A substantial literature over the years has focused on the relationship between political alienation and ideology. Much of this research contends that conservatives are more alienated than liberals because philosophically they believe that the best government is that which governs least. A close reading of this literature, however, reveals little consistency in the empirical findings. Both Wolfinger et al. (1964) and Rohter (1969), for example, support the argument that right wing conservatives tend to be the most alienated. Similarly, Wright (1976, 208) argues that "... the alienated are generally more conservative, more isolationist and more anti-communist than the population as a whole". On the other hand, Citrin, et al. (1975, 29), suggest that liberals are the most alienated. They conclude, "In sum, the conservative style and tradition promote political allegiance, whereas liberalism encourages criticism of political institutions, if not outright alienation". picture is even more confusing because Miller (1974), as well as Lipset and Schneider (1983), find that both liberals and conservatives, especially those on the extremes, are politically more alienated than the people who think of themselves as moderate or middle of the road.

These inconsistent and contradictory findings forestall any general or
definitive conclusion about the relationship between self-defined liberal or
conservative ideology and political alienation. In part, these contradictory

Resumé

A substantial literature over the years has focused on the relationship between political alienation and ideology. Much of this research contends that conservatives are more alienated than liberals because philosophically they believe that the best government is that which governs least. A close reading of the literature, however, reveals little consistency in the empirical findings. Survey data from Norway, Sweden, and the United States are used to provide a more extensive and consistent test of the hypothesis. Ideology is defined as both left/right self-identification and policy preference on economic and "new politics" issues. The evidence reveals that in Scandinavia higher levels of alienation are found among conservatives, whereas in the US the left has been consistently more alienated, except on "new politics" issues. The discrepancy between the citizen's preferred ideological orientation and that which the public perceives the government to take, is used to explain the different findings for the three countries and the shifts in the relationship between ideology and alienation across time.

findings arise because the previous studies were limited to a single time point or a restrictively small population, such as the Citrin et al. (1975) study which focused on only the Bay area of San Francisco. In addition, these previous studies were limited in that they focused largely on crosssectionrelationships liberal/conservative ideology and alienation rather than focusing on the dynamics of how this relationship changes across time. We suggest that utilizing both longitudinal and cross-cultural data, as well as a theoretical orientation focused on changes in governmental activities rather than shifts in the public's ideology, helps explain the contradictory findings of previous research.

Our main concern when examining the relationship between ideological self-identification and political trust is to determine whether ideology helps explain shifts in trust across time.¹ First we examine the relationship between left-right, or liberal-conservative self-identification and trust. In the second part of the article we investigate the content of left-right ideology and determine whether various meanings of left-right are differentially correlated with trust. We discuss both the traditional meaning of left-right in terms of support for government intervention in the economy, and more recent questions of "new politics", such as women's issues and concerns for the environment.

Before proceeding to consider the relationship between political ideology and alienation, however, we wish to present the overall trend for political trust in Norway, Sweden and the United States respectively. As illustrated in Figure 1, trust showed a particularly abrupt decline in all three countries between the mid-1960s and the mid-1970s (the survey questions measuring trust are presented in the Appendix to this article). After that, support recovered in Norway, while it continued to decline in Sweden and the United States. In the last two countries the downward trend was temporarily halted - between 1980 and 1984 in the United States and between 1979 and 1982 in Sweden - but resumed thereafter. Between 1985 and 1989 Norway also witnessed a marked decline in trust. Nevertheless, the Norwegian government remains far more positively evaluated than are the national governments of Sweden or the United States.

The similarities and differences in the trend for each country over the entire period for which data are available suggest that there were some common factors influencing trust early on, but that there were system differences which allowed Norway to escape from the more enduring deterioration in support which has afflicted both Swedish and American politics. During the years between 1964 and 1973 there were world events and social problems which affected all of these countries. These included the Vietnam War, growing Stagflation, increasing criticism of the welfare system, concerns about: taxation, and concern about the environment. At the same time, there were also issues that on the surface seem unique to each country, such as Watergate in the United States, the EEC vote in Norway, and regional politics and the nuclear power controversy in Sweden, which could help account for the trends.

An examinination of the alternative explanations for political trust in each country would be enlightening would help to promote the development of a general theory of perceived government responsiveness (for more on this see Miller & Listhaug 1990) Here the goal is more modest. The focus is not a general explanation Of the trend, but rather an examination of how political ideology contributes to change in political trust.

Explanatory Models

Two models must be considered when explaining the change in political trust. The first treats shifts in the aggregate distribution of ideological identification as a source of variation in trust levels. This model assumes that there is some static correlation between trust and ideological selfplacement,for that trust is lowest on the right or at the political extremes. If, for some reason, the public moves in a centrist direction, this should contribute to an increase in trust levels in the aggregate as the

relative numbers on the political fringes decline. The empirical support for this model would depend on the tendency of the distribution of left-right ideology to change over time. One might suspect that left-right ideology, or liberalism-conservatism as it is referred to in the US,² would form a relatively deep-rooted characteristic in the population, and that major changes over the actual time span in ideological balance is unlikely. This can, however, be decided only by examining the real empirical trends in the measures of ideology.

In the second model the key factor involves the ideological distance between the government in power and the governed. As the political and ideological bases of governments shift, various segments of the mass public might feel closer to or more distant from the political system, depending on their degree of agreement with the ideological leanings of the incumbent party. Governments do not normally represent extreme ideological positions, implement extreme policies, but seek a course on the moderate left or right. This tendency could lead to persistent political alienation on the far left and the far right. According to the second model, on the other hand, in periods with governments of the left in power, we would expect trust to increase among those who identify with the left and decrease on the right. In periods of Conservative dominance of government, by comparison, the opposite pattern would be expected to occur. Assuming that governments alternate between the left and the right, we might expect some equilibrium in trust levels to be achieved over time. But this would depend on the symmetry of the relationship, and how other factors affect trust. The focus of this article will be on the second model, but we will on occasion also assess the empirical evidence for model one.

Alienation by Left-Right Self-Identification

In Tables 1-3 we present the basic relationship between political trust and left-right self-identification for the three countries across time.³ In all cases, lower levels of trust in government are indicated by lower values of the percentage differences for the trust index.

Norway

In general for Norway, the lowest levels of political trust occur among those who think of themselves ideologically as farther to the right, although there are some differences in this general pattern across time (see Table 1). For example, the generally lower level of trust in 1973 was a reflection of the political volatility, evident during the preceding years, that resulted in unstable governments of shifting ideological leanings. The 1969 election

had returned the bourgeois coalition to power by the narrowest margin. But as talks with the European Economic Community (EEC) about a possible Norwegian membership progressed, this coalition broke down, primarily because of Center Party opposition to the talks, and a minority Labor government took power in 1971. Following the defeat of the membershipproposal the national referendum of 1972, the Christians, the Center Party, and the Liberals formed a minority government.

The elections of 1973 and 1977 returned Labor to power and they were to stay in control until 1981, but only because of an alliance, albeit a weakening one, with the Socialist Left. On the basis of the second model outlined above, this shift in power would lead to the prediction, which the evidence confirms, that the recovery from a low level of trust in 1973 would be strongest among the center-left and that these people would also display the highest levels of trust for the period from 1973 to 1981.

The 1981 election in Norway, however, marked the culmination of a shift to the right that characterized the second part of the seventies. The Conservative Party gained strength and was able to form a minority government after the election. In 1983 the Christian People's Party and the Center Party joined the Conservatives in a majority coalition that was to remain in power until 1986. Thus, the 1985 election marked only the second time in the post-war period that the non-socialist parties continued to stay in power at consecutive elections (the first was 1969). This constituted a

major shift in power and, according to the second model, would lead to a
reversal of the relationship between ideology and trust.

In fact, the overall trust levels remained fairly stable on the left (they increase somewhat on the far left and decrease marginally on the centerleft), among those identified with the right trust improved markedly, with an especially strong recovery among the small group who located themselves on the extreme right position of the scale. Following the shift of government power to the bourgeois parties, and the ensuing policy shifts in a free market direction (Listhaug 1986), trust levels were highest on the right, with even those on the far right markedly more trusting than citizens on the extreme left.

The three-party bourgeois government only lasted until 1986 when the Progressive Party, which was not a member of the government coalition, shifted its support to Labor in a crucial vote on the budget. The Labor Party subsequently ruled with a minority government for the remainder part of the electoral term. The non-socialist parties, however, came back into power in 1989, following an election where the big winners were the Progressives and the Socialist Left. At this point the problems of forming a viable non-socialist government - that were so apparent for the 1985-89 term - continued, and by 1990 a new minority Labor government was back in power. These latter developments, however, are outside the time frame of our study.

The decline of trust in Norway from 1985 to 1989 was strongest in the center and on the right (see Table 1). The sharpest decrease in trust occurred on the far right. The deterioration of political trust in 1989 must be interpreted as a response to the inability of the non-socialist government to function effectively, and the ultimate replacement of the government led by Kaare Willoch of the Conservative Party by a Labor cabinet led by Gro Harlem Brundtland. Given that the decline was much sharper on the far right than in the center suggests that the former group felt more ideologically alienated from the government. That trust fails to increase on the left in response to the shift of power to Labor may be understood as a response to the inability of Brundtland to implement a program much different from that of the non-socialist parties. This situation arose because she had to rely on at least one of these parties to survive in parliament. Moreover, the 1989 election returned the bourgeois parties to power, making that election less clear as a case of shifting power.

Sweden

Political life in Sweden, more than in the other Nordic countries, has been
dominated by the Social Democratic Party. In the years covered by our
data, the Social Democrats were out of power only for the period 1976-82,and

82,andthis was the first time since 1932 that the Swedish Social Democratic Party did not hold the reins of power. The Swedish Social Democratic Party has also been considered somewhat more radical than its Norwegian counterpart (Madsen 1984). From the second model outlined above it might therefore be assumed that the Swedish party would accommodate those with ideological positions on the far left more easily than the NorwegianLabor and that the peak of trust during periods of Social Democratic reign should occur among those further to the left than in the Norwegian case.

When discussing the findings for Sweden, it is useful to group the elections into three categories: (1) years with Social Democratic dominance of government; (2) years with non-socialist dominance; and (3) transition elections respectively. 1968,1973,1985, and 1988 constitute the first group, as the Socialists were in power both before and after these elections. By comparison, 1979 is the one election of non-socialist dominance inasmuch as the government was formed by non-socialist parties both before and after the election. The remaining two years in our data series (1976 and 1982) are transition elections.

In 1968 the relationship between trust and ideological self-placement was generally linear with a decreasing level of trust as one moves from left to right (see Table 2). In 1973 trust levels were still highest on the left, but the pattern was conspicuously weaker than in 1968 because of a sharp decline of trust on the left. This decline cannot be explained by shifts in government power since the Social Democratic Party remained in office, but could suggest that the far left had become increasingly disenchanted with the prevailing social democratic regime. Trust continued to deteriorate in 1976, the election in which the long-term government hegemony of the Social Democrats was broken.

The effects of the transfer of power should be seen more fully in 1979 when the non-socialist parties had enjoyed the opportunity to make their mark on government goals and policies. As seen above in Figure 1, trust declined more sharply in Sweden between 1976 and 1979 than in any other electoral period. Given this situation, it would perhaps be too demanding to expect trust to increase on the right, but it would still be interesting to discover whether the relative decline was weaker here than on the left. Even this weaker prediction, however, was not supported by the data (see Table 2) because trust deteriorated at about the same rate among those who located themselves on the left (in positions 0-4) as among individuals on the right (positions 6-10).

A possible explanation for trust failing to increase on the right was the inability of the non-socialist government to meet conservative expectations as an effective alternative to the long-term leftist dominance. The original three party government of Fålldin broke down in 1978 and was replaced

by a government of the small Liberal Party headed by Ullsten. A new three-party government, again headed by Fålldin, was launched after the non-socialist victory in the 1979 election. Later the Conservative Party left the ruling coalition in 1981, but the government stayed on with the Liberals and the Center Party for the remaining part of the period. In 1982, however, the Social Democrats recaptured the government and remained in power until 1991.

Another comparison which deserves attention in Sweden is that between the ideological underpinnings of trust in 1979 and 1985. The transfer of government power back to the Social Democrats should, according to the second model, enhance trust on the left. In contrast, an increasing ideological from government should have depressed trust among those on the right. Taking the difference in PDI values between 1979 and 1985 on the left and right respectively, there is at least some support for this prediction: all positions on the left, with the exception of position one, showed a weak increase in trust levels, and all but one of the positions on the right exhibited a somewhat stronger movement in the opposite direction.⁴ Finally, we note that trust levels remained fairly constant across the ideological in 1988. a year that again returned the Social Democrats to power.

United States

The United States, like to Norway and Sweden, has also seen a dramatic
growth in the size of the central government and the welfare state since

the realignment of the 19305. We might hypothesize, therefore, that conservativeswould be the most alienated of US citizens. Unlike Norway and Sweden, however, the United States, during the four decades of the fifties through the eighties, experienced frequent periods of divided government at the federal level. Often the presidency was controlled by Republicans while Democrats ruled congress, thereby making predictions about the relationship between ideology and trust more difficult. Yet, one prediction which is reasonable because of the divided government in the United States, is that the US pattern of relationships should be less consistent(less across the years than that observed for the other two countries. Furthermore, given that the president is a very visible policy initiator in the United States, we might predict that shifts in the relationship between ideology and trust will correspond to a change in the partisanship of the president.

The empirical evidence only partially supports these predictions. Contrary the Scandinavian countries and to the earlier United States studies (Wolfinger et al. 1964; Rohter, 1964), liberals in the United States tend to be more alienated than conservatives (see Table 3). Apparently, political distrust in the United States does not reflect a long-term disdain for an active central government, although "extremely" conservative people are consistently among the most alienated. Likewise, "extremely" liberal people are also consistently among the most alienated because they feel that government is not active enough.

When attempting to explain changes in the level of trust in the US, the shifts in the relationship between ideology and trust did correspond to a change in the partisanship of the president (see Table 3). While the Nixon/ Ford administration was in office liberals were relatively more distrusting. After Carter came to office in 1976, distrust grew disproportionately among conservatives. By 1980 those on the right were more alienated than most liberals, except for those on the extreme left. A moderating of distrust

among those on the right did occur after 1980 when a much more conservativeReagan/Bush came to power. Nevertheless, it was only in 1984 that trust became the dominant mood among conservatives. While distrust among conservatives during the eight years of the Reagan administration was not as deep as it was during the Carter years, Reagan failed to boost the confidence of conservatives any higher than where it stood after Watergate and the resignation of Richard Nixon.

In brief, a relatively persistent relationship between a conservative ideology alienation was found for Norway and Sweden. The relationship for the United States, by comparison, was less consistent across time. The shifts that appear in the relationship between ideological orientation and distrust for all three countries suggest that the observed cross-section relationships are less a function of ideological leanings and more a reflection of the perceived policies pursued by the government. When the policy orientation of the government changed, the relationship between trust and ideology also shifted. However, an entrenched liberal or conservatively oriented government did tend to produce a predominance of distrust among one ideological group more than another.

Alienation and "Traditional" vs. "New Politics" Ideology

To infer the ideological leanings of an individual from self-placement on a left-right scale is always rather dissatisfying, since location on an abstract scale does not indicate substantive knowledge about the content of ideology. that ideology, even cast in terms of left and right, can have many meanings, it would be useful to examine the effects of ideological content on trust. This is particularly important since recent literature has suggested that the possible alienating effects of a left ideological position might be associated with a "new politics" dimension rather than aspects of "traditional" leftist beliefs (Jennings & van Deth 1990). One reason for this is that while leftist governments, especially those of the Scandinavian social democratic variety, might successfully accommodate traditional concerns control of the economy, a high level of taxation, a large public sector, and a strong welfare state, they might not be as successful in securing the solution to questions raised by environmental politics, women's issues, and an assortment of minority group demands.

To further probe the ideological foundation of trust we therefore constructedtwo of left-right ideology, one tapping a traditional meaning, the other designed to capture aspects of "new politics". In contrast to left-right self-placement, which is measured in a standardized manner across elections, the construction of policy-based scales is much more

difficult to carry out in a way that guarantees reasonably reliable comparisonsacross Not unexpectedly, we have also had greater problems in finding measures for the "new politics" dimension than for the traditional aspects of left-right ideologies, since the latter issues have a longer history of creating political conflicts and ideological cleavages. Hence, more than for the preceding analysis we need to be cautious in making claims from these data.

Alienation and Traditional Left-Right Ideology

For Norway we constructed a traditional left-right policy index by adding the values of two items; one tapping support for the welfare state and one measuring the degree to which the respondent favored progressive taxation (a relatively higher level of taxation on high personal incomes). The resulting index yields a scale from 2 (respondents who want state social insurance to be expanded and say that taxes on high incomes should be raised) to 8 (respondents who want state social insurance to be cut and are strongly against higher taxation on high incomes).⁵

Overall, we find that the distribution of trust by traditional left-right policy positions in Norway resembles that of left-right self-identification. This finding is consistent with an interpretation that traditional policy content underlies left-right ideology among Norwegian citizens. It is interestingto that in the period 1973-81 there was no discernible tendency for trust to be markedly lower on the far left (see Table 4). With the shift of power to the non-socialists in 1981, and the ensuing policy shifts

by a Conservative-led government, trust rose on the right in 1985 while trust declined on the left. Following the breakdown of the bourgeois government in 1986 and the installation of a Labor cabinet, trust declined among all groups independent of their positions on the left-right policy scale. But the decline in trust from 1985 to 1989 was larger on the right than on the left.

For Sweden the analysis of the impact of traditional left-right policy concerns is somewhat more robust than in Norway since we have been able to construct a policy index based on three equivalent items across the entire time period 1968-88, thereby yielding a consistent left to right scale.⁶ The content of this index is similar to that used for Norway.

Our primary task here, as in previous sections, was to analyze how the relationship between trust and ideology shifted over time, not with the correlation of ideology and trust in any year. Nevertheless, it is fascinating to note the almost linear relationship between trust and ideology at the start of the period, as seen by PDI values in 1968 (see Table 5). These values are strongly positive among those taking policy positions in favor of the traditional goals of the left. On the other end of the ideological spectrum the PDI values are negative. The strength of the relationship in 1968 is compatible with an interpretation that the long term social democratic hegemony had significantly alienated those holding policy positions in opposition to those enforced by Swedish political authorities.

It is somewhat puzzling that trust declined more sharply on the left than on the right from 1968 to 1973 although the left remained in power. The finding for the policy scale is at this point similar to what we found for leftright (cf. Table 2). We do not have a good explanation for this shift, although we might speculate that the resurgence of a new radicalism in the youth movement and in sectors of the labor unions might have contributed to increased cynicism on the left. The data for those under 30 years of age reveal a much sharper decline in trust than for older people, thus providing some evidence to support this speculation.

Returning to the main hypothesis pertaining to shifts in the ideological bases of trust in government, we may compare the trust PDIs for 1973 and 1979. In the former election the Social Democrats remained in power, and in the latter year, the non-socialist parties were able to retain their hold on government. As we found for left-right self-identification in Sweden (see Table 2), there was no improvement in trust from 1973 to 1979 among those holding policy positions compatible with the goals of the non-socialist parties. It is interesting to note, nevertheless, that while the decline in trust was very sharp among those holding policy positions on the left and in the center, the decline was much lower, or non-existent, on the far right. This suggests that at least the shift in the political basis of government did have some influence on the relationship between trust and ideology in Sweden.

Following the return of power to the Social Democrats in 1982, we should, according to model two, expect an improvement of trust on the left, and a decline on the right. This expectation is supported by the data (see Table 5). From 1985 to 1988, with the government still controlled by the Social Democrats, the ideological basis of trust remained fairly constant and little change in trust occurred during that period. In general, therefore, while the Swedish case presents a few unresolved puzzles, the evidence suggests that shifts in traditional left-right policy concerns do partially account for the longer-term trends in political trust.

For the United States, a more consistent association is observed when the relationship between a traditional, economically based measure of leftright trust is examined than that which was found above with the general, liberal-conservative self-definition measure. Unfortunately, as Abramson (1990) has pointed out, the American National Election Studies have few survey items that can be used to measure economic orientation. Nevertheless, using an item that asks about whether or not the federal government should guarantee everyone a job and a good standard of living provides a measure for the US that is quite comparable to the economic orientation measures used in Norway and Sweden.

Using the guaranteed jobs indicator, we find that those favoring an active
central government were consistently more distrusting across the years than
those who prefer that each individual is responsible for his/her own standard

of living (see Table 6). In contrast to the Scandinavian countries, Americans opposed to government programs that ensure individual economic security were generally more trusting of government. This was true in every year except 1980 when there was no significant difference between those favoring or opposing government guaranteed jobs.

Across time this relationship is more stable than that found for liberalconservative Only two noteworthy periods of change appear in the data of Table 6. First, between 1972 and 1974 a sharper increase in distrust occurred among those preferring individual initiative than among those favoring government assistance. Part of this change may have been the reaction to Watergate, but that scandal cannot account for the disproportionate shift among those opposed to an active government. An additional explanation may have been the realization that without Nixon in the White House, the possibility of limiting the growth of the welfare state would be reduced.

The second major shift was the disproportionate growth in trust among those on the right between 1980 and 1984. This change more straightforwardly to be related to the policy change from the Carter to the Reagan administration, a point we document more fully below. But before moving on, however, it is important to note that during the entire period from 1972 to 1988 relatively little change in trust occurred among those more supportive of the welfare state. Even under Jimmy Carter these people were predominantly cynical. In comparison with Norway and Sweden this result suggests that the perceived policies of the US welfare state are insufficient to satisfy liberals, even under a Democratic administration.

Alienation and 'New Politics" Left-right Ideology

While the analysis so far demonstrates that support for, or opposition to, an active role of government in economic and social life, does influence the development of trust over time, the relevance of ideology is not fully exhausted by keeping within the boundaries of such questions. Another aspect of left-right ideology that has received increasing attention in recent years involves feminist issues and environmental protection. Since issues of relevance for this dimension have been less extensively covered in the election surveys, the data are, unfortunately, rather meager. This is especially true for the Swedish case where the analysis is limited to 1976 and 1979. Yet, this was a period marked by a precipitous deterioration in political trust, thereby making the incomplete analysis for Sweden more relevant. The analysis for Norway, by comparison, is somewhat more extensive, covering the period 1981-89.

To address this issue, a "new politics" index was constructed for both Norway and Sweden, by combining items tapping support for women's issues and protection of the environment. When constructing the index, individuals who were in favor of the women's movement, or who said that more should be done to protect the environment, were scored as left; those taking the opposite positions were scored as right. Creating the same index for the US is somewhat of a problem because the surveys lack a variety of items tapping the new politics dimension. A single item focusing on a broader role for women in government and industry was therefore used. Fortunately, this item was available for all years between 1972 and 1988, with the exception of 1986.

The data in Table 7 demonstrate that "new politics" preferences only weakly influenced trust in Norway. There is some tendency that those at both extremes were the least trusting. In 1989 this pattern was limited to the far right. As for the trend in trust, which is our principal pursuit, information about position on the "new politics" scale contributes little to an understanding of the change in trust from 1981 to 1985. In contrast to left-right policy position (see Table 4), which basically showed no movementin PDIs on the left and a slight improvement in trust for the center and right, a small improvement in trust is observed in 1985 regardless of position on the "new politics" scale. This suggests that "new politics" concerns did not play a significant role in assessments of the non-socialist government. The Labor government of 1986-89 reflected many of the goals articulated by the "new politics", setting a record for the number of cabinet positions held by women, and actively supporting pro-environment policies.

Despite this, there was no increase between 1986 and 1987 in trust among those with the most leftist policy positions on the "new politics" scale. But, as we observed for traditional left-right issues, the decline in trust from 1985 to 1989 was somewhat stronger on the right than on the left.

The analysis for Sweden includes only the elections of 1976 and 1979. The scale was constructed by combining items that tap support for women's issues and opposition to nuclear power. The latter issue was especially salient in the 19705, creating tensions within the established parties, and between traditional channels of political participation and unorthodox forms of influence that emerged during the campaign of the referendum on nuclear power.

In contrast to what we found for traditional left-right policy issues in Sweden in 1976, which showed trust to be highest on the left, the "new politics" scale shows that trust is lowest on the left (see Table 8). The change in trust patterns from 1976 to 1979 was also slightly different from that of traditional left-right ideology, which showed the decline in trust to be slightly stronger on the left than on the right. The opposite is true for the "new politics" dimension. Here the decline in trust was somewhat weaker on the left than on the right. This could suggest that the nonsocialist was perceived to be more receptive to "new politics" concerns of the left than to traditional leftist positions regarding the economy the welfare state. This is consistent with the fact that important parts of the non-socialist coalition, notably the Center Party, were closer to the environmental concerns of the new left than was the Social Democratic Party.

For the United States the relationship between the "new politics" dimensionand
is quite fascinating insofar as it is exactly the opposite of that
found earlier with the welfare state policy dimension. Across the years

those opposed to an equal role for women in government and industry have consistently been the most alienated (see Table 9). Also, the pattern where those on both the extreme left and right are the most cynical, is less evident with the "new politics" dimension. Indeed, only those who most strongly favor having women stay at home were regularly the least trusting. Furthermore, shifts in the relationship between alienation and the "new politics" dimension appear to have been influenced much less by Watergate or change in the partisanship of the administration. From this we can only speculate that the impact of "new politics" on the trend in distrust has been weaker than that of the more traditional economic concerns. In addition, the data suggest that most liberals have found "new politics" policies more satisfying than have conservatives.

Concluding Discussion: An Explanation

In summing up the analysis we may note the following points. In all three countries the ideological positions of citizens were related to their degree of alienation from government. Those who locate themselves on the ideological tend to be most strongly alienated. This is most likely explained by the fact that fringe groups, whether of the left or the right, cannot easily find their ideological positions rewarded by any government.

In general, we also find that for Norway and Sweden, but not the United States, those on the left are more trusting than citizens on the right. This pattern was most striking for Sweden; even in the period of non-socialist governments in the 19705, citizens on the left remained somewhat more trusting than those on the right. In Norway, on the other hand, the emerging control of government by non-socialist parties in the first half of the 1980s produced a reversal of the ideological correlates of trust. This shift is important because it contradicts the view that conservatives should be inherently distrusting of government and politics. Norway in 1985 and the US for the years 1972-76 and 1982-88 provide convincing cases that this is not a necessity. However, considering the long-term dominance of Social Democrats in the governments of Norway and, especially, Sweden, it is reasonable to expect that distrust would be stronger on the right in these countries. The difficulty that the non-socialist parties had in breaking the hegemony of the left, indicated by the instability and internal conflicts within their coalition governments, suggests that the inability of the conservatives govern effectively when they do grasp the reins of power might be the major reason for the continued political alienation on the right.

Finally, the limited evidence we have for the "new politics" dimension,
suggests that this aspect of left-right ideology might have a somewhat

different relationship to trust than traditional left-right concerns. The strong correlation between a leftist "new politics" position and cynicism in Sweden in the 19705, which is in sharp contrast to the pattern for traditional left-right ideology, might signify that Swedish governments have been less successful in accommodating pro-women and pro-environment policies. The much weaker relationship in Norway and the United States in the 1980s, with the most consistent distrust on the far right, by comparison, suggests that governments in these countries had adopted significant parts of the "new politics" agenda, and that groups most strongly in opposition to women's issues and environmental policies had become politically alienated.

How shall we interpret both the apparent similarities and discrepancies in the pattern of relationships between left-right orientation and trust in the three different countries? At the outset of this article we suggested two models for explaining change in political trust. The first focused on the relative distribution of liberals and conservatives in the population. The second involved the correspondence, or lack thereof, between government policies and citizen preferences. The foregoing discussion has given greater attention to the second model. The available empirical evidence tends to support this emphasis.

Although the cross-section relationships between different measures of left-right ideology and political trust are generally significant for all three countries and for a majority of the time points, there is virtually no evidence to suggest that a change in the population distribution of left-right can explain the trends in political trust presented in Figure 1. In fact, the proportion of liberals and conservatives in the three countries varied relatively little over the period from the late 1960's to the end of the 1980's. The average left-right positions for citizens in all three countries using the three different measures of political ideology are presented in Table 10. These data verify that for all three countries there was little consistent and systematic change in left-right self-identification.⁷

Likewise, the preferences reflected by a more traditional economic or social welfare dimension remained relatively stable across time (see Table 10). The overall social welfare preference of the population in Sweden, however, did undergo a significant change across time: there was a general although not monotonic shift toward the right. Since conservatives in Sweden were generally more alienated, this population shift in a conservative could explain to some extent the long-term trend in declining trust for Sweden. However, the year-to-year shifts in trust and traditional left-right attitudes were not highly correlated. Clearly, while increased anti-social welfare sentiment in Sweden may have contributed to the broader growth in distrust, it alone cannot account for the dynamics of this change.

In general, the trends for political trust evident in Figure 1 cannot be explained by change in the population's ideological orientation, because that orientation remained relatively stable over time. Neither can the rise of a "new politics" dimension explain the trends in trust. Only the US data on women's equality reveal a! significant change across time, and thus could be considered as a potential explanation for declining trust. However, the US shift in the population's policy preference has been uniformly toward greater equality for women, thereby systematically reducing over time the proportion of people who are most cynical - namely those who want to keep women at home. In short, despite the cross-section relationships between ideology and trust, change in the proportion of liberals and conservatives cannot explain the trends in political alienation.

What, then, does account for the shifts in political trust? We would suggest that a major explanation arises from the interplay between citizen preferences and what the government does or does not do in response to those preferences. Liberal and conservative orientations are important mediating factors that sensitize the public to certain government actions and contribute to how the public evaluates those actions. But, ideology, per se, does not determine political trust. Distrust develops when citizens

perceive a discrepancy between their preferred ideological orientation and that of the government. This model was generally assumed in the above discussion, but without empirical support that directly included measures of the public's perception of government policy and ideology.

Unfortunately there is little available empirical evidence to test the model. None of the Norwegian or Swedish election studies asked the respondents about their perceptions of the government's ideological orientation. the US data are very limited. Respondents were asked where they thought the federal government stood on the liberal-conservative scale only in 1988. Fortunately, the American National Election studies did ask where the respondents thought the federal government was located on the "guaranteed job" question in 1980, 1984 and 1988. Data for these three time points provide at least a partial test of the theory.

If our hypothesis is correct, there should be a correlation across time between distrust and the magnitude of the discrepancy between the average citizen's policy preference and the perceived government position. The empirical evidence necessary to test the theory is presented in Figure 2. This figure presents the mean positions on the guaranteed job seven-point scale for the respondents and where they perceived the federal government to stand in all three years. The figure clearly demonstrates that between 1980 and 1988 there was little change in the respondents' position, but a dramatic shift in where the government was perceived to stand on the guaranteed job dimension. Apparently people saw a clear shift in the government's ideological orientation as the Carter administration was replaced by that of Reagan after 1980. In addition, shifts in the magnitude of the discrepancy between the respondent's own position and that perceived the government corresponds closely to the trend in political trust (The magnitude of the policy discrepancy was .84, .27 and .39 in 1980, 1984 and 1988 respectively. The relative level of trust for the same years, as indicated by the cynicism PDI, was -29, +4 and -16). Figure 2 helps us understand why those on the right would have gained in trust disproportionately during the period from 1980 to 1988. Clearly, the Reagan government was perceived as better representing their conservative ideological

In conclusion, conservatives are not, it would seem, inherently more distrusting of government than liberals. The genuine independent variable in this relationship is the action taken by the government, not political ideology. One's liberal or conservative orientation is an important mediating variable that influences how citizens perceive and evaluate those government actions, but it is not, per se, a cause of political distrust. When a government pursues policies reflecting a particular ideological orientation over a period of time, it runs the risk of alienating those in the public who disagree with this policy orientation. After some time these unrepresented

individuals begin to feel that they have no say in government. This sense of powerlessness becomes particularly acute if a political party that presumablyreflects interests gains office, but then fails to produce any visible change. Within this process of interaction between citizens and government, when expectations go unfulfilled, neither liberals nor conservativesare to a deterioration in political trust.

---

NOTES

1. In this article we use "alienation" and "cynicism" as synonymous and "trust" as the opposite.

2. We treat the terms left/right and liberal/conservative as conceptually equivalent. We are not interested here in the concept of ideology as constraint or cognitive consistency (see Converse 1964). Rather, we are interested in ideological orientation, that is, whether people think of themselves as having a leftist/liberal or rightist/conservative orientation to politics. The measurement approach traditionally used when investigating self-identification with these orientations is to allow the citizen, or survey respondent, to define the terms however they think of them. Generally these definitions, when solicited, involve left or liberal to be associated with stronger central government and support for welfare programs whereas right or conservative is usually understood to mean support for fiscal constraint, weaker central government and less welfare spending. However, in recent years public understanding of these terms has also come to have some reference to lifestyle choices, thus those taking a left/liberal position tend to equate these labels with support for pro-environmental policies, support for the women's movement and a pro-choice preference on the issue of abortion. Definitions of right/ conservative, on the other hånd usually refer to industrial development having priority over environmental concerns, opposition toward the women's movement and a prolife on abortion. There is substantial literature on the topic of ideological selfidentification Niemi & Weisberg 1984, 317-390 for a review of this literature). This literature, in part, demonstrates that these ideological orientations overlap somewhat partisan identification, but that there is by no means a one-to-one correspondence. purpose here is to focus on the broader ideological labels, not partisan identification.

3. The data are from the national election studies of Norway, Sweden, and the United States. The Norwegian surveys were carried out by Henry Valen and Bernt Aardal ol the Institute for Social Research, Oslo, the Swedish studies by Soren Holmberg and colleagues at the Department of Political Science, University of Gothenburg, and the American surveys by the Center for Political Studies, University of Michigan. The data were made available by the Norwegian Social Science Data Services, the Swedish Social Science Data Service, and the Inter-university Consortium for Political and Social Research. For parsimony we report the sample size for each election survey below rather than in each table. No statistics for any cell of the various tables are reported if the sample size for that cell is less than 2 percent of the total sample. Norway Sweden United States 1973: 1225 1968: 2943 1964: 1557 1980: 1614 1977: 1730 1973: 2596 1968: 1337 1982: 1418 1981: 1596 1976: 2652 1970: 1502 1984: 2229 1985: 2180 1979: 2816 1972: 2689 1986: 1082 1989: 2195 1982: 2943 1974: 2505 1988: 1775 1985: 2883 1976: 2859 1988: 2668 1978: 2288

4. An important characteristic of the ideological distributions in the Swedish case is the low trust that is recorded at position five of the distribution. Five is the midpoint of the self-placement scale. The PDI values for this position are distinctively lower than those for adjacent positions on the scale. The most likely explanation for this pattern is that individuals with weak political involvement, and weak, or non-existent, ideological identifications, will tend to be politically alienated. (For more on this see Miller &: Listhaug 1990). Given that the low trust at position five of the left-right scale is consistent across time, it would have little impact on the aggregate levels of trust over time, unless the proportion of the population in that category grows. Since this has not been the case, it does not help explain shifts in trust across time.

5. Note that in 1973 a four-point tax item was administered as compared to a nve-point item for the other years, yielding seven as maximum value on the scale for this year. The distribution ol the policy scales, in contrast to the general self-placement scales, should be expected to change more over time as policy items are responsive to shift in public opinion, party strategies, and current problems in society. This can most clearly be seen for Norway in 1985 as in that election the Labor Party was able to capitalize on the public's concern about care for the elderly and other welfare state issues. This increased concern shows up in the rather sharp shift in favor of leftist policy positions for this year.

---

6. To construct the index we have relied on an analysis of policy items across all elections 1956-88 reported by Gilljam & Holmberg (1990, 275). They performed an extensive set of factor analyses of relevant issues, and constructed a left-right policy index from the three items that loaded most strongly on the left-right dimension for each of the years. With one minor exception, we utilized this index in the analysis for Table 5.

7. The Norwegian case is somewhat ill-suited to a discussion comparing average values across time because the interval of measurement for the left-right scale changed over the years. Nonetheless, the comparisons that we can make unequivocally point to a high degree of stability in the distribution (Table 10). This can be seen by comparing either 1977 with 1981 (in both years a nine-point scale was used) or 1985 with 1989 (where a ten-poinl scale was administered). In the first period, trust stayed about constant; in the second period, trust declined sharply.

ACKNOWLEDGEMENTS

We thank John Engel and Erik Reyerson for their assistance in the data analysis, Jeanne
Stoakes and Peggy Swails for their secretarial help. The research was partially funded by The
Norwegian Marshall Fund.

REFERENCES

Abramson, P. R. 1990. "The Decline of Over-Time Comparability in the National Election
Studies." Public Opinion Quarterly 54, 177-190.

Citrin, J., McClosky, H., Shanks, J. M. & Sniderman, P. M. 1975. "Personal and Political
Sources of Political Alienation." British Journal of Political Science 5, 1-31.

Converse, P. E. 1964. "The Nature of Belief Systems in Mass Publics", in Apter, D. E., ed.,
Ideology and Discontent. New York: Free Press.

Gilljam, M. & Holmberg, S. 1990. Rott, Blått, Grant. Stockholm: Bonniers.

Jennings, M. K. & van Deth, J. W. 1990. "Conclusion: Some Consequences for Systems and
Governments", in Jennings, M. K., et al. Continuities in Political Action. Berlin: de Gruyter.

Lipset, S. M. & Schneider, W. 1983. The Confidence Gap. New York: Free Press.

Listhaug, O. 1986. "The Norwegian Parliamentary Election of 1985." Electoral Studies 5, 79-83.

Madsen, H. 1984. Social Democracy in Postwar Scandinavia. Cambridge, Mass.: Department

of Government, Harvard University. Ph.D. dissertation.
•' Miller, A. H. 1974. "Political Issues and Trust in Government: 1964-1970", American Political
Science Review 68, 951-972.

Miller, A. H. & Listhaug, O. 1990. "Political Parties and Confidence in Government: A
Comparison of Norway, Sweden and the United States", British Journal of Political Science
20, 357-386.

Niemi, R. G. & Weisberg, H. F. 1984. Controversies in Voting Behavior. Washington D.C.:
CQ Press.

Rohter, I. S. 1964. "Social and Psychological Determinants of Radical Rightism", in Schoenberger,
ed., The American Right Wing. New York: Holt.

Wolfinger, R., Wolfinger, B K., Prewitt, K. & Rosenhack, S. 1964. "America's Radical
Right: Politics and Ideology", in Apter, D., ed., Ideology and Discontent. New York: Free
Press.

Wright, J. 1976. The Dissent of the Governed. New York: Academic Press.