This article focuses on the issue orientations of Danish voters at the most recent elections for the purpose of validating the existence of a New Politics dimension along with the customary Old Politics dimension, which is associated with the traditional left-right cleavage. In addition to studying the interrelations between issue positions, I proceed to examine some plausible questions concerning the relation of the Old and the New Politics dimension to party choice and social structure.

Resumé

A New Politics dimension of non-economic issue preferences emerged in Denmark during the 1980s. The new dimension is clearly separated from the traditional economic Old Politics dimension, though the two are correlated. The New Politics dimension has not given rise to viable new parties but is represented by three medium-sized parties: the Socialist People's and Radical Liberal parties constitute the New Left, and the Progressive Party constitutes the New Right. The three old class parties contribute little towards explaining the individual variance in the position of New Politics; as always, the Social Democratic Party represents the Old Left while the Liberal and Conservative parties represent the Old Right. In terms of social positions, New Left attitudes are correlated especially with level of school education and with employment in the reproductive sectors.

The Concept of New Politics

According to Miller & Levitin (1976, 5), the concept of New Politics came into use in the United States toward the end of the 1960s as a label for an interrelated set of "slogans, symbols, values, attitudes, and behaviors associated with the turmoil of the previous several years". However, in defining positions on the New Politics dimension, Miller and Levitin relied exclusively on attitudes, as they constructed an index of issue orientations and labelled the two polar positions the "New Liberals" and the "Silent Minority", respectively.

In a European context the concept of New Politics was introduced by Hildebrandt & Dalton (1978) in a study of the 1976 German election. They defined the New Politics dimension in terms of a mixture of issue orientations and value preferences, showing the affinity of the New Politics

dimension with Inglehart's (1971) materialist/post-materialist value dimension.

In later studies the label of New Politics has been associated with the emergence of new parties and new social movements (e.g. Miiller-Rommel 1990) and with attitudes toward these phenomena (e.g. Dalton 1988). Another result of the 1980s has been a more explicit recognition of the two poles of the New Politics dimension, the New Right as well as the New Left. This is in part a result of the actual appearance of New Right positions with considerable support in the electorates. Miller and Levitin polemically spoke of the "Silent Minority", showing the decline of that group between 1972 and 1976. During the 1980s, however, it seems to have grown in number and become less silent. In European countries, parties have been formed on particular New Right positions such as anti-immigration, as a counterpart to the previous wave of New Left parties based especially on en vironmentalism.

It seems prudent, however, to revert to the usage of Miller and Levitin, Hildebrandt and Dalton, which defines the Old and New Politics dimensions neither in terms of values nor in terms of parties and movements but in terms of issue orientations or policy positions. Broader value change, the emergence of cultural types and personality traits, etc., are then treated as analytically separated from Old and New Politics: they may or may not be shown to underlie the more specific policy positions of the Old and the New Politics. In turn these policy positions may or may not be taken up in party programmes and platforms and give rise to the formation of new parties. It therefore seems expedient to keep these processes apart from the question of defining the Old and New Politics dimensions.

Concerning the content of these dimensions, there seems to be a consensus Old Politics issues are mainly economic and New Politics issues mainly non-economic. But the distinction between economic and noneconomic is blurred. One and the same issue often has economic as well as non-economic aspects and its classification with respect to the two dimensions frequently depends on how the item was worded in the survey. For example, asking a respondent about the importance of environmental protection may elicit other responses than if the item attempts to make the respondent take into account the cost of environmental protection. The same holds true for some welfare issues, such as medical care and child allowance, which can alternatively be phrased as humanitarian issues or as economic transfers.

Another stipulation which is implied by some writers is that the New Politics issues should belong to the agenda of the mid-1960s or later. This usage creates certain problems, however. Some issues that are mainly economic have emerged during the 19705, such as the tax issue, the cost of the welfare state, and employee ownership of private firms. Conversely,

many non-economic issues date back to earlier time periods: issues of free speech, religious and moral issues, and security issues such as the NATO membership, to mention the most obvious examples. Finally, latecomers in the process of industrialization and democratization may at present be on their way towards a policy dimension that is similar to the Old Politics dimension in the advanced countries.

It therefore seems expedient to rely on the economic vs. non-economic distinction rather than the age of the issue in defining the content of the New Politics dimension. The label might thereby seem misleading; but if it can be shown that positions on non-material issues have begun to merge into one dimension as of late, it is still appropriate to label this emerging dimension "New Politics" even though some of its issues may have appeared on the political agenda of a previous time. For similar reasons the economic dimension can suitably be called the Old Politics dimension even though newer issues are still in the process of being added to it.

In accordance with usual practice, the New Politics as well as the Old Politics dimension will be conceived as running from left to right, thus giving rise to the polar positions of the Old Left and Old Right, the New Left and the New Right, respectively. However these labels should not be conceived as representing four distinct groups of voters. Since the same voters are asked for their positions on both the Old and the New Politics issues, a voter may be a member of either the Old Right or Old Left, and at the same time a member of the New Left or New Right. The overlap between these groups is a matter for empirical investigation.

The Two Dimensions in Denmark in 1990

Our first effort in the Danish context will be to validate the New Politics thesis, which implies that we demonstrate (1) the existence in the mass electorate of a coherent non-economic attitudinal policy dimension, (2) which is separate from - though not uncorrelated with - the classical economic left - right dimension, and (3) which is relatively permanent.

In order to investigate whether a New Politics dimension existed in the Danish 1990 election, four attitudinal items were selected to represent the Old Politics dimension and four to represent the New Politics dimension. The items making up the Old Politics dimension dealt with (1) social reforms, (2) economic equality, (3) state management of the economy, and (4) the expediency of wage increases. Those making up the New Politics dimension dealt with (5) foreign aid, (6) penalties for crimes of violence, (7) environmental protection, and (8) immigration. The full text and scoring of the responses to these items can be found in the Appendix. In the scoring process the responses on all eight items were scored from 1 (left) to 5 (right).

In order for the postulated two dimensions to emerge, a factor analysis should indicate high loadings of items 1-4 on one factor, the Old Politics dimension, and high loadings of items 5-8 on another factor, the New Politics dimension. The results of the factor analysis are presented in Table 1.

Two clear factors emerge, the New Politics dimension being represented by factor 1 and the Old Politics dimension by factor 2. Thus we have established that an internally coherent agenda of New Politics issues exists. In simple varimax rotation the factors are independent, and thus our analysis separates the New Politics agenda from the Old Politics agenda.

In the following we shall examine some correlates of respondents' positions the Old and New Politics dimension at the 1990 election. Therefore it becomes expedient to index these positions, and the result of the factor analysis in Table 1 suggests two additive indices, one for the respondents' Old Politics position using the responses to items 1-4 and one for the New Politics position based on the responses to items 5-8. Each index thereby ranges from 4 (extreme left) to 20 (extreme right).

The Emergence of the New Politics Dimension

The existence of a New Politics agenda in Denmark by 1990 comes as a surprise, since in her political development Denmark seems anything but a showcase for New Politics. After early triumphs of the New Left in 1960 and again in 1966, when the People's Socialist Party forced its way into the old four-party system, the 1973 election finally broke the party system wide open (Borre 1974). But that event hardly fits into the form recommended by the New Politics thesis. The most conspicuous addition to the party

system was the Progressive Party on its anti-tax and anti-bureaucracy platform; these could be recognized as populist variations on the theme of classical liberalism. The restructuring of the party system in 1960-66 and in 1973 therefore seemed solidly contained within the Old Politics framework.To large extent the two new major parties emerged with platforms overbidding the stands of the old parties: the Socialist People's Party overbid the Social Democrats on issues of welfare and state intervention, and the Progressive Party overbid the Conservatives and Liberals on tax reduction and economic liberalism. Although multidimensional scaling of partisan attitudes at the 1973 election exhibited a two-dimensional pattern (Rusk & Borre 1974), the second dimension that emerged was not a New Politics dimension but a protest or distrust dimension. On that dimension the presumed New Left and the presumed New Right scored in the same direction, while the parties having joined the past two governments scored in the opposite direction.

The political development after 1973 provides few indications of the emergence of New Politics in Denmark and even fewer of its rise to prominence. The issues debated during election campaigns were about the economy and the living standards of various social groups. The parties supposed to represent the New Politics (the People's Socialists on the left and the Progressives on the right) became largely isolated on the wings of the system, while the old parties sought to cope with problems of unemployment and inflation, taxes and budget deficits, and a poor balance of trade. Unlike the situation in the United States and West Germany, non-economic values were viewed largely as issues of lifestyle and morals and kept at arm's-length of parliamentary party politics. Not infrequently, the members of parliament were relieved of party discipline in roll-calls about these matters. A Green party did not participate in the general elections until 1987, where it obtained 1.4 percent of the vote, below the 2 percent threshold of representation.

Still, we might hope to find traces of a New Politics dimension emerging around 1979, at a time when the post-war generation and its issues had been given ample time to make an impression on party politics. The 1979 election survey contains a rich assortment of items on different issues, including those tapped in Table 1. Table 2 displays the factor-analytic solution for eight items, of which items 1-4 have exactly the same wording as in Table 1 while items 5-8, the presumed New Politics items, are worded differently but deal with the same issues as those in Table 1.

The Old Politics dimension comes to the fore in factor 1, but the New Politics dimension does not materialize as clearly. Factor 2 includes two of the issues, those of foreign aid and the rights of immigrants to retain their own culture and ways of life. Also the attitude towards penalties for crimes of violence loads relatively highly on this factor, whereas the goals of consumption vs. a more human environment have nothing to do with it.

Thus over the eleven years between 1979 and 1990, the different types of New-Leftists have come closer to one another, as have their opponents on the New Right. It would be interesting to know more precisely when and why this took place. A possible clue is provided by the partisan coalitions during the bourgeois four-party government of 1982-88. By and large, the bourgeois dominance was confined to the issues of Old Politics. As far as New Politics was concerned, the period saw an occasional parliamentary majority emerge on the issues of defence, civil rights, environment, and culture politics. This so-called "alternative majority" was established by the Radical Liberals siding with the Social Democrats and the People's Socialists against the bourgeois four-party government. These defeats were tolerated by the government as the minor political cost of carrying its economic policy through a very fragmented parliament. In other words, the New Left was permitted a victory as long as this did not interfere with the rule of the Old Right in "serious" matters.

The 1987 survey includes most of the items in Tables 1 and 2 but not the item on wage increase (item 4), nor that on foreign aid (item 5). The remaining six items can be followed through the three elections of 1979, 1987, 1990. The correlations between these items for these three elections are presented in Table 3.

The intercorrelations among the Old Politics items remained around r = 0.20 on all three occasions. The intercorrelations among the New Politics items, however, rose from r = 0.07 in 1979 to r = 0.22 in 1987 and r = 0.26 in 1990. Therefore the result suggests that in the main the crystallization of a New Politics agenda took place between 1979 and 1987. The third row shows that the correlations between Old Politics and New Politics items rose slightly from 1979 to 1987 but no further from 1987 to 1990. In 1979, the New Politics items were actually better correlated with the Old Politics

items than with one another; but by 1987 they formed a distinct cluster of
items.

By 1990, the distinctiveness of the New Politics dimension was suggested by Table 1. However, it should not be overlooked that the New Politics dimension emerges out of the Old Politics dimension and combines with it into a general concept of left and right in Danish politics. Even though the two factors in Table 1 were found by varimax rotation and thus were uncorrelated, the correlation between the New and Old Politics indices is r = 0.23. Their relations with respondents' left-right self-placement (on a ten-point scale) are almost linear and have a Pearson's r = 0.43 for the Old Politics index and r = 0.39 for the New Politics index. To some extent, therefore, Danish voters have tended to carry over their positions from Old Politics to positions on New Politics and to have adjusted their selfimage left and right voters accordingly. The parties can be suspected of playing a major role in that process, and we shall therefore turn to the partisan groups for a closer analysis.

The New Politics and the Party System

Scholarly tradition generally regards party systems of mature democracies as being slow and cautious in responding to new issues. The established parties are supposed to cling to old class and ideological cleavages long after these cleavages have disappeared or have been eclipsed by newer cleavages. On the other hand, the Lipset & Rokkan thesis about the "freezing" of European party systems in the early part of this century (Lipset & Rokkan 1967) ascribes to the party system a certain autonomy, a capacity for surviving economic, social and cultural change. According to this view, the "freezing" hypothesis could be criticized for failing to predict the emergence of new parties in the 19705. The Danish 1973 election seems to be a good example in supporting that criticism. At that election the three old class parties - the Social Democrats, the Liberals, and the

Conservatives - came down from 70 to 47 percent of the votes cast between
them.

However, one should also note the subsequent development. By 1984 these three parties were back at 67 percent of the votes, and in the recent election of September 1994 they received 75 percent of the votes between them. Thus the most plausible prediction made by the "freezing" hypothesis holds true: the old class parties are alive and well.

It seems paradoxical that the downfall of the old class parties occurred before the emergence of the New Politics dimension, and that the recovery of these parties took place simultaneously with the establishment of that dimension. Apparently there is a lack of hypotheses specifying how different systems will react to the challenge of New Politics issues. In particular, we need to study how party systems with a mixture of large, mainly class-based, and minor, mainly non-class parties, will respond to the challenge, since these party systems are rather typical of Western Europe.

In the early New Politics literature, the more or less implicit assumption was that New Politics attitudes would cause a permanent change in the voter base of the large parties. The old working-class party on the left would accommodate the New Left positions, and the old middle-class party on the right would take up the New Right positions. This of course entailed splits and inter-party conflicts, but permanent changes in the party systems were not predicted. This came true in the United States, for which Miller 6 Levitin showed (1976, 127) that 58 percent of the New Liberals but only 7 percent of the Silent Minority voted for the Democratic presidential candidate in 1972. In the case of Germany, the Hildebrandt & Dalton (1978) article dealt with the two main parties the Christian Democrats (CDU/CSU) and the Social Democratic Party (SPD), hardly mentioning liberal Free Democrats (FPD). The authors did not discuss whether a postmaterialist would have good chances, a possibility that materialized in 1980 when the Green Party was formed. Later, the New Politics dimension completed with the Republican Party and similar groupings in German politics.

More recent literature on post-material values and on new social movements gone to the other extreme by practically identifying New Politics with new parties. But as we have seen, the emergence of the New Politics dimension in Danish political attitudes has not been accompanied by the appearance of new parties in the party system. Most conspicuously, no viable Green party has been added. The new parties entered the stage during the 1960s and 19705, prior to the development of the New Politics dimension. Since we have defined New Politics with reference to the attitudinal dimension, it follows that new parties cannot be regarded as a valid indicator of the arrival of New Politics.

Since the "new party" hypothesis can be rejected in the Danish case, there remains the older thesis about realignment of the large parties. But a third hypothesis, in-between the other two, seems plausible. When the New Politics dimension began to emerge, there was already a field of small and medium-sized parties competing for attention. This competition was sharpened when the minor parties were squeezed between 1973 and 1984, losing 20 percent of the votes cast to the three big parties. New Politics would offer an opportunity for the minor parties to change the terms of competition, both among one another and between them and the large parties.

In turn, this strategy of the minor parties might well stabilize the monopoly the large parties on the Old Politics dimension. As voters who were attracted by the New Politics dimension shifted toward the new parties, those sticking to the Old Politics agenda would become a larger majority within the old parties. A sort of division of labour between the large and the small parties would ensue, in which the old and large parties would stand for continuity and stability, and for responsible government on the economic issues, whereas the smaller parties might serve as messengers innovators in the field of New Politics.

In order to gain more insight into the changes and positions of the party
groups on the two critical dimensions, we utilize the items from Tables 1
and 2 for the following diagrams.

Figure 1 plots the mean positions in 1979 of the voters for each party on the Old Politics index (X axis) and the New Politics index (Y axis). Figure 2 does the same for 1990. The continuity as well as the changes are very conspicuous over these eleven years. For one thing, voters of the Social Democratic and Socialist People's Parties have moved about three points towards the right on the Old Politics index, and those of the Conservative and Liberal Parties have moved over one point towards the right. With respect to the New Politics dimension, we unfortunately cannot compare the positions across the two years, since we used a different set of interview questions in 1990 from the set used in 1979.

Further, it should be noted that in both 1979 and 1990 the mean positions of the three large parties - the Social Democrats, the Conservatives, and the Liberals - differ considerably on the Old Politics dimension but very little on the New Politics dimension. With regard to the minor parties, however, dramatic changes have taken place. The Communist, Left Socialist Single-Tax Parties have disappeared from the parliament. The Socialist People's Party voters and the Radical Liberal Party voters have moved down towards the New Left, and the Progressive Party voters have moved up towards the New Right, while the voters of the small Christian People's Party have drifted toward a centrist position near the Social Democrats.

By 1990 the vertical dimension seems at least as prominent as the horizontal dimension in the party system. However, considering that the means of the three large parties, taking almost three-quarters of the vote, vary only on the Old Politics dimension, one may still be inclined to regard the Old Politics dimension as much more important than the New Politics dimension for the voting choice. We therefore need to know in quantitative terms to what extent the differences between the party means in Fig. 2 express the individual variance in Old Politics positions and in New Politics positions. This information is provided in Table 4.

We find in the upper section of Table 4 that the effect of the party mean upon the individual position is described by an Eta = 0.61 for the Old Politics dimension and Eta = 0.46 on the New Politics dimension. Thus, the Danish party system appears capable of expressing both dimensions of voter attitudes. Even though the New Politics dimension is represented in parliament almost exclusively by three medium-sized parties, their polarization large enough to absorb a sizeable portion of the variance (21 percent according to the squared Etas) of voter positions on that dimension.

The middle section of the table deals only with those 70 percent of the voters who voted for one of the three large-class parties. Within that subsample, the effect of party preference on the Old Politics position stays on a comparatively high level, Eta = 0.57, whereas its effect on the New Politics position is reduced to Eta = 0.09, which means that the three large parties absorb only one percent of the variance of their voters on the New Politics position.

Dalton (1991) has hypothesized a process whereby old issues tend to
show higher correlations with the partisan choice than with the left-right
identification of voters, and to retain their correlation with the partisan

choice longer than with left-right identification. By contrast, new issues
first acquire correlations with left-right identifications and only later with
the partisan choice.

In the Danish case, Dalton's prediction does not hold true, as we gather from the lower section of Table 4. In terms of Cramer's V coefficients, such as used by Dalton, the correlation of Old Left positions with party preference is V = 0.24 and with left-right identification V = 0.23. The correlation of New Politics position with party preference is V = 0.22 and that with left-right identification V = 0.21. There is no tendency for the party system to lag behind the left-right spectrum in expressing the new issues. However, it is obviously the small parties that add this flexibility to the party system.

It seems reasonable to conclude in agreement with our hypothesis that the party system is a dynamic force in establishing the New Politics dimension, as it is active in maintaining the Old Politics dimension. The latter function has mainly rested with the old class parties, which are still the largest in the Danish party system. The function of introducing the New Politics dimension to the voters has mainly rested with three minor parties which are less tied to the traditional economic classes: The Socialist People's Party and the Radical Liberals representing the New Left, and the Progressive the New Right.

Social Causes of New Politics Attitudes

New Politics theory is often associated with one or both of two broader perspectives, namely post-materialist theory and post-industrial theory. While the former sees New Politics issues and attitudes as caused by the values of the post-war generations of wealthy societies, the latter sees them as caused by organizational changes which began to be felt in the advanced industrial societies in the 19605.

Since the Danish election surveys do not include items asking respondents to show preference for certain values over others, they provide only inferential evidence about the post-material hypothesis. Values are supposed be learned early in life and remain stable over the individual's life cycle (Inglehart 1990, ch. 2). In so far as they cause long-run changes in the issue preferences of the electorate, such changes would materialize in cross-sectional data in the form of cohort differences.

Post-industrial theory, though it has many ramifications, almost invariably assumes education to be a critical stratification variable; the well-educated are pictured as forming a class or layer of technicians and professionals in post-industrial society (Bell 1973; Touraine 1974). The transition from industrial society toward post-industrial society implies that

personnel is shifted away from the primary and secondary sectors of production and into the tertiary sector, which includes such reproductive institutions as those of health care and education, together with information-handlingfunctions as administration, planning, research, and communicating with organizations in other settings. Education is the chief vehicle for removing people from the physical handling of goods and qualifying them for information handling. Thus, we may imagine new political cleavages to develop between the educated and the less educated, between the haves and the have-nots of the critical resource in postindustrialsociety. is not a generational cleavage, but since higher education is generally initiated while people are young, the cleavages will tend to rise in importance with the spread of higher education in new cohorts of young people.

In Table 5 a crude attempt is made to test these assumptions about the major social correlates of the Old and New Politics attitudes. The respondents' position on the two indices in 1990 was regressed on five variables: age, gender, school education, working-class membership, and membership of the old middle class (operationalized as self-employed status).

For the Old Politics position three variables are almost equally important, namely working class, gender, and old middle class. Workers are to the left and the old middle class to the right of other occupations, indicating that salaried employees (new middle class) are in the middle. Women are significantly to the left of men. However, these variables between them only succeed in producing a multiple correlation coefficient of r = 0.21.

For the New Politics position, by far the most important causal variable
is level of education. Class position is less important, and it is here characteristicthat

teristicthatboth the workers and the old middle class are located on the right; consequently the new middle class is on the left. However, the importance of education is seen from the fact that its correlation with the Old Politics position is r = 0.38 whereas the multiple correlation for all four variables is only slightly higher, r = 0.40.

It is of special interest that the beta effect of age on New Politics position is insignificant even though the crude correlation between age and New Politics position has an r = 0.22. The effect of age is therefore mainly a byproduct the educational effect, a finding indicative of post-industrialism rather than post-materialism.

Thus far, our data are in agreement with the hypothesis that the postindustrial establishes a new political cleavage between those with high and those with low education. But this does not in itself indicate the content of that cleavage. Why does the cleavage show up on these noneconomic rather than, for example, on the older set of issues? And why do the well-educated take the New Left rather than the New Right stands?

The likely answer to such questions takes us beyond the realm of election research and can only be hinted at here. To the extent that work organizations are gravitating from emphasis on production and material distribution towards emphasis on reproduction, administration, and technological they also change from supervised machine-like organizations professional organizations. This may entail a change in work ethics from individualistic competition and hierarchical relationships toward cooperation and egalitarian relationships. In a recent comprehensive Norwegian study, Jenssen (1991) distinguishes two interrelated dimensions of relational values, authoritarianism and individualism. He points out that New Politics issues tend to have a "thematic likeness" with these values such that New Right attitudes are associated with authoritarian individualism, Left attitudes with its opposite, egalitarian collectivism. Jenssen argues that the latter value type, which he calls radical humanism, is taught in higher education but can also be learnt by job experience and selection for positions in the new professional organizations. In contrast to postmaterial such values do not have to be transferred at an early stage; thus they can be shared by the young and older well-educated.

Thus, educational level and content as well as occupational role can serve as proxy variables for values in determining New Politics positions. The choice of education and occupation involves a choice of values, and vice versa. In a survey of Danish youth, Svensson & Togeby (1992) combine educational level with private vs. public employment so as to single out three new class groupings. The "new middle class" and the "yuppies" are characterized by high education and position in the public and private sector, respectively; the "plebs" are defined as those who have low educationand

cationandno permanent position on the job market. While the two former class groupings are shown to score high on political tolerance (roughly corresponding to our New Politics index), the plebs score low. On an index of egalitarian attitudes (roughly corresponding with our Old Politics index), the new middle class as well as the plebs are shown to be on the left, whereas the yuppies are on the right.

Our data from the 1990 election permit a distinction among various social groups. In Fig. 3 we map the position of four educational groups as well as a dozen different occupational groups on the Old and New Politics dimensions.

Groups defined by their educational level describe an almost straight,
vertical trace-line through the diagram. That is, people with different levels
of education differ with respect to the New Politics dimension but not with

respect to the Old Politics dimension. Those with 7 years of schooling place themselves on the New Right while those with 12 years of schooling are on the New Left, the difference between these two groups amounting to more than four index positions.

With respect to occupational groups, the old class divide separating the unskilled and skilled workers from the self-employed business class shows up as an almost horizontal axis in the middle of the figure. Above this old class axis, the farmers and those without occupation (mostly retired persons) stretch upward toward the New Right. Below it, the different white-collar groups spread out no less than four scale points downward, i.e. toward the New Left.

The large size of the white-collar group makes a subdivision possible along sectoral lines. The result is a picture consistent with the expectation that organizations with reproductive rather than productive objectives, and with non-profit objectives shielded from competition on market terms, favour the ideas of the New Left. Those forming the staffs in institutions of teaching and culture are found at the bottom of the figure, that is, to the extreme left of the New Politics dimension. They are followed by the personnel in the health and welfare sector, the students, and public administration personnel. Still higher up we find the white-collar personnel of private production and service companies. Finally, the personnel of sales and transport companies, who are most directly exposed to market forces, unfold right on the old class axis between the workers and the self-employed business class. The only occupational group which breaks the pattern is the small group of defence and police personnel; because of their preoccupation with physical safety, they cannot be expected to join the New Left inclinations other public service institutions.

Conclusions

New Politics theory states that in economically advanced nations, a set of non-economic issues will form a new dimension of political conflict. In this article an attempt has been made to formulate some hypotheses related to New Politics theory and to apply them to the Danish 1990 election. Our analysis has proceeded at three levels of political conflict, (1) the level of policy attitudes, (2) the level of party preferences, and (3) the level of social stratification and class groupings. A multilevel analysis was thought necessary in order to study policy attitudes in their context of political behaviour and social change. A fourth level, that of values, has only inferentially been drawn into the analysis as the Danish election surveys do not include items on values.

The case of the Danish 1990 election confirms some theses about New

Politics while disconfirming others. Above all it confirms the main thesis that on a number of non-economic issues of the New Politics variety, policy positions tend to form a distinct attitudinal dimension that is separated from (though not uncorrelated with) the economic Old Politics dimension. We also confirmed the overriding importance of the level of education in the formation of New Politics positions: the well educated tend to take positions on the New Left and the less educated, on the New Right. Among those with secondary or higher education, especially the new middle class of modern welfare societies (employed in health service, social work, education, and other reproductive organizations) constitutes the core of the New Left. The Danish case disconfirms that the New Politics dimension, defined in terms of attitudes, is necessarily connected with the rise of new parties or with splits in the established major parties. The positions of party groups on New Politics attitudes by 1990, in conjunction with the parliamentary development of the preceding decade, point toward a differentconclusion: competitive multiparty system (i.e. with parties strugglingfor and for larger vote shares) can be expected to accommodatethe Politics issues and even to take an active part in crystallizing positions on these issues. In particular, small and medium-sized parties that are not obliged to represent a national or religious minority may "mutate" into New Politics parties without necessarily giving up their roles in Old Politics.

REFERENCES

Adorno, T. W. et al. 1950. The Authoritarian Personality. New York: Norton.
Bell, D. 1973. The Coming of Post-Industrial Society. New York: Basic Books.

Borre, O. 1974. "The Danish Protest Election of December 1973", Scandinavian Political
Studies 9, 197-204.

Dalton, R. J. 1988. Citizen Politics in Western Democracies: Public Opinion and Political
Parties in the United States, Great Britain, West Germany, and France. Princeton, N.J.:
Chatham.

Dalton, R. J. 1991. "The Dynamics of Party Systems Change", in Reif, K. & Inglehart, R.,
eds., Eurobarometer- The Dynamics of European Public Opinion. London: Macmillan.

Flanagan, S. F. 1987. "Values in Industrial Society", American Political Science Review 81,
1303-1319.

Hildebrandt, K. & Dalton, R. J. 1978. "The New Politics: Political Change or Sunshine
Politics?", in Kaase, M. & von Beyme, K., eds., German Political Studies: Elections and
Parties, pp. 69-96. London: Sage Publication.

Inglehart, R. 1971. "The Silent Revolution in Europe: Intergenerational Change in Post-
Industrial Societies", American Political Science Review 65, 991-1007.

Inglehart, R. 1977. The Silent Revolution: Changing Values and Political Styles Among the
Western Publics. Princeton, N.J.: Princeton University Press.

Inglehart, R. 1990. Culture Shift in Advanced Industrial Society. Princeton, NJ: Princeton
University Press.

Jenssen, A. T. 1993. Verdival. Ny massepolitikk i Norge [Value Choice. New Mass Politics
in Norway]. Oslo: Ad Notam Gyldendal.

Lipset, S. M. & Rokkan, S. 1967. "Cleavage Structure, Party Systems, and Voter Alignments:
An Introduction", in Lipset, S. M. & Rokkan, S., eds., Party Systems and Voter Alignments.
New York: The Free Press.

Miller, W. E. & Levitin, T. E. 1976. Leadership and Change: Presidential Elections from 1952
to 1976. Cambridge, Massachusetts: Winthrop Publishers, Inc.

Muller-Rommel, F. 1990. "New Political Movements and 'New Politics' Parties in Western
Europe", in Dalton, R. J. & Kuechler, M., eds., Challenging the Political Order: New
Social and Political Movements in Western Democracies. Cambridge: Polity Press.

Rusk, J. M. & Borre, O. 1974. "The Changing Party Space in Danish Voter Perceptions,
1971-73", European Journal of Political Research 2, 329-361.

Stouffer. S. A. 1955. Communism, Conformity, and Civil Liberties. New York: Doubleday.

Svensson, P. & Togeby, L. 1992. "Post-Industrialism and New Social and Political Classes",
in Gundelach, P. & Siune, K., eds., From Voters to Participants. Århus: Politica.

Touraine. A. 1974. The Post-Industrial Society. Tomorrow's Social History: Classes, Conflicts,
and Culture in the Programmed Society. London: Wildwood House (first published 1969).