & Bakker, 2003) used in the present research (items 3, 7, 15 were reversely keyed ). Secure attachment style 1. I feel at ease in emotional relationships 3. I feel uncomfortable when relationships with other people become close 7. I avoid close ties 9. I trust other people and I like it when other people can rely on me 12. I find it easy to get engaged in close relationships with other people 13. I feel at ease in intimate relationships 16. I think it is important that people can rely on each other 20. I trust that others will be there for me when I need them Fearful attachment style 2. I would like to be open to others but I feel that I can't trust other people 4. I would like to have close relationships with other people but I find it difficult to fully trust them 18. I am afraid that I will be deceived when I get too close with others 21. I am wary to get engaged in close relationships because I am afraid to get hurt Preoccupied attachment style 6. I often wonder whether people like me 8. I have the impression that usually I like others better than they like me 10. I am often afraid that other people don't like me 15. I don't worry whether people like me or not 19. I usually find other people more interesting than myself 22. It is important to me to know if others like me Dismissing 5. I prefer that others are independent of me and I am independent of them 11. It is important to me to be independent 14. I like to be self-sufficient 17. I don't worry about being alone: I don't need other people that strongly

Appendix 2 Items and factor loadings (Varimax rotation) in scales of Identification and contact with the Dutch and native cultures obtained in SCA (see chapter 3). Polish Russian Hungarian sample sample sample I II I II I II I. Identification and contact with Dutch culture I feel a member of Dutch society .72 -.13 .73 -.02 .68 -.18 I like to chat with the Dutch .06 .05 .61 .63 .65 -.12 I am treated well by the Dutch .10 .31 .00 .59 .66 .69 There are values, typical of Dutch culture, which I appreciate a lot .02 .10 -03 .59 .59 .45 I am interested in issues important to Dutch people .06 .20 .08 .73 .73 .73 How often do you: watch a Dutch film? .06 .62 .53 -.03 .39 -.03 How often do you: make friends with Dutch people? .04 .00 .72 .67 .66 -.08 How often do you: listen to Dutch radio or to Dutch music? .10 .10 .01 .57 .57 .59 II. Identification and contact with a native culture I am proud of being Polish / Russian / Hungarian -.05 .51 .11 .17 .51 .54 I would like to speak Polish / Russian / Hungarian as often as possible -.04 .64 -.02 .62 -.06 .51 I am interested in issues important to Polish / Russian / Hungarian people .08 .00 .02 .75 .68 .62 I like traditional Polish / Russian / Hungarian cuisine -.01 .58 .17 -.01 .58 .71 I like to chat with Poles /Russian / Hungarians .13 .14 -.14 .65 .77 .73 I like to be among Poles /Russian / Hungarians .01 .23 -.16 .64 .77 .74 I follow news about Poland / Russia / Hungary .07 .06 -.05 .60 .67 .56 How often do you: go to Poland/ Russia / Hungary? .09 -.05 .28 -.18 .24 .31
Correlation between factors Explained variance .05 3.34 3.38 3.24 .14 3.41 2.58 -.08 3.16

Note. For each variable the highest loading in a sample is in bold.