Official final compilation of notes:

1. The Impact of Values By: Jan W. van Deth; Elinor Scarbrough Left-Right Materialist Value Orientations ODDBJRN KNUTSEN
According to the traditional model of industrial society, class conflicts were central for political identities and political preferences. The value orientations associated with the political conflicts of industrial society reflected these class conflicts in terms of left-right polarization. The key issues underlying this conceptualization were conflicts related to economic inequalities, the ownership of the means of production, and conflict over the desirability of a market economy. The ideology which supported emerging capitalism in the nineteenth century was economic and political liberalism. The value orientation associated with this ideology could also be called 'bourgeois' since these values were associated with the bourgeois class in the emerging capitalist society. Socialism and social democracy were the ideological response of the political forces of the left to the social and economic conflicts which characterized industrial society. At the mass level, these orientations grew out of the social experiences of industrial society and, in particular, the economic crises of the 1920s and 1930s. These value orientations will be labelled here--respectively--as right materialism and left materialism. Taken together they constitute the left-right materialist value orientation (cf. Von Beyme 1985: ch. 2; Lafferty and Knutsen 1984; Knutsen 1986a; Inglehart 1984: 25). A central belief of right materialism is the supremacy of the market. This entails an emphasis on economic competition between independent enterprises, personal freedom, and the integrity of private property, including private ownership of the means of production. In political terms, these values imply a relatively weak state, resistance against government regulation, and opposition to notions of social and conomic equality. A free market is socially desirable, both as the best way to organize the production and distribution of goods, and as a motivation system stimulating personal achievement. Moreover, in the long run, markets contribute to important collective interests. The market is also an effective information system for decentralizing decision-making, and an effective steering mechanism for distributing scarce resources. Left materialism is based on the notion that governments have an active role to play in achieving overarching goals such as economic security, solidarity, and equality in income and living conditions between social classes and strata. According to this value orientation, government regulation of markets and private enterprises, societal planning, government redistribution via progressive taxation and welfare reforms are necessary for obtaining important goals such as full employment and social equality. Left materialism

points to the failure of market mechanisms, which create social and economic inequalities, economic crises, and severe social conflicts. A regulated economy contributes to economic efficacy and individual work effort, and consequently to high productivity and economic growth. This clearly contradicts the bourgeois argument that there is an inevitable conflict between equality and strong government on the one hand, and efficiency and productivity on the other (cf. Castles 1978; Esping-Andersen 1985; Scharpf 1987). Left materialism and right materialism also take a different view on the control of industry and the workplace. Left materialism emphasizes societal control 'from above' through state intervention, and participation and control 'from below' by workers and other employees through representation systems and industrial democracy. Right materialism emphasizes that ownership implies responsibility: it is a property right of owners to have the final word in the decision-making of firms. Thus the value conflicts incorporated in left-right materialism relate to power and control in the sphere of production and the degree of redistribution in the sphere of consumption. They centre around opposing ideas of workers' control and state regulation of the economy versus an emphasis on private enterprise and the market economy; an emphasis on economic and social equality versus the need for differentiated rewards to stimulate efforts. In brief, the political conflicts embedded in left-right materialism are, at heart, economic in nature, referring in particular to the role of government in the economy. In this chapter, we focus on four aspects of left-materialist and right-materialist value orientations. First, the concepts of left and right materialism are compared to, and demarcated from, the more common notions of 'new' and 'old' politics and ideas referred to as the 'new right'. Secondly, we consider whether or not left-right materialism exists, empirically, at the mass level; that is, whether survey items which are assumed to tap left and right materialism are constrained at the level of the mass public. Thirdly, we examine the distribution of these value orientations over time and in comparative perspective. What are the long-term trends in the emphases on these values at the mass level? Which countries have the most leftist and most rightist population, and what explains the resulting comparative patterns? Finally, we look at the social structural antecedents of left and right materialism, again over time and comparatively. Are these value orientations primarily anchored in social class, as suggested by their association with the pattern of conflict in industrial society? And are there changes over time and between generations within countries with regard to these relationships?

Changing Value Orientations

Left-right materialism was a central political value orientation in industrial society, but is it characteristic of advanced industrial societies? New political values have been associated with advanced industrial society, primarily in the notion of 'new politics' and Inglehart's concept of materialist-postmaterialist orientations (see Chapter 5). Moreover, in the late 1970s and 1980s, there was a revival of conservative and libertarian ideas and movements, and increased support for political parties identified as 'new right'. How is the conceptualization of left and right materialism, as proposed in this chapter, related to these notions of 'new' and 'old' politics and the 'new right'? The values incorporated in left-right materialism are central to the ideological polarization between the 'old' left and right. The term 'old politics' refers to the

polarization patterns at the socio-structural and political level in typical industrial society and is based on the political conflicts between the 'old left' and the 'old right'. The class cleavage, focused on the labour market, was central in structuring the alignments of old politics, but old politics can also be conceptualized to include the pre-industrial cleavages related to religion ( Dalton 1988: 133). Left- right materialism, together with religious and secular value orientations, can be considered as the central conflicting orientations of old politics ( Inglehart 1984: 25-6; Dalton 1988: 133-49, 207-9). The relationship between materialism-postmaterialism and left-right materialism can be conceptualized in several ways. From a developmental perspective, we can regard a materialist orientation as the common values of individuals with leftist and rightist materialist orientations in industrial society. However, with the emergence of advanced industrialism and postmaterialist values, a new and cross- cutting set of orientations has arisen which challenge the centrality of left-right materialism in the political belief systems of mass publics. This conceptualization, however, only partly accords with conceptualizations of value change from a materialist to a postmaterialist orientation as a change from the values of 'old' politics to the values of 'new' politics. In this literature, there is a tendency to characterize the shift from materialist to postmaterialist orientations as identical to a shift from old politics to new politics, without acknowledging the persisting divisiveness of the value orientations of old politics; these divisions do not necessarily disappear with a shift from materialism to post- materialism (cf. Dalton, Beck, and Flanagan 1984: 20; Dalton 1988: 82). According to our conceptualization, left-right materialism and materialism-postmaterialism are separate sets of value orientations. Hence, individuals of a postmaterialist orientation have to take a standpoint regarding left-right materialism because it still appears to be central to political conflict in most Western democracies. 1 The conflict between the 'old' and 'new' left illustrates the relationship between the new and the older value orientations. In industrial societies, the left generally accepted a considerable degree of organizational discipline and hierarchy as necessary for effective political change. The left in advanced industrial politics is oriented more towards individual self-expression and tends to reject hierarchical organizations. Suspicious of the state, the new left tends to be more sympathetic to individualism, communalism, and variety in life-styles, while the old left tends to favour centralized, bureaucratic solutions, traditional life-styles and particularistic cultures--class culture, for example. The old left considers economic growth to be central for economic progress--to improve standards of living for the working class and other groups--to which the new left object. Even so, the new left also considers the traditions of the industrial left as central belief elements. Distinct from the debate on 'old' and 'new' politics, the 1970s and 1980s have seen a revival of conservative and liberal ideas. This heterogeneous movement has been labelled the 'new right'. Although accounts of the new right vary, most authors agree that the core ideas are economic in character, with an emphasis on individualism, a free market, and a limited state. Hence, the ideas of the new right have challenged the post-war consensus on the welfare state, state intervention, and citizenship rights in the social domain--that is, central elements of left materialism which have been implemented in most West European polities. The new right emphasises the superiority of market mechanisms as a promoter of both economic prosperity and individual freedom, and stresses that state

intervention does not work. Accordingly, new right thinking favours a minimal role for the state in both the economy and the social order ( King 1987; Levitas 1985). Thus, although the new right can best be understood as emphasizing right materialist values, other 'conservative' moral and social beliefs are also involved, particularly reaction against the social and sexual liberation of the 1960s and 1970s ( King 1987: 19-20)--that is, reaction against 'new politics' ( Flanagan 1987; Flanagan and Lee 1988). However, according to King ( 1987: 17), 'conservatism is secondary to liberalism in New Right ideology'. From this comparison of left and right materialism, on the one hand, and 'old' and 'new' politics and the 'new right' on the other, it becomes clear that the conflict of value orientations in advanced industrial societies in the 1970s and 1980s is related to divergent ideas grouped around our distinction between left and right materialism. There seems to be firm evidence that the shifts in value orientation in many countries can be understood in these terms. This brings us to the empirical part of our analyses.

Measuring Left-Right Materialism

Although the terminology is only rarely used, the empirical relevance of left and right materialist value orientations can be illustrated in many ways. For example, in Lijphart's study of twenty-one Western democracies, the 'socio-economic dimension' is the only one which is salient in all the countries examined. Moreover, compared to six other issue dimensions, the salience of the socio-economic dimension is 'high' in fourteen out of the fifteen West European countries examined ( Lijphart 1984: 128-32). Thus, in a comparative perspective, this dimension is clearly the most salient. This dimension is identical to our left-right materialist orientation, comprising four sets of oppositions: state versus private ownership of the means of production; a strong versus a weak role for government in economic planning; support versus opposition on redistributing wealth from the rich to the poor; and support versus resistance on expanding social welfare programmes. Lijphart's discussion remains somewhat speculative and subjective, but the overall observation is confirmed in Budge and Farlie ( 1983: ch. 2) study of political issue types in post-war elections in twenty-three democracies. 2 Only a few of the available comparative data sets have satisfactory indicators of left-right materialism, and it is impossible to find comparative data containing the same indicators in several countries which span one decade or longer. Our strategy then is, first, to present findings from some of the West European countries where identical items covering left-right materialism have been asked systematically over time in national surveys. Then, secondly, we look at evidence from comparative surveys featuring such items. 3 Our aim is to trace the degree of constraint among the items which, we propose, tap left-right materialist orientations and to construct indices to measure these orientations. The indices are unweighted additive indices, and, with one exception, all have eleven categories, ranging 0-10. The alternative index constructed from the Danish Election Studies 1979, 1984, and 1987 has seven categories, ranging 0-6. An indication of these items is given in the tables. 4

Country-Specific Evidence
We have country-specific data sets covering a time span of at least ten years for five countries: Denmark, Norway, Sweden, Britain, and the Netherlands. With only some

exceptions, successive election surveys or other regular national surveys in these countries have posed the same items. In addition, we have data for Germany with items on left-right materialism which cover a period of six years. Our strategy is to include as many left-right materialist items as possible in each analysis so long as they are not 'narrow' issues related to particular political objects. Since the published work we refer to outlines results from factor analyses similar to our analyses using the country-specific data sets, the results from our analyses are presented without the tables. The existence of left-right orientations consisting of items which tap attitudes and values related to what we call left-right materialism, has been documented in almost all election studies for the 1970s and 1980s in Norway and Sweden. In Norway, analyses of some thirty to forty items on 'political values' and 'issues' in the election surveys for 1977, 1981, and 1985 have established that conflict between notions of public control and private initiative covers several elements of left-right materialism ( Valen 1981: 248-9, table 11.3; Valen and Aardal 1983: 164-5, table 7.2; Aardal and Valen 1989: 60-1, table 4.2). The same applies to Sweden. Based on the 1979 election survey, Holmberg ( 1981: 262, table 12.8) found four factors: a left-right materialist factor which explains most of the variance in the items; a nuclear energy factor; a moral factor; and a factor related to social care. More recently, Gilljam ( 1990: 275) reports that an economic left-right dimension is found in every election survey since the first in 1956--altogether eleven studies spanning thirty-two years. Findings about constraint among traditional left-right values and issues in Sweden are summed up by Bennulf and Holmberg ( 1990: 176). Discussing whether there exists a green versus economic growth dimension among the Swedish mass public, they find that the strength of the correlations between several green items are only moderate compared to the strong correlations among left-right items (coefficients of 0.11-0.23 and 0.50-0.55 respectively). The authors doubt the existence of a green dimension, but to doubt the existence of a left-right materialist dimension would be absurd. Moreover, in a comparative study of the 1976 and 1977 elections, Petersson and Valen ( 1979) found an impressive similarity in the structuring of values among the Swedish and Norwegian electorates. They identified four dimensions, which they labelled government influence, energy and environment, equality (related to gender issues) and cultural outlook (related to religious and moral values). The first factor in both countries was the left-right materialist dimension which, the authors concluded, 'takes on a striking predominance' ( Petersson and Valen 1979: 319). Finally, left-right items have frequently been used to analyse the relationship between value orientations and party choice in Denmark. Borre ( 1984: 160) found quite strong correlations between four items among the Danish mass publics, of which three tap central elements of left-right materialism in each election survey between 1971 and 1981. To assess the structure of the items which, we anticipate, measure left-right materialist orientations in the three Scandinavian countries, we factor analysed data from election surveys conducted during the 1970s and 1980s. 5 Our results (not shown here) provide impressive support for the hypothesis that left-right materialism constitutes a major and distinctive value orientation among the mass public. In each survey, in each country, there is only one factor with an eigenvalue larger than 1.00, and the common factors have loadings of more than 0.50 on most of the items. In Norway and Sweden, the items which have a particularly high loading on the common factor are 'nationalization', 'risk of unemployment/public control', and 'societal control/ influence of leaders'. Our results for Denmark, too, provide strong support for the empirical reality of left-right

materialism. Moreover, analyses of a more comprehensive series of items from the election surveys for the late 1980s reveal that left-right materialism remains a dominant orientation in all three countries; the number of factors differs but the first and major factor in each country is always our measure of left-right materialism. Despite this clear patterning, however, the left- right materialist indices for Denmark and Norway could not be based on the same items for the whole period as some items have been dropped from surveys conducted after 1985. Hence, two indices were developed for these two countries, one based on items from the older surveys (the main index) and one based on items from more recent surveys (the alternative index). 6 In the case of Britain, the existence of left-right materialism at the mass level has been established by several scholars. The findings, however, are somewhat inconclusive when various types of social welfare issues are included in the analysis. Srlvik and Crewe ( 1983: 173-4) report five attitudinal dimensions for the 1974 (October) and the 1979 British Election Study. The dimensions are the same in the two surveys. The most important are, first, views on state control and other classic left-right issues, and, secondly, opinions on social welfare issues. These findings are confirmed by dimensional analyses of data from the 1974 (October) and 1983 election surveys by Rose and McAllister ( 1986: 119-21, 168, tables 7.1 and appendix D); Studlar and Welch ( 1981) present an even more convincing analysis of the structure of mass beliefs in Britain based on the British Election Study for 1974 (October). In their analyses of the 1983 election, Heath, Jowell, and Curtice ( 1985: ch. 8) propose a two-dimensional map of the British electorate, one dimension represented by attitudes towards nationalization and the other by attitudes towards nuclear weapons. In their subsequent review of electoral change in Britain, this simple map is elaborated to include attitudes on other issues. The authors conclude that there is substantial evidence that political attitudes in Britain have a two-dimensional character ( Heath et al. 1991: 172): On the one hand there is the conventional left-right dimension which is concerned with economic issues such as equality, nationalisation and the welfare state. And on the other there is a liberal-authoritarian dimension which largely cross-cuts the left-right one and which is concerned more with social issues such as law and order. We also used data from the British Election Study ( 1974 October, 1979, 1983, and 1987) to analyse left-right materialism in the British context. The analyses provide firm support for the notion of a single left-right dimension at the mass level. There is also a consistent pattern over time in the factor loadings. 7 A more comprehensive analysis of thirty items in the 1987 election survey reveals six factors. Left-right materialism is the first factor; support versus opposition to central aspects of the welfare state is the second factor ( Knutsen 1991). For the Netherlands, we review the findings reported by Middendorp ( 1978, 1991). Using data from the Dutch Cultural Change Survey, Middendorp ( 1978) examined the main dimensions of Dutch ideological controversies which he conceptualizes as a conflict between conservative and progressive ideology. In a more recent study, using data for 1970, 1975, 1980, and 1985, Middendorp analyses the indices tapping the philosophical and attitudinal levels. He finds that liberalism and socialism have opposite loadings on the economic left-right dimension at the attitudinal level, while conservatism has a high

loading only on the libertarian-authoritarian dimension. The pattern is impressively stable across the four surveys ( Middendorp 1991: 111, table 4.6), with the same two- and three-dimensional structure appearing in each survey. As regards the left-right dimension, Middendorp concludes ( 1991: 111): The left-right dimension is 'truly ideological', not only in terms of the underlying basic value of equality, but also because the two 'opposed' philosophies of socialism and liberalism load consistently negatively and positively, respectively, on this dimension. We selected five indicators from the data used by Middendorp, and our analysis revealed that left-right materialism clearly constitutes a common factor. No other factors emerged with an eigenvalue larger than 1.00 and all five indicators have loadings higher than 0.50 on the first factor in each analysis. Moreover, the factor structure and the loadings showed very high stability in the four surveys covering a period of sixteen years. Finally, the situation in Germany. The 1984-9 IPOS studies contain several traditional left-right materialist items as well as items measuring materialist-postmaterialist orientations. Analyses of the data for 1984, 1985, 1988, and 1989 show an impressively similar structure. Three of the four items which, we anticipate, tap central elements of left-right materialism have high loadings on the first factor in each analysis, while two items tapping the authoritarian-libertarian aspect of materialist-postmaterialist orientations have a high loading on the second or third factor. The item 'social welfare', which we expected to have a high loading on the left-right materialist factor, constitutes- together with an item on environmental protection--a separate factor. The three items with consistently high loadings on the first factor were employed in constructing the leftright materialism index for Germany.

Comparative Evidence
In addition to the country-specific data, we can also examine data from comparative surveys. Items which might tap left-right materialism have been included in several Eurobarometer surveys and in the 1990 European Values Study. Three surveys in the Eurobarometer series have items tapping left- right materialism: No. 11 ( 1979) has five items, and Nos. 16 ( 1981) and 19 ( 1983) have three items. Hence, because of the larger number of items, we use principally Eurobarometer No. 11, although the results obtained from Nos. 16 and 19 are referred to briefly. The results from the analysis of the five items tapping left-right materialism are presented in Table 6.1. These results show impressive support for the notion of a left-right materialist orientation. Only in Germany are there two factors with eigenvalues significantly higher than 1.00. 8 The left-right materialist factor explains 32-44 per cent of the variance in the eight countries, and has loadings higher than 0.45 on all variables in all countries. Most of the loadings are even higher than 0.60. Similar factor analyses based on the three items in the other two Eurobarometers show the same pattern: a single factor which has loadings of 0.50 and higher on each item in each country (not shown here). In these analyses, the common factor explains 40-55 per cent of the variance in the variables. Eurobarometer, No. 19, includes data from Greece, and here too, there is strong support for the notion of a single left-right materialist

dimension. The existence of an economic left-right dimension, based on the five indicators from Eurobarometer, No. 11, has previously been documented by Inglehart ( 1984: 34-42; 1990: 290-300), employing pooled data from the nine European Community countries. Inglehart also included a series of new politics items, and found two factors: one reflecting the classic economic left-right dimension; the other a 'new politics' noneconomic left-right dimension comprising items related to defence, penalties for terrorism, nuclear energy, and abortion. The same structure also appears in an lite sample of candidates to the European Parliament. Thus, the economic left-right dimension appears to be the predominant factor at both the mass level and the lite level. The 1990 European Values Survey contains seven items directly relevant to tapping leftright materialism. These data cover thirteen countries, so giving a wide comparative base. In six countries ( Britain, Italy, Spain, Portugal, Belgium, and Ireland), factor analysis of these seven items resulted in three factors with eigenvalues larger than 1.00; in the other countries, only two factors of this magnitude emerged. It appeared reasonable, therefore, to rotate only two factors for each country. 9 The factor loadings on the first rotated factor are presented in Table 6.2. The items 'wealth accumulation' and 'individual freedom/healthy economy' do not load significantly on the first factor, but the other five items have consistently high loadings on the first factor across ten countries. 10 In the southern countries of Italy, Spain, and Portugal only three or four of these items have loadings higher than 0.30 on this factor. In Britain and Ireland the items 'nationalization' and 'individual/ public responsibility' have much higher loadings than the other three items. Similar indices based on the five indicators were constructed for each country. Additional indicators based only on the three items 'nationalization', 'individual/private responsibility' and 'business management' were constructed for Italy, Spain, and Portugal since only these items loaded significantly on the first factor in these countries. In all, then, our findings provide strong evidence of left and right materialist orientations among West European mass publics. The orientations emerge in all the surveys examined, and there is evidence from many countries that the items which tap these orientations are strongly constrained. Moreover, although many central left materialist values were under attack in the late 1970s and in the 1980s, our analyses show that left and right materialism are still constrained in the traditional way in the late 1980s and in 1990. There is no evidence supporting the idea of weakening constraint.

Change and Stability

We now turn to consider change and stability in left and right materialism. In particular, has there been a value shift in a right materialist direction from the 1970s to the 1980s-as suggested by both increased support for liberal and conservative parties and the literature on the new right? For Britain, Denmark, the Netherlands, Norway, and Sweden, we can trace developments from the early or mid-1970s to the late 1980s, but the period for Germany is shorter. The distributions from the comparative data can be used for investigating the priority of left and right materialism cross-nationally.

As a simple indicator of change, an 'opinion balance' was computed by subtracting the percentage of respondents with a right materialist orientation from the percentage with a left materialist orientation. 11 Positive scores indicate that there is a majority with left materialist priorities. Table 6.3 shows the opinion balance for the items which had high loadings on the left-right materialism factor; most items have appeared in at least two surveys but, for some countries, we also include items asked only once in the most recent surveys. In addition to the opinion balances, Table 6.3 presents the mean scores on the left- right materialist index. The opinion balances confirm that there was a value shift in a right materialist direction around 1980 in most of these countries. In Denmark, there was a major shift from left materialist dominance among the mass public to more right materialist priorities. The major shift in the main index took place between 1979 and 1981, but there was some recovery in a left materialist direction between 1981 and 1984. According to the alternative index for Denmark, there was a marked tendency towards left materialism from 1984 to 1987. In all, then, by 1987, the Danish mass public was just as left materialist as in 1979. Similar trends can be observed in the Swedish data, where there is a change in a right materialist direction between 1979 and 1985. However, the change is not as large as in Denmark. In Norway, there is generally no significant change from 1977 to 1981 towards right materialism. However, there is a shift towards the right on the items 'public control' and 'nationalization'. Between 1981 and 1985 there was a general shift to a more left orientation but between 1985 and 1989 there was stability. In general, Norwegians were more left materialist in the late 1980s than a decade earlier. Shifts to the right can also be observed in the Netherlands between 1980 and 1985. In Britain, too, there was a rightwards shift between 1974 and 1979, but there was a gradual leftwards shift after 1979. According to the index, the British mass public was as leftist in 1987 as in 1974. Finally, Table 6.3 shows that there was a large and impressively stable right materialist majority during 1984-9 on each of the three items in the index for Germany. In general, during the 1970s and 1980s, the mass public in all these countries were negative about the nationalization of industry and public control of private enterprises. However, opinion was generally positive about the further equalization of incomes and standards of living. Over time, there appear to be larger differences in the emphasis on the leftist position on nationalization and the leftist position on the other items. We also computed the opinion balance using the comparative data. In Table 6.4, we present the opinion balance for the five items in Eurobarometer No. 11 ( 1979), the average opinion balance for the five items, and the mean scores on an additive index based on these items for each country. The table shows that there was a leftist majority in all eight countries in the late 1970s on all the items except 'nationalization'. The average opinion balance for the individual items across the eight countries (not shown) are +70 for 'economic equality' and 'public control of multinationals'; +46 for 'equal representation for employees'; +39 for 'government management of the economy'; and -14 for 'nationalization'. However, there are rather large differences between the countries. The ranking of the countries

according to the average opinion balance based on the five items, reveals that the average opinion balance varies from +70 for Ireland and +58 for France, down to +19 and +16 for Denmark and Britainrespectively. Using the Euroweight variable which weights for the size of national populations, 12 the correlation between country and the left-right materialist index is 0.30 (eta). Analyses of the differences between countries on the individual items reveals that the correlation is largest for 'economic equality' (0.35), followed by 'government management of the economy' (0.28), 'equal representation for employees' (0.25), 'public control of multinationals' (0.21). It is notably smaller for 'nationalization' (0.15). Two of the five items in Eurobarometer, No. 11--'nationalization' and 'economic equality'--also appear in Eurobarometer, No. 19 ( 1983), in the same format. So we can examine the development of priorities for these two items from 1979 to 1983. The average mean scores changed in the expected direction (data not shown), from -14 to -25 for 'nationalization', and from +70 to +64 for 'economic equality'. But this change was large in some countries: on 'nationalization', the change in Belgium was -41 (from +11 to -30), in France -37 (from +3 to -34), and in Ireland -22 (from +40 to +18). In Britain, however, the trend was in a leftist direction (+28). For 'economic equality', there were significant trends towards the right in Denmark (-20), Italy (-17), France (-14), and Britain (-13). The next time point when we can make direct cross-national comparisons is 1990, using data from the European Values Study. The opinion balances and mean scores for the indicators of left and right materialism are shown in Table 6.5. Although we cannot make direct comparisons between these results and those obtained for the 1979 Eurobarometer data, it is clear that the distributions in Tables 6.4 and 6.5 are very different. These differences are due--at least partly--to the fact that the question wordings are very different in the two surveys. Whereas all the items in Eurobarometer No. 11 invite respondents to 'agree' or 'disagree' with a left-worded statement, all the items in the European Values Survey are forced- choice questions. However, if we use a broader categorization of the countries, we find that our results for 1990 have much more in common with the Eurobarometer results than appears at first sight. In particular, the southern countries ( Spain, Portugal, Italy, and France) have the most left-leaning populations, while the populations in the economically more prosperous countries in the central and northern regions lean more to the right. The cross-national differences are generally large on each item and the ranking of the countries is quite similar for the five items. The grouping of countries corresponds nicely to the three-fold typology developed by Van Deth (see Chapter 3) based on GDP per capita and employment in industry. According to this typology, Denmark, Norway, Sweden, Belgium, Britain, and the Netherlands belong to the group which can be labelled advanced industrial societies with a rising and high GDP per capita accompanied by a declining proportion of the labour force working in the industrial sector. These countries, and Germany, are the countries where the population has the most rightmaterialist orientations. Two of the three countries where industrialization is continuing ( Spain and Portugal) have the most left-materialist population, while the middle-of-theroad countries ( France and Italy) have a middle position on the left-right materialist index. The only two exceptions, which precludes a perfect correspondence between the

index and the typology, are Ireland and Germany, where the population leans, respectively, more to the left and the right than expected. There are two kinds of explanation for such differences. The substantive explanation is a variant of the embourgeoisement argument: people in affluent countries with advanced welfare states have more rightist orientations than people in less advanced societies due to their higher income, and their more differentiated occupations and lifestyles. 13 The methodological explanation emphasizes the problems of explaining trends in one country because national situations change. For example, attitudes towards the nationalization of industry may depend on the extent of nationalization in a country at a particular time. Similar reasoning may apply to attitudes towards economic and social equality. People may retain the same orientation, but, as circumstances change or differ from one country to another, their response to survey questions may be different.

Social-Structural Antecedents
Left- and right-materialist value orientations are economic in character and refer, in particular, to the role of government in creating greater equality in society. We anticipate that the working class, those with low income, and those with less education are more inclined to emphasize left-materialist values, while employers, the self-employed, and the higher strata of the new middle class are likely to give the strongest support to rightmaterialist values. Consequently, we expect occupation and social class--the most direct indicators of economic interests--to be the strongest predictors of left and right materialism. In addition, as the most direct indicator of wealth, we expect income to be important. We expect these variables to be more important than education, which is less directly connected to economic values than the other hierarchical variables. 14 These relationships, however, will change. We would expect the impact of left and right materialism to decline with the development of advanced industrial society and the rise of the welfare state in Western Europe. Most social groups in these societies have experienced increased prosperity, and the gap between the income and lifestyles of the middle class and the working class has narrowed. This reduction of objective class differences may modify class conflicts and reduce the differences between the classes concerning left- and right- materialist orientations. Typically, industrial society was characterized by working-class and, to some extent, middle-class subcultures which were developed and strengthened by social networks structured along class lines. Left- and right-materialist orientations were largely derived from face-to-face contacts with family, friends, neighbours, and work colleagues. In contrast, advanced industrial society is characterized by residential mobility and increased residential heterogeneity, which tend to reduce the influence of group norms. Class-based social networks and community integration subsequently become less important for political orientations ( Dalton, Beck, and Flanagan 1984: 17-18). Institutional affiliations and loyalties have also declined in advanced industrial society. In industrial society, left-materialist values were 'learned', for example, in trade unions, and right-materialist values-- to some extent--in the churches. But the type of attachment to unions has changed in advanced industrial societies ( Goldthorpeet al. 1968). Particularly important for left materialism is the reduced sense of belonging and commitment to trade unions, the reduced sense of class solidarity, and reduced

dependence on the unions for political information. But, more generally, the erosion of group ties has undermined the influence of the traditional socialization agents for leftand right- materialist orientations; and with the political and social heterogeneity of primary networks has come a decline in structured political cues. In addition, the social mobility which characterizes advanced industrial societies tends to blur traditional class differences in value orientations. The rapid rise of the predominantly non-manual service sector has been accompanied by growing social mobility. Mobility from a working-class background to the new middle class is particularly relevant because it often implies moving from a left- to a right- orientated social environment. Some of these socially mobile individuals tend to keep their left materialist orientations, while others conform to their new environment, both of which tend to reduce the association between current socio-economic position and value orientations ( Thompson 1971; Stephens 1981: 186-91). The economic position of the new middle class is ambiguous, so probably we cannot expect to find that its members are decisively right- materialist in orientation. As a group with specialist knowledge and enjoying delegated authority in the workplace ( Goldthorpe 1982: 168), the new middle class has economic interests in economic and social differentiation. At the same time, as employees, the new middle class shares some of the same interests and problems as the working class. Although there are several theories about the value orientations of the new middle class, 15 its emergence contributes to blurring the distinctiveness of the class-based value orientations which characterized industrial societies. In particular, the new middle class is probably fairly heterogeneous in terms of left- and right-materialist values, and consequently more left oriented than the old middle class. The welfare state in advanced industrial society tends to produce more complex interests and value orientations than in typical industrial society. The income of a large segment of the population is a mixture of wage and welfare benefits, which often makes it difficult for individuals to define their political interests and their political value orientations. Thus, the welfare state also contributes to blurring class differences concerning left and right materialism ( Janowitz 1976: ch. 5). These factors imply that growing up in an advanced industrial society and a welldeveloped welfare state tends to reduce the impact of occupation, class, and income on left-right materialism. For example, coming to maturity during a period characterized by strong social tensions generated by economic and social differences generally leads to class and status differences with regard to materialist orientations, whereas being socialized during a less tense period tends to reduce these differences in values. Consequently, we anticipate that weakened correlations between the hierarchical variables and left-right materialism will be, in part, a consequence of generational replacement. Class-based differences will be greatest for generations which grew up during the pre-war period, and considerably less among post- war generations. However, period effects which influence all generations may also be present. In comparative terms, we expect these changes to be larger in the more advanced industrial societies of the northern and central regions of Western Europe because generational differences in experiences have been more evident in those countries. Finally, gender is also a relevant socio-structural antecedent of left-- right materialism.

Various explanations have been advanced of the 'gender gap' in values and voting patterns among women and men ( Hoel and Knutsen 1989). In relation to left-right materialism, we start from explanations which see changing values and attitudes among women as a result of economic and structural change. In particular, women have become increasingly independent economic actors. The earlier division of labour kept women in weaker positions in the education system and in the labour market, which has led to a concentration of women in the lower and middle levels in the education system and in the labour market. Thus, generally, we expect women to be more left materialist than men. Moreover, according to interest- based explanations, the effect of gender is likely to be indirect, making its impact via education and occupation. In a comparative perspective, we expect women to be more left materialist than men in the advanced industrial societies which are also characterized by high female labour participation. The empirical relevance of this discussion is tested in the next section, in which we analyse the impact of age, gender, education, income, occupation, and social class on left- and right-materialist orientations. 16 First, we examine the bivariate correlations between these variables and the indices for left-right materialism. Secondly, we perform multivariate analyses, testing a model in which all these variables are included, and examine possible interaction effects.

Bivariate Analyses
We start by examining the bivariate relationships in the country-specific data. The product moment correlations are presented in Table 6.6; in addition, we present the eta coefficients for age and education when eta is significantly higher than r. 17 Age and gender are seldom significantly correlated with the left-right materialist indices. Only in Germany and Britain (and Denmark on the alternative index) are there significant linear negative correlations, which indicate that the younger age groups are more leftist than older age groups. In Denmark, Norway, and the Netherlands, the eta coefficients are significantly higher than the product moment correlations in many instances. A closer examination shows that this is due to higher scores on the indices among the youngest and the older cohorts while the middle aged are more right materialist in all instances. The higher scores of the older cohorts may indicate the persistence of 'old left radicalism' among those who grew up in the 1920s and 1930s. The differences, however, are too small to analyse further. As to gender, the hypothesis that women are more left materialist than men in advanced industrial societies is not confirmed. Only in Norway and Germany do the most recent data show a modest correlation, in the expected direction, between gender and the left-right materialist indices. In these instances, women are more leftist than men, but this is not a general pattern. The correlations are considerably larger for the status variables-- occupation, class, education, household income--in all countries; and, apart from one exception, they are in the expected direction. Those with high income, high education, and those with middleclass occupations are more likely to be right materialist in orientation than those with lower income, less education, and those who belong to the working class. Except in the Netherlands, occupation and social class are the variables most strongly correlated with the left-right materialist indices. In the Netherlands, household income is somewhat more strongly correlated with the index than class and occupation. In the Scandinavian

countries, education is generally more strongly correlated with left and right materialism than household income, although this changes over time in Denmark and Norway. On the whole, education is more strongly correlated with the index than we expected--but with Britain and Germany as clear exceptions. In Sweden and Germany, the correlation with household income is relatively low. In general, the strength of the correlations between the three hierarchical variables and the left-right materialist indices have declined over time--which accords with our hypothesis. This is particularly evident in Denmark and Norway. In Denmark, between 1971 and 1984, the correlation with social class fell from -0.34 to -0.25; the correlation with education declined even more markedly, from -0.21 to -0.06. According to the alternative index for Denmark, based on the forced-choice items, the correlation with education is either positive but low, or not significant. In Norway, the decline in the correlations is largest for social class, which, between 1977 and 1989, fell by two- thirds (from -0.33 to -0.11), with most of the change occurring between 1985 and 1989. 18 The decline in the correlation for social class is somewhat smaller in the other countries, but still in the expected direction. Only in Britain do we observe more stability in the relationship. Turning to the comparative data, the results of our analyses largely confirm our findings from the country-specific data. The results are shown in Table 6.7. In the 1979 Eurobarometer data, age correlates with the left-right materialist index only in Italy, and gender is not significantly correlated with the index in any country. Education, household income, occupation, and social class are either moderately or quite strongly correlated with the index. In comparative terms, social class is most strongly correlated with the index in the Netherlands (-0.30), followed by Denmark, Belgium, Britain and Italy (about -0.20). The correlation is very modest in Ireland and France (-0.10 and 0.13), and lowest in Germany. Compared to these findings, the most important difference to emerge from the 1990 European Values data relates to gender. In seven of the thirteen countries, gender is significantly correlated with left-right materialism: in 1990, women are more left materialist than men. Moreover, as expected, this is most evident in the more northerly countries of Western Europe ( Iceland, Norway, Sweden, Germany, and the Netherlands), but is also evident in Italy and Portugal. Age, however, is only very weakly correlated with the index in most countries. There is some modest support for the hypothesis that the correlations between the hierarchical variables and left-right materialism are weakest in the advanced industrial democracies. The correlations with social class are weakest in the four Nordic countries (-0.11 to -0.18), and in Italy and Spain (-0.16). In the other countries, the correlations range between -0.20 and -0.26. For income there are larger and--according to our hypothesis--less systematic variations. Education is generally weakly correlated, and in some countries insignificantly correlated, with left-right materialism. The eta coefficients for occupation and the r for social class indicate that there are distinctive differences between various occupational groups, and between the working class and the middle class. But the fact that left-right materialist orientations follow the hierarchy of occupational positions cannot be detected directly from the eta coefficients since this indicator presupposes a nominal level measure for occupation. Thus a high

coefficient can be obtained by large differences in orientation between any of the occupational categories. However, examination of the means for different occupational categories show that left-materialist orientations decline monotonically from workers, via lower-, medium-, and higher-level non-manual workers, to the two self-employed and employer categories. We do not present these tables but report findings from a simple model based on comparing the eta and r coefficients. We have assumed a rank order between the occupational categories, from workers (1) through the three levels of non-manual workers (2, 3, and 4 respectively) to both the self-employed and employers (5). We call this variable 'hierarchical occupation'. By comparing the eta coefficients between occupation (as a nominal level variable) and the left-right materialist indices and the correlation coefficients between hierarchical occupation and the indices, and testing whether the differences are significant, we can assess whether or not hierarchical occupation captures all the correlation with left-right materialism. And we find that, apart from a few weak exceptions, the r coefficients are almost equal to the eta coefficients in Tables 6.6 and 6.7. This indicates that it is hierarchical occupation which captures all the correlation which occupation accounts for in the left-right materialist indices. Looking at all the surveys, there are two major exceptions to these models. The first relates to Denmark, Norway, and Sweden in surveys from 1985 to 1990. In these Nordic countries, middle-level non-manual workers are more left materialist than the model implies. For example, in the European Values data this group has the same mean score on the materialist index as workers in Denmark, while in Norway and Sweden the group approaches the lower-level non-manual workers in left- materialist orientation. This is an important finding which accords with previous research establishing that the radicalism of middle-level non-manual workers is generational, located primarily in post-war generations and related to public sector occupation ( Goul Andersen 1984, 1989; Knutsen 1986b; Hoel and Knutsen 1989). The second significant deviation from a linear model emerges for Spain and Portugal in the 1990 European Values data. Here, the group consisting of employers and the self-employed, and partly also the higher-level non-manual workers, are just as left materialist as non-manual workers. However, it is not clear how to explain the deviant patterns in these instances.

Multivariate Analyses
Our next step is to test a multivariate model of the socio-structural antecedents of leftand right-materialist value orientations. The causal model we employ is outlined in Figure 6.1. As the model includes some nominal-level variables and we anticipate some non-linear correlations, we use multiple classification analysis to examine the effects of the five background variables on left-right materialism. 19 We do not present the results in tables, but discuss some general trends. The analyses of the country-specific data show that the additive model generally explains 8-12 per cent of the variance in the left- right materialist indices. In Germany, however, the model explains only 3-4 per cent. There is no general trend towards the social background variables explaining less variance over time. As a rule, age and gender explain altogether about 1-3 per cent of the variance; the additional 6- 10 per cent is explained by the three hierarchical status variables. The patterns we found in the bivariate analyses of the hierarchical variables are generally little changed in these analyses. Since these variables are quite strongly intercorrelated,

some part of the bivariate correlation of occupation and household income is spurious due to the prior education variable (and the prior occupation variable in the case of household income). Even so, the effects of age and education generally increase when education is included in the second step. This is related to the fact that younger groups in most countries are more left materialist than older groups, and that more highly educated people generally have a more right-materialist orientation than people

FIGURE 6.1. Impact of socio-structural antecedents on left and right materialism with less education. At the same time, younger groups have higher education levels than older groups. When this education effect is controlled for, the direct effect of age increases. The same applies to the education variable when age is controlled for, which is more important since the controlled effect is the total effect of education. For example, in the British Election Study data, the bivariate correlations between left-right materialism and both age and education are around 0.10 (but 0.16 for age in 1974; see Table 6.6). The total effect of education increases to between 0.16 and 0.21 when age is controlled for, while the effect of age increases to between 0.17 and 0.27 when education is controlled for. We conclude from these results that education stands out as a more important predictor of left-right materialism than the bivariate relationships indicate. Compared to occupation, the effect of education is generally of about the same magnitude or even somewhat larger in Norway, Sweden, the Netherlands, and in Denmark until the 1980s. Only in Britain and Germany are the effects of education generally smaller than for occupation. Education has generally a larger effect than household income in almost all analyses. Thus the hypothesis that education has less impact than income and occupation is not confirmed in the multivariate analyses.

Interaction Effects
As regards interaction effects, the interesting questions are whether there are similar patterns across surveys from different time points, and whether there are consistent patterns across countries. We found generally few interaction effects, and in most surveys the same variables are involved--especially, age in relation to education and/or occupation. Moreover, in line with our hypotheses, there is a consistent tendency for education, occupation, and social class to be less important predictors of left and right materialism among the post-war generations. This can be illustrated by looking at the correlations between social class and education and the left-right materialist indices for those born before and after the Second World War, as presented in Table 6.8. Table 6.8 shows that the correlations are much higher for the pre-war generations compared to the post-war generations for all countries. For social class, the correlations for the pre-war generations are between -0.25 and -0.45 in most cases, while the correlations for the post-war generations are seldom more than -0.20. The pattern is even stronger for education. While there is a strong negative correlation between

education and the left-right materialist indices among the pre-war generations, the correlation seldom reaches the limit of -0.10 for the post-war generations. Denmark, however, appears to be an extreme case. Social class is very weakly correlated with the left-right materialist index among the post-war generations while quite strongly correlated among the pre-war generations. Education is not significantly correlated with the main index among the post-war generations but is quite strongly correlated among the pre-war generations. Moreover, in the 1984 data, education is positively correlated with the index in the post-war generations but shows a strong negative correlation in the pre-war generations. Opposite signs for the correlations in the two generations are also evident for the alternative index in both 1979 and 1984. The other interaction effects which appear in many countries, especially in the more recent data, involve gender in combination with age and education. In Denmark, Norway, Sweden, and Britain, more highly educated women are more left materialist than men; and in the Scandinavian countries, younger women are more left materialist than men. Among the pre-war generations, there are either insignificant differences between the sexes, or men are somewhat more left materialist than women. As to the comparative data, we report only results from the analysis of the 1990 European Values data. Generally, the results are very similar to those reported for the country-specific analyses. The five social background variables explain some 4-10 per cent of the variance in the left-right materialist index, and the three hierarchical variables explain most of this variance. The effect of education increases both absolutely and compared to the other hierarchical variables--and for the same reasons as explained above. However, the impact of education seldom approaches the impact of occupation and only approaches that of household income in about half the countries. In five countries, education has less effect than either occupation or income ( Britain, Germany, the Netherlands, Iceland, Denmark), and in three of these countries education is a less important predictor than occupation or income ( Denmark, Britain, and Germany). This is in line with our findings for the country-specific data. In two countries, gender is among the two most important predictor variables ( Iceland and the Netherlands). In Iceland, gender has the largest effect on the left-right materialist index compared to the other four background variables. This finding might be explained by the mobilization of Icelandic women by the Women's Alliance during the 1980s, which combined traditional demands for equality in the labour market and postmaterialist value orientations ( Styrkarsdottir 1986). However, our hypothesis about interest-based explanations for the radicalism of women, as revealed in the 1990 data for several countries, is not confirmed. Almost all the effects of gender are direct effects in the multivariate analysis, not indirect effects via the three hierarchical variables. In the relevant countries, the direct effects of gender are changed insignificantly compared to the bivariate correlations in all countries. Most of the significant interaction effects in these data also involve age and either occupation or education. We anticipated a generational pattern to the reduction of the correlations between the hierarchical variables and left-right materialism in advanced industrial societies. If this is correct, the differences between the generations in these countries should be

distinctive. This hypothesis is partly confirmed. Significant differences between the prewar and post-war generations are found in five advanced industrial societies ( Denmark, Iceland, Norway, Sweden, Belgium, and Germany) but not in the Netherlands and Britain. However, such differences are not present in any country which has not reached the advanced industrial stage ( Ireland, France, Italy, Spain, Portugal). In Portugal, the effects of education and social class are in the opposite direction to the findings for the advanced societies: social class and education are significantly more strongly correlated with left-right materialism among the post- war generations (-0.27 and -0.22 as against 0.16 and -0.08).

Conclusion
In this chapter we have brought together research findings on left-right materialism in order to further our understanding of the processes of value change in advanced industrial societies. First and foremost, we found strong evidence of left-right materialist orientations at the level of the mass public. Such orientations are manifest in all the surveys examined, and there is evidence from many countries that the items which tap left and right materialism are strongly constrained. There is no tendency for left-right materialist items to be less constrained in advanced industrial societies, although there is strong evidence for class de-alignment. In other words, left-right materialist orientations appear to be highly constrained in advanced industrial societies, although their anchoring in specific social groups seems to have declined. Our findings with regard to trends provide some support for the notion of a shift to the right around 1980. However, there is also evidence that priorities among the public changed in a left-materialist direction during the 1980s. But these changes are small compared to the large differences between the most advanced and the less advanced countries in Western Europe. People in the less advanced countries are more left materialist than people in the more advanced countries. Analyses of the relationship between background variables and left- right materialism indicate that these value orientations are principally anchored in hierarchical variables such as social class, education, and income. However, these factors appear to be less important over time, particularly among the post-war generations. Gender, however, appears to be a predictor of moderate importance in some advanced industrial societies: women are more left materialist than men. We can conclude, then, that left-right materialist value orientations are central features of political beliefs in West European societies. They mark out oppositions in the more advanced industrial societies no less than in the less advanced societies. This indicates a stability and continuity in political life in Western Europe which is often over- looked in studies of values and value change.

NOTES
1. This conceptualization differs from Flanagan ( 1987: 1304-6) division of the mass public into materialists and non-materialists, emphasizing economic and non- economic issues. For materialists, the central political conflicts are related to the 'old' right and the 'old' left alignments of class politics. Non-materialists are divided along a 'new left-new right' cleavage related to libertarianism versus authoritarianism.

2. Issues about 'socio-economic distribution', 'government control and planning', and 'individual initiative and freedom' dominate the agenda in nearly all democracies. Almost half of all issues (43%) relate to these domains, and there is no tendency for these issues to decline in salience during the post-war period. These issue domains constitute about 50% of all issues in the periods 1945-54 and 1965-81; in the period 1960-75 their dominance was less pronounced (35-40%). See Budge and Farlie ( 1983: ch. 2). 3. Most of the data sets used were made available by the Norwegian Social Science Data Services ( Bergen) from the ICPSR ( Ann Arbor, Michigan), the Danish Data Archives ( Odense), the Swedish Social Science Data Service ( Gteborg), and the ESRC Data Archive ( Essex). The Norwegian Election Studies were made available directly from the Norwegian Social Science Data Services. The Dutch Cultural Change Studies were made available by Steinmetzarchief ( Amsterdam). The German IPOS surveys were made available by Oscar Gabriel. Data from the 1990 European Values Study were made available by Ola Listhaug. 4. Details on the construction of the indices, and the wording of the questions used in the analyses, can be obtained from the author. 5. Only factors with an eigenvalue above 1.00 were rotated. This criterion was supplemented by the scree test, in which factoring stops when the graph of eigenvalues shows the eigenvalues beginning to level off ( Kim and Mueller 1978: 44-5). In all analyses, units with missing values ('don't know' answers and 'not answered') have been included by assigning the mean score of the relevant variable. All variables have been recoded so that the presumed left- and right-materialist alternatives have been assigned high and low values, respectively. 6. For Denmark, the main index covers 1971-84; the alternative index, based on completely different items, covers 1979-87. The alternative index for Norway consists of three of the four items used in all the surveys; one of these four items was not used in the 1989 survey. The index for the 1988 Swedish survey consists of some items which are different from the index based on the previous surveys. 7. Items relating to the misuse of welfare benefits or that too many people rely on welfare benefits are deliberately excluded. The welfare items seem to tap a separate dimension. 8. The eigenvalue of the second factor in the German sample is 1.15. In Denmark there is more than one factor with an eigenvalue higher than 1.00, but the eigenvalue here is 1.04 for the second factor. In both countries all variables have quite high loadings in the expected direction on the first unrotated factor (the lowest loading in the German sample was 0.48). 9. The first factor in each of the six countries had eigenvalues of 1.46 or higher; the second factor had eigenvalues between 1.04 and 1.17; the third factor had eigenvalues slightly lower than the second. Graphs show that a two-factor model fits in each country. 10. The explanation for this deviation might be that these items are neither so clearly related to the economic domain nor are they prescriptive beliefs, i.e. values. The item 'people can only accumulate wealth at the expense of others' is evaluative rather than prescriptive; 'more freedom for individuals' is not directly related to the economic domain, although

'individual freedom' is an economic argument. 11. The calculations are based on relative distributions where 'don't know' answers and missing cases have been excluded. Neutral positions ('do not agree with either' or 'does not matter') were included in calculating the percentages but not in calculating the opinion balance. When the items are scales where respondents were asked to indicate their position, we present the mean as well as the opinion balance based on collapsing all categories to the left and to the right of the middle position. If a middle category exists, all categories to the left and all categories to the right are summed. The summed percentage for the right materialist responses are then subtracted from the summed percentage for the left materialist alternatives. The 'neutral' category is included in calculating the percentages. For scales where the right-materialist alternatives have the highest value, the mean scores have been recalculated as if the leftist alternative had the highest score. 12. The Euroweight variable is used here because we consider the population of the eight countries as the universe, and examine how much of the total variance in the left-right materialist index and in the various items is explained by 'country'. 13. Inglehart has used a similar line of reasoning in analysing Eurobarometer, Nos. 11, 16, and 19. He explains the relatively large cross-national differences in left materialism on the grounds of diminishing marginal utility at both the societal and the individual level. See Inglehart ( 1990: 255-7). 14. Education may reflect cognitive factors in addition to class interests and value orientations to a larger extent than occupation and income. 15. The 'new middle class' refers to non-manual or white-collar workers. For a discussion of the structural characteristic of the new middle class, see Giddens ( 1973: ch. 10). For views about the political value orientations of the new middle class, see e.g. BruceBriggs ( 1979). Goldthorpe ( 1982), and Gouldner ( 1979). Ross ( 1978) presents a comprehensive review of influential French, Marxist-inspired, theories of the political orientations of the new middle class. 16. The variables occupation and social class are the respondent's own occupation and social class; income is household income. As occupation is coded differently in the several data sets, we constructed an occupation variable with the following categories: (1) workers; (2) lower-level non-manual workers; (3) middle-level non-manual workers; (4) higherlevel non-manual workers; (5) self-employed in business and professions; (6) selfemployed in primary sector (farmers, fishermen). The eta correlations between the value indices and occupation are based on these categories. The eta coefficient is based on the calculation of the total variance which occupation explains in the indices; i.e. occupation is treated as a nominal- level variable. In addition, we present the correlation between the value indices and a dichotomous variable for social class: (1) workers, (2) middleand higher-level non-manual workers and the two self-employed groups. The lower-level non- manual group is deliberately not included since it is unclear whether it should be placed in the working class or the middle class. 17. Such deviation indicates that the relationship is non-linear. 18. Since an alternative item is included in the index based on data from the 1989

Norwegian election survey ('market economy'), it may be argued that the indices comprise different items in 1977, 1981, and 1985 compared with 1989. An alternative index containing only the three items asked in each survey (see Table 6.3) provides the following correlations with social class for 1977, 1981, 1985, 1989 respectively: -0.30, 0.24, -0.26, and -0.16. These results are weaker than for the main index, but they confirm the impression that a major reduction occurred in the strength of the correlation with social class between 1985 and 1989. 19. The method employed is the 'effect change design', not the traditional path or effect transmittance design ( Hellevik 1989: ch. 5). The 'effect change design' means that the inclusion of variables starts with the prior variables and subsequently includes the intervening variables, but with the same dependent variable (here the left-right materialism index) in each step.

Slide 1:

2.

Lecture and class:

GV101 Lecture 5 Political Preferences, Ideologies and Dimensions It is important to recognize that the terms left and right are not ideologies in themselves, but dimensions of measuring and categorizing ideology. Use word political actors In political science we think about political preferences as any views that voters or parties or politicians or beaurecrats or central banks or judges or any kind of actor in the political process, the positions that those actors have on policy questions. Its not just economic questions but generally policy questions

Slide 2 and 3: Are political preferences malleable, do we have a particular set of views of one point in time and then change them? We have one view at one point regarding one policy view that changes back and forth or are they more concrete, frozen in time?

Slide 4: Key Questions for Today: Why do people have different political preferences or political views over a policy? How fixed are these preferences? What makes them change?

How can peoples preferences on a range of political issues be summarised in two dimensions or even a single dimension? How can we as political scientists simplify the range of political preferences that people have? What is political ideology and how does it relate to preferences? In political science we think about ideology as a group of views, a set of preferences and policy decisions. Slide 5: Outline 1. Where do political preferences come from? 2. Dimensions of politics and how these dimensions capture or summarize political preferences. 3. Locating political ideologies How do we locate political ideologies within this political space? 4. Some examples from across the world Slide 6: Where Do Political Preferences Come from? The classic idea is that peoples views or preferences on a policy is fundamentally related to economics Economic Class ? Karl Marx (mid 19th century) Individuals political interests are determined by their relationship to economic production (their role in the economic system in terms of their employment, their relationship to the means of production) Ex. Workers want redistribution of wealth while capital owners would Where these actors are located in the economic process. ATTITUDES Workers vs. owners of capital / land Max Weber (early 20th century) Individuals political interests are determined by their relationship to economic consumption Wealth and power are key factors Much more to do with your role in economic consumption. Ex. Higher income groups have certain political views as a result of the fact that they have large amounts of wealth. Simon Hix presents the example that a plumber or electrician, in Marxist views, would be workers, but in Webers views it would be depending on how much their income is. This means that even economics does not always have a clear connection to how you think of peoples political preferences, sometimes they can point some ways, other times they can point others. Slide 7: More Recent Versions of Economic Determinism

Given how complex, economically, society has become over the last recent years.

Ralf Dahrendorf (1959) In modern capitalist society there are many more complex economic goals, particularly with the growth of a rising new middle class. more complex network of economic interests. Growing economic fragmentation Could be closer to lower income or higher income levels depending on the political issue => Emergence of new middle class Herbert Kitschelt (1994) Herbert Kitschelt in the 90s starting to think about how preferences of voters who traditionally would vote for centre-left parties, how their actions have changed with change of roles in the economic system pointed mainly at the new middle class/new service sector. Preferences are determined by the type of jobs people do. Some new ones and argument different values: People processing jobs liberal values MORE SO THAN Data processing jobs conservative values CRITICISM PRESENTED BY SIMON HIX: YOU DO NOT KNOW WHICH WAY THE CAUSAL RELATIONSHIP IS GOING PEOPLE WHO ACQUIRE THOSE JOBS COULD ALREADY HAVE THOSE VIEWS OR IT COULD BE THE INFLUENCE OF THE JOB But what we do know is that there is a correlation in certain roles that people take in the new economic structure and the type of political preferences they have, not just economic preferences but other preferences like attitudes on social issues. Slide 8: Example: Current British Social Class Categories Office of National Statistics 1. Higher professional and managerial (A) 2. Lower managerial and professional (B) 3. Intermediate occupations (C1 and C2) 4. Small employers & non-professional self-employed (C1 and C2) 5. Lower supervisory and technical (C1 and C2) 6. Semi routine occupations (D) 7. Routine occupations (D) 8. Long term unemployed (E) => We will look at class voting next week It is not simple, it is not straightforward, to assume that we can assign people to a particular group in society and expect them to have a particular set of political attitudes purely by virtue of their economic role. It might have some influence, but not 100 percent accurate Many other factors influencing political preferences

Slide 9: From Economic Class to Post-Materialism Ronald Inglehart (1977) calls it post-material values. Materialism => focus on economic and physical security In the 1970s, he started to pioneer cross-country surveys to try to tap in to peoples political attitudes. Result: their economic income or social class wasnt a very accurate thing to reflect peoples political attitudes, particularly among higher-income groups. And so Inglehart, looking at this, came up with a new conception of how to think about it. He called it the contrast between, the differences between, materialist attitudes, where people primarily focus their political attitudes on economic security, prosperity, and physical security. Once they have those things, once society has reached a certain level of economic wealth and physical security, then people start to think about other values, what we call post-material values. They start to think about self-expression, wanting to be represented in the political process, attitudes towards equality in society, attitudes towards environmentalism, how to treat the environment, nuclear power, gender equality. These things in a sense, are secondary with the more primary needs being directly economic or physical security. His argument was that once a society reaches a certain level of wealth or wellbeing, the focus on material politics starts to shift and people start to focus on postmaterial politics. Post-materialism => focus on personal autonomy, self-expression, intellectual satisfaction e.g. civil liberties, gender equality, minority rights, environmentalism, anti-nuclear, anti-war etc. Hypothesis: economic wealth + mass university education -> shift from materialist to post-materialist values from one generation to next Slide 10: Simon Hix argues that this argument led to political scientists considering where individuals political attitudes come from. So, non-economic factors now matter ... (again?) Religion Gender Urban-rural Sexual orientation Family Language Education Nationality Ethnicity Media usage Age etc MY NOTE: JUST BECAUSE THESE INFLUENCE POLITICAL IDEOLOGY DOES NOT NECESSARILY MEAN THAT THEY ARE ACTUALLY FORMALLY USED TO DETERMINE LEFT-RIGHT CATEGORIES.

Slide 11: Left, Right and Left-Right Left-Right is a very useful term, as it is both a dichotomy and a continuum The aim is to finding out what peoples political preferences are and conceptualizing them in a simple way. But, Left and Right have no inherent substantive meanings. Their meanings are determined by the particular time and space. Their origins came from whether you sat on the left or right of the speakers chair during the French national assembly after the French revolution. 19th century Europe About freedom of the individual versus authority of the state, freedom of religion versus authority of the Catholic Church. It had almost nothing to do with economics. Left = individual freedom, religious tolerance, democracy, more radical usually Right = state and religious authority, protect landed property 20th century Europe shifted, where we started thinking about left and right as concepts that captured an economic conflict. Left = workers rights, wealth redistribution, higher taxes Right = economic freedom, free markets, lower taxes how we conceptualize politics as a dimension of politics changes over time; there is nothing inherent about what the left represents and what the right represents, it changes across time and across country. But its a very useful abbreviation at any time point in any country to conceptualize what the views are. Slide 12: Most modern political scientists think in two dimensions: 1) Intervention-Free Market (an economic Left-Right) How far should the state (or the political majority) be able to interfere in individual economic freedoms/choices? e.g. taxation, welfare spending, market regulation etc. Left The majority should have the right to do that, should be able to raise taxes and redistribute wealth. Right Advocates more individual choice, less intervention, less regulation of taxes. 2) Liberty-Authority (a social Left-Right) How far should the state (or the political majority) interfere in individual social freedoms/choices? AND HERE IS FLIPS e.g. abortion, gay rights, immigration etc. Left free to make own decisions Right Do have the right

=> The meaning of an integrated Left-Right dimension depends on how these two underlying dimensions relate to each other in a particular country in a particular period Slide 13: Locating Political Ideologies Ideology => a set of ideas about how you think the world should be organized, a worldview (in other words, a set of political preferences); a particular set of values that seem to fit together in a certain logic. from different sources. Liberalism EMERGED IN 19th CENTURY Social freedom + Economic freedom Conservatism (19th century tended to oppose liberalism by defending traditional European views), Church, ruling elites. Social authority + Economic freedom Led to economic conflicts but emerged in favor of economic freedom but social authority, ex. Thatcher Clause 28, anti-gay in classroom. Socialism Social authority (originally) + Economic intervention Recently: more socially liberal. Environmentalism most recent. Social freedom + Economic intervention Slide 14: from Herbert, where to put social groups in economy based on 2 dimensions. Slide 15: Slide 16: Slide 17: Slide 18: USE DIAGRAMS IN APPENDIX TO SHOW POINTS. Slide 19: MY NOTE: THE WORD IS ARE. Meaning of Left-Right Across Countries Russell Dalton (2006) WORLD VALUE SURVEY INGLEHART Question in World Values Survey: In political matters, people talk of the Left and the Right. How would you place your views on this scale, generally speaking? 1 -- 2 -- 3 -- 4 -- 5 -- 6 -- 7 -- 8 -- 9 -- 10 Left Right Research question: What other political preferences correlate with peoples Left-Right position across different regions of the world? Slide 20: TABLE ---

Conclusion: Where you place yourself on the left-right scale is context dependent. PUT TABLE IN APPENDIX, GET FROM DALTON READING. He uses example of South Korea, use example from the Middle East of what determines which scale youre on. Slide 21: In sum: Political preferences are actors (e.g. voters, parties etc.) views on political issues Because preferences on many issues are related to each other, preferences can usually be summarised in a single- or two-dimensional space Modern political scientists often use two dimensions: (1) economic intervention vs. free market (2) social liberty vs. authority The Left-Right is often used as a useful summary of positions on these two underlying dimensions But, the substantive meaning of Left and Right varies across time and country

Class notes:
Class notes: week 5:

Core question of this session: What are the strengths and weaknesses of the Left/Right distinction of political preferences? Why do the content and relevance of Left/Right distinctions vary across countries and over time? Which factors help to explain individual political attitudes? Why are we interested in mapping political views? Inter alia because: Mapping political views helps to identify political issues that are of particular relevance to the individuals in a given society and the manner in which political discourse is organised (cf. Dalton 2006); the distribution of political preferences among voters affects party strategies and their chances to win elections (cf. Schofield, Miller and Martin); in the same vein, patterns of political preferences affect the relevance of different political actors and the possibility and direction of policy change (cf. Hix, Noury and Roland). From the lecture Political preferences are actors views on political issues.

 Modern political scientists often use two dimensions [to map these preferences]: (1) economic intervention vs. free market; (2) social liberty vs. authority. The Left-Right distinction is often used as a summary of positions on these two underlying dimensions.

The meaning of Left and Right in the contemporary Western discourse

Left

Right

Economic dimension

Social dimension

Locating Political Ideologies Your task: Locate these political ideologies (as presented in the lecture) on an economic and a social Left-Right dimension: Liberalism Social freedom + Economic freedom Conservatism Social authority + Economic freedom Socialism Social authority (originally) + Economic intervention Environmentalism Social freedom + Economic intervention Your task Split into groups of three and briefly address the following questions: What do Left and Right in your adopted countries (primarily) stand for? Why do the content and relevance of Left/Right distinctions vary across countries and over time? Defining ambiguous concepts: Left and Right By its very nature the leftright scale, having no fixed definition in terms of its substantive policy content, is likely to vary in

meaning as we move from country to country. (Benoit and Laver 2006:131) For older citizens ... Left means support for social programs, working-class interests and the influence of labor unions. Right is identified with limited government, middle-class interests, and the influence of the business sector. Among the young ... Left means opposition to nuclear energy, support for sexual equality, an internationalist orientation, or endorsement of multiculturalism. Right means a preference for traditional lifestyles, moral values, and a traditional sense of national identity and interest. (Dalton 2006:4) The varying content of Left-Right preferences Economic polarization along the Left/Right scale is strongest in the advanced industrial democracies. ... Europe was, after all, the birthplace of Marxism, socialism and the ideological movements that generated the traditional meaning of Left. On the other hand, longitudinal evidence and generational comparisons suggest that economic controversies are moderating in Europe. A generation ago, economic polarization in Western Europe probably was even sharper. Economic polarization also is a significant basis of political cleavage in East Europe. Certainly East Europeans were familiar with this ideological debate, since it was embedded in the communist political order. But this cleavage was blurred under the old regimes, since there was no ideological competition between alternative positions. ... However, after the transition to democracy, the traditional Left/Right economic alignment is emerging in East European party systems and is only slightly weaker than in the West (Kitschelt et al., 1999). When one moves beyond Europe, however, economic controversies over the role of the state, income inequality, and market competition are relatively independent of Left/Right orientations. In Latin America, for instance, polarization on economic issues is barely linked to Left/Right attitudes. Economic polarization is also substantially weaker in Asian democracies, Arab nations, and African nations. Marxism, and its attendant political controversies, may have structured political conflict in Europe, but such controversies are less central to the ideological framework of mass publics in the rest of the world. (Dalton 2006:12-13) The varying content of Left-Right preferences Our theoretical expectation for postmaterial issues is quite clear. Since these issues supposedly enter the political agenda once

socio-economic development addresses traditional economic and security concerns, postmaterial issues should be a significant source of ideological orientations only in advanced industrial democracies. Inglehart's (1977, 1990) analyses of postmaterial values support such a prediction. ... Rather than postmaterialism being the distinct basis of ideological polarization in advanced industrial democracies, these values are strongly related to ideology on a global scale. In most regions, postmaterialists are disproportionately Leftist, and this is the pattern in advanced industrial democracies, Latin America, Asian democracies, and Arab nations. Among these four regions, the relationship is actually weaker in the Western democracies. At the same time, postmaterialism has the opposite impact on ideology in Eastern Europe and Africa. In these two regions, postmaterialists are more likely to locate themselves on the Right end of the scale. (Dalton 2006:14) ... citizens in developing nations are more likely to orient themselves to politics in terms of religion or national identity issues, which generally outweigh economic issues as a basis of Left/Right identity. (Dalton 2006:16) See also Benoit and Laver for some more empirical evidence on the varying content of Left-Right preferences across countries. Factors influencing the meaning and relevance of Left/Right distinctions Inter alia: Changing salience of particular policy dimensions (cf. Benoit and Laver); length of democratic history (cf. ibid.); social modernisation (cf. Dalton 2006 Dalton lists a variety of reasons why social modernisation affects the meaning and relevance of Left/Right distinctions; make sure to read them carefully); design of formal political institutions, in particular electoral systems (cf. Schofield, Miller and Martin); degree of party system polarisation (Dalton 2011). Is Osama bin Laden left-wing or right-wing? How about Robert Mugabe? ... Consider Fidel Castro. He persecutes homosexuals, crushes trade unions, forbids democratic elections, executes opponents and criminals, is a billionaire in a country of very poor people and has

decreed that a member of his family shall succeed him in power. Is Castro left-wing or right-wing? ... Some owl, from the Economist, I think, wrote, 'The right believe in economic freedom, the left in personal freedom.' Very well, a key economic freedom is free movement of labour and a key personal freedom is the right to own a firearm. So, does a right-wing Englishman believe people from Africa should have unlimited right to enter Britain looking for work, and does a left-wing Englishman believe all Britons should have the right to carry revolvers? ... Is internationalism more right-wing than nationalism? Internationalist people and organisations include Adam Smith, Coca Cola, Karl Marx, McDonald's, Trotsky, Microsoft, the United Nations, Toyota and the World Trade Organisation. Those opposed to internationalism include Hitler, the anti-globalisation demonstrators, Verwoerd, Stalin and Naomi Klein. ... Consider personal habits. Is it leftwing or right-wing to be a vegetarian, teetotaller and animal lover (Hitler)? To enjoy boxing and shooting animals (Nelson Mandela)? ... I could go on and on. The fact is that the terms 'left-wing' and 'right-wing' are meaningless. ... People can neither define 'left' and 'right' nor recognise left-wing or right-wing philosophies when they see them, because they never see them. They do not exist. Andrew Kenny, The Spectator, 5th February 2005.

Percent of public placing selves on Left/Right scale. Source: Dalton 2006:7.

Your task Split into groups of three and briefly discuss the following question: How useful is the Left/Right distinction of political preferences? When discussing this question, consider in particular: The context-dependent nature of the Left/Right distinction;

 methodological issues regarding the measurement of Left/Right preferences [see especially the Kroh reading for this week, or Simons argument that the Left/Right distinction is both a continuum and a dichotomy]; and whether two dimensions (an economic and a social one) are enough to understand and classify political attitudes.

3. Excerpts from Andrew Kenning the End of Left and Right

Kenning refers to the formation of the terms left and right as the great intellectual curse of the French Revolution, which has crippled political thought for more than two centuries. From the beginning it was an infantile notion that replaced rational argument with a playground division into two gangs who understood nothing clearly except how much they hated each other. Nobody has ever been able to define the beliefs of 'left' and 'right' or the differences between them. Quoting someone from the Economist: 'The right believe in economic freedom, the left in personal freedom.' Very well, a key economic freedom is free movement of labour and a key personal freedom is the right to own a firearm. So, does a right-wing Englishman believe people from Africa should have unlimited right to enter Britain looking for work, and does a left-wing Englishman believe all Britons should have the right to carry revolvers?

Kenning later makes reference to seemingly simple political habits Is it left-wing or right-wing to be a vegetarian, teetotaller and animal lover (Hitler)? To enjoy boxing and shooting animals (Nelson Mandela)?

The fact is that the terms 'left-wing' and 'right-wing' are meaningless. ... People can neither define 'left' and 'right' nor recognise left-wing or rightwing philosophies when they see them, because they never see them. They do not exist.

4.

Social Modernization, Dalton

The twin pillars of economic and religious cleavages remain important in European states; cultural values and nationalism provide stronger bases of ideology in Asia and the Middle East. As Bell suggested, social modernization does seem to transform the extent and bases of ideological polarization within contemporary societies. In the halcyon days of the early 1960s, Daniel Bell (1960) made a provocative claim about the End of Ideology. Bell maintained that In the Western world, therefore, there is a rough consensus among intellectuals on political issues: the acceptance of a Welfare State; the desirability of decentralized power; a system of mixed economy and of political pluralism. In that sense, too, the ideological age has ended (p. 373). He also claimed that while ideological debates had been exhausted in the West, new ideologies were emerging and driving politics in Asia and Africa. MY VIEW: Its not that the polarization decreased due to economic developments, but that it shifted towards more social issues. So basically it went: social, economic (industrial revolution), social. As the old cleavages apparently waned, however, new forms of political cleavage emerged in the advanced industrial democracies. This created a new debate over whether ideology was ending, or merely shifting the content of ideological competition. Most notably, Ronald Inglehart (1977, 1990) and others have argued that new types of postmaterial issues were repolarizing Western publics, stimulating new conflicts over environmental quality, gender equality, and life style choices. The rise of Green parties and other social movements injected new ideological debates into the politics of advanced industrial democracies. More recently, a New Right reaction to these issues has further polarized contemporary politics. This debate has largely focused on Bells claim about the End of Ideology

in the West, but not his comparison between the developed and developing world. In fact, there has been little systematic research on how social modernization may have affected the bases of ideological cleavage as Bell suggested. Our paper takes a broad international view of the End of Ideology debate. The End of Ideology thesis argues that ideological differences will moderate as nations experience social modernization. This occurs because increasing affluence provides the resources to address some of the most pressing social needs that have long been a primary goal of government: providing economic sustenance and security. In addition, the increasing complexity of a developed industrial society leads to a more differentiated social structure, more complex patterns of social and economic relations, and more interactions between members of the polity. Black/white political differences might become muted into shades of grey by the complex structure of modern societies and cross-cutting interests. Indeed, this was implicit in much of the literature on the impact of modernization on political conflict. We test the End of Ideology and Postmaterial hypotheses with a new wave of data from the World Values Survey (WVS). The fourth wave of the WVS includes an unprecedented set of nations spanning the six inhabited continents and representing the diverse cultural, political and economic variations across nations. Over 70 nations are available for analysis, an unequaled resource in the social sciences. We develop our analyses in several steps. First, we discuss the argument and logic underlying the End of Ideology hypothesis, and the rival Postmaterial hypothesis. Second, we use cross-national aggregate data to test the core hypothesis of whether ideological positions are less polarized in advanced industrial democracies, while continuing to divide the publics in the developing world. Third, we examine the whether the correlates of ideology and hence the meaning of ideological cleavage vary systematically across nations. These empirical findings provide the basis for discussing the relationship between social modernization and ideology, and the likely consequences of this relationship for contemporary political systems. The End of Ideology thesis Daniel Bell premised the End of Ideology hypothesis on a set of social changes that were transforming Western democracies. One factor was the tremendous economic progress of the mid twentieth century, and the concomitant transformation of the employment patterns and living conditions. In a later work, Bell (1973) articulated this position in more detail, forecasting the emergence of post-industrial societies as the endproduct of this transformation. The development of the welfare state, expanding employment in the tertiary sector, increasing geographic and

social mobility all contributed to the blurring of traditional ideological divisions. Similarly, scholars such as Lane (1965) and (Beer, 1978) discussed how the age of affluence would lessen attention to the economic controversies of the past and lead to a new period of political consensus. A second element of Bells (1960) thesis was that modern societies were steadily becoming more secular. This trend was lessening the moral content of political debate. In addition, he argued that political ideologies had traditionally had to compete with religion for public support. As religious attachments moderated, so also could the emotional attachments to a political position. Religion remains an important element in many Western democracies, but its influence has waned as a consequence of social modernization (Norris and Inglehart, 2004). Moreover, in contrast to the West, Bell held that ideology continued to be a driving political force in developing nations. He concluded that the extraordinary fact is that while the old nineteenth-century ideologies and intellectual debates have become exhausted [in the West], the rising states of Asia and Africa are fashioning new ideologies with a different appeal for their own people (Bell, 1960: 373).1 He emphasized the importance of nationalism, ethnicity, Pan-Arabism, and other ideological conflicts in the developing world. In a recent update to his initial book, Bell (2000) stressed the role of ethnicity and nationalism as the source of division in developing nations. At the same time, one might add that the struggles over economic well-being and individual rights still existed in the developing world, even if advanced industrial democracies had made substantial progress in addressing these concerns. In contrast, the Postmaterial Hypothesis challenged the accuracy of the End of Ideology thesis as applied to advanced industrial societies. Ronald Inglehart (1977,

1A

contrasting position questions whether publics in lower income and less developed nations were sufficiently engaged to have ideological orientations toward politics (Almond and Verba, 1963; Mainwaring, 1999: 131). In these societies, politics might primarily involve elite competition, since many voters are politically unaware and do not develop an ideological position.

1984, 1990) agreed that the traditional bases of ideological cleavage were eroding, especially visible in the class cleavage and the economic values underlying this framework. Inglehart explicitly stated that there was a withering away of Marxian politics (1990: ch. 9). In Marxs place, however, new political controversies over life style issues, quality of life, and self-expression were emerging in postindustrial societies. This directly led to research on the changing content of Left and Right in these societies (Inglehart and Klingemann, 1976; Inglehart, 1984; Fuchs and Klingemann, 1989; Knutsen, 1995; Evans et al., 1996). For older citizens,

these terms appear largely synonymous with socioeconomic polarization: Left means support for social programs, working-class interests and the influence of labor unions. Right is identified with limited government, middle-class interests, and the influence of the business sector. Among the young, however, postmaterial or libertarian issues provide a new basis of ideological identity. Left means opposition to nuclear energy, support for sexual equality, an internationalist orientation, or endorsement of multiculturalism. Right means a preference for traditional lifestyles, moral values, and a traditional sense of national identity and interest. Public opinion surveys from several Western democracies demonstrated the existence of these two separate dimensions of cleavage, and the generational patterns implied by the Postmaterial thesis (Inglehart, 1984; Evans et al., 1996). In summary, the Postmaterial hypothesis holds that ideology did not end, but the content of ideology changed with social modernization. A second critique of the End of Ideology hypothesis involves Bells assumptions about the developing world. He wrote at a time when decolonialization and national independence movements were transforming the Third World. During this period, nationalism and independence were powerful symbols in these nations. Furthermore, the political ideologies of these regimes were often portrayed in stark terms because the superpowers used the developing world as a surrogate for direct competition. Thus, political elites often stressed communist or Western orientations. However, it was less clear whether these geopolitical choices motivated the thinking of the populace. And with the end of the Cold War, this polarization also quickly dissipated. Research also implied that these publics have limited ideological orientations and are relatively unengaged in politics (e.g., Almond and Verba, 1963; Pye and Verba, 1965). The limited empirical research on ideological orientations among Third World publics has not resolved these contrasting images (e.g., Nathan and Shi, 1996; Mainwaring, 1999; Shin and Jhee, 2004). In short, the nature and content of ideological attachments among publics in the developing world is imprecisely understood. We examine these rival theories using the data from the newest wave of the World Values Survey. We first test whether ideological extremism varies systematically with socio-economic characteristics of the nation. Does social modernization moderate political polarization? Second, we examine the correlates of Left/Right attitudes within nations to determine whether the content of ideological orientations also changes with social modernization. The World Values Survey The World Values Survey (WVS) is a worldwide investigation of sociocultural and political change. An international network of social

scientists at leading universities and research centers around the world conducts the WVS, coordinated by a directorate board. The board develops the questions to include in the survey, and this is translated into the national language by each research institute. The WVS spans four waves since 1981. This paper is largely based on data from the fourth wave of the WVS, which includes representative national surveys from more than 65 societies on all six inhabited continents. Virtually all the nations of Western and Eastern Europe are included, along with most other OECD democracies. The unusual feature of the fourth wave is the expansion of the project to a set of developing nations that were previously not included in international survey projects. The East Asian surveys, for example, include Vietnam, Singapore, and Indonesia, as well as the more commonly surveyed nations of China, Japan, Korea, and the Philippines. The WVS includes a new set of Arab nations, such as Iran, Jordan, Egypt, and Morocco. There are additional surveys in several sub-Saharan nations and extensive surveying in Latin America. The data from the fourth wave nations, and from some additional nations surveyed in the third wave, have been released as a merged cross-national data file (Inglehart et al., 2004).2 The WVS provides a unique resource to look at broad questions of social modernization and the comparison between developed and developing nations. Measuring ideological position Political scientists may disagree on the content and nature of ideological competition, but there is general agreement that some ideological framework or core political identity is used to organize political discourse in a nation and the individual belief systems of the citizens. Typically, such broad orientations are described in terms of Left/Right attitudes (Fuchs and Klingemann, 1989; Barnes, 1997). Political issues are discussed or summarized in terms of Left/Right or liberal/conservative philosophies, parties are summarized by their position along this continuum, and politicians are evaluated by their political tendencies. The ability to think of oneself in Left/Right terms does not imply that citizens possess a sophisticated conceptual framework or theoretical dogma. For many individuals, Left/Right attitudes are a summary of their positions on the political issues of greatest concern. Survey data from developing nations is limited. However, Huber and Inglehart analyzed elite perceptions of political cleavages across 42 nations and concluded: The leftright dimension, then, can be found almost wherever political parties exist, but it
2 Additional

information on the nations included, survey methodology, questionnaires, and other technical points of the World Values Survey are available in Inglehart et al. (2004) or on the projects website: www.worldvaluessurvey.org.

as an amorphous vessel whose meaning varies in systematic ways with the

underlying political and economic conditions in a given society (1995: 110). The World Values Survey adopted the common question of asking respondents to position themselves along a ten-point scale, where 1 is labeled as Left and 10 is labeled as Right: In political matters, people talk of the Left and the Right. How would you place your views on this scale, generally speaking? 1 2 3 4 5 6 7 8 9 10 Left Right In some nations, the political discourse is different, and so alternate wording is used for the endpoints of the scale.3 We are less concerned with the labeling of the scales poles, as long as these labels reflect the shorthand of political polarization in the nation. When there were deviations, the national teams made these decisions to produce maximum comparability to the theoretical construct. We will use the term Left/Right as a shorthand for this scale in the World Values Survey, although we recognize the exact terminology for this scale may vary in some nations. This operationalization of ideology is certainly different from the more rigorous meaning embedded in Bells writings. Mass publics typically lack the type of strict ideological reasoning that exists among political elites and intellectuals. Instead, we are tapping a framework of political thinking that is closer to Anthony Downs (1957) conceptualization of Left/Right as a cognitive framework for orienting political debate and mass beliefs. Some indication of the basic validity of such orientations comes from the large majorities of the public who can position themselves on this scale (see Table 1). Across the 75 nations in our analyses, roughly three-quarters of the public place themselves on this scale. This scale appears most problematic in the Arab and Middle Eastern nations. Relatively small proportions in Pakistan (12 per cent), Morocco (27 per cent), Jordan (36 per cent), Algeria (46 per cent), and Iran (59 per cent) locate themselves in Left/Right terms. The methodological appendix for the Jordanian survey, for instance, flags this as a problematic question in their survey. The terms Left and Right lack relevance in the Middle East, where political divisions certainly exist but are expressed using different terminology. In addition, the percentage of the public who can locate themselves on the Left/Right scale tends to be lower in some of the new democracies of Eastern Europe where the lines of political competition are still forming. On the whole, however, most citizens in most nations can describe themselves in Left/Right terms.
3 For

instance, the US survey used the terms Liberal and Conservative; the Japanese survey used progressive versus conservative; Vietnam used continuing the reform and no reform. The question was not asked in the surveys in China, Singapore, and Egypt.

Table 1. Percent of public placing selves on Left/Right scale

Cross-national comparisons of ideological polarization The essence of the End of Ideology hypothesis is that social modernization moderates ideological polarization, providing a more centrist and moderate political debate. The tensions in advanced industrial societies are not between survival and starvation or between opposing moral absolutes, but between more modest differences in political means and ends. Thus, the most direct test of the hypothesis is to see if ideological polarization moderates with social and political development.4 We began our analysis by calculating the percentage in each nation that scored at either the two most Leftist categories on the ideological scale, or the two most Rightist categories on the scale.5 These percentages are based only on those who positioned themselves on the scale as an indicator of the politicized public. Figure 1 presents the relationship between national affluence (GNP per capita, adjusted for purchasing power parity) with Left and Right extremism. Clearly, there is a strong negative curvilinear relationship between extremism and affluence. For Left extremism, the relationship has a Multiple R of 0.50; for Right extremism the R is 0.56. Right extremism is especially high in less affluent nations, reaching over 20 per cent in the poorest nations. Both Left and Right extremism average only about 5 per cent of the public in the most affluent societies. The nature of the political controversies certainly varies across these nations, so a Left extreme position likely taps a different subset of issues. Thus, we are not making claims about the content of these ideological positions at this point (although see our discussion below). Rather, we are asking if the polarization of mass publics is systematically related to economic development and the empirical evidence answers with a strong yes. Furthermore, it is not the case that the Left is polarized in one nation, and the Right in another. There is a significant positive relationship (R = 0.34) between the percentage of Left and Right extremists across nations. The cumulative nature of these patterns is even more evident if we combine Left and Right extremists together; the Multiple R with national income increases to 0.64. Thus, independent of the content of political controversy, citizens in lower income nations are more likely to divide themselves into sharply opposing ideological camps. To make sure that these patterns were not unique to the World Values Survey because of the particular group of nations in the study or instrumentation effects, we replicated these analyses with data from the Comparative Study of Electoral Systems (CSES). Module I of the CSES surveyed 22 democracies and included a question on Left/Right self-placement. We calculated the percentage of the public that took the

For the less developed nations in our study, a contrasting explanation suggests the causal arrow might go in the opposite direction. That is, sharp polarization in a nation may hinder national development, thus retarding economic growth and democratization.
5

We excluded Vietnam and Tanzania because in both instances a large majority was positioned in one category, and thus we presumed this represents a nation-specific interpretation of this scale. At the same time, since these are both low income nations, their inclusion would only strengthen the pattern in Figure 1.

Figure 1. The percent of Left and Right extremists by national affluence

two extreme points on either the Left or Right end of the continuum. The percentage of extremists was strongly correlated (0.68) with a measure of national affluence virtually an identical result to the WVS. A similar pattern occurs if we use an indicator of political development, measured by the Freedom House scores. There is a strong negative relationship between the percentage of extreme Leftists (R = 0.49), extreme Rightists (R = 0.40), and total extremism (R = 0.52) in the World Values Survey. This common pattern should be expected since the GNP and Freedom House scores are themselves so strongly correlated (R = 0.73). Our findings thus broadly affirm Bells thesis of the convergence of ideology with socio-political development. Similarly, Mainwaring (2005) has recently used data from the Comparative Study of Electoral Systems to show that the Left/Right polarization of party sympathizers is substantially weaker in more developed nations. In other words, less affluent and less democratic nations face an electorate that is often sharply polarized on the dominant ideological divisions of the nation. The centrifugal forces created by such polarization likely produce strains in governing, which may lessen the performance of the government and even weaken the stability of the regime.6 Thus, regardless of the content of the ideological dimension, this debate is characterized by a sharply divided public in many developing nations. Conversely, ideological polarization is more moderate in advanced industrial democracies. Far fewer individuals place themselves at the extreme positions on the Left/Right scale in these nations. This suggests that political competition is more manageable, because a large moderate center provides a middle ground for political discourse and cooperation. Even when elite debates may become intense between the ideological extremes, the lack of an equally polarized public moderates these controversies. Thus, while the centrifugal forces of polarization create political strains in many developing nations, a large moderate center exerts centripetal forces in most advanced industrial democracies. Social modernization and the content of ideology The second element of the End of Ideology debate involves the potential

shift in the content of ideology because of social modernization. Ingleharts Postmaterial thesis holds social modernization addresses many of the basic economic and sustenance needs that traditionally have been the major policy goals of citizens and their governments. In European political systems, these needs were typically expressed in Marxian class- based issues such as the nationalization of industry, redistribution of income, and the governments role in the economy. However, as these goals were addressed, public and eventually government attention shifted attention toward a new set of postmaterial goals. Issues such as environmental protection, social equality, self-expression and life style choices typified this new postmaterial issue agenda. He demonstrated this pattern with data from Western Europe, which showed that Left/Right identities were a mix of traditional economic issues and postmaterial issues (Inglehart, 1990: ch. 9). Moreover, the connection between Left/Right ideology and postmaterial issues was substantially stronger among the young, suggesting a generational shift in ideological orientations was occurring. As we have noted above, the empirical evidence on the content of Left/Right orientations in developing nations is quite limited. In many of these nations we find that a high percentage of the public is able to place themselves on this scale. Our next task is to determine the content of these orientations by examining the correlates of Left and Right. The World Values Survey includes three sets of issues that are broadly discussed in the literature on ideology.

6 The

World Bank combined several indicators to measure perceptions of the likelihood that the government in power will be destabilized or overthrown by possibly unconstitutional and/or violent means, including domestic violence and terrorism (www.worldbank.org/wbi/governance/ govdata2002/index.html). This index is strongly related to the combined percentage of Left and Right extremists in a nation (R = 0.68), which suggests that ideological polarization does weaken a regime.

Economic orientations: The End of Ideology debate typically focuses on the question of whether social modernization lessens Marxist-based ideological controversies. This theme is highly visible both in Bells writings and in Ingleharts. We selected four measures from the WVS that seem to best represent this cleavage:7 . Government ownership of business and industry should be increased . Competition is good. It stimulates people to work hard and develop ideas . People should take more responsibility to provide for themselves . We need larger income differences as incentives for individual effort Some of these issues are linked to socialist debate about the relationship between

government and the economy. Other questions focus on issues of inequality and individual responsibility that may have equal relevance in developing nations. Postmaterial orientations: Ingleharts (1977, 1999) concept of postmaterial value change led us to select the following four items as tapping these orientations: Four-item postmaterial values index Sustenance vs self-expressive values Support for environmental protection Support for gender equality Other orientations: While much of the debate has focused on hypothesized transition from economic to postmaterial issues, the literature also discusses other potential basis of polarization. For instance, religion is regularly seen as stimulating moral issues that are separate from either economic or postmaterial orientations (Norris and Inglehart, 2004; Huntington, 1996). We include support for democracy to see if regime form provides a basis of political cleavage in less-democratic nations (see Shin and Jhee, 2004). Bell and others have claimed that nationalism may serve as a basis for political identity in developing nations. Therefore, we included four other measures into our analyses: Importance of God in ones life Consider oneself religious Democratic regime index National pride We are not interested in the distribution of these issue opinions in each nation. Rather, we are asking to what extent these different issue dimensions are related to Left/Right orientations, signifying the content of Left/Right positions in the nation. Because economic and political development tend to be strongly related, and we have a large number of nations, we group nations into six broad cultural regions to provide the broad perspective encapsulated in the End of Ideology debate. We identify six regions with common characteristics: advanced industrial democracies, the

7 For

the combined set of nations, these four items were strongly interrelated and formed a common dimension in a factor analysis. The factor loadings were: government ownership (0.70), competition is good (0.62), individual responsibility (0.59), income inequality (0.33).

post-communist nations of East Europe, Latin American nations, Asian democracies, Middle Eastern nations, and a diverse set of African nations.8

Economic orientations We first analyze the relationship between economic orientations and Left/Right self-placement. The literature provides only partial evidence on what we might find across global regions. Inglehart demonstrated that economic issues are still significantly related to Left/Right orientations in Europe, even if postmaterial issues also are part of contemporary Left/Right divisions. Kitschelt and his colleagues (1999) also found that economic issues were often linked to Left/Right orientations among East European publics in the early 1990s. Both of these studies, however, are limited to a single geographic region, and the comparison to other regions is implicit. The top panel of Table 2 presents the relationship between four economic issues and Left/Right orientations across the six regions. The table presents the unstandardized regression coefficients for all the nations in a region pooled together. We used the unstandardized coefficients because the previous analyses demonstrated the substantial differences in Left/Right variance across regions. Economic polarization along the Left/Right scale is strongest in the advanced industrial democracies. On the one hand, these controversies have been institutionalized in the structure of the party system and elite competition for more than a century, and this persisting importance is still apparent. Europe was, after all, the birthplace of Marxism, socialism and the ideological movements that generated the traditional meaning of Left. On the other hand, longitudinal evidence and generational comparisons suggest that economic controversies are moderating in Europe.9 A generation ago, economic polarization in Western Europe probably was even sharper. Economic polarization also is a significant basis of political cleavage in East Europe. Certainly East Europeans were familiar with this ideological debate, since it was embedded in the communist political order. But this cleavage was blurred under the old regimes, since there was no ideological competition between alternative positions. The communist class structure also transformed the traditional class alignment; despite the claim that the regime represented the working class, the intelligentsia and middle class held a privileged status. However, after the transition to democracy, the traditional

Not all nations listed in Table 1 are available on all the items we compare. Most categories are self- explanatory. The category of Asian democracies includes Japan, Taiwan, South Korea, India, the Philippines, and Turkey. For South Korea we use the 1999 data, even though there is not a separate missing data code to separate respondents not placing themselves on the scale. Turkey was included in this group to distinguish it from the non-democracies in the Middle East. Japan is included both among the advanced industrial democracies and among the Asian democracies, because it reflects elements of both groups. There were a few remaining non-democratic nations in Asia, but the number

was too few to make regional generalizations.

We ran separate generational comparisons for the advanced industrial democracies. The Multiple R for these four issues predicting Left/Right orientations was slightly stronger for citizens over 50 (R = 0.25) than for those under 30 (R = 0.22).

Table 2. The correlates of Left/Right orientations

Left/Right economic alignment is emerging in East European party systems and is only slightly weaker than in the West (Kitschelt et al., 1999). When one moves beyond Europe, however, economic controversies over the role of the state, income inequality, and market competition are relatively independent of Left/Right orientations. In Latin America, for instance, polarization on economic issues is barely linked to Left/Right attitudes.10 Economic polarization is also substantially weaker in Asian democracies, Arab nations, and African nations. Marxism, and its attendant political controversies, may have structured political conflict in Europe, but such controversies are less central to the ideological framework of mass publics in the rest of the world. These results demonstrate that the socialist/capitalist controversies that structured politics in Europe, do not drive politics in the developing world. This

10 One

possibility is that the low coefficient occurs because we averaged quite different relationships across the ten Latin American surveys. We examined the government ownership question as one example. Only two of these ten nations match the average correlation for advanced industrial democracies, and most display a very weak relationship. Thus, it is the absence of a relationship, rather than conflicting relationships, which primarily produces the weak overall coefficient in Latin America (0.02).

does not preclude other economic issues from framing Left/Right attitudes among these publics, but we have not yet been able to systematically identify such economic issues using the WVS. However, it is clear that Left/Right divisions in the developing world are not strongly linked to the ideological conflicts over economic growth that structured European politics for the past century. Postmaterial orientations Our theoretical expectation for postmaterial issues is quite clear. Since these issues supposedly enter the political agenda once socio-economic development addresses traditional economic and security concerns, postmaterial issues should be a significant source of ideological orientations only in advanced industrial democracies. Ingleharts (1977, 1990) analyses of postmaterial values support such a prediction. The second panel of Table 2 presents the relationship between several postmaterial measures and Left/Right orientations. Rather than postmaterialism being the distinct basis of ideological polarization in advanced industrial democracies, these

values are strongly related to ideology on a global scale.11 In most regions, postmaterialists are disproportionately Leftist, and this is the pattern in advanced industrial democracies, Latin America, Asian democracies, and Arab nations. Among these four regions, the relationship is actually weaker in the Western democracies. At the same time, postmaterialism has the opposite impact on ideology in Eastern Europe and Africa. In these two regions, postmaterialists are more likely to locate themselves on the Right end of the scale.12 One explanation for this apparent anomaly is that postmaterialism is related to broader value differences linked to modernization, which Inglehart and Welzel (2005) describe as the shift from survival to self-expressive values. These concerns are relevant beyond just advanced industrial democracies, and we might expect that survival concerns are even more salient in less affluent nations. To illustrate this broader relationship, the second row in this set of variables presents the relationship for self- expressive values. At the same time, it appears that historical conditions have created different linkages between these value orientations and Left/Right ideological labels. Another example of the postmaterial agenda is environmental protection. The WVS contained a number of environmental questions, and we selected an item on the willingness to pay more for environmental protection because it was included in the largest number of surveys. Environmentalism is significantly linked to Leftist orientations in the Western democracies (b = 0.16), with weaker relationships in most
.
11

Because these are unstandardized regression coefficients, one cannot directly compare the magnitude of effects across variables. The comparisons should focus on the same variable across regions. In judging the differences for the postmaterialism variables, the standard errors of these coefficients were in the 0.020.04 range. Also see the multivariate analyses in Table 3 below. For the pooled set of East European nations, postmaterialists were 910 per cent more common on the far Right compared with the far Left, the opposite of the relationship in Western democracies.

12

social modernization and the end of ideology debate 15 other regions (and sometimes the relationship runs in the opposite direction).13 Thus, this postmaterialist issue does mobilize Western voters as Inglehart hypothesized. Our final postmaterial example is support for gender equality. The emergence of a feminist movement in advanced industrial democracies is typically linked to the postmaterial agenda. However, Table 2 demonstrates that gender roles are commonly linked to Left/Right orientations even beyond the advanced industrial

democracies. The strongest relationships are found in Arab nations, Asian democracies, and Latin America regions where the role of women remains intensely debated (Inglehart and Norris 2003). Perhaps the greatest formal equality for women was achieved in Eastern Europe, and here gender role has the weakest link to ideological orientations. In summary, these examples of postmaterial issues yield ambiguous regional patterns. As postmaterial issues, their impact should be strongest in advanced industrial democracies, where their impact is actually quite modest.14 Instead, these issues have a stronger impact in developing nations where questions of material well-being and gender roles are salient to many citizens. It may be that these issues are more visible in advanced industrial democracies, because in these nations the two sides are relatively balanced. However, where differences in opinion exist in these other regions, it serves as a basis of polarization. The other anomaly is Eastern Europe. The communist heritage of these nations has apparently created a different Left/Right alignment on these issues, with the Left becoming the advocate for materialist interests. The semantics of Left and Right have different meanings to post-communist electorates, and these meanings also likely vary across nations within the region (Kitschelt et al., 1999: 282288). Other orientations Bell suggested that factors such as decolonialization, nation-building, religion, and ethnicity might provide the basis of Left/Right orientations in the developing world, as these nations faced different political problems than those that shaped the history of European societies. Therefore, we selected a set of measures from the World Values Survey that might tap such orientations. These analyses are even more tentative than the two prior sections, because research on the ideological orientations of publics in developing nations is much less extensive. Nevertheless, our exploratory analyses can help chart this territory and increase understanding of the nature of political cleavages in the developing world. Several studies have demonstrated that religion is strongly related to Left/Right orientations in less developed societies (Norris and Inglehart, 2004: ch. 9; Mainwaring, 1999). However, it is difficult to study religious attachments in a comparable manner

13

A notable exception to the pattern is the African nations, where environmental attitudes have a strong relationship (in the opposite direction to the Western democracies). We suspect this is because environmentalism is indirectly tapping economic issues because of the wording of the question. We combined three variables postmaterial values, environmentalism, and gender equality to predict Left/Right orientations within each region. The Multiple R in Western democracies is only half that of economic issues in these same nations.

14

around the globe. The number of religious denominations across these 70+ nations confounds the broad comparisons we are seeking, and the patterns of religious observance are equally varied. The typical frequency of attendance at religious ceremonies varies widely across religions. Even monotheistic beliefs are not universal. Therefore, we relied on two questions to tap religious orientations. The first question asked about the importance of God in the respondents life. The second asked whether the respondent considered himself/herself a religious person, regardless of whether he/she went to formal religious services. The last panel of Table 2 shows that religious attachments are related to Left/Right orientations in most regions. In Latin America and Asian democracies, both religious items are strongly related to Left/Right selfplacements. In the former, it undoubtedly reflects the role of the Catholic Church in Latin American politics. The importance of God item has its strongest relationship in Arab nations (b = 0.206), roughly doubling this variables impact in Western or Eastern Europe. Africa is the one region where neither religious question displays a substantively large coefficient. On a global scale, however, religion does matter, and its seems to matter more in the developing nations of Latin America, Asia, and the Middle East. Another potential ideological division involves orientations toward various aspects of nation building. We first compare whether support for a democratic regime structures Left/Right orientations.15 There is a clear interaction of ideology and national history. In post-communist East Europe, for example, the democratic orientations of the citizenry are linked to Rightist orientations because of the communist legacy. Similarly, in Arab nations the Left is tied to anti-Western and anti-democratic views, and democratic values lead to Rightist orientations. In Latin America and Asia the effects of democratic values is even stronger, and these orientations lead to Leftism. In short, regime choice is part of the ideological structure in new democracies and non-democratic nations, but the nature of this relationship varies with the historical alignment of Left/Right forces. Our final measure taps nation building on a different level. The WVS asked about feelings of national pride, which we interpret as tapping nationalist orientations. National pride typically is related to Rightist orientations, except in Eastern Europe and Arab nations. We suspect that this reflects domestic political alignments similar to those displayed for democratic values. These findings suggest that citizens in developing nations are more likely to orient themselves to politics in terms of religion or national identity issues, which generally outweigh economic issues as a basis of Left/Right identity. We suspect that ethnicity may also play a greater role in shaping political identities in developing nations (Elkins

15 We

combined four questions on regime preferences, support for government by: (a) strong leaders, (b), experts, (c), army rule, and (d), support for democracy. We simply summed together responses to the four items (reversing the polarity of the democracy item) and divided the total by four. The resulting scale runs from (1) prefer nondemocratic regimes and disapprove of democracy, to (4) prefer a democratic regime and disapprove of non-democratic regimes.

social modernization and the end of ideology debate 17 and Sides, 2004; Bell, 2000). The different bases of political identities in developing nations can also explain why politics may become so divisive and polarizing. While policy differences over economics or even environmental quality are more susceptible to compromise, divisions over religion, regime and political community are more fundamental, almost inviolable to their adherents. Combining dimensions One advantage of the Left/Right scale is that it provides a summary of the issues of contention in a political system. However, our analyses show that Left/Right orientations do not derive from a single source, but reflect multiple dimensions. Even in the broad regional patterns we have presented, both economic and cultural factors shape political identities, and the mix of dimensions reflects both the social development of the nation, and its cultural/political history. Therefore, the last step in our analyses combines the several dimensions to examine their joint and independent influence in forming Left/Right orientations. We selected a subset of six items from Table 2 to avoid multicollinearity among items tapping a similar concept. We began with the item on government ownership of business to measure the classic socialist/capitalist divide. As we have shown, this continues to shape the identities of some citizens. We selected support for the environment to tap potential postmaterial concerns. Although the gender role scale was initially analyzed to measure postmaterial orientations, based on the patterns in Table 2 we included it as a measure of traditional values in many developing nations (and thus distinct from postmaterial concerns). Belief in the importance of God provides a measure of religiosity. Support for democratic values is one element of political development that may be important in democratizing nations (e.g., Shin and Jhee, 2004). National pride provides an indicator of how nationalism may shape Left/Right orientations. We included all six predictors in an OLS regression model, including all nations within each cultural region in separate analyses (Table 3). We consider these analyses as a preliminary attempt to map broad ideological orientations in terms of global regions, with the expectation that further research should probe these patterns on a nation-by-nation basis and with additional predictors. Our findings for the advanced industrial societies are very consistent with previous research. In these nations, the two strongest correlates of

Left/Right orientations are economics ( = 0.143) and religion ( = 0.138). These have been the two major pillars of social and political competition in Europe and other Western democracies, and their impact remains to the present. In addition, the postmaterial issue of environmental protection has greater weight in shaping the identities of these publics, relative to the other five regions in our analyses. Thus, the content of Left/Right identities remains quite full and rich in Western democracies, even if the intensity of these sentiments have been moderate during the later half of the twentieth century. East European publics generally mirror the pattern among Western publics. Economics and religion are the strongest correlates of Left/Attitudes in the East,
Table 3. A multivariate model of Left/Right orientations across regions

albeit with significantly weaker coefficients in both instances. With our single cross- sectional survey we cannot tell whether the impact of these correlates is changing with democratization, and this is important in interpreting the present results. But indirect evidence suggests that ideological polarization is increasing along these cleavage lines. It is also significant that democratic values do not divide these publics. Immediately following the democratic transition, this may have been an important factor in structuring party competition and charting the institutional structure of the new system (Kitschelt et al., 1999). This division is no longer linked to Left/Right identities, however. Consistent with our earlier analyses, people in developing nations are more likely to base their political identity on gender and religion. Gender and religion tap traditional value orientations based on the social patterns of pre-industrial societies. The importance of gender roles in shaping political identities in Latin America, the Middle East and East Asian democracies is a striking example of the persistence of these value cleavages. Orientations toward gender appear to tap feelings of equality between the sexes, as well as elements of tolerance and modernization that transcend the specific relationship between men and women (Inglehart and Norris, 2003). Religion also displays a significant relationship with Left/Right orientations in Latin America and the Middle East. The nature of Eastern religion lessens its role as a political cleavage in East Asian democracies. Perhaps what is the most striking in Africa is the narrow base of Left/Right orientations. The poverty of these nations precludes the strong class alignments as observed in European democracies as a consequence of industrialization. Neither religion nor gender roles taps into traditional value cleavages. The only substantial predictor of Left/Right is attitudes toward environmental protection; but this relationship runs in the opposite direction as in the advanced industrial democracies. We suspect that this question taps economic concerns (because of its reference to taxes in the

question) rather than environmental quality per se. The other potential predictors of Left/Right orientations display only weak effects. Social modernization and ideology Does social modernization transform the ideological basis of mass belief systems? Daniel Bell (1960, 1973) offered a broad theoretical model of how social modernization would affect mass beliefs. First, Bell presumed that the degree of ideological polarization was narrowing in advanced industrial societies. Second, he claimed that the sources of ideological polarization also changed as a consequence of modernization. We have used the unprecedented resources of the World Values Survey to empirically examine Bells theory to an extent never before possible. We have painted on a large canvas, and used broad brush strokes. Before discussing our findings and implications, we want to stress the need for more extensive and detailed analyses across individual nations and using different issue dimensions. We examined a broad theory of social change, analyzing patterns of Left/Right polarization across six regions of the globe. Some of the specific national patterns within these regions undoubtedly vary from the overall pattern, and these national differences can provide insights into how national histories structure the framework of political competition.16 And it would be valuable to build upon our findings with other methodologies, such as open-ended inquiries into the meaning of Left and Right. Thus, our findings provide an outline of the impact of social modernization on ideology that can be refined with more detailed analyses. We treat the Left/Right attitudes of mass publics as an indicator of the ideological orientations that Bell described. In large measure, the degree of ideological polarization fits Bells predictions. Polarization along the Left/Right dimension is substantially greater in the less affluent and less democratic societies than in advanced industrial democracies. Left and Right extremism reaches over 20 per cent in the least developed nations, but averages only about 5 per cent among the public in Western democracies. The extent of political polarization is important because it can shape the political process of a nation. A long theoretical tradition holds that the centrifugal forces generated by polarization strain the political consensus (Sartori, 1976: 131173; Mainwaring, 1999: 131135). Bingham Powell (1982) for instance, demonstrated that support for extremist parties in a nation was a significant predictor of political violence. Similarly, Dalton and van Sickle (2004) found a significant relationship between the percentage of Left and Right extremists in a nation and the level of political protest. Just as the End of Ideology hypothesis linked ideological convergence to the moderation of political conflict, divided publics may strain the political order and the ability of states to govern.

Another major finding is that the correlates of Left/Right orientations vary systematically across regions. The twin pillars of economic and religious cleavages remain important in European states. These two bases of polarization are strongest in Western democracies ( = 0.143 and = 0.138 respectively), which runs counter both to Bells End of Ideology hypothesis and Ingleharts Postmaterial hypothesis. We suspect that a previous partisan competition on these two cleavages generated even sharper differences in the mid twentieth century, which remain apparent in our contemporary data. In other words, polarization was presumably even greater a generation ago (Kirchheimer, 1966; Inglehart, 1977, 1990). In addition, environmental issues have their greatest weight in Western democracies, suggesting that this new basis of cleavage is developing within the advanced industrial democracies. In contrast, Left/Right orientations in other regions are typically not linked to the economic issues that divide European publics. In our set of Asian democracies, traditional social values (represented by attitudes toward gender roles) and political community (national pride) divide these publics. Thus, the social dimensions of class
16 For

instance, we found that both economic and religious issues were correlated with Left/Right orientations in Eastern Europe. In more detailed comparison of four Central European nations, Kitschelt et al. (1999: ch. 8) finds that religion has a strong impact in Poland, but weak effects in the other three nations. At the same time, economic issues are significant correlates of Left/Right in each nation.

social modernization and the end of ideology debate 21 and religion that structure Western political systems are largely muted in the East, and postmaterial issues are also less salient. Few among Arab publics express a Left/Right identity, since this nomenclature is not commonly used in political discourse. Where such Left/Right identities do exist, they are strongly related to attitudes toward gender roles. We see this as partially a reflection of the strong gender divide in Arab societies, as well as broader feelings of tolerance and acceptance of diversity that are tapped by the gender question (Inglehart and Norris, 2003). Finally, African publics appear to have diffuse Left/Right orientations, with all the correlates we examined. In his recent comment on the End of Ideology, Bell (2000) claims that ethnicity and linguistic cleavages are strengthening in the developing world, providing a new basis of division even if these divisions are not fully expressed in a broader worldview or ideology. The differing bases of ideology may explain why developing societies are so polarized, because political orientations are shaped by deeply seated questions of moral and national identity. This pattern also suggests that these divisions will be more difficult to manage than the economic competition in advanced industrial democracies. This evidence on the correlates of ideological polarization is important at

several levels. Ideological frameworks broadly define the content of politics and the nature of political competition. If politics is framed in terms of a socialist/capitalist or a nationalist ideology, this will influence the policy choices presented to the public and the nature of political discourse. The dominance of one ideological framework can also exclude other issues from the agenda, as when the New Deal debate in America excluded race from consideration. Or, when nationalism issues submerge discussions of class interests. Thus, the ideological structures we described are important in modeling processes of coalition formation, political representation, and electoral competition. In summary, it is premature to argue that ideology is ending in any region of the globe. Citizens in affluent and less-affluent societies still rely on broad orientations such as Left/Right identities as heuristics for political action. But social modernization does transform the content of ideological polarization and the degree of this division. Indeed, one of the major underdiscussed consequences of social modernization may be this transformation of the ideological debate and all that this implies.

5.

Right and Left, Needham

Hertz is an example of a political scientist who categorized left and right based on reasons of the social kind. Ed Needhams Right and Left compiles different perspectives of determinants of the concepts based fully on cultural context and over time, illustrating how context-dependent the terms are. Essentially, he presents empirical evidence through the consideration of the cultural signification from different civilizations what the terms mean to various different cultures and their origin. CULTURE EXAMPLE: Arabic Culture Religious polarity Extreme degree the characteristics of a social institution

6. Party Policies in Modern Democracies, Benoit and Laver

There are two fundamentally different approaches to estimating positions on a left-right scale. The first is to assume that the substantive meaning of left and right in any given political system is self-evident to all In this sense, the left-right spectrum is sui generis for a given political system, forming one of the primitives

of the system. The second approach is to build up a left-right scale from component parts, each part having a more precise substantive meaning than the more general underlying notions of left and right. Thus the left-right scale in a given political system can be seen as having to do with economic policy where economic policy might include policies on the trade-offs between lower taxation and higher public spending, for example, or between the regulation and deregulation of business and industry. It may also be seen as having to do with social policy on matters such as military spending, overseas aid, and dealings with international organizations such as the United Nations. In other words, there is a lot of substantive policy content that people typically regard as being natural to associate with the left-right spectrum in politics, and one approach to defining and estimating a left-right scale is to construct this scale from its substantive content. In building this so-called scale, Benoit states that there are some systems that are primarily about economic policy and others that have exrepion of economic and social policy. By its very nature the left-right scale, having no fixed definition in terms of its substantive policy content, is likely to vary in meaning as we move from country to country. While most people using the terms left and right probably feel that these terms do indeed have substantive policy content in any given setting, it seems very unlikely indeed that all of the policy areas in which we are interested have the same importance in all political systems we want to investigate. Thus the relative contributions of different policy areas to the meaning of left and right seem likely to vary from setting to setting. For this reason at the very least, the substantive meaning of left and right is almost certain to very between political systems. RESULT OF SURVEYS THE SUBSTANTIVE CONTEXTS OF LEFT AND RIGHT IN DIFFERENT CONTEXTS THE LEFT-RIGHT SCALE AS A DESCRIPTION OF SOCIO-ECONOMIC POLICY POSITIONS: Perhaps the most common way of imputing substantive policy content to the left-right scale is to describe it as a left-right scale of socio-economic policy. This implies that the meaning of left and right is some bend of a dimension contrasting interventionist and laissez-faire economic policies, with a

dimension contrasting liberal with conservative positions on matters of social and moral policy. At one end of the combined left-right scale we have people who favor liberal views on social policy as well as economic policies involving higher levels of state intervention in the economy. At the other end we have those who favor conservative views on social policy and lower levels of state intervention in the economy. Combining substantive policy dimensions in this way is an empirical generalization aboyt the way many people think about the world implying that social liberals have tended to favor state intervention in the economy and social conservatives have tended to favor laissez-faire economic policies. Some of the empirical studies conducted by the author, however, fail to consider the causal relationship between the social aspects and the economy. For example, he studies immigration laws and deregulation as separate policies but fails to see how Base on table 6.a.1: the classic socio-economic definition of left and right seems to be a particularly western way of looking at things. Going beyond the information in Table 6.a.1, it is also the case that the existence of important local policy dimensions such as national identity in Japan, religion in Turkey, security and Palestinian statehood in Israel, or the Northern Ireland question in both parts of Ireland lessen the degree to which purely socioeconomic policy can fully explain left and right placements in a particular context. NOTE: SUBSTANTIVE NOT PROCEDURAL The detailed regression results in Table 6.a.1 confirm our expectation that there is huge variation in the substantive correlates of left and right when we move from country to country. These findings are summarized in Figure 6.1., which shows, by broad region, the differences between the regression coefficients for economic and social policy. The regression coefficient of 1.0 means that there is essentially a one-to-one correspondence between party positions on the policy dimension under consideration and positions of the same parties on the leftright scale; a smaller coefficient means that there is an association but that the correspondence is less direct. Figure 6.1 shows the coefficients for the taxes/spending dimension on the right, and those for social liberalism on the left. Nearly all coefficients were

positive in sign, indicating that social conservatism and a preferences for lowering taxes were indeed consistent with the local definitions of political right. The only exceptions are in cases where the bars cross over to the other side. This happened in Slovakia for social liberalism, and for taxes/spending policy in Croatia and Malta although only in the case of Malta was the coefficient estimate statistically significant. Comparing the coefficient for economic policy with the coefficient for social policy the only two independent variables in each analysis gives us a measure of the relative impact of economic and social policy positions on predictions of left-right policy positions. Even the briefest of glances at Figure 6.1 shows us that the substantive meaning of left and right, at least in the minds of country specialists who are probably sources as authorative as any on such matters, is indeed very different in different countries. At the top section of each region, where the taxes/spending bars are longest, we find countries such as Cyprus, Iceland, Norway, the Ukraine, Belarus, Latvia, and Estonia, for which left-right positions are almost entirely explained by economic policy. At the bottom of the table we find countries such as Portugal, Austria, Hungary, Japan, and Turkey, for which leftright positions are by contrast dominated by positioning on social policy. Between these extremes, we find countries such as Germany, Britain, Poland, Israel, the United States, and Australia, for which economic and social policy contribute to left-right positions in relatively equal measure. Economic and social policy positions are not the only substantive correlates of left and right in each country. Yet viewed by themselves, what emerges beyond any shadow of a doubt from Figure 6.1 is that their relative contribution to the left and right placements and hence the substantive meaning of the left-right political dimension itself varies widely across countries This raises the possibility that movements in left-right positions may either be real underlying movement in party positions, or changes over time in substantive meaning of the scale. CONCLUSION: it is indeed possible to predict expert placements of parties on a very general left-right scale from placements of the same parties on a set of substantive policy dimensions. This is true even when we confine the set of substantive dimensions in

two economic policy, measured in terms of the trade-off between lower taxes and higher public spending, and policy on moral issues such as abortions, gay rights, and euthanasia. Using only these two substantive scales, OLS regressions predicting leftright placements had adjusted R2 values of over 0.67 for half of the countries we investigated. The more substantive policy dimensions are considered further improves the fit. The relative contribution of different substantive policy dimensions to our ability to predict parties left-right positions varies quite considerably from country to country. In some countries, left and right are overwhelmingly about economic policy. In others, left and right are primarily associated with social liberalism and conservatism. Furthermore, once we look inside the general realm of economic policy, there are countries where it is the taxes/spending dimension that is associated with left and right, and others where left and right are more closely associated with deregulation. Similarly, in some countries it is social liberalism and conservatism that allows us to classify parties as being on the left or the right; in others it is immigration policy. The pessimistic conclusion, then, is that we may well be treading on thin ice methodologically when comparing left-right policy positions across space or time. Because the substantive meaning of the left-right dimension is so context-dependent, it may be impossible for any single scale to measure this dimension in a manner that can be used for reliable or meaningful cross-national comparisons. It is later advised that maybe having a onedimensional model of some aspect of party competition is deemed more important.