Resumé

The topic of the article is a comparison of voter reactions in Sweden and Norway towards some current political issues. The analysis is based upon nationwide election surveys in 1976 (Sweden) and in 1977 (Norway). Despite the fact that the party systems are different in several ways, the patterns of political attitudes in the two electorates are surprisingly similar. The analysis suggests that political cleavages are basically the same in these two neighbouring countries. Likewise the location of corresponding parties in the two systems is approximately the same in relation to specific cleavages.

The topic of the present paper is to compare reactions of voters in Sweden and Norway towards some current political issues. The two countries reflect considerable similarities with regard to social structure and political institutions. In both countries some 95 percent of the population belong to the Lutheran State Church. And except for tiny Lapp minorities in the North, both countries are characterized by ethnic homogeneity. Economic and social differences are left as the main areas of political dispute. Both nations are constitutional monarchies and unitary states with a parliamentary form of government. The electoral systems are almost identical, applying a modified version of the Sainte-Lagué method of proportional representation. Furthermore, the structure and functioning of mass media in the two countries have much in common: in both, for example, radio and television are run as a state monopoly.

The framework for political contest is rather similar in Sweden and Norway. However, the content of political dispute in any given society depends upon the character of competing electoral alternatives. A comparison of the party systems of Norway and Sweden reveals striking similarities as well as interesting differences. Due to variations in historical experiences earlier stages of party development were not quite parallel in the two countries. But around 1920 the Scandinavian five-party core model was established in both (Berglund & Lindström 1978). The parties

were from left to right: the Communists, the moderate Socialists (Social Democrats), two parties at the centre - Liberals and Agrarians - and finally the Conservatives. However, subsequent developments turned out to be quite different. Sweden in fact is the only Nordic country which has maintained the Scandinavian five-party model. The only exception is a tiny Christian Democratic party which emerged in 1964. Not only has the number of parties remained stable in the Swedish system, the proportion of the electorate supporting each one of the parties is basically the same today as 50 years ago. Indeed, the Swedish party system reflects a stability during recent decades which is perhaps unequalled in the world. Norway on the other hand has experienced substantial changes. The establishment of the Christian People's party in 1933 marks the first deviation from the five-party model. Another split occurred in 1961 when a leftwing group broke away from Labour to form the Socialist People's party which ideologically was located between Labour and the Communists. Finally, in 1972-73, after the intensive dispute over Norway's entry into the European Communities, fragmentation occurred in several parties (Valen & Rokkan 1974). The old Liberal party was split and reduced to two mini-parties, Liberals and the New People's party. On the left there was both a split and a merger: an anti-EC faction broke out from Labour and joined the Communists and the Socialist People's party in a Socialist Electoral Alliance (SV). Finally, a populist protest party emerged on the rightwing, named Anders Langes party (for the reduction of taxes and public expenditures). In 1976, after the death of the founder, the party changed its name to the Progressive party (FRP).

A common feature of Swedish and Norwegian politics is a strong labour movement. Since the 1930's the Social Democrats have been dominating parties in both countries, occupying government office for most of the time. As for the strength of other parties, considerable variations are evident between Sweden and Norway. Table 1 describes the support for various parties at the parliamentary elections of, respectively, 1976 and 1977. The most striking similarity is found in the division of the vote between socialist and non-socialist parties. It is almost identical in the two countries. But: among bourgeois parties the differences are great. Jointly, the parties at the centre enjoy much greater strength in Sweden than in Norway, while the Norwegian Conservatives are nine percentage points ahead of their Swedish sister party.

The question now is: how are variations in partisan alternatives in
Sweden and Norway reflected in the formation of public opinion? Because
political parties tend to set the agenda for political debate, one should

expect similarities as well as differences between the two electorates in
their reactions towards current political issues.

At first sight, the political agenda in Sweden may appear to be quite different from that in Norway. Whereas Norwegian politics during the 1970's has revolved around the EC and oil issues, Swedish political debate of recent years; has concerned itself mainly with nuclear energy and wage earners' funds. Nevertheless, a closer investigation will disclose similarities of major significance between the two nations. A great number of political issues have been the subject of debate in both countries. This fact became evident when the Norwegian election survey of 1977 was in the planning stages. In its eventual formulation one will find that approximately one-half of the attitudinal questions included were drawn directly from the equivalent Swedish survey of 1976. More specifically, there are 15 questions that are exactly identical. Another four are comparable, but differ somewhat in their wording and response alternatives. These 19 questions cover a broad spectrum of current political issues, and it is this data base which will help us illustrate some similarities and differences between Swedish and Norwegian party politics.

We have presented the 19 attitudinal questions in Table 2: because of space limitations, some of the questions have been abbreviated in the table, which shows in percentages the response distributions, plus three summarizing statistical measures. The first of these is an index showing the balance of opinion: the proportion of positive answers minus the proportion of negative ones. 'Don't know' and ambivalent answers are

included in the percentage base, but not in the index. The other two statistical measures are arithmetic means and standard deviations. These measures have been calculated under interval scale assumptions, whereby the values 1 through 5 have been ascribed to the response alternatives: 'Don't know' responses have here been excluded from our calculations.

The questions have been ranked in Table 2 according to their degree of similarity between Norwegians and Swedes. First we present those questions where the distributions of opinions in the two countries are highly uniform. The further down the list, the greater the differences in attitudes between Sweden and Norway. As a summarizing measure of the degree of similarity we employ the absolute amount of difference between the balance of opinion figures.

The question on which the Norwegian and Swedish electors are in closest correspondence is the proposal to build more day care centres for children - the balance of opinion is exactly the same. With regard to the thought of reducing taxes on low incomes we find almost the same correspondence of attitudes; an overwhelming majority supports this idea in both countries. The same is true with respect to the question of furthering equalization in terms of incomes and conditions of work. On these matters there is a widespread inclination toward a favourable view of equalization in both Sweden and Norway.

Out of the 15 directly comparable questions, the one on which attitudes diverge most sharply is the idea of equal representation of women in parliament through sex quotas. In Norway 44 per cent are negative towards this proposal, but only 20 per cent in Sweden. Another area where opinion differences between Norwegians and Swedes are particularly pronounced is defence issues - in Sweden there are comparatively more respondents who wish to reduce defence appropriations. Also with respect to pornography there are noticeable differences between the countries: more Norwegians than Swedes would like to prohibit all forms of pornography.

The principal picture which emerges through this form of comparison of the response distributions is that there are no major differences between the political climates of Norway and Sweden. On the whole, the responses in the two countries approximately coincide. On some questions, where divergences are more substantial, the latter may possibly be related to particular circumstances in the respective countries. The more favourable disposition of Norwegians toward their military defence may be explained by Norwegian membership in NATO. The somewhat more positive attitudes of Swedes toward nuclear power may be related to the opinion-formation

activities of the Swedish Social Democratic Party. The more averse reaction against pornography in Norway may be traced to the fact that Norwegian culture, as regards moral issues, is somewhat more conservative. But such interpretations should not be carried too far - it is not possible to find any general tendencies to the effect that one nation is more radical, conservative, or favourable toward decentralization than the other.¹

So far we have only examined the questions individually. By an analysis of pair-wise relationships we gain the possibility of a closer investigation of attitude structures in the two countries. If we subject the ensuing correlation pattern to factor analysis, we may discern a number of more general attitude dimensions. Tables 3 and 4 are presentations of two factor analyses: one from Sweden in 1976, and one from Norway in 1977.² The figures in the tables, the factor loadings, express the correlations between a particular question and an underlying factor. These values are our guide in interpreting the political contents of the factors.

In both Sweden and Norway, the first and most important factor explains about 20 per cent of the total variance, and roughly 60 per cent of the common variance. Not only does it have equal importance in both countries; interpretations of its contents lead to nearly identical results. The factor loadings have the same structure in both countries. The only exception is the question of social reforms, which loads considerably higher in Sweden, but the difference might well spring from differences in measurement techniques. It is Indisputable that factor number one captures the left-right dimension, measures opinion conflicts between voters who favour a great deal of government influence over economy and social life versus those who wish the opposite. Given the 19 attitudinal questions, this left-right dimension takes on a striking predominance. It accounts for more than twice as much of the total variance as any of the other factors.

The government-influence dimension, however, is not the only ideological cleavage in the electorate of either Norway or Sweden. The remaining three factors provide important complementary information about opinion cleavages. Admittedly, the interpretation of factors is not always self-evident. To ascribe to a factor a general political content, frequently summarized under a single category or label, is in effect to simplify reality. But in this particular case we find ourselves in a favourable situation. With respect to both election surveys we have previously carried out factor analyses on all of the collected material, not only on that particular subset of questions which are common to both countries. The outcome of these factor analyses has provided detailed descriptions of attitude structures,

with which the present solutions can be compared. The Swedish analysis contained 29 questions which yielded 5 interpretable factors. In the Norwegian case, 35 variables gave as a result 6 different factors (Petersson 1977, Table 3.11; Valen 1978, Table 3).

The question which loads highest on factor two in Sweden concerns nuclear energy. Attitudes toward nuclear power also characterize the second factor in the original Swedish factor analysis. However, in the reduced solution in Table 3 this factor is not distinguished on the basis of nuclear power alone. Questions of energy conservation and new sources; of energy also attain relatively high factor loadings. The characterization of the original factor analysis as purely a nuclear energy dimension is due to the fact that three separate interview questions were contained in the survey. The correlations between these three questions were sufficiently high for the constitution of a separate factor. But when, as in this case.

only one nuclear energy question is included, then other and related questions will also load relatively high. Even though the factor now warrants a somewhat broader interpretation, it is still obvious that it captures an underlying dimension which has to do with positioning on issues of energy policies. One end of this scale represents a strongly positive attitude toward nuclear power combined with skepticism toward energy conservation and alternative sources of energy. At the opposite end we find nuclear power opponents, ardently supporting a new set of energy policies.

In Norway also the second factor is characterized by energy questions.³ The pattern is the same as in Sweden. Nuclear power is also in Norway attitudinally opposed to energy conservation and alternative sources of energy. Furthermore, the extended analysis had previously demonstrated that this dimension also captures controversies in questions of oil exploitation

in the North Sea. The factor loadings are not exactly identical in Norway and Sweden. The question of discontinuing unprofitable railroad services has a high loading in Norway, somewhat lower in Sweden. However, despite certain minor differences, the overall picture remains convincing: in both Sweden and Norway there is a separate attitude dimension in matters of energy and environment policies.

Even the third factor has similar political contents. In both countries it represents attitudes toward equality. Obviously there are also in this case certain national peculiarities, yet one is struck by the similarity between the two countries. What is represented is not only equality with regard to socio-economic matters and equality between groups of employees, but also sexual equality. In both countries the questions of women representatives in parliament and of day care centres for children have high loadings. This interpretation of the factor is supported by the fact that it also appeared in the original analyses, those based on the entire material in the two countries.

However, the fourth factor is, on the face of it, more difficult to interpret. Particularly in Sweden, it seems that questions of various contents have high loadings. But on the background of previous analyses we venture an interpretation, fully conscious of its less-than-solid foundations. The fact that the pornography and day-care-centre questions have high loadings, on opposite sides, leads us to the conclusion that in both countries this factor reflects differences in something which we, for lack of a better term, may call cultural outlook. In Sweden this attitudinal dimension corresponds to those two factors which in the previous analysis were designated moral and social policies. In the Norwegian analysis the equivalent factor has previously been referred to as a liberal view of culture. A liberal, modern permissive cultural attitude is represented on this scale, as its opposite - a morally conservative and traditional view.

This latter factor reflects a cultural difference between generations. Attitudes on these types of moral issues are strongly correlated with age. Young respondents express liberal opinions, whereas older people are more conservative on moral issues (Table 5). With respect to the remaining three attitude dimensions, age differences are considerably less pronounced. True enough, there is in both countries a tendency that younger respondents are more critical toward nuclear energy and favourable toward equalization than their elders, but these correlations are weak. And with regard to the most important political dimension, government influence, age differences are quite negligible.

Thus far we may summarize our impressions from the comparison of Sweden and Norway. Indisputably, the predominant impression is one of similarity. It is hardly surprising that certain minor details in the factor solutions diverge. All the more striking is the fact that marginal distributions , correlation patterns, as well as factor structure and age correlations, are so similar. Those cleavages we may observe in one country, are also in evidence in the other. Despite historical incongruities and despite the fact that on the surface political developments have at times been rather different, there is nonetheless a great deal of common ground in these two political systems. The political poles, between which political life in both Sweden and Norway unfolds, are government versus private control over the economy; industrial centralization/growth vs. environmental protection/decentralization; conservation vs. active elimination of social cleavages; and liberal vs. conservative views on cultural and moral issues.

In our analysis thus far we have considered the electorate as a whole. But behind the averages there may be hidden considerable differences between sub-groups. An analysis of cleavages in political systems of the Nordic type would be incomplete if it did not address itself to one such type of collectivities: the political parties. The role of political parties in the democratic process is decisive, not only as options at elections, but also through their important role in the process of opinion formation. The parties are not just passive listeners to the expressed opinions of the electorate: they also take an active part in the formation of attitudes.

What differences of opinion, then, can we find between the supporters of the various parties? Tables 6 to 9 present an overview of opinion divergencies between the parties in Sweden in 1976 and in Norway in 1977. Tables 6 and 7 present the balance of opinion figures for the 19 questions. The figures may be compared with Table 2, which gives the equivalent

information for the entire electorate, including non-_voters. Tables 8 and 9 are based on the four general attitude dimensions, which have been described in Tables 3 and 4. Tables 8 and 9 show attitude averages within parties. These means have been computed from factor scores derived from the four factors.

Tables 6 and 7 provide us with ample opportunities for detailed comparisons between our two countries. Opinions may here be compared, not only between parties within one country, but also between parties in different countries. At this point one should recall that four of the questions are not comparable between countries. In Tables 6 and 7 these have been marked by asterisks. The remaining 15 questions are, as we have previously mentioned, identical.

The nuclear energy question presents an interesting example of a possible comparison. In Norway there is hardly any disagreement between the parties; they are all decidedly negative toward nuclear energy. The only exception is the Conservative (H) voters, who, though preponderantly negative in their views, assume a somewhat more sympathetic position. The pattern in Sweden corresponds to the Norwegian one in that the supporters of most parties are critical, with Moderate voters somewhat more positive. But the great exception is the Social Democrat party. In Sweden this party had taken a position in favour of nuclear energy before the 1976 election, conducting an intensive information campaign for its views, thereby pulling its supporters towards more favourable opinions on nuclear energy. In Norway, the nuclear energy issue has not been politicized in this manner. The results are evident in our data: the balance of opinion index for the Social Democrats (S) is -2, for the Labour Party (A) -64.

Tables 8 and 9 present average positions for each party on the four ideological dimensions from our factor analysis. As one would expect, the attitudes toward government influence distinguish sharply between parties. In Sweden the average scores vary from +116 (VPK) to -83 (M), in Norway from +119 (SV) to -83 (H). The parties fall along the classical left-right dimension. Not only is their order parallel in the two countries, but the average figures for 'sister parties' are amazingly similar. The Moderate and Conservative voters express attitudes which place them at exactly the same index figure on this left-right scale. VPK and SV also assume a practically identical position; the difference between the Social Democrats and the Labour Party is also negligible. The remaining 'parties of the centre' average in Norway between -28 and -42, and in Sweden -33 and -43 (KDS in Sweden -14). With respect to the most important political

cleavage, the left-right dimension, Norwegian and Swedish voters are in
essence faced with the same party system.

As regards location on the second factor, the energy and environment dimension, the similarities between Sweden and Norway remain great, but party conflicts form somewhat divergent patterns in this case. In Sweden the position as the most 'environmentalist' and anti-nuclear-energy party is occupied by the Centre party (and KDS). In Norway it is SV which has the equivalent extreme position. At the opposite pole there is in Norway a cluster of three parties: the Labour Party (+10), the Conservatives (+8), and the Progress Party (4-13). In Sweden the Social Democrats form incontestably the most pro-nuclear-energy party (+32). In Sweden the energy policy battle, particularly the nuclear energy issue, has been fought with somewhat different alignments from those which at present characterize the Norwegian party system on these issues.

The conflict patterns along the third dimension, the equality factor, are similar in both countries. As in the case of controversies over government influence, it is the left-right dimension that dominates. One notable exception is the Norwegian Centre Party, which ends up close to one pole of the scale, with the same average index as the Conservatives.

The distinguishing feature of party conflicts over cultural and moral
issues is the fact that a single party in each country assumes an extreme
position. This is the religious party - in Sweden KDS and in Norway KRF

- whose supporters have extremely conservative views. Between the other parties, however, differences are not significant; for them, as previously shown, the gulf in attitudes between generations is the most important factor. In the case of most voters, year of birth is a better predictor of opinions on moral and cultural issues than party affiliation.

As evidenced in this analysis, it is obvious that certain parties resemble each other quite strongly, that their supporters have approximately the same attitudes. Other parties, on the other hand, seem to be opposites; their attitude profiles are entirely contrary to each other. These vaguely based impressions may find exact quantitative expressions. As we shall now see, the preceding description of attitude structures can form the basis of a pair-wise comparison between the parties of Sweden and Norway. This comparison, in turn, permits a configuration analysis of the Norwegian and Swedish party systems. Dimensional analyses of party systems are nowadays common in political science. But to our knowledge, this is the first time anybody has carried out one common dimension analysis of the party systems of two countries.

Our configuration is based on the data presented in Tables 8 and 9, the party averages on our four factor indices. We take advantage of the fact that our factor analysis yielded parallel results in the two countries, that the four factors are pair-wise comparable. Thus, Tables 8 and 9 may be regarded as a data matrix. Our four factors constitute the units of analysis. The 14 political parties (6 in Sweden and 8 in Norway) form the variables. Each party can now be compared to any other party. One must only provide a measure which yields a high positive value if the attitude profiles of the two parties are highly congruent. Inversely, this measure should have a strongly negative value if attitudes diverge sharply, and it should take a value close to zero when attitude profiles are neither very similar nor very dissimilar. One measure which fulfills these criteria is the wellknown product moment correlation coefficient. A system of coordinates is formed for each pair of parties; the four data points - the attitude scores - are plotted in; and the correlation is computed. Naturally, four data points

are not much, but in this case they should suffice. Significance problems are small, as this analysis is based on a universe in its entirety. As regards the relative weights of our four attitude dimensions, they will vary from party to party. In the case of a party which assumes an extreme position on any factor, this data point will be located far away from origin, and hence it will take on a comparatively great weight in the computation of correlations. Reciprocally, when a party occupies a centre position on a scale, this dimension will influence the correlation coefficient to a relatively lesser extent. Thus, factor 4 has the greatest significance for the extreme: parties KDS 2tnd KRF, and less for the other parties.

In Table 10 we have listed the highest positive correlations, identifying those parties which are most ideologically akin, according to the opinions of their electors. Also shown are those pairs of parties with the highest negative correlations, those parties which are furthest apart.

The results so far are hardly surprising. The closely related parties are either non-socialist parties, such as the Conservatives and the Progressive Party (+ 0.985), or socialist parties, such as VPK and SV (+ 0.967). Those parties which are characterized by the greatest distance, which constitute the extremes of the party system, are VPK and the Moderates (-0.965), and SV and the Conservatives (-0.963). Thus far, all is familiar. It sounds as if the old left-right dimension should be able to summarize the correlation pattern. It is possible to subject this conjecture to a more rigorous test. One may, for example, enter the correlation matrix into a factor analysis. The results are shown in Table 11.

The result of this configuration analysis is not one, but three factors.
The first factor, however, is by far the most important, accounting for
nearly two thirds of the total variance. It captures the left-right scale: at

the far left we find SV (factor loading -0.992) and VPK (-0.935), at the extreme right the Conservatives (0.985), FRP (0.966), and the Moderates (0.918). But the second dimension also provides important information, it complements the picture of the party systems. The ideological similarities and dissimilarities between the supporters of various parties cannot be totally captured by the left-right model alone. In the two-dimensional map in Figure 1, based on the factor matrix in Table 11, we notice the existence of two distinct party blocs: the socialist and non-socialist blocs. Within these blocs, however, there are interesting variations, as the second dimension reflects ideological opinion differences which chiefly concern energy and environment issues as well as cultural and moral questions. The polar ends are constituted by, on the one hand the Social Democrats in Sweden, and on the other hand a cluster of four parties: the Swedish Centre Party, the Norwegian Centre Party, KDS, and KRF. The rurally dominated and the religiously oriented parties are thus set off in a separate group, whereas the more 'modern' and urban bourgeoisie are represented by the Conservatives, the Moderates, the Liberal Parity (in Sweden), and the Progressive Party. The Norwegian Liberals (V) and New People's Party (DNF) do not quite belong to either of these clusters, but float somewhere in between, though certainly still on the non-socialist side.

Thus there emerges a common picture of the party systems of Sweden

and Norway, a picture which confirms the principal antagonism between left and right, which illustrates certain partial deviations from this main model, but which above all demonstrates the kinship of the Norwegian and Swedish party systems.

---

NOTES

1. These 15 questions may be classified on the one hand according to the conflict dimension which they reflect (further below), on the other according to the degree of opinion similarity between the two countries (Table 2). These two properties are not related to each other.

2. This table presents the factor matrix of an orthogonal solution after varimax rotation. In a preparatory analysis Kaijser's criterion gave five factors in both Sweden and Norway. Since the fifth factor was uninterpretable, the number of Factors was limited to four.

3. It should be emphasized that for reasons of comparability the second and third factor in Norway have been interchanged. In our factor analysis the energy policy factor originally came out number three. However, the eigenvalues of factors two and three are relatively equal both in Sweden and Norway.

REFERENCES

Berglund, S. and Lindström, U. 1978. The Scandinavian Party System(s). Lund: Studentlitteratur.

Petersson, O. 1977. Väljarna och valet 1976. Valundersökningar. Rapport 2, Statistiska
centralbyrån. Stockholm: Liber.

Valen, H. 1978. 'Konfliktlinjer og stridsspørsmål i norsk politikk 1977'. Paper presented at
the Nordic Political Science conference, Sandbjerg, Denmark.

Valen, H. and Rokkan, S. 1974. 'The Election to the Storting in September 1973'. Scandinavian
Political Studies 9, 203-210.