In 1809, Finland was granted a national diet in accordance with the Constitution of Sweden. This assembly of. four estates was elected twelve times during the years 1863—1904. The change-over to the present unicameral Parliament (in Finnish, Eduskunta, and in Swedish, Riksdag) came in 1906 in accordance with the Parliament Act of that year. Using d'Hondt's method of proportional representation, two-hundred members are elected to the Parliament. The term is four y^ars (up to 1954, three years), but the head of state has the right to dissolve the Parliament

---

6 Erling Bjøl, La France devant l'Europe. La politique européenne de la IVe République, (Munksgaard, Copenhague, 1966).

and order new elections. Since 1906, both men and women have had equal suffrage. During the years 1907—1917, before Finland obtained her independence, eight elections were held. During the years 1919—1939, there were nine elections, and after World War 11, there have been seven elections (1945, 1948, 1951, 1954, 1958, 1962, 1966). Official results of all these twenty-four elections have been published.

The President of the Republic of Finland is elected every six years through indirect voting. In January, through proportional elections, three-hundred members are elected to an Electoral College, who in turn elect the President on February 15. There is no limitation on how many terms a President may serve. The Republic's eighth President was elected in 1956 and reelected in 1962. In municipal elections, equal voting rights were acquired in 1917 when election to the communal councils also became mandatory, replacing an earlier system of inhabitants' meetings. The municipal elections are held every fourth year in October. These too have acquired a general political significance. The three above-mentioned elections for Parliament, President's Electoral College and communes never occur at the same time.

The Finnish electoral and party system offers some special opportunities for study. Because the elections for the Electoral College and the Parliament use the same organizational base, voters' behavior can be compared in these elections. From the point of view of ecological studies it is significant that throughout the last decades there has been little change in the election system, the election results also being reported since 1907 by communes. Important parties and the major political ideologies have been the same throughout this time. Because the elections take place on two consecutive days, it is possible to compare each day's voter participation. A distinctive feature of Finnish political life (different in this regard from other Scandinavian countries) is the relatively strong support of the Communist movement.

A short English language summary of Finland's parliamentary elections, election statistics and election studies will appear this year.¹ It is sufficient here to mention that historical election studies have been carried out and that voting behavior has been explained ecologically and through the interview method.

The first significant ecological studies appeared in 1956. Jaakko Nousiainen compared two constituencies in the elections between 1922—1954 in order particularly to explain the bases of Communist support.² This study is in the tradition of French political geography and is similar to Sven Rydenfelt's study of Communist support in Sweden.³ Erik Allardt used data from all the communes of Finland in order to explain the influence of various social factors on voter participation/ Later he used new commune data to study the ecology of political radicalism which he first connected with the societal class structure and later with a more general theory of structural tension.⁵ In his ecology he moved in 1965 further into a comparative framework in order to analyze structural tensions on a theoretical level (210).⁶

---

1 Chapter on Finland by Onni Rantala in the International Guide to Electoral Statistics, to be published in 1967 by the International Committee on Social Science Documentation.

2 Jaakko Nousiainen, Kommunismi Kuopion läänissä (Communism in the Kuopio province, Joensuu: 1956).

3 Sven Rydenfelt, Kommunismen i Sverige (Communism in Sweden, Lund: 1954). 4 Erik Allardt, Social struktur och politisk aktivitet (Social structure and political

4 activity, Borgå/Helsingfors: Söderström, 1956).

5 See the bibliography in Scandinavian Political Studies, volume I (1966), items 470, 480, 546-553, 598, 599.

6 These references in the text are made to the bibliography for 1965, published in the present volume. — See also items 523—524 of the bibliography in volume I of Scandinavian Political Studies.

The first example of the use of the interview method was a panel interview study in conjunction with ;he 1956 presidential election which dealt with the voting behavior of university students.⁷ Its purpose was to make comparisons with panel studies made in the United States and Britain to see if their findings could be useful in analyzing the Finnish electorate. The results showed significant similarities of behavior.

The bibliography for 1965 is included in this volume. In that year, two works on voting behavior were published. Onni Rantala finished the first part of his basic study of political regions (200); his paper on pages 117—140 is based on this study Pertti Pesonen published a general account of his 1958 election study (241). The two programs will be presented below. These will be followed by information on studies analyzing the 1966 parliamentary election.

It should be noted that the following review does not include all election studies being made in Finland. Other programs include an analysis of interview material gathered after the 1962 elections in Helsinki,⁸ as well as a general presentation by Tuttu Tarkiainen of electoral systems, voter participation, and election results through the years 1907—1962, which belongs to the history series on the legislature of Finland. There is also Risto Sankiaho's ecological analysis of voting behavior in the constituency of Uusimaa.

1. Basic Study of Finland's Political Ecology

The political-ecological mapping of Finland takes in the whole country. In order to get the necessary historical perspective the areas seceded to the Soviet Union in the 1947 Peace Treaty are also included. The support of the parties and participation in elections are determined on the basis of the parliamentary elections held during the years 1907—1958. The analysis comprises the period's 22 elections. Because of the general static character of relative party strength, it might have been possible to define the political regions on the basis of a representative sample of elections. But when changes are analyzed, measures of support are inevitably needed from nearly all the elections.

The commune was found ',o be the most suitable unit for such an extensive study in area and time-span; values of the political and social variables were thus calculated for each commune. Since the number of communes in the country varied during the years 1907—1958, (1907: 514, 1958: 549) the communes that had been separated or amalgamated during this period were treated as integral units starting from the beginning of the study period (all-in-all 503). Area additions, which sometimes crucially changed the basic political color of a commune, had to be omitted for these sometimes did not even follow the boundaries of voting districts inside the communes.

From the point of view of this study, the official election statistics of Finland
have been useful to present the election results by party and commune. Only ä
couple of inadequacies in the time series create difficulties. In the earlier statistics,

---

7 Pertti Pesonen, Valitsijamiesvaalien ylioppilasäänestäjät (Student voters in the Election of Presidential Electors, Vammala/Helsinki: Tammi, 1958); summarized in Pesonen, The Voting Behaviour of Finnish Student's, Democracy in Finland (Helsinki: The Finnish Political Science Association, 1960), pp. 93—104.

8 Only some articles have been published: Olavi Borg, Monipuoluejärjestelmien ideologisten peruserojen kuvaE.misesta (On the concepts describing basic ideological cleavages in a multi-party system), Politiikka 1963, pp. 1—13; Borg, Helsinkiläisten äänestäjien mielipiteet sosiaalisen ja taloudellisen kehityksen esikuvista (Voter's opinions on models for social and economic progress), Politiikka 1964, pp. 207—234; and Mikko Valtasaari, Helsirikiläiset äänestäjät presidentin valitsijamiesvaaleissa 1962 (Helsinki voters in the 1962 presidential elections), Politiikka 1963, pp. 113—131.

7 Pertti Pesonen, Valitsijamiesvaalien ylioppilasäänestäjät (Student voters in the Election of Presidential Electors, Vammala/Helsinki: Tammi, 1958); summarized in Pesonen, The Voting Behaviour of Finnish Student's, Democracy in Finland (Helsinki: The Finnish Political Science Association, 1960), pp. 93—104.

8 Only some articles have been published: Olavi Borg, Monipuoluejärjestelmien ideologisten peruserojen kuvaE.misesta (On the concepts describing basic ideological cleavages in a multi-party system), Politiikka 1963, pp. 1—13; Borg, Helsinkiläisten äänestäjien mielipiteet sosiaalisen ja taloudellisen kehityksen esikuvista (Voter's opinions on models for social and economic progress), Politiikka 1964, pp. 207—234; and Mikko Valtasaari, Helsirikiläiset äänestäjät presidentin valitsijamiesvaaleissa 1962 (Helsinki voters in the 1962 presidential elections), Politiikka 1963, pp. 113—131.

parties which formed joint electoral alliances have not been treated separately. This created a problem, for example, because in those elections where there were most considerable shifts in party votes there were unusually numerous joint electoral alliances. In addition, the statistics concerning the communes lack the absentee ballot count. This to a certain extent distorts the final results, especially, of course, for those parties which receive relatively many absentee ballots.

The percentage of valid votes was used as the measure of the parties' standings in the elections. The total number of enfranchized citizens was not used as a base because the strong parties in the islands and the sparsely populated areas would have suffered due to difficult participation and consequent low turnout, particularly in the initial period of the study.

Since this is a basic study, parties and groupings were selected as the objects of analysis which cover the whole political scene. The study concerned itself with four, approximately equal groupings: the Social Democrats, Communists, Agrarian Union (now Center Party), and the rest of the bourgeoisie (made up mainly of conservative and liberal parties). With the exception of the Communists, all the above-mentioned groupings have functioned on the Finnish political scene during the total observed period. The Communists first participated in the election of 1922; in the elections of 1933—1939, the Communist movement was outlawed. Nevertheless, the Communist group is such a unique phenomenon on the Finnish political scene that it deserves to be treated separately. In this way the study, in a sense, is symmetrical with both the right (bourgeoisie) and the left split into two groups.

The object of analysis is the politically formed region. Only general comments and observations are made about such other types of area which are of secondary importance to political behavior. Therefore, relatively little attention is paid to such administratively limited groupings as cities, city-adjacent communes and rural communes; Swedish, bi-lingual, and Finnish speaking communes; the islands; or the single member and multi-member constituencies (of the former Finland has had two, Lapland 1907—1936, and Aland since 1948).

This research is divided in three parts and will be published (in Finnish) accordingly: (a) Finland's Political Regions, which has appeared in a mimeographed edition in 1965 (200) and is presently at the press; (b) Regional Changes in Party Support, presently being completed; and (c) The Emergence and Background of Finland's Political Regions. A brief account of the procedure of each part follows below.

(a) Finland's Political Regions

In the first part of the series, attention is centered on the static features of the political scene, such as the geographic distribution of party support and a description of political regions and their internal structure.⁹ By looking at the two right and the two left groupings side by side, the purpose vas to find out their mutual relationships. This method proved unsuccessful, for in many cases a right grouping had to be compared with a left one and vice-versa. Also this group comparison was not necessary for analyzing the total bourgeois or leftist support, because there are no uniquely bourgeois or left districts in Finland; the left is characterized primarily by the presence of Social Democrats and the bourgeois either by that of Agrarian support or that of other bourgeois parties. Experience was also gained which showed that "the other bourgeois" was too heterogeneous to be treated as one group, regardless of the fact that these parties' areas of support were found to be near each other.

---

9 Some findings are reported by Qnni Rantala, The Political Regions in Finland, on pp. 117—140 in the present volume.

9 Some findings are reported by Qnni Rantala, The Political Regions in Finland, on pp. 117—140 in the present volume.

9 Some findings are reported by Qnni Rantala, The Political Regions in Finland, on pp. 117—140 in the present volume.

The groupings' support in each commune in the parliamentary elections, 1907— 1958, is represented by the arithmetical percentage support average. In order to determine the stability of party support which is a prerequisite of political regions, the dispersion was also calculated for the elections in question. With the help of the averages, the extensiveness of each political grouping was determined throughout the country. Here the basic criterion was a percentage representing an even ten close to the empirical lower quartile. In the extension areas, one or more support areas were found in which a group's popularity has remained continually stronger than the average — the average in this case being represented by an even ten near the median. Even in the support areas, certain regional differences and groupings were observed. This lead to the establishment of a particular core area, or areas. The minimum support level for these was represented by an even ten near the upper quartile. That part of the support area which does not belong to the core area is called the core's area of influence. The communes which obtained a similar level of support, the so-called separate pockets, were catalogued so they could be used as a comparison group of the core areas.

The structure of the core areas was analyzed on the basis of outward symbols and political dimensions. Because the core area does not form an exact unit (due to the general nature of polilical regions), in the mapping of border lines a rough outline rather than a detailed boundary was drawn. At the same time an effort was made to see to what extent the core areas correspond with economic regions and administrative districts (such as provinces or constituencies), and in what way are their borders shaped by geographical features such as large water barriers. In seeking to determine the district's central places of importance, the historial aspects were not considered as important as the functions of a place in political activity. The criteria for this were: the location of the Parliament member's home, the location of party district headquarters, and the commune where the parry paper is published. The number of central places, the type of commune and location within the core area added characteristic features for the study of the parties and regions. The political basis of the core area was analyzed in order ;o determine the relative strength of the parties and voting turnout. By looking at the basis of party support, it was possible to tell how the area's political orientation and location affect the possibilities of different groups to establish a bridgehead there. By looking at the relationship between voting activity and political power, an effort was made, among other things, to observe under which preconditions do cross pressures lessen voting turnout in politically static and homogeneous areas.

In the final, coordinating chapter of part one of the study, all four groups are viewed parallel. In this way it was easier to obtain a general picture of the basic ecological characteristics of the political field. The political map on which was marked the largest (and absolute majority) group in the commune, 1907—1958, tells us about the leading pclitical orientations around the country and creates a basis for studying the general political climate. The core-area map, in turn reveals what parts of the country are politically static and homogeneous, and which are the unstable or labile and heterogeneous regions outside any core areas. The map also shows to what extent both groupings of communes vary in relation to geographical factors and the prevailing political atmosphere. Taking as a base each group's country-wide support median, these were compared to find out how many groups in each commune achieved this figure, and accordingly which parts of the country could be listed as one, two, three, or four party regions. The question was also asked which group» in which regions and in which coalition get along well with each other.

(b) Regional Changes in Party Support

In the second part of the study dealing with changes in party support, the four groups are compared parallel with each other. In addition to the experience gained in the first part, this method was followed because of the finding that in the parliamentary elections an increasing movement has taken place across the traditional bourgeois-left cleavage.

The changes in party strength are analyzed on the basis of long periods from which temporary fluctuations are eliminated, and short periods reflecting the impact of current situations. Characteristic longer periods include the total time span 1907—58 and its three subperiods, 1907—1917 (last decade of Russian rule), 1919—1939 (independent Finland's first two decades), and 1945—1958 (postwar political atmosphere). The effect of the current developments on the political map is clarified by such factors as the work-stoppages in the beginning of the century, the periods of unemployment, renewal of the party system, the Civil War, Winter and Continuation War, as well as the movement of people from the areas seceded to the Soviet Union.

For the longer periods, 25—30 social variables have been selected to explain the political transitions. In all the time periods the most essential variables were repeated, such as the occupational structure, landownership and area under cultivation, income and education, living space and social care, population movement and voter participation. The political changes are compared to the social changes during the same periods and not to such values of the structural variables which would represent a given point of time.

The change is measured by the difference of the percentages at the end and at the beginning of a period. To understand the relationship between political and social changes, a factor analysis is first performed covering the whole country. When the best explanatory variables are found, they will have to be arranged in rank order through discrimination analysis separately for four or five areas which represent different levels of political change.

In addition to delineating the degrees, directions, and reasons for change, the study maps out the routes of ideas, as well as the time required for change to take place in regional power relationships. Besides the political and social environment, the direction in which ideas travel may depend on local contact systems such as travel connections and population density. In seeking to find out how long the process of change takes, the following questions will be asked: how long does it take to change the basic political coloration of the community, what differences can be observed in the dynamics of political expansion and contraction, does change come in stages, is change preceded by faltering, and what influence is excerted by the relative strength of the parties, the population structure, and the character of a constituency?

In addition to looking at areas where change has taken place, other areas of the land are mapped out where the relative strength of the parties has faltered. Thereafter, when one compares the politically stable, changed and faltered areas, one can deduce what kind of problems help bring about conditions of stability, consistent change, or lability. Moreover, it will be observed how much ground the three areas, representing different modes of behavior, have acquired from each other.

(c) The Emergence and Background of Political Regions

The third part of this research project will seek the reasons for regional
concentration and stability of party support.

The drawing up of an inventory of cause factors is difficult, because in Finland
as in many other countries it is hard to find statistical time series from an early
enough period. For this reason the study must partly rely on qualitative analysis.

In addition to explaining the causes of party support, the study seeks to relate
the support to later events which have strengthened accepted political traditions,
or are self-explanatory in the context of the political environment.

With the aid of relevant variables the aim is to further explain the present characteristics of the political regions and see how they might be related to the original political background of these areas. At the same time, information is gained about the character of the parties and their social roots and what changes have taken place in these relationships. The methods of determining the correlation between political and explanatory variables will be similar to those used in the second part of the study.

The ecological research program which has been outlined above, was launched in 1960. The first part was largely completed in 1965. The second dynamic study will be ready in 1967 or 1968, and the final explanatory analysis is scheduled to be published in 1969.

2. A Study of Party Activities and Voter Reactions

Finland's second panel interview study on electoral behavior (241)¹⁰ dealt with the parliamentary elections held July 6 and 7, 1958. This study is in many respects a continuation of the study mentioned in the early part of this article which compared the voting behavior of Finnish students with findings of studies done in the United States and Britain. This time a random sample of 501 Tampere (Finland's second largest city) inhabitants were interviewed twice. A smaller sample from the rural commune of Korpilahti, which is situated in the same constituency, were also interviewed. The pxtrpose here was broader than just a study of voting behavior, for the attempt was made to outline a comprehensive picture of the parliamentary elections.

This study has two general theoretical points of departure. First a distinction is made between the concept of the 'normal' political process and that which interrupts it, the election campaign. An effort is made to determine the length of the election campaign through various qualitative methods as well as through quantitative analysis of newspapers. Secondly, the study accepts the view that the political system makes authoritative decisions. The system itself is pictured as a multifaceted communications network, where the political powers send communications in that direction which has the authority to make decisions concerning them. By combining these two viewpoints the election campaign is defined as a limited period, distinguishable as separate from the 'normal' political process, during which the political powers turn their attention from the official decision makers to the "people", the electorate, in order to influence vote decisions.

The newspaper analysis gave empirical support to this view: the relative
emphasis in the content of ths political material changed near election time. For
example, the two local newspapers of the left, one Social Democratic, the other

---

10 Earlier article? include iter.is 526, 533 and 571 of the bibliography in volume I of Scandinavian Political Studies. A new English version of the book by P. Pesonen, entitled An Election in Finland, will be published by Yale University Press in 1967.

10 Earlier article? include iter.is 526, 533 and 571 of the bibliography in volume I of Scandinavian Political Studies. A new English version of the book by P. Pesonen, entitled An Election in Finland, will be published by Yale University Press in 1967.

Communist, intensified their criticism of the other competing leftist party. This may have stemmed from the competition between these parties who were partially wooing the same voters. Simultaneously, the party papers left the central government party, the Agrarian Union, to a relatively lesser treatment of criticism for they did not have to compete so strenuously with it for the local voters.

In analyzing the citizens' vote decisions, the premise of the study was correspondingly twofold. The final voting decisions were explained, on the one hand, by behavioral predispositions observed before the election campaign and, on the other hand, by the influences of the actual campaign. This applies to both the decision to take part in the election and to the choice of one's party.

On this basis, it can be said that the 1958 election study is divided into three parts: (1) As a point of departure, the 'normal' political process is analyzed particularly on the level of the localities under study. The book presents the results in the chapters entitled: "An Industrial City and a Rural Commune", "The Indifferent and the Interested Citizens", "Party Identification", and "Newspaper Content and the Citizens' Opinions". (2) The election campaign is then studied as a kind of 'intervening variable'. This was essentially the campaign work of parties, candidates, and private persons as well as the public's direct reactions toward it. This presentation is mainly in the book's chapters "The Framework of the Campaign" and "The Content of the Campaign". (3) The voting itself is analyzed in the two chapters: "Participation in the Election" and "The Choice of Party and Representative". Throughout the study attention was given to the center-periphery dimension, or the differences between the capital, the industrial city, and the rural commune. Finally, the interview findings of some individuals and small groups could only be explained by considering that the general political situation even influenced political action at the micro-level.

In this connection, we can only briefly point to some of the findings of this study. As one could expect on the basis of several earlier studies, it proved of central importance to divide the supporters of different parties into strong party identifiers and weak identifiers. Especially, the strength of identification with one's party was related to one's interest in politics: The strong party identifiers were most interested, and the non-partisans were the least interested. Among the party supporters, the strong identifiers tried to be near the group's official norm sources showing a special tendency to read their party's main organ published in the capital. In reviewing separate political opinions, three levels also became evident. First, as the political newspapers presented their party's 'official' viewpoint, the strong indentifiers sometimes held a different opinion but did not deviate as often as the party's weak identifiers. There was actually a four-level radicalism within the left: furthest away from each other were the strong Communists and strong Social Democrats, while the weak Communists and weak Social Democrats were kind of in-between groups. But among the whole city population the Social Democrats were in a difficult ideological squeeze between the Communists and the non-Socialists.

It is impossible to determine precisely when the election campaign started. In any case, measuring the length of this campaign by different methods revealed that the study's first wave of interviews two months before the election had been administered before the parties' campaign activity had begun to influence people's opinions. The election campaign reached its peak only 10—11 days before the elections. Practically every citizen was in some way an object of its activity. Yet the campaign was generally viewed indifferently, even with annoyance. The parties received the more campaign help from their supporters the better established their local organizations were in the city. Regarding the external

form, the 1958 election in Fir.land was an historical event in that it was the last
before television established its;elf as the third great mass medium of communication.

In studying election participation it was observed that the active part of the electorate was interested in politics 'normally' as well as during the campaign. Along with the high turnout of this part of the electorate there was a tendency to decide how to vote long before the elections and also to cast a ballot early, preferably on the first voting day. In analyzing non-voting this study does not point to lack of interest and external difficulties as alternative explanations of non-voting, as is usually done. Instead, participation is considered the result of two interrelated factors. Those people vote whose motivation is stronger than the resultant bother it takes 1o vote. Correspondingly, the more special difficulties reduce voting participation the: less motivation there exists. The difficulty of choice was one factor which reduced the desire to vote. On the other hand, the factors which did motivate to vote included the perceived importance of the election and, especially, a person's strong party identification.

The great majority of the voters had made their party choices before the election campaign. Only a small minority either changed parties or crystallized the choice close to election dates. Thus the main result of the campaign was to activate the party's supporters and to increase their loyalty. One reason for this was that informational effort reached those voters best whose view was most easily changable. These findings correspond to those made in two-party countries.

Personal influence showed effectiveness as compared to direct mass communication. The findings of the study corresponded to the well-known hypothesis of the "two step flow of communication". But, in addition, there was reason to suggest that the hypothesis be made more exact as a "two and one-half step" model. To this has been added a new phase called the 'horizontal' level of information exchange where then? is no leader-follower relationship. On the other hand, the last phase is seen as incomplete because a part of the population has not been politically influenced by personal contacts or the mass media. The influential persons were divided into two classes: Those that did campaign work on their own initiative and those who through the initiative of their small groups were opinion leaders. A hypothesis was also presented for later studies proposing that the demand for oponion leaders was dependent on the particular political situation.

The generalization could be made that the findings explaining voting behavior which have usually concerned themselves with two-party countries only can in themselves be well adapted to Finnish conditions which are characterized by the multi-party system and proportional representation.

A brief mention might be added about the research methods of the 1958 study. The general design of the interview study was primarily a two-wave panel interview concentrating on on«; community. Its new feature was the simultaneous use of recorded source materials. Some interview findings were verified with the use of appropriate name lists, such as the names of the leaders of local party associations and the signatures from the nomination petitions of candidates. It was possible to check first on their voting district and then on their participation in the election. The interviewees' answers concerning their voting participation were similarly compared to the information available on the registers (poll books) marked by the election boards. Other information about the interviewees was obtained from the city taxboard. In cooperation with election boards in certain voting districts, the time when people voted was observed. Perhaps the most noteworthy methodological exi>eriment was to check whether those who were interviewed twice in 1958 took part in the 1962 elections. This extended time

span of the panel study made it possible to verify many of the conclusions concerning political participation. In addition, it led to the observation that one group of voters who backed a new, unsuccessful party in 1958 later became alienated from politics.

3. The 1966 Election Study

Since the aforementioned studies, parliamentary elections were last held in Finland March 20 and 21, 1966. In 1958, the Communists and the Conservatives were victorious, but the Agrarian Union suffered a reversal. The Social Democratic Party started its downhill course which continued in the 1962 elections. On the other hand, the 1966 election proved a comeback for the Social Democrats but a reversal for the Conservatives and the Communists particularly.¹¹

At the Institute of Political Science, University of Tampere, some interrelated studies are underway as well as some separate studies which try to explain voting behavior and other aspects of the 1966 election. Of these the following deserve to be mentioned:

(a) For a secondary analysis it is possible to to use material which the Gallup Institute gathered primarily for the purpose of predicting the vote by interviewing in February, 1966, a national sample of over 2,000 persons. Initially this work was done as for a single non-party newspaper, the Helsingin Sanomat. With the same sample and at the request of the same newspaper, a follow-up interview was conducted a half year after the election adding considerably to the scientific value of the data. For this secondary analysis, the main concerns are the switching of party loyalties and the effects of mobility on party choices.

(b) In April, 1966, or just after the election, interview data covering the whole country were gathered. This was the first national sample interviewed for the exclusive purpose of election studies. The interview phase of this was also done by Gallup of Finland. Unfortunately, it was not possible to conduct a two-wave panel interview. Altogether 1,264 interviews were conducted. In the sparsely populated areas, the sampling interval was twice the general one; for this reason the analysis includes 1,550 cases.

The three main purposes of this study are (1) to obtain accurate information about the social background of the voters of Finland's various parties; (2) to compare different mass media in these elections, in which for the first time television evidently was greatly more influential than radio; and (3) to create possibilities for comparisons with studies conducted in other countries, such as was started in Norway in 1965. At the same time it is possible to continue the analvsis of certain problems from the 1958 study, such as comparing the choices made of parties with those of the individual candidates.

(c) In connection with radio and T.V. coverage of the elections, an experimental quick prognosis of voting results was conducted based on a carefully selected samole of some 300 voting districts. This was planned and conducted by Risto Sänkiaho. The experiment was quite successful. With the help of computers, the election results were known fairly accurately in two hours. This was considerably ahead of the official count.¹²

---

11 Cf. the review bv Maenus Lembere on nn. 247—252 in this volume.

12 Histo Sänkiaho, Vuoden 1966 eduskuntavaalien ennustaminen (The Use of Computers in Predicting the Results of the Finnish Parliamentary Election of 1966), Politiikka 1966, pp. 57-66, 83-85.

11 Cf. the review bv Maenus Lembere on nn. 247—252 in this volume.

12 Histo Sänkiaho, Vuoden 1966 eduskuntavaalien ennustaminen (The Use of Computers in Predicting the Results of the Finnish Parliamentary Election of 1966), Politiikka 1966, pp. 57-66, 83-85.

(d) The study is also concerned with the ecological aspects of the election results with the commune as the basic unit, analyzing such aspects as regional party support in 1966 and the areal changes that took place 1962—1966. Party loyalties and the shifting of party support is explained by both political variables (including political tradition dating back to 1929, turnout and the support of other parties), and by social variables (including occupation, social class, income level, development of the communes, availability of the mass media and how the communes stand in regard to the movement of population). The ecological part of the study, which uses factor analysis, has been planned in such a manner that it helps to test many of the interview findings.

(e) A study of the campaign work of the individual candidates to the Parliament is also under way.¹³ Added to this is the comparing of social backgrounds of parliamentary candidates with those members who were elected. This work's problem structuring reminds us of a Norwegian study on the same subject.¹⁴

(f) In order to develop a more effective method of measuring the direct influence of mass communication, the semantic differential was tried in connection with the "great election debate" conducted on T.V. and radio. Supplemented with transformation analysis, the method makes it possible to demonstrate especially the immediate influences of Ihe mass media on those to whom the transmission is directed. Kaarle Nordenstreng and Osmo A. Wiio report on this work in a separate article on pages 223—244 of this volume.

As can be noticed, the research on the 1966 election is rather diffused. It has not been feasible to plan and carry out an extensive but unified research program. Yet, an attempt is made to bring together the findings of the various projects in order to form a unified presentation.