Resumé

This article is based on the 1991 Swedish Election Study and sets out to analyse gender differences in voting behaviour. To increase our understanding of the dynamics between the parties and women and men in the Swedish electorate, an analysis of gender-based "party images" is undertaken. Party images refer to policy orientations and the analysis thereof is empirically based on three sets of open-ended questions about the most salient issues in the election. The main purpose is to investigate to what extent women and men who vote for the same party in an election, in their mind, have differing images of the party they cast their vote for. With large differences in party images it can be put into question if women and men vote for the "same" parties even if the party label is the same on the ballots they choose on election day. The results show that for the Social Democratic Party and the Conservative Party there was to a large extent an agreement between the male and female voters in their party images in the 1991 Swedish Election. This was also true for the Christian Democratic Party. Among women and men who voted for the other five parties analysed in the article, a larger degree of gender-based disagreements in the party images was found.

Introduction

Despite many differences, there are parallels between the 1991 Swedish Election and the Presidential Election in the United States in 1980. In the United States, the election of Ronald Reagan as President was the first time since women won the suffrage that a significant proportion of women voted differently from men. Eight percent fewer women than men voted for Reagan, the Republican candidate.¹ This voting difference was later labelled the "gender gap" by the National Organization for Women (NOW). (Mansbridge 1985, 166; Mueller 1991, 24). Since this declaration, the gender gap notion has become popular among scholars and in the public debate.² (Recently, for example, much attention was devoted to women candidates and to women's issues in the 1992 election in the United States.)

In the Swedish Election of 1991, women's representation decreased from 38 percent to 33 percent at the same time as parties with a majority of men among their voters advanced. In the debate after the election, these were the two reasons why women were highlighted as a losing group. Similar to the United States after the Reagan election, a women's organization (in Sweden composed of a network of influential women) has successfully

used the election outcome to set the agenda for further discussion. This
discussion, in various ways, is working towards bringing the situation of
women to the centre of the Swedish electoral campaign in 1994.

However, the awareness of gender-related differences in politics originated much earlier, and it does indeed cover a wider field of political behaviour than the question of party choice. As a background, this article examines various ways of analysing the gender gap. This leads to a discussion of what is meant by "party-choice" in the question of gender and voting. The analyses of the gender gap in voting often stop with the voter's decision on which ballot to choose on election day. The purpose of the empirical analysis in this paper is to move one step further. The influence of gender is not restricted to the choice between different parties. To have a deeper understanding of the dynamics between the parties and the women and men in the electorate, it seems necessary to try to capture to what extent there are different gender gaps within the parties. The empirical analysis is based on data from the 1991 Swedish Election Study.³

Gender Gaps in Voting Behaviour

In the 1980s, the concept of a gender gap had become established among scholars of political science. Starting out as a reference to differences in party preference between women and men, the term continues to have its base in the field of voting behaviour. But, today, "gender gap" has been used in so many different settings that it has become a catch-all phrase referring to a variety of phenomena. In her article Feminists and the Gender Gap, Pamela Johnston Conover (1988) seeks to sort out the discussion by differentiating between four kinds of gender gaps; in levels of mass participation, in electoral behaviour, in partisanship and, finally, in public opinion on political issues.

Studies in both the United States and Europe have revealed narrowing, and even disappearing, differences in levels of mass participation between men and women. As a consequence, this gap currently attracts little attention. More work is being done in exploring the electoral and partisan gaps. In Europe, differences in party choice between men and women have long been recognized in electoral studies. Even though the term "gender gap" has not always been used exclusively, results have shown that the gap can take different forms; both in direction and in magnitude, dependent on country-specific characteristics. Following societal changes, the gender gap in party choice can also change over time. In several countries, including the Scandinavian countries, Great Britain and West Germany there has, in the past decades, been a shift in party choice from women being more Conservative in their vote than men, to men being more Conservative than women.⁴

In the United States, the electoral gender gap has continued to exhibit the same trend as when it was first recognized in the 1980 election; with women more likely than men to support Democratic candidates. However, in the following elections the magnitude of the difference has not reached the 1980 level of eight percentage points. The partisan gap, which refers to differences in party identification, has in the United States had the same direction as the election gap. During the 1980s, women came to make up a majority of Democratic identifiers (Randall 1987, 50-94; Conover 1988; Kenski 1988).

When it comes to the fourth gender gap, the gap in public opinion,
Conovers highlights this as the most important political difference between
men and women. The reasons for this, she states, are as follows:

For one thing, to the extent that they are issue based, the full explanation of the electoral and partisanship gaps becomes dependent upon understanding the gender gap in political preferences. More generally, the gender gap in public opinion may alter the salience of various political issues thus prompting changes in the political agenda. And, as gender gap widens on particular issues, there may be aggregate shifts in public's preferences which in turn create pressure to change the direction of public policy (Conover 1988, 986).

In electoral studies, the focus on differences in attitudes has mainly been in the way they can help to explain differences in party choice. The focus has to a lesser extent been on the political issues as such. However, when the differences between women and men regarding attitudes on political issues have been the main focus, a wider set of questions on the gender gap has been revealed. One is the fundamental question on how to explain gender-based gaps in public opinion. Demographic sources have failed fully to explain such differences and other factors have come to light, stressing the distinctive aspects of women's and men's lives and the values fostered by separate life experiences (Miller 1986; Conover 1988; Gilens 1988; DeVaus & McAllister 1989; Cook & Wilcox 1991; Rusciano 1992; Welch & Hibbing 1992). Other questions concern the way in which gender gaps in the electorate are transferred to the arena of political representation. As women still constitute a small minority in many political forums, the empirical results in this field are quite vague. However, more and more research is being done which concentrates on the ways, and to what extent, women are influencing the top levels of the political arena⁵ (Carroll 1984; Kathelene 1989; Norris & Lovenduski 1989; Rapoport et al. 1990; Skjeie 1992).

A Discussion of the Notion of Party Choice

The distinctions made by Conover are very clear at the analytical level,
and thus many empirical studies of gender differences more or less explicitly

utilize one of the outlined themes, concentrating on understanding it in particular. This means, for instance, that the question of gender-based differences in political attitudes in Swedish studies to a large extent has been explored issue by issue. Recent differences have been found in attitudes towards Swedish membership in the EC, the use of nuclear power, environmental questions and in the attitudes towards the public sector. Women voters in Sweden are clearly more negative than men towards the EC and the use of nuclear power. Regarding attitudes towards environmental questions and the public sector, the gender differences are reversed, with men being more negative than women (Oskarson 1991, 1992; Gilljam & Holmberg 1993, 189).

These issues have almost achieved a separate status in the analysis of Swedish politics, a status which delineates them from other issues seen as more tightly bound to the Swedish party system. However, studies which explore, or explain, gender differences in political attitudes are only of a limited interest if the differences do not at the same time have any impact on political behaviour. And with rather small, however interesting, variations in party choice between men and women in Sweden, studies of attitude differences have been kept at the outskirts of Swedish election studies.

Feminist scholars have reacted against this, and in various ways tried to bring to the forefront gender-based cleavages.⁶ However, if the goal is a wider understanding of the relationship between gender and voting, there is a need to transcend the classifications that are mainly used today in this field.

One way of doing this is to review the concept of party choice. In their classic work Political Change in Britain - the Evolution of Electoral Choice, Butler & Stokes (1974, 338-350) discuss the question of differing "images of the parties" in the electorate. The point is that one and the same party can be pictured in different ways in the minds of different voters. The central question for Butler and Stokes is whether one party manages to shape a more appealing image than other parties, and thus attracts more votes. In a slightly different, but analogous way, the concept of "party images" can be used in the understanding of gender and voting. If men and women hold different "party images", even though they cast their vote for the same candidates, one could question to what extent the choice really is for the "same" party.

An Empirical Analysis of the 1991 Swedish Election

The traditional way to register an eventual gender gap in party choice is to

look at the whole electorate to see how women and men cast their votes. It was in this way that the notion of a gender gap first appeared. A complementary method is to analyse the electorate of each party, to discern to what extent it is constituted by a majority of either men or women or if it is a "gender mixed" electorate. In this way it is possible to classify parties as being more or less "gender homogeneous" or "gender heterogeneous" in their electoral structure. Homogeneous parties are then understood as parties which to a large extent attract male or female voters. Heterogeneous parties refer to those parties which attract votes from both women and men to an equal extent. The ambition here is not to analyse women's and men's attitudes on different issues in order to explain how they did cast their votes in the election in September 1991. Instead, the results of the election are analysed in order to classify the Swedish parties on the gender homogeneous-gender heterogeneous continuum.

The purpose of this classification is only to form a base for further analysis. The main purpose is to focus on the question of gender and voting in order to capture to what extent the images of the parties are uniform or conflicting among women and men in each parties' electorate. The underlying idea is that the images of the parties can be more or less "gender uniform" or "gender conflicting". A gender uniform party is a party where there is a consensus among women and men voters as to how they picture their party. A gender-conflicting party refers to a party where women and men voters have contradictory images of their party. (As for the classification of the parties as being gender homogeneous or gender heterogeneous, it is a question of a continuum between being gender uniform or gender conflicting.)

What, then, is an image of a party? In order to analyse this question empirically I use three questions from the 1991 Swedish Election Study; what the voters consider to be important issues in the election, what are the most important reasons for their party choice and, which party they consider to have either a positive (good) or negative (bad) policy on various issues.⁷ The image of the party then comes to consist of the following three parts; what the voters think are the most salient issues in the election, their own reported reasons for their party choice and the judgements they make about the party policy on different areas. There are other possible ways of analysing party images, for instance in terms of group representation. In this article, however, an issue-oriented image is utilized. By choosing this set of questions, it is possible to capture the differences in attitudes which in the voters' minds connect closely to their party choice.⁸

The analysis is on an aggregate level. Gender differences among each parties' electorate; can take two forms; they can be seen as differences in priorities and as differences in emphases. For an example of priority differences, one could note that women place environmental issues as the

highest priority, whereas economic issues are the highest on men's ranking list. Differences in emphases refer to results, such as 15 percent more women than men think that family issues are important issues in the election.

Gender-Homogeneous and Gender-Heterogeneous Parties

Gender differences in party choice in Sweden have been analysed mainly in terms of men and women voting for different blocs. Up until the 1970s women voted for the bourgeois parties to a larger extent than men, then the difference disappeared, while in the 1985 election it reappeared. However, in the 1980s, the difference has been in the opposite direction, with women voting for socialist parties to a larger extent than men. These differences have never exceeded about 4 percent.

In the 1991 Swedish Election, there was no gap between women and men in the support for different party blocs. The gender differences in voting all emerged within the bourgeois bloc (Gilljam & Holmberg 1993, 185-186).

The Conservative Party and New Democracy were the two parties which attracted more votes from men than women. Five percentage points more men voted for the Conservative Party, and for New Democracy the gap was four percentage points. Women, on the other hand, tended to favour the three other bourgeois parties; the Centre Party, the Liberal Party and the Christian Democratic Party. For each of these three parties the gender gap was two percentage points. The Green Party, which did not succeed in reaching the 4 percent cut-off level for representation, also gained more

votes from women than men, with a gap of three percentage points. These
differences also make themselves felt on the gender structure in the parties'
electorates.

The Green Party and New Democracy were the two most genderhomogeneous parties. Women voters made up two-thirds of the Green Party's electorate. In New Democracy, the magnitude of the gap was about the same, but the direction was the opposite with two-thirds of the party's voters being men. The difference of 14 percentage points in the Conservative Party's electorate, with a predominance of male voters, is of special interest because the party strongly advanced in the election and is today the leading party within the bourgeois bloc. In the Liberal, Centre and Christian Democratic Parties there was a slight predominance of women voters. The two socialist parties, the Social Democratic Party and the Left Party, can best be described as completely gender-heterogeneous parties. The next step is to evaluate whether it was the more heterogeneous or homogeneous parties that were the most gender uniform or gender conflicting in the election.

Gender-Uniform and Gender-Conflicting Parties

The Electorate as a Whole

In the 1991 Election Study, 16 different issues were mentioned for the question concerning the most important issue in the election. On the question of the most important reasons for party choice, one further issue was mentioned, thus resulting in a total of 17 issues. The distribution in the electorate as a whole on these two questions is presented in Table 3.

In the 1991 Election, 25 percent of the Swedish voters mentioned the

environment as being an important issue, which made this the most frequently mentioned issue for the choice of party. Following the environmental issue, most frequently mentioned were employment, social, economic and pension issues. All of these issues were mentioned by at least 20 percent of the voters. The three most common issues mentioned as reasons for party choice were social, economic, and family policy issues; with social issues mentioned by 14 percent of the voters, and the two others each by 12 percent of the voters. When analysed by gender, the picture turns out to be somewhat different.

For men, the most commonly mentioned issue on the question of important issues in the election was the economy.⁹ This was mentioned by 24 percent of the male voters. Following this the ranking was employment, environment, taxes and social issues. For women, the environment was the most commonly mentioned single issue, as it was for the electorate as a whole. The next four issues for women were social, family, pensions and employment concerns.

Similar differences can be noted on the question of men's and women's
reasons for party choice. For men, economic concerns were the most

important ones, mentioned by 16 percent of the male voters, followed by social, employment, taxes and family issues. For women, social concerns were ranked highest on the list of reasons for their party choice, mentioned by 16 percent of the women voters, followed by issues of family policy, the environment, taxes and the economy.

As discussed earlier, gender differences in issue orientation can be analysed as differences in priorities and as differences in emphases. From Table 5 it can be seen that in response to the question of important issues in the election, women put greater emphasis on issues of family policy, social policy, pensions, and the environment. Men, on the other hand, emphasized economy and foreign politics to a greater extent. Also, on the question of the voters' reasons for their party choice is the noteworthy difference that men put greater emphasis on the economy, while women, again, put greater emphasis on issues of family policy.

Thus, there is evidence of a gender gap in the 1991 Swedish Election, regarding which issues were given the highest priority as salient questions in the election, as well as on reasons for the voters' party choice. The main result is that men gave higher priority to economic issues, and women to

issues of family policy. These two areas were also those where there existed clear gender differences in emphasis. When gender differences are analysed in the electorate as a whole, it would appear that two very traditionally "male" and "female" issues are the most differentiating. Further analysis will show whether this is also true when gender differences are analysed party by party.

How Widespread is the Gender Gap in Attitudes?

To this point, only the issues that are the most commonly mentioned among the voters have been discussed. Of the 16 issue areas which were mentioned as important issues in the election and the 17 which were mentioned as reasons for party choice, quite a few were mentioned by less than 10 percent of the voters. However, it is of interest to take into account the complete list of ranking in the analysis. Spearman's coefficient of rank-order correlation shows how well the different ranking lists "match" between women and men in each party. Included in the analysis now is also the question of in which issue areas the voters think their party has a good policy. The voters were asked to make judgements of the party's policies in eleven different areas.

The high correlations show that the gender gap in issue orientation should not be understood as differing priorities covering the whole spectrum of politicized issues in the election. Of 18 correlations, eight are higher than 0.90. Analysed in this way by party and gender simultaneously, high correlations are not a surprise. It seems plausible to suppose that men and women voters of a party have a set of issue priorities in common, and

therefore their judgements about the policy of the party they vote for is not totally contradictory. There are some additional correlations worth noting. The lowest correlation, 0.58, is for the Left Party on the question of evaluations on the party's policies. Correlations lower than 0.75 are also found for the Green Party on two of the three aspects, and for New Democracy on three aspects. However, these results should be handled with care, since all three are rather small parties.

The correlation analysis does not take into account the salience of the issues being mentioned. Issues mentioned by only a few percent of the voters are treated in the same way as the most commonly mentioned issues. However, the gender differences are of a higher relevance, the higher priority the voters give to an issue. In the following, the analysis focuses on the five top ranked issues on the priority list of: (1) which issues men and women of each party see as the most important in the election, (2) their main reasons for their party choice, and (3) the five policy areas on which they give the party they voted for the best evaluation. Comments are confined to only those areas where the differences in emphases are the largest.

The Left Party

For both female and male voters of the Left Party, environmental and social issues were regarded as the two most important issues in the election. These two issues were also listed first as important reasons for their choice of the Left Party. However, as reasons for party choice, the issues have changed places. Social issues were at the top of the ranking list for both women and men in the Left Party electorate, followed by environmental issues.

Larger differences were not evident until further down in the priority list of the Left Party voters. The most marked difference was that women voters in the Left Party gave higher priority to international issues than men. Foreign policy and international issues ranked as the third most important issue in the election and third as the reason for their party choice for women. For the male voters of the Left Party, these issues did not rank among the top five on either question. As important issues in the election, men, instead, put forward economic and employment issues. And as a reason for their party choice, the employment issue was highly ranked. These differences were also visible in the emphasis on various issues. The international issues (including the EC) were to a larger extent emphasized by women whereas economic issues were emphasized by men. Issues of family policy also constituted an area which was differently emphasized by gender, with a high support level from women.

Regarding in which areas the Left Party voters gave the most positive

evaluations on their party's policy, men voters were the most positive towards the Left Party's employment and taxation policies. Women voters were, instead, more positive towards the refugee policy and the party's environmental policy. In the areas of social security and energy, both women and men voters for the Left Party evaluated the party positively. The gender differences were, as mentioned, mainly on the issues of employment, taxation, refugees and environment.

The Social Democratic Party

The two totally dominating issues for Social Democratic voters, regardless of gender, were employment and social issues. These were the two issues that were mentioned as the most important in the election, and also as the main reasons for party choice of the Social Democratic voters. In addition, these issues were on the top of the ranking list of evaluations. The difference between women and men voters was in the order of importance. Men voters ranked the employment issue higher and also placed greater emphasis on this issue, whereas women voters ranked the social issue highest and emphasized this more strongly.

Apart from this, gender differences among the Social Democratic voters mirrored those in the electorate as a whole. The economy was ranked higher among men voters and was more strongly emphasized, whereas it was the issue of family policy that was ranked higher and more strongly emphasized among women voters. The differences were mainly visible with regard to issues considered as important in the election, with reasons for party choice exhibiting a narrower gap.

In evaluating the areas in which the party had good policies, the main result is that men voters of the Social Democratic Party were more positive towards the party's policy than women voters. This was most noticeable on the evaluations of foreign policy, employment and the EC issue.

The Centre Party

For male voters of the Centre Party, the party was seen as one based on environmental and farming issues, with the emphasis on agriculture. The most clearly dominating reason for men voters to favour the Centre Party was the farming issue. Similar to the male voters, the women also had a clearly dominant reason for favouring the Centre Party, that of the environment. For women voters of the Centre Party, the party was viewed as a party based on environmental issues and family policy issues, with an emphasis on the environment.

The differences between men and women who vote for the Centre Party
were most strongly marked in which issues they mentioned as important in

the election. The women ranked family issues and pension issues higher and with more emphasis, whereas men placed higher priority and emphasis on farming and rural issues. On the question of reasons for their party choice the ranking differences had narrowed. In the emphases they were still visible.

In the list ranking how the voters evaluated their party's policy, there was a very limited difference between men and women of the Centre Party. Placed at the top were environmental, energy and nuclear power, and child care issues.

The Liberal Party

For male voters of the Liberal Party, economy was the most important issue in the election, and the main reason for party choice. For women voters, the economic issue was among the top five as important issues in the election, but both social and environmental issues were ranked higher. The social issue was clearly the most important reason for the women voters for their choice of party. For men, economic issues were both ranked higher and emphasized more strongly, whereas for women, it was the social issue that was ranked higher and emphasized more clearly.

When it comes to mentioning reasons for their party choice, men and women who voted for the Liberal Party agree to a large extent, particularly about the order in which the issues should be ranked. The exception was mainly on foreign policy issues, which were clearly ranked higher among men. In emphases the difference was evident again, with economic and foreign policy issues being more salient for men voters. The women voters did not have one issue of a similar salience, rather they seemed to put greater emphasis on a collection of more "soft" issues.

In evaluating the policies of the Liberal Party, the most obvious picture was that male voters were far more positive toward their party's policies than the female voters. For no other party was there evidence of a corresponding level of positive evaluations by the male voters. For all the eleven different issue areas the men voters evaluated the party policies more positively them the women voters did. Furthermore, for a majority of these areas, the differences were greater than 18 percentage points.

The Conservative Party

As with the Social Democratic voters, there were two clearly dominating issues for men and women voters of the Conservative Party. For the Conservative voters economic and tax issues were the two most important issues in the election. In addition, these two issues were the main reasons

for the voters' party choice, and it was in these two areas that the Conservative
voters gave their party the best evaluations.

However, some differences appeared between men and women voters for the Conservative Party. The gender differences were most noticeable in the emphasis placed on issues by men and women. Pensions, social, family, and environmental issues were to a larger extent emphasized by women voters, whereas male voters emphasized issues of foreign policy, economy, energy and nuclear power. On the question of reasons for party choice, the main difference was that men emphasized the economy more strongly than women did.

The most marked difference in how Conservative women and men voters evaluated their party's policies was that men were clearly more positive towards the party's energy and nuclear power policy than women were. However, this issue was not among the top five on either the men's or women's ranking lists. The EC was another issue where men gave a clearly more positive evaluation of party policy than women voters did. This male level of support for positive judgements was evident in other areas as well; however, it was not of the same magnitude as with the Liberal Party.

The Christian Democratic Party

The issue of family policy was clearly dominating among women and men voters for the Christian Democratic Party. The family issue ranked first as the important issue in the election, as well as first as reason for party choice. This issue also attracted the most positive evaluation from the voters. The gender difference appeared only in that women voters emphasized this issue even more strongly than men did.

Pension, morals (including religion), environmental and social issues followed the family issue on the ranking list of important issues in the election and also as reasons for party choice. Social issues were somewhat more highly prioritized, and somewhat more strongly emphasized among women voters, especially as important issues in the election.

The men and women voters of the Christian Democratic Party also primarily agreed upon their evaluations of the party's policies on different issues. Child care and social security received the most positive judgements, followed by environmental issues for the male voters and refugee issues for the female voters. The gender differences on evaluations were, together with those for the Centre Party, the smallest among the eight parties in this analysis.

The Green Party

The Green Party voters were, naturally, united around the environmental

issues. For men as well as for women voters, environment was the most important issue in the election, the main reason for party choice and the area which drew the most positive judgements from the voters. Over 90 percent of the party's female voters mentioned environment as the important issue in the election and 100 percent answered that they thought the party had a good policy on this issue.

Gender differences among Green Party voters were mainly reflected in that men both gave higher priority and put greater emphasis on issues of foreign policy than women did. For women voters, the issue of family policy was emphasized to a larger extent as an important issue in the election. Regarding the question of reasons for party choice, the main gender difference was that women voters almost entirely focused on the environmental issue, while men mentioned a broader range of issues.

In the evaluations of the party's policies, it was the issues of environment, energy and nuclear power, and the EC that were ranked highest. Women strongly emphasized the environmental issues, otherwise it was the male voters who gave the most positive judgements about the party's policies in the other areas.

New Democracy

While both men and women voters for New Democracy mentioned taxation as the most important issue in the election, the refugee issue was seen as the main reason for party choice and it received the most positive evaluations regarding the party's policies. Taxation was ranked as number four on the list of reasons for party choice for men voters; for women voters it was ranked as number two. As an important issue in the election the refugee issue was not ranked among the top five for women; for men it was number two. The women voters mentioned instead social, family, and pensions issues.

The list ranking in areas which the voters of New Democracy thought
their party had a good policy was headed by refugee policy, and followed
by taxation, economy, and law and order, for both men and women.

Concluding Discussion

The Social Democratic Party, the Conservative Party and the Christian Democratic Party appear to be the most gender uniform parties. These parties had one, or a few, central issues that linked their women and men voters together. For the Social Democrats, employment and social issues provided such a link. The economy and taxation issues were the corresponding linking issues for the Conservatives, and for the Christian

Democrats it was the issue of family policy. The gender differences in party
images for these three parties must be seen as minor in the 1991 Swedish
Election.

The picture of the Centre Party, the Liberal Party and the Left Party, in comparison, reflects a larger degree of gender difference. The Centre Party was for women voters primarily an environmental and family-oriented party, whereas for men voters it was undoubtedly a farmer's party. However, these differences were foremost related to the different emphasis that men and women in the Centre Party put on divergent issues. Nevertheless, there was an agreement on how they evaluated their party's policy which gives an impression of a yet quite well-united party.

For the Liberal Party the situation was somewhat different. The level of support on the male side for positive evaluations about the party's policy was the heaviest among the eight parties analysed. Furthermore, the salience that men voters of the party put on economic issues was not matched to the same extent by any issue that women voters supported.

Gender conflicts were most obvious in the Left Party. Even though the voters agreed upon their placement of social issues as an important issue for the party, the divergences were evident on the issues of employment, taxation, environment and foreign policy. For the male voters, the Left Party was more a party oriented around employment and taxation issues than for the women voters. Instead, the women were oriented towards environmental issues and foreign policy.

The picture of the Green Party and New Democracy is more divided. For women voters of the Green Party, the party appeared to be a one-issue party. The party's male voters seemed to take into account a larger spectrum of issues. New Democracy is even more difficult to comprehend. For instance, what the voters mentioned as important issues in the election were not connected to their reasons for their party choice in the same way as for other parties.

How the continuum between being more or less gender uniform or gender conflicting in the party images, corresponded to the continuum between being more or less gender homogeneous or gender heterogeneous in electoral composition was not obvious in the 1991 Swedish Election.

The two largest parties in the election have been described as gender uniform in their parly images; however, the Conservative Party was leaning towards the gender homogeneous side in electoral composition, whereas the Social Democratic Party was gender heterogeneous.

Among the more gender-conflicting parties in party images were two of the newcomers in Swedish politics, the Green Party and the New Democracy Party. In the election these two parties were also the two most gender homogeneous in their electorate - the New Democracy Party with a predominance of male voters, and the Green Party with a predominance

of female voters. The Christian Democratic Party with slightly more women
than men in their electorate, and also a newcomer in the 1991 Election,
did not, however, have the same kind of conflict in the party image.

In the election the Left Party, the Centre Party and the Liberal Party attracted votes both from men and women to an, almost, equal extent. However, in their gender-heterogeneous electorate there were tendencies towards gender conflict in their party images.

The analysis in this article is based on a restricted set of questions. And within this set of questions it is only a restricted set of issues which gives rise to gender differences. In order to have a deeper understanding of the dynamics of gender and voting, the analysis has to be combined with a wider set of factors influencing the party choice, as well as with theories explaining gender differences in political attitudes. However, the results still demonstrate that in the discussion of gender and voting, one cannot easily assume that women and men vote for the "same" parties even if the party label is the same on the ballots they choose on election day.

---

NOTES

1. The gap of eight percentage points was noted in the National Election Study programme at the University of Michigan.

2. In this article the notion of gender gap is used in order to connect to the gender gap discussion. However, in the empirical analysis it would be more correct to use "sex", as I refer to women and men and not explicitly to the socially constructed gender.

3. The 1991 Swedish Election Study was lead by Mikael Gilljam and Sören Holmberg at the Department of Political Science, University of Gothenburg, and has been reported in Gilljam & Holmberg 1993.

4. Duverger (1955) and Tingsten (1937) are among the classics in this field. For a discussion on changes in the voting gender gap, see for example Björklund (1986), Listhaug et al. (1985), Norris (1986), Oskarson (1990, 240-241) Randall (1986, 1987, 70-73) and Rusciano (1992).

5. Another important area which is not touched upon in this article is how women are influencing the arena of policy output.

6. Jönasdöttir (1985) is a case in point.

7. The exact wording of the questions was as follows: (1) Thinking about this year's election, is there any issue or issues that are especially important to you when it comes to choosing which party you are going to vote for in the parliamentary election on 15 September? (2) You say you are going to vote for (... Party) in this year's election. What are the most important reasons for you planning to vote for (... Party)? (3) I would like to ask you a few questions concerning your opinion about the policies of the parties on various issues. On this card we have put some issues which have been important in Swedish politics in recent years. Is there, in your opinion, any party or parties with a good policy on (... Issue). Is there, in your opinion, any party with a bad policy on (... Issue)? On nos. 1 and 2 it was possible to mention up to five issues, for issues on no. 3, see Appendix. Numbers 1 and 2 only include issues that were mentioned by at least one percent of the respondents. In this study no. 3 only includes the positive evaluations.

---

7. The exact wording of the questions was as follows: (1) Thinking about this year's election, is there any issue or issues that are especially important to you when it comes to choosing which party you are going to vote for in the parliamentary election on 15 September? (2) You say you are going to vote for (... Party) in this year's election. What are the most important reasons for you planning to vote for (... Party)? (3) I would like to ask you a few questions concerning your opinion about the policies of the parties on various issues. On this card we have put some issues which have been important in Swedish politics in recent years. Is there, in your opinion, any party or parties with a good policy on (... Issue). Is there, in your opinion, any party with a bad policy on (... Issue)? On nos. 1 and 2 it was possible to mention up to five issues, for issues on no. 3, see Appendix. Numbers 1 and 2 only include issues that were mentioned by at least one percent of the respondents. In this study no. 3 only includes the positive evaluations.

8. The risk with this approach is that it gives the impression that only a few issues determine the election turnout. Even with leaving out other factors like social background or party identification, this is a simplification. However, with this focus it is possible to analyse attitudes among voters of different parties and of different genders that can be assumed to be among the most politicized in the election.

9. Differences that are stressed in the text are significant at the 0.05 level.

Election Commentary

ACKNOWLEDGEMENTS

Thanks are extended to Mikael Gill jam, Sören Holmberg, Susan Marton and Maria Oskarson
for valuable comments on earlier versions of this article.

ACKNOWLEDGEMENTS

Complete terminology for the 17 issue areas are: environment, employment, welfare/social policies/health care, economy, pensions/elderly care, taxes, family policies/child care, foreign/defence/international/EC, immigrants and refugees, housing, left-right ideology, education, energy/nuclear power, agriculture/rural policies, law and order, religion/moral, equality, public versus private sector. Eleven issues on which the voters evaluated the party's policies as being good or bad: employment, environment, taxes, issues concerning the Swedish economy, energy and nuclear power issues, issues concerning Swedish foreign policy, issues concerning social security, issues concerning child care, issues concerning the European Community, immigration/refugee issues, issues concerning law and order.

REFERENCES

Björklund, T. 1986. "Kvinners og menns partipreferanse", Tidsskrift for Samfunnsforskning
5, 417-43.

Butler, D. & Stokes, D. 1974. Political Change in Britain. London: MacMillan.

Carroll, S. J. 1984. "Women Candidates and Support for Feminist Concerns: the Closet
Feminist Syndrome", Western Political Quarterly 2, 307-323.

Conover, P. J. 1988. "Feminists and the Gender Gap", Journal of Politics 4, 985-1010.

Cook, E. A. & Wilcox, C. 1991. "Feminism and the Gender Gap- A Second Look", Journal
of Politics 4, 1111-1122.

DeVaus, D. & McAllister, I. 1989. "The Changing Politics of Women: Gender and Political
Alignment in 11 Nations", European Journal of Political Research 17, 241-262.

Duverger, M. 1955. The Political Role of Women. New York:UNESCO.

Gilens, M. 1988. "Gender and Support for Reagan: A Comprehensive Model of Presidential
Approval", American Journal of Political Science 1, 19-49.

Gilljam, M. & Holmberg, S. 1993. Väljarna inför 90-talet. Stockholm: Nordstedts Juridik.

Jönasdöttir, A. G. 1985. 'Kvinnors intressen och andra värden", Kvinnovetenskaplig tidskrift
2, 17-33.

Kathelene, L. 1989. "Uncovering the Political Impacts of Gender: An Exploratory Study",
Western Political Quarterly 2, 397-421.

Kenski, H. C. 1988. "The Gender Factor in a Changing Electorate", in Mueller, C. M., ed.,
The Politics of the Gender Gap. Newbury Park: Sage.

Listhaug, 0., Miller, A. H. & Valen, H. 1985. "The Gender Gap in Norwegian Voting
Behaviour", Scandinavian Political Studies 3, 187-206.

Mansbridge, J. J. 1985. "Myth and Reality: The ERA and the Gender Gap in the 1980
Election", Public Opinion Quarterly 1, 164-178.

Miller, A. H. 1986. "Gender Politics in the United States". Paper presented at: ECPR Joint
Sessions of Workshops, Gothenburg, Sweden, April 1986.

Mueller, C. 1991. "The Gender Gap and Women's Political Influence", Annals of the American
Academy of Political and Social Science 515, 23-37.

Norris, P. 1986. "Conservative Attitudes in Recent British Elections: An Emerging Gender
Gap?", Political Studies XXXIV, 120-128.

Norris, P. & Lovenduski, J. 1989. "Women Candidates for Parliament: Transforming the
Agenda?", British Journal of Political Science 19, 106-114.

Oskarson, M. 1990. "Klassröstning på reträtt" in Gilljam, M. & Holmberg, S., eds., Rött
blått grönt. En bok om 1988 års val. Stockholm: Bonniers.

Oskarson, M. 1991. "Kvinnor, män och kärnkraft", Statens kärnbränslenämnd: SKN Rapport
50.

Oskarson, M. 1992. "Kvinnor - den offentliga sektorns försvarare", in Holmberg, S. &
Weibull, L., eds., Trendbrott. SOM-rapport 8. Statsvetenskapliga institutionen, Göteborgs
universitet.

Randall, V. 1987. Women & Politics: An International Perspective. 2nd ed. London: Macmillan
Education.

Randall, V. 1986. "Women and the Left in Western Europe: A Continuing Dilemma",
Western Political Quarterly 2, 307-313.

Rapoport, R. 8., Stone, W. J. & Abramowitz, A. I. 1990. "Sex and the Caucus Participant:

Rusciano, F. L. 1992. "Rethinking the Gender Gap. The Case of West German Elections,
1949-1987", Comparative Politics 3, 335-357.

Skjeie, H. 1992. Den politiske betydningen av kjonn: En studie av norsk topp-politikk. Oslo:
Institutt for samfunnsforskning, rapport 92:11.

Tingsten, H. 1937. Political Behavior: Studies in Electoral Statistics. Stockholm: P. S. King &
Son.

Welch, S. & Hibbing, J. 1992. "Financial Conditions, Gender, and Voting in American
National Elections", Journal of Politics 1, 197-213.