Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

"We don’t want immigrants because they don’t integrate ... and steal our jobs" : comparing economic and… Murakami, Go 2006

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
831-ubc_2006-0592.pdf [ 5.97MB ]
Metadata
JSON: 831-1.0092730.json
JSON-LD: 831-1.0092730-ld.json
RDF/XML (Pretty): 831-1.0092730-rdf.xml
RDF/JSON: 831-1.0092730-rdf.json
Turtle: 831-1.0092730-turtle.txt
N-Triples: 831-1.0092730-rdf-ntriples.txt
Original Record: 831-1.0092730-source.json
Full Text
831-1.0092730-fulltext.txt
Citation
831-1.0092730.ris

Full Text

"WE DON'T W A N T IMMIGRANTS B E C A U S E T H E Y DON'T INTEGRATE. . . A N D STEAL OUR JOBS": COMPARING ECONOMIC A N D C U L T U R A L INFLUENCES ON XENOPHOBIA C A N A D A , A U S T R A L I A A N D NEW Z E A L A N D by GO M U R A K A M I B. Policy Studies, Chuo University, 2003 A THESIS SUBMITTED IN PARTIAL F U L F I L L M E N T OF THE REQUIREMENTS FOR THE DEGREE OF M A S T E R OF ARTS in THE F A C U L T Y OF G R A D U A T E STUDIES (Political Science) THE UNIVERSITY OF BRITISH C O L U M B I A August 2006 © Go Murakami, 2006 ABSTRACT Immigration has recently become a salient political issue in liberal democracies. Many political scientists have analyzed causes of emerging support for anti-immigrant parties, and the development of immigration control in Europe. Compared to the analysis of voting behavior and policy analysis, however, public opinion concerning immigration has not been fully examined in political science literature. Thus this study systematically the public opposition to immigration in opinion surveys by applying the theory of prejudice, theory of perceived threat to group position, social identity theory and contact theory. The analysis uses the election studies of Canada, Australia and New Zealand in varied years, which are merged with census data to examine contextual effects. The broad research question is who opposes immigration, why they oppose and under what conditions they are more likely to do so. The focus of the analysis in this thesis is on comparing the influences of economic and cultural factors. A brief review of the history of accepting immigrants in three countries reveals that both factors have significantly influenced the development of immigration'policy. Although economic situation strongly influences the level of public's opposition to immigration over time, the analysis of five election studies of three countries repeatedly shows that cultural concerns increase individuals' probability of opposition more strongly than economic concerns. The analysis also finds that the local economic situation does not stably influence opposition to immigration, even if where immigrants are geographically concentrated. On the contrary, local economic situation increases the likelihood of opposition to immigration only where immigrants are not concentrated. Geographic concentration of immigrants does not have an interaction effect with economic concerns, but it magnifies the influence of one's cultural concerns on opposition to immigration during the economic downturn. Accordingly, the examination of interactions effects reveals that even when contextual factors influence opposition to immigration, they interact with cultural concerns. Thus cultural concerns should be paid more attention for the politics of immigration for future research. ii T A B L E O F C O N T E N T S Abstract ii Table of Contents iii List of Tables v List of Figures vi Acknowledgements viii 1 Introduction 1 2 Theory and Research Strategy 4 2.1 Overview 4 2.2 Theory of prejudice and political ideology 4 2.2.1 Symbolic racism, modern racism, and racial resentment theory 5 2.2.2 Prejudice or political ideology? Criticisms of Symbolic Racism Theory 7 2.2.3 Application of the theory of prejudice to opposition to immigration 8 2.3 Theory of group position and identification 10 2.3.1 Perceived threat to group position/realistic conflict theory 10 2.2.3 Social identity theory 12 2.3.3 Application of the theory of group position and identification to opposition to immigration 13 2.4 Contact theory and contextual effects 15 2.4.1 Contact theory 15 2.4.2 Proportion of social groups as a contextual effect 16 2.5 Integration of theories and research strategy: Comparing economic and cultural threat 19 3 Historical Background 23 3.1 Overview 23 3.2 The history of immigration policy 24 3.2.1 Economic incentives and cultural concerns: Implementing a racially discriminatory immigration policy 24 3.2.2 Abolishment of the racial discriminatory immigration policy 26 3.3 Anti-immigrant sentiment after the abolishment of discriminatory immigration policy 28 3.3.1 Public discussion on immigration in 1980s 28 . 3.3.2 Emergence of anti-immigrant party and cultural concerns ; 30 3.4 Economic situation, racial/ethnic composition, and the public opposition to immigration 31 4 Data, Operationalizations and Model Specifications 40 4.1 Data used for analysis 40 4.2 Dependent variable 40 4.3 Independent variables 41 4.3.1 Socio-demographic variables 41 4.3.2 Contextual variables 43 4.3.3 Attitudinal variables and interactions with contextual variables .....45 4.4 Model specification and problems in measurement 49 4.4.1 Regression model 49 4.4.2 Problems in measurement 50 5 Results and Interpretations 55 5.1 Canada 55 5.2 Australia ..66 5.3 New Zealand 80 5.4 Summary and comparison : 87 6 Conclusion, Implications and Further Questions 92 Bibliography ••• 96 Appendices 110 Appendix A : Question wordings and coding 110 A . l Canada HO A.2 Australia 114 A . 3 New Zealand 117 Appendix B: Descriptive statistics 121 B. l Canada 121 B.2 Australia 122 B.3 New Zealand 123 LIST OF TABLES Table 4.1 The predicted directions and attributes of the independent variables 49 Table 5.1 Logistic regression on opposition to immigration, Canada (Models with socio-demographic and contextual variables) 56 Table 5.2 The estimated influence of single-variable manipulations on opposition to immigration (y= 1), Canada: Model 2 (1993) and Model 4 (2004) 57 Table 5.3 Logistic regression on opposition to immigration, Canada (Models with attitudinal and interaction variables) 62 Table 5.4 Ordered logistic regression on opposition to immigration, Australia (Models with socio-demographic and contextual variables) 67 Table 5.5 The estimated influence of single-variable manipulations on opposition to immigration (y = 0, 1 and 2), Australia: Model 2 (AES1996) and 4 (AES2001) 68 Table 5.6 Ordered logistic regression on opposition to immigration, Australia (Models with attitudinal variables and interactions) 72 Table 5.7 Ordered logistic regression on opposition to immigration, New Zealand (Models with socio-demographic and contextual variables) 81 Table 5.8 The estimated influence of single-variable manipulations on opposition to immigration (y = 0, 1 and 2), New Zealand: Model 2 82 Table 5.9 Ordered logistic regression on opposition to immigration, New Zealand (Models with attitudinal variables and interactions) 85 Table 5.10 Directions of influence and the frequency of statistical significance 89 Table 5.11 Summary of the overall results of statistical analyses and national contextual data 90 Table B . l Descriptive statistics of variables used in models, Canada 121 Table B.2 Descriptive statistics of variables used in models, Australia 122 Table B.3 Descriptive statistics of variables used in models, New Zealand 123 v LIST OF FIGURES Figure 3.1 National unemployment rate and public opinion against further intake of immigrants in Canada 1 (Environics Focus Canada survey) 33 Figure 3.2 National unemployment rate and public opinion against further intake of immigrants in Canada II (Gallup survey) 34 Figure 3.3 National unemployment rate and public opinion against further intake of immigrants in Australia (various surveys) 35 Figure 3.4 Rate of immigrants from Asia and public opinion against further intake of immigrants in Canada 37 Figure 3.5 Rate of immigrants from Asia and public opinion against further intake of immigrants in Australia 38 Figure 5.1 The estimated influence of unemployment rate changes on opposition to immigration (interaction with RIP), 1993 CES (y = 1) 59 Figure 5.2 The estimated influence of unemployment rate changes on opposition to immigration (interaction with RIP), 2004 CES (y = 1) 60 Figure 5.3 The estimated influence of common social/cultural value scale on opposition to immigration (interaction with RIP), 1993 CES (y = 1) 64 Figure 5.4 The estimated influence of common social/cultural value scale on opposition to immigration (interaction with RIP), 2004 CES (y = 1) 65 Figure 5.5 The estimated influence of unemployment rate changes on opposition to immigration (interaction with RIP), AES 1996(y = 2) , 70 Figure 5.6 The estimated influence of unemployment rate changes on opposition to immigration (interaction with RIP), AES 2001 (y = 2) 71 Figure 5.7 The estimated influence of economic anxiety scale on opposition to immigration I (interaction with UR), AES 1996 (y' =2) 73 Figure 5.8 The estimated influence of economic anxiety scale on opposition to immigration I (interaction with UK), AES2001 (y =2) 74 Figure 5.9 The estimated influence of economic anxiety scale on opposition to immigration 11 (interaction with RIP), AES 1996 (y =2) 75 Figure 5.10 The estimated influence of economic anxiety scale on opposition to immigration II (interaction with RIP), AES 2001 (y =2) 76 Figure 5.11 The estimated influence of cultural essentialism scale on opposition to immigration (interaction with RIP), AES 1996 (y =2) 77 Figure 5.12 The estimated influence of cultural essentialism scale on opposition to immigration (interaction with RIP), AES2001 (y =2) (y =2) 78 Figure 5.13 The estimated influence of unemployment rate changes on opposition to immigration (interaction with RAPO), 2002 NZES (y = 2) 84 Figure 5.14 The estimated influence of opposition to economic types of migrants changes on opposition to immigration (interaction with UR), 2002NZES (y =2) 86 vii ACNKNOWLEDGEMENTS I offer my deepest appreciation to the faculty, staff and my friends at the U B C . I owe particular thanks to Professor Fred Cutler, who continuously gave me many insightful comments on my thesis as my academic advisor. I also thank Professor Richard Johnston, Benjamin Nyblade, Alan Jacobs, Lisa Sundstrom, Angela O'Mahony for many suggestions on my paper in the coursework and at the presentation of the earlier version of this thesis in Comparative Canadian Workshop in November 2005. I especially appreciate Scott Waston for being a discussant at the workshop. I also give my best appreciation to Colin Trehearne for continuously discussing the topic to develop my idea. Special thanks to Professor David Ewick and Professor Steven Reed, who spent a tremendous amount of time for my education and helped me developing the basic idea of this thesis. Also, I would like to express special thanks to Masao Takeuchi, Kiichiro Arai, and other students in Professor Reed's zemi, as well as Professor David Walker and Yutaka Daijogo for commenting on the earliest version of this thesis. Lastly, I would like to appreciate Australian Social Science Data Archive for providing the Dataset of Australian Election Study 1996 and 2007, Bob Burge at Canadian Opinion Research Archive at Queens University, staffs at Australian Embassy Library in Japan, and Sekiguchi Natsue at Chuo University for helping me with collecting necessary dataset for analysis and ' literature. viii 1 INTRODUCTION Immigration has been one of the hottest topics in the public debate in immigrant-receiving liberal democracies in 1990s and 2000s. In many of those countries, "anti-immigrant" political parties gained a substantial public support, and public opinion polls often showed people's reluctance to accept more immigrants. On the other hand, the governments have also sought to attract more economically advantageous, skilled and affluent immigrants, and gradually modified racially/ethnically discriminatory immigration policy to become less discriminatory. Thus while immigration policy has been liberalized to meet domestic labor demands as the movement of people crossing over national borders has grown in recent decades (Castles and Miller 2003; Freeman 1995; Cornelius and Tsuda 2004), people generally do not welcome many immigrants (Simon and Lynch 1999; Money 1999; Lahav 2004: 86-87). Several empirical studies explain voting behavior of anti-immigrant party supporters, and the history and development of immigration policies have been well described, but public opposition to immigration has not been analyzed sufficiently in the political science literature. Most particularly, a systematic and comparative examination of opposition to immigration in public opinion polls based on relevant theories is rare. A number of questions remain unanswered: Under what conditions are people more likely to oppose immigration? Who is more likely to oppose immigration? Why do they oppose it? What aspect of immigration acceptance is of the greatest concern? The aim of this study is to systematically analyze and compare the influence of economic and cultural factors on opposition to immigration among the public by applying several theories of intergroup relations. I aim to contribute to the understanding of politics of immigration. The study uses several election studies of Canada, Australia and New Zealand in different years, which are merged with census data to examine contextual effects. One of the important findings of this study is that both economic and cultural factors strongly increase opposition to immigration, but that cultural concerns influence opposition more strongly than 1 economic concerns. The three countries are chosen because they accept a large number of immigrants with various racial/ethnic and cultural backgrounds, they have had a similar historical and political experience on immigration policy, and the issue became politically salient in recent years. The percentages of permanent settlers in their total population, as well as the stock of foreign-born population is relatively larger in these countries than many other O E C D countries,' and the country of origin of accepted immigrants is quite diverse (OECD 2006: 32, 45, 235-243). They all changed immigrant policy drastically from a "White" Canadian/Australian/New Zealand policy to a non-discriminatory one, officially or unofficially adopted an idea of multiculturalism,2 and established a point system to assess the economic viability of immigrants as a selection criterion (about the comparability of immigration policy of three countries, see Ongley and Pearson 1995). Thus, there are many similarities. The method of comparing similar countries is called a "method of difference" or "most similar system" design, in which the different outcomes are more likely to be explained by the different characteristics of the countries i f Mil l ' s method is applied (Przeworski and Teune 1970: 32-46). This means, at a general level, that the differences of the three countries are less likely to be explained by the shared characteristics of them including their higher level of immigration acceptance, the diverse racial/ethnic and cultural backgrounds of immigrants they have accepted, liberalized immigration policy and its history, official (and unofficial) multiculturalism policy, and so on. 3 Fortunately, the major findings are stably the same in three countries, with some differences in 1 The percentage of annual permanent settlers intake in the total population in 2004 is 0.74 in Canada, 0.83 in Australia, and 1.02 in New Zealand, while the average of 17 O E C D countries is 0.46 (OECD 2006: 32). The stock of foreign-born population in 2004 is 18.0% in Canada, 23.6% in Australia, and 18.8% in New Zealand, while the average of 26 O E C D countries is 10.6% (OECD 2006: 45). 2 While Canada and Australia adopted multiculturalism as an official policy, New Zealand has never officially done so. New Zealand's immigrant settlement policy, however, has leaned towards the direction of multiculturalism since 1986, and its policy is comparable to Australian case (Fletcher 1999: 6-7, 16-22). 3 Certainly, there is a problem of applying Mi l l ' s method to a small number of cases in comparison (Lieberson 1991; Collier, Mahoney, and Seawright 2004; Ragin 2004), and the similar historical experiences of immigration policy and acceptance patterns could lead to different outcomes. 2 the interaction effects of some variables. The differences in the way in which the interaction variables work in three countries are explained not by the similarities described above but by the contextual economic situation. This thesis is composed of five chapters. First, major theoretical explanations of intergroup relations are introduced in Chapter 2, with an emphasis on the theory of prejudice and perceived threat theory. Considering the existing theories, I suggest a research strategy of comparing economic and cultural factors in the last half of Chapter 2. In Chapter 3, the historical background of immigration policy and public attitudes to immigration in the three countries is briefly reviewed. The comparison of their historical paths suggests that both economic interests (mainly domestic labor demands) and cultural concerns (often driven by racism) have shaped immigration policy and the public perception of it. Chapter 3 also examines the relationship between the indicators of economic and cultural concerns and the level of opposition to immigration over time. Chapter 4 discusses operationalization of concepts, model specification, and the problems in measurement of the main statistical analysis, and Chapter 5 shows the results of the analyses. Despite the varied measurement of attitudinal variables in three countries, I repeatedly obtained the same results, which suggests that cultural concerns have a stronger influence on opposition to immigration in three countries. Chapter 6 provides a brief conclusion of the findings and poses further research questions in order to better understand public's opposition to immigration. 2 THEORY AND RESEARCH STRATEGY 2.1 Overview The purpose of this chapter is to provide the theoretical background on race and ethnic relations and discuss possible applications thereof in explaining opposition to immigration. Many theories were developed in the United States and Europe to explain attitudes towards other social groups as well as related policy preferences. Sears, Hetts, Sidanius, and Bobo (2000: 16-31) put the theories in three categories: sociological approaches to racial prejudice; social structural approaches focusing on the power and resources of social groups; and approaches in terms of race-neutral political values. Forbes (1997: 22-36) classifies the theories of ethnic conflict into three clusters: contact, realistic conflict, and social identity theory. This chapter touches on most of them, but for the purpose of focusing on immigration issues, the emphasis is placed on the theory of prejudice, group position and social identity, and contact. The possible implications of the theories are discussed when applied to explain opposition to immigration: In the last part of this chapter, the intersection of those theoretical trends is discussed, in order to restructure the whole argument and generate a research strategy for this thesis. From the literature review in this chapter, I argue that the theory of prejudice and perceived threat to group position and identification can jointly be used to specify the sources of opposition to immigration, and that the contact theory can also be applied to examine the influence of contact in a competitive situation. I further argue that a systematic comparison of economic and cultural factors in determining individuals' attitudes towards immigration policy can provide a good understanding of why people oppose immigration. 2.2 Theory of prejudice and political ideology The theory of prejudice argues that individuals' racial or ethnic prejudice, or their negative images of other social groups, which developed in their youth through a socialization process, makes them oppose liberal policies on race and ethnic relations (Sears, Hetts, Sidanius, 4 and Bobo 2000: 16). Allport (1954) defines negative ethnic prejudice as "an antipathy based upon a faulty and inflexible generalization," (9) which can be conceptually differentiated from group conflict (8). Link and Oldendick (1996), for example, find that whites' "social construction differentials" of the racial groups, measured by subtracting the degree of negativity of their self-group (white) image from that of the other racial groups (Asians, Hispanics and blacks), affect their perception of the groups' effect on the economy, politics, and society in general in the United States (161-162). Since their operationalization of the social construction differentials is based on the questions asking how lazy/hardworking, unintelligent/intelligent, and violent/peaceful the other and their own racial groups are (155-157), their findings suggest that whites' racial prejudice is positively related to their perception of, and attitudes to, the outgroups Although racial prejudice still persists in the US, the literature often suggests that the "old fashioned," "biological," "blatant," or "Jim Crow" racial prejudice against blacks or racial discrimination in principle declined noticeably since 1960s (Schuman, Steeh, and Bobo 1985: 73-86; Kinder and Sanders 1996: 95-98). Thus more and more whites have agreed to racial equality in principle, however, most whites still oppose some specific race related policies which help other racial groups and ethnic minorities (Schuman, Steeh, and Bobo 1985: 86-104; Feldman and Huddy 2005). According to Steeh and Krysan (1996: 130-133), for example, about 60 to 80 percent of whites have consistently opposed affirmative action policy in jobs, special aid, and certain quotas for blacks and other ethnic minorities. 2.2.1 Symbolic racism, modern racism, and racial resentment theory Some political scientists provide a good explanation to this puzzle in 1970s, by arguing that a new type of anti-black prejudice has emerged, replacing the old "Jim Crow" type of racism. The new racism is called modern/subtle/symbolic racism, or racial resentment (Sears 1988; Henry and Sears 2002; Kinder and Sanders 1996; Kinder and Mendelberg 2000). According to this theory, this new racism is composed of "a blend of anti-black affect and the kind of 5 traditional American moral values embodied in the Protestant Ethic" (Kinder and Sears 1981: 416). Specifically, Henry and Sears (2002: 254) argue that this racism can be described as a belief that; ... racial discrimination is no longer a serious obstacle to blacks' prospects for a good life; that blacks' continuing disadvantages are due to their own unwillingness to take responsibility for their lives; and that, as a result, blacks' continuing anger about their own treatment, their demands for better treatment, and the various kinds of special attention given to them are not truly justified. The important point is that the concept not only concerns the covert racial prejudice or anti-black affects but also their general political views on whether certain groups deserve receiving help through policy, or if their claims are legitimate. As Kinder and Sanders (1996: 109) write, i f "prejudice is thinking il l of others without sufficient warrant," what they mean by symbolic racism or racial resentment is not simply a measure of prejudice. Another important component of this theory is that the symbolic racism emerges as a result of symbolic predispositions acquired in one's childhood (Sears 1988: 58); in other words, an instant emotional reaction to a racial symbol which they learned to like or dislike before adulthood. Students of this theory also point out that this racism should be conceptually distinct from the tangible threat or calculation of self-interest, and show that opposition to race-related policies such as affirmative action to assists blacks is not driven by whites' individual self-interest or benefit (Sears, Hensler, and Speer 1979; Kinder and Sears: 1981; Sears and Kinder 1985). For example, Kinder and Sears (1981) find that whites' opposition to the policy of bussing children to integrated schools is unrelated to whether they have children at that age; neither is it related to living in an area where they feel racial threat. Since the early 1970s, the new racism theorists have accumulated evidence that this new racism strongly determines whites' attitudes towards race-related policies in the US (Sears, Hensler, and Speer 1979; Kinder and Sears 1981; Huddy and Sears 1995; Kinder and Sanders 1996; Sears, Van Laar, Carrillo, and Kosterman 1997; Kinder and Mendelberg 2000. For a review of the literatures, see Sears, et al. 2000: 16-20; Sears 1988: 59-61). For example, Sears, et 6 al. (1997) test the effects of symbolic racism scales on whites' opposition to: (a) the policies of providing equal opportunity for blacks (such as the government's assurance of fair treatment of blacks in.jobs and facilitating school integration); (b) the policies which financially assist blacks, and; (c) affirmative action policy in employment and education, using several surveys from 1986 to 1995. Although the items used to measure symbolic racism vary by survey,4 their findings are consistent: symbolic racism is a stable, strong predictor of all three dependent variables, after controlling for the old-fashioned racism, anti-black affect, and political ideology (28-31). It is worthwhile noting, however, that the symbolic racism can be predicted by the last two components of its definition (34-36). 2.2.2 Prejudice or political ideology? Criticisms of Symbolic Racism Theory The new racism theory has been criticized for its obscure definition and labeling, inconsistent operationalization of the concept with heterogeneous variables, and narrow understandings of "self-interest" (Bobo 1983; Schuman, Steeh, and Bobo 1985: 178-189; Sniderman and Tetlock 1986). The symbolic racism theorists fire back by arguing that the complexity of the concept and inconsistency of measurement are due to the struggle to develop a consistent notion of the new racism based on a secondary analysis of general survey questions (Sears and Kinder 1985: 1145; Sears 1988: 55, 63). Most recently, however, Henry and Sears (2002: 265-269) show the measurement of symbolic racism has been internally consistent and validly discriminated from the effect of mere traditional anti-black affects and political ideology in predicting their attitudes towards race-related policy (see also Kinder and Sanders 1996: 109-115; Tarman and Sears 2005). The most trenchant criticism could be, however, that symbolic racism is just a compound scale of two different components, prejudice and political ideology, which should be conceptually differentiated, and whose relationship should be clearly investigated (Sniderman 4 They are all composed either of "denial of continuing racial discrimination," "absence of positive emotions toward blacks," a normative view that "blacks should work harder," "excessive demands" are made by civil rights leaders, and the sense that blacks have received "underserved advantage" (Sears et al. 1997: 25-26). 7 and Tetlock 1986: 132). The critics emphasize the role of race-neutral political ideology or the principle of fairness, which underlies the attitude to race-related policies, and argues that such ideology is not necessarily driven by racial prejudice (Sniderman and Piazza 1993: 5-8, 176-178; Sniderman, Carmines, Layman and Carter 1996; Sniderman and Carmines 1997: 6-11, 144-146; Sears et al. 2000: 27-31; Sniderman, Crosby, and Howell 2000: 253-262). They do not deny that whites' racial prejudice, whether they are sophisticated or not, influences opposition to the affirmative action (Sniderman and Piazza 1993: 51-56; Sniderman and Carmines 1997: 17-18), but they emphasize that some conservatives oppose affirmative action policies because of their political standpoint that the government should not spend extra money on welfare, regardless of their level of prejudice against blacks (Sniderman and Piazza 1993: 136-165; Sniderman and Carmines 1997: 61-97). A recent study by Feldman and Huddy (2005) endorses this point, when they find that the effect of racial resentment on the level of support for race-related policies (the college scholarship program for the whites, blacks, middle class, and poor) is different for conservatives and liberals. Their experimental surveys show that the conservatives are more likely to oppose the scholarship program for either racial or economic group (except for the middle class blacks), whereas liberals are more likely to oppose the program only when blacks are referred as the recipients, as the racial resentment scale increase (176-178). This suggests that for conservatives, racial resentment does not represent their racial prejudice, but has more ideological content, whereas for liberals, racial resentment scale functions as their level of prejudice, but does not represent their ideology.5 Thus the critiques of the new racism question the measurement validity of racial resentment and symbolic racism. 2.2.3 Application of the theory of prejudice to opposition to immigration There are several problems in applying the theory of prejudice developed in the field of 5 Feldman and Huddy also analyze the determinants of racial resentment for liberals and conservatives, and find that the overt racial prejudice does not play a decisive role for conservatives, while it clearly does for liberals. They try to specify the ideological origins of the racial resentment, and find that the sense of equality and individualism are the significant predictors (178-180). 8 racial relations directly to predicting the level of opposition to immigration. Two points stand out: (a) immigrants are not necessarily composed of different races or ethnic groups from the native-born people, and; (b) it is unclear how, and what kind of ideological value is related to the acceptance of immigrants. First, while some prejudice about immigrants could be generalized in a country, "immigrants" are still an elusive social group as a target of prejudice compared to specific social groups such as "blacks" in the US. In other words, the prejudice against immigrants could be completely different depending on what racial or ethnic group individuals imagine as immigrants. Second, as opposed to the conventional view which sees blacks as the target of affirmative action, equality promotion and welfare assistance, people may not have a strong consensus that immigrants are the recipients of those benefits, and accepting immigrants does not necessarily mean assisting them. One can imagine that conservative whites oppose a policy to assist blacks, because the conservatives often think that the government should not spend extra money on welfare assistance to anybody. Yet it is hard to imagine that only the conservatives oppose further immigration by the same logic. 6 Having said that, some political scientists apply ideas from the theory of prejudice and attempt to explain the level of opposition to further immigration. For example, Citrin, Green, Muste, and Wong (1997) find that individuals with more negative feelings to ethnic minorities tend to favor a restrictive immigration policy in the US: the whites who rated Hispanics and Asians relatively lower than whites in the feeling thermometer are much more likely to oppose further immigration (866-867, 872-875, 879). If we can assume that people in the US know or think that recent immigrants are composed of Hispanics and Asians, it is certainly legitimate to expect that the racial or ethnic prejudice against those groups of immigrants causes opposition to further immigration. Citrin, Reingold, and Green (1990: 1136-1142) find that the normative American identity, such as "Americans should speak and write in English" (1135) significantly 6 One possible interpretation, however, is that conservatives oppose accepting immigrants more because they think,the government should not spend money or resources on immigrants, but rather on the people living in a country. However the attitude that resources should be distributed among the native people rather than immigrants is widely shared by the natives and not particularly strong among conservatives. 9 influences the degree of negative reactions to the impacts of other ethnic groups on Californian society, while personal economic well-being and the respondents' perceptions of Californian economy do not explain it well . 7 According to them, this is because opposition to immigration is caused by symbolic predisposition about immigrants or the attitudes to some symbols that distinguish immigrants from native people, which can be represented by the political norms about what Americans should be like. Accordingly, the theory of prejudice expects that individuals' sophisticated racial prejudice and traditional.values, which are acquired through political socialization in early years, affect attitudes to race-related policies, whereas their self-interests do not. Concerning attitudes to immigration policy, it implies that opposition to immigration could be determined by individuals' prejudice against particular social or cultural immigrant groups, as well as their normative views about natives. 2.3 Theory of group position and identification 2.3.1 Perceived threat to group position/realistic conflict theory The perceived threat to group position theory suggests that individuals' perceived threat from outgroups, which share interests in conflict with their own group, facilitates their hostility towards the outgroups as a "defensive reaction against explicit or (usually) implicit challenges to the dominant group's exclusive claim to privileges" (Quillian 1995: 588). In contrast to the theory of prejudice, this theory emphasizes the role of perceived threat to one's identified groups' status, privileged position, power, and other social and economic resources in determining one's political attitudes, whereas prejudice plays a minor role (Coser 1956; Blumer 1958; Bobo 1983; Giles and Evans 1986; Bobo 1988; Glaser 1994; Bobo and Hutchings 1996; Bobo 1999; Bobo 2000). According to Blumer (1958: 3-7), racial prejudice is just a function of a sense of group position, and it emerges when the members of a racial group perceive threats from 7 Citrin, Green, Muste, and Wong (1997) also present a similar finding that the respondents' self-interest, measured by the evaluation of their financial situation or employment status, is an irrelevant factor in predicting opposition to further immigration. 10 other racial groups. As a classic example of this claim, Sherif, Harvey, White, Hood, and Sherif (1961) provide persuasive experimental evidence. They took children (all protestant, middle-class white) to a summer camp and let them join one of two groups to do some group activity. This was done so that they have no contact with the members of the other group at the initial stage. Then they provided some games which let each member of the group cooperate within their own group but the two groups must compete for some exclusive rewards. They found that the children were hostile to members of the other group with some prejudiced views about them, only after they perceived a threat from the other group in a competitive situation. The group competition and the subsequent realistic threat should be considered as the only reason why the children held hostility and prejudice against the other group, because there were no racial, ethnic, or cultural differences between the groups, and no conflict between those groups before they started to compete. This basic idea of perceived threat to group position is shared by the realistic conflict theories, which emphasizes the role of conflicting material interests (LeVine and Campbell 1972: 29-31; Bonacich 1972; Olzak 1992). According to LeVine and Campel (1972: 29), "real conflict of interests causes intergroup conflict" and "real threat," (perception of realistic threat), which causes "ingroup solidarity" (31) and "awareness of own ingroup identity" (32). Importantly, the threat may or may not be "real," but what matters in forming the attitudes is the perception or belief that outgroups threaten their own group (LeVine and Campbell: 41-42; Golder 2003: 438). Olzak's (1992: 76-84) extensive research on ethnic conflicts, protests, and other forms of ethnic collective actions reported in major US newspaper from 1877 to 1914 shows a clear pattern: when the economic resources for white Americans are threatened, because of increased immigrant inflow and business climate change, racial violence and ethnic conflicts took place between whites and blacks more often than during their economic prosperity. Bobo and Hutchings (1996) and Bobo (2000: 147-154) provide micro-level evidence that individuals' perceived costs and benefits to respondents' identified racial groups about preferential treatment of blacks in employment, housings and political representation leads them to negatively react to affirmative action programs. 2.3.2 Social identity theory Social identity theory shares similar ideas with perceived threat theory in that both emphasize conflict among groups and the role of members' attachment to their own groups (Forbes 1997: 29). It argues that any group categorization, however meaningless it might be, can lead its members to ingroup favorism and outgroup antagonism (Tajfel 1970; Tajfel and Turner 1979; Tajfel 1982; Hogg and Abrams 1988; Hogg, Terry and White 1995). The important difference between social identity theory and perceived threat theory is that the former does not require objective threats to be involved in the formation of attitude to the ingroup and outgroups: Social identity theory sees that social categorization and identification is a sufficient condition for group conflicts, and perceived threat is not necessary. In short, it could be argued that the perceived threat theory is more economic oriented, whereas the social identity theory is more psychological oriented (Forbes 1997: 29). TajfePs (1970) classic experiment well illustrates this point. After asking respondents to judge their preferences on several pictures drawn by "Kandinsky" and "Klee," he told them that they belong either to Kandinsky or Klee group in a way that nobody knows which group other respondents are assigned to. By doing so, they hold a kind of social identification with their assigned group, but the group difference is meaningless in fact, because they are'actually randomly assigned to either group regardless of their preference of the pictures. Tajifel then asked each respondent individually to distribute some money to two randomly selected other respondents, whose names were not identified, but so that respondents only knew which group they belong to. In this situation, although the respective social category was completely meaningless (subjects were randomly assigned), and no objective group interests or threats existed in their action (they didn't know who exactly received the money, and there are no personal or group rewards from any action), he still found that respondents distributed the money significantly in favor of the ingroup, and disadvantageously to the outgroup. There are, however, some criticisms of the social identification theory. First, Insko, 12 Schopler, Kennedy, Dahl, Graetz, and Drigotas (1992) run a more sophisticated experiment similar to Tajifel's, using both ideas of social identity and realistic conflict (or competition among groups), and find that the mere social categorization did not cause subjects to maximize their group difference, whereas the subjects who experience competition with the other group did. The implication is that a mere social categorical differentiation may not be sufficient for the conflict. The level of intergroup contrast and attitudes of hostility, however, should vary by the level of strength of social identity, so that the individuals who see a certain social category as important for their identification.may express stronger hostility towards the other social category. Riketta (2005) for example shows that both university students with higher group identification with their study major and those with a stronger perception of intergroup competition with the students of other study major, try to contrast the differences of social groups. This implies that whether social categorization matters or not depends on the strength of a particular social category in the society. The second criticism is then that the social identity theory does not clearly theorize which group boundaries are more likely to be salient than others (Forbes 1997: 37). To be sure, some social categories are more powerful sources of identification than others in a particular context, and some social categories overarching smaller categories may attenuate the effect of intergroup hostility caused by smaller categories (Gaertner, Dovidio and Bachman 1996; Marcus-Newhall, Miller, Holtz and Brewer 1993; Brewer 1996). This means that what types of social categories are salient at a certain period of time is important in assessing the effect of social identification on the intergroup conflict. 2.3.3 Application of the theory of group position and identification to opposition to immigration The perceived threat theory can be directly applied to predict attitudes towards immigration policy: it implies that the level of perceived threats from immigrants, distinct from racial or ethnic prejudice, should increase opposition to further immigration. This application is especially straightforward when it is economic. For example, Quillian (1995: 597-603) reports that a perceived economic threat to the people in a country, measured by the interaction of the proportion of immigrant inflow and the inverse G D P growth rate, increased the racial bias scale in 12 European countries. At the macro or aggregate level, a strong correlation between the temporal fluctuations of the public's opposition to accepting immigrants and the national unemployment rate, the level of the immigrants inflow, or the interaction of them can be best explained by the perceived economic threat (for example McAllister 1993 in Australia; Palmer 1996: 182-184 in Canada, and; Barkan 2003 in California, the US. As examples of explaining votes for anti-immigrant parties, Lewis-Beck and Mitchell 1993; Jackman and Volpert 1996; Fetzer 2000: 81-86; and Golder 2003). Thus Coenders and Scheepers (1998) are correct when they argue that the temporally static characteristic of racial prejudice does not provide a good explanation to why opposition to further immigration fluctuates over time (See also Monroe, Hankin, and Van Vechten 2000). The assumption for the aggregate data analysis is that individuals actually perceive economic threat from immigrants through the level of national economic indicators (unemployment rates) and the level of immigrant inflow. Palmer (1996: 186-189) provides individual level evidence that perceived economic threat such as "immigrants taking jobs away" from natives strongly triggers the respondents' opposition towards further immigration in Canada. On aggregate (by country) level, Lahav (2004: 189-194) also shows that the public's perceptions about the employment situation is highly correlated with the perception of the number of foreigners in a country. Applying the social identification theory to explain opposition to further immigration is also possible. First, the social category of "immigrants" signals natives that they belong to a large social group of state citizens, as opposed to people from foreign countries. Thus the distinction between immigrants and natives is one of the possible boundaries to promote opposition to immigration, so that the level of opposition to immigration is different among the native-born and foreign-born. Goot (1984: 28-29) sees that more Australian-born oppose immigration than 14 foreign-born by about 20 to 30 percent points in 1984 surveys, and a similar result is obtained in 2003 surveys (Goot and Watson 2004: 265). Second, respondents can imagine any possible social attributes of immigrants which differentiate them from native people, as they regard immigrants as a different group to their own, such as "the other" ethnic, religious or linguistic group. In this case, "immigrants" become a coded category of such salient subgroup, and those subgroups that are believed to be predominantly native are expected to oppose immigration more strongly than other minority subgroups. For example, Goot (1984: 21-27) sees that the immigration issue became strongly associated with the "problem of Asians" after the Australian national identity and Asian immigrants were politicized during the Blainey debates in 1984. In this regard, opposition to immigration could mean opposition to Asians, which are considered to form a different social group from Australians in most Australians' minds. 2.4 Contact theory and contextual effects 2.4.1 Contact theory Contact theory starts with a simple notion that individuals' contact experience with members of outgroups, under certain conditions, will produce or reduce their prejudice about them (Allport 1954: 261-282). In the early development of the research, psychologists repeatedly found that mere contact does not improve or reduce ones' prejudice, but that the changes in attitudes follows a number of situational, psychological, and institutional factors in the contact (Allport 1954: 261-262; Amir 1969: 338; Amir 1976; Stephan 1987: 14-17; Pettigrew 1998: 67-69). For example, contact is likely to reduce the prejudice against outgroup when the groups in contact are perceived to have an equal status (Tsukashima and Monterol976: 156-159; see also Robinson and Preston 1976 for a modification); the contact is institutionally supported by some authority; the contact is more physically and personally intimate (Jackman and Crane 1986); the members of each group feel pleasant and the contact is rewarding; the ratio of a group in contact is well balanced, and so on. Among others, one of the strongest factors influencing the direction of attitudes is whether the contact is in a competitive or cooperative situation and 15 whether the social groups have a common or different goal to pursue (Cook 1985: 453-457; Stephan 1987: 21). Marcus-Newhall, Miller, Holtz and Brewer (1993) assigns either one of two different roles (1 or 2) to the subjects who identify themselves with either one of two social categories (A or B) to accomplish a common goal as a team. She created two superordinate experimental groups, one in which the different social group members do the same role (a cross-cutting model, represented by A l , B l , A2 and B2), and the other in which the same social group members do the same role (a convergent model, represented by A l A l , B2, and B2). They find that the intergroup bias (A against B or vise versa) is reduced after the contact in a cooperative situation, and the effects are particularly large for the cross-cutting model (see also Brewer 1996: 297-300). More precisely, the result shows that where members of different social groups must cooperate more closely by playing the same role, attitudes to the outgroup are improved through their contact experience, more than those who do not have to cooperate as much in the same contact situation. Thus, the contact theory suggests that contact magnifies the effects of the preexisting conditions of contact: favorable or hostile attitudes to outgroups before the contact are amplified by contact. A n interesting criticism has been made as the contact theory becomes more fine-grained about the conditions under which attitudes to outgroups improve. Stephan (1987: 31-33) points out that while psychologists have found a long list of conditions so that intergroup contact improves attitudes to outgroups, they fail to explain why the interracial relations in the United States has improved significantly over the few decades. He argues that while many required conditions for a favorable contact in laboratory research implies that the actual improvement of attitudes is difficult to achieve in a short time, the effect of the intergroup contact in the actual social settings may have a different effect in a long term. 2.4.2 Proportion of social groups as a contextual effect Some scholars apply the idea of contact theory to examine the effect of racial or ethnic heterogeneity in a geographical unit on the level of prejudice or attitudes to outgroups. They 16 often adopt other theories and argue that racial or ethnic heterogeneity improves or worsens ones' racial/ethnic attitudes, when racially or ethnically heterogenous neighbors provide more opportunities for contact with outgroups. The results are mixed: some report a negative relationship (Giles 1977; Giles and Evans 1986; Glaser 1994), some find a qualified positive one (Ford 1973; Kalin and Berry 1982; Kalin 1996), and others (Schissel, Wanner and Frideres 1989; Lubbers and Scheepers 2000) find no relationship between the contextual proportion of outgroups and the ingroups' racial or ethnic attitudes or prejudices. Furthermore, others offer a more complicated explanation between them with a different interpretation of the contextual variables (Kinder and Mendelberg 1995; Oliver and Mendelberg 2000; Blake 2003). Many scholars who find a negative relationship apply a perceived threat to the group position theory to explain its causal relationship. Glaser (1994), for example, in his study of attitudes to race related policy in the US from 1980 to 1988, finds that the more blacks in white respondents' residential area lead to less positive attitudes to blacks in the US South. He explains that this correlation is due to a racial group conflict: "as southern whites live in areas with greater proportions of blacks, they will perceive a stronger connection between black political gains and white political losses for the likelihood of these political gains in greater in such an area" (35). On the other hand, Kalin (1996) uses the "mere exposure" hypothesis, which claims that repeated exposures to stimulus objects develop a sense of acceptability of them, leading to a positive feeling about the objects (Zajonc 1968), to explain the positive relationship between ethnic presence and the comfort ratings of particular ethnic groups.8 There are at least two problems in logically connecting the proportion of social groups to the contact theory and other theoretical trends. First, the proportion of social groups may not itself be a valid measurement of perceived threat or opportunity for contact. Giles and Evans (1986) find that the concentration of blacks does not necessarily mean a higher level of perceived threat for whites Americans. The proportion of social groups itself should not be directly 8 It must be noted that Zajonc cautions against applying the mere exposure hypothesis to the interracial relationship, because the subjects' antecedent attitudes (affects) to the social groups are always involved in the relationship (2, 23). 17 translated as perceived threat, especially when it provides an opportunity to dispel prejudiced views on outgroups by contact (Kinder and Mendelberg 1995: 420). On the other hand, the higher opportunity for contact does not necessarily lead individuals to have a close contact with the members of outgroups. Sigelman, Bledsoe, Welch and Combs (1996: 1321-1324) find that living in a racially mixed neighborhood increases the chance of whites' having a casual contact with blacks, while for most blacks, living in a racially heterogeneous neighbors does not necessarily lead them to a casual contact with whites. Other socio-demographic variable such as the contact experiences with outgroups in their youth, ages, and religious attendances are the important determinants of whether to contact with other social groups. Second, a natural selection bias or endogeneity problem, as well as the omitted variable bias may distort the true causal effects of racially/ethnically heterogeneous environment on racial/ethnic attitudes. In other words, the choice of residential areas may reflect ones' level of prejudice (endogeneity), and the proportion of the outgroups may correlate with other socio-demographic or economic variables, which affect ones' racial or ethnic attitudes. For example, Oliver and Mendelberg (2000: 574-576) claim that the perspective of environmental determinants on racial attitudes should be broadened from a mere proportion of different race in a community to a wider socio-economic context, and find that the higher percentage of the population with a tertiary degree, the lower the individuals' opposition to immigration in that area in the US. (in Canada, see Blake 2003). This means that some contextual variables other than the proportion of outgroups, such as socio-economic and educational level of residents in certain areas are, can influence individuals' attitudes to immigration policy as a reaction to their socio-economically stressful environment (Oliver and Mendelberg 2000: 583-586). Alternatively, whites' more tolerant attitudes to blacks in higher socio-economic contexts can be conformity effects by educated people, since the more educated are more likely to show tolerance to outgroups (Bobo and Licari 1989; Golebiowska 1995), and the larger size of them can shape other individuals to the dominant norm, converging their opinion (for theory of conformity and group size, see Stang 1976; Bond 2005). In any case, other variables than proportion of 18 immigrant population must be examined to specify how it affects opposition to immigration. 2.5 Integration of theories and research strategy: Comparing economic and cultural threat Recent arguments on race relations suggest that the abovementioned theories are not mutually exclusive (Huddy and Sears 1995; Bobo and Hutchings 1996; Forbes 1997; Bobo 1999 Sears, et al. 2000). Not only can each theory predict a responsive negative behavior to outgroups independently, but there is conceptual overlap among some ideas used in the theories. To be precise, some perceived threat or source of social identification can be a product of prejudice towards outgroups, and some part of prejudice also stems from perceived threats or social identifications. The answer to a hypothetical question, "do you feel immigrants are taking jobs away from the people already living in this country?"—often an indicator of perceived economic threat from immigrants—may just reflect ones' negative prejudice about immigrants. Rosenberg (1956) shows that once individuals hold a negative affect to certain objects, they are likely to believe that the objects tend to block individuals' important values. If this holds true about immigrants in general as the object, then the individuals with a negative image or prejudice about immigrants tend to perceive any sort of threat from immigrants, even though their material beings are not actually worsened.9 By the same token, a question measuring a negative prejudice about immigrants, such as "immigrants are lazy" or "immigrants are too demanding" can be interpreted as a result of perceived threat, i f respondents link their imaginary laziness or demands to deprivation of social welfare or erosion of their countries' economic success. Stephan, Ybarra, 9 The economic impact of immigrants, as expected, is not simple. Economists have examined the impacts of immigrants on natives' unemployment and wage level for a long time, but their mixed findings suggest that their economic impacts are indeterminate or infinitessimal (LaLonde andTopel 1991; Simon, Moore and Sullivan 1993; Foster and Withers 1994; Shan, Morris, and Sun 1999; Espenshade 2000; Addison and Worswick 2002). The mechanism of economic impact of immigrants on unemployment and wage is complicated. Borjas (1995: 5-11) for example, argues that immigrants are can create economic surplus in a country only when they lower the wage rates of native workers, and when the skill levels of immigrants are different from that of natives (and he stresses that their skill level should be higher. 13-16; see also Borjas 1999). The point here is, however, that the idea that immigrants steal jobs from native people often comes from the predisposed negative attitudes towards immigrants. 19 Martinez, Schwarzwald and Tur-Kaspa (1998: 560-561) use the theory of prejudice but re-interpret it as a "symbolic threat," or the "perceived group differences in morals, values, standards, beliefs and attitudes" (Stephan, Stephan and Gudykunst 1999: 619). They find that realistic threat moderately correlates with symbolic threat (bivariate correlation was between .53 to .55; Stephan et al. 1998: 568-569). Thus in this thesis, existing theories are not compared to argue which theory is more appropriate or powerful to explain the variation. This thesis aims to understand who opposes immigration and under what conditions and why do they oppose, and existing theories are used to predict the influences of socio-demographic, contextual, and attitudinal variables. Instead of examining which theory can better predict the level of opposition to immigration, this study poses a question that existing theories have not answered yet. Which factor, economic or cultural, more strongly determines the probability of opposition to immigration? Although this question is vital to the understandings of opposition to immigration as well as profoundly suggestive to the studies of voting behavior for anti-immigrant parties, it has been completely neglected until recent years.1 0 To be sure, the distinction between economic and cultural factors is not absolute as the existence of the term "socio-economic" suggests, and the factors that can be characterized neither by economic or cultural factors also influence opposition to immigration." I believe, however, that two factors are the major concerns in the immigration debate, and specifying relative strength of their influence at a general level will clarify exactly what they concerns about immigration. There are three ways to justify this argument. First, the notion of "threat" or "prejudice" is too broad to examine exact reasons why people oppose further immigration. Glaser (1994: 23) appropriately illustrates whites' perception of threats from blacks in the US as "a zero-sum view of politics," in which they feel "lose something valued to a rival group." In this case, what they think they would lose may not only be 1 0 The studies that explicitly mention and examine this question, which are discussed in detail later, include Fetzer (2000), Sniderman, Hagendoorn and Prior (2004) and Mughan and Paxton's (2006). 1 ' For example, concerns about law and order, environment, and concentration of population are related to opposition to immigration. 20 economic resources, but their imaginary cultural unity in a community or country, when they encounter "different" people of different skin color, language and religion. Since "threat" includes all these notions, the perceived threat to group position theory does not tell us in what aspect they feel threatened by blacks. Similarly, prejudice to outgroups is also an amorphous concept, whereby researchers can extend its meaning to cover a wide range of general misperception and negative affects about outgroups. Second, although the distinction between economic and cultural factors cuts across the theoretical distinction between theory of prejudice and perceived threat to group position and ( identification, each theory has certain affinities: theory of prejudice and social identity theory tend to emphasize the role of cultural factors, whereas the theory of perceived threat to groups and realistic conflict theory stress the economic factors in determining individuals' political attitudes (Forbes 1997; Stephan et al 1998; Stephan, Stephan and Gudykunst 1999; Fetzer 2000). Distinguishing between economic and cultural factors may contribute to the theoretical argument in a way to find which theory can better explain particularly opposition to immigration. Third and most importantly, systematic research that compares the impacts of economic and cultural factors on opposition to immigration is scarce. Sniderman, Hagendoorn and Prior (2004) conduct an experimental survey to compare the influences of economic and cultural threat on group hostility and stereotypes against immigrants and ethnic minority in The Netherlands in 1997 to 1998. In their strict coupled and decoupled experiments, they show that cultural (or national) identity was a much stronger trigger than economic anxiety in determining opposition to immigration. Though their study excels in comparing the influence of economic well being and national identity on exclusive attitudes to ethnic minorities at the cognitive level, it did not assess them in socio-demographic and contextual dimensions. In a systematic examination of such variables, Fetzer (2000) compares the effects of economic interests and cultural marginality using surveys and aggregate economic and demographic data. As Sniderman and others did, he also finds that the cultural marginality theory can better explain opposition to immigration at a cross-sectional level, but the temporal fluctuation of the level of opposition to 21 immigration highly correlates with the economic situation in the US, France and Germany. He concludes that both theories have consistent and distinct effects on opposition to immigration. Mughan and Paxton's (2006) elaboration of the psychological model to explain voting for the Australian anti-immigrant One Nation Party in 1998 shows that the "material threat" has a stronger impact on restrictive immigration policy than the "assimilationist threat" (347-354), suggesting the economic threat is a stronger determinant. The research interest of this thesis is in line with that of the abovementioned research. This thesis aims to systematically compare the impacts of economic and cultural factors using individual socio-demographic/economic, contextual, and attitudinal characteristics in surveys. By introducing socio-demographic categories and economic backgrounds, and by classifying them to economic and cultural factors, who opposes immigration can be carefully examined. Under what conditions people are more likely to oppose immigration can be assessed by introducing local contextual variables such as unemployment rate and proportion of immigrants of the total population, and they are interpreted either as economic or cultural (or both) influences. Finally, the influences of attitudinal variables which represent individuals' economic concerns and cultural concerns, are compared. In short, the individuals in an economically deprived position and situation, or those who feel economic anxiety are expected to oppose immigration more in economic terms. On the other hand, those in the social category which distinguishes themselves as an ingroup from immigrants as an outgroup, as well as those who perceive cultural anxiety are expected to oppose immigration more in cultural terms. 22 3 HISTORICAL BACKGROUND 3.1 Overview The purpose of this chapter is to provide a broader historical context for the statistical analysis of opposition to immigration in later chapters. A l l three countries have a long history of accepting a large number of immigrants, and there are certain similarities in their history of immigration policies and the public reactions to it (Ongley and Pearson 1995). First, they are a former colony of the British Empire and the majority of immigrants in the early history of nation-building came from the United Kingdom. In order to maintain the racial and ethnic homogeneity of the country, they adopted a racially/ethnically discriminatory immigration policy to exclude "unwanted" immigrants and favor British and Irish origins. After the World War II, they found that they needed to accept more immigrants from other countries to cover the labor shortages in the rapid economic development. Since then, the country of origin of immigrants became quickly became diverse including Southern and Eastern Europe, Asia and Caribbean, and the three countries finally abolished the discriminatory immigration system mainly for economic purposes. Opposition to immigration surged when the countries experienced economic downturn from the oil shocks of the early 1970s, but this also coincided with the diversifications of immigrants' racial/ethnic origins. From time to time the immigration issue rises to prominence in public discourse and racism and cultural concerns often appear. The issue of immigration became suddenly salient in elections during the 1990s in three countries, when an "anti-immigrant party" started to "play a race card" in the three counties. The level of salience and the way each party responded to the issue was different, but the idea of restricting and reducing the number of immigrants from the current level was widely supported by the public. Most importantly, such public's attitudes, as well as that of the political parties, involved not only economic concerns but also racial/ethnic and cultural concerns about immigrants. In the final section of this chapter, the relationship between the public opinion and economic and cultural indicators is analyzed. The positive 23 correlation between the national unemployment rate and the level of opposition to immigration shows that the national economic situation strongly influences the level of opposition to immigration among the public over time in Canada and Australia. On the other hand, no relationship is found between the rate of Asian immigrants and the level of opposition to immigration in Canada, while a strong relation is found in Australia. This is confusing; however, proxying cultural concerns by the rate of Asian immigrants is less than ideal. To what degree cultural concerns influence opposition to immigration over time is unclear from the preliminary analysis, mainly because no good indicator is available over time. By comparing the histories of immigration policy and the public reactions to the policies, however, this chapter argues that cultural concerns have played an important role in shaping the public opposition to immigration, while there is no doubt that economic situation strongly determined the level of opposition. 3.2 The history of immigration policy 3.2.1 Economic incentives and cultural concerns: Implementing a racially discriminatory immigration policy Immigration policy is developed along with the population planning, national identity formation and economic development in three countries (Hardcastle, Parkin, Simmons and Suyama 1994: 95-96; Fleras and Spoonley 1999: 150-152). While the government enthusiastically sought for immigrants' labor during the economic boom and restricted their entry during the recession, cultural concerns should also be considered as an important causal factor in the early days of immigration policy. Knowles (2000: 27-28), for example, introduced then vice-president of Trades, James Wilks' letter to the Prime Minister Laurier in Canada in 1900, which strongly insisted that the government should adopt the Alien Labour Act to restrict the entry of "ignorant, unfortunate... non-English-speaking aliens" who were perceived to provide serious social damages to the Canadian community. Racism and cultural concerns were a striking feature of the restrictive immigration policy in three countries, and it was later called White Canada/Australia/New Zealand policy. Such restriction was legally endorsed in the late 24 19th to the early 20th century, which coincided with a period of their economic downturn. Within ten years after Wilks' policy recommendation, The Immigration Act of 1906 and 1910 were stipulated in Canada to restrict the entry of immigrants with the racial background "unsuited" to the Canadian society (Knowles 2000: 32-34). In Australia, the Commonwealth Immigration Restriction Act of 1901 was enforced to formalize the exclusions of non-British and non-European immigrants by introducing an English dictation test, which succeeded the racial screening practices in place since the mid-19th century against Chinese and other ethnic minority immigrants (Curthoys 2003; Tavan 2005: 7-11). Similarly in New Zealand, the head tax for Chinese immigrants was introduced in 1880s, and the Immigration Restriction Act of 1899 and its amendment in 1910 provided a legal base for discriminatory immigration policy during the economic recession (Phillips 2006; Ongley and Pearson 1995: 773). Racially and ethnically discriminatory immigration policy continued through World War II, until the countries experienced a sustained period of economic development. Although the government was careful about using immigration as a tool for population and economic growth in Canada and New Zealand (Ongley and Pearson 1995: 767-768), strong labor demand pushed the government to allow a large number of immigrants to enter the country in the 1950s and 1960s. This is best described in a famous slogan often cited by the politicians at that time: "populate or perish" - John Diefenbaker used it for his electoral campaign in 1957 (Economic Council of Canada 1991: 13), Arthur Calwell "revived" it in his Parliamentary speech in 1945 (Jupp 2002a: 10-11), and the political activist Alfred E. Mander echoed it in New Zealand (Phillips 2006). Interestingly, however, the motivation for accepting more immigrants under the slogan of "populate or perish" is not merely economic: concerns about the smaller population and demands for a strong population growth are often expressed in the context of fear of invasion and social disorder. In Canada, Prime Minister Mackenzie King supported enlarging the number of immigrants in 1947 by referring to "the danger that lies in a small population attempting to hold so great a heritage as ours" (Reitz 2004: 100). He also said "I am sure [...] people of Canada do 25 not wish, as a result of mass immigration, to make any fundamental alteration in the character of our population" (Foster 1999: 70; 87). Five years later, King's administration enforced the Immigration Act of 1952, by which the government could formally restrict the entrance of immigrants based on the racial/ethnic backgrounds, social classes, place of birth, and cultural practices. Racial anxiety, which was entangled with the concerns about the national defense in immigration policy, was more frankly expressed in Australia, whose people still strongly remembered the Japanese air attack to Darwin during the World War II. According to Walker (2003), people's anxiousness about or fear of invasion by some Asian countries to their territory has caused Australia to accept more immigrants from particular sources of countries or regions and reject others (See also Tavan 2005: 11-29). Many Australians perceived that their large, "empty" lands cannot be defended with their small number of white population. Indeed, then Immigration Minister Arthur Calwell, who advocated the White Australian immigration policy, tried to recruit "only 'white' and preferably British immigrants" in this period, and agreed to accept a large number of displaced persons from Europe (Jupp 2002a: 11-12; Ozolins 1994: 203-204). An analogous situation applies in New Zealand. The immigration assistance scheme funded by the government was predominantly awarded to Britons and to some "white" Europeans, mostly Dutch (Phillips 2006). British immigrants were the most appreciated because they were considered as "instant citizens," who were perceived as not having a cultural difference from most of the New Zealand citizens at that time (Fleras and Spoonley 1999: 152).12 The stipulation of Immigration Act of 1964 endorsed the exclusion of non-European, mainly Asian immigrants and preference for British-origin immigrants (Ongley and Pearson 1995: 773). 3.2.2 Abolishment of the racial discriminatory immigration policy As the countries sought to accept more immigrants from the 1950s to 1970s for economic reasons, however, they noticed that they need to abandon their racially/ethnically discriminatory immigration policy and gradually changed their practices to non-discriminatory 1 2 Certainly,' Maori and other Aboriginal people were totally neglected from this prototype of New Zealand citizens. 26 ones. This shift is often described as a "triumph of economics over discrimination" (Ongley and Pearson 1995: 770). In Canada, the new regulation on immigration policy, which virtually abolished the racial/ethnic discriminatory selection practices was implemented in 1962, and the discriminatory selection criteria were replaced by the universal point system based on the economic and humanitarian characteristics in 1967 (Knowles 2000: 81-84). The older version Of Immigration Act was replaced in 1976, which practically systematized and formally liberalized Canada's immigration policy (Hawkins 1989: 70-75). In Australia and New Zealand, accepting Asian university students under the Colombo Plan in 1951 questioned the legitimacy of rejecting permanent residence of non-European immigrants (Markus 2003: 180-181; Phillips 2006). Since this is also coincided with the strong labor demands which were not fully sufficed by immigrants from the U K and Ireland, the Liberal government of Australia relaxed the racial/ethnic restriction in 1966, when it opened the gate for skilled non-Europeans who showed the ability to integrate to the Australian society (Jupp 1995: 209; Ozolins 1994: 206-207; Tavan 2005: 156-166). The White Australian immigration policy was formally terminated in 1973, when the Labour government announced that immigration selection would not involve racial/ethnic backgrounds, and a universal point system similar to Canadian one was introduced (Hawkins 1989: 103-106; Tavan 2005: 199-206). In New Zealand case, the liberalization process of the immigration policy was a bit prolonged. Long after the Labour government promised a nondiscriminatory immigration policy in its election campaign in 1972, the White New Zealand policy was formally abolished in 1986, when the country eliminated the regional preference clause and introduced a point system in 1991, whose criteria are mainly economic skills (Ongley and Pearson: 775; Fleda and Spoonley 1999: 162-163). This policy change and the idea of using immigration for covering the short term labor demands were well supported by the business sectors at that time (Ongley 2004: 205). Thus, the client politics model suggests, the relatively high volume of immigrants were accepted in response to the demands to pro-immigration bodies in the traditional immigration countries (Hardcastle, Parkin, Simmons and Suyama 1994; Freeman 1995; Cornelius and Tsuda 27 2004: 11-12). This implies that the economic incentive to accept more economically viable immigrants strongly influenced the liberalization process of the immigration policy in three countries. 3.3 Anti-immigrant sentiment after the abolishment of discriminatory immigration policy 3.3.1 Public discussion on immigration in 1980s The central interest of this research is on people's attitudes to immigration in three countries, not the immigration policy itself. The important distinction should be made between two questions: what causes government to adopt certain regulations on accepting immigrants, and what causes individuals to support or oppose accepting immigrants. The public's interest in immigration policy in general, is much lower than in other issues such as economic management, education, tax and healthcare (Tienhaara 1974: 4, 42-43). Still, immigration policy has been hotly debated among the public sporadically, and both economic and cultural concerns have been expressed. In Canada, the arrivals of 150 Tamils at the shore of Newfoundland in 1986 and 170 Sikhs in Nova Scotia in 1987 seeking an asylum raised the public awareness of immigraion and refugee policy (Foster 1998: 96). The issue became salient when the first Sikh student requested for wearing his turban hat for his graduation from the Royal Canadian Mounted Police training college instead of wearing the traditional Stetson hat (Holton and Lanphier 1994: 129). Troper (1994: 279-280) reasons that the hardening attitudes of the public to immigration policy in the early 1990s under the economic recession was the result of the increasing anxiety about the changing ethnic composition of the Canadian society due to the inflow of new immigrants from the non-traditional sources of countries. Economic concerns about the impact of immigration were also shared among the public, when a journalist Stoffman (1993) criticized the government for accepting "too many" family-class immigrants who had not been assessed by the point system. Since the debate about immigration policy under the economic downturn almost always involved the cultural concerns about the recent immigrants, it is difficult to distinguish whether 28 opposition to immigration is mainly driven by the economic or cultural concerns. The same is true for Australia. Immigration policy was hotly debated among the public at least twice during the 1980s. In 1984, a Professor of Australian History, Geoffrey Blainey expressed his concerns about the cultural and economic influence of Asian immigrants on Australian society in his book and media, which was soon criticized for its racist connotation (Hawkins 1989: 274-275; Goot 1984). In 1988, Fitzgerald Report advised the government to adopt a stricter immigrant selection criterion based on economic skills rather than accepting immigrants for humanitarian reasons so that immigrants could best contribute to the Australian society (FitzGerald 1988: 14-17, 51-56; Jupp 2002a: 47-48). Soon after this advice was announced, John Howard, then the opposition leader of the Liberal Party criticized the multiculturalism of Australia, and said that Asia immigrants should be reduced due to consideration of their capacity for social integration. The media critically reported his comment, which raised the public awareness about immigration policy, and then, Howard was replaced by Andrew Peacock for his unpopularity (Ozolins: 212; Jupp 2002a: 110-111). The average national unemployment rate was higher in the 1980s than in 1970s, and the general concerns about immigration policy, opposition to accept more immigrants appeared after the ethnic composition of coming immigrants significantly changed to Asians. Thus the economic concerns about immigrants were often entangled with the concerns about Asian immigrants (Goot 1984: 25), which cannot be separated from the cultural concerns. In New Zealand, too, the immigration debate during the 1990s was inseparable from the cultural concerns about rapidly increasing Asian immigrants, when Winston Peters politicized the immigration policy during his 1996 campaign (discussed in the following). 1 3 1 3 The public debate over immigration policy appeared later in the 1990s in New Zealand, partly because the liberalization of immigration policy took place in the late 1980s, which also proves the point that cultural concerns influenced the salience of the issue. The issue of concerns about immigration in the late 1980s was the legal legitimacy of visa-free entry of certain groups of immigrants (mainly from Pacific Irelands), when the media covered their overstay problem (New Zealand Planning Council 1991: 52). 29 3.3.2 Emergence of anti-immigrant party and cultural concerns The debate on immigration became most prominent politically, when an anti-immigrant party gained a substantive support. In Canada, although the extreme anti-immigrant message was largely suppressed by the leader of Preston Manning (Soberman 1999: 258; Laycock 2002), and the official party platform for immigration policy was not clear, Reform Party gained a substantive support in 1993 general election, partly because many Reform Party supporters are more strongly oppose further immigration than others (Lustzig and Wilson 2005: 119-122).14 According to Soberman (259-266), the media covered the Reform Party's criticism of the government's immigration policy during the election campaign, which sparked a wider debate on immigration. Many news articles appeared during this period representing public concerns about economic impacts of immigrants, but the difficulties of social integration of non-European immigrants were often implied in the story (Soberman: 262-266). In Australia, Pauline Hanson politicized the issue of immigration in the 1996 election, which successfully led her to be elected as an independent member of the Parliament (Goot 1998: 68-71; Jupp 2002a: 127-131; Goot and Watson 2001; and Gibson, McAllister and Swenson 2002 for the analysis the voting behavior of Pauline Hanson's One Nation Party supporters in 1998 election). It was clear that her message and the nature of the public discussion about immigration policy involved cultural concerns about Asian immigrants: in her first speech in the parliament, she said, I believe we are in danger of being swamped by Asians. [...] They have their own culture and religion, form ghettos and do not assimilate. [...] Immigration must be halted in the short term so that our dole queues are not added to by, in many cases, unskilled migrants not fluent in the English language" (Commonwealth of Australia Parliamentary Debates 1996). In New Zealand, too, the immigration policy attracted the public's attention when Winston Peters, 1 4 I should note that the main conclusion of Lustzig and Wilson's work is that the strong moral traditionalism is the single most stable, statistically significant attitudinal character of Reform/Alliance supporters, by which they can be differentiated from the supporters of other parties. According to their analysis, the attitudes towards immigration, crimes, economic management are not significantly significant (122-125), but this is simply because they include the regional control and the moral traditionalism factor in a regression to predict the vote for the Reform. 30 the party leader of New Zealand First, started his election campaign by promising a significant reduction of the immigration level at that time to less than 10,000 a year in 1996 election campaign (Miller 1998: 205; Trlin and Watts 2004: 116). He was much more careful about his reference to racial issue than Hanson, but his policy stance was severely criticized as racist by the media as soon as he mentioned that Asian immigrants are "invading" New Zealand. The electoral success of the New Zealand First in 1996, however, revealed his anti-immigrant campaign was not totally rejected by New Zealanders (Trlin and Watts 2004: 116). More importantly for the interest of this research, the emphasis was more strongly put on the changing racial/ethnic composition of immigration (Asianization) rather than the economic and influence of immigration in the public debate which appeared in the newspaper in 1996 (Munshi 1998). To summarize, the debate over the immigration policy appeared most saliently in the 1990s, and opposition to immigration was phrased in both economic and cultural terms. The popularity of the anti-immigrant parties in the three countries reflect the public's cultural concerns over the drastic change of the ethnic composition of immigrants from the traditional U K and European countries to Asian countries after the abolishment of the discriminatory immigration policy. 3.4 E c o n o m i c si tuat ion, racial /ethnic composi t ion , and the publ ic opposit ion to immigra t ion The national economic situation and immigration policy has a close relationship. The link is mainly caused by the nature of the labor market where fewer people try to immigrate to a country for fewer jobs available during the economic downturn, but possibly by the government's attempt to control the level of immigrant intake depending on the national economic situation, although the effectiveness of such practices is critically questioned (Hardcastle, Parkin, Simmons and Suyama 1994: 120-124; Cornelius and Tsuda 2004). In all three countries, the oil shock brought about a policy change to shrink the level of immigration by controlling the economic skills of immigrants (Economic Council of Canada 1991: 15-17; Money and Cole 1999: 168-172; Ongley and Pearson 1995: 769). Probably because of the both 31 factors, the annual immigrant intake is highly correlated with the national economic climate indexed by the unemployment rate. Again, the immigrant inflow, however, is not a good indicator of the public attitudes to immigration. Direct measurement of the public opposition to immigration is preferable, but the availability of such data is confined to two countries. Below is a preliminary analysis of the relationship between the level of opposition to immigration in the public opinion and economic and cultural indicators in Canada and Australia.' 5 Interestingly, many studies report that the public support of immigration policy declines when the economic situation is in decline (McAllister 1993; Goot 2000; Palmer 1996: 182-184; Fetzer 2000: 80-90). Figures 3.1 and 3.2 validate this point in Canada. The dashed line in Figure 3.1 shows the percentage of people who agree and strongly agree that "there is too much immigration to Canada" in different years of surveys. Figure 3.2 is from a different set of surveys, but both figures illustrate the close relationship between the aggregate level of opposition to immigration and the unemployment rate in Canada. Figure 3.3 is the Australian version showing the same trends, although the question wordings vary in different years of surveys. Thus, the figures show that the national economic situation strongly influences the general level of the public opposition to immigration, which implies that people who perceive general economic anxiety are more likely to oppose accepting further immigrants.16 1 5 The sufficient number of public opinion surveys is not available to conduct a similar analysis for New Zealand. 1 6 Holton and Lanphier (1994: 139) provide a different, unique interpretation of the link between economic condition and opposition to immigration. They argue that "existence of the recession is interpreted by public opinion as rendering existing level s of immigration and refugee intake as untenable and irresponsible," which well explain the attitudes of people whose economic status is not threatened or in insecure position. 32 Figure 3.1 National unemployment rate and public opinion against further intake of immigrants in Canada I (Environics Focus Canada survey)" Unemployment rate - -o- - "strongly agree' + "somewhat agree" 13 80.0 12 - 75.0 11 T V / o \» 70.0 _ _ 10 - /•' V / ; \x CO £ 65.0 to "to loyment r 9 8 a •1 \ • r \ -O / \ 1 f" ' \ J V2/° 1 60.0 + "some QL E 55.0 Une 7 Q t _ ro 6 -N: 13 Pearson's R: .70 » % % 50.0 Strongly 5 -% o 45.0 4 40.0 1975 1978 1981 1984 1987 1990 1993 1996 1999 2002 2005 year a. The questions wordings for public opinion are "Do you strongly agree, somewhat agree, somewhat disagree or strongly disagree with each of the following statements?—Overall there is too much immigration to Canada." The figures are the combined percentage of the answer, "agree" and "strongly agree." Source: For the public opinion surveys, Environics. Various years. Environics Focus Canada. [On-line]. Provided by Canadian Opinion Research Archive, Queens University, (no renewal date), Retrieved July 31, 2006: http://jeff-lab.queensu.ca/poadata/info/fc/fclist.shtml For the unemployment rate, Statistics Canada. Various years. Labour Force Survey Estimates. (Table 282-0002). [On-line]. Page updated Jan 4, 2006, retrieved July 31: http://estat.statcan.ca/cgi-win/CNSMCGl.EXE?regtkt=&C2Sub=&ARRAYID=2820002&C2DB=EST&VEC= &LANG=E&SDDSL0C=%2F%2Fwww.statcan.ca%2Fenglish%2Fsdds%2F*.htm&R00TDIR=ESTAT%2F& RESULTTEMPLATE=ESTAT%2FC1I_P1CK&ARRAY_PICK=1&SDDS1D=&SDDSDESC=; and Statistics Canada. Labour Force Characteristics. [On-line]. Page updated Jan 24, 2006, retrieved July 31, 2006: http://www40.statcan.ca/101/cst01/econl0.htm Figure 3.2 National unemployment rate and public opinion against further intake of immigrants in Canada II (Gallup survey)" a. The questions wordings for the public opinion are "If it were your job to plan an immigration policy for Canada at this time, would you be inclined to increase immigration, decrease immigration, or keep the number of immigrants at about the current level?" The figures are the percentage of the answer, "decrease immigration." Source: For the public opinion survey, Gallup Canada, Inc. Various years. Canadian Institute of Public Opinion (CIPO). [On-line]. Provided by Carleton University Library Data Center. Updated July 28, 2006, retrieved July 31, 2006: http://wwwJibrary.carleton.ca/ssdata/surveys/pop_gallup.html,'7 and; Jedwab, Jack. 2006. "Canadians Remain World Leaders in Openness to Immigration: Tomorrow we mark international day for the prevention of racial discrimination." [Online, press release report of Association of Canadian Studies, March 21, 2006] ACS Research & Polling - Canadian Opinion Online. (No update date). Retrieved July 31, 2006: http://www.acs-aec.ca/Polls/Openness%20to%20Immigration.pdf For the unemployment rate, see Table 3.1. 1 7 Gallup Canada collected the data, and Carleton University Library Data Center provided dictionaries and cleaned data under the auspices of National Archives of Canada. None of these organizations are responsible for the analyses and interpretations of data. 34 Figure 3.3 National unemployment rate and public opinion against further intake of immigrants in Australia (various surveys)" 12.0 10.0 6.0 3 4.0 2.0 0.0 - • Unemployment Rate -O- '- "too many," "reduce" or "too high" N = 24 Pearson's R = 0.84 1964 67 70 73 76 79 82- 85 91 94 97 2000 2003 80.0 70.0 60.0 i? 50.0 40.0 30.0 20.0 10.0 0.0 year a. The questions wordings and answer choices vary by survey, but all of them ask about the number of immigrants. The figures are the percentage of respondents who oppose to further immigration intake (the answer of "too many,' "reduce," "too high" and so on). If there are multiple choices to express their opposition to immigration, they are combined. Sources: For the public opinion survey, Gallup Australia, Inc. 1964. Opinion Poll D0317: Australian Gallup Poll, August 21] 1964. [On-line], Provided by Australian Social Science Data Archive (ASSDA), The Australian National University (ANU), (no renewal date), Retrieved December 6, 2003 from the World Wide Web: http://assda.anu.edu.au/polls/D0317.html; Gallup Australia, Inc. 1966. Opinion Poll D0324: Australian Gallup Poll, February 5, 1966. [On-line]. Provided by ASSDA, ANU, (no renewal date), Retrieved December 6, 2003 from the World Wide Web: http://assda.anu.edu.au/polls/D0324.html; National Multicultural Advisory Council, 1999. "Statistical Appendix." Australian Multiculturalism for a new Century: Towards Inclusiveness. [On-line], (pp. 39-44) Page updated November 30, 2003. Canberra: Commonwealth of Australia. Retrieved December 6, 2003 from the World Wide Web: http://www.immi.gov.au/multicultural/_inc/publications/nmac/statistics.pdf (for 1967, 1968, 1970, 1971, 1981, 1985, 1992 and 1995 data) Gallup Australia, Inc. 1977. Opinion Poll D0254: Australian Gallup Poll, March, 1977. [On-line], Provided by ASSDA, ANU, (no renewal date), Retrieved December 6, 2003 from the World Wide Web: http://assda.anu.edu.au/polls/D0254.html; Gallup Australia, Inc. 1984. Opinion Poll M0004: Morgan Gallup Poll, May, 1984 (Computer Reports). [On-line]. Provided by ASSDA, ANU, (no renewal date), Retrieved December 6, 2003 from the World Wide Web: http://assda.anu.edu.au/polls/M0004.html; Gallup Australia, Inc. 1988. Opinion Poll M0054: Morgan Gallup Poll, August, 1988 (Computer Reports). [On-line]. Provided by ASSDA, ANU, (no renewal date), Retrieved December 6, 2003 from the World Wide Web: http://assda.anu.edu.au/polls/M0054.html; Gallup Australia, Inc. 1989. Opinion Poll M0066: Morgan Gallup Poll, August, 1989 (Computer Reports). [On-line]. 35 Provided by ASSDA, ANU, (no renewal date), Retrieved December 6, 2003: http://assda.anu.edu.au/polls/M0066.html; McAllister, Ian, et al. 1990, Australian Election Study, 1990. [Computer file]. Canberra: ASSDA, ANU; Irving Saulwick and Associates. 1991. Opinion Poll DO'/'55: Saulwick Poll, October, 1991. [On-line], Provided by ASSDA, A N U , (no renewal date), Retrieved December 6, 2003: http://assda.anu.edu.au/polls/D0755.html; Jones, Roger, et al. Australian Election Study, 1993. [Computer file], Canberra: SSDA, A N U ; Jones, Roger et al. Australian Election Study, 1996. [Computer file]. Canberra: ASSDA, ANU; Newspoll. (No Title). [On-line], (no renewal date), Retrieved December 6, 2003 from the Worid Wide Web: http://newspoll.com.aU/image_uploads/cgi-lib.27322.l.04031MMI.pdf; Bean Clive, et a l . , 1998, Australian Election Study, 1998. [Computer file], Canberra: ASSDA, ANU; Gow, David, et a l . 2000. Australian Constitutional Referendum Study 1999. [Computer file]. Canberra: ASSDA, ANU; Bean, Clive, et al. 2002. Australian Election Study, 2001. [Computer file]. Canberra: ASSDA, ANU. Gibson, Rachel, et a l . 2003. The Australian Survey of Social Attitudes, 2003 [Computer file]. Canberra: ASSDA, ANU; Bean, Clive, et al. 2005. Australian Election Study, 2004. [computer file]. Canberra: ASSDA, ANU. For the unemployment rate, International Labour Organization. 1970-2005. Year Book of Labour Statistics, various editions. Geneva: ILO. There is, on the other hand, no research that suggests that public opposition to immigration policy fluctuates along with indicators of cultural concerns. The lack of such research comes about because obtaining a valid and reliable measurement of cultural concerns over time is difficult. Since the debate over immigration in Canada and Australia often involves the concerns about immigrants from Asia, I plotted rate of Asian immigrants as a percentage of the total permanent settlers in both countries. Figure 3.4 shows that there is no clear relationship between the annual level of Asian immigrants' inflow and the level of opposition to immigration among the Canadian public. The rate of Asian immigrants grew gradually from 30% in the mid 1970s to about 65% in the late 1990s, while the level of opposition fluctuates during that period. The percentage of Asian immigrants stayed at around 60% from 1995 to 2005, but the opposition to immigration significantly dropped during the same period. Thus the data shows that the rate of Asian immigrant inflow does not influence opposition to immigration in Canada. 1 8 I also tried rate of immigrants from other countries, including Africa, South and Central America, and Oceania. Only there is a moderate correlation (Pearson's R of .5) with each rate of the regions, but the relationship with Africa is negative. 36 Figure 3.4 Rate of immigrants from Asia and public opinion against further intake of immigrants in Canada." •"too much immigration" "decrease immigration" c o o - o - - Rates of immigrants from Asia 80.0 70.0 60.0 50.0 •s 40.0 30.0 20.0 "too much immigration" and unemployment rate N: 13, Pearson's R:-.21 » o °- a "decrease immigration" and unemployment rate N: 18, Pearson's R: -..32 70.0 65.0 60.0 ?^ 2 55.0 < B 50.0 45.0 C 2 OO 40.0 E B 35.0 o ID 30.0 <* 25.0 20.0 1975 78 81 84 87 90 93 96 99 2002 2005 year The country of Asian immigrants include Afghanistan, Bahrain, Bangladesh, Bhutan, Brunei, Cambodia, Republic of China, Cyprus, Hong Kong, India, Indonesia, Iran, Iraq, Israel, Japan, Jordan, Korea, Kuwait, Laos, Lebanon, Macao, Malaysia, Mongolia, Myanmar, Nepal, Oman, Pakistan, Palestinian Authority, Philippines, Qatar, Saudi Arabia, Singapore, Sri Lanka, Syria, Taiwan, Thailand, Tibet, United Arab Emirates, Vietnam, and Yemen. Source: For rate of immigrants from Asia, Citizenship and Immigration Canada. 1974-1996. Citizenship and Immigration Statistics various years. Cjttawa: Minister of Supply and Services Canada, and; Citizenship and Immigration Canada. 2004. Facts and Figures 2004: Immigration Overview: Permanent and Temporary Residents. [On-line]. Page updated July 11, 2005. Retrieved July 31, 2006 from the World Wide Web: http://www.cic. gc.ca/english/pub/facts2004/permanent/9. html. For the public opinion survey, see Figure 3.1 and.3.2. 37 Figure 3.5 Rate of immigrants from Asia and public opinion against further intake of immigrants in Australia." - o - - Percentage of "too many" •Rate of immigrants from Asia E o a. n. O 30.0 50.0 45.0 40.0 35.0 < £ o E E 30.0 c 25.0 20.0 1984 86 90 92 94 96 98 2000 2002 2004 year a. The calculation of rate of immigrants from Asia is not precise, because of the limited availability of the data. They include immigrants from Vietnam, Hong Kong, Philippines, China, Malaysia, India, Sri Lanka, Lebanon, and Chinese Taipei. Source: For the rate of immigrants from Asia, Organization for Economic Co-Operation and Development (OECD). 1995-2004. Trends in International Migration: Annual Report Various Editions. Paris: OECD, and; OECD. 2006. International Migration Outlook: Annual Report 2006 Edition. Paris: OECD. For the public opinion survey, see Figure 3 .3 . On the other hand, Figure 3.5 shows a strong correlation between the two in Australia. It shows that the peak of opposition to immigration in the early 1990s coincides with the massive Asian immigrants' inflow. Soon after this, the rate of Asian immigrants quickly subsided as rate of immigrants from New Zealand (about 10% in 2004) and U K (about 16% in 2004) increased. Although Australian data is a little distorted because the finer data was not available, the trajectories of the two lines clearly show a positive relationship. The problem, again, is the implication of this correlation. Does this suggest that cultural concerns influence the level of opposition to immigration in Australia? Do two contrasting results in Canada and Australia show the difference of salience and the degree of perceived threat posed by immigrants from Asia? 38 The answer to those questions is probably no, mainly because of measurement error. It could be possible that people would know more about the composition of immigrants when the immigration issue becomes salient in a country, and such salience is made because of the higher level of Asian immigrant inflow. Compared to national unemployment rate, however, the annual rate of immigrants from Asia is not as valid or reliable an indicator of cultural concerns people would perceive about immigrants. Unlike in the case of unemployment rate, people in aggregate or at individual level would be able to correctly guess how many percentages of immigrants are coming from Asia. Furthermore, those Asians and other ethnic minorities would permanently live in the country, which mean that they could possibly pose a cultural threat to some native people in later years than they immigrated. The stock of foreign-born population only accumulates over time, so the rate of people with certain ethnic or cultural backgrounds do not function as cultural concerns which should fluctuate over time. Accordingly, it is possible that the Figure 3.5 suggests a spurious correlation, while no correlation in Figure 3.4 is due to the measurement error of the level of cultural concerns. A valid and reliable indicator of cultural concerns about immigrants, like unemployment rate for the economic concerns, is necessary to compare the influence of both factors, but obtaining such indicator is quite difficult. Thus the preliminary analysis of the relationship between the economic and cultural factors and the public opinion about further immigration intake reveals that economic situation strongly influences the level of opposition, while racial composition of immigration does not. The problematic measurement of cultural concerns, however, biases the result of this analysis, and this is hard to be solved, because a good aggregate indicator of cultural concerns that varies overtime is difficult to obtain. The analysis and comparison of influences of those factors will be possible at the individual level cross-sectionally in the following chapter. The historical overview of immigration policy and the public reaction to immigration policy still suggests that cultural concerns often emerged in the public debate over immigration policy, even when the economic concerns and situation strongly influenced the level of opposition. 39 4 DATA, OPERATIONALIZATION AND MODEL SPECIFICATION 4.1 D a t a used for analysis For the analysis of opposition to immigration, I use the 1993 Canadian Election Study Incorporating the 1992 Referendum Survey on the Charlottetown Accord (hereafter shortened as 1993 CES. Northrup and Oram 1994), 2004 Canadian Election Study (2004 CES. Blais, Everitt, Fournier, Gidengil and Nevitte 2005), Australian Election Study 1996 and 2001 (AES1996 and AES2001. Jones, McAllister and Gow 1996; Bean, Gow, and McAllister 2002), and the 2002 New Zealand Election Study (2002 NZES. Vowles, Aimer, Miller, Banducci, and Karp 2002). For contextual information I use the Census of Canada 1991 and 2001, whose data is aggregated by Forward Sortation Areas (FSA) in Canada 1 9; the Census 1996, and 2001, both sized for 2001 electoral districts in Australia 2 0 (Kopras 7 November 2000; Kopras 3 March 2003); and the 2001 Census, sized for 2002 electoral districts for New Zealand (Statistics New Zealand 2002). The total number of respondents varies from 2010 (AES 2001) to 5783 (2002 NZES), and response rates also vary by survey method and category (e.g. telephone interview or mail-back), but overall, the response rates exceeds 50 percent in all the studies. 4.2 Dependent var iable In order to measure opposition to further immigration, this study uses a question asking how the number of immigrants accepted by the respondents' country should be—reduced, increased, or kept at the present level— as a dependent variable. By defining and operationalizing "opposition to (further) immigration" as such, the study can evade the otherwise ambiguous argument. First, the variable indicates individuals' political preference on immigration. Strictly speaking, the concept should be distinguished from the perception about 1 9 I appreciate Professor Fred Cutler for providing the 1993 CES dataset with the contextual variables. 2 0 Strictly speaking, the Census 1996 is adjusted for the 2000 electoral boundary, but the difference between 2000 and 2001 is negligible. 40 the number of immigrants, which is often captured by asking whether the current number of immigrants allowed to a country is too many or too few. 2 ' Second, the variable is also distinct from the affective evaluation of immigrants (anti-immigrant sentiment), racial/ethnic prejudice, or racism, which is often operationalized by asking how "lazy" or "diligent" (respondents think that) immigrants are, or by the position of a feeling thermometer on immigrants in general.22 Fortunately, all the election studies have an indicator of opposition to immigration as a policy with similar wordings in three countries, despite some slight difference in scaling in Canada, where the choices of answer are either "more," (immigrants) "fewer," or "depends/stay the same" (should be accepted). Since the interest here is the opposition to immigration, and it is unclear whether the respondents who answered "depends/stay the same" oppose immigration more strongly than those who answered "more," I recoded them into a binary variable for Canadian data. In Australia and New Zealand, the respondents can choose either (the number of immigrants should be) "increased a lot," "increased a little," "remain about the same as it is," "reduced a little" or "reduced a lot." For the same reason as for Canadian data, I recoded them to vary from 0 to 2, 0 for the first three categories, 1 for "reduced a little," and 2 for "reduced a lot" so that the number represent the ordered difference of anti-immigrant sentiment. For the detail of these variables, see Appendix A and B. 4.3 Independent variables 4.3.1 Socio-demographic variables Following the abovementioned theories in Chapter 2 and several results of the previous 2 1 They are strongly correlated, however: For example, the Pearson's correlation coefficient between the perception of the number of immigrants (whether respondents feel that they are "gone [much] too far," "about right" or "not gone/nearly far enough") and opposition to immigration mAES is 0.67 in 1996 and 0.69 in 2001. Using this variable as a dependent variable slightly changes the result, but it does not change the general conclusion of this thesis. 2 2 The statistical relation between the two is not so strong, but certainly correlated. The Pearson's correlation coefficient between the feeling thermometer on immigrants (respondents' feeling about immigrants, measured by the scale ranging from 0 [cold] to 100 [warm]) and opposition to immigration in 1993 CES is -0.29. There is no feeling thermometer for immigrants in 2004 CES, but the correlation with the feeling thermometer of racial minority is -0.27. 41 research that examined the causes of opposition to immigration (Tienhaara 1974; Goot 1984; Schissel, Wanner, and Frideres 1989; Citrin, et al. 1997; Palmer 1996; Fetzer 2000; Scheve and Slaughter 2001; Goot and Watson 2004), several socio-demographic variables are hypothesized to influence opposition to immigration. They are categorized into four groups: economic, cultural, both economic and cultural, and neither economic or cultural factors. First, many scholars repeatedly find that the better-educated show more tolerance to accepting immigrants. Then tertiary education (a binary variable, i f respondents have received university level education or above regardless of degree, 1 is assigned, and 0 i f not) is hypothesized to have a negative influence on opposition to immigration. Some find that gender (binary, 1 i f male) and age (absolute number) or age category (in 10s), and Quebec residency (binary, 1 i f living in Quebec, only in Canadian data) also influence opposition to immigration, but the direction of the influence is not stable. Since it is hard to classify the effect of these variables either to economic or cultural factor, they are introduced as control variables.2 3 As for economic factors that may influence opposition to immigration, 1 use unemployed (binary), manual, unskilled or lower economic class labor (binary), self-evaluated working class (binary), and yearly household income (interval variable with eight to eleven intervals, coded in an ascending order such as "less than $10,000" = 1, "$10,001 to $ 20,000" = 2, and so on), the assumption being that economically disadvantaged respondents are more likely to feel economic threat or hold economic prejudice against immigrants. Except for household income, all these binary variables should have a positive influence on opposition to immigration, because the unemployed, lower economic class labor and working class citizens are more economically vulnerable in terms of economic competition with immigrants. As for the cultural factors, born in a native country (for Canada and Australia, 1 i f 2 3 Tertiary education represents the degree of economic security when the more educated are more economically advantageous than the less educated in labor market or on wage (Scheve and Slaughter 2001: 141-142), but it also represents the degree of tolerance to other social/cultural groups and their ideas (Bobo and Licari 1988). Quebec residency is introduced because the perception of cultural threat, prejudice towards immigrants, and the sense of national identity as Canadian could be quite different in French Canada, and the ethnic and cultural characteristics of immigrants they accept are different because the province has its own immigration policy. 42 native-born) or self-evaluated ethnic identification with New Zealand and European (binary, for New Zealand only) and Maori (binary) is hypothesized to have a positive influence on oppose immigration, because they would feel threatened by their perceived immigrants' different, conflicting cultures and social values from the mainstream. For a similar reason, born in English speaking countries (binary) and Christian (binary) are hypothesized to have a positive influence, because they are also considered to have a stronger nominal identification with the cultural characteristics. English is the dominant, official language (except for Quebec) and Christianity is a dominant religion in three countries, whereas many recent immigrants do not speak English as their first language nor have a religious affiliation with Christianity.2 4 Interestingly, a previous study shows that the frequency of religious attendance is negatively related to the opposition to immigration. Soper and Fetzer (2000) find a negative effect of a church attendance on opposition to Islam immigrants in France and U K , and explain that the religious attendance provides a chance for exposure to religious elites' speech, which openly criticizes xenophobic attitudes and intolerance to other religions. For "nominal" Christians, their religious affiliation is more like an "ethno-cultural identity which she or he must defend, Crusader-like, against the Muslim, Middle Eastern 'invaders'" (182). Therefore, religious attendance (ordinal variable with six ordered choices from "never" go to religious services at all =0 to "at least once a week" = 5 for Australia and New Zealand) is hypothesized to have a negative influence on opposition to immigration, because a frequent religious socialization lessens their cultural prejudice and threat. For the detail of these variables, see Appendix A and B. 4.3.2 Contextual variables Contextual characteristics are introduced from each Census. First, as shown in Chapter 3, unemployment rate (UR) is highly correlated with the level of opposition to immigration over time. The conventional interpretation of this relationship is that UR represents an economic 2 4 Certainly there are various kinds of Christianity, and the effects of religious identifications on the opposition to immigration may be different by sect (see Fetzer 2000 in detail). The main purpose of this study, however, is to compare the two factors in a broad sense, so difference in the overarching cultural characteristics are not compared here. 43 threat of immigrants to the people already living in the country. If this is true, local UR should have a positive influence on opposition to immigration at individual level. In a same manner, rate of immigrant population (RIP) is hypothesized to have a positive influence on opposition in an economic sense, because higher RIP in the respondents' residential area poses an economic threat to their local economic resources. The literature provides good evidence for this when the interaction of UR and RIP is positively correlated with the probability of voting for anti-immigrant political party or opposing further immigration. (Lewis-Beck and Mitchell 1993; Quilian 1995; Coenders and Scheepers 1998; Money 1999; Golder 2003). Hence UR, RIP, and the interaction of them are introduced to represent an economic threat factor, and are expected to have a positive influence on opposition to immigration. Unfortunately, RIP is not available for New Zealand, so rate of Asian, Pacific peoples, and "other" ethnic groups (RAPO) is introduced as a proxy variable.2 5 Second, the influence of cultural threat can also be operationalized with some contextual variables. Rate of ethnic minority origins/birthplace is an ideal variable for this purpose, since immigrants contain a large number of British and Ireland origins, who are generally considered to share similar cultural values and characteristics. Introducing this variable, however, will cause a multicollinearity problem, because it is highly correlated with RIP (Pierson's correlation coefficient often exceeds .9). Therefore, RIP is assumed to represent a cultural threat, and hypothesized to have a positive influence as a cultural threat, i f individuals perceive that concentration of immigrants causes a cultural tension in a country.26 Interpreting RIP as a cultural threat is a little problematic, however, as discussed in Chapter 2. The physical presence of immigrants in the residential areas also provides a good chance to contact with them to improve their attitudes, and i f this is true, the predicted influence of the same variable is opposite to the cultural threat. If the concentration of immigrants provides a chance to contact which improves attitudes towards immigrants, RIP is hypothesized to have a negative influence 2 5 "Other" ethnic groups is a category where they are not categorized either as European, Maori, Pacific peoples, or Asians (Statistics New Zealand 2002). 2 6 A model with rate of ethnic minority origins/birthplace is tested in place of rate of immigrants, and the effect is generally weaker, but the overall result is the same. 44 on opposition to immigration. Third, some scholars find that rate of the tertiary educated has a strong impact on the attitudes to immigration (Oliver and Mendelberg 2000; Blake 2003), but what this variable signifies is not clear in the previous literature. With UR and RIP controlled, this variable should not represent local economic situation or a contact situation by immigrants, but rather represents an influence similar to the religious attendance: If the tertiary educated respondents are less opposed to immigration, having more highly educated people in the residential areas of respondents will have a conformity effect. Since the influence of tertiary education cannot be classified either to economic or cultural, the effect of rate of the tertiary educated is also introduced as a control variable. 4.3.3 Attitudinal variables and interactions with contextual variables Critical variables that distinguish economic and cultural factors of opposition to immigration in this study are in an attitudinal dimension. Due to the limitation and differences of available questions in each survey, this study interprets several comparable variables to represent respondents' reactions to economic and cultural factors, which has influence on opposition to immigration. First, the evaluation of national and personal economic situation is adopted to represent individual's general level of economic anxiety, following the Lahav's (2004: 191) findings of its strong linkage with the rejection of immigrant population. In Canada and Australia, the prospective and retrospective evaluation of their own financial situation and provincial and states' economy (both for Canada and Australia), the evaluation of current labor market situation and anxiety about family members' losing jobs in future (only for Australia) are considered to represent this concept. A factor analysis is conducted by principal factor method, and only one factor explains each variation with the factor loadings above .35 in both countries except for AES 1996. In AES 1996, a factor analysis yields two factors, retrospective and prospective evaluation of economy, which has no relationship each other. In an ancillary analysis (not shown here but 45 available upon request) separately introduced them in a regression model, but the overall result is the same. Cronbach's alpha is .58 (.49 in 2004) for Canadian and .18 (.82 in 2001) for Australian data. The answers to each question are simply summed and rescaled to vary from 0 to 1 so that its coefficient can be easily compared to the cultural one. 0 signifies a complete satisfaction with the economic and financial condition and labor market situation (those who perceive the household, national economy and labor market situation became better and willimprove, and do not worry about family members' unemployment), and 1 signifies the greatest concern. 2 7 Following the previous research based on perceived threat and realistic conflict theory, economic anxiety is expected to have a positive influence on opposition to immigration. 2 8 The measurement of cultural concerns is different in each country. As Murray and Watson (2005), Verkuyten and Brug (2004), and Citrin, Reingold, and Green (1990) find, nativism or the respondents' belief about the essential part of the native people's social or cultural characteristics, positively relates to opposition to immigration. In Canada, all the questions used to measure this concept ask how desirable the respondents think the social and cultural homogeneity or traditional values are for their society: whether the members of the society should better have similar social values and ideas, recent people don't respect traditional values, and the members of the Parliament (RCMP) should wear the same hat or not regardless A simple summation of answer points is used just for simplicity, but the factor scores (ballet) did not make any difference in overall results, and the correlation between the two are above .98. 2 8 Since the retrospective evaluation of provincial or state economic situation, as well as their own financial situation may involve the part of respondents' attitudes to the government's management of economy, the model needs to partial out the effect of the degree to which the respondents think government's economic policy affected the state of economy or their financial situation. This control is possible by introducing the variables that ask how worse off or better off the government made respondents' financial situation or provincial and national economy. For example, assume hypothetical respondents who think that the government's policy worsened the national economy. If they evaluate the state of the economy based on the bad economic management of the government, their retrospective evaluations of economy do not simply represent their economic anxiety, but they also include their evaluation of the government's economic performance. If the former part is correlated with opposition to immigration somehow, the estimate of the effect is over or underrepresented without a control. Since we do not know how respondents' evaluation of the government's economic management is related to opposition to immigration, the variable measuring respondents' perceived effect of the government economic policy on the national economy and financial situation is included just as a control variable without a prediction. The results show that the effect of economic anxiety becomes a little weaker without this control, but it does not make much difference in most cases. 46 of their religion (in CES 1993), newer lifestyles are detrimental to society, and more emphasis on the traditional values would produce fewer problems in this country (in CES 2004). In Australia, questions asking whether "Truly Australian" culture is characterized by native-born, living in Australia for a long time, English-speaking and being Christian are used to measure how important it is to share the same cultural characteristics. There is certainly a slight difference in the measured concept by these variables in Canada and Australia due to the available questions, but both variables operationalize respondents' degree of cultural concerns which are distinct from economic anxiety. In a similar manner as for the economic anxiety scale, they are summed and rescaled to vary from 0 to 1,0 suggesting no or minimum cultural concerns and 1 the maximum. I call these scales a common social/cultural value scale for Canadian data, and cultural essentialism scale for Australian data. I confirm that only one factor explains the variation for both countries, but the result of the factor analysis and reliability test for CES 1993 is weak, with the smallest factor loadings being .31 (similar social value desirable) and alpha of .34. The analysis is conducted with disaggregated variables, but the overall result is the same.2 9 For Australian data, the lowest factor loading is .43 and the alpha is .66 in 1996, and .48, and .70 in 2001. In both datasets, the scales that represent cultural concerns are hypothesized to have a positive influence on opposition to immigration. For New Zealand data, a different set of questions are used, mainly because there are no comparable cultural factor questions. They ask whether the following immigrants "be allowed to immigrate freely, should there be restrictions, or should they not be allowed to immigrate at al l ," offering four categories of "people from Muslim countries," "people from Asia," "people with technical skills," and "people willing to do manual labour."30 The former two questions prime respondents with cultural groups, whereas the latter two prime respondents with economic groups. In order to measure the strength of the opposition to each group of immigrants, 0 is 2 9 When three variables are introduced separately, preference to shared social value and opposition to RCMP' s wearing turban have a statistically significant influence at 99.9% level, and perception that few people respect tradition is statistically significant at 90%) level. 3 0 There is another type of "people seeking political asylum," but this variable is eliminated because it is not clear if they are perceived as economic or cultural groups. 47 assigned to "immigrate freely" and "don't know" answers, 1 for "restrict immigration," and 2 for "not allowed to immigrate," and the answers to the questions on Muslim and Asian immigrants are summed, as well as those to the questions on skilled manual labor migrants to vary from 0 to 4 (see Appendix A in detail). I assume that the relationship between opposition to economic types of immigrants and economic anxiety, as well as the relationship between opposition to cultural types of immigrants and respondents' concern about culture are parallel, and both are hypothesized to have a positive influence opposition to immigration in general. The variable of opposition to social/cultural types of immigrants has more coherence than the variable of opposition to economic types of immigrants (alpha for each is .82 and .58), so they are introduced separately, but the overall result is the same.31 Interaction effects between attitudinal and contextual variables are crucial in examining how the local economic and cultural situation relates to respondents' economic and cultural concerns, as well as attitudes to immigration. More specifically, the economic and cultural situations are expected to amplify or attenuate the impact of economic and cultural concerns. The following three interaction terms are hypothesized from the theory: First two are drawn from contact hypothesis, which suggests that contact with outgroups in competitive situation produces hostility. If we can assume that RIP represents the frequency of contact, and that economically and culturally concerned respondents are more sensitive to economic and cultural competition with immigrants, the interaction between RIP and economic anxiety scale (or R A P O and opposition to economic types of immigrants in New Zealand), and the interaction between RIP and cultural essentialism scale (or common social/cultural value scale or opposition to cultural types of immigrants) are hypothesized to have a positive influence on opposition to immigration. Second, realistic conflict theory predicts that actual economic interests and conditions cause realistic economic perception of threat, which causes hostility towards outgroups. If UR represents actual economic condition and economic anxiety signifies respondents' perceived economic threat, the interaction of them is hypothesized to have a positive influence on 3 1 A l l four variables are statistically significant at 99.9% level, but the impact of two variables of opposition to cultural types of immigrants is stronger than that of opposition to economic types. 48 opposition to immigration. A l l the operationlized hypotheses are summarized and classified in Table 4.1. Since the purpose of this thesis is to compare the relative size of the influence of economic and cultural factors, the stability (measured by the frequency of statistical significance and its direction) and the strength of impact of variables are assessed, individually and comparatively. Table 4.1. The predicted directions and attributes of the independent variables Variables Attribute Socio-demographic Attitudinal Economic concern Cultural concern Interactions U R x economic attitude RIP x economic attitude RIP x cultural attitude Economic Cultural Economic Economic Cultural Prediction Tertiary educated Control -Gender male Control ? Age Control ? Unemployment Economic + Factory or unskilled worker Economic +• Self-identified working class Economic + Household Income Economic -Native-born Cultural + ESC-born Cultural + Christian Cultural + Religious attendance Cultural -ntextual Unemployment rate (UR) Economic + Rate of immigrant population (RIP) Economic and cultural +/-Rate of the tertiary educated Control -U R x R I P Economic + 4.4 Model specification and problems in measurement 4.4.1 Regression model A l l the dependent variables used in the analysis are binary or take a limited number of responses, and they are ordinal in a sense that the higher number represents stronger opposition 49 to immigration. Since the limited probability model by ordinary least square regression would yield a heteroskedasiticity problem, maximum likelihood estimation (logistic regression form) is applied for the analysis.3 2 The simplified equation for the logistic regression is Pr (opposition to immigration, ory = l|x) = A (Jj0 + f}kXik). for Canadian data (binary response), and Pr 0 = 0) = A (So - IfiA) = 1 / [ 1 + exp (Y.pkXlk - S0)) Pr 0 = 1) = A (<3, - T./3kXik) - A (<50 - £ » ) • = 1 / [ 1 + exp (ZpkXik - <5,)] - 1 / [ 1 + exp - <50)] Pr (y = 2) = 1- A (<5, - X/3kXik) = 1- 1 / [1+ exp ( I / J A - <$,)] for Australian and New Zealand data, where k is the number of independent variables, X and P represent each independent variable and its coefficient in the logistic equation, and 8 0 and 5, are the estimated cutoff points. Because the coefficients of logistic regression do not provide enough information about the size of the impact of each independent variable and its statistical significance, I used CLARIFY to estimate the probability of taking each value (either;; = 1 [the number of immigrants should be reduced a little] ory = 2 [reduced a lot]) developed by Tomz, Wittenberg and King (2003).3 3 4.4.2 Problems in measurement Several problems should be mentioned regarding measurement validity. First, concerning the measurement reliability of the dependent variable, some respondents may not answer the question honestly, when they think the question is inducing a socially desirable answer, or when they intend to show their generosity to immigrant groups, whereas in fact they 3 2 Other forms (limited probability model with robust standard errors and probit model) are tested, but he result is virtually the same. 3 3 For this software and the idea of presenting statistical information, see also King, Tomz, and Wittenberg (2000). 50 actually oppose to the further acceptance of immigrants. The effect of this bias can be estimated to overrepresent the positive answer, and this problem is serious especially when the bias is systematically linked to respondents' choices of answers in the independent variables used to predict opposition to immigration. The problem is indeterminable, however, unless there are some good clues or sophisticated experimental method to distinguish those respondents who answer the question sincerely and those who do not. 1 assume that answering the questions asking what the number of immigrants that country accepts should be, are not as'strongly affected by the general sense of social desirability as the other sensitive questions such as desirable racial or ethnic components, feeling thermometers, or negative or positive image about immigrants. Second, some other respondents may still answer "don't know" to the question, because they do not want the interviewers know their real political attitudes (Berinsky 1999: 1211-1212). In this case, an examination of systematic bias of the answer "don't know" from the other answers to the question is possible, as Berinsky found in the opinions on racial integration. Compared to Berinsky's dataset, where one third of respondents refused to answer the question (1213), the "don't know" and "refused" are quite few in the dataset. In 1993 and 2004 CES, only 5% answer "don't know," and only 0.5% (0.4% in 2004) refused. In AES 1996 and 2001, there is no "don't know" category, and only 1.2% (1.8% in 2001) did not answer the question. NZES 2002 data does not code "don't know" and "refuse" respondents separately from missing answers, and interviewers did not ask some questions (including the immigration-related question) to a large number of respondents due to the time constraint of short telephone surveys (Vowles, Aimer, Miller, Banducci, Karp and Miller 2004: 195). Assuming that telephone interviewees did not ask the question about immigration, only 2.4% of all the respondents who answered the questions by mail answered "don't know." In the ancillary analyses, including those respondents to "oppose to immigration" category did not change the overall results in three countries.34 Thus, the bias of representing the positive response to immigrant inflow based on 3 4 Further, no variables used to predict opposition to immigration, except for the respondents' age in 2002 NZES, can predict the binary response of "don't know," "refuse" and/or "missing" choices against the other choices as a reference category at 95% statistically significance level in 51 the respondents' social desirability can be assumed to be low enough to neglect, or it should not critically damage the interpretation of the overall result at least. Second, it is possible that the socio-demographic categories that are hypothesized to represent either economic or cultural characteristics may contain the opposite factor. For example, imagine Maori respondents answered "people from Asia" shouldn't be allowed to immigrate at all for some reason. On the one hand, we can guess that they answered so because they felt Maori culture would be threatened by Asian immigrants, but on the other hand, we can equally imagine that their answer was economically driven, fearing that their labor resources were threatened by Asians immigrants. Thus it is possible to imagine that the native-born perceive economic threat by immigrants or they are more economically prejudiced against immigrants than the foreign-born. Several regression analyses are conducted to test whether only those categorized as an economic factor express more economic anxiety or opposition to economic types of immigrants, and those categorized as a cultural factor express more cultural essentialism more often, and the results are a little confusing. For instance, Christians in Canadian and New Zealand data, and self-identified working class in Australian data are more economically anxious as well as culturally essentialists. This implies that socio-demographic categories cannot be simply counted as strictly either economic or cultural factor, although most of the variables predict either type of concerns as expected. Third, operationalized variables in the attitudinal dimension are indirect proxies of economic and cultural threat or prejudice and their size of impact is not fairly comparable to each other in Canadian and Australian data. Intuitively, comparing the seize of impact of the variables that ask respondents' perception of economic and cultural threat from immigrants, or the impression of economic and cultural impact of immigrants on the society with exactly the same wordings is the best method. Unfortunately such variables are unavailable in every dataset, and comparing the influences of those variables may not be ideal, either.35 Sniderman, et al. (2004) a logistic regression analysis. 3 5 There is a legitimate measurement of economic threat, asking whether immigrants take jobs away from Australians in AES 2001, but the question whether they think immigrants make 52 caution that using these kinds of questions (a direct measurement of threats) can cause some unexpected bias in the estimation of the effect of threat because respondents would answer that they are threatened regardless of the type of threat involved. In their words, Just because people say that they perceive another group to threaten their cultural identity, it does not follow that they actually fear for their cultural identity. They may say they do to mask their true motives. So they may say that their concern is to protect the well-being of the national culture and way of life when their concern, in truth, is to preserve their own economic well-being (41, emphasis is in original). Further, using the words "immigrants" in independent variable could also cause bias by making respondents imagine either economic or cultural factors of immigration. As Gilens (1996; 1999: 60-79) argues that welfare and poverty is a "race-coded" issue, "immigration" can be also a "ethnic-coded" or specific race-coded issue: when respondents hear the words "immigration" or "immigrants," they may instantly associate with immigrants with specific ethnic or cultural backgrounds. This is supported by the close examinations of Australian polls by Goot (1984: 34), when he concludes that "attitudes to immigration and attitudes to Asian immigration are connected" and it is strengthened when the issue of the labor market competition slips in. The answer "immigrants should not be allowed because they take jobs away from native people" do not only expresses respondents' economic threat, but it may also measure cultural opposition to their imagined immigrants. Accordingly, introducing variables that do not contain "threat" or "immigrants" can avoid this problem, and the operationalized variables of economic anxiety and cultural essentialism (or common social/cultural value) scales can suffice for this purpose. In New Zealand dataset, where such variables are not available, the variables that directly ask what types of immigrants are acceptable are introduced as a second-best solution. Australia more open, a single available question in the dataset to assess cultural threat of immigrants, does not equivalently correspond as a measurement of cultural threat. Because AES 2001 does not have a question such as "immigrants deprive Australian cultural value" or "immigrants erodes Australian cultural unity," it is impossible to construct a strictly contrasting, valid cultural threat scale of immigrants. In Canadian data, there is a good measurement of cultural threat, asking whether respondents agree or disagree with the idea that "too many recent immigrants just don't want to fit into Canadian society," but there are no comparable economic threat questions 53 In any case, comparability of two variables is a problem. In Canadian and Australian datasets, it is not clear whether the economic anxiety scale measures respondents' economic concerns as deeply as the cultural essentialism (or common social/cultural value) scale measures their cultural concerns. The number of items used to measure economic anxiety scale, however, is always larger than that to measure cultural essentialism (or common social/cultural scale) scale. This does not still solve the problem of comparability, but the problem always exists as long as the different concepts (economy and culture) is compared. In the New Zealand dataset, while the statistical correlation between attitudes to Muslim and Asian immigrants is high (Pearson's correlation coefficient, or r = .70), that for skilled and manual labor is much lower (r = .41). This means that those respondents who oppose Muslim immigrants are highly likely to oppose to allow Asians to immigrate, but those who oppose skilled immigrants do not necessarily oppose manual labor immigrants (or less likely to oppose than in the former case). The different attitudes may reflect the respondents' economic situation and characteristics in a way that economically vulnerable respondents (unemployed, factory workers, manual labor, working class, and lower household income earners) oppose manual labor immigrants because they perceive that manual labor immigrants have more competing economic interests than skilled immigrants who compete in a different labor market. I performed some regression analyses to test this, but no clear evidence was found to support this argument. I also distinguished respondents in three groups: respondents who answer any types of immigrants shouldn't be allowed to immigrate (indistinguishable opposition to economic and cultural types of immigrants); those who answer that cultural groups of immigrants should not be allowed, but economic groups of immigrants can immigrate; and those who answered that economic groups of immigrants can immigrate, but cultural groups of immigrants should not be allowed to immigrate. Several possible patterns of recoding variables are tried, but they produced the same or even stronger results as the main findings. In Chapter 5, results with four disaggregated variables and two aggregated scales in an attitudinal dimension are shown. 54 5 RESULTS AND INTERPRETATIONS 5 .1 Canada The result of logistic regression of socio-demographic and contextual variables in 1993 CES and 2004 CES are shown in Table 5.1. First on socio-demographic dimension, variables of tertiary education, male and Quebec resident are less likely to oppose, while semi- and unskilled labor,3 6 Canada-born and Christian are more likely to oppose immigration in both years. Unemployment and household income are not always statistically significant, but their influence is in an expected direction respectively. Age, unemployment of main income earner (only available in 1993) and English-speaking country-born do not influence attitudes to immigration. The statistically insignificant socio-demographic variables are dropped in Model 2 and 4, and the probability change of opposing immigration (the probability of answering "Canada should admit fewer immigrants") is shown in Table 5.2 to compare the size of influence of each variable. A typical respondent's characteristics are assumed as a "default" by taking an absolute value for dichotomous variables, because "48% male" or "25% living in Quebec" does not make any realistic sense. The assumed typical respondents are Canada-born Christian female who have not received any tertiary education, and lives outside Quebec. In CES 1993, they are also assumed not unemployed and not semi- and unskilled workers. Table 5.2 shows that the overall level of opposition to immigration is much higher in 1993 than in 2004. For typical respondents with similar socio-demographic characteristics, the probability of their opposing immigration (y = 1) is approximately 75% in 1993 and 41% in 2004. The table also shows that tertiary education has a strongest negative influence on opposition to immigration in both years, the effect of which reduces the probability by about 20% points. This could be both because tertiary educated respondents are more economically competitive in a labor market even after controlling other economic characteristics, and because they are more generous to accepting other social outgroups and their ideas, as discussed. Being unemployed 3 6 No variable is available that can represent lower-class job in 2004 CES. 55 and unskilled labor are more likely to oppose immigration by 5% points, and $40,000 more income in addition to the average income decreases the probability by 4% points in 1993, while these influences are not confirmed at all in 2004. Canada-born and Christian stably increase the probability of opposition in both years, approximately by 11% points and 8% points respectively. It is too early to conclude anything from this result, but the overall impression of the results is that the variables of social and cultural categories have more stable, a little stronger influence on opposition to immigration than variables of economic categories. Table 5.1 Logistic regression on opposition to immigration, Canada (Models with socio-demographic and contextual variables) 1993 2004 Model 1 Model 2 Model 3 Model 4 b P b P b P b p Socio-demographic Tertiary educated -.93 .000 -.94 .000 -.78 .000 -.79 .000 Male -.19 .014 -.19 .016 -.21 .009 -.31 .000 Age category .02 .392 .02 .325 Quebec resident -.75 .000 -.75 .000 -.26 .010 -.26 .005 Unemployed .39 .016 .36 .026 .05 .803 Main income earner unemployed in past 1 year .08 .486 - -Semi- and unskilled labor .25 .007 .24 .009 - -Household income category -.04 .018 -.04 .009 -.01 .318 Native-born .60 .000 .51 .000 .64 .000 .54 .000 ESC-born .21 .346 .02 .926 Christian .49 .000 .50 .000 .28 .005 .28 .002 Contextual Unemployment rate (UR) .02 .041 .02 .038 -.0.1 .359 -.01 .148 Rate of immigrant population (RIP) .03 .002 .03 .003 .03 .002 .02 .022. Rate of the university degree holders -.02 .024 -.01 .035 -.02 .000 -.02 .000 URxRIP -.002 .005 -.002 .007 -.001 .077 -.001 .427 Constant -.12 .26 -.09 .22 -1.14 .26 -.90 .18 N 3067 3069 3471 3990 Log likelihood -1864. 6 -1867.0 -1971.9 -2285.0 Pseudo R 2 .09 .09 .05 .04 56 Table 5.2 The estimated influence of single-variable manipulations on opposition to immigration 0 = 1 ) , Canada: Model 2 (1993) and Model 4 (2004) Independent variables Default" Change 1993 {y= \f 2004 (y= \f Default • \41.0%-___.J j Tertiary educated 0 . ' Oto 1 54.1% 23.9% Male 0 Oto 1 71.4% 33.6% Quebec resident 0 Otol 58.7% ' 34.9% Unemployed 0 Oto 1 81.2% -Semi- and unskilled labor 0 l 0to 1 79.2% -Income category mean ($48,300) mean to 9 ($90,000) 70.4% -Native-born 1 1 toO 64.3% 28.9% Christian 1 1 toO 64.8% 34.5% a. The default setting for this probability prediction is a respondent who has not received any tertiary education, female living outside Quebec, employed, not semi or unskilled labor, with her household income of about $48,300, Canada-born Christian who lives in the FSA area with the unemployment rate of 10.6%, the rate of immigrant population of 13.4% and the rate of university degree holders of 8.8% (mean position) b. The default setting for this probability prediction is a respondent who has not received any tertiary education, female living outside Quebec, Canada-born Christian who lives in the FSA area with the unemployment rate of 7.8%, the rate of immigrant population of 14.9% and the rate of university degree holders of 18.8% (mean position). Table 5.1 includes interesting results about the influence of contextual variables,-too. First, it is noticeable that rate of university degree holders has a negative influence on opposition to immigration in both years, after the economic situation is controlled by unemployment rate (UR). With UR and rate of immigrant population (RIP) controlled, rate of university degree holders should be interpreted as a conformity effect than economic or cultural threat, as discussed in the previous chapters. The negative influence means that i f typical respondents are surrounded by more highly educated people, who are much less likely to oppose immigration, their opinion about immigration tend to conform to a more relaxed view. Alternative explanation comes from the endogenous relationship of these variables: the respondents with more relaxed views on accepting immigrants are more likely to live in areas with more highly educated people, who share similar political opinions. The same idea can also be applied to the interpretation of the effect of UR and RIP, i f less anti-immigrant respondents tend to live in an area with more 57 immigrants with the lower unemployment rate.37 Second, the interaction effect of UR and RIP is statistically significant at the 95% level in 1993 but insignificant in 2004. Further, the interaction effect is negative on opposition to immigration, which contradicts perceived economic threat theory, because it means that the effect of UR is attenuated by the level of RIP. Figure 5.1 and 5.2 depict differences in probability changes of answering "admit fewer immigrants" when RIP of the typical respondents' FSA is at its minimum (solid line) and maximum (dashed line) value. A solid line in Figure 5.1 means that respondents are more likely to oppose immigration as UR grows when there are no immigrants around them, while a dashed line suggests that they are less and less likely to oppose immigration as UR grows, when they are surrounded by many immigrants. Although the effect is insignificant, the same trend can be confirmed from Figure 5.2: respondents become less opposed to immigration as UR increases when immigrants are concentrated in the respondents' residential area. This does not make any economic sense: If a perceived economic threat comes from the local labor competition with immigrants, the effect of higher unemployment rate must be magnified by the geographic concentration of immigrants, but the negative interaction effect provides opposite evidence. One of the possible interpretations of this mysterious relationship is that local economic situation matters only where immigrants are not concentrated. Under the situation where the local UR is higher, people may think that accepting immigrants worsens their local economy or "take jobs away from Canadians" only when there can see no or fewer immigrants in their areas. Surrounded by many immigrants, however, UR becomes a less important factor. Why higher UR decreases the chance of opposition when immigrants are concentrated is not clear, but the wide 95% confidential intervals for dashed lines suggest that its influence is insignificant. In addition, the effect of economic situation is unstable in both datasets, 3 7 It should be noted that the areas with higher rate of university degree holders are likely to have less unemployment rate with more immigrant populations. Bivariate correlation coefficient (Person's R) between rate of university degree holders and UR is roughly -0.3, and that for rate of university degree holders and RIP ranges from 0.4 to 0.5. I assume that introducing these variables together does not cause multicollinearity, when none of the variation inflation factors of these variables exceed the threshold of 10 in ordinary least square regression analysis. 58 Some other models (not shown here) without an interaction term and rate of university degree holders are tested, but the effect of UR is either quite weak, statistically insignificant or even negative on opposition to immigration. The overall result does not change when UR is replaced by another economic contextual variable, the rate of families with the annual household income below $20,000, which represents the degree of economic disadvantage/8 Figure 5.1 The estimated influence of unemployment rate changes on opposition to immigration (interaction with RIP), 1993 CES (y=\f R I P = 0 % h RIP = 62% c ,-ro CD I" •o to g* _ co CO M— O CM s ^ o o — T T T H 95% CI of RIP = 0% H 95% CI of RIP = 62% i 1 1 10 20 30 Unemployment rate (0 - 47) 40 50 The other variables are held constant in the same manner as in footnote a of Table 5.2. These results are available by contact. 59 Figure 5.2 The estimated influence of unemployment rate changes on opposition to immigration (interaction with RIP), 2004 CES (y=\f RIP = 0% RIP = 72% H 95% CI of RIP = 0% i 95% CI of RIP = 72% 0 10 20 30 Unemployment rate (0 - 46%) 40 50 a. The other variables are held constant in the same manner as in the footnote b of Table 5.2. Attitudinal variables and their interactions with the contextual variables are included in Table 5.3. By introducing attitudinal variables, some socio-demographic variables become statistically insignificant at the 95% level, and the size of the coefficients of most of the variables declines from those in Model 2 and 4, which reveals that the part of the economic and cultural factors represented by socio-demographic categories are explained or mediated by attitudinal variables.3 9 The most important result in this table is the relative size of coefficient of economic anxiety and common social/cultural value. Although a strict comparison is impossible, Model 6 suggests that common social/cultural value scale has a stronger influence on opposition to immigration than economic anxiety. More concretely in 1993 CES data, the probability of 3 9 These variables are included in a model anyway, because other significant variables would become statistically insignificant once they are dropped. A separate analysis shows that the individually insignificant variables are jointly statistically significant. 60 answering "admit fewer immigrants" is about 40% when typical respondents do not agree at all that having common social and cultural values in a society is desirable (that is, when common social/cultural value scale takes a minimum value of 0), while the probability increases to 82% when they agree with the idea (when common social/cultural scale takes 1.00). On the other hand, the probability for the same typical respondents who do not worry about the economic or financial situation at all (when economic anxiety scale takes a minimum value of 0) is 55%, and it increases to 78% when they are fully economically anxious (when economic anxiety scale takes 1.00). In Model 7 of 2004 CES, this difference in size of effects is not as obvious as the one in 1993 dataset, however, the probability increase when common social/cultural value scale moves from its minimum to maximum is 24% points, compared to 20% point increase for the same move in values of economic anxiety scale. 4 0 1 The weaker impact of common social/cultural value in 2004 CES is partly because only two items are used to construct this scale. 61 Table 5.3 Logistic regression on opposition to immigration, Canada (Models with attitudinal and interaction variables) 1993 2004 Model 5 Model 6 Model 7 Model 8 b P b P b P b P Socio-demographic Tertiary educated -.62 .000 -.63 .000 -.79 .000 - -.79 000 Male -.46 .667 -.06 .583 -.31 .012 -.30 014 Quebec resident -.93 .000 -.91 .000 -.43 .011 -.43 .010 Unemployed .30 .159 .27 .211 Semi- and unskilled labor .43 .000 .42 .001 Household income category -.02 .312 -.02 .306 Native-bom .43 .006 •45 .005 .74 .000 .78 .000 Christian .20 .136 .19 .159 .19 .211 .18 .257 Contextual Unemployment rate (UR) .04 .003 .03 .510 -.01 .457 -.02 .624 Rate of immigrant population (RIP) .03 .018 .001 .970 .03 h000 .01 .440 Rate of the university degree holders -.01 .202 -.01 .228 -.02 .006 -.02 .006 URxRIP -.003 .023 -.003 .022 Attitudinal Economic anxiety 1.04 .017 .87 .453 .87 .022 .85 .290 Government's influence on economy -.20 .556 -.21 .551 1.21 .005 1.21 .005 Common social/cultural value 1.91 .000 1.22 .004 1:07 .000 .64 .061 Interactions UR x economic anxiety .010 :909 .021 .787 RIP x economic anxiety .005 .874 -.009 .669 RIP x common social/cultural value .050 .024 .029 .076 Constant -1.76 .41 -1.26 .77 -3.02 .41 -2.74 .56 N 1826 1826 1603 1603 Log likelihood -1098. 4 -1096.2 -834.4 -832. 6 Pseudo R 2 .11 .11 .10 .10 Model 6 and 8 include interaction terms between: UR and economic anxiety; RIP and economic anxiety and; RIP and cultural essentialism, in order to examine the complicated relationship between the effects of contextual and attitudinal variables. Both data suggest roughly the same result: economic anxiety does not have any strong interaction effect with UR or RIP, while common social/cultural value has a strong interaction effect with RIP. The statistically 62 insignificance of the first two could be attributed to the multicollinearity, because the items used for interactions are highly correlated one another.41 The results of several ancillary analyses (not shown here but available upon request) reveal that there is a weak but statistically insignificant interaction between UR and economic anxiety when the interaction between RIP and economic anxiety is dropped, but no substantial, statistically significant influence of the interaction between RIP and economic anxiety is found after dropping all or some of the other interaction terms in both datasets. Even i f their standard errors are marred by the multicollinearity, the size of interaction is tiny. The trajectories of two lines that represent the influence of economic anxiety under the different contextual settings (lowest and highest level of UR and RIP, not shown here but available upon request) are not substantially different, or sometimes almost identical. This means that the effect of economic anxiety is constant regardless of the contextual situation: typical respondents oppose accepting immigrants as they are more economically anxious, whether the actual local economy is good or bad. Theoretically, this further implies that the effect of perceived economic threat does not depend on a realistic (or actual) economic situation. In contrast, Model 6 and 8 contains a statistically significant, positive and substantial interaction effect between RIP and social/cultural value scale. In 1993 for example, Figure 5.3 shows that the probability change in answering "admit fewer immigrants" when the social/cultural value scale moves from minimum to maximum value (0 to 1), is +27% points (from 49% to 76%), when there are virtually no immigrants in respondents' residential area. This probability change is magnified to +66% points (from 25% to 91%) when the same respondents live in an area where almost half of the population are immigrants. Similarly, Figure 5.4 suggests that the probability change is +13% points (from 23% to 36%) in the area with virtually no immigrants, while the probability change in the area where 72% of the population are immigrants increases to +53% points (from 35% to 88%). 4 1 Their bivariate correlation ranges from 0.7 to 0.9.1 tested variation inflation factor (VIF) in an ordinary least square regression model, and the all the VIF scores for all the interaction terms as well as UR and RIP exceed the threshold of ten (in the worse case, 32). 63 Figure 5 The estimated influence of common social/cultural value scale on opposition to immigration (interaction with RIP), 1993 CES (y = 1 ) a RIP = 0% i 1 95% CI of RIP = 0% RIP = 46% h—• —! 95% CI of RIP = 46% Q _ i i 1 1 r 0 .2 .4 .6 .8 1 Common social/cultural value (0-1) a. The assumed characteristics is not having received tertiary education, female living outside Quebec, not unemployed, not semi or unskilled labor, with her household income of about $50,300, Canada-born Christian who live in the FSA area with RIP of 13.4% and the rate of university degree holders of 8.8% (mean). 64 Figure 5.4 The estimated influence of common social/cultural value scale on opposition to immigration (interaction with RIP), 2004 CES (y=\f — RIP = 0% i 1 95% Cl of RIP = 0% RIP= 72% 95% CI of RIP = 72% 1 H o_ ' i i i 1 r 0 .2 .4 .6 .8 1 Common social/cultural value scale (0-1) a. The assumed characteristics is not having received tertiary education, female living outside Quebec, Canada-born who live in the FSA area with UR of 7.6% and the rate of university degree holders of 18.8% (mean). This positive interaction effect means that the effect of the attitudes to common social/cultural value on opposition to immigration is magnified by the degree of concentration of immigrants in respondents' residential area. In other words to explain the same idea is that the effect of the level of concentration of immigrants is increased by how strongly respondents believe that common social/cultural values are desirable for the society. The latter interpretation is theoretically interesting, because it suggests that higher RIP means a perceived cultural threat for those who worry about the nation's social and cultural unity, while the same high RIP does not pose any threat for those who do not share such concerns. If we can assume that higher RIP leads to more frequent contact with them, the results confirms a part of contact theory: contact with outgroups in a culturally competitive situation (gauged by respondents attitudes towards 65 common social/cultural values) leads to a group hostility. In summary, analysis of two datasets shows that both economic and cultural factors represented by some socio-demographic and attitudinal variables have a substantial influence on opposition to immigration. The impacts of variables that represent cultural concerns, however, look a little larger than those of economic concerns in both dimensions. The negative interaction effect between UR and RIP implies that UR increases the opposition to immigration only where immigrants are not concentrated. RIP amplifies the effect of cultural concern, while neither UR nor RIP enhances the effect of economic concerns. 5.2 Australia Most of the results of Australian data confirm the findings in Canadian case. Model 1 and 3 in Table 5.4 show the estimated coefficients and each cut point in the ordered logistic equation when all the hypothesized socio-demographic and contextual variables are introduced. Tertiary educated, working class, household income, Australia-born, and religious attendance have a stable influence on opposition to immigration with an expected sign, while male and the unemployment do not influence opposition in both AES 1996 and 2001. Born in English-speaking countries and Christian are not always significant, but the sign of the both variables are in an expected direction. In Model 2 and 4, those statistically insignificant socio-demographic factors are dropped, and the probability of taking each value of the dependent variable (y = 0, 1 or 2) is shown in Table 5.5, when one socio-demographic factor changes its value in Model 2 and 4. As in the analysis of Canadian data, Table A U S assumes typical Australian respondents as their default characteristics to compare the influence of independent variables (for the detail of the characteristics, see the footnote a and b of Table 5.5). 66 Table 5.4 Ordered logistic regression on opposition to immigration, Australia (Models with socio-demographic and contextual variables) 1996 200 Model 1 Model 2 Mode 3 Model 4 b P b P b P b P Socio-demographic Tertiary educated -.43 .003 -.44 .001 -.28 .037 -.34 .010 Male -.21 .088 -.09 .453 Age .002 .629 -.01 .001 -.02 .000 Unemployed -.08 .790 .11 .772 Working class .51 .000 .45 .001 .45 .001 .45 .001 Household income category -.08 .001 -.07 .012 -.07 .001 -.08 .000 Native-born .53 .010 .45 .002 .96 .000 1.01 .000 ESC-born .24 .362 .58 .050 .62 .034 Christian .24 .089 .30 .030 .23 .087 Religious attendance -.09 .006 -.10 .002 -.12 .002 -.11 .003 Contextual Unemployment rate (UR) .04 .423 .05 .331 .05 .356 .03 .604 Rate of immigrant population (RIP) .04 .108 .04 .080 .06 .010 .05 .025 Rate of the tertiary educated -.02 .056 -.02 .036 -.04 .000 -.04 .000 URxRIP -.004 .023 -.005 .014 -.005 .059 -.003 .148 Cutting point I -.634 .650 -.630 .573 .497 .636 .071 .201 Cutting point 2 .642 .650 .641 .573 1.424 .638 .999 1.107 N 993 1038 1263 1281 Log likelihood -1027 0 -1076 2 -1031 .2 -1049 4 Pseudo R 2 .05 .05 .06 .06 67 Table 5.5 The estimated influence of single-variable manipulations on opposition to immigration (y = 0, 1 and 2), Australia: Model 2 (AES 1996) and 4 (AES 2001), 1996a 200 l b Independent variables Default Change Increase a lot/little or same (v = 0) Reduce a little Reduce a lot (y = 2) Increase a lot/little or same (v = 0) Reduce a little 0'=1) Reduce a lot (V=2) : Default IlltS; IIBIIIII 34.0% -.- 29 ; ' 6 % • -35.6% -~18J% Tertiary educated 0 Oto 1 44.2% 29.5%d 26.2% 72.0% 14.4% 13.6% Age Mean (45.5) Mean to 90% point (67) - - - 71.6% 14.5% 13.9% Working class 0 Oto 1 24.6% 29.0% 46.5% 54.0% 20.3% .25.8% Household income category Mean Mean to 10 ($95,000) 41.3% 29.9%d 28.7% 71.7% 14.5% 13.8% Native-born 1 0 44.4% 29.4%d 26.2% 83.0% 9.2% 7.6% ESC-bornc 0 1 - - - 72.9% 14.0% 13.2% Christian 1 0 41.0% 30.0%d 29.0% - - -Religious attendance 2(once a year) 2 to 5 (at least once a week) 43.0% 29.6%d 27.4% 73.6% 13.7% 12.7% a. The default personal characteristics in AES 1996 assumes not having received tertiary education, not working class with his/her household income of about $38,400, native-born Australian, Christian, who attend religious services at least once a year and live in an electoral district with the unemployment rate of 9.3%, the rate of immigrants of 21.5%, and the rate of the tertiary educated of 16.6% (all at their mean position). N b. The default personal characteristics in AES 2001 assumes not having received tertiary education, aged 45.5, not working class with his/her household income of about $51,400, native-born Australian, who attend religious services at least once a year and live in an electoral district with the unemployment rate of 7.4%, the rate of immigrants of 21.9%. and the rate of the tertiary educated of 19.7% (all at their mean position). c. As the status of ESC-born changes from 0 to 1, Native-born is set to 0. d. Change in values of the independent variable is statistically insignificant at the 95% level. All the other changes shown here are statistically significant. With almost the same characteristics, the probability of typical respondents' answering "increase immigrants a little/lot" and "remain about the same as it is" altogether (y = 0) is 34% in 1996 and 65% in 2001, which means that the level of opposition to immigration among the public subsided greatly in 5 years. Although tertiary education substantially reduces a probability of opposing immigration in both years, its influence is weakest among the three countries: the tertiary educated are less likely to oppose immigration by 8 to 10 percent points. The size of the influence of comparable independent variables looks quite similar in both years, and the variables that represent economic and cultural factors influences opposition to immigration in an equally substantial manner, except for Australian-born in 2001. For economic characteristics, working class increases the probability of answering "reduce a lot" by about 10%> points and additional $ 61,000 ($43,000 in 2001) to the annual income reduces the probability of answering "reduce a lot" by about 9% points. For social and cultural characteristics, Australian-born increases the probability by 9% to 18% points, Christian increases it by 6% in 1996, English-speaking country-born increases it by 6%> in 2001, and religious attendance decreases the probability by about 7% in both years. Thus, there is not much difference in the size of influence between economic and cultural characteristics in a socio-demographic dimension in Australia. The contextual effects in Table 5.4, Figure 5.5 and 5.6 confirm the findings in Canadian data. First, a substantial conformity effect of the concentration of the highly educated people can be seen in both years as in Canadian data. The interaction effect of UR and RIP is negative and statistically significant at the 95% level in 1996, which also endorses the findings in Canadian data. Figure 5.5 and 5.6 show that the influence of UR on the probability of answering "reduce a lot" (y = 2) declines as the RIP increases.42 The slope of the solid line (when RIP is fixed at its minimum level) is slightly positively tilted, while that of the dashed lines (when RIP is fixed at its maximum level) is steeply negative. Again, this result suggests that local unemployment increases opposition to immigration only when respondents do not see many immigrants. Further 4 2 The probability of answering "reduce a little" is omitted because of the available space. The result is available by contact. 69 as in the Canadian case, the effect of UR is not stable in both datasets. The models without an interaction term and rate of the tertiary educated are tested, but the effect of UR is weak, statistically insignificant or negative in both years. I also tested rate of families with the weekly household income below $500 instead of UR, but the result is virtually the same.43 Figure 5.5 The estimated influence of unemployment rate changes on opposition to immigration (interaction with RIP), AES 1996 (y = 2) a RIP =4.6% H CO * o co o ."O CO <u D) c iq 1— <D •"vT c (0 CO — CNj £ o T -CL o H RIP= 51.3% ^ 95% Cl.of RIP = 4.6% 95% CI of RIP = 51.3% T • j N \ N J - _ - \ I 7^  -i -~ J . - -- p - I I— 5 10 15 Unemployment rate (3-18) 20 The other variables are held constant in the same manner as in the footnote a of Table 5.5. These results are not shown here but available by contact. 70 Figure 5.6 The estimated influence of unemployment rate changes on opposition to immigration . (interaction with RIP), AES 2001 {y = 2f r~-o as CO 0 o TJ i n CD _1— CD C i _ •sr (U g a> an CO o CN CO o 0_ RIP = 4.4% RIP = 52.8% H 95% CI of RIP = 4.4% H 95% CI of RIP = 52.8% 1 i 10 15 Unemployment rate (3-15) 20 a. The other variables are held constant in the same manner as in the footnote b of Table 5.5. Attitudinal variables and interactions are included in Table 5.6, and the main conclusion stays the same: the effect of cultural essentialism is stronger than that of economic anxiety in both years. The expected probability of the typical respondents44 answering "reduce a lot" (y = 2) increases from 19% to 46% as their economic anxiety increases from its minimum to maximum position, while the probability increases from 6% to 63% for the same change in the cultural essentialism in Model 5 in AES 1996. The size of influence diminishes in Model 7 (AES 2001), however, its increase is 28%) points (4% to 32%) for economic anxiety compared to 39% points (2% to 41%) for cultural essentialism. Thus the impact of cultural factors is found stronger than economic factors on attitudinal variables. For the detail of the assumed characteristics, see the footnote a and b of Figure 5.7 and 5.8. 71 Table 5.6 Ordered logistic regression on opposition to immigration, Australia (Models with attitudinal variables and interactions) 1996 2001 Model 5 Model 6 Model 7 Model 8 b P b P b P b P Socio-demographic Tertiary educated -.34 .024 -.35 .022 -.16 .286 -.16 271 Age -.03 .000 -.03 000 Working class .29 .043 .32 .027 .28 .057 .26 075 Household income category -.02 .484 -.02 .449 -.02 .455 -.02 387 Native-born .27 .089 .26 .115 .77 .003 .76 .003 ESC-born .66 .044 .65 .047 Christian .15 .307 .15 .295 Religious attendance -.15 .000 -.15 .000 -.16 .000 -.16 .000 Contextual Unemployment rate (UR) .05 .298 .12 .218 .01 .873' .04 .650 Rate of immigrant population (RIP) .05 .029 .05 .134 .04 .085 .04 131 Rate of the tertiary educated -.02 .085 -.02 .100 -.03 .006 -.03 008 URxRJP -.006 .007 -.006 .006 -.002 .307 Attitudinal Economic anxiety 1.36 .015 4.72 .011 ' 2.42 .000 3.37 028 Government's influence on economy .01 .982 .13 .807 -.80 .067 -.81 063 Cultural essentialism 3.30 .000 1.22 .105 3.56 .000 3.56 000 Interactions UR x economic anxiety -.110 .488 -.001 972 RIP x economic anxiety -.117 .007 -.040 203 RIP x cultural anxiety .104 .002 -.0001 978 Cutting point I 2.22 .73 2.79 1.16 2.70 .76 3.11 1.17 Cutting point 2 3.60 .74 4.19 1.16 3.722 .76 4.13 1.17 N 932 932 1149 1149 Log likelihood -909.4 -901. I -875.0 -874.Z Pseudo R 2 .10 .11 .12 .12 72 Figure 5 .7 The estimated influence of economic anxiety scale on opposition to immigration I (interaction with UR), AES 1996 {y =2)a co . co 0) o Q) c co CO o — CM . Q CO O -r-UR = 3.5% UR = 17.9% H 95% Cl of UR = 3.5% -i 95% Cl of UR = 17.9% .4 .6 Economic anxiety (0-1) The assumed characteristics is not having received tertiary education, not working class with the household income of about $39,300, native-born Australian, Christian, who attend religious services at least once a year and live in an electoral district with RIP of 21.6% (mean), and the rate of the tertiary educated of 16.6% (mean). Cultural essentialism scale and its interaction with RIP are set at is mean. 73 ;ure 5.8 The estimated influence of economic anxiety scale on opposition to immigration I (interaction with UR), AES 2001 (y =2f UR = 3.2% i 1 95% C l of UR = 3.2% UR = 15.1% i 1 95% C l of RIP = 15.1% CD -CO LO _ . . . (D o T3 OH i i i 1 1 r~ 0 .2 .4 .6 .8 1 Econom i c anxi ety (0-1) The assumed characteristics is not having received tertiary education, not working class, with his/her household income of about $52,900, native-born Australian, who attend religious services at least once a year and live in an electoral district with RIP of 22.0% (mean). Cultural essentialism scale and its interaction with RIP are set at is mean. 74 Figure 5.9 The estimated influence of economic anxiety scale on opposition to immigration II (interaction with RIP), AES 1996 (y =2) a RIP = 4.6% i 1 95% CI of RIP = 4.6% RIP = 51.3% 95% CI of RIP = 51.3% *".1 o 4 i 1 1 1 1 r 0 .2 .4 .6 .8 1 Economic anxiety (0-1) a. Other variables are held constant in the same manner as in the footnote a of Figure 5.7. 75 Figure 5.10 The estimated influence of economic anxiety scale on opposition to immigration II (interaction with RIP), AES 2001 (y =2)a a. Other variables are held constant in the same manner as in the footnote a of Figure 5.8. 7 6 Figure 5. 11 The estimated influence of cultural essentialism scale on opposition to immigration (interaction with RIP), AES 1996 (y =2f .a. Other variables are held constant in the same manner as in the footnote a of Figure 5.7. 77 Figure 5.12 The estimated influence of cultural essentialism scale on opposition to immigration (interaction with RIP), AES 2001 (y =2)a oo J ro T3 CD in J 03 — CN ro o T -RIP = 4.4% RIP= 52.8% H 95% CI of RIP = 4.4% : 95% CI of RIP = 52.8% 1 1 .4 .6 Cultural essentialism (0-1) a. Other variables are held constant in the same manner as in the footnote a of Figure 5.8. Model 6 and 8 include the same three interaction terms between contextual and attitudinal variables as in Canada. 4 5 First, compared to the strong interactions in AES 1996, all the interaction effects are quite weak and often statistically insignificant in AES 2001. The statistically insignificance is certainly caused by the multicollinearity, when the contextual variables and interaction variables are often highly correlated, 4 6 but each interaction effect is still statistically insignificant even when tested separately by dropping other highly correlated interaction variables in 2001. Further, their impacts are also quite weak: except for Figure 5.12, 4 5 The interaction between UR and RIP is dropped in AES 2001, because the effect is statistically insignificant at the 95% level in Model 7, and introducing them exacerbates the problem of multicollinearity. Their bivariate correlation ranges from 0.6 to 0.8. VIF scores in an ordinary least square regression model are examined, and all the scores for all the variables involving the interaction terms exceed the threshold of ten except for cultural essentialism scale in both years. 46 78 the difference in the slope of the curve between the solid and dashed lines is not large in 2001 (Figure 5.8 and 5.10) compared to the obvious difference found in AES 1996. Second, contrary to the expectations in realistic conflict theory and contact theory in a competitive situation, all the interaction effects are negative except for the interaction between RIP and cultural anxiety in AES 1996. For example, Figure 5.7 shows that while respondents who live in an electoral district with the lowest UR of 3.5% are more likely to oppose immigration as their economic anxiety grows, but the probability of opposition stays at the lower level regardless of their economic anxiety, i f they live in an electoral district with the highest UR of 17.9%. Further, Figure 5.9 shows that economic anxiety increases the probability of opposing immigration when there are few immigrants around respondents, but it decreases the chance when the immigrants are concentrated in an electoral district of respondents. These results push forward the theoretical implication obtained in Canadian cases: the influence of economic anxiety emerges only when economic situation is good, or when immigrants are not concentrated. When the contextual parameters are higher, what matters is the effect of cultural essentialism, as the steep dashed lines in Figure 5.11 and 5.12 suggest. When RIP is at highest level of 15.1%, the predicted probability of the respondents who are fully economically anxious (economic anxiety = 1.00) is approximately 14% in 1996 and 35% in 2001, while that with the strongest cultural essentialist attitudes (cultural essentialism = 1.00) is 83% in 1996 and 57% in 2001. Although the interaction effect between RIP and cultural essentialism is statistically insignificant at the 95% level in 2001, a different trajectory of solid and dashed lines implies that the influence of cultural essentialism is magnified when RIP increases. Thus Figure 5.11 and 5.12 confirm the results obtained in Canadian data, which implies that the same RIP could mean a cultural threat to some respondents while not to the others. Moreover, i f the RIP is assumed to represent frequency of contact with immigrants, the figures imply that more frequent contact with immigrants with culturally competitive situation produces hostility to them. Thus Australian data repeated some of the basic results in Canada: the stronger influence of cultural concerns than those of economic concerns; the local economic situation 79 c contributes to opposition to immigration in an expected direction only when there are no or fewer immigrants; and geographic concentration of immigrants magnifies the effect of cultural concerns. Moreover, i f a negative interaction effect between RIP and economic concerns is a true relationship, it suggests that economic anxiety increases opposition to immigration when immigrants are not concentrated, which is in line with the explanation of the negative interaction between UR and RIP. 5.3 New Zealand Results of the New Zealand dataset sometimes contradict the findings in Canadian and Australian data. First for the socio-demographic variables, Table 5.7 shows that male, unemployed, manual workers and Christian are not statistically significant at the 95% level, but other variables assumed to present either economic or cultural factors significantly influence opposition to immigration in an expected direction. The statistically insignificant variables are dropped in Model 2, and the probability change of the dependent variable is shown in Table 5.8. In the same manner as in Canadian and Australian data, the table assumes typical respondents' characteristics.47 See a footnote a in Table 5.8 for the detail. 80 Table 5.7 Ordered logistic regression on opposition to immigration, New Zealand (Models with socio-demographic and contextual variables) Independent variables Model 1 b P Model 2 b P Socio-demographic Tertiary educated -.70 000 -.72 000 Male .06 401 Age -.01 000 .01 000 Unemployed .20 326 Factory worker . .57 .005 .50 012 Manual worker .10 .542 Working class .44 .000 .45 000 Household income category -.05 .017 -.05 009 NZ and European identity .64 .001 .65 000 Maori identity .91 .001 .90 000 Christian .10. .263 Religious attendance -.08 .000 -.08 000 Contextual Maori electorate (dummy) .52 .028 .51 024 Unemployment rate (UR) .05 .028 .05 010 Rate of Asian + Pacific + Other ethnic group (RAPO) .01 .321 .002 863 Rate of the tertiary educated -.01 .144 U R x R A P O -.001 .634 .0003 786 Cutting point I ' .30 .30 .37 .29 Cutting point 2 1.60 .30 1.67 .29 N 2940 3149 Log likelihood -2853.4 -3059.6 Pseudo R 2 .06 .06 81 Table 5.8 The estimated influence of single-variable manipulations on opposition to immigration (y = 0, 1 and 2), New Zealand: Model 2 a b Probability of the answer of Independent variables Default Change • Increase a lot/little Reduce a Reduce a or same (y = 0) little O = 1) (v = 2) :s.9.% 22.2% Tertiary educated 0 • Oto 1 66.3% 21.5% . 12.2% Age mean (48.3) mean to 90% point (70) 54.0% 27.1% 19.0% Factory worker 0 Otol 36.7% 31.1% 32.2% Working class 0 Oto 1 37.9%. 31.2% 30.9% Household income category mean ($38,100) mean to 6 ($88,400) 51.6% 28.0% • 20.4% NZ and European identity 1 0 64.4% 22.3% 13.2% Maori identity 0 Oto l c 42.7% 30.3% 27.0% Religious attendance 1(once a 1 to 5 (at least 57.0% 25.9% 17.1% year) once a week) a. The default personal characteristics for this probability prediction is not having received tertiary education, aged 48, not factory labor, not working class, with his/her household income of about $38,100 self-identified with New Zealand and European ethnic groups, who attend religious services once a year and live in a non-Maori electoral district with the unemployment rate of 8.3% (mean) and the rate of Asia, Pacific, and other ethnic groups of 10.8% (mean). b. All the influences of independent variables shown here are statistically significant at 95% level. c. As the Maori identity changes from 0 to I, NZ and European identity is set to 0. Table 5.8 suggests that about the half (49%) of the assumed typical respondents would answer the question "increase immigrants a little/lot" or "remain about the same as it is" (y = 0), another 30% would answer "reduce a little" (y — 1) and the rest (22%) would answer "reduce a lot" (y = 2). As found in Canadian and Australian data, receiving tertiary education reduces the chance of opposing immigration significantly: the probability of answering "reduce a little" (y = 1) decreases by about 7% points, and the probability of answering "reduce a lot" (y = 2) decreases by 10% points. The table also shows that both economic and cultural characteristics have a substantial influence on opposition to immigration. For example, as a cultural factor, the probability of opposing immigration increases by 6% points for the answer "reduce a little" (y = 1) and 9% points for the answer "reduce a lot" (y = 2), when respondents' ethnic identification changes to New Zealand and European from other ethnic groups (but not Maori). If the same 82 respondents identify themselves as New Zealand Maori, the probability increases by about 8% and 14% points respectively for the answer "reduce a little" and "reduce a lot." In a similar manner, being a factory worker increases the probability by about 10% points, and working class also increases the probability by 8% points for the answer "reduce a lot." This is a quite rough comparison of the impact of economic and cultural factors, but their effects of socio-demographic variables are constantly confirmed in New Zealand dataset. The coefficients of the contextual variables give contrasting evidence to Canadian and Australian findings. First, Model 1 shows that rate of the tertiary educated does not have a statistically significant effect on opposition to immigration, although the direction of its influence is as expected. With the control of UR, this means that the conformity effect to the attitudes of the tertiary educated is weak in 2002 NZES. Second, the effect of UR is relatively large and statistically significant, while rate of Asian, Pacific and other ethnic origins (RAPO) is not at the 95% level. Furthermore, an interaction term of UR and R A P O is positive but the effect is tiny and statistically insignificant. Strongly positive effect of UR can be seen in Figure 5.13, but the size of influence is not dependent on the level of R A P O : when typical respondents lives in an electorate with R A P O of 1.4%, the probability of answering "reduce a lot" (y =2) changes from 17%o to 32% as UR ranges from 3% to 18%>. Other things being equal, when 60% of the population are Asians, Pacific and other ethnic origins, the probability changes from 20% to 42% as UR changes in the same manner. The trajectory of these two lines looks similar, which means the positive influence of UR is not limited to the situation where only fewer ethnic minorities reside.4 8 This is an opposite result confirmed in Canada and Australia, implying that the local economic situation matters regardless of the concentration of immigrants.4 9 VIF is examined after running an OLS regression, and the VIF for R A P O and the interaction term are found relatively high, 15.0 and 13.5 respectively, so it is possible that the statistical insignificance of positive interaction effect is biased. 4 9 Different contextual variables are used for New Zealand case (RAPO) and for the other two countries (RIP). R A P O represents ethnic and cultural heterogeneity of the area more, while RIP may not always do so, because the majority of immigrants to Canada and Australia are from the U K and Ireland. Rate of ethnic minorities is tested in both countries by replacing RIP, but the result is the same. 83 Figure 5.13 The estimated influence of unemployment rate changes on opposition to immigration (interaction with RAPO), 2002 NZES (y = 2) a CD O x> m CD __ .E CD CD "S jS'CN X ! ro x i ^_ o ^ RAPO = 60.4% H RAPO =1.4% !•••• ^ 95% CI of RAPO = 60.4% 95% CI of RAPO = 1.4% 10 Unemployment rate (3-18) 15 a. The other variables are held constant in the same manner as in the footnote a of Table 5.8 except for unemployment rate (UR) and rate of Asian + Pacific + Other ethnic group (RAPO) 20 In Model 3 and 4 of Table 5.9, attitudinal variables are added to Model 2, and Model 5 includes some interaction effects between contextual and attitudinal variables. The attitudinal variables are quite different from those in Canada and Australia, but again, the influence of opposition to social/cultural groups (immigrants from Muslim countries and Asia) is much stronger than that of opposition to economic groups (skillful and manual labor immigrants), while all four variables strikingly increase the chance of opposing immigration. This is true when the variables are aggregated to construct a unified scale in several different ways as shown in Model 4. • ~ 84 Table 5.9 Ordered logistic regression on opposition to immigration, New Zealand (Models with attitudinal variables and interactions) Model 3 Model 4 Model 5 b P b P b P Socio-demographic Tertiary educated -.56 000 -.56 000 -.57 .000 Age -.01 000 -.01 000 -.01 .000 Factory worker .33 .135 .36 103 .37 .104 Working class .31 .001 .32 001 .32 .001 Household income category -.03 .227 -.03 168 -.03 .173 NZ and European identity .46 .024 .49 016 .49 .015 Maori identity .66 .019 .69 014 .69 .014 Religious attendance -.05 .012 -.06 .014 -.06 .014 Contextual Maori electorate (dummy) .34 .171 .36 .148 .34 .171 Unemployment rate (UR) .05 .007 .05 .008 .04 .305 Rate of Asian + Pacific + Other ethnic group (RAPO) .01 .011 .01 .017 .01 .214 Attitudinal Oppose economic types of immigrants .47 .000 .39 .003 Oppose immigrants with technical skills .60 .000 Oppose immigrants willing to do manual labor .32 .000 Oppose social/cultural types of immigrants .94 .000 .97 .000 Oppose immigrants from Muslim countries .72 .000 Oppose immigrants from Asia 1.29 .000 Interactions UR x oppose economic types of immigrants .009 .519 RAPO x oppose economic types of immigrants -.003 .856 RAPO x oppose cultural types of immigrants -.001 .784 Cutting point I 2.88 .32 2.79 .32 2.70 .40 Cutting point 2 • 4.47 .33 4.37 .32 4.29 .40 N 3050 3050 3050 Log likelihood -2583.7 -2591. 8 -259 .5 Pseudo R 2 .18 .18 .18 85 Figure 5.14 The estimated influence of opposition to economic types of migrants changes on opposition to immigration (interaction with UR), 2002 NZES (y =2)a UR = 3.1% UR = 18.1% H 95% Cl of UR = 3.1% i 95% Cl of UR = 18.1% co CD T3 Q) LO J cn c TO CO •s • — CM CO jQ O i-~ 1 : I I I 1 0 1 2 3 4 Strength of opposition to economic types of immigrants (1-4) The assumed characteristics are not tertiary educated, aged 48, not factory worker, not working class with the annual household income of about $38,300, self-identified with New Zealand and European ethnic groups, who attend religious service once a year, live in the electoral district with the RAPO of 10.6%. Other attitudinal and interaction variables are set at its mean. Model 5 includes interaction terms between contextual and attitudinal variables, and the interaction effect between UR and opposition to economic types is depicted Figure 5.14, when UR is held constant at its minimum and maximum point. First, none of these interaction effects are statistically significant at 95%, even when they are introduced separately.50 Despite its insignificance, a positive interaction effect of UR and opposition to economic types of immigrants is seen in Figure 5.14, when the slope of dashed line (when UR is at the maximum of 18.1%) is much steeper than the solid line (when UR is at the minimum of 3.1%). When the UR Multicolinearity is not a prolem: VIF is examined, but only the interaction between UR and economic types of immigrants has the higher score of 12.3, and that for the other variables is lower than ten. 86 is 3.1 %, the probability of answering "reduce a lot" (y = 2) increases from 7% to 32% by 25% points, while this increase is magnified to about 42%> points (from 17%o to 59%>) when the UR is 18.1 % as the strength of opposition to economic types of immigrants changes from minimum (0) to maximum (4) point. In contrast, there is no significant interaction effect between R A P O and attitudinal variables. There are at least three theoretical implications of these results: first, in New Zealand case, actual local economic situation measured by UR enhances the effect of perceived economic threat from immigrants. Second, no interaction effects involving the variable of R A P O as well as its weak positive influence imply that R A P O does not pose a cultural threat especially to those who have cultural concerns. Third, contact theory is not confirmed, i f we can assume that the higher concentration of ethnic minority groups means more frequent contact with them. Respondents who have stronger cultural and economic concerns are not more likely to oppose immigration by more frequent contact with those minorities. These results can be contrasted to the previous results obtained in Canadian and Australian cases. Why New Zealand dataset produced a different result is not clear at this point, but part of the reason may be a different operationaization of the attitudinal variables. Accordingly, the results suggest that both economic and cultural factors have strong influence on opposition to immigration, the stronger influence of cultural factor is found in attitudinal dimension, even when the effect of economic factor is intensified by the local economic situation. In contrast to Canadian and Australian case, higher concentration of minority groups or more frequent contact with immigrants does not intensify the effect of economic nor cultural attitudes on opposition to immigration. 5.4 Summary and comparison Accordingly, the logistic regression analyses in three countries show the following • results and implications; 1. Some socio-demographic variables that represent both economic and cultural characteristics substantially influence the level of opposition to immigration; 87 2. Cultural concerns are usually stronger than economic concerns; .3. The effect of the local economic situation is relatively weak, and interaction between UR and RIP, which is assumed to represent the labor competition with immigrants, is not significantly positive, but sometimes significantly negative, which implies that UR has a positive influence only in areas with fewer or no immigrants; 4. RIP often magnifies the effect of cultural concerns, which suggests that immigrant concentration poses a cultural threat especially to those who are concerned about cultural homogeneity, while such a threat is weakly perceived by those who are not strongly concerned about it; and 5. If RIP represents frequency of contact with immigrants, a statistically significant, positive interaction effect between RIP and cultural concerns suggests that the frequent contact with immigrants in a culturally competitive situation leads to a hostility towards them. Table 5.10 summarizes the results of all the regression analyses. It shows that tertiary education, working class, household income, native-born and religious attendance has a robust and consistent influence on opposition to immigration in three countries in different years. On the other hand, the influence of age, unemployment and ESC-born is weak and unstable, while the influence of Christianity is moderate but consistent. The contextual effects are not simple, and sometimes confusing. Rate of the tertiary educated (or university degree holders), however, consistently influences opposition to immigration, which implies that there is a conformity effect. The influence of UR and RIP is unstable, while RIP often positively contributes to opposition to immigration. The interaction effect between UR and RIP is negative in two out of five analyses. The inconsistent effect of these variables may be because opposition to immigration also depends on the level of national economic situation and the degree of salience of immigration issue at national level. The same could be said to the inconsistent effect of the interaction terms between contextual and attitudinal variables. Although the statistical insignificance of these 88 interactions may be caused by multicollinearity and the small number of observations, the national economic situation and salience of immigration issue may influence the viability of these influences. Therefore, the macroeconomic situation and the overall opposition to immigration level are compared in Table 5.11. Table 5.10 Directions of influence and the frequency of statistical significance' Direction and frequency of statistical significance Variables Prediction Canada Australia • NZ Total Socio-demographic Tertiary educated - -2 -2 -1 -5/5 Male -2 -0 +0 -2/5 Age +0 -0 -1 -1/5 Unemployed + + 1 ±0 +0 +1/5 Unemployment of the main income earner + +0 n.a. n.a. +0/1 Factory or unskilled worker + +0 n.a. + 1 +1/2 Working class + n.a. +2 + 1 +3/3 Household Income - -1 .. ' -2 -1 -4/5 Native-born + +2 +2 + 1 +5/5 ESC-born + +0 + 1 +1/4 Christian + +2 + 1 +0 +3/5 Religious attendance n.a. -2 -1 -3/3 Contextual Unemployment rate (UR)b + +0/-1 -1 + 1 +1,-2/5 Rate of immigrant population (RIP)b +/- + 1 -1/+1 + 1 +3, -1/5 Rate of the tertiary educated -2 -2 -0 -4/5 UR x RIP + - i -1 -0 -2/5 itudinal Economic concern + +2 +2 + 1 +5/5 Cultural concern + +2 +2 + 1 +5/5 Interactions UR x economic attitude + +0 -o • +0 +0/5 RIP x economic attitude + ±0 -0 -0 -0/5 RIP x cultural attitude + + 1 + 1-0 ' -0 +2/5 a. The signs in the table represent the direction of the influence, and figures are the frequency of statistical significance. For example, -2/5 in male and total section means that the variable has a negative statistically significant effect for twice out in 5 different dataset which include this variable. These results are based on Model 1 and 3 (in New Zealand, only Model 1) for socio-demographic and contextual variables, on Model 5 and 7 (in New Zealand Model 4) for attitudinal variables and on Model 6 and 8 (in New Zealand Model 5) for interactions. b. The individual statistical significance of UR and RIP is counted when the interaction effect is omitted. This result is not shown in previous Tables, but available by contact with the author. 89 Table 5.11 Summary of the overall results of statistical analyses and national contextual data' Variables 1993 Canada 2004 Australia 1996 2001 NZ 2002 Opposition toiimmigro 2S 9"., 63.4% National unemployment rate * 11.4% 7.2% / 10.6% 6.8% 5.2% ' Immigrant population 16.6%' 18.0% 23.3% 23.1% 18.4% Stronger socio-economic variable Cultural Cultural (On par) (On par) (On par) Interaction effect of UR and RIP * - * - + Stronger attitudinal variable Cultural (On par) Cultural Cultural Cultural Interaction effect of UR and economic attitude + + - - + Interaction effect of RIP and economic attitude + - - - -Interaction effect of RJP and cultural attitude + * + + * - -a. The sign in the table signifies the direction of the effect, and asterisk means statistical significance. For example, "-*" for the interaction variable between UR and RIP in 1993 in Canada means that the effect is negative and statistically significant. Source: For the national unemployment rate, OECD. 2006. OECD Economic Outlook. (June) no. 79. Paris: OECD, p. 178. For immigrant population, OECD. 2006. International Migration Outlook. Paris: OECD, p. 262. 'The data is of 1995. According to the table, the statistical significance of the negative interaction effect of UR and RIP and the positive interaction effect of RIP and a cultural attitudinal variable appears when the national unemployment rate is higher: in 1993 in Canada (11.4%) and 1996 in Australia (10.6%). In addition, because UR is positively correlated with the general level of opposition to immigration among the public in time series, they are also the time when more people oppose immigration. Furthermore, it should be noted that these are the years when immigration issues are salient in the national elections as discussed in Chapter 3. In Canada, the Reform Party made huge gains in the general election in 1993, and when the issue of the immigration was debated more often than in previous elections. In Australia, too, Pauline Hanson was elected as an independent MP, and her nationalistic and often xenophobic political messages about Asian immigrants were often the topic of the political debate in 1996. Although this simplistic analysis based on a mere covariation of the findings and events, i f this relationship is true, it is suggestive to understanding the strong link between the fluctuation of national economic situation and the general level of opposition to immigration among the public: economic competition with immigrants matter only when the economy is good. The local economic situation influences opposition to immigration in areas with few immigrants, and the influence of cultural essentialism is magnified in areas with more immigrants concentrated - and this appears when the national economy is in recession, and immigration issue is salient. It seems that what increases opposition to immigration during the recession is not local economic situation, but cultural concerns of the people who live in areas with many immigrants. 91 6. CONCLUSION, IMPLICATIONS AND FURTHER QUESTIONS Immigration has been hotly debated among the public in Canada, Australia and New Zealand in the 1990s and 2000s. In the debate, opposition to immigration is expressed for various reasons, but it almost always includes concerns about the economic and cultural impacts of immigrants, especially on employment, wage and social (dis)integration. The analysis of the election studies in various years in three countries shows that cultural concerns are more likely to increase the chance of opposition to immigration than economic concerns, while both factors play important role in determining attitudes to immigration policy. While there is a strong link between unemployment rate and the level of public opposition to immigration over time, the relationship between those variables is not confirmed in most of the cross-sectional analyses of the election studies. Further, the influence of unemployment rate is not amplified by the proportion of immigrant population. Instead, the result shows that geographic concentration of immigrants positively enhances the influence of cultural concerns when the level of opposition to immigration and unemployment rate are higher, but it does not relate to the influence of economic concerns. These findings suggest many implications for the understanding of politics of immigration. First, the psychological mechanism of support for anti-immigrant parties should include cultural concerns as a key variable to predict their opposition to immigration. Many studies on this topic include economic indicators such as unemployment rate, and often operationalize economic competition by rate of immigrant inflow to predict votes for anti-immigrant party (Lewis-Beck and Mitchell 1993; Jackman and Volpert 1996; Lubbers and Scheepers 2000; Golder 2003), but they often neglect variables that represent cultural concerns or attitudes toward social integration.52 Brechon and Mitra(1992) examined the relationship between concentration of immigrants and the percentage of votes for the Front National in 1988 5 2 A n exception to this is the work by Mughan and Paxton (2006). 92 legislative election in France and find that their positive correlation disappears as the size of the district of analysis becomes smaller. They conclude that supporters of anti-immigrant parties are not driven by the actual level of immigrants, but individuals' "fear of the immigrant population" (Brechon and Mitra 1992: 70). If this "fear" cannot be mapped geographically but is tightly connected with cultural concerns over immigration, operationalizing it by some contextual variables is difficult as discussed in Chapter 3. The future aggregate, time-series studies need to find out a good indicator of it, or use the change in attitudinal variables over time. Second, the results are also suggestive for the existing literature on the geography of immigration politics. Some previous studies on immigration argue that opposition to immigration does not crystallize, because immigrants provide geographically concentrated benefits and diffused costs, which would give people few incentives to organize opposition (Freeman 1995: 885). Money (1999: 64) finds that the opposition to immigration concentrated on several electoral districts, and their experiences of economic downturn and massive immigration inflow in the 1970s provided an electoral incentive to restrict the number of immigrants in U K , France and Australia. As far as the public opinion concerns, geographical concentration of immigrants might increase the chance of opposition, especially through cultural concerns, but the local economic situation does not intensify opposition to immigration.5" In Australia, Jupp (2002a: 129) points out that Oxley, the electoral district that elected Pauline Hanson in 1996 election by a swing of 19%, has only a few percent of Asian immigrants, but voters are often economically depressed. He also finds that a large electoral swing against Labor in 2001 election took place either in the districts characterized by many working class voters without ethnic diversity, or "a border-line between ethnic and 'Anglo ' areas" (Jupp 2002b: 265-266). Economic concerns certainly influence the level of opposition, but these two electoral characteristics imply that if the Money's indicators of immigrant pressures (1999: 96-97, 129-131, 179-180) have a problem, when she calls it "the measures of anti-immigrant pressure" because the indicators are just a product of the number of immigrants and unemployment rate, which does not represent the level of opposition to immigration per se. Further, it is tautological to see the correlation between the indicator of "anti-immigrant pressure" and unemployment rate or the number of immigrants as an evidence for the link between the concentrated costs of immigration and opposition to it. 93 concentration of immigrants matters, it is related to the influence of cultural concerns of immigration. The concentration of immigration, however, may not always increase the influence of cultural concerns on opposition to immigration, because it also provides a chance to contact and learn about immigrants' culture. For example, some recent research shows that Mosque construction in European cities was first resisted and refused by the local community and authority for fear of Muslim immigrants and their "different" culture, but such a resistance gradually subsided as the negotiation with local residents and authorities proceeded (Cesari 2005a: 1018-1019; Cesari 2005b; 1025-1029; Landman and Wessels 2005; Saint-Blancat and Schmidt di Friedberg 2005). Although the influence of the change in rate of immigrant population, as well as the length of time for which local communities have held racially/ethnically diverse immigrants is not examined in this study, a longer contact period with immigrants may decrease opposition to immigration. If the public's attitudes to immigration change in that way, their negative reaction to further immigration based on cultural concerns may better be understood as one phase of a long negotiation process of their social and cultural identity (Kastoryano, 2002). Many questions concerning opposition to immigration remain unanswered yet. First, do other concerns about immigrants than economic and cultural ones, such as fear of crimes or terrorism, independently explain the variation in opposition to immigration? What is the relationship among those variables? Is there any overlap among those different types of concerns? The measurement of cultural concerns is not uniform in this thesis, and is an amalgam of specific concerns about several cultural characteristics or people: what is termed as "traditional values" and certain types of hats in Canada, essential linguistic, religious, and native characteristics in Australia, and Muslim and Asians in New Zealand. What would be the best way to measure cultural concerns, and would a uniform measurement strategy still produce the same result? If so, are there any specific cultural concerns that increase opposition the most? For example, an influence of the concerns about linguistic difference may be different between in some de-facto multilingual countries like India and Indonesia and linguistically rather homogeneous countries 94 like Japan and Korea. In the United States, racial difference has been the most prominent social division than religion and language, while religious concerns may be more relevant to opposition to immigration in France. Specific symbols for cultural distinction and rejection may vary in different place and times, and these characteristics are often interrelated. Even so, each symbol, and concerns about it, may perform in a similar manner when it comes to opposition to immigration. Thus there are a lot of questions unanswered in this field and I hope the result of this thesis can provide some insight to further research on the determinants and mechanism of opposition to immigration in other times and places. 95 BIBLIOGRAPHY Addison, Thomas, and Christopher Worswick. 2002. "The Impact of Immigration on the Earnings of Natives: Evidence from Australian Micro Data." The Economic Record 78, no. 1:68-78. Allport, Gordon. 1954. The Nature of Prejudice. Cambridge, M A : Addison-Wesley. Amir, Yahuda. 1969. "Contact Hypothesis in Ethnic Relations." Psychological Bulletin 71, no. 5: 319-342. Amir, Yehuda. 1976. "The Role of Intergroup Contact in Change of Prejudice and Ethnic Relations." In Phyllis A . Katz, ed., Towards the Elimination of Racism. New York: Pergamon Press, 245-308. Barkan, Elliott R. 2003. "Return of the Nativists?" Social Science History 27, no. 2: 229-83. Bean, Clive, David Gow, and Ian McAllister. 1998. Australian Election Study, 1998. [Computer file]. Canberra: A S S D A , A N U . Bean, Clive, Davd Gow, and Ian McAllister. 2002. Australian Election Study 2001. [Computer file]. Canberra: Social Science Data Archives, The Australian National University. Bean, Clive, Ian McAllister, Rachel Gibson, and David Gow. 2005. Australian Election Study, 2004. [computer file]. Canberra: A S S D A , A N U . Blais, Andre, Joanna Everitt, Patrick Fournier, Elisabeth Gidengil, and Neil Nevitte. 2005. Canadian Election Study, 2004. [Computer file]. Montreal: Universite de Montreal. Blake, Donald. 2003. "Environmental Determinants of Racial Attitudes among White Canadians." Canadian Journal of Political Science 36, no. 3: 491-509. Blumer, Herbert. 1958. "Race Prejudice as a Sense of Group Position." Pacific Sociological Review 1: 3-7. Bobo, Lawrence. 1983. "Whites' Opposition to Busing: Symbolic Racism or Realistic Group Conflict?" Journal of Personality and Social Psychology 45, no. 6: 1196-1210. Bobo, Lawrence. 1988. "Group Conflict, Prejudice, and the Paradox of Contemporary Racial Attitudes." In Phyllis A . Katz and Dalmas A . Taylor, eds., Eliminating Racism: Profiles in Controversy. New York: Plenum Press, 85-114. Bobo, Lawrence. 1999. "Prejudice as Group Position: Microfoundations of a Sociological Approach to Racism and Race Relations." Journal of Social Issues 55, no. 3: 445-472. Bobo, Lawrence, and Frederick C. Licari. 1989. "Education and Political Tolerance: Testing the Effects of Cognitive Sophistication and Target Group Affect." The Public Opinion Quarterly 53, no. 3: 285-308. Bobo, Lawrence and Vincent Hutchings. 1996. "Perceptions of Racial Group Competition: 96 Extending Blumer's Theory of Group Position to a Multiracial Social Context." American Sociological Review 61, no. 6: 951 -972. Bonacich, Edna. 1972. " A Theory of Ethnic Antagonism: The Split Labor Market." American-Sociological Review 37, no. 5: 547'-559. Bond, Rod. 2005. "Group Size and Conformity." Group Processes and Intergroup Relations 8, no. 4:331-354. Borjas, George J. 1995. "The Economic Benefits from Immigration." Journal of Economic Perspectives 9, no. 2: 3-22. Borjas, George J. 1999. Heaven's Door: Immigration Policy and the American Economy. Princeton, NJ : Princeton University Press. Brewer, Marilynn B. 1996. "When Contact is not Enough: Social Identity and Intergroup Cooperation." International Journal of Intercultural Relation 20, no. 3: 291-303. Brechon, Pierre and Kumar Mitra. 1992. "The National Front in France: The Emergence of an Extreme Right Protest Movement." Comparative Politics 25, no. 1: 63-82. Castles, Stephen and Mark J. Miller. 2003. The Age of Migration. Basingstoke: Palgrave Macmillan. Cesari, Jocelyne. 2005. "Mosque Conflicts in European Cities: Introduction." Journal of Ethnic and Migration Studies. 31, no. 6: 1015-1024. Cesari, Jocelyne. 2005. "Mosques in French Cities: Towards the End of Conflict?" Journal of Ethnic and Migration Studies. 31, no. 6: 1025-1043. Citizenship and Immigration Canada. 1974-1996. Citizenship and Immigration Statistics various years. Ottawa: Minister of Supply and Services Canada. Citizenship and Immigration Canada. 2004. Facts and Figures 2004: Immigration Overview: Permanent and Temporary Residents. [On-line]. Page updated July 11, 2005, retrieved July 31, 2006: http://www.cic.gc.ca/english/pub/facts2004/permanent/9.html Citrin, Jack, Beth Reingold, and Donald P. Green. 1990. "American Identity and the Politics of Ethnic Change." The Journal of Politics. 52, no. 4: 1124-1154. Citrin, Jack, Donald P. Green, Christopher Muste, and Cara Wong. 1997. "Public Opinion toward Immigration Reform: The Role of Economic Motivations." Journal of Politics 59, no. 3: 858-881. Coenders, Marcel and Peer Scheepers. 1998. "Support for Ethnic Discrimination in the Netherlands 1979-1993: Effects of Period, Cohort, and Individual Characteristics." European Sociological Review 14, no. 4: 405-422. Collier, David, James Mahoney, and Jason Seawright. 2004. "Claiming Too Much: Warnings about Selection Bias." In Henry E. Brady and David Collier, eds., Rethinking Social Inquiry: Diverse Tools, Shared Standards. Lanham, M D : Rowman & Littlefield, 85-102. 97 Commonwealth of Australia. 1996. Parliamentary Debates. House of Representative Official Hansard (Ho. 8, 1996. 9, 10, 11, 12 September 1996). GPO, Commonwealth of Australia. Cook, Stuart. 1985. "Experimenting on Social Issues: The Case of School Desegration." American Psychologist 40, no 4: 452-460. Cornelius, Wayne, andTakeyuki Tsuda. 2004. "Controlling Immigration: The Limits of Government Intervention." In Wayne Cornelius, Takeyuki Tsuda, Philip Martin, and James Hollifield, eds., Controlling Immigration: A Global Perspective (Second edition). Stanford, C A : Stanford University Press, 3-48. Coser, Lewis. 1956. The Functions of Social Conflict. Glencoe, IL: Free Press. Curthoys, Ann. 2003. "Liberalism and Exclusionism: A Prehistory of the White Australia Policy." In Laksiri Jayasuriya, David Walker, and Jan Gothard, eds., Legacies of White Australia: Race, Culture and Nation. Crawley: University of Western Australia Press, 8-32. Economic Council of Canada. 1991. Economic and Social Impacts of Immigration. Ottawa: Minister of Supply and Services Canada. Espenshade, Thomas J. 2000. "Immigrants, Puerto Ricans, and Earnings of Native Black Males." In Gerald D. Jaynes, ed., Immigration and Race: New Challenges for American Democracy. New Haven: Yale University Press, 125-142. Environics. Various years. Environics Focus Canada. [On-line], Provided by Canadian Opinion Research Archive, Queens University, (no renewal date), retrieved July 31, 2006: http://jeff-lab.queensu.ca/poadata/info/fc/fclist.shtml Feldman, Stanley, and Leonie Huddy. 2005. "Racial Resentment and White Opposition to Race-Conscious Programs: Principles or Prejudice?" American Journal of Political Science 49, no. 1: 168-183. Fetzer, Joel. 2000. Public Attitudes toward Immigration in the United States, France, and Germany. New York: Cambridge University Press. Fitzgerald, Stephen (Chair). 1988. Immigration: A Commitment to Australia: The Report of the Committee to Advise on Australia's Immigration Policy. Canberra: Australian Government Publishing Service. Fleras, Augie and Paul Spoonley. 1999. Recalling Aotearoa: Indigenous Politics and Ethnic Relations in New Zealand. Auckland: Oxford University Press. Fletcher, Michael. 1999. Migrant Settlement: A Review of the Literature and Its Relevance to New Zealand. Report commissioned by the New Zealand Immigration Service. New Zealand Immigration Service, Department of Labor. [On-line government publications]. (No renewal date), retrieved July 31, 2006: http://www.dol.govt.nz/publication-view.asp?ID=47 Forbes, Hugh D. 1997. Ethnic Conflict: Commerce, Culture, and the Contact Hypothesis. New 98 ' Haven: Yale University Press. Ford, W. Scott. 1973. "Interractial Public Housing in a Border City: Another Look at the Contact Hypothesis." The American Journal of Sociology 78, no. 6: 1426-1447. Foster, William, and Glenn Withers. "Macroeconomic Consequences of International Migration." In Gary P. Freeman and James Jupp, eds., Nations of Immigrants: Australia, the United States, and International Migration. Melbourne: Oxford University Press, 90-110. Freeman, Gary. 1995. "Modes of Immigration Politics in Liberal Democratic States." International Migration Review 29, no. 4:881 -902. Gaertner, Samuel L. , John F. Dovidio, and Betty Bachman. 1996. "Revisiting the Contact Hypothesis: The Induction of a Common Ingroup Identity." International Journal of Intercultural Relation 20, no. 3: 271 -290. Gallup Australia, Inc. 1964. Opinion Poll D0317: Australian Gallup Poll, August 21, 1964. [On-line]. Provided by Australian Social Science Data Archive (ASSDA), The Australian National University (ANU), (no renewal date), retrieved December 6, 2003: http://assda.anu.edu.au/polls/D0317.html Gallup Australia, Inc. 1966. Opinion Poll D0324: Australian Gallup Poll, February 5, 1966. [On-line]. Provided by A S S D A , A N U , (no renewal date), retrieved December 6, 2003: http://assda.anu.edu.au/polls/D0324.html . Gallup Australia, Inc. 1977. Opinion Poll D0254: Australian Gallup Poll, March, 1977. [On-line]. Provided by A S S D A , A N U , (no renewal date), retrieved December 6, 2003: http://assda.anu.edu.au/polls/D0254.html Gallup Australia, Inc. 1984. Opinion Poll M0004: Morgan Gallup Poll, May, 1984 (Computer Reports). [On-line]. Provided by A S S D A , A N U , (no renewal date), retrieved December 6, 2003: http://assda.anu.edu.au/polls/M0004.html Gallup Australia, Inc. 1988. Opinion Poll M0054: Morgan Gallup Poll, August, 1988 (Computer Reports). [On-line]. Provided by A S S D A , A N U , (no renewal date), retrieved December 6, 2003: http://assda.anu.edu.au/polls/M0054.html Gallup Australia, Inc. 1989. Opinion Poll M0066: Morgan Gallup Poll, August, 1989 (Computer Reports). [On-line]. Provided by A S S D A , A N U , (no renewal date), retrieved December 6, 2003: http://assda.anu.edu.au/polls/M0066.html Gallup Canada, Inc. Various years. Canadian Institute of Public Opinion (CIPO). [On-line]. Provided by Carleton University Library Data Center. Updated July 28, 2006, retrieved July 31, 2006: http://www.library.carleton.ca/ssdata/surveys/pop_gallup.html Gibson, Rachel, lan McAllister, and Tami Swenson. 2002. "The Politics of Race and Immigration in Australia: One Nation Voting in the 1998 Election." Ethnic and Racial Studies 25, no. 5:' 823-844. 99 Gibson, Rachel, Shaun Wilson, David Denemark, Gabrielle Meagher, and Mark Western. 2003. The Australian Survey of Social Attitudes, 2003. [Computer file]. Canberra: A S S D A , A N U . Gilens, Martin. 1996. '"Race Coding' and White Opposition to Welfare." American Political Science Review 90, no. 3: 593-604. Gilens, Martin. 1999. Why Americans Hate Welfare: Race, Media, and the Politics of Antipoverty Policy. Chicago: The University of Chicago Press. Giles, Michael W. 1977. "Percent Black and Racial Hostility: An Old Assumption Reexamined." Social Science Quarterly 58, no. 3: 412-417. Giles, Michael W., and Arthur Evans. 1986. "The Power Approach to Intergroup Hostility." Journal of Conflict Resolution 30, no. 3: 469-486. Glaser, James M . 1994. "Back to the Black Belt: Racial Environment and White Racial Attitudes in the South." The Journal of Politics 56, no. 1:21-41. Golder, Matt. 2003. "Explaining Variation in the Success of Extreme Right Parties in Western Europe." Comparative Political Studies 36, no. 4: 432-466. Golebiowska, Ewa A . 1995. "Individual Value Priorities, Education, and Political Tolerance." Political Behavior 17, no. 1: 23-48. Goot, Murray. 1984. Constructing Public Opinion: The Polls and the Asian Immigration Debate. [Occasional Papers no. 5, September issue, published by The Ethnic Affairs Commission of NSW]. Sydney: The Ethnic Affairs Commission of NSW. Goot, Murray. 1998. "Hanson's Heartland: Who's for One Nation and Why." In Tony Abbott, et al., eds., Two Nations: The Causes and Effects of the Rise of the One Nation Party in Australia. Melbourne: Bookman, 51-74. Goot, Murray. 2000. "More 'Relaxed and Comfortable': Public Opinion on Immigration Under Howard." People and Place 8, no. 3: 46-60. Goot, Murray, and Ian Watson. 2001. "One Nation's Electoral Support: Where Does It Come From, What Makes It Different and How Does It Fit?" Australian Journal of Politics and History. 47, no. 2: 159-191. Goot, Murray, and Ian Watson. 2004. "Immigration, Multiculturalism and National Identity." In Shaun Wilson, Gabrielle Meagher, Rachel Gibson, David Denemark, and Mark Western, eds., Australian Social Attitudes: The First Report. Sydeney: University of New South Wales Press, 182-203., Gow, David John, Clive Bean, and Ian McAllister. 2000. Australian Constitutional Referendum Study, 1999. [Computer file]. Canberra: A S S D A , A N U . Hardcastle, Leonie, Ansrew Parkin, Alan Simmons and Nobuaki Suyama. 1994. "The Making of Immigration and Refugee Policy: Politicians, Bureaucrats and Citizens." In Haward Adelman, Allan Borowski, Meyer Burstein, and Lois Foster, eds., Immigration and 100 Refugee Policy: Australia and Canada Compared (volume 1). Melbourne: Melbourne University Press, 95-124. Hawkins, Freda. 1989. Critical Years in Immigration: Canada and Australia Compared. Kingston and Montreal: McGill-Queen's University Press. Henry, P. J., and David O. Sears. 2002. "The Symbolic Racism 2000 Scale." Political Psychology 23, no. 2: 253-283. Hogg, Michael A . and Dominic Abrams. 1988. Social Identifications: A Social Psychology oflntergroup Relations and Group Processes. London: Routledge. Hogg, Michael A . , Deborah J. Terry, and Katherine M . White. 1995. " A Tale of Two Theories: A Critical Comparison of Identity Theory with Social Identity Theory." Social Psychology Quarterly 58, no. 4: 255-269. Holton, Robert, and Michael Lanphier. 1994. "Public Opinion, Immigration and Refugees." In Haward Adelman, Allan Borowski, Meyer Burstein, and Lois Foster, eds., Immigration and Refugee Policy: Australia and Canada Compared (volume 1). Melbourne: Melbourne University Press, 125-148. Huddy, Leonie, and David O. Sears. 1995. "Opposition to Bilingual Education: Prejudice or the Defense of Realistic Interests?" Social Psychology Quarterly 58, no. 2: 133-143. Insko, Chester A . , John Schopler, James F. Kennedy, Kenneth R. Dahl, Kenneth A . Graetz, and Stephen M . Drigotas. 1992. "Individual-Group Discontinuity from the Differing Perspectives of Campbell's Realistic Group Conflict Theory and Tajfel and Turner's Social Identity Theory." Social Psychology Quarterly 55, no. 3: 272-291. International Labour Organization. 1970-2005. Year Book of Labour Statistics, various editions. Geneva: ILO. Irving Saulwick and Associates. 1991. Opinion Poll D0755: Saulwick Poll, October, 1991. [On-line]. Provided by A S S D A , A N U , (no renewal date), retrieved December 6, 2003: http://assda.anu.edu.au/polls/D0755.html Jackman, Mary R., and Marie Crane. 1986. '"Some of My Best Friends Are Black.. . ' Interracial Friendship and Whites' Racial Attitudes." The Public Opinion Quarterly 50, no. 4: 459-486. Jackman, Robert W., and Karin Volpert. 1996. "Conditions Favouring Parties of the Extreme Right in Western Europe." British Journal of PoliticalScience 26, no. 4: 501-521. Jedwab, Jack. 2006. "Canadians Remain World Leaders in Openness to Immigration: Tomorrow We Mark International Day for the Prevention of Racial Discrimination." [Online, press release report of Association of Canadian Studies, March 21, 2006] ACS Research & Polling - Canadian Opinion Online. (No update date), retrieved July 31, 2006: http://www.acs-aec.ca/Polls/Openness%20to%20Immigration.pdf 101 Jones, Roger, Ian McAllister, David Denemark and David Gow. 1993. Australian Election Study, 1993. [Computer file]. Canberra: SSDA, A N U . Jones, Roger, Ian McAllister, and David Gow. 1996. Australian Election Study, 1996. [Computer file]. Canberra: Social Science Data Archives, The Australian National University. Jupp, James. 1995."From 'White Australia'to'Part of Asia ' : Recent Shifts in Australian Immigration Policy Towards the Region." International Migration Review 29, no. 1: 207-228. Jupp, James. 2002a. From Australia to Woomera. Cambridge: Cambridge University Press. Jupp, James. 2002b. "Ethnicity and Immigration." In John Warhurst and Marian Simms, eds., 2007: The Centenary Election. St Lucia: University of Queensland Press: 261-270. Kalin, Rudolf, and John W. Berry. 1982. "The Social Ecology of Ethnic Attitudes in Canada." Canadian Journal of Behavioural Science 14, no. 2: 97-109. Kalin, Rudolf. 1996. "Ethnic Attitudes as a Function of Ethnic Presence." Canadian Journal of Behavioural Science 28, no. 3: 171-179. Kastoryano, Riva. 2002. Negotiating Identities: States and Immigrants in France and Germany. Translated by Barbara Harshav. Princeton: Princeton University Press. Kinder, Donald R., and David O. Sears. 1981. "Prejudice and Politics: Symbolic Racism versus Racial Threats to the Good Life." Journal of Personality and Social Psychology 40, no. 3: 414-431. Kinder, Donald R., and Lynn Sanders. 1996. Divided by Color: Racial Politics and Democratic Ideals. Chicago: University of Chicago Press. Kinder, Donald R., and Tali Mendelberg. 1985. "Cracks in American Apartheid: The Political Impact of Prejudice among Desegregated Whites." The Journal of Politics 57, no. 2: 402-424. Kinder, Donald R., and Tali Mendelberg. 2000. "Individualism Reconsidered: Principles and Prejudice in Cotemporary American Opinion." In David O. Sears, Jim Sidanius and Lawrence Bobo, eds., Racialized Politics: The Debate about Racism in America. Chicago: The University of Chicago Press, 44-74. Knowles, Valerie. 2000. Forging Our Legacy: Canadian Citizenship and Immigration, 1900-1977. Ottawa: Public Works and Government Services Canada. Kopras, Andrew. 2000. "Electorate Rankings: Census 1996 (2000 Electoral boundaries)." Research Paper, 11, 2000-2001. [On-line paper posted on Australian Parliamentary Library Website]. Canberra: Commonwealth of Australia. Page updated September 27, 2001, retrieved November 30, 2004. http://www.aph.gov.au/library/pubs/rp/2000-01/01 RP11 .htm Kopras, Andrew. 2003. "Electorate Rankings: Census 2001." Research Paper, 2, 2002-03. [On-line paper posted on Australian Parliamentary Library Website]. Canberra: 102 Commonwealth of Australia. Page updated February 24, 2003, retrieved 30, 2004: http://www.aph.gov.au/library/pubs/rp/2002-03/03RP02.htm Lahav, Gallya. 2004. Immigration and politics in the New Europe: Reinventing Borders. Cambridge: Cambridge University Press. LaLonde, Robert J., and Robert H. Topel. 1991. "Immigrants in the American Labor Market: Quality, Assimilation, and Distributional Effects." The American Economic Review 81, no. 2:297-302. Landman, Nico and Wendy Wessels. 2005. "The Visibility of Mosques in Dutch Towns." Journal of Ethnic and Migration Studies 31, no. 6: 1125-1140. Laycock, David. 2002. The New Right and Democracy in Canada: Understanding Reform and the Canadian Alliance. Ontario: Oxford University Press. Levine, Robert A . , and Donald T. Campbell. 1972. Ethnocentrism: Theories of Conflict, Ethnic Attitudes, and Group Behavior. New York: John Wiley & Sons. Lewis-Beck, Michael, and Glenn E. Mitchell II. 1993. "French Electoral Theory: The National Front Test." Electoral Studies 12, no. 2: 112-127. Lieberson, Stanley. 1991. "Small N's and Big Conclusions: An Examination of the Reasoning in Comparative Case Studies Based on a Small Number of Cases." Social Forces 70.2: 307-320. Link, Michael W., and Robert W. Oldendick. 1996. "Social Construction and White Attitudes Toward Equal Opportunity and Multiculturalism." The Journal of Politics 58, no. 1: 149-168. Lubbers, Marcel, and Peer Scheepers. 2000. "Individual and Contextual Characteristics of the German Extreme Right-wing Vote in the 1990s: A Test of Complementary Theories." European Journal of Political Research 38, no. 5: 63-94. Lusztig, Michael, and Matthew Wilson. 2005. " A New Right? Moral Issues arid Partisan Change in Canada." Social Science Quarterly 86, no. 1: 109-128. Marcus-Newhall, Amy, Norman Miller, Rolf Holtz, and Marilynn B. Brewer. 1993. "Cross-cutting Category Membership with Role Assignment: A Means of Reducing Intergroup Bias." British Journal of Social Psychology 32: 124-146. Markus, Andrew. 2003. " O f Continuities and Discontinuities: Reflections on a Century of Australian Immigration Control." In Laksiri Jayasuriya, David Walker and Jan Gothard, eds., Legacies of White Australia: Race, Culture and Nation. Crawley: University of Western Australia Press, 175-189. McAllister, Ian. 1993. "Immigration, Bipartisanship and Public Opinion." In James Jupp and Marie Kabala, eds., The Politics of Australian Immigration. Canberra: Australian Government Publishing Service, 161-178. 103 McAllister, Ian, Roger Jones, Elim Papadakis, and David Gow. 1990, Australian Election Study, 1990. [Computer file]. Canberra: A S S D A , A N U . Miller, Raymond. 1998. "New Zealand First." In Hans-Georg Betz and Stefan lmmerfall, eds., The New Politics of the Right: New-Populist Parties and Movements in Established Democracies. New York: St. Martin's Press, 203-210 Money, Jeannette and Kimberly Cole. 1999. "Immigration Control in Australia." In Jeannette Money. Fences and Neighbors: The Political Geography of Immigration Control. Ithaca: Cornell University Press, 157-205. Money, Jeannette. 1999. Fences and Neighbors: The Political Geography of Immigration Control. Ithaca: Cornell University Press. Monroe, Kristen R., James Hankin, and Renee B. Van Vechten. 2000. "The Psychological Foundations of Identity Politics." Annual Review of Political Science 3: 419-447. Mughan, Anthony and Pamela Paxton. 2006. "Anti-Immigrant Sentiment, Policy Preferences and Populist Party Voting in Australia." British Journal of Political Science 36, no. 2: 341-358. Munshi, Debashish. 1998. "Media, Politics, and the Asianisation of a Polarised Immigration Debate in New Zealand." Australian Journal of Communication 25, no. 1: 97-110. National Multicultural Advisory Council, 1999. "Statistical Appendix." Australian Multiculturalism for a new Century: Towards Inclusiveness. [On-line], (pp. 39-44). Canberra: Commonwealth of Australia. Page updated November 30, 2003, retrieved December 6, 2003: http://www.immi.gov.au/multicultural/_inc/pubIications/nmac/statistics.pdf Newspoll. (No Title). [On-line], (no renewal date), retrieved December 6, 2003: http://newspoll.com.aU/image_uploads/cgi-lib.27322.l.0403IMMI.pdf New Zealand Planning Council. On the Move: Migration and Population -Trends and Policies (Report No. 6). Wellington: Population Monitoring Group, New Zealand Planning Council. Northrup, David A . , and Anne E. Oram. 1994. The 1993 Canadian Election Study, Incorporating the 1992 Referendum Survey on the Charlottetown Accord. [Computer file]. Toronto: Institute for Social Research, York University. Organization for Economic Co-Operation and Development (OECD). 1995-2004. Trends in International Migration: Annual Report Various Editions. Paris: O E C D . O E C D . 2006. International Migration Outlook: Annual Report 2006 Edition. Paris: O E C D . O E C D . 2006. OECD Economic Outlook (June, no. 79). Paris: O E C D . Oliver, J. Eric, and Tali Mendelberg. 2000. "Reconsidering the Environmental Determinants of White Racial Attitudes." American Journal of Political Science 44, no. 3: 574-589. Olzak, Susan. 1992. The Dynamics of Ethnic Competition and Conflict. Stanford: Stanford University Press. 104 Ongley, Patrick. 2004. "Ethnicity, Migration and the Labour Market." In Paul Spoonley, Cluny Macpherson and David Pearson, eds., Tangata Tangata: The Changing Ethnic Contours of New Zealand. Southbank Victoria: Thomson Dunmore Press, 199-220. Ongley; Patrick, and David Pearson. 1995. "Post-1995 International Migration: New Zealand, Australia and Canada Compared." International Migration Review 29, no. 3: 765-793. Ozolins, Uldis. 1994. "Immigration and Immigrants." In Judith Brett, James Gillespie, and Murray Goot, eds., Developments in Australian Politics. Hong Kong: Macmillan Education Australia, 202-216. Palmer, Douglas. 1996. "Determinants of Canadian Attitudes Toward Immigration: More Than Just Racism?" Canadian Journal of Behavioural Science 28, no. 3: 180-192. Phillips, Jock. 2006. "History of Immigration." TeAra — Encyclopedia of New Zealand. Ministry for Culture and Heritage. [On-line government publication]. Page updated June 9, 2006, retrieved 31, July 2006: http://www.teara.govt.nz/NewZealanders/NewZealandPeoples/HistoryOflmmigration/ Przeworski, Adam, and Henry Teune. 1970. The Logic of Comparative Social Inquiry. New York: John Wiley & Sons. Quillian, Lincoln. 1995. "Prejudice as a Response to Perceived Group Threat: Population Composition and Anti-immigrant and Racial Prejudice in Europe." American Sociological Review 60, no. 4: 586-611. Ragin, Charles. 2004. "Turning the Tables: How Case-Oriented Research Challenges Variable-Oriented Research." In Henry E. Brady and David Collier, eds., Rethinking Social Inquiry: Diverse Tools, Shared Standards. Lanham, M D : Rowman & Littlefield, 123-138. Reitz, Jeffrey. 2004. "Canada: Immigration and Nation-Building in the Transition to a Knowledge Economy." In Wayne Cornelius, Takeyuki Tsuda, Philip Martin, and James Hollifield, eds., Controlling Immigration: A Global Perspective (Second edition). Stanford, C A : Stanford University Press, 97-133. Riketta, Michael. 2005. "Cognitive Differentiation Between Self, Ingroup, and Outgroup: The Roles of Identification and Perceived Intergroup Conflict." European Journal of Social Psychology 35: 97-106. Robinson, Jerry W., and James D. Preston. 1976. "Equal-Status Contact and Modification of Racial Prejudice: A Reexamination of the Contact Hypothesis." Social Forces 54, no. 4: 911-924. Rosenberg, Milton. 1956. "Cognitive Structure and Attitudinal Affect." The Journal of Abnormal and Social Psychology 3: 367-372. Saint-Blancat, Chantal and Ottavia Schmidt di Friedberg. 2005. "Why Are Mosques a Problem? Local Politics and Fear of Islam in Northern Italy." Journal of Ethnic and Migration 105 Studies 3\, no. 6: 1083-1104. Scheve, Kenneth F., and Matthew J. Slaughter. 2001. "Labor Market Competition and Individual Preferences over Immigration Policy." The Review of Economics and Statistics 83, no. 1: 133-145. Schissel, Bernard, Richard A . Wanner, and James S. Frideres. 1989. "Social and Economic Context and Attitudes toward Immigrants in Canadian Cities." International Migration Review 23, no. 2, 289-308. Schuman, Howard, Charlotte Steeh, and Lawrence Bobo. 1985. Racial Attitudes in America: Trends and Interpretations. Cambridge, M A : Harvard University Press. Sears,'David O. 1988. "Symbolic Racism." In Phyllis A . Katz and Dalmas A . Taylor, eds., Eliminating Racism: Profiles in Controversy. New York: Plenum Press, 53-84. Sears, David O., Carl P. Hensler, and Leslie K. Speer. 1979. "Whites' Opposition to 'Busing': Self-interest or Symbolic Politics?" The American Political Science Review 73, no. 2: 369-384. Sears, David O., and Donald R. Kinder. 1985. "Whites' Opposition to Busing: On Conceptualizing and Operationalizing Group Conflict." Journal of Personality and Social Psychology 48, no. 5: 1141-1147. Sears, David O., Colette Van Laar, Mary Carrillo, and Rick Kosterman. 1997. "Is It Really Racism?" 777e Public Opinion Quarterly 61, no. 1: 16-53. Sears, David O., John J. Hetts, Jim Sidanius, and Lawrence Bobo. 2000. "Race in American Politics: Framing the Debate." In David O. Sears, Jim Sidanius, and Lawrence Bobo, eds., Racialized Politics: The Debate about Racism in America. Chicago: University of Chicago Press, 1-43. Shan, Jordan, Alan Morris, and Fiona Sun. 1999. "Immigration and Unemployment: New Evidence from Australia and New Zealand." International Review of Applied Economics 13, no. 2: 253-260. Sherif, Muzafer, O; J. Harvey, B. Jack White, William R. Hood, and Carolyn W. Sherif. 1961. Intergroup Conflict and Cooperation: The Robbers Cave Experiment. Oklahoma: University of Oklahoma Institute of Intergroup Relations. Sigelman, Lee, Timothy Bledsoe, Susan Welch, and Michael W. Combs. 1996. "Making Contact? Black-White Social Interraction in an Urban Setting." The American Jourrnal of Sociology 101, no. 5: 1306-1332. Simon, Rita J. and James P. Lynch. 1999. " A Comparative Assessment of Public Opinion toward Immigrants and Immigration Policies." International Migration Review 33, no. 2: 455-467. Simon, Julian L., Stephen Moore, and Richard Sullivan. 1993. "The Effect of Immigration on Aggregate Native Unemployment: A n Across-city Estimation" Journal of Labor Research 106 14, no. 3: 299-316. Sniderman, Paul M . , and Philiop E. Tetlock. 1986. "Symbolic Racism: Problems of Motive Attribution in Political Analysis." Journal of Social Issues 42, no. 2: 129-150. Sniderman, Paul M . , and Thomas Piazza. 1993. The Scar of Race. Cambridge: The Belknap Press Harvard University Press. Sniderman, Paul M . , Edward G. Carmines, Geoffrey C. Layman, and Michael Carter. 1996. "Beyond Race: Social Justice as a Race Neutral Ideal." American Journal of Political Science 40, no. 1: 33-55. Sniderman, Paul M . , and Edward G. Carmines. 1997. Reaching Beyond Race. Cambrirdge, M A : Harvard University Press. Sniderman, Paul M . , Gretchen C. Crosby, and William G. Howell. 2000. "The Politics of Race." In David O. Sears, Jim Sidanius, and Lawrence Bobo, eds., RacializedPolitics: The Debate about Racism in America. Chicago: The University of Chicago Press, 236-279. Sniderman, Paul M . , Louk Hagendoorn, and Markus Prior. 2004. "Predisposing Factors and Situational Triggers: Exclusionary Reactions to Immigrant Minorities." American Political Science Review 98, no. 1: 35-49. Soberman, Liane. 1999. "Immigration and the Canadian Federal Election of 1993: The Press as a Political Educator." In Harold Troper, and Morton Weinfeld. Ethnicity, Politics, and Public Policy: Case Studies in Canadian Diversity. Toronto: University of Toronto Press, 253-281. Soper, Christopher, J., and Joel Fetzer. 2002. "Religion and Politics in a Secular Europe." In Ted Gerard Jelen and Clyde Wilcox, eds., Religion and Politics in Comparative Perspective: The One, the Few, and the Many. Cambridge University Press, 169-191. Stang, David J. 1976. "Group Size Effects on Conformity." The Journal of Social Psychology 98:175-181. Statistics Canada. Labour Force Characteristics. [On-line]. Page updated Jan 24, 2006, retrieved July 31, 2006: http://www40.statcan.ca/101/cst01/econl0.htm Statistics Canada. Various years. Labour Force Survey Estimates (Table 282-0002). [On-line]. Page updated Jan 4, 2006, retrieved July 31: http://estat.statcan.ca/cgi-win/CNSMCGI.EXE?regtkt=&C2Sub=&ARRAYID=2820002& C2DB=EST&VEC=&LANG=E&SDDSLOC=%2F%2Fwww.statcan.ca%2Fenglish%2Fs dds%2F*.htm&ROOTDIR=ESTAT%2F&RESULTTEMPLATE=ESTAT%2FCII_PICK& ' ARRAY_PICK=1&SDDSID=&SDDSDESC= Statistics New Zealand. 2002. 2001 Census of Population and Dwellings: Electoral Profde (2002 Boundaries). Auckland: Statistics New Zealand. Steeh, Charlotte, and Maria Krysan. 1996. "Trends: Affirmative Action and the Public, 107 1970-1995." The Public Opinion Quarterly 60, no. 1: 128-158. Stephan, Walter G. 1987. "The Contact Hypothesis in Intergroup Relations." In Clyde Hendrick, ed., Group Processes and Intergroup Relations . Vol. 9 of Review of Personality and Social Psychology. Newbury Park, C A : Sage Publications, 161-178. Stephan, Walter G., Oscar Ybarra, Carmen M . Martinez, Jeseph Schwarzwald, and Michal Tur-Kaspa. 1998. "Prejudice Toward Immigrants to Spain and Israel: An Integrated Threat Theory Analysis." Journal of Cross-Cultural Psychology 29, no. 4: 559-576. Stephan, Walter G , Cookie White Stephan, and William B. Gudykunst. 1999. "Anxiety in Intergroup Relations: A Comparison of Anxiety/Uncertainty Management Theory and Integrated Threat Theory." International Journal of Intercultural Relation 23, no. 4: 613-628. Stoffman, Daniel. June 1993. Toward a More Realistic Immigration Policy for Canada. Toronto: Howe Institute. Tajfel, Henri. 1970. "Experiments in Intergroup Discrimination." Scientific American 223: 96-102. Tajfel, Henri. 1982. "Social Psychology of Intergroup Relations." Annual Review of Psychology 33: 1-39. Tajfel, Henri, and John C. Turner. 1979. " A n Integrative Theory of Intergroup Conflict." In William G. Austin and Stephen Worchel, eds., 777e Social Psychology of Intergroup Relations. Monterey, C A : Brooks/Cole, 33-48. Tarman, Christopher, and David O. Sears. 2005. "The Conceptualization and Measurement of Symbolic Racism." The Journal of Politics 67, no. 3: 731-761. Tavan, Gwenda. 2005. The Long, Slow Death of White Australia. Carlton North: Scribe Publications. Tienhaara, Nancy. 1974. Canadian Views on Immigration and Population: An Analysis of Post-war Gallup Polls. Ottawa: Manpower and Immigration (Canada). Trlin, Andrew, and Noel Watts. 2004. "Immigration Policy and Immigrant Settlement: A Flawed Relationship at the Turn of the Millennium." In Paul Spoonley, Cluny Macpherson, and David Pearson^ eds., Tangata Tangata:, The Changing Ethnic Contours of New Zealand. Southbank Victoria: Thomson Dunmore Press, 111 -134. Troper, Harold. 1994. "Canada's Immigration Policy Since 1945." International Journal 48, no.2: 255-281. Tsukishima, Ronald T., and Darrel Montero. 1976. "The Contact Hypothesis: Social and Economic Contact and General Chnages in the Study of Black Anti-Semitism." Social Forces 55, no. 1: 149-165. Verkuyten, Maykel, and Peary Brug. 2004. "Multiculturalism and Group Status: The Role of 108 Ethnic Identification, Group Essentialism and Protestant Ethic. European Journal of Social Psychology 34 no. 6: 647-661. Vowles, Jack, Peter Aimer, Raymond Miller, Susan Banducci, and Jeffrey Karp. 2002. 2002 New Zealand Election Study. [Computer file]. Auckland: New Zealand Election Study, (no renewal date), retrieved 10 Apr i l , 2006. http://www.nzes.org/exec/getdata/nzes02.zip Vowles, Jack, Peter Aimer, Susan Banducci, Jeffrey Karp and Raymond Miller. 2004. Voters' Veto: The 2002 Election in New Zealand and the Consolidation of Minority Government. Auckland: Auckland University Press. Walker, David. 2003. "Race Building and the Disciplining of White Australia." In Laksiri Jayasuriya, David Walker and Jan Gothard, eds., Legacies of White Australia: Race, Culture and Nation. Crawley: University of Western Australia Press, 33-50. Zajonic, Robert B. 1968. "Attitudinal Effects of Mere Exposure." Journal of Personality and Social Psychology Monograph Supplement 9, no. 2, part 2: 1 -27. 109 APPENDICES Appendix A : Question wordings and coding A . l Canada ) Opposition to immigration (variable code: cpsg5 in 1993, cpsp9 in 2004) i Question: "Do you think Canada should admit MORE immigrants or FEWER immigrants than at present?" Coding: "More," "depends," "stay the same," "don't know," and refused = 0, and "Fewer" = i . Tertiary educated (variable code: cpsolO and refn2 in 1993, cpss3 in 2004) Question: "What is the highest level of education that you have completed?" Coding: "Some University," "Bachelors," "Master," "Professional degree or Ph.D." = 1, other answers that are lower level of educations than university = 0 Male (code: cpsrgen and refrgen in 1993, cpsrgen in 2004) No question wordings. Interviewer recorded respondent's gender. Coding: Male = 1, female = 0. Age category (code: cpsage in 1993, cpssl in 2004) Question: "In what year were you born?" Coding calculation: floor [{1993 or 2004 - (answered year of birth)} /10]. "Don't know" answer and refused are treated as missing. Quebec residence (code: cpsprov in 1993, province in 2004) No question wording. Interviewer recorded the province of interview. Coding: "Quebec" = 1, other province = 0. Unemployed (code: cpsjobl in 1993, cpss4 in 2004) Question: "Are you presently working for pay, are you unemployed, retired, a student, or a homemaker?" (1993); "Are you currently self employed, working for pay, retired, unemployed or looking for work, a student, caring for a family, or something else?" (2004) Coding: "Laid off," "unemployed/looking for work" = 1, "working now," "retired," "disabled," "student," "homemaker," "self employed," "working for pay," "caring for a family," "don't know," and refused = 0. Main income earner unemployed in past 1 year (code: cpsjobl and cpsjob7, only available in 1993) Question: "Were you/Was your family's principal wage earner out of work or laid off at any time during the last year?" Coding: "Yes" = 1, "no," "don't know," refused or missing = 0 Semi-or unskilled labor (code: cpspinpr only available in 1993) Question: No question wording recoded in the codebook. Reported classification is based on Pineo, Porter and McRoberts' (1985) socio-economic index classification. Coding: "semi-skilled clerks and sales," "semi-skilled manual," "unskilled clerks and sales," or "unskilled manual" = 1, other = 0. Income category (code: cpsol8 and cpso 18a in 1993, cpssl8 in 2004) Question: "Could you please tell me your total household income. Be sure to include income FROM A L L 110 SOURCES such as savings, pensions, rent, as well as wages, TO THE NEAREST THOUSAND DOLLARS, what was your TOTAL HOUSEHOLD INCOME before taxes and other deductions for 1992?" (cpsol 8); "We don't need the exact amount; could you tell me which of these broad categories it falls into ..." (cpsol 8a); "Now your last year's total household income before taxes. That includes income FROM A L L SOURCES such as savings, pensions, rent, as well as wages. Was it. . ." (cpssl 8) Coding calculation for cpsol8: floor [((answered income values in thousands)}/10]. If the result is zero, 1 is assigned. Coding for cpsol8a and cpssl8: <$20,000 = 1, $20,000-529,000 =2, $30,000-$39,000 =3, $40,000-$49,000 =4, $50,000-$59,000 =5, $60,000-$69,000 =6, S70,000-$79,000 =7, $80,000-$89,000 =8, $90,000-5100,000 =9, >$ 100,000 = 10, "don't know" and refused are treated as missing values. Native-born (code: cpsol 1 and refn!2 in 1993, cpssl2 in 2004) Question: "In what country were you born?" Coding: Canada = 1, other countries and "don't know" and refused = 0 ESC-born (code: cpsoll and refnl2 in 1993, cpssl2 in 2004) Question: "In what country were you born?" Coding: U.K., Ireland, U.S.A., India = 1, other countries and "don't know" and refused = 0 in 1993; US, Australia, Bahamas, Barbados, Belize, Bermuda, Britain, Dominica, Hong Kong, Ireland, India, Nigeria, Philippines, Scotland, Trinidad, Wales = 1, other countries and "don't know" or not codable = 0 in 2004. Christian (code: cpso9 in 1993, cpss9 in 2004) Question: What is your religious affiliation? Is it Protestant, Catholic, Jewish, another religion, or none?" (1993); "Please tell me what is your religion, if you have one?" (2004) Coding: "Protestant" and "Catholic" = I, "Jewish," "other," "none," "don't know," and refused = 0 in 1993; "Anglican," "Baptist," "Catholic/Roman Catholic," Greek/ Ukrainian Orthodox/ Russian Orthodox/ Eastern Orthodox," "Jehovah's witness," "Lutheran," "Mormon/ Latter Day Saints," "Pentecostal/ Fundamentalist," "Presbyterian," "Protestant (only after probe)," United Church of Canada," "Christian (only after probe)," "Christian Reform," "Salvation Army," "Mennonite" = 1, "none, don't have one," "Buddhist,/ Buddhism," "Hindu," Jewish/ Judaism/ Jewish Orthodox," "Muslim/ Islam," "Sikh/ Sikhism," "other," "don't know/ agnostic" = 0 in 2004. Unemployment rate (by Forward Sortation Area, or FSA) Coding calculation: Unemployed population above age 15 divided by labor force above age 15 and multiplied by 100. Rate of immigrant population (by FSA) Coding calculation: Total immigrant population divided by total population and multiplied by 100. Rate of the university degree holders (by FSA) Coding calculation: Total population with university degree divided by total population and multiplied by 100. Retrospective evaluation of personal financial situation (code: cpscl, cpscla and cpsclb in 1993, cpsfl in 2004) Question: "Would you say that you are BETTER off or WORSE off financially than you were a year ago?" (cpscl); if the answer is "BETTER," "Is that MUCH better off or SOMEWHAT better off? (cpscla), and if the answer is "WORSE," "Is that MUCH worse off or SOMEWHAT worse off? (cpsclb) is 111 asked in 1993; "Financially, are you better off, WORSE off, or about the same as a year ago?" in 2004. Coding: "Much better" = 1, "somewhat better" = 2, "same," "don't know," and refused = 3, "somewhat worse" = 4, "much worse" = 5 in 1993; "better off = 1 "about the same," "don't know" and refused = 2, and "worse of f = 3 in 2004. Prospective evaluation of personal financial situation (code: cpsc2, cpsc2a and spcs2b in 1993, cpsO in 2004) Question: "Do you think that a year from now you will be BETTER off financially, WORSE off, or just about the same as now?" (cpsc2); "Is that MUCH better off or SOMEWHAT better off?" (cpsc2a); "Is that MUCH worse off or SOMEWHAT worse off?" (cpsc2b) in 1993; "Do you think that a year from now you will be BETTER off financially, WORSE off, or about the same as now?" in 2004. Coding: Same as in Retrospective evaluation of personal financial situation Retrospective evaluation of provincial economy (code: cpse 1, cpce 1 a and cpse 1 b only available in 1993) Question: "What about economic conditions in [province name]? Would you say that OVER THE PAST YEAR economic conditions in [province name] have gotten better, stayed about the same, or gotten worse?" (cpsel); "Is that MUCH better or SOMEWHAT better?" (cpsela); "Is that MUCH worse or SOMEWHAT worse?" (cpselb) Coding: Same as in Retrospective evaluation of personal financial situation. Retrospective evaluation of national economy (code: cpshl, cpshla and cpshlb in 1993; cpsml in 2004) Question: "Now, 1 want to ask you about the economy in all of CANADA. Would you say that over the PAST YEAR Canada's economy has GOTTEN BETTER, STAYED ABOUT THE SAME, or GOTTEN WORSE?" (cpshl); "Is that MUCH better or SOMEWHAT better?" (cpshla); "Is that MUCH worse or SOMEWHAT worse?" (cpshlb) in 1993; "Now, I want to ask you about the economy. Over the PAST YEAR, has CANADA'S economy: gotten better, gotten worse, or stayed about the same?" in 2004. Coding: Same as in Retrospective evaluation of personal financial situation in 1993; "gotten better" = 1, "stay about the same," "don't know," and refused = 2, "gotten worse" = 3 in 2004. Prospective evaluation of national economy (code: cpsh2 in 1993; cpsm3 in 2004) Question: "What about the NEXT 12 months? Do you expect Canada's economy to get BETTER, STAY ABOUT THE SAME, or GET WORSE?" in 1993; "What about the NEXT 12 months? Will CANADA'S economy: get better, get worse, or stay about the same?" in 2004. Coding: "Better" = 1, "Same," "don't know," and refused = 2, and "worse" = 3 in 1993; "get better" = 1, "stay about the same" and "depends who wins the election" = 2, "get worse" = 3 in 2004. Federal government made respondents worse off (code: cpsc3 in 1993, cpsf2a and cpsf2b in 2004) Question: "Have the policies of the FEDERAL government made you BETTER off, WORSE off, or HAVEN'T THEY MADE MUCH OF A DIFFERENCE either way?" in 1993; "Have the policies of the Federal government made you BETTER off, or haven't they made much difference?" (cpsf2a), "Have the policies of the Federal government made you WORSE off, or haven't they made much difference?" (cpsf2b) in 2004. Coding: "Better off = 1, "No difference," "don't know" and refused = 2, "Worse of f = 3 in 1993; "better off = 1, "haven't made much difference," "don't know" and refused = 2, "worse off = 3, and all the • respondents who are not answered this question are coded 2 in order to maintain the number of observations. 112 Federal government made respondents' provincial economy worse (code: cpse2, cpse2a and cpse2b only available in 1993) Question: "Have the policies of the FEDERAL government made the Respondent's province> economy BETTER, WORSE, or HAVEN'T THEY MADE MUCH DIFFERENCE either way?" (cpse2), "Is that MUCH better or SOMEWHAT better?" (cpse2a), "Is that MUCH worse or SOMEWHAT worse?" (cpse2b) Coding: Same as in Retrospective evaluation of personal financial situation in 1993. Federal government made Canadian economy worse (code: cpsh3, cpsh3a and cpsh3b in 1993; cpsmla and cpsmlb in 2004) Question: "Have the policies of the FEDERAL government made Canada's economy BETTER, WORSE, or HAVEN'T THEY MADE MUCH DIFFERENCE either way?" (cpsh3), "Is that MUCH better or SOMEWHAT better?" (cpsh3a), "Is that MUCH worse or SOMEWHAT worse?" (cpsh3b) in 1993; "Have the policies of the FEDERAL government made Canada's economy better or haven't they made much difference?" (cpsmla), "Have the policies of the FEDERAL government made Canada's economy worse, or haven't they made much difference?" (cpsmlb) in 2004. Coding: Same as in Retrospective evaluation of personal financial situation in 1993; "better" = 1, "haven't made much difference," "don't know" and refused = 2, "worse" = 3, and all the respondents who are not answered this question are coded 2 in order to maintain the number of observations. Society better with the same social value (code: mbsg2 in 1993; mbsa9 in 2004) Question: (Do you strongly agree, agree, disagree, or strongly disagree?) "Society would be better off if we all had similar values and ideals" in 1993; "This country would have many fewer problems if there was more emphasis on traditional family values" in 2004. Coding: "Strongly disagree" = 1, "Disagree" = 2, "No opinion," "not sure" = 3, "Agree" = 4 and "Strongly agree" = 5. Few people respect traditional value (code: mbsal7, only available in 1993) Question: "People today don't have enough respect for traditional values" (Do you strongly agree, agree, disagree, or strongly disagree) Coding: Same as in Society better with the same social value. RCMP should wear the same hat (not turban) (code: mbse7,.only available in 1993) Question: "(In this section we are interested in your views about various aspects of Canadian society. Please indicate which answer best reflects your opinion) 'Members of the RCMP: 1. Should have the right to wear a turban for religious reasons, 2. Should all wear the same hat regardless of their religion'" Coding: "Right to wear a turban" = 1, "Undecided" = 2 and "Wear the same hat" = 3. Newer lifestyles break society (code: mbsa7 only available in 2004) Question: "Newer lifestyles are contributing to the breakdown of our society." Coding: Same as in Society better with the same social value. Economic anxiety scale Recoding calculation: Scores on Retrospective evaluation of personal financial situation, Prospective evaluation of personal financial situation, Retrospective evaluation of provincial economy (only in 1993), Retrospective evaluation of national economy and Prospective evaluation of national economy are summed, and 5 (or 4 in 2004) is subtracted and divided by 18 (or 8 in 2004), so that the score ranges from 0 to 1. 113 Government's influence on economy scale Recoding calculation: Similarly, scores on Federal government made respondents worse off, Federal government made respondents' provincial economy worse (only in 1993) and Federal government made Canadian economy worse are summed, and 3 (or 2 in 2004) is subtracted and divided by 10 (or 4 in 2004), so that the minimum score would be 0, and maximum would be 1. Common social/cultural values scale Recoding calculation: Similarly, scores on Society better with the same social value, Few people respect traditional value (in 1993), RCMP should wear the same hat (in 1993) and Newer lifestyles break society (in 2004) are summed, and 3 (or 2 in 2004) is subtracted and divided by 10 (or 8), so that the minimum score would be 0, and maximum would be 1. A.2 Australia Opposition to immigration (variable code: G8 for 1996, F6 for 2001) Question: "Do you think the number of immigrants allowed into Australia nowadays should be reduced or increased?" Coding: "Increased a lot," "increase a little," Remain about the same as it is," = 0, "Reduced a little" = 1, "Reduced a lot" = 2. Tertiary educated (variable code: H2 in 1996 and 2001) Question: "In all, how many years of tertiary study have you completed since you left secondary school? If your tertiary study was part-time, give the number of years of equivalent full-time study." Coding: More than one year of tertiary education = 1, none = 0 Male (code: Jl in 1996, II in 2001) "What is your sex?" Coding: Male = 1, female = 0. Age (code: J2 in 1996,12 in 2001) Question: "When were you born? Just the year will do." Coding calculation: 1996 or 2001 - (answered year of birth). Used absolute age. "Don't know" answer and refused are treated as missing. Unemployed (code: h4 in 1996 and 2001) Question: "Now some questions about the work you are doing now. Last week, what were you doing?" Coding: "Unemployed - FT work," "unemployed - PT work" = 1, "working FT for pay," "working PT for pay," "retired (from paid work)," "at school or uni," "a full-time school or university student," "keeping house" and "other," = 0 , Working class (code: J7 in 1996,115 in 2001) Question: "Which social class would you say you belong to?" Coding: "Working class" = 1, "upper class," "middle class" and "none" = 0 Household income category (code: J12 in 1996,117 in 2001) Question: "What is the gross annual income, before tax or other deductions, for you and your family living with you from all sources? Please include any pensions and allowances, and income from interest or dividends." Coding for J12 (1996): less than $3,0000 = 1, $3,000 - $8,000 = 2, $8,000 - $20,000 =3, $20,001-$30,000 =4, $30,001-$40,000 =5, $40,001-$50,000 =6, $50,001-$60,000 =7, $60,001-$70,000 =8, more than 114 $70,001 = 9, and refused are treated as missing values. Coding for 117 (2001): $less than $10,000 = 1, $ 10,001-$20,000 = 2, $20,001 -$30,000 =3, $30,001-$40,000 =4, $40,001-$50,000 =5, $50,001-$60,000 =6, $60,001-$70,000 =7, $70,001-$80,000 = 8, $80,001-$90,000 = 9, $90,001-$100,000 = 10, more than $100,000 = 11, and refused are treated as missing values. Native-born (code: J30WN in 1996 and 130WN in 2001) Question: "Where were you born?" Coding: Australia = 1, other countries, "other" and "other, unspecific" = 0 ESC-born (code:J30WN in 1996,130WN in 2001) Question: "Where were you born?" Coding: New Zealand, United Kingdom, Republic of Ireland, and Northern America = 1, "Australia," "Italy," "Germany," "Greece," "Malta," "Netherlands," "Poland," "(former) Yugoslavia," "Vietnam," "Oceania," "Southern Europe," "Western Europe," "Northern Europe," "Eastern Europe," "The USSR and Baltic States," "The Middle East and North Africa," "Southeast Asia," "Northeast Asia," "Southern Asia," "South and Central America and Caribbean," "Africa excluding North Africa," "China" and "other, unspecific" = 0. Christian (code: J5 in 1996, 15 in 2001) Question: "What is your religion or faith?" Coding: "Roman Catholic," "Catholic-Roman," "Anglican," "Church of England," "Uniting Church/ Methodist," "Orthodox," and "Presbyterian' - 1, "other" and "no religion" = 0. Religious attendance (code: J6 in 1996 and 2001) Question: "Apart from weddings, funerals and baptisms, about how often do you attend religious services?" Coding: "Never," = 0 "less than once a year" = 1, "at least once a year" = 2, "several times a year" = 3, "at least once a month" = 4, "at least once a week" = 5. Unemployment rate (by electoral district) Table 34b of "Ranked by unemployment rate" in Kopras (2000) in 1996: Table 33b of "Ranked by Unemployment Rate" in Kopras (2003) in 2001. Rate of immigrant population (by electoral district) Table 17b of "Ranked by Proportion of Persons Born Overseas" in Kopras (2000) in 1996; Table 17b of "Ranked by Proportion of Persons Born Overseas" in Kopras (2003) in 2001. Rate of the tertiary educated (by electoral district) Table 46b of "Ranked by Proportion of Persons with Tertiary Education Qualifications" in Kopras (2000) in 1996; Table 44b of "Ranked by Proportion of Persons with Tertiary Education Qualifications" in Kopras (2003) in 2001. Retrospective evaluation of personal financial situation (variable code: D5hhold in 1996 and 2001) Question: "How does the financial situation of your household now compare with what it was 12 months ago?" Coding: "a lot better," = 1, "a little better" = 2, "about the same" = 3, "a little worse" = 4, and "a lot worse" = 5. Prospective evaluation of personal financial situation (code: D7hhold in 1996 and 2001) Question: "Compared to now, what do you think the financial situation of your household will be in 12 months time?" 115 Coding: Same as in Retrospective evaluation of personal financial situation. Retrospective evaluation of national economy (code: D5cntr in 1996 and 2001) Question: "How do you think the general economic situation in Australia now compares with what it was 12 months ago?" Coding: Same as in Retrospective evaluation of personal financial situation. Prospective evaluation of national economy (code: D7cntr in 1996 and 2001) Question: "What do you think the general economic situation in Australia as a whole will be in 12 months time?" in 2001. Coding: Same as in Retrospective evaluation of personal financial situation. Worry about family member's unemployment in future (code: D9 only in 2001) Question: "How worried are you that in the next 12 months you or someone else in your household might be out of work and looking for a job for any reason-very worried, somewhat worried, or not worried at all?" Coding: "not at all worried." = 1, "somewhat worried" = 2 and "very worried" = 3. How easy to get a job (code: D10 only in 2001) Question: "In your community these days, how easy is it for someone who is trying to find a job to get a good job at good wages-very easy, somewhat easy, somewhat hard, or very hard?" Coding: "very easy" = 1, "somewhat easy" = 2, "somewhat hard" = 3 "very hard" = 4. Job availability better in past or future (code: Dl 1 only in 2001) Question: "When it comes to the availability of good jobs for Australian workers, some say that the best years are behind us. Others say that the best years are yet to come. What do you think?" Coding: "best years definitely behind" = 1, "best years probably behind" = 2 "best years probably yet to come" = 3 and "best years definitely yet to come" = 4. Own standard of living since 1998 Fed election (code: D13p7 only in 2001) Question: "Thinking back to the Federal election in 1998, when John Howard won against Kim Beazley, would you say that since then the following have increased or fallen?" then indicates, "Your own standard of living." Coding: "increased a lot" = 1, "increased a little" = 2, "stayed the same" = 3, "fallen a little" = 4 and "fallen a lot" = 5. Federal government made respondents worse off (code: D6hhold in 1996 and 2001) Question: "Compared with 12 months ago, would you say that the Federal government's policies have had a good effect, a bad effect, or that they really have not made much difference to the financial situation of your household?" Coding: "A good effect" = 1, "not much difference" = 2, "a bad effect" = 3. Federal government made Australian economy worse (code: D6cntry in 1996 and 2001) Question: "What effect do you think they [the Federal government] have had on the general economic situation in Australia as a whole?" Coding: Same as in Federal government made respondents worse off. Federal government will make respondents worse off (code: D8hhold in 1996 and 2001) Question: "Do you think that, 12 months from now, the Federal government's policies will have had a good effect, a bad effect, or that they really will have not made much difference to the financial situation of your household?" Coding: Same as in Federal government made respondents worse off. 116 Federal government will make Australian economy worse (code: D8cntry in 1996 and 2001) Question: "What effect do you think they will have had on general economic situation in Australia as a whole?" Coding: Same as in Federal government made respondents worse off. Truly Australian-born in Australia (code: G1 bnaust in 1996; G6p 1 in 2001) Question: Four questions below start, "Some people say the following things are important for being truly Australian. Others say they are not important. How important do you think each thing is?" then indicates, "Being born in Australia." Coding: "not at all important" = 1, "not very important" = 2, "fairly important" = 3 and."very important" = 4. Truly Australian-live Australia most of life (code: G1 livaus in 1996; G6p3 in 2001) Question: (The same question as the above and) "living in Australia most of one's life." Coding: Same as in Truly Australian-born in Australia. Truly Australian-being Christian (code: G1 christ in 1996; G6p5 in 2001) Question: (The same question as the above and) "being Christian." Coding: Same as in Truly Australian-born in Australia. Truly Australian-speak English (code: Glspkeng in 1996, G6p4) Question: (The same question as the above and) "being able to speak English." Coding: Same as in Truly Australian-born in Australia. Economic anxiety scale Recoding calculation: Scores on Retrospective evaluation of personal financial situation, Prospective evaluation of personal financial situation, Retrospective evaluation of national economy and Prospective evaluation of national economy, Worry about family member's unemployment in future (only in 2001), How easy to get a job (2001), Job availability better in past or future (2001) and Own standard of living since 1998 Fed election (2001) are summed, and 4 (or 8 in 2001) is subtracted and divided by 16 (or 28 in 2001), so that the score ranges from 0 to 1. Government's influence on economy scale Recoding calculation: Scores on Federal government made respondents worse off, Federal government made Australian economy worse, Federal government will make respondents worse off, and Federal government will make Australian economy worse are summed, and 4 is subtracted and divided by 8, so that the minimum score would be 0, and maximum would be 1. Cultural essentialism scale Recoding calculation: Scores on Truly Australian-born in Australia, Truly Australian-live Australia most of life, Truly Australian-being Christian and Truly Australian-speak English are summed, and 4 is subtracted and divided by 12, so that the minimum score would be 0, and maximum would be 1. A.3 New Zealand Opposition to immigration (variable code: C14) Question: "Do you think the number of immigrants allowed into New Zealand nowadays should be - " Coding: "Increased a lot," "increase a little," Remain about the same as it is," = 0, "Reduced a little" = 1, "Reduced a lot" = 2. Tertiary educated (variable code: G6) 117 Question: "Which one of the following indicates your highest formal educational qualification? Please tick one box." Coding: "Incomplete university degree," "undergraduate university degree" and "postgraduate university degree" = 1, "incomplete primary/none^" "primary completed," "secondary without UE or 6th form certificate," "complete secondary," and "non degree professional trade or technical tertiary qualification." = 0 Male (code: Gl) Question: "Are you" Coding: Male = 1, female = 0. Age (code: G2) Question: "In what year were you born?" Coding calculation: 2002 - (answered year of birth). Used absolute age. "Don't know" answer and refused are treated as missing. Unemployed (code: G7) Question: "This question is about the work you are now doing. Which of the following best describes your present position? Please tick all that apply." Coding: "Unemployed," "laid off," and "looking for job" = 1, "working full-time for pay or other income (32 or more hours a week)," "working part-time or other income (less than 32 hours a week)," "retired," "temporarily or permanently disabled, unable to work," "at school, university, or other educational institution," "unpaid work outside the home" and "unpaid work within the home" = 0 Factory worker (code: Gl 1) Question: "Would you describe your present or last job (if you do not presently have one) as: Please tick only one box." Coding: "Factory work" = 1, "managerial," "professional," "semi-professional or technical," "clerical service or sales," "skilled manual or trade," "manual labouring" and "farming" = 0 Manual worker (code: Gl 1) Question: "Would you describe your present or last job (if you do not presently have one) as: Please tick only one box." Coding: "Manual labor" = 1, "managerial," "professional," "semi-professional or technical," "clerical service or sales," "skilled manual or trade," "factory work" and "farming" = 0 Working class (code: G16) Question: "Some people see themselves as middle class, and others as working class, while some don't see themselves as belonging to a class at all. Do you see yourself as: Please tick one box." Coding: "Working class" = 1, "middle class," "belong to no class" and "Don't know" = 0 Household income category (code: G22) Question: "What was your personal income before tax between 1 April 2001 and 31 March 2002? What about the total income before tax of all members of your household in the same year?" Coding: No income = 0, less than $14,900 = 1, $14,900 - $20,699 = 2, $20,700 - $32,399 = 3, $32,400-$51,099 = 4, $51,100-$75,699 = 5, $75,700-$ 101,099 = 6, more than $ 101,099 = 7. "don't know" are treated as missing values. NZ and European identity (code: G14a) Question: "Do you identify yourself as any of these? You can tick as many as apply." Coding: A NZ European = 1, "A NZ Maori," "pacific island" and "others" = 0 118 ) Maori identity (code: G14a) Question: "Do you identify yourself as any of these? You can tick as many as apply." Coding: A NZ Maori = 1, "A NZ European," "pacific island" and "others" = 0 Christian (code: G18) Question: "What is your religion, if you have one?" Coding: "Anglican," "Presbytarian," "Catholic," "Methodist," "Baptist," "Latter Day Saints," "Ratana," "Independent-fundamentalist church," and "other Christian" = 1, "No religion" and "other non Christian" "atheist" "Bahai," "Brethren," "Buddist," "Greek Orthodox," "Hare Krishna," "Hindu," "Humanist," "Jehovahs Witness," "Judaism," "Lutheran," "Muslim," "Paganism/ Spriitualism," "Pai Mari," "Quaker," "Ringatu," "Russian Orthodox," "Salvation Army," "Seventh Day Adventist," "Sikhism," "Sufism" and "Taoism" = 0. Religious attendance (code: G17) Question: "Apart from weddings, funerals, and baptisms, about how often if at all do you attend religious services these days? Please tick one box." Coding: "Never," = 0 "Once a year" = 1, "2-11 times a year" = 2, "Once a month" = 3, "2-3 times a month" = 4, "at least once a week" = 5. Unemployment rate (by electoral district) Unemployment rate in Table 14 in Statistics New Zealand (2002: 142-148). "People unemployed as a percentage of people in the labour force," and people in the labor force are "aged under 15 years and people aged over 64 years" Rate of Asian, Pacific and other ethnic group (by electoral district) Maori Ethnic Group, Pacific Peoples Ethnic Groups, Asian Ethnic Groups and Other Ethnic Groups in Table 5 in Statistics N ew Zealand (2002: 48-53). Those who "stated each ethnic group, whether as their only ethnic group or as one of several ethnic groups," and "a person reported more than one ethnic group ... have been counted once in each applicable group." Rate of the tertiary educated (by electoral district) Bachelor Degree and Higher Degree in Table 8 in Statistics New Zealand (2002: 80-85). Oppose technically skilled migrants (code: C15) Question: Four questions below start, "Thinking about the follovying groups that might want to immigrate to New Zealand, do you think they should be allowed to immigrate freely, should there be restrictions, or should they not be allowed to immigrate at all? Please tick one box in each row," and indicate "People with technical skills" Coding: "Immigrate freely and don't know" = 0, "restrict immigration" = 1, "Not allowed to immigrate" = 2. Oppose immigrants willing to do manual labor (CIS) Question: (The same question as the above) "People willing to do manual labour" Coding: Same as in Oppose technically skilled migrants. Oppose immigrants from Muslim countries (C15) Question: (The same question as the above) "People from Muslim countries" Coding: Same as in Oppose technically skilled migrants. Oppose immigrants from Asia (C15) Question: (The same question as the above) "People from Asia" Coding: Same as in Oppose technically skilled migrants. 119 Oppose economic types of immigrants Recoding calculation: Scores on Oppose technically skilled migrants and Oppose immigrants willing to do manual labor are summed. Oppose social/cultural types of immigrants Recoding calculation: Scores on Oppose immigrants from Muslim countries and Oppose immigrants from Asia are summed. 120 Appendix B: Descriptive Statistics B.l Canada Table B . l Descriptive statistics of variables used in models, Canada 1993 2004 Mean S.D. N .. Mean S.D. N Dependent variable Opposition to immigration (0 to 1) .60 .49 3775 .29 .45 4323 dependent variable Tertiary educated .27 .44 4853 .34 .47 4267 Male .50 .50 4860 .46 .50 4323 Age category 3.85 1.60 3758 4.41 1.71 4268 Quebec resident .27 .44 3775 .24 .43 4323 Unemployed .07 .26 3775 .04 .19 4255 Main income earner unemployed in past 1 year .14 .35 3775 - - -Semi- and unskilled labor .31 .46 3428 - - -Household income category 4.19 2.70 3497 4.81 3.01 3723 Native-born . .87 .34 4869 .86 .35 4301 ESC-born .04 .19 4869 .05 .22 4301 Christian .78 .41 3772 .76 .43 4236 Unemployment rate (UR) 10.61 5.40 3611 7.91 4.79 4453 Rate of immigrant population (RIP) 13.58 12.04 3611 14.77 14.30 4453 Rate of the university degree holders 8.94 6.91 3611 18.79 11.20 4453 U R x R I P 131.59 145.46 3611 103.82 122.20 4453 Economic anxiety .55 .14 3775 .49 .20 4323 Government's influence on economy .65 • 17 3775 .54 .16 4323 Common social/cultural value .62 .20 2171 .58 .29 1662 U R x economic anxiety 5.86 3.44 3611 3.85 2.94 4105 RIP x economic anxiety 7.37 7.05 3611 6.93 7.53 4105 RIP x common social/cultural value 8.13 8.01 2105 8.12 9.42 • 1633 B.2 Australia Table B.2 Descriptive statistics of variables used in models, Australia 1996 2001 Mean S.D. N Mean S.D. N Dependent variable Opposition to immigration (0 to 2) .97 .84 1775 .57 .80 1973 dependent variable Tertiary educated .70 .46 1172 .65 .48 1469 Male .48 .50 1796 .47 .50 1979 Age 46.48 16.65 1707 48.13 16.50 1887 Unemployed .04 .20 1717 .05 .21 1945 Working class .45 .50 1737 .45 .49 1939 Household income category 5.04 2.27 1674 5.11 3.11 1832 Native-bom .78 .42 1726 .76 .43 1978 ESC-born .10 .31 1726 .09 .29 1978 Christian .76 .43 1745 .69 .46 1960 Religious attendance 1.75. 1.83 1747 1.64 1.81 1965 Unemployment rate (UR) 9.40 2.99 1786 7.54 2.34 2000 Rate of immigrant population (RIP) 21.24 10.64 1786 21.45 10.63 2000 Rate of the tertiary educated 16.04 7.11 1796 18.89 8.74 2000 U R x R I P 196.68 128.75 1786 157.61 99.18 2000 Economic anxiety .53 .15 1638 .56 .19 1750 Government's influence on economy .52 .15 1658 .54 .25 1869 Cultural essentialism .63 .22 1706 .61 .23 1912 U R x economic anxiety 4.92 2.18 1627 4.27 2.09 1742 RIP x economic anxiety 11.01 6.32 1627 11.97 7.41 1742 RIP x common social/cultural value 13.07 8.05 1695 12.75 8.03 1902 122 B.3 New Zealand Table B.3 Descriptive statistics of variables used in models, New Zealand 2002 Mean •. S.D. -N Dependent variable Opposition to immigration (0 to 2) .73 .82 4690 Independent variable Tertiary educated .28 .45 5546 Male .47 .50 5781 Age 48.84 16.27 5728 Unemployed .03 .17 4656 Factory worker .04 .19 4413 Manual labor .05 .22 4413 Working class .21 .40 4518 Household income category 3.76 1.98 3790 N Z and European identity .83 .37 5320 Maori identity .11 .31 5320 Christian .73 .44 4244 Religious attendance 1.30 1.80 4537 Maori District .13 .34 5533 Unemployment rate (UR) 8.38 3.39 5533 Rate of Asian + Pacific + Other ethnic group (RAPO) 10.73 10.59 5533 Rate of the tertiary educated 7.21 5.30 5533 U R x R A P O 93.27 110.69 5533 Oppose economic types of immigrants 1.57 1.00 4603 Oppose immigrants with technical skills .58 .57 4628 Oppose immigrants willing to do manual labor .99 .63 4642 Oppose social/cultural types of immigrants 2.07 1.00 4618 Oppose immigrants from Muslim countries 1.11 .58 4648 Oppose immigrants from Asia .96 50 4626 U R x oppose economic types of immigrants 13.70 12.08 4419 R A P O x oppose economic types of immigrants 16.46 22.05 4419 R A P O x oppose cultural types of immigrants 21.44 25.42 4434 123 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0092730/manifest

Comment

Related Items