What's Driving the Public? A Cross-Country Analysis of Political Attitudes, Human Values and Political Articulation
by Joakim Kulin and Alexander Seymer
Umea University; University Salzburg
Sociological Research Online, 19 (1) 14
Received: --- Accepted: 12 Jan 2014 Published: 28 Feb 2014
This study addresses how political attitudes are shaped across national contexts. It does so by investigating the influence of nation-specific political articulation and framing on the relationship between human values and political attitudes. Based on the literature, two attitude dimensions can be identified. First, the socioeconomic dimension captures the tension between economic equality and equity (rewarding achievements and effort). Second, the sociocultural dimension captures the tension between individual/civil liberties and traditions/conservative norms. Very few comparative studies systematically investigate the influence of a coherent structure of more basic and abstract motivations (values) on political attitudes. We fill this gap by examining the influence of basic human values on the socioeconomic and sociocultural attitude dimensions across national contexts. To investigate the impact of human values and political attitudes, individual-level data from the European Social Survey (ESS 2008) from 2008 are analyzed using multi-group structural equation modeling (MGSEM). Moreover, we also explore political discourse as a key contextual factor at the country level modifying the relationships between values and attitudes. Specifically, we use data from the Comparative Manifesto Project to investigate the moderating influence of political articulation, i.e., the articulation of socioeconomic and sociocultural issues in political party manifestos, on the relationship between values and political attitudes across countries. Results indicate substantial cross-national variation in the link between values and sociopolitical attitudes, and that this variation can be partly explained by the articulation of sociopolitical issues.
Keywords: Political Attitudes, Basic Human Values, Political Articulation, European Social Survey, Comparative Manifesto Project Database
Introduction1.1 For anyone attempting to understand political development in democratic societies, studying public opinion is crucial. It is particularly important to identify the determinants of political attitudes, as well as the circumstances under which they play key roles. In this study, we examine the influence of the political discourse on political attitude formation, by investigating how the articulation of sociopolitical issues by the political elite impinges on the relationship between human values and political attitudes among citizens.
1.2 Whether seeking to implement, maintain, or roll back social policy or civil rights legislation, policymakers and politicians are constrained by public opinion. One reason why sociopolitical attitudes are quite resilient is that they are grounded in deeper and more fundamental values (e.g., Feldman 1988; Feldman and Zaller 1992). Despite the theoretical relevance of values, their role has received relatively little attention in cross-national research into political attitudes (Feldman 2003). One possible reason for this neglect is that the value basis of political attitudes is often either implicitly or explicitly assumed a priori; for instance, it is assumed that people who support redistribution do so based on egalitarian values, and that people who support same-sex marriage do so based on liberal values. However, the relationship between values and political attitudes does not exist in a vacuum. Political debates differ across countries: in some countries, political discussion is dominated by socioeconomic issues such as inequality and redistribution, whereas in others it is more concerned with socioculturally oriented issues such as immigration and minority rights. This is reflected, for instance, in substantial cross-national variation in the articulation of sociopolitical issues in party manifestos (see, e.g., Klingemann et al. 2006; for a substantive application, see Kumlin and Svallfors 2007). Previous research demonstrates that political elites often evoke fundamental values when promoting political issues (Sears 1980; Sears and Funk 1991) and that citizens are disposed to take their cues from political elites (Jacoby 2000; Schneider and Jacoby 2005). Therefore, the public's ability to connect specific political issues to underlying values very likely depends on the articulation of political issues. Indeed, while values may exert a general influence on attitudes toward core political issues, the strength of this influence should depend on the extent to which political elites articulate these issues and their value underpinnings. Since political elites often evoke fundamental values to promote political issues, we posit that in countries where certain political issues are more salient in the political debate, attitudes toward these issues should be more clearly value driven.
1.3 We cross-nationally examine the relationship between basic value orientations and sociopolitical attitudes, and explore the moderating influence of political articulation on the value-attitude relationship across countries. More specifically, we use data from the 2008 European Social Survey and employ multi-group structural equation modeling (MGSEM) to estimate the effects of values on political attitudes in several Western European countries. To investigate the role of political articulation as a moderator of value-attitude relationships across countries, the Comparative Manifestos Project dataset is used. We focus on the effects of four higher-order value types from the 'theory of basic human values' (Schwartz 1992) on socioeconomic and sociocultural political attitudes, and consider the extent to which articulation of socioeconomic and sociocultural issues by the political elite can explain cross-country differences in the effects.
Theoretical framework2.1 Economic self-interest has long been viewed as the main driver of political attitudes. For instance, those espousing political economy perspectives often claim that self-interest is a key determinant of attitudes toward inequality and redistributive policies (Meltzer and Richard 1981; see also Iversen and Soskice 2001; Cusack et al. 2006). However, considerable research in the fields of political sociology and political science now suggests that interest-oriented perspectives have rather limited explanatory power to predict attitudes compared with that of more fundamental predispositions such as values and norms (see, e.g., Sears and Funk 1990; Svallfors 1996, 2007; Rothstein 1998; Ullrich 2002; Mau 2003; Kumlin 2004; Brooks and Manza 2007). This is also in line with findings in experimental economics and evolutionary biology, which stress the importance of other-regarding motives such as beliefs about justice and reciprocity (Bowles and Gintis 2000; Fehr and Fischbacher 2002; Gintis et al. 2004)
2.2 Previous US research demonstrates that people form their views on various political issues based partly on personal and deeply held values (see e.g., Feldman 1988; Feldman and Zaller 1992; Feldman and Steenbergen 2001; for an overview, see Feldman 2003). Although cross-national research into the impact of values on political attitudes is scarce, recent comparative studies demonstrate that individuals holding egalitarian values are also more likely to hold typical left-wing socioeconomic attitudes supporting, for example, redistribution and government responsibility for social protection (Kulin 2011; Kulin and Svallfors 2013). Furthermore, research also demonstrates that values are related to more socioculturally oriented attitudes, in that conservative values are related to negative attitudes toward immigration and related policies (Davidov et al. 2008a; Davidov and Meuleman 2012).
2.3 Whereas values can play an important role in attitude formation, the extent to which they do must be understood by taking into account the moderating influence of contextual factors (Sagiv and Schwartz 1995: 447). For instance, the impact of values on political attitudes is likely to be stronger if related policies have important consequences for the sociopolitical issues they are intended to address. Feldman argues that:
The hypothesized relationship between core beliefs and values and public opinion rests on such a simple mechanism. Political evaluations may be based, in part, on the extent to which policies and actions are consistent or inconsistent with certain important beliefs and values. . . .To some extent, policies and actions are simply judged right or wrong because of their implications for deeply held values. (Feldman 1988: 418)
2.4 Results from previous cross-national research support these claims. For instance, Kulin (2011) demonstrates that the impact of self-transcendence (e.g., egalitarian) values on attitudes toward redistribution is stronger in countries where redistributive institutions are more generous, i.e., more effective in equalizing incomes. This suggests that when people perceive policies as more effective in addressing the issue they are meant to tackle, the general relationship between attitudes toward these issues and their underlying values becomes stronger. This is in line with a growing body of literature identifying more general policy feedback effects of political institutions on public opinion (Pierson 1993; Mettler and Soss 2004; Svallfors 2007). Moreover, the strength of the relationship between values and attitudes is likely to increase with the level of political sophistication (Zaller 1992). Recent evidence confirms this claim, as the impact of egalitarian values on redistributive attitudes is stronger in members of higher social classes, who generally display higher levels of political sophistication (Kulin and Svallfors 2013).
2.5 The argument that values influence political attitudes is based on the assumption that people make cognitive connections between policies and more fundamental principles and beliefs such as basic values. Furthermore, individuals likely perceive similar political issues, and the policies intended to address them, very differently depending on the political climate and political debate in their particular countries. We believe that the political discourse may have crucial implications also for the extent to which people draw on more fundamental values when forming their political attitudes. Hence, we expect the articulation of political issues to moderate the relationship between values and political attitudes cross-nationally.
2.6 Among the first to introduce the idea of political elite framing was Converse (1964), whose seminal work proposed that public opinion is structured by belief systems organized and presented to the mass public by the political elite. It has been difficult to identify the degree of consistency in public opinion required to corroborate the idea that elite-imposed ideologies are upheld by the public. In fact, public opinion, seen as a complex system of attitudes, rarely mirrors the rhetoric of political elites. Still, the influence of elite framing on public opinion has been well documented in relation to a wide range of specific issues and across many contexts (Jacoby 2000; Schneider and Jacoby 2005). According to Jacoby (2000), framing effects appear mainly indirectly rather than as immediate responses to an articulated issue, triggering different sets of influences on citizens' attitudes. This is in line with Feldman (1988: 417), who argues that 'political attitudes and opinions are not simply accepted on the basis of their packaging by elites, but in a general way are consistent with certain core beliefs and values'. Whereas the relationship between values and attitudes is related to the perceived efficiency of policies and to the level of political sophistication, 'the impact of people's value predispositions always depends on whether citizens possess the contextual information needed to translate their values into support for particular policies and candidates' (Zaller 1992: 25). Hence, citizens do not connect their political attitudes to applicable values independently of their political context. Results from experimental studies confirm this, indicating that values can have important consequences for attitudes and behavior, but that they often must be cognitively activated by external factors (Verplanken and Holland 2002). Previous cross-national survey research has found a direct relationship between political articulation and sociopolitical attitudes. Kumlin and Svallfors (2007) demonstrate that class differences in socioeconomic attitudes are related to cross-country variation in the articulation of socioeconomic issues. Relatively little is known, however, about the values underlying sociopolitical attitudes cross-nationally, or about the influence of political articulation on this relationship across countries. Given cross-national differences in the political articulation of sociopolitical issues (see also Budge et al. 2001), we believe that there are good reasons to expect differences in the value-attitude relationship across countries.
2.7 The idea that political articulation influences the value-attitude relationship is consistent with the 'symbolic politics' perspective, since it highlights the role of political elite framing in structuring political attitudes, stressing that such attitudes 'are formed mainly in congruence with longstanding values about society and the polity' (Sears et al. 1980: 671). Regarding the influence of politics, Sears and Funk (1991: 14) argue that the principal goal of politicians is to 'code political symbols in terms of what will evoke widespread and supportive predispositions in the citizenry'. According to this perspective, politicians aim to articulate sociopolitical issues and symbols in a way that elicits values that strengthen electoral support. The extent to which values influence political attitudes should therefore, at least partly, depend on the articulation of political issues and symbols associated with the attitudinal object. While the literature frequently theorizes about these relationships, empirical studies of them across varied, yet comparable, political contexts are rare. One possible reason for this lacuna is that values and attitudes have been treated as indistinguishable concepts (Rohan 2000). Moreover, the longstanding absence of an agreed-on theory of values and a validated measurement instrument (Davidov et al. 2008) has resulted in a lack of empirical cross-national research into the link between values and political attitudes. We contribute to this emerging research field by conducting cross-national research that (i) makes a clear theoretical distinction between values and attitudes, and (ii) uses an established theory and employs a validated measurement instrument to measure values.
Distinctions and clarifications
2.8 Attitudes are psychological tendencies to evaluate a certain object or situation in a favorable or unfavorable manner (see also Rokeach 1968: 550; Eagly and Chaiken 1993: 1; Ajzen 2001: 28). Values, on the other hand, are abstract and trans-situational goals that motivate and guide attitudes and behavior. Accordingly, values do not refer to specific objects, actions, or situations, but instead to the more abstract criteria used to evaluate them. Whereas a person can have as many attitudes as there are objects in the world (Ajzen and Fishbein 1975), values constitute a limited set of trans-situational goals motivated through the response to three universal requirements of human existence: 'needs of individuals as biological organisms, requisites of coordinated social interaction, and survival and welfare needs of groups' (Schwartz 1992: 3). Furthermore, whereas attitudes are more volatile, values are more stable and occupy a more central position in the personality of an individual (Hitlin 2003; Hitlin and Piliavin 2004).
|Figure 1. Circular continuum representing the structure of individual values and higher-order value types.|
2.9 A theory of human values that has received increasing attention is the theory of basic human values (Schwartz 1992). Starting from a holistic desire to give a complete account of the structure of values, the scale itself was empirically derived by a smallest-space analysis revealing a quasi-circular structure of ten distinctive values. As shown in Figure 1, adjacent values are more closely related than are distant ones, while opposite values in the circular structure conflict with each other. A complete description of the ten values and their motivational emphases can be found in Table 1. To our knowledge, the theory of basic human values is the only theory that gives a complete account of the structure of human values in a holistic and universal value space. Moreover, an intriguing property of the scale is the structure of four higher-order value types along two orthogonal dimensions - i.e., self-transcendence versus self-enhancement and conservation versus openness to change - with each of the four types comprising two or three of the ten values. The structure of the ten values and the orthogonal relationships between the higher-order value types have been repeatedly validated across many cultures (e.g., Schwartz et al. 2001; Schwartz and Boehnke 2004; Davidov et al. 2008).
|Table 1. Value types defined in terms of their motivational goals and the values that represent them.|
2.10 The self-enhancement higher-order value type refers to motivational goals promoting individual success, achievement, social recognition, and power (i.e., power and achievement values). The openness to change higher-order value type refers to motivational goals promoting individual freedom, independence of thought, and novelty (i.e., self-direction and stimulation). The self-transcendence higher-order value type refers to motivational goals promoting self-transcendence in favor of a collective focus on equality and helping others (i.e., universalism and benevolence), thereby opposing self-enhancement values. Finally, the conservation higher-order value type captures motivational goals promoting social order and security while conforming to conservative norms and traditions (i.e., conformity, tradition, and security), thereby opposing openness to change values in the circular schema.
2.11 Turning to political attitudes, no single dimension adequately reflects the whole spectrum of political attitudes (Kinder 1998). In surveying the literature, at least two key political dimensions can be distinguished in relation to which attitudes may form. While the terminology varies - for example, economic/non-economic liberalism (Lipset 1979), equality/freedom (Rokeach 1973), classical liberalism, and economic egalitarianism (Schwartz 1994) - the inherent meanings are similar. Accordingly, we study political attitudes along two sociopolitical dimensions: a socioeconomic and a sociocultural dimension (cf. Svallfors 2005). The socioeconomic dimension focuses on the tension between economic equality and equity (i.e., rewarding achievements and effort). This dimension encompasses attitudes toward distributive arrangements in society (e.g., taxation and redistribution), inequality, and poverty. The sociocultural dimension focuses on the tension between individual and civil liberties versus upholding traditions and conservative norms. This dimension encompasses attitudes toward, for instance, same-sex marriage and adoption, immigration, and submission to authorities. Meanwhile, as Schwartz (1994) points out, the two higher-order value dimensions - i.e., self-transcendence versus self-enhancement and openness to change versus conservation - correspond to the abovementioned political dimensions. Economic egalitarianism 'refers to whether the government devotes itself more to promoting equality by redistributing resources or to protecting citizens' ability to retain the wealth they generate in order to foster economic growth and efficiency', whereas classical liberalism 'refers to whether the government should devote more to guarding and cultivating individual freedoms and civil rights or to protecting the societal status quo by controlling deviance from within or enemies from without' (Schwartz 1994: 39 - 40).
2.12 Based on these accounts, one can formulate more specific expectations as to the general relationship between the four higher-order value types and socioeconomic/sociocultural political attitudes. First, self-transcendence and self-enhancement values should be particularly relevant to socioeconomic attitudes, whereby self-transcendence values should positively affect egalitarian socioeconomic attitudes due to their focus on transcending the self and helping others. Likewise, self-enhancement values should have a negative effect due to their focus on promoting individual power and success. Second, conservation and openness to change values should be more closely related to sociocultural attitudes, whereby conservation values should negatively affect liberal sociocultural attitudes due to their focus on upholding traditions and conservative norms. Openness to change values, on the other hand, should have a positive effect due to their focus on promoting individual autonomy and stimulation. Research into values and voting corroborates these claims. When political conflict concerns socioeconomic issues such as distributing material resources, the most important values are self-transcendence and self-enhancement, whereas when politics concern traditional morality versus modernization (i.e., when sociocultural issues are more salient), the most important values are conservation and openness to change (Barnea 2003; Barnea and Schwartz 1998).
2.13 Previous research also shows that people make tradeoffs between more fundamental values and principles when forming their preferences in the arena of politics. According to Tetlock (1986), people make tradeoffs between important yet conflicting values when making up their minds about political issues. Moreover, 'complex trade-off forms of reasoning are most likely to emerge in policy domains that activate important, and approximately equally important, conflicting values' (Tetlock 1986: 826). While value conflicts and tradeoff-type reasoning are more likely to occur in relation to single political dimensions eliciting multiple and opposing values, some scholars argue that value conflicts can also occur between broader political dimensions, for example, between freedom and equality (see, e.g., Lipset 1979; McClosky and Zaller 1984; Feldman and Zaller 1992). This suggests that tradeoffs may also appear between broader value and attitude dimensions. Yet, the question of how people respond to these conflicts and make tradeoffs, particularly across national political contexts, has largely gone unanswered.
2.14 However, Piurko et al. (2011) have recently studied the values underlying subjective left-right orientation, though it is unclear exactly what political dimensions the unidimensional left-right scale reflects across different national political contexts. The authors focused on individual value types rather than on the higher-order dimensions, which further obscures the value bases of broader political traditions (e.g., freedom and equality) and their related attitude dimensions across countries. Furthermore, they did not link the emerging patterns in the relationship between values and left-right orientation to key contextual factors assumed to moderate this relationship cross-nationally. We apply a different approach by relating higher-order value structures to two central attitude dimensions - socioeconomic and sociocultural attitudes - corresponding to the political traditions discussed above, which we believe cannot be collapsed into a single measure (e.g., left-right). In addition, we relate the relationship between values and attitudes to the political context by introducing political articulation as a potential moderator of the value-attitude relationship across countries.
2.15 Based on the discussion above, we formulate more specific hypotheses about the relationship between values and attitudes across countries, and about the moderating influence of political articulation. In particular, we expect the relative importance of the two higher-order value type dimensions - self-transcendence versus self-enhancement values, and conservation versus openness to change values - for socioeconomic and sociocultural attitudes, respectively, to differ across national political contexts. While the impact of value types/dimensions is likely universal in terms of the directions of their effects on socioeconomic and sociocultural attitudes, we would expect the strength of the effects to display cross-country differences. If socioeconomic issues are more prominently articulated in a given country, its people are more likely to see these issues as more pressing and consequently to draw on values relevant to these attitudes. Therefore, in countries where socioeconomic issues are more clearly articulated, self-transcendence and self-enhancement values should be more important for socioeconomic attitudes. In countries where sociocultural issues are more clearly articulated, they are more likely to be more strongly driven by conservation and openness to change values. Meanwhile, in some countries, a more diverse spectrum of political issues may be articulated, which leads us to expect the two higher-order value dimensions to exert a moderate and approximately equally strong influence on socioeconomic and sociocultural attitudes in these countries.
2.16 We see at least two empirical manifestations that can confirm our arguments. First, to the extent that our argument is valid, we expect that in countries where socioeconomic attitudes are more strongly driven by relevant values, sociocultural attitudes will be less value driven, and vice versa (H1). Hence, we expect a tradeoff relationship. We start by internally validating our argument by using individual-level data on values and political attitudes, focusing on the relative strength of the relationship between self-transcendence versus self-enhancement values and socioeconomic attitudes as well as conservation versus openness values and sociocultural attitudes. Second, a more crucial empirical manifestation would be a clear relationship between the relative importance of the higher-order value types for political attitudes (i.e., socioeconomic and sociocultural) and political articulation. In countries where socioeconomic issues are more salient in politics, the link between self-transcendence versus self-enhancement values and socioeconomic attitudes will be stronger. Likewise, in countries where sociocultural issues are more clearly articulated, the impact of conservation versus openness values on sociocultural attitudes will be stronger (H2).
Data and methods3.1 To investigate the relationship between higher-order value dimensions and attitude domains, individual-level data from the fourth wave of the European Social Survey (ESS) were employed. The fourth wave includes socioeconomic indicators in the rotating module on welfare state attitudes not covered in other rounds. Furthermore, the human value scale (Schwartz 1992) is a permanent part of the survey. The data on political articulation originates from the Manifesto Database Project (https://manifestoproject.wzb.eu/). We study these relationships in ten European countries: Switzerland (CH, n=1696), Germany (DE, n=2659), Denmark (DK, n=1549), Finland (FI, n=2143), France (FR, n=2031), the United Kingdom (GB, n=2278), Ireland (IE, n=1734), the Netherlands (NL, n=1722), Norway (NO, n=1534), and Sweden (SE, n=1751). For group comparisons, the samples were weighted according to ESS guidelines with pairwise deletion applied.
3.2 In this study, we focus on how political articulation moderates the relationship between values and sociopolitical attitudes across national contexts. Since we believe that these factors are relatively stable and that the influence of political articulation occurs over longer time periods, we focus on older, mature, and relatively stable political contexts. Hence, our country selection is based on a comparable-cases strategy, the aim of which is to compare politically and historically similar political contexts (e.g., Western countries with similar electoral systems and relatively stable political traditions) rather than to maximize the number of countries.
3.3 Both attitude domains are measured as latent variables with four indicators each. The sociocultural domain consists of attitudes toward gay people, immigrants, authority, and gender equality. With respect to gay people, respondents were asked to specify, using a five-point Likert scale, whether gay men and lesbians should be free to live their lives as they see fit. Similar scales were applied for the authority and gender equality items, asking whether schools should teach children to obey authority and whether women should reduce paid work in favor of family responsibilities. The final item about immigrants, measured using an 11-point scale, asked whether immigrants make the country a better or worse place to live. Supporting gender equality and rights of gay people, as well as positive attitudes toward immigrants and the rejection of authority, would reflect the most progressive and open (i.e., left-wing) attitudes; conversely, low scores on all these indicators reflect more traditional (i.e., right-wing) attitudes.
3.4 The socioeconomic attitude domain is slightly more complex in composition and consists of four items. The first two are five-point Likert scales concerning whether or not the government should reduce income differences, and whether little inequality in the standard of living is necessary for a society to be just. We also use composite scores indicating a general preference for government responsibility for social protection in areas such as pensions, parental leave benefits, and unemployment insurance. The reason for calculating composite scores is to prevent this category of socioeconomic attitudes from having too much weight in the final latent construct. The items are summed up in an index reflecting weak preferences for government responsibility in these areas at the lower end, and strong preferences for government responsibility at the upper. In other words, somebody with maximum scores on all six dimensions thinks that the government should act as the main welfare provider, while somebody with minimum scores sees little need for welfare state provision. Finally, the last item captures taxation principles, or the degree of progressiveness in the taxation system. As taxation is the main source of state finance, we capture it on the socioeconomic dimension. The respondents were asked where their preferred taxation system was situated on a scale ranging from a flat-rate model at one end to a progressive taxation model at the other.
3.5 Finally, value types are measured in line with Schwartz's (1992) value theory. We focus on the higher-order value structures, which consist of two to three values, measured by two or three items each in the ESS. In line with research employing the higher-order value types (Davidov 2008, 2010), we operationalize the types by loading each value item directly on the latent construct. Consequently, each value type was measured by at least four items. Conservation is measured by six items capturing conformity, tradition, and security values. Self-enhancement is measured by four items capturing power and achievement values. Self-transcendence is measured by five items, universalism being captured by three items and benevolence by two. Finally, openness to change is measured by four items capturing self-direction and stimulation.
|Figure 2. A generalized schematic of the estimated MGSEMs.|
3.6 With the multiple-item measures, the effects of the higher-order value types on each attitude domain are estimated using two latent constructs along a multi-group structural equation model (MGSEM or SEM, if not applied to multiple samples). Figure 2 shows a generalized schematic of all models. The multiple manifest measures, represented by rectangles, capture the two latent constructs in two separate measurement models. The structural model is a simple regression model with the value structure as explanatory variable and the political attitude as response variable (or more details on model specification and fit, see the Appendix).
3.7 To test whether the relationship between values and political attitudes is contingent on the political context, we use data on the political articulation of socioeconomic and sociocultural issues, and link the political articulation to the value-attitude relationship across countries. As an indicator of political articulation, we use data from the Comparative Manifesto Project (CMP) dataset (see Klingemann et al. 2006). In the CMP data, the program contents of election manifestos have been indexed for most Western countries and for all elections since 1945. The dataset is a classification of the smallest significant units ('quasi-sentences') of each election manifesto on a wide range of topics distributed among several policy areas. The measure of political articulation used here is one that indexes the share of program contents among all political parties dealing with socioeconomic and sociocultural issues between 1980 and 2003. See the Appendix for a description of the construction of the measures.
Results4.1 We start our analyses from two sets of hypotheses. First, according to our tradeoff hypothesis (H1), we expect that in countries where self-transcendence (ST) and self-enhancement (SE) values have a greater impact on socioeconomic attitudes, the impact of conservation (CO) and openness to change (OC) values on sociocultural values will be weaker, and vice versa. Second, according to our articulation hypothesis (H2), we expect this balance to at least partly result from the extent to which socioeconomic and sociocultural issues are politically articulated. Hence, in countries where socioeconomic issues are more salient in the political debate, self-transcendence and self-enhancement values will more strongly affect socioeconomic attitudes, at least relatively speaking. Conversely, in countries where sociocultural issues are more clearly articulated, the impact of conservation and openness to change values on sociocultural attitudes will be stronger.
|Table 2. Country-level effect sizes, standard errors (S.E.), and explained variances (i.e., squared multiple correlations, SMCs) for the effect of ST and SE values on socioeconomic attitudes and of CO and OC values on sociocultural attitudes.|
4.2 Before turning to the results concerning our more specific hypotheses, we examine the general relationship between higher-order value types and the two sociopolitical attitude dimensions. In Table 2, the unstandardized effects, standard errors, and explained variances (i.e., squared multiple correlations, SMC) are presented for each model. In nearly all countries, we find that the higher-order value types have the expected effects on political attitudes. Self-enhancement values negatively affect socioeconomic attitudes in all countries except Switzerland, France, and the UK (where the effects are not significant), while self-transcendence values have positive and significant effects in all countries. Conservation values have a significant negative effect, and openness to change values have a significant positive effect on sociocultural attitudes in all countries (with the exception of openness to change values in Denmark). This means that people who think that everyone should be treated equally and who care for other people are more likely to hold left-wing socioeconomic attitudes (i.e., support redistribution and small income differences), while in most countries, people who think that it is important to acknowledge achievements and success are less likely to hold these attitudes. Turning to sociocultural attitudes, those who think that it is important to uphold traditions and the status quo are more likely to hold right-wing attitudes (e.g., oppose same-sex marriage and immigration), while those who think that it is important that people should be free to live life as they see fit and to lead stimulating lives are less likely to hold these attitudes. Consequently, our results are consistent with those of previous research into the impact of higher-order value types on more specific sociopolitical attitudes (see, e.g., Davidov et al. 2008a; Kulin and Svallfors 2011) and voting (Barnea and Schwartz 1998; Barnea 2003). We next investigate our more specific hypotheses about value-attitude links across countries and the role of political articulation.
Tradeoff in the value-attitude relationship (H1)
4.3 To test the hypothesized tradeoff relationship between the value bases of the two sociopolitical attitude dimensions, measures indicating the overall impact of the relevant values on the attitude dimensions were calculated using the explained variances. The mean SMCs, i.e., the explained variances for each model, were calculated for each country as (ST + SE)/2 and (CO + OC)/2. These measures indicate the average strength of the relationships between the higher-order value dimensions and the attitude dimensions to which they relate. In Figure 3, these relationships are plotted against each other, with the relationship between the value dimension self-transcendence versus self-enhancement and socioeconomic attitudes on the y-axis and the relationship between the value dimension conservation versus openness to change and sociocultural attitudes on the x-axis. Based on the plot, we identify a clear tradeoff (i.e., negative correlation) between the impact of conservation versus openness to change values on sociocultural attitudes, on the one hand, and the impact of self-transcendence versus self-enhancement values on socioeconomic attitudes, on the other (r = -0.75, p < 0.05; R2 = 0.56).
|Figure 3. The impacts (mean SMCs) of the two higher-order value dimensions (CO vs. OC and ST vs. SE) on sociocultural and socioeconomic attitudes, respectively.|
4.4 In line with our hypothesis, in countries such as Sweden and Norway where self-transcendence and self-enhancement values have a stronger impact on socioeconomic attitudes, conservation and openness to change values have a comparatively weaker impact on sociocultural attitudes. Similarly, in countries such as Switzerland and Germany where conservation and openness to change values have a comparatively stronger impact on sociocultural attitudes, self-transcendence and self-enhancement values have a relatively weaker impact on socioeconomic attitudes. This means that when one political attitude dimension is more strongly driven by deep value orientations, the other dimension is comparatively less value driven. Hence, there seems to be a tradeoff between the two sociopolitical attitude dimensions in terms of the extent to which they draw on underlying values. We next examine to what extent this tradeoff can be linked to political articulation across different national contexts.
The influence of political articulation on the value-attitude relationship across national contexts (H2)
4.5 Political articulation can be conceptualized and measured in several ways. One way is to focus on the absolute articulation of either socioeconomic or sociocultural issues. However, since we are interested in explaining cross-country differences in how values affect socioeconomic and sociocultural attitudes in relative terms, it is more suitable to also focus on the relative salience in the articulation of socioeconomic and sociocultural issues. Moreover, since we want to relate the relative strengths of the value-attitude relationships along two dimensions (i.e., self-transcendence and self-enhancement values with socioeconomic attitudes, and conservation and openness to change values with sociocultural attitudes), we construct a single measure reflecting the relative articulation of socioeconomic and sociocultural issues in the party manifestos of each country. Similarly, for the value-attitude relationships, we use the measures from the previous analysis (see Figure 3) and calculate the difference between the two sets of relationships. Hence, we calculate the mean SMCs for the impact of self-transcendence versus self-enhancement values on socioeconomic attitudes, and subtract the mean SMCs for the impact of conservation versus openness to change values on sociocultural attitudes, i.e., (ST + SE)/2 - (CO + OC)/2.
4.6 As seen in the scatterplot (Figure 4), the results indicate that political articulation is related to the value-attitude relationships in a relatively clear and consistent manner (r = 0.59, p = 0.07; R2 = 0.35). In countries where socioeconomic issues are more clearly articulated, the influence of self-transcendence and self-enhancement values on socioeconomic attitudes is relatively strong, whereas in countries where sociocultural issues are more salient, the influence of conservation and openness to change values is stronger. For instance, in Sweden and Finland, where the political articulation of socioeconomic issues is particularly salient, the impact of self-transcendence and self-enhancement values is comparatively strong, and in countries such as Switzerland and Germany, where political articulation is more concerned with sociocultural issues, the impact of conservation and openness to change values is stronger. Hence, the extent to which the sociopolitical attitude dimensions are driven by applicable values is highly dependent on which sociopolitical dimension is more clearly articulated.
|Figure 4. The balance between the value-attitude relationships (y-axis) and political articulation (x-axis).|
4.7 In these types of small-n comparisons, the overall pattern is more important than the statistical significance indicated by formal testing. This is due to the volatility of significance testing given the moderate deviations from the patterns of single cases. Of all the studied countries, Norway is the only one that can be considered an outlier, since it deviates from an otherwise clear relationship. This is due to the relatively weak articulation of socioeconomic issues in Norwegian party manifestos. Excluding Norway makes the correlation much stronger, with the explained variance rising from 35% to 63% (r = 0.79, p < 0.01; R2 = 0.63). An important question, which we cannot answer definitively here, is why Norway deviates from an otherwise very clear trend. On the one hand, the fact that politicians in Norway have not articulated socioeconomic issues as strongly as they do in, for example, Sweden and Finland, is conspicuous, not least considering Norway's comparatively generous socioeconomic policies (Scruggs 2006). On the other hand, the lack of articulation of socioeconomic issues in Norway could be seen as a consequence of the rather uncontroversial nature of spending in Norway, whose government and people have benefited greatly from North Sea oil revenues via state-owned oil companies and the Government Petroleum/Pension fund. Hence, socioeconomic issues may be as important in Norway as in, for instance, Sweden, though they are not as explicitly articulated. Meanwhile, socioeconomic politics in Norway may be characterized by consensus rather than conflict, which would result in less salient articulation. This could explain why Norway is an outlier.
Conclusions5.1 Our results clearly support the interpretation that the articulation of sociopolitical issues by the political elite influences political attitude formation, i.e., the extent to which values drive political attitudes. In essence, the results indicate that socioeconomic attitudes are more strongly driven by relevant underlying values in countries where socioeconomic issues are more clearly articulated in political manifestos, whereas sociocultural attitudes are more strongly driven by relevant values in countries where sociocultural issues dominate political manifestos.
5.2 While many studies have focused on the feedback effects of policies on public opinion, very few studies have examined the influence of political articulation on attitudes. We believe our contribution is important since it also identifies a more direct influence of political elites on attitude formation, more specifically, on the value basis of sociopolitical attitudes. We do realize that political attitudes also constitute input into the political process (see, e.g., Lipset and Rokkan 1967) and we have by no means been able to determine the exact nature of this bidirectional relationship. Nevertheless, our results highlight a relationship between the political context and public opinion that goes beyond pure policy feedback effects.
5.3 We are aware that using data on political manifesto content is not ideal when studying political articulation and issue framing. At best, these data constitute a reasonably good proxy for what the public perceives about the content of politics and the state of political debate in any given country. Certainly, many people do not read political manifestos, but instead obtain political information through the mass media and other sources. However, we believe that political manifestos greatly reflect the political content received by the public through the mass media, especially given the unequivocal nature of our results. In the absence of comparative and representative data on political articulation and framing through the media, political manifestos constitute a viable alternative.
5.4 One finding of our study is that the occurrence of one set of values (i.e., higher-order value type) driving attitudes usually implies the simultaneous activation of opposing higher-order value types (within each dimension) having opposite effects. Hence, activation of one value dimension to facilitate electoral support may very well simultaneously activate opposing values that undermine the desired support. For instance, in countries where socioeconomic issues are salient in politics, self-transcendence values have a relatively strong positive effect, while self-enhancement values also have a comparatively strong negative effect. This implies that political articulation directs the attention of electorates to relevant sociopolitical dimensions, but not necessarily to specific value types within each dimension. Therefore, political articulation seems to influence public views on whether political conflict should concern socioeconomic or sociocultural issues - i.e., which sociopolitical dimension is more important - though it does not seem to influence public views on how this conflict should be resolved. In this sense, the public does not seem to passively internalize ideological messages conveyed by the political elites. Perhaps this is one of the reasons why Converse (1964) and others after him failed to link the ideological reasoning of the public to political elite framing.
AcknowledgementsA previous version of the article was presented at the 5th European Survey Research Association (ESRA) Conference, Ljubljana, 15-19 July in 2013. We thank the commentators at the ESRA conference, as well as three anonymous reviewers, for critical and helpful comments.
Notes1The questions were implemented along with the Portrait Value Questionnaire (PVQ), which has been validated by various studies (e.g. Schwartz et al. 2001, Schmidt et al. 2007).
Appendix 1: Model specification and fit
Estimating an SEM has two main advantages. The first advantage is that SEM provides a way to work with latent variables and relate them to each other in a one-step estimation process controlling for measurement errors. To be more precise, our model is based on multiple-item constructs (i.e., latent variables) that build exclusively on the shared variance of all indicators in the items, thereby eliminating the item-specific variance by placing it in the measurement error. This procedure, which improves measurement quality, is something that survey researchers have only recently begun to acknowledge (Saris and Sniderman 2004). Hence, we avoid relying on a more problematic operationalization, such as indices including disputable applications of Cronbach's alpha (Revelle and Zinbarg 2009; Sjitsma 2009). In doing so, we can ensure that our latent constructs reflect only the common variance, which, in the context of the political attitudes measured here, is especially relevant since we do not want spurious variation in any of our items to influence our latent measurement construct.
In our context, such SEM supports our theoretical claim that there exist underlying abstract attitude dimensions based on more concrete attitude measures. In SEM terminology this is 'structural equivalence', which compares the imposed structure including all specified associations between variables, as seen in Figure 2, with the independence and the saturated model. Situating the model somewhere between perfect fit to the data and theoretical independence of all variables provides grounds for model comparison based on fit statistics. There is a great variety of fit statistics with specific thresholds. We provide Bentler's comparative fit index (CFI) with a threshold of >0.9 (Bentler 1990), the root mean square error of approximation (RMSEA) with a cutoff value of <0.05 (Steiger 1990), AIC, and minimum discrepancy measures (for an account of fit indices, see Kline 2011: 193ff). To conclude this discussion of the first advantage, SEM lets us confirm the structural relationship in a more complex way implying a higher measurement quality.
The second advantage of SEM relates to our interest in cross-national comparisons. When applying latent constructs, we must be sure to measure the same construct across countries. Using SEM provides unique opportunities to test the same model across multiple samples using a multi-group framework (i.e., MGSEM). MGSEM can be specified to impose equality constraints across groups on various parameters, starting with testing the equivalence of variable relationships - often referred to as structural invariance - and increasing the number of parameters constrained to be equal. Theoretically, all parameters could be made equal, which would simultaneously test for the equivalence of the groups with regard to all variables and parameters. In our case, we are interested in the association between two latent constructs and therefore have no interest in constraining this association to be equal. To interpret the association between values and attitudes, we imposed measurement invariance on the values across countries (Steenkamp & Baumgartner, 1998; Vandenberg & Lance 2000). For instance, self-transcendence is based on five indicators, which will have separate weights in relation to the latent construct for this value type, but each weight is constrained to be equal across countries (groups). Consequently, if an acceptable model fit is achieved, the higher-order value types can be considered comparable across countries. For theoretical reasons, we abstained from imposing measurement invariance on the political attitudes, as we assume that similar but not identical conflicts exist across societies. Hence, we imposed only structural invariance, as we assume that the sociocultural dimension in all countries includes our set of indicators.
As we are interested in the relationships between the two higher-order value dimensions and sociocultural and socioeconomic attitudes, four models were estimated. Two models were estimated for each attitude measure, each using one of the two underlying higher-order value types. We decided to separate the analyses of the different higher-order value types due to the known problems with implementing the human value scale in the ESS. Various previous papers have indicated the low discriminatory validity of the reduced Portrait Value Questionnaire covering only 21 of the original 52 value items (Davidov 2008, 2010; Knoppen & Saris 2009; Schmidt et al. 2007). Therefore, we decided to estimate the effects of each higher-order type separately, as our interest is less in testing the value structure itself than in the types' relationships with attitudes. Nonetheless, the antagonistic relationship will be evident in the direction and strength of the effects on attitudes as well.
The decision to estimate distinct models for both attitude domains is motivated by the idea that both concepts are not equally exclusive in each national context. Considering both domains in one model would impose assumptions about the relationship in order to interpret the regression weights. Such a claim is beyond our hypothesis and would tremendously increase model complexity. In the four estimated models, the assumptions tested for are measurement invariance across higher-order value structures and structural equivalence along the attitude domains. In other words, the higher-order value structures are identically composed along all factor loadings across all country samples. Meanwhile, countries may vary in the relevance of the four dimensions to each attitude domain. In some countries, the sociocultural dimension may be driven more by gender issues than immigration, while in a country where immigration is a contentious issue, the importance of gender may be less relevant. Hence, we cannot claim measurement equivalence for attitude domains, but the model supports the claim that, in each tested country, the domains consist of these four dimensions, and the analysis is based only on the shared variance of the latent variables in each country. Such a conclusion satisfies our claims in our hypotheses, in which we refer only to attitude domains and not to common understandings of higher-order attitudes.
|Table A1. Model comparisons (model fit measures) for measurement models (CFA) and structural models (MGSEM).|
We tested for measurement invariance and structural equivalence in order to compare the structural reliability of our models across countries. Measurement invariance was imposed on the value construct to ensure that the higher-order structures were similar in meaning, as previously tested (e.g., Davidov et al. 2008; Kulin and Svallfors 2013). Table A1 shows the relevant fit measures for the confirmatory factor analysis of the two latent attitude concepts and the final four value-attitude models. All models meet the good fit criteria, with CFI > 0.9 and RMSEA < 0.05 for structural equivalence, confirming the similarity of the model structure across countries (for an account of fit indices, see Kline 2011: 193ff). 
Appendix 2: Indicators of political articulation
To capture political articulation of the socioeconomic and sociocultural dimensions, we use measures originally developed by Kumlin and Svallfors (2007, 2008), whereby the proportion (i.e., the relative total number of quasi-sentences) of the party manifestos dedicated to each theme is calculated. Based on this strategy, the different dimensions in this analysis have been constructed by merging multiple categories as follows:
I. Socioeconomic dimension, "left-wing themes": per403 ("Market Regulation") + per404 ("Economic Planning") + per405 ("Corporatism") + per412 ("Controlled Economy") + per413 ("Nationalization") + per503 ("Social Justice") + per504 ("Welfare State Expansion") + per701 ("Labor Groups: Positive")
? II. Socioeconomic dimension, "right-wing themes": per401 ("Free Enterprise") + per402 ("Incentives") + per414 ("Economic Orthodoxy") + per505 ("Welfare State Limitation") + per702 ("Labor Groups: Negative") + per704 ("Middle Class and Professional Groups").?
III. Sociocultural dimension, "left-wing themes": per 201 ("Freedom and Human Rights") + per602 ("National Way of Life: Negative") + per604 ("Traditional Morality: Negative") + per607 ("Multiculturalism: Positive") + per705 ("Underprivileged Minority Groups") + per706 ("Non-Economic Demographic Groups")?
IV. Sociocultural dimension, "right-wing themes": per601 ("National Way of Life: Positive") + per603 ("Traditional Morality: Positive") + per 605 ("Law and Order") + per608 ("Multiculturalism: Negative").?
For a detailed account of the contents of each category, see Budge et al. (2001: App III).
The measure that we finally use and report in the analyses is the balance (i.e., difference) between the share of program content in each country that deals with socioeconomic issues and the share of program content in each country that deals with sociocultural issues, i.e., [I + II] - [III+IV]. See Klingemann et al. (2006: App. I and II) for a detailed account of each category.
ReferencesAJZEN, I. (2001). Nature and Operation of Attitudes. Annual Review of Psychology, 52(1), 27-58. [doi://dx.doi.org/10.1146/annurev.psych.52.1.27]
BARNEA, M. F. (2003). Personal Values and Party Orientations in different Cultures. Unpublished doctoral dissertation, The Hebrew University of Jerusalem, Israel.
BARNEA, M., & Schwartz, S. H. (1998). Values and Voting. Political Psychology, 19, 17-40. [doi://dx.doi.org/10.1111/0162-895X.00090]
BOWLES, S., & Gintis, H. (2000). Reciprocity, self-interest, and the welfare state. Nordic Journal of Political Economy, 26. Available from: http://www.nopecjournal.org/NOPEC_2000_a02.pdf [Accessed 12 October 2013]
BROOKS, C., & Manza, J. (2007). Why welfare states persist: the importance of public opinion in democracies. University of Chicago Press.
BUDGE, Ian, Hans-Dieter Klingemann, Andrea Volkens, Judith Bara, Eric Tanenbaum with Richard C. Fording, Derek J. Hearl, Hee Min Kim, Michael D. McDonald, Sylvia M. Mendes (2001). Mapping Policy Preferences. Estimates for Parties, Electors, and Governments 1945-1998. Oxford: Oxford University Press.
CONVERSE, P. E. (1964). The Nature of Mass Belief Systems. In D. E. Apter (Ed.), Ideology and Discontent. New York: Free Press.
CUSACK, T., Iversen, T., & Rehm, P. (2006). Risks at Work: The Demand and Supply Sides of Government Redistribution. Oxford Review of Economic Policy, 22(3), 365 - 389. [doi://dx.doi.org/10.1093/oxrep/grj022]
DAVIDOV, E., & Meuleman, B. (2012). Explaining Attitudes Towards Immigration Policies in European Countries: The Role of Human Values. Journal of Ethnic and Migration Studies, 38(5), 757 - 775. [doi://dx.doi.org/10.1080/1369183X.2012.667985]
DAVIDOV, E. (2010). Testing for Comparability of Human Values across Countries and Time with the third Round of the European Social Survey. International Journal of Comparative Sociology, 51(3), 171 - 191. [doi://dx.doi.org/10.1177/0020715210363534]
DAVIDOV, E. (2008). A Cross-Country and Cross-Time Comparison of the Human Values Measurements with the Second Round of the European Social Survey. Survey Research Methods, 2(1), 33 - 46. Available from: https://ojs.ub.uni-konstanz.de/srm/article/view/365/1297 . [Accessed 5 April 2013]
DAVIDOV, E., Meuleman, B., Billiet, J., & Schmidt, P. (2008a). Values and Support for Immigration: A Cross-Country Comparison. European Sociological Review, 24(5), 583 - 599. [doi://dx.doi.org/10.1093/esr/jcn020]
DAVIDOV, E., Schmidt, P., & Schwartz, S. H. (2008b). Bringing Values Back In: The Adequacy of the European Social Survey to Measure Values in 20 Countries. Public Opinion Quarterly, 72(3), 420. [doi://dx.doi.org/10.1093/poq/nfn035]
EAGLY, A. H., & Chaiken, S. (1993). The Psychology of Attitudes. Fort Worth, Texas: Harcourt Brace Jovanovich College Publishers.
ESS Round 3: European Social Survey Round 3 Data (2006). Data file edition 3.4. Norwegian Social Science Data Services, Norway - Data Archive and distributor of ESS data.
FISHBEIN, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior?: an introduction to theory and research (p. xi, 578 p.). Reading, Mass.: Addison-Wesley Pub. Co.
FEHR, E., & Fischbacher, U. (2002). Why social preferences matter-the impact of non?selfish motives on competition, cooperation and incentives. The economic journal, 112(478), C1 - C33. [doi:10.1111/1468-0297.00027]
FELDMAN, S., & Zaller, J. (1992). The Political Culture of Ambivalence: Ideological Responses to the Welfare State. American Journal of Political Science, 36(1), 268 - 307. [doi://dx.doi.org/10.2307/2111433]
FELDMAN, S. (2003). Values, identity, and the structure of political attitudes. In D. O. Sears, L. Huddy, & R. Jervies (Eds.), Oxford handbook of social psychology (p. 477 - 508). New York: Oxford University Press.
FELDMAN, S. (1988). Structure and Consistency in Public-Opinion: the Role of Core Beliefs and Values. American Journal of Political Science, 32(2), 416 - 440. [doi://dx.doi.org/10.1111/ 10.2307/2111130]
GINTIS, H., Bowles, S., Boyd, R., & Fehr, E. (2005). Moral sentiments and material interests: The foundations of cooperation in economic life. Economic learning and social evolution. Cambridge, Mass.: MIT Press.
HITLIN, S. (2003). Values as the Core of Personal Identity: Drawing Links between two Theories of Self. Social Psychology Quarterly, 66(2), 118 - 137. [doi://dx.doi.org/10.2307/1519843]
HITLIN, S., & Piliavin, J. A. (2004). Values: Reviving a dormant Concept. Annual Review of Sociology, 30, 359 - 393. [doi://dx.doi.org/10.2307/29737698]
IVERSEN, T., & Soskice, D. W. (2001). An asset theory of social policy preferences. American Political Science, 95(4), 875 - 893.
JACOBY, W. G. (2000). Issue Framing and Public Opinion on Government Spending. American Journal of Political Science, 750 - 767. [doi://dx.doi.org/10.2307/2669279]
KINDER, Donald R. (1998). Opinion and Action in the Realm of Politics. In Gilbert, D.; Fiske, S. T.; Lindzey, G. (Eds), The Handbook of Social Psychology. Oxford University Press.
KLINE, R. B. (2011). Principles and practice of structural equation modeling (3rd ed.). New York: The Guilford Press.
KLINGEMANN, Hans-Dieter, Andrea Volkens, Ian Budge, Judith Bara, and Michael D. McDonald (2006). Mapping Policy Preferences II: Parties, Electorates and Governments in Eastern Europe and the OECD 1990 - 2003. Oxford: Oxford University Press.
KNOPPEN, D., & Saris, W. E. (2009). Do we have to combine Values in the Schwartz' Human Values Scale? A Comment on the Davidov Studies. Survey Research Methods, 3(2), 91 - 103. Available from: https://ojs.ub.uni-konstanz.de/srm/article/download/2601/2778 . [Accessed 5 April 2013]
KULIN, J., & Svallfors, S. (2013). Class, Values, and Attitudes Towards Redistribution: A European Comparison. European Sociological Review, 29(2), 155 - 167. [doi://dx.doi.org/10.1093/esr/jcr046] [doi://dx.doi.org/10.1093/esr/jcr046]
KULIN, J. (2011) Values and welfare state attitudes: The interplay between human values, attitudes and redistributive institutions across national contexts. PhD thesis. Umeå: Umeå University. Avalaible from: http://urn.kb.se/resolve?urn=urn:nbn:se:umu:diva-49853 [Accessed 12 April 2013]
KUMLIN, S., & Svallfors, S. (2008). Social Stratification and Political Articulation: Why Attitudinal Class Differences Vary Across Countries. Luxemburg Income Study Working Paper Series, Working Paper No. 484.Available from: http://www.lisproject.org/publications/liswps/484.pdf . [Accessed 5 April 2013]
LIPSET, S. M., & Rokkan, S. (1967). Party Systems and Voter Alignments: Cross-National Perspectives. New York: Free Press.
LIPSET, S. M. (1979). The first new nation: The United States in historical and comparative perspective. New York: Norton.
MAU, S. (2003). The Moral Economy of Welfare States: Britain and Germany Compared. (M. Rhodes & M. Ferrera, Eds.) Social Research (p. 238). New York, London: Routledge.
MCCLOSKY, H., & Zaller, J. (1984). The American Ethos: Public Attitudes toward Capitalism and Democracy. Cambridge, MA: Harvard University Press.
METTLER, S., & Soss, J. (2004). The Consequences of Public Policy for Democratic Citizenship: Bridging Policy Studies and Mass Politics. Perspectives on Politics, 2(1), 55 - 73. [doi://dx.doi.org/10.1017.S1537592704000623] [doi://dx.doi.org/10.1017/S1537592704000623]
MELTZER, A., & Richard, S. (1981). A rational theory of the size of government. The Journal of Political Economy, 89(5), 914-927. Avaiable at: http://www.jstor.org/stable/1830813 [Accessed 12 October 2013] [doi://dx.doi.org/10.1086/261013]
PIURKO, Y., Schwartz, S. H., & Davidov, E. (2011). Basic Personal Values and the Meaning of Left-Right Political Orientations in 20 Countries. Political Psychology, 32(4), 537 - 561. [doi://dx.doi.org/ 10.1111/j.1467-9221.2011.00828.x] [doi://dx.doi.org/10.1111/j.1467-9221.2011.00828.x]
REVELLE, W., & Zinbarg, R. E. (2008). Coefficients Alpha, Beta, Omega, and the glb: Comments on Sijtsma. Psychometrika, 74(1), 145 - 154. doi:10.1007/s11336-008-9102-z [doi://dx.doi.org/10.1007/s11336-008-9102-z]
ROKEACH, M. (1968). Beliefs, Attitudes, and Values: A Theory of Organization and Change. San Fransisco: Jossey-Bass.
ROKEACH, M. (1973). The Nature of Human Values. New York: Free Press.
SAGIV, L., & Schwartz, S. H. (2007). Cultural Values in Organisations: Insights for Europe. European Journal of International Management, 1(3), 176 - 190. Available from: http://inderscience.metapress.com/content/t76855265p328878/fulltext.pdf . [Accessed 5 April 2013] [doi://dx.doi.org/10.1504/EJIM.2007.014692]
SARIS, W. E., & Sniderman, P. M. (2004). Studies in Public Opinion. Princeton, N.J.: Princeton University Press.
SCHMIDT, P., Bamberg, S., Davidov, E., Herrmann, J., & Schwartz, S. H. (2007). Die Messung von Werten mit dem 'Portraits Value Questionnaire'. Zeitschrift für Sozialpsychologie, 38(4), 261 - 275. [doi://dx.doi.org/10.1024/0044-3518.104.22.1681]
SCHNEIDER, S. K., & Jacoby, W. G. (2005). Elite Discourse and American Public Opinion: The Case of Welfare Spending. Political Research Quarterly, 58(3), 367 - 379. [doi://dx.doi.org/ 10.1177/106591290505800301]
SCHWARTZ, S. H. (2009). Culture Matters: National Value Cultures, Sources and Consequences. In Wyer, R. S., Chiu, C. and Hong, Y. (Eds). Understanding Culture: Theory, Research and Application. Psychology Press.
SCHWARTZ, S. H., & Boehnke, K. (2004). Evaluating the structure of human values with confirmatory factor analysis. Journal of Research in Personality, 38(3), 230 - 255. [doi://dx.doi.org/doi:10.1016/S0092-6566(03)00069-2] [doi://dx.doi.org/10.1016/S0092-6566(03)00069-2]
SCHWARTZ, S. H., Melech, G., Lehmann, A., Burgess, S. M., Harris, M., & Owens, V. (2001). Extending the Cross-Cultural Validity of the Theory of Basic Human Values with a Different Method of Measurement. Journal of Cross-Cultural Psychology, 32(5), 519 - 542. [doi://dx.doi.org/ 10.1177/0022022101032005001] [doi://dx.doi.org/10.1177/0022022101032005001]
SCHWARTZ, S. H., & Sagie, G. (2000). Value Consensus and Importance: A Cross-National Study. Journal of Cross-Cultural Psychology, 31(4), 465 - 497. [doi:10.1177/0022022100031004003] [doi://dx.doi.org/10.1177/0022022100031004003]
SCHWARTZ, S. H., & Sagiv, L. (1995). Identifying Culture-Specifics in the Content and Structure of Values. Journal of Cross-Cultural Psychology, 26(1), 92. [doi: //dx.doi.org/10.1177/0022022195261007] [doi://dx.doi.org/10.1177/0022022195261007]
SCHWARTZ, S. H. (1994). Are There Universal Aspects in the Structure and Contents of Human-Values. Journal of Social Issues, 50(4), 19 - 45. [doi://dx.doi.org/10.1111/j.1540-4560.1994.tb01196.x] [doi://dx.doi.org/10.1111/j.1540-4560.1994.tb01196.x]
SCHWARTZ, S. H. (1992). Universals in the Content and Structure of Values: Theoretical Advances and Empirical Tests in 20 Countries. Advances in Experimental Social Psychology, 25(1), - 65. [doi://dx.doi.org/10.1016/S0065-2601(08)60281-6] [doi://dx.doi.org/10.1016/S0065-2601(08)60281-6]
SEARS, D. O., & Funk, C. L. (1991). The Role of Self-Interest in Social and Political Attitudes. Advances in Experimental Social Psychology, 24(1), 1 - 91. [doi://dx.doi.org/10.1016/S0065260108603275] [doi://dx.doi.org/10.1016/S0065-2601(08)60327-5]
SEARS, D., & Funk, C. (1990). The limited effect of economic self-interest on the political attitudes of the mass public. Journal of Behavioral Economics, 19(3), 247 - 271. [doi://dx.doi.org/10.1016/0090-5720(90)90030-B]
SEARS, D. O., Lau, R. R., Tyler, T. R., & Allen Jr, H. M. (1980). Self-Interest vs. Symbolic Politics in Policy Attitudes and Presidential Voting. The American Political Science Review, 670-684. [doi://dx.doi.org/10.2307/1958149] [doi://dx.doi.org/10.2307/1958149]
SIJTSMA, K. (2009). On the Use, the Misuse, and the Very Limited Usefulness of Cronbach's Alpha. Psychometrika, 74(1), 107 - 120. [doi:10.1007/s11336-008-9101-0] [doi://dx.doi.org/10.1007/s11336-008-9101-0]
STEENKAMP, J. E. M., & Baumgartner, H. (1998). Assessing Measurement Invariance in Cross?National Consumer Research. Journal of Consumer Research, 25(1), 78 - 107. [doi:10.1086/209528] [doi://dx.doi.org/10.1086/209528]
STEIGER, J. H. (1990). Structural model evaluation and modification: An interval estimation approach. Multivariate behavioral research, 25(2), 173 - 180. Available from: http://www.statpower.net/Steiger Biblio/Steiger90b.pdf [Accessed 12 October 2013] [doi://dx.doi.org/10.1207/s15327906mbr2502_4]
SVALLFORS, S. (2007). The Political Sociology of the Welfare State: Institutions, Social Cleavages, and Orientations. Stanford: Stanford University Press. [doi://dx.doi.org/10.11126/stanford/9780804754354.001.0001]
SVALLFORS, S. (2005). Class and Conformism: A Comparison of four Western Countries. European Societies, 7(2), 255 - 286. [doi://dx.doi.org/10.1080/14616690500083493] [doi://dx.doi.org/10.1080/14616690500083493]
SVALLFORS, S. (1997). Worlds of Welfare and Attitudes to Redistribution: A Comparison of Eight Western Nations. European Sociological Review, 13(3), 283-304. [doi:10.1093/oxfordjournals.esr.a018219] [doi://dx.doi.org/10.1093/oxfordjournals.esr.a018219]
TETLOCK, P. E. (1986). A Value Pluralism Model of Ideological Reasoning. Journal of Personality and Social Psychology, 50(4), 819 - 827. [doi://dx.doi.org/10.1037/0022-3522.214.171.1249] [doi://dx.doi.org/10.1037/0022-35126.96.36.1999]
ULLRICH, C. G. (2002). Reciprocity, justice and statutory health insurance in Germany. Journal of European Social Policy, 12(2), 123 - 136. [doi:10.1177/0952872002012002111] [doi://dx.doi.org/10.1177/0952872002012002111]
VANDENBERG, R. J., & Lance, C. E. (2000). A Review and Synthesis of the Measurement Invariance Literature: Suggestions, Practices, and Recommendations for Organizational Research. Organizational Research Methods, 3(1), 4 - 70. [doi:10.1177/109442810031002] [doi://dx.doi.org/10.1177/109442810031002]
VERPLANKEN, B., & Holland, R. W. (2002). Motivated Decision Making: Effects of Activation and Self-Centrality of Values on Choices and Behavior. Journal of Personality and Social Psychology, 82(3), 434 - 447. [doi://dx.doi.org/10.1037//0022-35188.8.131.524] [doi://dx.doi.org/10.1037/0022-35184.108.40.2064]