University of Helsinki

Introduction

"The finding that men are more likely to participate in politics than women is
one of the most thoroughly substantiated in social science. ... Economic and
social modernization is slowly eroding this sex difference, however", writes
Lester Milbrath (1965, 135—136). According to Maurice Duverger, the last generalization
applies mainly to voting participation, in certain countries, not to
women's participation in political leadership: "The percentage of women members
of parliament, for instance, is hardly increasing. On the contrary, it tends to
fall after the first elections in which women have had the suffrage, and to become
stabilized at a very low level" (Duverger 1955, 123).

Seymour M. Lipset et al. have classified explanatory factors related to original statistical regularities in rates of voting turnout (1954, 1128—1135). According to this classification, men are an example of groups which tend to have a high voting turnout for several reasons, women of groups with a low voting turnout. Three out of four of their main propositions can be tied to sex roles. The second proposition concerns access to information about the relevance of government policies. This includes among others occupational training and experience making for general insight. "The housewife is at a great disadvantage in this respect, a fact that may help to account for the lower voting rate of women in general", Lipset et al. write (p. 1131). The amount of leisure is also related to political activity. "The group most affected by lack of leisure is probably the woman. The role of housewife and mother is particularly demanding. 'Man works from sun to sun, but woman's work is never done'. In every country women have a lower turnout than men. This sex differential increases as one goes down the income scale. This suggests its relation to the number of children and the absence of laborsaving devices or servants" (p. 1132).

The third proposition by Lipset et al. is related to group pressures to vote: Group pressures to vote are stronger in the case of men than of women. "There are some cases in which group pressures are directed in exactly the opposite way — against voting. The universally lower vote of women may be due in part to norms of a 'woman's place', which disapprove of political participation of women. In the United States these do not seem to prevent voting among

women who are interested; however, women with little interest feel free not to
vote, while men with little interest still feel called on to go to the polls" (p. 1133).

The fourth proposition refers to cross-pressures which have been found in Finnish studies also to affect women's voting more than that of men (Allardt 1956, 125—130). Lipset et al. state (p. 1134) that "in European countries particularly, the working-class women are more concerned with religious values and more exposed to church influences than are their husbands, who tend to belong to leftist unions. The conflict between the pressure of their class position and of their husbands, in favor of leftist voting, and the opposition of the churches to the antireligious left parties may lead to a withdrawal from political choice and contribute to the lower turnout of women, which is particularly marked in the working class".

Apart from studying the influence of these kinds of explanatory factors, sex roles in politics, their change with time and their variation in different circumstances, can also be understood and explained in terms of ideologies concerning the roles of men and women.

According to the traditional and romantic sex role ideologies, politics belong to the man's role (Dahlstrom 1963, 18—19). These ideologies are deeply rooted in Scandinavia. Even today in Finland only about a third of the people in both urban and rural areas consider it important for a woman to be interested in politics, whereas two thirds of urban and more than 50 °/o of rural people consider it important for a man. The Marxian sex role ideology is based on the principle of total equality of men and women in society, including the spheres of work life and politics. In practice women in the socialist countries have attained a fairly large representation on the lower political levels but in the important positions men heavily dominate. The early women's emancipation movement in the late 1800's and early 1900's — which is still represented by many women's organizations, nowadays dominated by elderly women — has two successors. The radical sex role ideology in Scandinavia is, like the Marxian ideology, based on the principle of total equality of the sexes. It proceeds further than the Marxians and the early suffragettes in trying to bring about changes in the man's role, too. It encourages men to accept some traditionally feminine responsibilities like child care and housework. This would make it easier for women to be active in work life and politics. There are also men actively engaged in work for this movement. The goal is total integration of sexes. The woman power or liberation movement in the United States is based on the notion of the conflict of interests of men and women. It compares the fight for women's power with the class-struggle and race-struggle. It uses methods of direct: action to achieve its goal, woman power. It assumes that integration is not possible but revolt and fight are necessary to bring women into power. In some ways it resembles the fight of the early suffragettes, but it is also tied in with the Marxian ideology.

We may assume that sex roles in politics are more traditional in those times
and groups in which traditional ideologies concerning the roles of men and

women prevail. Groups influenced by the Marxian or modern radical sex role
ideologies should show less traditionalism in the political roles of the sexes.

In this article sex roles in Finnish politics will be examined on the basis of
hypotheses based on the generalizations by Lipset and on assumptions about the
possible influence of different ideologies.

The study is composed of two main parts, in which both behavior and attitudes are examined. In the first part political participation of men and women as voters, political discussers and party workers, their political interest and expectations concerning it are studied. The second part describes men and women as political representatives: how candidates are nominated in the elections, how many votes they get and how many of them get into the decision-making representative bodies. In this section attitudes toward equality of the sexes in political leadership are also studied. In the final chapter the results of both parts are combined.

The study is based on official statistics and interview data. The official statistics of Finnish parliamentary and municipal elections in 1907—68 are fairly extensively used. Sex differences in voting activity are also analyzed in relation to other characteristics of Finnish society. This was possible on the basis of data collected by Erik Allardt, who kindly placed the original partly unpublished results of his correlation and factor analyses of Finnish communes around 1950 at my disposal. Some comparisons to other Scandinavian countries are made, but they are quite fragmentary. The interviews referred to were made in 1966 in Helsinki, two small towns and five rural communes. A representative sample of men and women aged 15—64 were interviewed in Helsinki and the rural communes. In the small towns only women were questioned. The communities studied do not represent Finnish communes in general. It is, however, possible to draw comparisons between western and eastern Finland.

I. Political participation

Scandinavian Comparisons

Full enfranchisement of all adult men and women took place in Finland in 1906. Finnish women were the first in Europe to get the vote in parliamentary elections. Of the Scandinavian countries, Norway came next with a partial right to vote in 1907; in 1913 women were enfranchised on the same conditions as the men. In Denmark and Iceland political franchise was granted to women in 1915. Universal suffrage for men and women was introduced in Sweden at elections to the lower chamber of the parliament in 1921.

The difference between men's and women's voting turnout at the beginning
of this century was lower in Finland than in the rest of Scandinavia as shown
by the following figures.

Voting rates of men and women in parliamentary and municipal elections in Norway and Finland during this century are shown in Figure 1. The rate of voting during the first half of the century was lower in Finland than in Norway. However, sex differences in the voting rates up to the 1940's were smaller

in Finland. It has generally been found that the difference between men and
women non-voters seems to be smaller when the total vote is higher (Duverger
1955, 24). The cases of Finland and Norway are exceptions to this rule. The sex

difference in Finland compared to that in Norway is especially noticeable in rural communes: in Finland the women's voting rate in 1908—50 was only about 10 %, in Norway about 20 % lower than the men's. The small difference is mainly due to the men's laziness in voting. The low rate of voting among Finnish men may be partly explained by deficiencies in the registration of eligible voters; in some parts of Finland (see p. 218) persons who had emigrated were kept in the voting register for a long time, and since the majority of emigrants in the beginning of the century were men, this in a way artificially lowered the voting rate of men.

In the 1950's and the 1960's the national differences have disappeared and
the voting percentages of both sexes in all Scandinavian countries are high and
very close to each other:

Voting in parliamentary elections is in general more common than in municipal
elections, and the sex difference is smaller.

Rural — Urban Differences in Voting Rates

Sex differences in political participation are diminishing with time. New ideas and behavior patterns are usually adopted first in the center of the society and later transmitted to the periphery (Galtung 1964, pp. 206—216). Urban areas belong to the center of society, rural areas to the periphery. We may thus expect that voting activity will be higher and sex differences in political behavior and attitudes smaller in towns and cities than in rural areas. This assumption is also based on Lipset's et al. formulations one and two:

"A group will have a higher rate of voting if its interests are more strongly
affected by government policies", and

"A group will have a higher rate of voting if it has more access to information
about the relevance of government policies to its interests".

Voting turnout of men and women in Finland in parliamentary and municipal
elections in urban and rural communes before and after the second world
war were the following:

During the latter time period, rural-urban residence had no influence on the voting rates of men, but according to the hypothesis, rural women voted less than urban women and thus the sex difference was smaller in towns than in the country. In the earlier time period voting rates in parliamentary elections were about the same in urban and rural areas — contrary to the hypothesis, the voting rate of urban men was even a little lower than that of rural men. Harriet Holter reports a similar finding from Norway (1970, 101—102). The small sex difference in urban communes is thus due to the low voting rate of men. In municipal elections, place of residence had earlier a great influence on voting activity. Townspeople seemed to be more concerned about local community matters than country people; they went more often to the polls in municipal elections.

Sex differences are more significant in the countryside than in the towns also
as regards other types of political behavior. In personal interviews about interest
in politics the following results were achieved:

Political interest of women both in Helsinki and in the rural communes is
weaker than that of men. Women in Helsinki are as interested as rural men.

Urban people are politically more active than rural people, and this applies especially to women. This may be due to their central position in society, the greater relevance of government politics to their lives and their easier access to political information, as was earlier assumed. Lipset's et al. third condition for a high voting turnout might be appropriate here, too:

"A group will have a higher rate of voting if it is exposed to social pressures
demanding voting."

On the basis of the interview data it is possible to study pressures toward political activity in town and country. Expectations concerning men's and women's political interest were, in the interview study, measured by asking how important the respondents considered certain activities to be for men and women. The items were: temperance, faithfulness in marriage, interest in culture, going to church, interest in politics, membership in associations, speaking at meetings, holding elected offices, getting a good education, and getting ahead in one's career. The first four were considered to be more important for women than for men, the last six more important for men than for women. The item concerning interest in politics differentiated the sexes more than any of the others. The following percentages give the distributions of answers.

An interest in politics is more often expected from men than from women. This is agreed upon by men and women in urban and rural communes. For a man to be interested in politics is considered more important in the city and towns than in the country. Role expectations concerning women are quite unanimous in different groups of respondents. When we compare actual and expected interest in politics, we find that the behavior of women in Helsinki has changed more rapidly than role expectations.

Coming back to the proposition of Lipset et al., men are clearly exposed to more social pressures demanding political activity than women. Expectations concerning men's political interest are more demanding in urban than in rural districts. This may partly explain the higher political interest of urban men, which is reflected in their higher voting activity (except in the general elections

in 1907—39). In the case of women this explanation cannot be used: urban norms do not demand women to be politically more interested than rural norms. A change in the reference group of urban women might serve as an explanation. Urban women may be concerned with what is expected from people in general, rather than merely from women, when they act politically. Their reference group may be the whole society or the group of men, not that of women. Here I am grateful for a discussion with Rita Liljestrom. Urban women are willing to interact with men more than rural women (Haavio-Mannila 1968, 235). They want to identify themselves with the men's, not with the women's group.

We may conclude by saying that modernization and urbanization increase as expected the political activity of both sexes and decrease the difference between men and women. This may be due to the greater involvement of government politics in the interests of modern urban people, and to their easier access to information about government policies relevant to their interests. Urban men were found to be more exposed to social pressures demanding voting than rural men, but no difference in the expectations concerning women's political interest could be seen between urban and rural areas. The urban norms do not demand women to be more active than rural norms, but urban women are so anyway. Other factors than social pressures demanding political interest from women may account for their higher voting turnout.

This conclusion as to the importance of urbanization in leveling off the sex differences in political behavior has support from, among others, Harriet Holter's study in Norway. She found that "the degree of industrialization of residence is the attribute that most discriminates between men's and women's rates of political activity" (1970, 103). She studied simultaneously the effect of occupation, education and place of residence on the political activity of men and women.

Regional Differences in Voting Rate

The social system in different parts of Finland has, due to geographical factors, developed differently. The agrarian populations of western and southern Finland are inhabitants of compact villages, whereas the rural population of eastern Finland lives in a scattered landscape of many lakes and forests. In northern Finland the population density is very low but people are concentrated into villages somewhat more than in eastern Finland. "In historical studies it has been shown that political traditions seem to affect the political climates of the two halves of Finland. The northern and eastern parts have always had more radicalism, even within the agrarian and bourgeois blocks. In the south and west the social system of earlier times was heavily village-centered, whereas in the north and east, with its extremely low density of population, it was very much based on the extended family. There are some studies indicating that the villagecenteredness in the south and west created extremely strong village cohesiveness as late as the latter part of the nineteenth century. This was particularly true in South Pohjanmaa where the political climate is very conservative. On theoreti

cal grounds one can assume that strong village cohesiveness results in support
for the status quo, but no studies of the impact of these differences in the social
structure have been carried out" (Allardt and Pesonen 1967, 350).

In the following we shall study the voting rates of men and women in different
regions of Finland during two periods of time, before and after the second
world war on the basis of map 1 and Haavio-Mannila 1970, appendix 1.

Voting rates of both men and women in 1908—39 were high in the southwestern parts of Finland. The only exception was the rural communes of the province of Vaasa in western Finland, where men voted very little. The difference between the voting rates of men and women was only about 1 °/o. A closer analysis of the voting rates year by year reveals that this exceptional situation prevailed from 1908 to the beginning of the 1930'5, when the rate of voting of men in Pohjanmaa (province of Vaasa) began to rise and soon reached a level above the national average. The exceptionally low voting rate of men in 1908 —30 was due to the great migration from this area, which reached its peak in 1901—10 but continued to be common until 1930, when it stopped.

The emigrants were not always removed from the registers of eligible voters. When we subtract the absent potential voters, from the number of persons registered eligible to vote, and then compute the rate of voting in the rural communes in the province of Vaasa, we get the following results for the elections in 1910 and 1919:

The actual rural population in the province of Vaasa was voting even more actively than people in other parts of Finland. The difference between the sexes was still smaller than in the whole country, but not quite as small as the official voting rates indicate. A new statute about the registration of migrants was passed in 1930. This explains the sudden change in the official voting rate figures. Simultaneously emigration ceased so that possible omissions in registration were no longer caused by it. The same correction in the registration figures could be made for the provinces of Turku and Pori, Uusimaa and Hame, where migration was also fairly large. It did not at all, however, reach such large proportions as in the province of Vaasa.

The voting turnout of both men and women during the period 1908—39 in
eastern Finland was low, both in urban and rural communes. There the difference
between men's and women's voting turnout was great.

In northern Finland the voting turnout during the period preceding the Second World War was much below the national average. The sex difference in voting participation was high. Soon after the war, however, the voting rates increased and the sex difference diminished especially in the northern part, in the province of Lapland. Before the war the percentage poll was above the national

average only in the elections of 1929. That was the last year before the war when the Communist party was officially allowed to participate in elections. In northern and eastern Finland, Allardt has found an ecological association between high voting participation and "Backwoods Communism" (1964, 118). A rise in voting activity and a decrease in the sex difference in northern Finland may be attributable to increased support for the communists.

During the second period, in 1945—66, the voting turnout has continued to
be high in the south-western parts of the country. The eastern provinces still lag
considerably behind: voting participation is low and sex differences high.

To explain the regional differences in voting turnout we may apply the
fourth proposition by Lipset et aL:

"A group will have a higher rate of voting of the pressures to vote are not
directed in different political directions so as to create conflict over which way
to vote."

Possibilities of cross-pressure vary in different areas of Finland, as has been pointed out in many earlier studies (Allardt 1956, Riihinen 1965, Martikainen and Sankiaho 1969, for example). To cite Martikainen and Sankiaho (p. 22): "In the scatteredly and sparsely populated eastern Finland social groups were open and heterogeneous. High mobility and migration between and inside groups created preconditions to cross-pressures, which have been reflected in the tradition of political participation." Allardt has shown how a fairly even competition between the bourgeois and the socialist parties diminishes the rate of voting (1956, 52—91). Competition between left and right is characteristic of eastern Finland, notably of the provinces of Mikkeli and Kuopio, where the voting rates are lowest.

In western Finland the social structure was uniform and the system was characterized by demands for strong conformity in behavior (Riihinen 1965, 60—68). Under these circumstances high political activity had become a tradition. Social and political structures were homogeneous and without cross-pressures.

Ecological Analysis of Variables Explaining Sex Differences in Voting Turnout Men's and women's voting turnout in relation to each other in an area can be compared with the social characteristics of the same area. I had an opportunity to reanalyze the results of the correlation and factor analyses conducted by Erik Allardt. They were based on information about various social characteristics of Finnish communes around 1950. In 1954 there were 550 communes in Finland. The communes form the basic units of administration and of official statistical analysis.

Allardt has divided the communes into groups in two ways. One analysis was performed in five groups: 1. Towns and cities; 2. Rural communes with a Swedish-speaking majority; 3. Rural communes in southern and western Finland; 4. Rural communes in eastern Finland; and 5. Rural Communes in northern Finland (Allardt 1964, 119—120). The other analysis was made in two groups:

all communes, both urban and rural, 1. in the south and west and 2. in the north and east. The Åland Islands, which have a large degree of self-government and elect only one representative to Parliament, have been omitted from the analysis (Allardt and Pesonen 1967, 350).

The variable measuring the sex difference in voting turnout was calculated so that a high value means a large difference in the voting turnout. Thus it indicates that the men's rate of voting is high compared to that of women. We can call it either "men's voting turnout compared to women's" or "low women's voting participation", or simply "sex difference in voting turnout".

Sex difference in voting turnout is correlated with different variables in different regions. Thus it is most informative to examine the groups of communes one by one by presenting the correlations which are statistically significant at the .05 level at least.

In towns and cities the sex difference in voting turnout is correlated with the
following variables:

The variables can be divided into three groups: those related to 1. religious
traditionalism, 2. low economic standard, and 3. migration.

According to Allardt's factor analysis, women's low voting turnout compared to men's had high factor loadings on the factor of Socialist Traditions (+.55) and on the factor of Religious Traditionalism (+.52) when rotation was by the cosine solution, an analytical oblique rotation method invented by Ahmavaara and Markkanen (Allardt 1964, 120—122). Varimax rotation gave a similar result. Sex difference in voting turnout had a loading (—.47) on Factor V composed of variables related to low religious participation, high income, speed of housing construction, and large size of the town or city. This factor is a combination of the Metropolitan and Religious Traditionalism Factors in the oblique rotation. Low women's voting turnout had a loading (+ .31) also on Factor 111, which was composed of variables related to migration, fascist vote, frequency of thefts, Communist increase, lack of people on social relief, and lack of people separated from the church. This factor resembles the factor of Migration in the oblique rotation.

Decrease in traditional religiousness, high economic standard, and stability of
population seem to be preconditions for high voting participation among women
in urban areas. A combination of modernity and stability creates a political

climate without cross-pressures and makes it easy for women to vote. A secularized,
economically developed urban commune probably also contains pressures
toward voting which are absent in less developed towns.

In rural communes with a Swedish-speaking majority very different variables
explain the sex difference in voting turnout. Correlations are many:

Correlations related to the linguistic and political structure are highest. A homogeneous Swedish-speaking commune where people vote for their own party, not for socialist parties provides a uniform culture without cross-pressures on how to vote. The liability of women to cross-pressures has often been used as an explanation for women's low voting participation (Lipset et al. 1954, 1134, Allardt 1956, 126—128). Lack of cross-pressures obviously explains the small sex difference in the Swedish-speaking rural communes.

Other correlations point in the same direction; variance in farm size, people on social relief, assaults, people working in retail trade mean a lack of homogeneity and cohesiveness. Communes characterized by a lack of uniformity have a high sex difference in voting turnout. A negative correlation with the proportion of small farms and a positive one with income per capita point to one characteristic of these uniform Swedish-speaking communes: they are relatively poor or economically underdeveloped. Traditionally cohesive, integrated poor communes provide in the Swedish-speaking rural districts the most secure basis for women to vote.

In the factor analysis sex difference in voting turnout had a high loading (+.80) on the factor of Socialist Traditions in the oblique rotation (Allardt 1964, 122). According to the Varimax rotation, low women's voting participation was also closely tied to socialist traditions and disintegration (Factor VI), with a high loading (+ .76) on this factor. All kinds of disruptive factors in this traditional,

mechanically integrated culture lower the voting participation of women compared
to men.

In the rural communes of southern and western Finland the following
variables are correlated with sex difference in voting turnout:

In urban and in Swedish-speaking rural communes no correlation was found between general voting activity and sex difference in turnout. Here they have the highest correlation. Political activity thus explains women's high voting turnout in the rural communes of the south and west.

Many of the variables correlating with sex difference in voting turnout are in some way related to class-conflicts and inequality in the past: proportion of "Reds" killed in the Civil War, proportion of farms settled by crofters, Social Democratic vote, variance in farm size, proportion of small farms. On the other hand, present and past security (lack of unemployment) decreases the women's voting rate compared to the mens. A low economic level and religiousness also tend to lower it.

According to the published results of the oblique rotation, sex difference in voting turnout had no high loadings on any of the factors in the Finnish-speaking rural communes. According to the unpublished results of the Varimax rotation, sex difference in voting turnout has fair loadings (—.45) on Factor 111 and V. Factor 111 is an equivalent of Factor IV in the oblique rotation. It can be called the factor of Lack of Class Conflicts in the Past. The highest negative loadings are on variables related to "Reds" killed in the Civil War, farms settled by crofters, Social Democratic vote, migration, people on social relief and library loans. When class conflicts have been lacking in the past, women go to the polls. Factor V corresponds to Factor 111 in the oblique rotation, the factor of Insecurity with loadings indicating a high proportion of crowded dwellings, lack of telephones, unemployment, variance in farm size, Social Democratic vote, increase in Communist vote, proportion of persons separated from the church and low amount of library loans. In communes with insecure social conditions,

women's voting turnout compared to men's is high. It is possible that insecure
conditions actually, instead of increasing women's voting turnout, decrease that
of the men, and the sex difference may become small in this way.

A large difference in the voting turnout of men and women in the rural communes of the south and west can be explained by traditionally unequal social structure. This may be due to the low voting rate of working class women. Homogeneous farmer communes lack both sex and class differences.

In the rural communes of eastern Finland variables related to the economic
level are correlated more than anywhere else with the sex difference in political
activity:

Sex differences are large in the poor, sparsely populated areas where people live on forestry and the Communist vote is large. Religiousness is, unlike in the urban communes, negatively correlated with sex difference in voting turnout. In the communes, where people attend religious meetings, the communications may be better than elsewhere, and this may be reflected in voting activity. Somewhat unexpectedly, outflow migration and property crimes correlate negatively with sex difference in voting activity. This may point to a situation in which the traditional backwoods culture is beginning to change. A rise in the economic level is accompanied by migration and an increase in the crime rate but also by a rise in women's voting participation. A stagnating underdeveloped commune with emerging backwoods communism does not activate women to vote.

According to the Varimax rotation, sex difference in voting rate had a fairly high loading (+ .56) on Factor I, which may be called Backwoodsness. It includes almost all of the variables listed above: forestry as an occupation, low income and population density, many small farms, few electrified homes, few telephones, many crowded dwellings, Communist vote, increase in Communist vote, low voting turnout, little migration, increase in income per capita, low variance in income per capita, few "Reds" killed in Civil War, small proportion of industrial and retail trade workers. The only thing correlated with sex difference in voting activity but not belonging to this factor is religiousness.

A decline in backwoodsness and an increase in active religious participation explain the high women's voting turnout in rural eastern Finland. Industrialization and economic development have often been found to correlate with an increase in women's voting (e.g. Milbrath 1965, 136, Holter 1970, 103). Eastern Finland is a typical example of this tendency.

In the rural communes of northern Finland highly scattered variables are
correlated with sex difference in voting:

Some similarities with the results obtained earlier can, however, be seen. First high sex difference is connected with political passivity as in Finnishspeaking rural communes in general. Second, past class-conflicts and Social Democratic traditions but present security and religiousness increase the sex difference. According to the factor analysis, men's voting turnout compared to women's has fairly high loading (—.61) on Factor VII in Varimax rotation, which can be called the factor of Political Activity.

The other grouping of communes made by Erik Allardt divides them into those in (1) the south and west and (2) the north and east. Reference will be made here only to the results of the published factor analysis (Allardt 1967, 354—356) in these two groups of communes. In the south and west, sex difference in voting participation has a loading of +.37 on the factor of National Conservatism, consisting of variables such as the vote of the largest party in each commune (unfortunately this variable was not included in the analysis of the five groups), Fascist vote in 1936 and vote for prohibition in

1931. The proportion of votes cast for the largest party means the strength of a single party in a commune. It may be used as an indicator of political structure without cross-pressures, and is correlated with sex difference in voting (—.37).(^—.37). Women vote most in those communes where one party dominates. This result can be combined with those obtained in the Swedish-speaking rural communes.

In the north and east sex difference in voting activity has a high loading (—.56) on the factor of Political Activity, on which the other high loadings are voting turnout in 1954 and 1958 and a vote for prohibition in 1931. In the north and east, voting for prohibition and an Agrarian vote correlate negatively (—.35 and —.20) with sex difference in voting turnout. In the north and east conservatism does not decrease women's voting participation as it does in the towns.

To summarize the results, we may first point out that different variables are associated with sex difference in voting turnout in different areas. There is not a single variable which would have a significant correlation with sex difference in all groups of communes. Nearest to it come the following:

Class conflicts and inequality in the past were still in 1951 everywhere ecologically connected with uneven participation of men and women in the election of representatives to parliament. Inequality in class-structure reflects itself in inequality in sex roles. Traditions of equality in the society are the best precondition for equality of the sexes.

Other factors explain the difference in the voting turnout of men and
women differently in different areas. To summarize, some of the highest correlations
in each group of communes will be presented.

Some kind of modernization and egalitarianism combined with a high economic level lies in the majority of cases behind a small difference in men's and women's voting turnout. There are, however, some exceptions. In the Swedishspeaking rural communes traditionalism and a low standard of living are connected

with low sex difference. In the eastern rural communes conservatism and outflow migration in addition to improving economic conditions lessen the sex difference. In the north, economic insecurity takes even women to the polls, or keeps men away from them.

II. Political representation

Comparison of Men and Women as Nominated and as Successful Candidates A high voting turnout of women does not automatically mean that women become elected to political offices. Maurice Duverger has schematized the political roles of men and women by a figure which shows how women's influence progressively declines as the higher levels of leadership are reached. "The political role of women grows smaller as we approach the 'centre' of political leadership. There are few women candidates at elections, fewer women members of parliament, still fewer women ministers, and no women heads of governments", he wrote in 1955 (p. 123). Nowadays the situation in the world has somewhat

changed: there are women acting even as prime ministers. But that is not the
case in Finland. There is usually one or at most two female members of the
Cabinet.

Duverger's figure is shown below, compared to the situation in Finland.

The degree to which women are represented in parliament corresponds very nearly to the schematized proportion in the figure by Duverger. In the municipal councils there are fewer women than one would have expected on the basis of Duverger's scheme. The main difference between the Finnish data and Duverger's figure is the small difference in the percentage of women (1) between eligible and actual voters, and (2) between nominated candidates and those elected. The proportional representation system in Finland does not function against women to as great an extent as in countries with other kinds of electoral systems, such as majority representation systems. We may add to the percentages presented in the figure that in the parliamentary elections of 1958—66 women got 15.4 °/o of the votes and in the municipal elections of 1968, 16.3%. In the parliamentary elections women were elected in the same proportion as they received votes, but in the municipal elections the proportion of votes cast for women exceeded the proportion of women elected (10.7%).

The proportion of women elected to parliament has not increased with time quite as steadily in Finland as it has in Sweden. The proportion of women among the candidates is presented here on the basis of Tuula Kaurinkoski's careful investigation for the years 1907—54. Official Statistics give the data for subsequent years:

At the beginning of the century, in 1907—24, the proportion of women

elected to parliament was greater than that of women candidates. Since then men candidates have had a slightly better chance of being elected. Tuula Kaurinkoski found that during 1907—54 female candidates represented 9.3 °/o of all candidates and 9.5 °/o of those elected. There were on the average 3.8 female and 3.9 male candidates per each representative elected. Women were thus elected to parliament in the same proportion as they stood as candidates. Since the war the proportion of those elected among all candidates, according to sex has been the following:

Sex differences have also recently been very small. Women are somewhat handicapped in the elections when they stand as candidates for small parties in small electoral districts. For example, in the Swedish People's party the proportion of women receiving more votes than the minimum necessary for election in the country as a whole has been larger than in other parties. In 1958—66 there were 23.5 % women in this group in the Swedish People's party, but only 13.7 °/o in the country as a whole. At the same time 14.7 °/o of the members elected to parliament were women. Women thus do not belong to the group of "nearly elected" more often than men, except in some small parties. The proportional representation system with single person lists does noi; discriminate against women. Data from countries with other electoral systems show that women are elected to a lesser degree than they are nominated as candidates (Duverger 1955, 87-88, Kaurinkoski 1958, 33-34, Valen 1966, 131).

In Scandinavia Finland has the highest proportion of women in parliament. In the municipal councils there are more women in Sweden. The proportion of women in both bodies is lowest in Norway. This corresponds to the low rate of economically active women in Norway and to the low percentage of girls at

universities there. In Finland there are more women in the labor force and a
the universities than elsewhere in Scandinavia, especially than in Norway (Hoi
ter 1966, 9-11, Haavio-Mannila 1969, 340-341).

Rural—Urban Differences in Voting for Men and Women

3i o / The election of women to parliament is an increasing trend. We may thus expec that women are more often elected as political representatives in the more mod ernized urban than in the rural areas.

The proportion of women among members elected according to their place oi
residence, and the percentage of votes given to women in urban and rural communes
in the parliamentary elections of 1958—66 were as follows:

More women receive votes and are elected in urban than in rural communes. In general, urban residents are overrepresented in the parliament (Noponen 1964, 223, 317). This seems to apply especially to women. As many as 61.4 °/o of the female representatives were of urban origin in the parliamentary elections of 1958—66. Among the men the percentage was 55.4. Tuula Kaurinkoski found that female candidates from urban communes had better chances of being elected than those from rural communes. In 1907—54 only 57.9 °/o of women candidates but 67.2 °/o of women elected lived in towns or cities (Kaurinkoski 1958, 52).

The rural-urban difference can be seen in the proportion of women elected
to the municipal councils since 1945:

The proportion of women in rural municipal councils has been growing more steadily than in the urban ones. The percentage of women in urban councils was highest in 1950 and 1953. In 1968 a sharp increase can be seen. This may be due to the active, radical sex role discussion in the mass media since 1965.

It is also mostly urban women who in practice give women the possibility
of being political leaders. Women in Helsinki vote for women more than members
of the other groups interviewed.

However, a majority of women everywhere votes for men. Only a very tiny proportion of men, 5 °/o, casts votes for women candidates. If sex roles were totally abolished from political life, half of the voters of both sexes would vote for men, half for women. As a matter of fact, women in Helsinki almost represent this "ideal".

Attitudes towards equality of the sexes in political leadership and toward women as leaders in general are more positive among urban than rural women. The rural—urban difference among men is minimal. As an example the distribution of answers to one item measuring this dimension will be presented:

Men in the capital are most hesitant, in their opinions. Rural men have the
most negative opinions, urban women the most positive.

An attitude scale was constructed on the basis of answers to four items measuring women's participation in politics and leadership. The items were intercorrelated for Helsinki on the average by .26. The items were: "Women ought to participate as much as men in the affairs of politics and government", "Women ought to keep themselves in the background when politics and public problems are discussed", "Women ought to get as many chances for leading positions in occupational life as men", and the above-mentioned "Women in general ought not to have managing or leading positions". The scale means for different groups of men and women were:

The radicals in the sex role discussion seem to be urban women. In this case
men have not adopted new ideas before women, as they generally do.

In the first part of this article it was found that sex differences in voting were smallest in urban communes. Now we have shown that the proportions of men and women among political representatives are more even in towns and cities than in rural areas. This is mainly due to the fact that urban women vote for women candidates. Men are as traditional in Helsinki as in the rural communes. To eliminate sex roles in politics, both men and women should vote both for men and women. If we only urge women to vote for women, we in a way create a "Women's party". But to achieve a more even representation of the sexes in parliament and other political bodies, women probably have to fight for themselves, as the "woman power" movement believes. At present men seem to favor candidates of their own sex, even in the most urbanized city in Finland.

Differences between Parties in Voting for Men and Women

The proportion of women among candidates, among those receiving votes and among those elected to the parliament is and has traditionally been highest in the socialist parties especially among the Communists (Tables 1 and 2), lowest in the Agrarian party (recently renamed as The Center party) and in the Swedish People's party. Recently the National Coalition party (Conservatives) has had many women candidates, but the proportion of women among those elected is not as high. The proportion of women among candidates has been lower than among the elected notably in the Social Democratic party and in the small Social Democratic Union of Workers and Small Farmers.

The proportion of votes cast for women in 1948—66 has decreased in both socialist parties, among Communists as well as Social Democrats. In the Agrarian party, on the contrary, a steady increase has taken place since 1954. In the Conservative, Liberal and Swedish People's parties the proportion of votes for women decreased from 1954 to 1962, but in the election of 1966 a small increase took place. This was most noticeable in the Swedish People's party.

This increase in the number of votes cast for women in 1966 in the burgeois parties may be related to the beginning of the radical sex role debate in Finland in autumn 1965. It was especially lively in the Swedish press but also in some other bourgeois newspapers and journals. Since then the other mass media have also joined in the discussion. The effects of this discussion could be seen in 1970 when as many as 21.5 % of those elected to the parliament were women. Unfortunately the detailed results of the last election were not available for this article.

Table 3 shows that this increase of votes cast for women in the bourgeois parties has mainly taken place in urban communes. The overall decline in the proportion of votes for women has, however, also occurred mostly in urban communes. This is mainly due to the decrease of votes cast for socialist women.

Women are elected to municipal councils most frequently in the Social party. Then follow the non-socialist parties, and finally the Communists. The increase in the proportion of women elected to municipal councils has been most noticeable in the non-socialist parties (Haavio-Mannila, 1966, 115). This may be due to the increase of women working in "white collar" occupations. Women represent typical "middle class" groups in other ways also: there are more men at the top and at the bottom of almost any social rank or

other dimension. Voting for non-socialist parties is typical middle-class behavior. Because many women vote for middle-class parties, and women vote for women, women there get many votes. This assumption receives support from the larger number of votes cast for women in the Liberal party compared to the Conservative party. However, the old socialist tradition of the equality of the sexes in political participation still keeps the proportion of votes cast for women high in the Social Democratic party in municipal elections and in both the Social Democratic and the Communist party in parliamentary elections. In parliamentary elections, the Communist party has had some very successful candidates, for example, Hertta Kuusinen, the daughter of the great Communist leader, Otto Wille Kuusinen, who herself has a very high position in the party.

Regional Differences in Voting for Men and Women

Sex differences in voting turnout are largest in eastern Finland. If voting for women follows the same rules as women's voting, as it does in the case of ruralurban differences, we would expect a low proportion of votes given for women candidates in eastern Finland.

The results are, however, quite contradictory to this hypothesis. Women have been elected to parliament, to municipal councils, and to the elementary school boards in the nineteenth century (which are taken as an example of traditional practices) more in eastern than in western Finland (map 2 and Haavio-Mannila 1970, 34—36). This applies to urban as well as to rural communes. But when we re-define this result according to party (map 3), an interesting phenomenon can be seen. In parliamentary elections it is only in the Agrarian and Conservative parties and in municipal elections in the non-socialist parties that this "eastern Finland rule" holds true. In the socialist parties women candidates get most votes in those areas where socialism is deeply rooted and traditional. In traditionally socialist areas the principle of the equality of the sexes seems to be followed better than in the areas of "emerging radicalism".

But why do the Agrarians and the Conservatives vote for women in eastern
Finland? Interviews give some indication of the reasons or background for this.

The attitudes of men toward women's participation in public life are most favorable in eastern Finland. Among women it would seem that the southnorth dimension or that of industrialization count more. Thus it looks as though the attitudes of men toward women as leaders would "explain" the emergence of women political leaders. This result is, however, based on a very small number of communes and must be treated only as a hypothetical one. Some indications pointing in the same direction have been found in the results of a national survey by the Finnish Broadcasting Company in which one item about the equality of the sexes was included (Haavio-Mannila 1969, 345^—346).

We may also refer to the different social and political climate in eastern and western Finland. In western Finland the pressure toward conformity includes traditionalism in the division of labor between the sexes. Women can and must vote, but cast their votes for men. In eastern Finland, people, especially women, need not vote, but when they do, they have a greater choice whom to vote for. There is a more variable distribution of parties and more possibilities of voting for a "deviant" candidate. The parties are aware of this tendency to vote for women, and put forth plenty of women candidates. These are elected in greater proportions than elsewhere.

A change is apparently on the way. Regional differences are diminishing. Levels of industrialization and urbanization are beginning to count more. The southern and central areas of Finland are more developed than northern and peripheral Finland. The dividing line is more that between center and periphery

than between east and west. Considering the traditional importance of the difference between eastern and western Finland it seems permissible to say that a change is occurring in the social structure, and that the change in sex differentiation is an indicator of this more general change.

Summary

The results of this study on the voting activity of men and women in Finland, on the election of men and women and on the attitudes and expectations related to this activity are, in general, consistent with those of previous studies and with the propositions set forth in the beginning. Some additions to the facts known earlier can, however, be reported.

Apart from the obvious effects of urbanization, of a rise in the socioeconomical level, and of a decrease in religious and conservative traditionalism, some other factors also help to explain small differences in the voting turnout of men and women. For example, in some areas, notably in the Swedish-speaking rural communes, economic development increases sex differences by bringing with it social disorganization and cross-pressures. An important result of the ecological analysis was that all over the country past class-conflicts, traditions of inequality of social classes are reflected in present inequality or difference in sex roles in voting participation. Socialist traditions, which according to Marxian ideology stress the equality of the sexes in political life, and which are strong in the same communes where class conflict and inequality of the sexes prevail, are unable to decrease sex differences when the social structure is characterized by class conflicts.

The social and political climates in eastern and western Finland differ remarkably from each other. In eastern Finland women's voting activity is low but a high proportion of women is elected to parliament and municipal councils, notably in the bourgeois parties. This is due to a lack of uniform norms as to whether and for whom to vote. Pressure towards conformity in western Finland brings women to the polls but to vote for men. This shows that women's voting and voting for women are not one-dimensional phenomena in all social circumstances. On the other hand, it was found that urbanization increases both women's voting and voting for women. — The liberal attitude of men in eastern Finland towards women's participation in politics and leadership, reflects lack of homogeneous conservatism and provides a good foundation for those women who want to be politically active. Women in eastern Finland are not, however, as already mentioned, active as voters but when they are, they often vote for women candidates. The east—west cultural difference seems to be losing importance at present and the south—north or center—periphery difference which is related to economic development, is gaining in importance.

The influence of sex role ideologies could be seen primarily in voting for women, not that of women. The Marxian principle of the equality of the sexes is reflected in the larger proportion of women among candidates and those elected in the socialist parties. This proportion has for some reason been declining

since 1954. On the other hand, voting for women candidates in bourgeois parties
has increased. This may be due to the growing number of women working
in middle class occupations. Women get votes almost always from other women

— only a tiny proportion of men votes for women — and these educated middleclass women increasingly cast their votes for women candidates. Another reason for the increase in the proportion of votes cast for women in bourgeois parties particularly in towns and cities, may be the radical sex role discussion which began in 1965 and which may have influenced the voting in the elections of 1968 and 1970. This assumption receives support from the sharp increase already in 1966 in votes cast for women in the Swedish People's party, whose supporters included the initiators of this discussion, imported from Sweden.

Socio-economic development, conservative conformity, traditions of equality in the social structure and ideologies concerning equality of the sexes are on the basis of this study some of the factors influencing sex roles in Finnish politics. There is however considerable variation in their importance in different social groups and geographical areas. They also have different effects on the voting turnout of men and women and on the voting for men and women.

REFERENCES

Allardt, Erik. Social struktur ocb politisk aktivitet, Borgå, Soderstroms, 1956.

Allardt, Erik. "Patterns of Class Conflict and Working Class Consciousness in Finnish Politics"
in Erik Allardt and Yrjo Littunen (eds.). Cleavages, Ideologies and Party Systems, Transactions
of the Westermarck Society, vol. X, Turku, 1964.

Allardt, Erik and Pertti Pesonen. "Cleavages in Finnish Polities' in Seymour M. Lipset and
Stein Rokkan (eds.). Party Systems and Voter Alignments, New York, The Free Press, 1967.

Braunias, Karl. Das Parlamentarische Wahlrecht, 11, Berlin, Walter de Ghuyter & Co. 1932.

Dahlstrom, Edmund. "Analys av konsrolldebatten" in Kvinnors liv och arbete, Stockholm, Studieforbundet
Naringsliv och Samhalle, 1963.

Duverger, Maurice. The Political Role of Women, Paris, Unesco, 1955.

Galtung, Johan. "Foreign Policy as a Function of Social Position", Journal of Peace Research,
no. 3—4, 1964, pp. 206—231.

Haavio-Mannila, Elina. "Aktivitet och passivitet bland kvinnor i Finland", in Kynne eller kon?
Om konsrollerna i det moderna samhdllet, Stockholm, Raben & Sjb'gren 1966, pp. 105—131.

Haavio-Mannila, Elina. Suomalainen nainen ja mies, Porvoo, Werner Soderstrb'm Osakeyhtio,
1968.

Haavio-Mannila, Elina. "The position of Finnish Women, Regional and Cross-National Comparisons",
Journal of Marriage and the Family, no. 2 (May), 1969, pp. 339—347.

Haavio-Mannila, Elina. Sex Roles in Politics, Research Reports, University of Helsinki, Institute
of Sociology, no. 139, 1970.

Holier, Harriet. "Women's Occupational Situation in Scandinavia", International Labour Rewiew,
no. 4, 1966, pp. 383—400.

Holter, Harriet. Sex Roles and Social Structure, Oslo, Universitetsforlaget, 1970.

Jutikkala, Eino. Suomen talonpojan historia (History of the Finnish Farmer), Helsinki, Suomalaisen
Kirjallisuuden Seura, 1958.

Kaurinkoski, Tuula. Naisehdokkaat Suomen eduskuntavaaleissa, Master's thesis in political science,
University of Helsinki 1958.

Lipset, Seymour M., Paul F. Lazarsfeld, Allen H. Barton and Juan Linz. "The Psychology of
Voting: An Analysis of Political Behavior", Handbook of Social Psychology, vol. 11, Cambridge,
Mass., Addison-Wesley Publishing Company, Inc., 1954, pp. 1124—1175.

Martikainen, Tuomo and Risto Sånkiaho. Åånestysaktiivisuus v. 1968 kunnallisvaaleissa, Research
Reports, Institute of Political Science, University of Helsinki, no. 16, 1969.

Milbrath, Lester W. Political Participation, Chicago, Rand McNally, 1965.

Noponen, Martti. Kansanedustajien sosiaalinen tausta Suomessa (The Social Background of the
Members of Finnish Parliament), Porvoo, Werner Soderstrom Osakeyhtio, 1964.

Official Statistics of Finland, VI and XXIX, Helsinki, Central Statistical Office.

Riihinen, Olavi. Teollistuvan yhteiskunnan alueellinen erilaistuneisuus, Porvoo, Werner Soderstrom
Osakeyhitio, 1965.

Statistical Yearbooks, Central Statistical Offices of Finland, Denmark, Norway and Sweden.

Tingsten, Herbert. Political Behavior, Stockholm Economic Studies no. 7, Stockholm, P. A. Norstedt
& Soner, 1937.

Valen, Henry. "The Recruitment of Parliamentary Nominees in Norway", Scandinavian Political
Studies, vol. I, 1966, pp. 121—166.