1. Local government and the first local elections: an introduction

2. Theoretical basis.

2.1 The basis model

2.2 Voting

2.3 Socio-demographic determinants of turnout

2.3.1 Ethnicity

2.3.2 Life-cycle argument

2.3.3 The socio-economic status argument

2.3.4 The size and character of a settlement

2.4 Political attitudes

2.4.1 Political efficacy

2.4.2 Trust and satisfaction

2.4.3 Party sympathy

2.4.4 Interest in politics

3. Empirical analysis

3.1 Findings

4. Discussion

5. Summary

6. References

THE FIRST LOCAL ELECTIONS IN RE-INDEPENDENT ESTONIA:

DETERMINANTS OF VOTING ACTIVITY

 

  1. Local government and the first local elections: an introduction
  2. On October 13-17, 1993, elections were held to local governing bodies in Estonia. These elections were characterised by public apathy -- the national level turnout was 52.6% (Table 1).

    Table 1. Official statistics of the local elections in Estonia, October 13-17, 1993

     

    Population

    Voters

    Actually voted

    Percentage

    COUNTIES

    Harju

    578 433

    316 441

    185 430

    58.6%

    Hiiu

    11 817

    8 079

    4 324

    53.5%

    Ida-Viru

    213 114

    92 772

    61 140

    65.9%

    Jõgeva

    42 454

    27 880

    12 814

    46.0%

    Järva

    43 736

    28 695

    14 345

    50.0%

    Lääne

    33 333

    20 277

    10 486

    51.7%

    Lääne-Viru

    77 404

    47 947

    22 717

    47.4%

    Põlva

    36 269

    24 518

    11 460

    46.7%

    Pärnu

    99 810

    63 834

    31 113

    48.7%

    Rapla

    39 975

    24 840

    10 672

    43.0%

    Saare

    40 588

    26 387

    11 892

    45.1%

    Tartu

    158 858

    97 970

    37 952

    38.7%

    Valga

    40 585

    26 271

    14 613

    55.6%

    Viljandi

    64 885

    44 823

    19 751

    44.1%

    Võru

    44 955

    29 564

    14 734

    49.8%

             

    TOWNS

    Tallinn

    452 840

    253 289

    151 210

    59.7%

    Narva

    80 491

    31 478

    20 931

    66.5%

    Kohtla-Järve

    57 287

    23 203

    15 097

    65.1%

    Pärnu

    52 085

    31 136

    14 367

    46.1%

    Tartu

    109 133

    64 012

    21 859

    34.2%

    TOTAL

    1 526 216

    880 296

    463 443

    52.6%

    Source: Kohaliku omavalitsuse volikogu valimised 17 oktoober 1993. Dokumente ja materjale. 1994: 41-46

    In national elections in September 1992 voted 68% of those eligible to vote; in Lithuania in November 1992 voted 75% and in Latvia in June 1993 voted 89% (Alatalu 1993).The local elections turnout percentage is not 'alarming', but on the other hand these were the first local elections in re-independent Estonia.

    The role of local government in Estonia before 1993 can be characterised shortly: "/.../ for many people the very existence of local government in their municipality or county has been unnoticeable" (Rosimannus 1993).

    During 1993, the governing coalition undertook major changes in the area of local governments. Riigikogu, Estonian parliament, adopted a number of laws on local government. The main aim of these changes was to change the two level structure of local government into unilevel local government in Estonia. A former centralised system of regional government was to be replaced by stronger and more independent local authorities. 'More independent' in a sense of increased financial independence -- a chronic lack of money was the main problem. Another feature that increases the importance of local governments for the public is that local governments were planned to become central institutions of regional development ('Isamaa' 1994: 23-24; Mäeltsemees 1994: 78). The changes were planned to start with newly elected local authorities.

    According to the Election Law people vote for a particular person; paradoxically the law strongly unfavours single candidates. Although the current governing coalition ran their own lists in many districts, characteristic to competing lists was that they were formed mainly on local basis. The opposition preferred to support local initiative and submitted their candidates into various local electoral unions (Savisaar 1993). The coalition blamed the opposition for avoiding political responsibility (Laar 1993), but one feature of the party system in Estonia was its limitedness mainly to urban areas, making it complicated to run separate lists elsewhere but in the capital. During the election campaign, the opposition attempted to raise the argument that the act of voting in local elections was an act of expressing unsatisfaction with the coalition performance; such reasoning was well received by pensioners and the Russian population (Lukas 1993). The governing coalition also viewed elections as an indicator of their performance (Laar 1993).

    The aim of this paper is to offer insight into public passiveness. The analysis is mainly based on the data from the public opinion polls that were conducted by the Estonian Market and Opinion Research Centre, Ltd., EMOR, in October 1993, right after the elections. The study uses 1438 interviews with Estonians and Russians. The data is weighted to approximate the sample to whole population on the dimensions of sex and age. Along with the public opinion polls also some materials from national daily newspapers have been used.

     

  3. Theoretical basis.
    1. The basis model
    2. Explanations of political behaviour have taken into account both environmental and personal factors (Milbrath & Goel 1977: ch. 3-5; Verba, Nie, Kim 1978: 80-93; Kaase 1990: 26). The basis model is shown on Figure 1. Here the point of departure is that attitudes affect participation, and attitudes in turn are affected by exogenous characteristics of a person and by characteristics of an environment. While exogenous characteristics and environment have also direct effect on participation, both attitudes and socio-demographic characteristics of a person are affected by ethnicity.

      Continuos lines present relationships that are of interest in this paper. Relationships presented with dotted lines are also important regarding voting activity but are not of primary interest now.

       

       

       

       

       

       

       

       

       

       

       

       

       

       

       

       

       

    3. Voting
    4. Voting has been treated as a form of conventional political activity (Milbrath & Goel 1977: 11; Kaase & Marsh 1979: 42; Conway 1991: 1; Dennis 1991: 23; Crotty 1991: 1; Martinussen 1977: 29). Moreover, for analytical purposes voting has been considered as a convenient indicator of participation: "Voting has some special advantages for generalising about the political aspects of human behaviour. People who vote are many and diverse." (Dennis 1991: 23). However, voting has been recognised as a fairly particular form of activity. Marsh and Kaase have written that voting is "a unique form of political behaviour in the sense that it occurs only rarely, is highly biased by strong mechanisms of social control, is accompanied by rain-dance ritual of campaigning and does not involve the voter in major informational or other costs" (Marsh & Kaase 1979a: 86). Voting has also been recognised as an act "by which the citizen affirms his/her loyalty to the system rather than an act by which he makes demands to the political system" (Milbrath & Goel 1977: 12).

      Voting in Estonia does not involve any major extra costs. No personal registration is required, all residents of a municipality or a county are automatically registered to vote. In local elections 1993, all Estonian citizens who were registered as residents of a municipality or a county in 01. 01. 1993 were eligible to vote, as well as those non-citizens who were permanent residents of a municipality or a county -- had resided there at least 5 years by 01. 01. 1993. Automatic registration procedures caused some confusion: some people did not appear in the lists of voters although they claimed to be eligible to vote. This problem touched mainly non-citizens and the number of persons who could not vote was several hundred (Alatalu 1993). Since polling stations are located mostly at local schools, access to them is fairly easy. People who could not vote on the election day -- October 17 -- had an opportunity to vote in advance, from October 13 to 16.

      In the next section I point out some expectations regarding the relevance of ethnicity and other factors.

       

    5. Socio-demographic determinants of turnout
      1. Ethnicity
      2. Since the establishment of Estonian citizenship, the Russian population had been largely excluded from national level policy and political processes. During the years of Estonian re-independence Russians in Estonia experienced the shift from the position of a somewhat privileged category in the society to the position of ethnic minority. The elections analysed in this paper were the first opportunity to regain some influence over policy processes. These considerations suggest that Russians were strongly motivated to take part in the elections and that their participation rate would be higher than that of Estonians.

         

      3. Life-cycle argument
      4. Although middle-aged persons appear to participate most actively (Kaase & Marsh 1979b: 111) it can be argued that this is only because of the effects of other social characteristics. When these are controlled voter turnout increases with age (Conway 1991: 17). Here the 'life-cycle effects' theory offers an explanation of why young people tend to be politically passive. As individuals grow older they start to work and become taxpayers, they marry and have children (which involves the need for larger residences and child care facilities). These concerns of daily life are to a considerable extent dependent on governmental policies, and this tends to stimulate awareness of politics and political activity. During their retirement years people are more free from family raising and occupational duties; free time combined with increased involvement in community affairs facilitates participation and voting. Although different life-cycle experiences among men and women have served to establish a gender gap regarding participation in politics, for electoral participation this gap has diminished to a large extent in the last years. These considerations suggest that voting activity is higher among older and married people (Pettersen & Rose 1995).

        I consider age, marital status, and sex as indicators of life experience.

         

      5. The socio-economic status argument
      6. One may argue that the social structure of the former Soviet Union country did not correspond to that of developed industrial countries and thus it is questionable to apply the approach that relates political participation to persons' socio-economic status (SES). In Estonia, all main strata included in the concept of SES are represented: clearly distinguishable upper class (~4%), the middle class (~25%) and the lower class (~55%) (Einasto 1994: 77). Although their proportions do not exactly correspond to that of industrial countries and the occupational structure differs from the income structure (Einasto 1994: 81-89) I consider the presence of the main strata as a sufficient justification for incorporating the concept of SES into the study.

        A number of studies confirm the hypothesis that participation, voting, and socio-economic status (SES) are strongly correlated. The common explanation of this tendency is that political participation generally requires political resources and that those with higher socio-economic status can more readily afford such investment (Pettersen and Rose 1995). Voting in Estonia clearly does not require any resource except time wasted on casting a ballot. Here I argue that it is reasonable to include this dimension into analysis since consequences of an election and decision whether to vote or not are relevant in the context of SES.

        I chose educational attainment, personal income, income per family member, the possession of various consumer and property items, occupation, and self-reported social class to indicate person's placement in the socio-economic structure.

         

      7. The size and character of a settlement

      Here two contradicting models posit the relationship between participation and type of settlement. In the mobilisation model the key variable is the stimulation that comes from an environment: "/.../ exposure to more communications, interaction with others involved in politics, support from peers for such activity, and the development of personality traits compatible with political activity" (Verba & Nie 1972: 231). Since increased stimulation is one of the concomitants of urbanisation higher voting activity can be expected in urban areas.

      The Decline-of-Community Model predicts a decline in participation as one moves from small town or village to city. In the larger units, politics is more complicated, impersonal and distant. Individuals are less tied to political, economic and other institutions, and borders of a local community become more ambiguous. For increased complexity and reduced personality reasons the attention of individuals becomes more diffuse and participation loses its meaning (ibid.). According to this model higher voting activity can be expected in rural areas.

      The argument that the size of a community has little explanatory power for participation analysis is based on the assumption that all local governments operate according to fairly standardised routines, independent of size and other characteristics of a community (Pettersen & Rose 1995). In the case of Estonia where local governments underwent major changes this argument perhaps does not have that much relevance as for political systems where local governments have clearly established their role. Of course, the analysis of the data tells what is the role of the type of a settlement.

      The variable that I have used is the type of a settlement.

       

    6. Political attitudes
    7. To select the attitudes I have partly relied on a political system support approach (Farah, Barnes & Heunks 1979: 432). In the next section I present different aspects of the approach that includes political efficacy, trust and satisfaction.

       

      1. Political efficacy
      2. The sense of political efficacy includes beliefs that one can understand politics and that ordinary people can influence policy making (Conway 1991: 33). One may think of the same attitude as of "/.../ the feeling that individual political action does have, or can have, an impact upon the political process /.../" (Abramson 1983: 135). Another important attitude is that toward responsiveness of officials and governmental institutions. This attitude is an outcome of a performance of political actors regarding interests of ordinary people or 'people like oneself' (Conway 1991: 33.). A number of empirical investigations confirm that the beliefs in the responsiveness of a system and self-confidence in political matters facilitate voting (Abramson 1983: 298).

        I have used four questions to measure both aspects of political efficacy.

         

      3. Trust and satisfaction
      4. It is natural to assume that individuals evaluate various components and outcomes of a political system and possess feelings toward various parts of a political system. A distrusting citizen may hold an attitude that certain parts of the political system need to be rearranged while a trusting citizen does not feel like rearrangements are necessary. It is not clear toward what parts of a system -- incumbents, institutions or an entire regime -- trust or distrust is pointed; this debate, however, rages over the meaning and interpretation of certain questions (cf. Abramson 1983: 193-206). These questions are concerned with honesty and competence of governmental officials in the area of allocating public resources.

        Another set of attitudes derives from the discrepancy between legitimate expectations and what an individual actually receives. Evaluations of responses of government and other political actors to policy demands bring about feelings of policy dissatisfaction or satisfaction. Satisfaction has two ingredients -- diffuse and specific. The diffuse aspect refers more to evaluations of a regime, and the specific aspect is largely a response to events and performance specific institutions (cf. Farah, Barnes & Heunks 1979: 381-444). Empirical evidence suggests that the more one is satisfied, the more one participates (ibid.: 435).

        In the present study I have included satisfaction with the state of affairs on the national level, satisfaction with and trust in national government. Regarding the county level satisfaction with and trust in local government are included. Also general trust in political parties is included.

        Supportive persons possess relatively high levels of trust, satisfaction and self-confidence -- they are psychologically more involved -- and this brings about increased participation (including increased voting activity) -- they are also behaviourally more involved.

         

      5. Party sympathy
      6. Along with feelings of trust and self-confidence I found it reasonable to include also a variable "that measures an individual's sense of attachment to a political reference group" (Abramson 1983: 71). In the United States the concept of party identification refers to tendency to vote for the party of which one reports to be member of and to support its incumbency (Niemi & Weisberg 1993: 214-216). The same logic must not necessarily hold in other countries (for example Netherlands (Thomassen 1993)) but what can be expected is that those with stronger positive feelings toward some (or one) of competing parties vote more likely than those who lack feelings of sentiment.

        I have used a variable that reveals whether an individual feels him- or herself attached to some political party or movement.

         

      7. Interest in politics

    Approaches to political interest rely on the assumption that interest is a measure of the degree of motivation for political participation. Thus is interest a motivational 'interlink' between attitudes and behaviour; actually are levels of participation and levels of political interest strongly related (van Deth 1990: 277).

    Interest is affected by a variety of factors but the basis premise for interest to occur is a person's ability to understand what is going on in the political sphere. Otherwise, if events and developments carry little or no meaning for a person they are also irrelevant for the person and the person can not be expected to develop high levels of interest (ibid.: 278-281).

    I have used individuals' self-placement on the interest scale, or the measure of subjective political interest to indicate the levels of individuals' interest in politics.

     

  4. Empirical analysis
  5. The purpose of data analysis is to establish direct effects of variables on voting activity. For data analysis I employed ordinary least squares (OLS) multivariate linear regression. OLS regression technique has been originally developed as a technique to analyse metric data. However, in practical situations regression with nominal data is performed successfully. More serious objection to the OLS technique rises from the dichotomous nature of the dependent variable. In the case of OLS technique unstandardised regression coefficients may be interpreted as probabilities, but these may rise above 1 and fall below 0. Such illogical outcomes have no substantial interpretation.

    Here the primary goal lies in establishing direct effects of the variables, not in predicting outcomes. Results may be faulty while observing change in the dependent variable as a result of unit change in an independent variable, or while observing probability that individuals might belong to a category of voters given a certain combination of values of the independent variables. The results are correct in the sense that they signal about presence of significant relationship.

    To establish direct effects I have employed blockwise regression; the blocks are comprised of attitudinal, socio-demographic and ethnicity indicators respectively. The regression is run twice for each block: once only the variables of a certain block included and then once again with all variables included. By comparing the results from the two regressions it is possible to get additional conformation to the existence of the effects that might occur in ‘bivariate’ -- when only the variables of a certain block are involved -- relationship. If the coefficients of the variables in a certain block remain stable after other variables have been introduced then direct effects of the variables in the block obviously exist. If the effects of the variables included in a certain block drop (and perhaps become insignificant) when all variables are included then the turnout is affected primarily by later introduced characteristics, which in turn are affected by the former variables. Building on the aforementioned arguments, Pettersen and Rose used roughly similar technique (Pettersen & Rose 1995).

     

    1. Findings

    Table 2 presents empirical findings. The stability of the regression coefficients suggests that three socio-demographic characteristics did have direct impact on voting activity. Out of a variety of attitudes only one is of importance with respect to the turnout.

     

    Table 2. Determinants of voting activity (N = 1438)

    Ethnicity

    Socio-demographic

    Attitudes

    All

    Ethnicity

    .07*

    .00

    Gender

    -.04

    -.06

    Age

    .14*

    .12*

    Education

    .04*

    .03

    Type of a settlement1:

    medium-size settlements

    .03

    .03

    rural settlements

    .09*

    .09*

    Time lived in a community

    .04*

    .04

    Personal income

    .02

    .02

    Income per household member

    -.01

    -.02

    Marital status

    .01

    .00

    Number of consumer and property items

    .02

    .01

    Occupation2:

    mid-rank position

    .10*

    .11*

    high position

    .13

    .12

    no job

    .02

    .04

    Self-reported social class

    .04

    .04

    Satisfaction with the state of affairs, national level

    .02

    .04

    Satisfaction with Government

    .04

    .03

    Satisfaction with local government

    .00

    .00

    Trust in Government

    .05

    .04

    Trust in local government

    .05

    .04

    Local identity preference3:

    weak

    .08*

    .05

    medium

    .06

    .05

    strong local identity preference

    .04

    .04

    Trust in political parties

    .01

    .01

    Party attachment

    .06

    .06

    Interest in politics

    .07*

    .04

    Politics is complicated

    .01

    .00

    People like me have no say

    -.02

    -.02

    Voting is the only opportunity

    .13*

    .12*

    Politicians don't care about us ordinary people

    -.01

    -.02

    Adj. R2

    .00

    .09

    .08

    .16

    Table entries are unstandardised regression coefficients.

    1, 2, 3 -- these variables were recoded into sets of dummy variables, and their respective reference categories are following:

    1 -- the capital or a big town

    2 -- low-rank job

    3 -- no local identity preference

    Level of significance * = .05

    Adj. R2 is the adjusted R2 -- the coefficient of determination adjusted for the number of independent variables in the equation

    Amongst several findings two are most interesting. At first, ethnicity did not have considerable effect on the turnout. The effect that occurred in bivariate relationship faded away when other variables were introduced. This result suggests that ethnicity does affect socio-demographical and attitudinal characteristics of a person, and has indirect effect on the turnout, but it does not influence voting activity directly. The second important finding is that attitudes had virtually no impact on the turnout. This result is more important from the theoretical point of view -- for analysts of voting behaviour the domain of attitudinal variables has been the key to explain the results of an election.

    Which variables discriminated most strongly between voters and non-voters? Here four indicators stand out significantly -- age, occupation, the type of a settlement and an opinion related to elections. Casting a ballot was more common among older people, higher occupational position, those living in smaller settlements and with stronger belief that elections were the only way to affect social life. The insignificance of the coefficient of the high ranking job group is probably due to a relatively small number of cases in the category. Four socio-demographic dimensions that have been found important in various countries and times had weak or virtually no impact on the turnout -- these are gender, education, income and marital status.

    Except for one, none of the attitudes affected significantly the turnout. The opinion about elections was related to voting activity in a very natural way -- those with stronger belief in the power of elections were more likely to vote. Opinions reflecting satisfaction with various parts of the political system did not have impact on behaviour.

    One may find traces of influence of two attitudinal dimensions. At first, if there was difference in the turnout then people having more confidence in either national or local governments were more active. And second, the turnout was slightly higher in the group comprised of people possessing stronger sense of political involvement.

     

  6. Discussion
  7. The finding that ethnicity had no influence on the turnout reflects the situation where differences in social position override differences in ethnicities. Most of the Russian population had lived in Estonia already for years or tens of years, and they were not more troubled with regular everyday concerns than Estonians. Due to liberal policies of the Government and the legislative body, Russians did not major specific problems. Both Russians and Estonians faced very similar problems in everyday life: worsening economic performance of the state hit equally both Estonians and Russians. Thus, both groups were largely in the same situation.

    The minor role of attitudes can be explained by the two following reasons. The basic reason is that political attitudes were relatively poorly developed. First free elections in re-independent Estonia were held one year before -- clearly solid attitudes could not be developed either toward parties which were virtually their first term in Parliament or toward Government. People simply lacked reference points like consequences of certain policies or performance of certain political actors to be able to locate clearly political figures. Local governments were perfect examples in this sense -- in the year of elections their role in Estonia was completely reconsidered. From this perspective people could have only weak and confused attitudes that were of little help for a voting decision.

    Another reason is that due to irrelevance of local governments these weak attitudes were not sufficiently actualised (partly also because of their weakness) and thus they could not play decisive role in voting activity. The relevance of the second reason is also based on the irrelevance of the satisfaction with the state of affairs on national level -- this opinion does not presume clear location of various parts of the political system.

    The elections actualised only strongly held beliefs about elections as the mean of influencing social life -- a rather concrete opinion. Causal link between this opinion and voting is probably rather two-way than one-way relationship. The opinion certainly facilitated voting, but since the poll was conducted after the elections, one should take into account the possibility that the opinion itself gained strength as a result of the voting.

    What stands behind the importance of age, occupation and settlement type? Since several possible explanations are rejected by the data, no clear answer is given. If one is willing to accept the SES argument then the impact of the incomes and the effect of possession of various items is expected to occur, but this was not the case. Moreover, how to explain the activity of older people who are mostly pensioners and do not rate high on the SES dimension? If one is willing to accept the argument that people in higher occupational positions possess more developed mental skills, then the effect of education is expected to occur, but this was not the case. If one is willing to accept the argument that older people, people in rural settlements, and these in higher occupational positions are more involved in the life of local communities, then they are expected to have developed also local identity preference, but that was not the case.

     

  8. Summary
  9. The paper considered voters’ turnout at the first local elections in Estonia, the primary aim was to establish the main determinants of the turnout. For this purpose a wide variety of appropriate theoretical arguments were employed.

    The results suggest that the voting activity was largely determined by respondents’ age and occupation, as well as by the type of a settlement, where they lived. The data did not allow to reveal the logic behind the importance of these variables, several possible explanations did not find sufficient conformation. Neither ethnic background nor attitudes in general had strong effect on the turnout. Several dimensions that have been found important in different countries and times -- income, education, marital status, and gender -- did not have substantial effect on the turnout.

  10. References

Abramson, Paul R. Political Attitudes in America. Formation and Change. San Francisco: W. H. Freeman and Company, 1983.

Alatalu, Toomas. Valimiste võti on Imavere imes /The Key to the Elections is in Imavere Miracle -- in Estonian/. "Päevaleht", 16. 11. 1993.

Conway, Margaret M. Political Participation in the United States. Washington: A Division of Congressional Quarterly Inc., 1991.

Crotty, William. Political Participation: Mapping the Terrain. In: W. Crotty /ed./. Political Participation and American Democracy. Greenwood: Greenwood Press, 1991, pp. 1-22.

Dennis, Jack. Theories of Turnout: An Empirical Comparison of Alienationist and Rationalist Perspectives. In: W. Crotty /ed./. Political Participation and American Democracy. Greenwood: Greenwood Press, 1991, pp. 23-66.

Einasto, Mart. Majandusliku kihistumise mõõtmine Eestis /The Measurement of Economic Stratification in Estonia -- in Estonian/. unpublished master thesis University of Tartu, 1994.

Farah, Barbara G., Samuel H. Barnes & Felix Heunks. Political Dissatisfation. In: Samuel H. Barnes and Max Kaase /eds/. Political Action. Mass Participation in Five Western Democracies. Beverly Hills: Sage, 1979, pp. 409-448.

"Isamaa". Kaks aastat valitsust /Two Years in Government -- in Estonian/. Tallinn, 1994.

Kaase, Max & Marsh, Alan a. Measuring Political Action. In: Samuel H. Barnes and Max Kaase /eds./. Political Action. Mass Participation in Five Western Democracies. Beverly Hills: Sage, 1979, pp. 57-96.

Kaase, Max & Marsh, Alan b. "Background of Political Action" In: Samuel H. Barnes and Max Kaase /eds./. Political Action. Mass Participation in Five Western Democracies. Beverly Hills: Sage, 1979, pp. 97-136.

Kaase, Max. Mass Participation. In: Jennings, Kent M. and van Deth, Jan W. /eds./. Continuities of Political Action. A Longitudinal Study of Political Orientations in Three Western Democracies. Berlin: Walter de Gruyter, 1990, pp. 23-66.

Kohaliku omavalitsuse volikogu valimised 17 Oktoober 1993. Dokumente ja materjale. koostajad Alo Heinsalu ja Arvi Lundver /Local Elections in October 17, 1993. Documents and Materials. -- in Estonian. Compiled by Alo Heinsalu and Arvi Lundver/. Tallinn: Huma, 1994.

Laar, Mart. Opositsioon, koalitsioon ja kohalikud valimised /The Opposition, the Coalition and the Local Elections -- in Estonian/. "Postimees", 08. 03. 1993.

Lukas, Tõnis. Tartlasi ei huvita tulevik /Inhabitants of Tartu Are Not Interested in Their Future -- in Estonian/. "Postimees", 25. 10. 1993.

Martinussen, Willy. The Distant Democracy. Social Inequality, Political Resources and Political Influence in Norway. London: John Wiley and Sons, 1977.

Milbrath, Lester W. & Goel, M. L. Political Participation Chicago: Rand McNally, 1977.

Mäeltsemees, Sulev. Kohalikud omavalitsused Eestis. Põhiinformatsioon. Kogumikus: Kohalikud omavalitsused Kesk- ja Ida-Euroopas ning SRÜ-s 1994 a. /Local Governments in Estonia. In: Local Governments in Central and Eastern Europe and in CIS in 1994 -- in Estonian/. Budapest: The Institute of Local Government and Public Service, 1994, pp. 73-84.

Niemi, Richard G. & Weisberg, Herbert F. How Meaningful is Party Identification? In: Richard G. Niemi & Herbert F. Weisberg /eds./. Classics in Voting Behavior. Washington: A Division of Congressional Quarterly Inc., 1993, pp. 210-223.

Pettersen, Per Arnt & Rose, Lawrence E. Participation in Local Politics in Norway: Some Do, Some Don't; Some Will, Some Won't. 1995 /manuscript of the forthcoming article/

Rosimannus, Rain. Kas kõik on nii ilmselge? /Is Everything So Evident? -- in Estonian/. "Rahva Hääl", 02. 11. 1993.

Savisaar, Edgar. Mida näitasid 17. oktoobri valimised? /What Did the 17 October Elections Demonstrate? -- in Estonian/. "Rahva Hääl", 30. 10. 1993.

Thomassen, Jacques. Party Identification as a Cross National Concept: Its Meaning in the Netherlands. In: Richard G. Niemi & Herbert F. Weisberg /eds./. Classics in Voting Behavior. Washington: A Division of Congressional Quarterly Inc., 1993, pp. 263-266.

van Deth, Jan W. Interest in Politics. In: Jennings, Kent M. and van Deth, Jan W. /eds./. Continuities of Political Action. A Longitudinal Study of Political Orientations in Three Western Democracies. Berlin: Walter de Gruyter, 1990, pp. 275-312.

Verba, Sidney & Nie, Norman H. Participation in America. Political Democracy and Social Equality. New York: Harper and Row, 1972.

Verba, Sidney, Nie, Norman H. & Kim, Jae-on. Participation and Political Equality. A Seven-Nation Comparison. Cambridge: Cambridge University Press, 1978.