Resumé

This article focuses on women's representation in the Swedish Riksdag. The theory of the politics of presence serves as a point of departure. The aim is to underpin empirically - or to test empirically - the assertion that female politicians, to a greater extent than male politicians, represent the interests of women. The concept of women's interests divides, on a theoretical level, into three components: the recognition of women as a social category; acknowledgement of the unequal balance of power between the sexes; and the occurrence of policies to increase the autonomy of female citizens. On the empirical level this corresponds to measurements indicating female versus male MPs' attitudes and behaviour in areas such as gender equality and social welfare policy. The data used are parliamentary survey studies from 1985, 1988, and 1994. The analysis controls for effects of politicians' gender when other factors - e.g. party affiliation, age, education, and parliamentary experience - are taken into account. The main result is that the theory of the politics of presence gains strong empirical support. What this study contributes is a significant measure of stability for the feminist critique of more established of representative democracy.

Introduction

Does it make any difference if elected bodies are made up of women or men? The feminist research that has emerged within political science during recent years has actualized issues of social representation. There is today a rather widespread critique the nucleus of which is that a male-dominated parliament does not give equal consideration to the interests of male and female citizens. One of the foremost representatives within the field is the British political scientist Anne Phillips. In The Politics of Presence (1995), she formulates the critique as follows:

There are particular needs, interests, and concerns that arise from women's experience, and these will be inadequately addressed in a politics that is dominated by men. Equal rights to a vote have not proved strong enough to deal with this problem; there must also be equality among those elected to office. (Phillips 1995, 66)

---

* Lena Wangnerud, Statsvetenskapliga institutionen, Goteborgs universitet, Box 711, SE 405 30 Goteborg, Sweden. E-mail: lena.wangnerud(« pol.gu.se

This critique is founded upon a number of observations, the most important which have to do with the function of the parliamentary system and the differences in everyday life experiences between women and men. If we begin with the parliamentary system, Phillips points out that the political process can never be planned entirely in advance. Even if bills and programmes have been thoroughly worked out before the parties take a vote, elected representatives still have a certain measure of autonomy in their daily work in parliament:

New problems and issues always emerge alongside unanticipated constraints, and in the subsequent weighing of interpretations and priorities it can matter immensely who the representatives are ... representatives do have considerable autonomy, which is part of why it matters who those representatives are. (Phillips 1995, 44)

In this article, I shall not be further developing the question of degree of autonomy in the parliamentary process. Phillips' point is a reasonable point of departure for delving into matters of social representation. I shall instead devote my energies to the question of gender differences in everyday life experiences. Using surveys among members of the Swedish parliament, I shall empirically underpin Phillips' assertion that the gender of politicians is an important factor in representative democracy.

The results that emerge will not surprise anyone who has followed the feminist field of research. They show, for example, that issues of social welfare are weighted more heavily on the agendas of female politicians than on those of male politicians. We see as well that it is almost exclusively female politicians who pursue issues of equality between the sexes. What this study contributes, however, is a significant measure of stability. The purpose of the article is to present analysis that can provide the field with greater convincing force, which is necessary if the feminist perspective is to become integrated into the core of political science.

Women's Interests: Theoretical Considerations

One of the most common, and yet most controversial, concepts in feministoriented science is that of women's interests. Some of the most important issues in the debate deal with the relationship between objective and subjective interests - is it possible to ascribe interests to women if they have not themselves expressed them? - and the relationship between gender and other social lines of demarcation: do women have interests that are shared across borders such as class and ethnicity?¹

The questions that arise are complex. Several leading researchers have
chosen to deal with them by working with various forms of minimal
definitions of women's interests. Anna Jonasdottir (1991, 156) writes that

it is possible to maintain 'some sort of minimal common denomination:
the interest in not allowing oneself to be oppressed as a woman, or, in
fighting patriarchy.' Hege Skjeie (1992, 39) argues for the feasibility of
'establishing a minimum formulation of common women's interests, seen in
that case in direct relation to public political decisions.'3

The problem with definitions of this kind is that they form a rather weak foundation upon which to build. I believe that two mistakes are easily made here; first, that one will be seduced into taking an unnecessarily defensive posture and, second, that one will fail to differentiate adequately the theoretical the operational level in research.⁴

What is at Stake?

It is important to remember that the concept of interests is not the province solely of feminist research. In her classic work The Concept of Representation, Pitkin (1967, 156) writes that the concept of interests is übiquitous the debate on representation. It is related to a perception that there is something at stake in the representative process. To differentiate interests is a matter of concretizing that which various groups can expect to gain through democratic participation.⁵

One way of approaching the question of what is at stake with respect to women as a group is to illustrate a society not regarded as favourable to the situations of female citizens. If one reads political scientists such as Hernes (1987), Jonasdottir (1985; 1991), Pateman (1988; 1989; 1992), and Phillips (1991; 1995), the following picture stands out as a negation of a womanfriendly a society in which there is a sharp differentiation between private and public; a society where children, the sick, and the elderly are cared for as unpaid work in the home; a society where there is no discussion of the circumstance that this division essentially coincides with a division between feminine and masculine; a society where violence against women and children is seen as a private matter. The authors emphasize women's position of dependency in private life and the inadequacies of their political citizenship. Oppression of women is maintained through the circumscription their autonomy (self-determination).

The picture that emerges in the literature can be likened to a political world of silence. It is also a picture that bears many features of a vicious circle: if the political system does not deal with women's reality, women will not concern themselves with politics and, consequently, the system will not change. Pateman sees the 'democratization of everyday life' as a key to a changed relationship between the sexes (Pateman 1989, 169).

The assumption upon which my study is based is that women as a group
have a common interest in increasing their autonomy. I see the politicizationof

ationofwomen's everyday life experiences as a first step in that process. It
is in women's interests that their experiences, to the same extent as those of
men, are allowed to leave their mark on political decision making.

The question is then how this politicization will come about. We shall once again forge links to pivotal themes within feminist-oriented research. One of these deals with men as the norm in society. The debate on women's interests started as a reaction to the characterization of the political system as gender neutral. Feminist research disagrees. Seemingly neutral categories often conceal a male-dominated reality.⁶ Another pivotal theme is the subordination women and possible measures for its abolition. There is no watertight bulkhead here; rather, the acknowledgement that the political system is not gender neutral may be seen as a first step towards change. A second step would then be acknowledgement that the relationship between the sexes is not equal and a third step would be to present measures designed to change this.⁷

The more fleshed-out definition I wish to make of the concept of women's interests follows the division above. In my opinion, women's interests can be narrowed down to three concerns: the recognition of women as a social category; acknowledgment of the unequal balance of power between the sexes; and the occurrence of policies designed to increase the autonomy of female citizens.

Previous Empirical Studies

I wrote above that the results presented in this article are not surprising. However, it would be wrong to give the impression that research agrees that the gender of politicians is an important factor in representative democracy. Most researchers agree that gender has a certain impact; the issue is how much significance is ascribed to this factor.⁸ One may obtain a thoughtprovoking of diverse views by comparing Pippa Norris and Joni Lovenduski (1995) with Peter Esaiasson and Soren Holmberg (1996).

In their study of the British parliament, Lovenduski and Norris group the members of parliament according to values, priorities, and legislative roles. They find that women have more feminist and leftist radical values than men. Women also award higher priority to social policies and put more emphasis on working with cases taken from their own constituencies. Lovenduski and Norris conclude their review by saying:

Lastly, the evidence suggests that gender influences all three dimensions. This gender difference not be exaggerated, since party proved by far the stronger indicator of values and policy priorities. The gender gap among politicians was modest and not evident on every indicator. Nevertheless, this suggests that the election of more women to Westminster has the potential to make more than just a symbolic difference. (Norris & Lovenduski 1995, 224)

This should be compared with the conclusions of Esaiasson and
Holmberg, who, after a review of policy attitudes among women and men
in the Swedish parliament, sum up their results with the following words:

The main result of our MP study is that the difference between men's and women's attitudes was small on most of the issues investigated. On a small number of issues, however (those having to do with social welfare, morals and the environment), female representatives tended, irrespective of party, to hold somewhat different views from male MPs. Gender is thus associated with attitudes on certain policy issues among Riksdag members. It is not meaningless whether a man or a woman is elected to the Riksdag. (Esaiasson & Holmberg 1996, 37)⁹

Whereas Lovenduski and Norris emphasize that they find differences along all dimensions, Esaiasson and Holmberg stress that their measurements small differences related to a small number of issues. Their concluding words that gender is not meaningless have a different connotation Lovenduski and Norris' talk of potential.

I shall not overemphasize the role that the researcher's own gender may play in this context. However, lan McAllister and Donald Studlar (1992, 403) found only limited differences between female and male parliamentarians Australia: 'In each of these instances, women were found to be very similar in their political orientations to men in the same party, with only policy differences limited to gender-related issues.' However, Sue Thomas (1994, 105) found that women had made a distinct contribution to policy among U.S. politicians at the state level: 'Women's distinctive contribution focused on policy not procedure. They have broadened existing legislative agendas by increasing attention given to issues of women, and children and families.'

The differences in interpretation that exist between various authors highlight the need for stability in research.¹⁰ In order to progress, we must have knowledge of the extent to which results depend upon differences of definition, methods, or something more substantial. It is difficult to make good comparisons without a firm foundation.

Stability in Research

When I seek more stability in research, I am aiming in part to establish a clear anchoring in theory. In this article I wish to arrive at the route by which the experiences of female citizens are transmitted to the political world. The theory of the politics of presence says that female politicians are necessary to create room on the political agenda for women's interests. Is this true?

By seeking stability in research, I am also concerned with methodological
implementation. Many previous empirical studies have been based upon
results from single instances of measuring. Other studies have been made in

contexts where the gender perspective is so pronounced that there is a risk that the politicians - consciously or unconsciously - have couched their replies. Politicians are sensitive to the politically correct norm, and the relationship between the sexes is a loaded question in many Western European democracies.

The Current Study

The current study is based on surveys of Swedish members of parliament on three occasions: in 1985, 1988, and 1994. The studies were done within the framework of the Swedish election study programme.¹¹ The exact questions presented with the individual analyses.¹² Before we get to that point, however, I should like to take the theoretical discussion of women's interests one step further.

The division of women's interests into three components - the recognition of women as a social category; acknowledgment of the unequal balance of power between the sexes; and the occurrence of policies designed to increase the autonomy of female citizens - shall be seen as generally valid. It can be used to study policy at various levels, in different countries or in various epochs. However, in an empirical study each component must be operationalized. resulting indicators may naturally differ depending upon the time and space in which the concrete research project takes place. In this study of the Swedish parliament the following indicators will be used.

Women as a group. The recognition of women as a social category is studied in two ways. First, via the parliamentarians' definitions of their job: do they separate women as a particular group which it is important for them to represent? Second, via the contacts that the parliamentarians with various groups in their political work: are women's organizations included in this network?

Gender equality policy. Acknowledgement of an unequal balance of power between the sexes is studied in a similar manner. This involves the views of parliamentarians on the issue of whether more gender equality is desirable, and the extent to which they pursue issues of gender equality in their parliamentary work.

Social welfare policy. The occurrence of policies designed to increase the autonomy of female citizens is the component that is most difficult to operationalize. However, from my point of view the self-determination of female citizens is greatly dependent upon social welfare policies. In current society, women are more vulnerable than men in the conflict between the spheres of reproduction and production. It is therefore in their particular interest that issues related to this conflict are given room on the political

agenda. Otherwise, we are back in the political world of silence for women
about which several feminist-oriented political scientists have written.

The first empirical analysis, however, will not deal with the significance of gender among politicians, but rather with that among voters. I shall begin by asking one self-critical question: is there actually any reason to speak of differences between everyday life experiences for women and men? And if so, do these differences have any relevance for the political system?

The Policy Areas of Greatest Importance to Voters

What makes interest theories (feminist and otherwise) difficult to manage is that, as a researcher, one risks ending up in an elitist position where one harshly ascribes particular points of view to people. Many political scientists to align themselves with, for example, Robert Dahl's (1989) view of democracy, which is that it is the right of every citizen to have the last word with respect to his or her preferences, priorities, and needs.

The following analysis is based on a question asked in the voter surveys that have been carried out within the Swedish election study programme: 'Thinking about this year's election, is there any issue or issues that is especially important to you when it comes to choosing which party you are going to vote for?' The question is open ended and the respondents may state up to five areas of importance to them. This measurement can be seen as one way of indicating the validity of the definition of women's interests made earlier. Table 1 reports the three most frequently named policy areas among female and male voters between 1982 and 1994.

The results show that female and male voters have similar priorities to a certain extent. Jobs and the environment were two of the policy areas most frequently mentioned by both sexes during the studied period. Yet while there are similarities, several systematic differences emerge as well. Family policy is an important area among female voters, but not among male voters. Family policy was one of the three most frequently mentioned areas among women in 1985, 1988, and 1991, but was not mentioned by male voters at all during the studied period. Social policy and health care is given higher priority by female voters than by male voters. Social policy and health care was one of the three most frequently mentioned areas among female voters in four of the five studied elections (1982, 1988, 1991, and 1994), but only upon two occasions (1988 and 1994) among male voters. The economy and taxes are prioritized more highly by male voters than by female voters. The economy and/or taxes was one of the three areas most frequently mentioned by male voters throughout the entire period 1982-1994. areas do not appear at all among female voters.

The analysis underpins the view that there are grounds for adopting a gender perspective on the political process. We obtain confirmation that the theory of the politics of presence involves pivotal issues in representative democracy. To narrow it down, we can say that it is hardly defensible that the priorities of female citizens are given less scope in politics than the priorities of male citizens. The studies in the following sections will deal with the situation as it currently stands in the Swedish parliament. Are there gender differences there that are similar to those that emerged among voters? The report will follow the operational division into the three components within the concept of women's interests: women as a group, gender equality policy, and social welfare policy.

Women as a Group

The theoretical aspect in focus here is the recognition of women as a social category. This recognition is a first step in the politicization of women's everyday life experience. The indicator that we shall begin with is a question that involves the views of parliamentarians about their representative duties.

The following question was asked in the surveys carried out in 1985,
1988, and 1994: 'How important are the following duties to you personally

as a member of the Riksdag?' The member of parliament was then asked to state his or her position on about ten different duties, such as representing individual voters, the party, etc. Table 2 shows how the duty to 'promote the interests/views of women' was assessed by female versus male members of parliament. The responses are reported in two categories: those who considered the duty very important and those who considered it unimportant ¹³

The results show that there are great differences between female and male parliamentarians in their views about the duty to promote the interests / views of women.¹⁴ Among female members of parliament in all years, more than half answered that this is very important to them personally. In 1985, the proportion was 55 percent, in 1988 56 percent, and in 1994 54 percent. Among male members of parliament, significantly fewer made corresponding In 1985, 10 percent of male members of parliament said that the duty to promote the interests/views of women is very important to them personally, while 9 percent in 1988 and 6 percent in 1994 agreed.'⁵

Before we draw any conclusions from the differences that emerged, we should look at the results from other indicators that concern the component of women as a group. In the studies carried out in 1985 and 1994, questions were asked about what contacts the parliamentarians had made (in their capacities as politicians) with various women's organizations. They were asked to state what degree of contact they had had during the year preceding study. Table 3 reports the responses among female versus male members of parliament in two categories: those who had had frequent contact with a women's organization (at least once a month) and those who had had no contact with women's organizations.

The results show that there are distinct differences here as well. Among the men in parliament, 9 percent in 1985 and 4 percent in 1994 said that they had been in frequent contact with a women's organization. Among the women in parliament, the corresponding figures were 55 percent in 1985 and 51 percent in 1994.

The main result of both of these analyses is thus that in the course of their political work female politicians to a greater extent differentiate women as a special group. More detailed studies have also shown that the genderrelated remains in broad terms when consideration is given to other background factors such as party affiliation, age, education, and parliamentary (Wangnerud 1998, 125ff).¹⁶ We shall now proceed to see whether the results become equally clear when the other two components the concept of women's interests are analysed.

Gender Equality Policy

One way of interpreting the analysis in the preceding section is to say that it measures whether politicians are content with the relationship between the sexes as it stands today: if one believes that things are going well, why should one differentiate women as a special group? We shall now explore this theme further and study the attitudes of parliamentarians towards proposals for more equality between the sexes.

In the surveys, members of parliament were asked to state their opinions
about a number of conceivable future societies: what kind of society did they
consider it worthwhile/not worthwhile to pursue? The question measures a

general ideological approach.¹⁷ In the study carried out in 1994, this suite of items included for the first time a question about the attitudes of members of parliament towards the suggestion 'To work towards a society with more equality between women and men.' The answer lies on a scale that ranges from 0 (very bad proposal) to 10 (very good proposal).

However, the formulation above is rather general and there is a risk that the measurement will seem shallow. Accordingly, I have chosen, in a complementary to pose a number of questions that specifically concern the relationship between women and men in the politicians' own workplace: the Riksdag. In the 1994 study, members of parliament were presented with four different arguments for an equal distribution of women and men in the parliament.¹⁸ The members of parliament were asked to take a position on each argument, designating it as very important, fairly important, not very important, or not at all important. The questions formed the basis of an index where the value 0 means that all arguments for an equal gender distribution were regarded as not at all important and the value 12 means that all arguments were regarded as very important.¹⁹

Table 4 shows the mean values for the general ideological attitudes towards gender equality held by female and male members of parliament and with respect to their attitudes towards the more specific question about gender distribution in parliament. The analysis also took into account the factors of party affiliation, age, education, and parliamentary experience.

We must be cautious with level estimation in reports of this kind - statements there is in an absolute sense a small or large percentage of respondents who are for or against a suggestion. That which we can draw conclusions about with any certainty is to what extent there exist differences between groups. However, I shall deviate from this principle for a moment: the mean value of 9.6 indicates that women in the Swedish parliament are very positive towards a society with more equality between women and men. Among male members of parliament, the corresponding mean value was 8.2, which also indicates a generally positive attitude.

The main result in Table 4 is not, however, this level estimation, but rather that women in all groups studied were more positive about working towards more gender equality than were men in all comparable groups. This applies to both indicators. It should be noted that the spread of responses increased somewhat in the index based upon questions that more specifically conditions in the parliament. My interpretation is that we have come a step further towards bypassing ritual answers where everyone claims to be in favour of equality. The analysis has become more incisive, since it deals with parliament rather than society at large, and through the lessening of emphasis upon the future. The index is also based on weighting of responses to several different questions.

A more detailed analysis shows that gender and party affiliation are the

most important factors when we attempt to explain views on equality betweenthesexes. a regression analysis (see the tables in the appendix), neither the age of the parliamentarians nor the number of years in parliamentyieldedany effect. Education has a certain effect, in that

highly educated members of parliament are more positive towards equality than those with less education. However, this result applies only to male politicians and the effect is negligible when compared with the significance of party affiliation.

Regression analysis also shows that when it comes to the more general question about equality, party affiliation is somewhat more decisive than gender. However, the opposite is true when it comes to the question of various arguments for an equal distribution between women and men in parliament; here gender is more decisive than party affiliation. What is notable in this context is, nevertheless, that gender has an effect that is comparable to the effect of party affiliation. This result supports the theory of the politics of presence. We shall see in the next step whether this assessment up when we study the actions of parliamentarians in the area of gender equality. The results in Table 5 are based upon two open-ended questions asked in the studies of 1985, 1988, and 1994, where the members of parliament were asked to state which issues_/problems they emphasized most when campaigning and which political areas they were personally most interested in. Replies that included 'the woman issue,' 'sex discrimination,' 'affirmative action,' etc. were sorted into the gender equality category.

It is easy to declare that equality between the sexes is not a major issue in the Swedish parliament. It is just as easy to declare the absence of male politicians. In 1985, 14 percent of female politicians stated that they emphasized equality in their campaigns. The corresponding figures were 12 percent in 1988 and 16 percent in 1994. The levels are about the same - around 10 percent - if one looks at the personal areas of interest among female politicians. Among male politicians, however, all the results in Table 5 are perilously close to 0.

The theoretical aspect of the concept of women's interests upon which we have focused here is an acknowledgement of an unequal balance of power between the sexes. The underlying assumption has been that the more favourable a member of parliament is about working towards gender equality and the more actively he or she pursues gender equality issues, the stronger the acknowledgement of inequality between the sexes. If we judge solely upon the behaviour of members of parliament, the result is that the issue is of no overwhelming importance in the parliamentary process: the group that actively pursues gender equality issues is relatively small. None the less, the main result supports the theory of the politics of presence. Although the group that acts in this arena is not very big, it is made up almost exclusively of women.

Social Welfare Policy

We shall now address the component in the concept of women's interests that deals with the occurrence of policies designed to increase the autonomy of female citizens. The study is based upon the premise that it is in women's interests that issues of reproduction and caring services are given room on the political agenda. This does not mean that I intend to declare a certain type of solution as being better or worse for women. The sole premise is that if these issues are absent from the political agenda, then we are back in the political world of silence that Pateman and others have written about (see also Halsaa 1987).

Nor is it my ambition to state any particular level at which one can say that these issues must be given x amount of scope in order for women's interests to be regarded as having been satisfied. As before, the focus is on female and male politicians and any differences there may be between them with respect to issues of representation.

The following analysis is based upon a number of survey questions where there is nothing in the formulations themselves that can be said to stimulate parliamentarians to start thinking in terms of gender. This is an important point. We do not wish to arrive at answers where there is any suspicion that the respondents have adapted themselves to expectations of 'femininity' or 'masculinity'.²⁰ We approach the parliamentary process from three directions: that has to do with the parliamentarians' encounters with voters; one with their personal priorities; and one with their assessments for the future. The choice to include several indicators should be seen here, as before, in the light of the endeavour to achieve stability; if several measurements yield similar results, the conclusions become more credible.

Two of the questions occurred previously: 'Which issue/s or problem/s
did you emphasize most in your campaign work before this year's election?'

and 'Which political issue area/s are you personally most interested in?' The third question - 'Thinking ahead, about the next few years, which issues or problems do you personally think are most important for the parties, Riksdag and government to try to solve?' - is new.

The responses have been coded according to a detailed code scheme. About 30 issue areas were mentioned (more than one issue could be named).²¹ We shall examine here the extent to which social welfare policy issues were found among those mentioned. Included in this category are social policy, family policy, elder care, and health care.²² The percentages in Table 6 show the proportion that mentioned one or more of the social welfare issues in 1985, 1988, and 1994. It makes no difference whether one or several issues were mentioned. The line of demarcation is drawn between having addressed social welfare policy issues and not having done so.

Table 6 shows several important results. The first is that there is a connection between the gender of politicians and the extent to which they pursue social welfare policy issues. In 1985, 75 percent of female members of parliament addressed issues of social policy, family policy, elder care, or health care in their election campaigns. The corresponding figure among male members of parliament was 44 percent. In that same year, 52 percent of female members of parliament stated welfare policy as an area of personalinterest, the proportion of male members of parliament who did so was 11 percent. With respect to their assessments of important future

issues, the figure was 52 percent among the women and 31 percent among
the men.

The second important result is that, although gender differences were found on the two subsequent survey occasions, the gap has closed somewhat over time. If we look at the election campaign, the gender difference was 31 percent in 1985, 24 percent in 1988, and 16 percent in 1994. In the area of personal interest, the gender difference was 42 percent in 1985, 34 percent in 1988, and 19 percent in 1994. With respect to important future issues, the gender difference was 21 percent in 1985, 11 percent in 1988, and 8 percent in 1994.

The third significant result in Table 6 has to do with comparison of the various arenas. Gender differences were consistently greater in the politicians' agendas - those that show up when they state their own areas of interest - than in the agendas that appear in the election campaign or the statement of important future issues. In the theory of the politics of presence, the arenas with the least measure of outside control are the most pivotal. Based upon this reasoning, one can assume that the 'personal interest' indicator is the measure that best corresponds to the assumptions put forth by Anne Phillips. The election campaign, for example, is an arena that is centrally controlled to a great extent by the parties.²³

Table 7 shows the extent to which the differences that appeared remain once we take additional factors into account. Included in the analysis are, as previously, party affiliation, age, education, and parliamentary experience. The scope of the analysis is now limited to the politicians' personal agendas.

The main result is that female politicians have consistently been the group that have pursued social welfare policy issues to the greatest extent in their parliamentary work. There are only two results in Table 7 that indicate reverse gender differences (more men than women who mention social welfare The difference in the Left Party was —1 percent in 1988 and the difference was -18 percent in the Green Party in 1994. Upon one occasion there is a zero result (an equal number of men and women mention welfare policy): the Liberal Party in 1994. In all other cases, the results support the theory of the politics of presence.

The question then is how conspicuous the differences are. In 1985, gender differences were consistently 30 percent or more, which one can certainly call a clear result. In 1988, gender differences were 20 percent or more as a rule, also a distinct difference. But the basis was weaker in 1994 - the differences modest for several groups. In conclusion, the results prove two things: generally speaking, there are still more women than men in parliament who have social welfare policy as an issue on their personal agenda, but the pattern has become weaker with time. The significance of gender was not as palpable in 1994 as it was in 1985.

Concluding Discussion

If we accept the definition upon which the analyses presented here are based, it is difficult to repudiate the conclusion that women's interests are primarily represented by female politicians. The theory of the politics of presence gains strong empirical support. It is always possible to discuss which indicators are most appropriate, but I believe that the measurements available here are reasonable from the validity standpoint and that they fulfil the goal of stability in several ways. Beyond that already mentioned, it is worth pointing out that the measurements address both opinions and behaviour among politicians.

One issue of primary importance for further discussion is the possible consequences of these gender differences. The results challenge an unreflective naive 'gender neutrality,' which I believe characterizes much of the research on representative democracy. The responsible party model is the most influential model today.²⁴ The party affiliation of politicians is at the hub, while it can be said that factors such as gender are still situated on the outskirts of political science.

Another issue, which also involves consequences, is how we, in a deeper sense, should interpret the patterns found in this study of Swedish members of parliament. When certain issues are more important on the agendas of female politicians than on those of male politicians, does it mean that the more female representatives are elected, the more woman-friendly politics will become? I shall begin by saying that the question posed is difficult to answer. However, it is possible to submit some views based upon the results. In all years during the period in focus (1985-1994) the proportion of women in the Swedish parliament was greater than 30 percent. (In 1985, there were 32 percent women in the Riksdag, with 38 percent in 1988 and 41 percent in 1994.²⁵) In other words, gender-related patterns exist at the same time that female representation numerically is strong.²⁶

I should like to denote the two perspectives that are relevant here as the static perspective and the dynamic perspective on the more qualitative, or content-based, aspects of female representation.²⁷ From a static perspective, interpretation of the pattern that emerged is that it involves a division of labour between female and male politicians. When women enter the political arena, they take over certain areas from men, but nothing becomes fundamentally different as a result. From a more dynamic perspective, interpretation is rather that the 'feminine' side is given greater scope because there are more female politicians.

The problem with giving any unambiguous answer to the question of which perspective is 'truest' has to do with the prerequisites for research within all social sciences. It is impossible to tinker with the Swedish Riksdag and find out how it would be if we had one parliament made up solely of

men and another consisting only of women. Nor can we shoot 'take 2' of Swedish politics and discover what the situation would have been today had female representation levelled off at the 10-15 percent of the parliament of 20-30 years ago.

However, I find the dynamic perspective to be the most credible. One of the things that supports such a point of view is the clear finding that few male politicians award priority to women's interests in their parliamentary work. There is no significant reason to believe that the male politicians of today differ in that respect from male politicians in previous decades. Thus, when women take over seats from men, it should mean that we gain a greater number of politicians who prioritize issues such as gender equality and social welfare.

Another result that underpins a perspective for change comes from the voter surveys carried out by the Swedish election study programme. Since the 19605, these studies have asked voters to rate their subjective level of interest in-politics. The results show that 32 percent of female voters in 1960 said that they were very or fairly interested in politics. By 1994, the corresponding had risen to 52 percent, an increase of 20 percent. During the same 30-year period, there was only a marginal increase among male voters, from 57 to 63 percent (Oskarson & Wångnerud 1995, 40). In other words, there is still a higher proportion of men than women who claim an interest in politics, but the difference has declined markedly over the years.

A third result that supports the dynamic perspective comes from the study of members of parliament carried out in 1994. They were asked the following question: 'In the past 20 years, the proportion of women has increased in most parties in the Riksdag. Are there concrete issues on which you believe the position of your party has changed owing to the increased representation of women?'²⁸ The respondents were asked to indicate yes or no and the results show that 74 percent of the women and 50 percent of the men said yes.

Thus, a majority of members of parliament believe that a change has taken place. We also asked which issues the politicians perceived as having changed. Three areas emerge most strongly in the responses: gender equality was mentioned by 40 percent, family policy by 34 percent, and social policy by 33 percent of the parliamentarians (the question was open ended and the respondents were allowed to mention more than one area). This means that the politicians' own picture agrees well with the other indicators presented in this article.²⁹

To sum up the discussion, I should like to say that women's increased participation has meant that a shift of emphasis has occurred in the political agenda: women's interests have been given greater scope and become more central. This result is supported by studies carried out by Hege Skjeie (1992) in Norway. However, to gain a solid basis for such a conclusion, we would really need to carry out further studies in which countries like Sweden and

Norway are contrasted with other countries where female representation is
much lower.

Finally, I would like to refer back to the static perspective on female representation. There is, naturally, a risk that the patterns described here conserve rather than change the prevailing order of gender and power. However, it is likely that what we have at hand is a development that can be described in terms of different phases. The fact that women have historically contributed to putting greater emphasis on the social welfare aspects of politics does not necessarily mean that they should confine themselves to this area for all time. In the longer term, it would be rather peculiar if women and men in the future did not exercise power and influence in all political domains to roughly the same extent. However, I believe it is almost impossible to go from a low proportion of female parliamentarians to a high proportion without going through a stage wherein patterns of 'masculine' and 'feminine' appear in the content of politics.

---

NOTES

1. The debate has been going on at least since the early 19705, when Jessie Bernard (1971) published Women and the Public Interest. An Essay on Policy and Protest. During the 1980s, Irene Diamond and Nancy Hartsock argued against the use of interest theories for analysis of women's conditions; they wished instead to refer to women's particular needs (Diamond & Hartsock 1981). There has been continuing debate in Scandinavian research, primarily between Jonasdottir (1985, 1991) and Halsaa (1987) on aspects of the content and form of women's interests. Jonasdottir has stressed participation (form) itself as the most pivotal interest, while Halsaa has more strongly emphasized the content of politics. A closely related theme in the debate involves the extent to which interests are basedon biology or socialization. Modem research in political science is characterized by a pragmatic approach in which we do not entirely exclude biological grounds but put greatest emphasis on the significance of social/ cultural norms for the creation of separate gender categories. Phillips (1998) gives a good introduction to the debate on women's interests. See also Young (1990).

2. Anne Phillips herself can be said to work with a minimalist approach. She writes: '[T]he argument from interest does not depend on establishing a unified interest of all women: it depends, rather, on establishing a difference between the interest of women and men' (Phillips 1995, 68).

3. The quote from Skjeie (1992) has been translated from Norwegian.

4. I believe that the use of minimal definitions has partly come about through a wish to distance ourselves from a biological perspective on gender dimensions. However, my approach is that as soon as we problematize the significance of gender, we have already shown that factors other than the purely biological are meaningful.

5. In her book The Politics of Presence (1995), Anne Phillips also discusses the interest concept in relationship to class and ethnicity. Norris & Lovenduski (1995) is an example empirical analysis of the significance of gender, class, and ethnicity in political decision making.

6. In Nordic research, the anthology Det uferdige demokratiet. Kvinner i nordisk politikk by Haavio-Manilla et al. (1983) is an early work that problematized the man as the social norm. In Sweden, Eduards (1977. 1988). Jonasdottir (1985), and Peterson (1984) are examples of significant early works.

---

7. The implementation of political decisions that reflect women's interests may be seen as a fourth step.

8. However, there are few recent books on Swedish politics in which political scientists explicitly repudiate the significance of gender in the parliamentary process. See Larsson (1993, 121).

9. However, see also chapter 4 in Esaiasson and Holmberg (1996) for a modifying point of view.

10. For further empirical research on the significance of gender in the parliamentary process, for example Darcy et al. (1994), Lovenduski & Norris (1996), Norris et al. (1992), Purdy (1991), Sclozmanet al. (1995). Vega & Firestone (1995).

11. The Swedish election study programme was built up during the 1950s under the leadership of Professor Bo Sårlvik and Professor Jorgen Westerståhl. Professor Soren Holmberg currently heads the programme. Within the election study programme, studies of both voters and parliamentarians are carried out.

12. Where possible, I include all years (1985, 1988, and 1994). However, certain questions were asked upon only one or two occasions. The response rate has been above 95 percent each of the parliamentary surveys, which means that the data can be håndled as a complete sample.

13. A detailed analysis of the 1988 study shows that the party is the primary representative duty for both female and male parliamentarians. Among female members of parliament, women's interests views are given second place in a ranking of duties, assessed as being very important, followed by the duty to represent the region_{/constituency} individual voters. Among male members of parliament, individual voters and the region/constituency are awarded second and third place, respectively, in a corresponding (Oskarson & Wångnerud 1995, 129).

14. See Esaiasson (2000) for similar results among parliamentarians in all Nordic countries.

15. Sue Thomas (1994, 69) has carried out studies on the state level in the US. In her study. there was a similar question that showed that 57 percent of female politicians and 33 pareen i of m;ile poln icians considered represent ing women to be a very import an t duty. Thus, the proportion of women is about equal in Sweden and the US, while the level for male politicians is significantly higher in the US. The question is how much this may be attributed to substantial differences, or whether it reflects the circumstance that Thomas' study was conducted in a context wherein the gender perspective was significantly prominent than in the Swedish studies (and that male politicians in the US have to a greater extent adapted their responses to what is seen as politically correct).

16. Analyses of the significance of gender runs a risk of falling flat if we do not include other background factors as well. Party affiliation, age, education, and parliamentary experience (the number of years in parliament) are some of the most common factors in representation studies. Conducting controls for factors in addition to gender can be seen as one way of further increasing the stability of conclusions.

17- For uonien's and incif> attitude (both ol \oUTs and ai |>nrli..mienlar\ Ic\ cl) lowanK a number of different future societics, see Oskarson & Wangncrud (1995; 1996).

18. Since the number of women in the Riksdag is lower than the number of men, an equal distribution is equivalent to an inercase in women's representation. An assessment of the four different arguments is shown in the table below. I have included here the proportion who responded "very important' among both female and male parliamentarians. The question was inspired partially by Hernes (1987. 23), who makes a distinction between justice, resource, and interest arguments with respeel lo female representation, and partially by Thomas (1994), who differentiates between women's infiuence upon legislative procedures and legislative products in representative contexts. See also Phillips (1995) and Ok in (1995) for various arguments in the debate on increased female representation. Attitudes towards Four Different Arguments for an Equal Distribution of Women and Men in the Riksdag Very important argument (percent) Women Men Difference Parliament should reflect the most important groups in society in its composition 73 40 +33 Women have different life experiences 74 26 +48 Working procedures and the climate of discussion willchange 60 16 +44 There will be consequences for policies 50 8 +42 Comment: The least number interviewed was 135 women and 192 men. Source: 1994 Parliament Study.

---

18. Since the number of women in the Riksdag is lower than the number of men, an equal distribution is equivalent to an inercase in women's representation. An assessment of the four different arguments is shown in the table below. I have included here the proportion who responded "very important' among both female and male parliamentarians. The question was inspired partially by Hernes (1987. 23), who makes a distinction between justice, resource, and interest arguments with respeel lo female representation, and partially by Thomas (1994), who differentiates between women's infiuence upon legislative procedures and legislative products in representative contexts. See also Phillips (1995) and Ok in (1995) for various arguments in the debate on increased female representation. Attitudes towards Four Different Arguments for an Equal Distribution of Women and Men in the Riksdag Very important argument (percent) Women Men Difference Parliament should reflect the most important groups in society in its composition 73 40 +33 Women have different life experiences 74 26 +48 Working procedures and the climate of discussion willchange 60 16 +44 There will be consequences for policies 50 8 +42 Comment: The least number interviewed was 135 women and 192 men. Source: 1994 Parliament Study.

19. Cronbach alpha is 0.78. whicli indicaies good reliability for the index.

20. This design differs from e.g. Hege Skjeie's study. in which she asks Norwegian parliamentarians they consider to be women's versus men's interests in politics (Skjeie 1992, 108). Skjeie was studying perceptions of femininity and masculinity among politicians.

21. Oskarson & Wångnerud (1995; 1996) provide a broad examination of the impact of gender on different issue areas in Swedish politics at both electoral and parliamentary levels. The purpose of this article, however, is not to ask in which. or in how many, areas there are gender differences. The selection reflects issue areas of specific concern for women's autonomy in contemporary society.

22. What should be included under the category of social welfare policy is far from selfevident. wider definition that includes education and the environment is certainly conceivable. My final choice was. however to interpret issues of reproduction and caring services in narrower terms. The subcategory of social policy includes. in addition to the general formulation of social policy. responses such as the public sector. the welfare etc. The subcategory of family policy includes all responses that concern family issues - including childcare allowance, parental leave insurance. and daycare centres. The subcategory elder care involves all types of support for the aged. The subcategory health care includes. in addition to the general formulation of health care. responses such as health insurance, waiting periods, etc. For an exact code scheme. see the documentation of the studies in Esaiasson et al. (1995).

23. Note. however. that this point is not very important. since similar results appear for all three indicators.

24. For a description of the responsible party model and its strong status, see e.g. Esaiasson & Holmberg (1996) as well as Klingemann et al. (1994).

25. Since the last general election (1998). 43 percent of members of the Swedish parliament have been women. Half of the cabinet ministers in the government have been »omen since the 1994 election.

26. In comparison to the situation worldwide the number o( women in the Swedish p;i ilu niL'nl !:¦ liiL'h ITu'iv \v;i> ;i n inU'i na_[ ional average of 13 percent women in national parliaments in 1998. For statistics on female representation in various countries. see the homepage of the Inter-Parliamentary linion at: http: www.ipu.org .

27. Soine describe this line of demarcation in terms of an 'optimislic" versus a "pessimistic" perspective. See e.g. Kanonen & Selle (1995) and Siim (1990. 15).

28. The question is modelled on Skjeie (1992).

29. These quesiions uere also asked in p,n hameiiiarKins m the other Nordie countries. The results agree to a great exlent with the Swedish results (Wångnerud 2000).

REFERENCES

Bernard, J. 1971. Women and the Public buerest. An Essay on Policy and Protest. Chicago:

Aldine/Athertone.

Dahl, R. 1989. Democracy and its Critics. New Haven: Yale University Press.

Darcy, R., Welch, S. & Clark, J. 1994. Women, Elections and Representation. Lincoln and
London: University of Nebraska Press.

Diamond, I. & Hartsock, N. 1981. 'Beyond Interests in Politics: a Comment on Virginia
Sapiro's "When Are Interests Interesting? The Problem of Political Representation of
Women",' American Political Science Review 3, 717-23.

Eduards, M. 1977. Kvinnor och politik. Fakta och forklaringar. Stockholm: Liber.
Eduards, M. 1988. 'Att studera politik ur ett konsperspektiv,' Statsvetenskaplig tidskrift 3,
207-21.

Esaiasson, P. 2000. 'Focus of Representation,' in Esaiasson, P. & Heidar, K., eds., From
Westminister to Congress and Beyond. The Nordic Experience. Columbus: Ohio State
University Press.

Esaiasson, P. & Holmberg, S. 1996. Representation from Above. Members of Parliament and
Representative Democracy in Sweden. Aldershot: Dartmouth.

Esaiasson, P., Holmberg, S. & Brothén, M. 1995. Riksdagsenkdt 1994. Dokumentation. Statsvetcnsks.pl
3. institutionen Gdteborss universitet.

Haavio-Mannila, E. et. al. 1983. Det uferdige demokratiet. Kvinner i nordisk politikk. Oslo:
Nordisk Ministerråd.

Halsaa, B. 1987. 'Har kvinnor gemensamma intressen?' Kvinnovetenskaplig tidskrift 4,42-56.
Hernes, H. 1987. Welfare State and Woman Power: Essays in State Feminism. Oslo: Oslo
Norwegian Press.

Jonasdottir, A. 1985. 'Kvinnors intressen och andra varden,' Kvinnovetenskaplig tidskrift 2,
17-33.

Jonasdottir, A. 1991. Love Power and Political Interests. Towards a Theory of Patriarchy in
Contemporary Western Socities. Orebro: Orebro Studies, No. 7.

Karvonen, L. & Selle P., eds. 1995. Women in Nordic Politics. Closing the Gap. Aldershot:
Dartmouth.

Klingemann, H.-D., Hofferbert, R. & Budge, I. 1994. Parties, Policies and Democracy.
Boulder: Westview Press.

Larsson, T. 1993. Det svenska statsskicket. Lund: Studentlitteratur.

Lovenduski, J. & Norris, P., eds. 1996. Women in Politics. Oxford: Oxford University Press.
McAllister, I. & Studlar, D. 1992. 'Gender and Representation among Legislative Candidates

in Australia,' Comparative Political Studies 3, 388-411.

Norris, P. & Lovenduski, J. 1995. Political Recruitment. Gender, Race and Class in the British
Parliament. Cambridge: Cambridge University Press.

Norris, P., Vallance, E. & Lovenduski, J. 1992. 'Do Candidates Make a Difference? Gender,
Race, Ideology and Incumbency,' Parliamentary Affairs 4, 496-517.

Okin, S. 1995. 'Politics and the Complex Inequalities of Gender,' in Miller, D. & Walzer,
M., eds., Pluralism, Justice and Equality. Oxford: Oxford University Press.

Oskarson, M. & Wångnerud, L. 1995. Kvinnor som vdljare och valda. Om betydelsen av kon i
svensk politik. Lund: Studentlitteratur.

Oskarson, M. & Wångnerud, L. 1996. 'Vem representerar kvinnorna?' in Rothstein, B. &
Sårlvik, 8., eds., Vetenskapen om politik. Festskrift till professor emeritus Jorgen Westerståhl.
institutionen, Goteborgs universitet.

Pateman, C. 1988. The Sexual Contract. Cambridge: Polity Press.

Pateman, C. 1989. The Disdorder of Women. Democracy, Feminism and Political Theory.
Cambridge: Polity Press.

Pateman, C 1992. 'Equality, Difference, Subordination: the Politics of Motherhood and
Women's Citizenship,' in Bock, G. & James, S., eds.. Beyond Equality and Difference.
London: Routledge.

Peterson, A. 1984. 'The Gender-Sex Dimension in Swedish Politics,' Acta Sociologica 1, 3-17.
Phillips, A. 1991. Engendering Democracy. Cambridge: Polity Press.

Phillips, A. 1995. The Politics of Presence. Oxford: Oxford University Press.
Phillips, A., ed. 1998. Feminism and Politics. Oxford: Oxford University Press.

Pitkin, H. 1967. The Concept of Representation. Berkeley: University of California Press.

Purdy, E. 1991. The Representation of Women and Women's Issues: Differences in Voting Patterns
of Male and Female Members of the House of Representatives. Ann Arbor: UMI.

Sclozman, K., Burns, N., Verba, S. & Donahue, J. 1995. 'Gender and Citizen Participation:
Is There a Different Voice?,' American Journal of Political Science 2, 267-93.

Siim, B. 1990. 'Feministiska tolkningar av samspelet mellan kvinnor och vålfårdsstaten,'
Kvinnovetenskaplig tidskrift 2, 13-25.

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

Thomas, S. 1994. How Women Legislate. Oxford: Oxford University Press.

Vega, A. & Firestone, J. 1995. 'The Effects of Gender on Congressional Behavior and the
Substantive Representation of Women,' Legislative Studies Quarterly 2, 213-22.

Wångnerud, L. 1998. Politikens andra sida. Om kvinnorepresentation i Sveriges riksdag.
Goteborg: Goteborg Studies in Politics, No. 53.

Wångnerud, L. 2000. 'Representing Women,' in Esaiasson, P. & Heidar, K., eds.. From
Westminister to Congress and Beyond. The Nordic Experience. Columbus: Ohio State
University Press.

Young, I. 1990. Justice and the Politics of Difference. Princeton: Princeton University Press.