Content analysis in Lithuania: the state of the art

1 Draft only Content analysis in Lithuania: the state of the art Algis Krupavicius, Associate Professor of Kaunas University of Technology and Vytautas Magnus University E-mail: akr239@kaunas.omnitel.net Paper for a delivery at the 27 th Joint Sessions of Workshops of the European Consortium for Political Research, Workshop Estimating the Policy Positions of Political Actors, Mannheim, March 26-31, 1999. Abstract Content analysis in political science of Lithuania is very new method of investigation. In this introductory paper on a status of content analysis in the country are discussed only basic questions. The first part of paper describes issues and problems of introduction of content analysis. Among them are as following: How ideas and methods of content analysis came to Lithuania s universities? How to solve a beginner s dilemma in content analysis: to imitate or to improve old schemes? The second part stops on research projects which are already done, or close to be finalized using content analysis technique. In particular on party manifests research, or to what extent Western variables can be applied in new democracy. Political symbols and political discourse in a mirror of content analysis. The third part of paper discuss a change from manual expert coding to computerized content research, and opportunities of new research projects. However, the aim of this paper is not to present results of content analysis extensively rather to make an overview of efforts, problems and some achievements in the field in Lithuania through very individual experience and personal reflections.

2 I. The state of affairs in the field of content analysis in Lithuania cannot be separated from a general context of development of political science as new academic discipline and research area since the late 80-s. Political science still is very young academic subject and research area in Lithuania. In 1999 Lithuanian political scientists will celebrate only the 10 th anniversary of introduction of political science into universities curricula. Of course, we can find some historical roots of Lithuanian political research in the inter-war period in 20-s and, especially, 30-s, namely at the universities of Vilnius and Kaunas. Even during the years of the Soviet domination, there were some opportunities of less ideological political research, especially studying international relations and politics of foreign countries 1. Unfortunately, all historical reflections cannot enlighten enough problems of content analysis in the country, especially, that the content analysis is a relatively new analytical tool even in Western political science. Many scholars from Eastern and Central Europe might tell very individual stories of their first touch with quantitative and qualitative methods in political research. Of course, some similarities are existing in our experiences and our career trajectories to become involved in such sort of political investigation. I hope that one day we will have an opportunity to compare our entry into the content analysis in detail. However, few colleagues who are involved in content analysis in my home country would tell quite different stories at the moment. For sure, most of us were forced to enter this unknown field of knowledge mines simply by circumstances. In my case, I was asked in early 1993 to start to teach methods of political research at the Vytautas Magnus University for undergraduate students in political science. Of course, even at that time it was clear that no way to develop qualified and substantive studies in political science without teaching political analysis, or methods of political research. Another problem is related to myself as a representative of founders generation in the discipline that most of us were self-educated and entered into political science through other very different subjects as law, history, sociology etc. As usual the initial understanding of research methods and approaches in political science was brought by every newcomer to the field from their native discipline. As far as many of newly born political scientists came from humanities but not exact sciences, the research perspective that analysis in political science is a descriptive and comparative one in best case was very common. Only few of us, who has had one or another contact with application of the aggregate social or individual level sociological data in previous studies, have had some but very general knowledge and understanding about a very existence of quantitative and qualitative methods. Few years ago I asked Hans-Dieter Klingemann (WZB), why he developed good research connections predominantly with sociologists in Eastern Europe in early 90-s, instead of establishing fruitful research relations with political scientists. He told me that if most of among would-be political 1 About development of political science in Lithuania see: Krupavicius A. (1997) The development of political science in Lithuania: Years of the breakthrough. European Journal of Political Research, 31: 499-517.

3 scientists from Eastern Europe were unable to understand his professional language adequately, the absolute majority of sociologists from the region did so and they were familiar with a discourse of Western sociology, as well spoke the same professional language. Looking back, I need to agree with Hans-Dieter Klingemann: yes, it was so. But the explanation of the situation is rather simple: the sociology was established academic subject since the 60-s, as political science only started its existence in the late 80-s. Fortunately, that we were rather young (many of us only in the mid- twenties) at the moment of introduction of political science in Lithuania, and we were able to learn a lot by trying and doing research in politics. The first step of every beginner was to educated himself in basic ideas and methods of political research through taking and reading all the available Western text-books from almost empty shelves in political methodology literature at the university library. At the same time I was introduced with very general understanding of content analysis. However, till now I wonder that we still lack a good text-book to introduce content analysis to our students, and to explain its relationship with practical political research worldwide. I should be very happy to know that I am wrong saying this. Meanwhile, a general interest of Lithuanian political researchers to content analysis and understanding of it, as foremost, as a study of written political texts was and is rather low. Looking historically, nevertheless, we could find some examples of political text research as early as in the 30-s. Despite, that majority of such studies were focussed exclusively on style of political texts, and namely on style of A. Smetona, the autocratic president of Lithuania in 1926-1940, writings 2. Till the mid-nineties in content analysis of political issues specialists in linguistics, especially, from the University of Vilnius were ahead compared to political scientists. L. Bielinis has mentioned as one of most interesting studies 3, a study done by E. Lassan, a linguist from the Vilnius University, who analyzed oppositions of collectivity and individualism 4 in the discourses of the official authorities and dissent movement during the Soviet period. The fact that researchers from very different academic fields are entering into content analysis might be confirmed by a recent study on contents of crime information in Lithuanian newspapers done A. Dobryninas (University of Vilnius), who is by education a philosopher. 2 See: Salys A. (1931) Smetona mus rašto moko (pastabos apie jo raštų kalbą).- Jaunoji Karta, Nr.23.; Ambrazevièius J. (1934) Stiliaus kultūra. A. Smetonos stilius.- Židiinys, Nr.8-9.; Evenšteinas M. (1934) Dvasinis Antano Smetonos veidas jo raštų šviesoje. In: Antanas Smetona. Kaunas; Alantas V. (1931) Antano Smetonos raštų stilius.- Vairas, Nr. 10.,pp.184-194. 3 Bielinis L. (1996) Kalbejimo strategijos politiniame tekste (Lietuvos Respublikos Prezidento kalbu pavyzdziu). Vilnius: TSPMI, pp. 9-10. 4 Lassan E. (1995).Valdžios ir disidentijos diskursas Tarybų Sąjungoje: kognityvinė-retorinė analizė. (in russian). Vilnius: Vilniaus Universitetas.

4 On the other hand, rather early interest into content analysis by some Lithuanian linguists and philosophers was not an accidental event keeping in mind that Professor Algirdas Greimas was worldwide known specialist in semiotics, who spent most of his life teaching and studying semiotics in Paris. Fortunately, he established contacts with Lithuanian philosophers and linguists during the late Soviet period, and encouraged them to study this relatively new academic field. Among those, who did an extensive study in philosophy of language in the 70-s and 80-s was Rolandas Pavilionis, current President of the Vilnius University. Of course, semiotics is far enough from the methodology of content analysis in contemporary political science, but both deal with meanings, discourse or text structure. But my first serious example of how to use content analysis came with papers and materials on the ECPR Party Manifests project in late 1994. It was a source from which I am taking ideas till now. As far as I am convinced fundamentally in Aristotelian motto that we are not only homo sapiens but also homo politicus, I believe that political systems, in spite, of their differences could be and are comparable, I always perceived a problem of adaptability of external instruments and methodologies of scientific investigation as a minor issue. Of course, it does not mean that mechanical imitation, or methodological and technological transplants of investigation will work efficiently in every new environment without knowledge about specifics of that environment. A real problem is only to find or to develop suitable operational tools of investigation, and every coherent and systemic methodology of research can reveal one or another aspect of reality with great degree of reliability. Briefly speaking, this is an argument for a constructive application of content analysis schemes in new democracies which originally were born in the West. II. One of the first attempts to apply content analysis exceptionally through a prism of political science was a Ph D thesis presented by Lauras Bielinis (Institute of International Relations and Political Science, University of Vilnius) on Speech strategies in the political text (On an example of speeches by A. Brazauskas, the President of Lithuanian Republic) in 1996. Main imperative of L. Bielinis to focus his research on political text and political speech strategies was based on understanding that political text might be named into an instrument of political power by which politicians make influence on a given society: through structuring of its political behavior, mobilizing similarly minded people, implementing political decisions, organizing ideological structures 5. Moreover, according to L. Bielinis, the political text always reflect existing power relations [ ], but because it is a product of open political space, all the political realities, environment in general, and, especially, its social and psychological context are influencing making and presentation of political text 6. 5 Bielinis L. (1996) Kalbejimo strategijos politiniame tekste (Lietuvos Respublikos Prezidento kalbu pavyzdziu). Vilnius: TSPMI, p.3. 6 Bielinis L. (1996) Kalbejimo strategijos politiniame tekste (Lietuvos Respublikos Prezidento kalbu pavyzdziu). Vilnius: TSPMI, p. 4.

5 From instrumental point, for the analysis Bielinis selected 29 texts of A. Brazauskas, the President of Lithuania (1993-98), from a period of 1993-1995. Methodologically he based himself heavily on a tradition and instruments of research of so-called pragmatics school 7. All the presidential speeches were classified along the lines of 12 speaking strategies: generalization, using of illustration (example), correction, reinforcement, softening, compromise, repetition, contrast, change, avoiding, intimacy and appeal as means to increase role and improve political positions of the president 8. However, most successful part of Bielinis study was related to analysis of pronouns distances in the speeches of A. Brazauskas. All the pronouns (I, We, You, Yours, He etc.) used in Brazauskas speeches were classified along their relationship with five value orientations (as described by the author): neutrality, belief (trust), promise, obligation and responsibility 9. Main conclusion of Bielinis analysis was that A. Brazauskas, as the President, preferred to be active and not a politically neutral political leader in office 10. At the end of 1997 L. Bielinis initiated and implemented a short study of programs of seven presidential candidates in the President elections of 1997-1998. As a coding unit here was selected a paragraph. The focus of this project was to explore a strength of democratic orientations of presidential candidates, and whether individual candidate s program conforms the ideology he declares. The coding, or rather text search, was based on looking for 15 so-called basic categories: Freedom; Equality; Citizen; Society; Community; Man; Nation; State; Rights; Duties; Elections; Constitution; Law; Representation; 7 See: Maitland K., Wilson J. (1987), Ideological Conflict and Pronominal Resolution. Journal of Pragmatics,. Nr.11, pp. 495-512; Wilson J. Politically speaking. N.Y., 1990; Leech G. Principles of Pragmatics. London., 1983; Levinson S. Pragmatics. Cambridge, 1983; Verchueren J. Norwood, 1985; London. What People Say They Do with Words. Horn, Pragmatics Theory. In: Newmeyer F. (ed.) Survey of Lingvistics, Vol. 1.Cambridge, 1988; Green G. M. Pragmatics and Natural Language Understanding. Hillsdale, 1989. 8 Bielinis L. (1996) Kalbejimo strategijos politiniame tekste (Lietuvos Respublikos Prezidento kalbu pavyzdziu). Vilnius: TSPMI, p. 78. 9 Bielinis L. (1995) Ivardziai politiniame tekste ir politiko nuostatos. Politologija, vol. 6 (1), p. 81. 10 Bielinis L. (1995) Ivardziai politiniame tekste ir politiko nuostatos. Politologija, vol. 6 (1), p. 82.

6 Democracy 11. In spite that the project was interesting enough as the attempt to analyze presidential candidates programs, but its framework needed more of theoretical and methodological consistence and serious updating. Moreover, many findings of the project were irrelevant to other evaluations of presidential candidates, or even misleading in understanding of political positions of individual candidates. For example, in the Bielinis project an extension of state s role was identified as a main theme in election programs of Paulauskas and Adamkus, who were described respectively as the Centre-left and Centreright candidates by most Lithuanian and foreign experts 12. Of course, it is absolutely misleading conclusion. Another striking mistake was a conclusion that Landsbergis, as a candidate of the Conservative party, undermined rule of law in his electoral manifest 13. At least, such conclusion is not correct looking on all issues raised in the Landsbergis campaign for the President in 1997. These failures of the Bielinis project in the estimation of political orientations of the presidential candidates in 1997 reflect well- how systemic and objective methodological and technological framework of content analysis needs to be in order to get reliable and valid results. A different focus was chosen in developing of content analysis in Kaunas compared to Vilnius based scholars. The strategy here was based on a different approach- don t reinvent a wheel, and as a result party manifests research project was started at the Vytautas Magnus University (Kaunas) in 1996. A subject of the initial analysis were two programs declared by the Homeland Union (Lithuanian Conservatives), i.e. a successor party from the Right wing of Sajudis, in 1994-1996. A framework of analysis was taken from the ECPR Manifest Research Group. However, due to realities of the postcommunist situation the framework required some updating in order to reflect more precisely specifics of transitional society and newly emerged political parties. On the other hand, it was important not to go too far from an original design of research for comparative reasons as well as to escape from possible mistakes in coding procedures through using limited and approved number of codes. After prolonged training in coding procedures an initial enthusiasm of some undergraduate students disappeared and a final work was done by D. Mackevicius, who was a fourth- year student in political science at that time. Final framework of the HU(LC) manifesto analysis included only 5 new codes compared to the ECPR scheme, i.e. 205- the Soviet heritage; 306- Electoral mobilization; 417- Policy of taxes; 418- Investments (foreign and domestic) ; 419- Policy of energy 14. Speaking about some findings from this pilot analysis of the HU(LC) programs of 1994-1996, it is important to mention that was no a radical shift in political priorities of the 11 Bielinis L. (1998) Prezidento rinkimai 97: Kandidatu i Lietuvos Respublikos Prezidentus rinkiminiu nuostatu ir stereotipu lyginamoji analize.- Politologija, vol. 12 (2), p. 28. 12 Bielinis L. (1998) Prezidento rinkimai 97: Kandidatu i Lietuvos Respublikos Prezidentus rinkiminiu nuostatu ir stereotipu lyginamoji analize.- Politologija, vol. 12 (2), p. 29, 31. 13 Bielinis L. (1998) Prezidento rinkimai 97: Kandidatu i Lietuvos Respublikos Prezidentus rinkiminiu nuostatu ir stereotipu lyginamoji analize.- Politologija, vol. 12 (2), p. 33. 14 Mackevicius D. (1997) Tevynes Sajungos (Lietuvos Konservatoriai) programu turinio analize. Kaunas: VDU, p. 22.

7 party. However, the 1996 program was more oriented to economic policy issues, instead of putting emphasis on a functioning of political system, or issues of morality, national way of life, i.e. categories covered by a group of fabric of society. Table 1. Policy preferences of the HU(LC) in 1994-1996 (in percentages) 1994 1996 External relations 11 17 Freedom and democracy 5 6 Political system 12 4 Economy 29 38 Welfare and quality of life 16 14 Fabric of society 20 12 Social groups 7 8 Source: Mackevicius (1997), Tevynes Sajungos (Lietuvos Konservatoriai) programu turinio analize, p. 44. A new phase in Lithuanian party manifesto research has started in late 1997 and it is continued till now. Two major tasks were established for this particular phase: 1) to code and analyze the founding and parliamentary election programs of all major Lithuanian parties: the HU(LC), Social Democratic, Christian Democratic and Labour Democratic parties, as well as the Centre and Liberal Unions using manual expert coding; 2) to do the same work using NUD*IST computer program. As far as the expert coding of the HU(LC) programs in 1996 was rather an experimental one and covered only part of the Conservative party programs since 1993, all the texts of this party are coded again using the updated expert coding frame. As a supplementary part in this scheme, there was initiated a subproject, which covers electoral programs of presidential candidates in the elections of 1993 and 1997-1998. This time (since late 1997) expert coding scheme of party manifests research was substantially updated and total number of codes has reached 95 categories instead of the 56 original standard codes. Main problem was to limit ourselves with manageable and comparable number of categories to the ECPR framework. Due to these considerations all 7 domains of the ECPR coding frame were preserved and majority of new codes were introduced as subcategories to the existing codes (see: Table 2.). Still main problem of the expert coding is define correctly and in comparable way a quasi-sentence as main coding unit. In order to escape major coding errors is not enough to follow recommendations of the ECPR Manifest Research Group 15, but as well as it seems efficient enough to work in a group, when every expert checks and discusses coding results of own and his colleagues texts. Of course, it is time-consuming effort. Another and additional way to check validity of expert coding is by using and comparing the results with computer-based coding. A possibility to obtain some already coded texts from the ECPR Manifest Research Group also might be an efficient way to check correctness of coding results, especially, for those, who are relative newcomers in the field. Table 2. Full expert coding frame of Lithuanian political party programs (finalized in 1998) 15 See: Volkens A. (1992) Content Analysis of Party Programmes in Comparative Perspective: Handbook and Coding Instructions. Berlin: WZB.

8 EXTERNAL RELATIONS 101:Foreign Special Relationship :Positive 1011: Russia: Positive 4014: Control of Economy : Negative 402: Incentives 403: New Market Regulation 1012: Western States : Positive New 4031: Mixed (Social Market) Economy: Positive 1013: Baltic States : Positive 404: New Economic Planing 102: Foreign Special Relationship : Negative 1021: Russia : Negative 405: Corporatism 406: New Protectionism : Positive 1022: Western States : Negative New 1023: Baltic States : Negative 407: New Protectionism : Negative 103: Anti-Imperialism 1031: Russian Army : Negative 408: Economic Goals 409: New Keynesian Demand Management 1032: Independence : Positive 410: New Productivity 1033: Rights of Nations : Positive 411: New Technology and Infrastructure 104: Military : Positive 105: Military : Negative 412: Controlled Economy 4121: Social Ownership: Positive New 106: Peace 107: Internationalism : Positive 413: Nationalization 4131:Privatization : Negative New 1071: International Party Partnership New 108: European Community : Positive 109: Internationalism : Negative 110: European Community : Negative 111: National Security 4132: Property Restitution : Negative New 414: Economic Orthodoxy 415: Marxist Analysis 416: Anti-Growth Economy 417: Tax Policy New 418: Investments New FREEDOM AND DEMOCRACY 419: Energy sector New 201: Freedom and human rights 420:Full employment New 202: Democracy 421: Low inflation New 2021:Transition to Democracy New 2022: Communist : Positive WELFARE New AND QUALITY OF LIFE 2023: Communist : Negative 501: New Environment Protection 203: Constitutionalism : Positive 2031: Strong Presidentialism : Positive New 204: Constitutionalism : Negative 2041: Strong Presidentialism : Negative New POLITICAL SYSTEM 301: Decentralization 302: Centralization 502: Culture 5021: Private-public Mix in Culture 503: Social Justice 5031: Private-public Mix in Soc. Justice 504: Welfare state Expansion 5041: Private-public Mix in Welfare 505: Welfare State Limitation 506: Education Expansion 303: Governmental and Administrative Efficiency (including administrative 507: Education and state Limitation reform) 304: Political Corruption 305: Political Authority 3051: Public Situation : Negative FABRIC OF SOCIETY 601: New National Way of Life : Positive ECONOMY 401: Free Enterprise 4011: Privatization : Positive 602: National Way of Life : Negative 603: Traditional Morality (Christian values) : Positive 604: New Traditional Morality (Christian values) : Negative 4012: Privatization Vouchers : Positive New 4013:Property Restitution: Positive New 605: Law and Order 606: Social Harmony 6061: Political Coalitions New

9 607: Multiculturalism : Positive 608: Multiculturalism : Negative SOCIAL GROUPS 701: Labour Groups : Positive 702: Labour Groups : Negative 703: Agriculture and Farmers 704: Middle Class and Professional Groups 705: Underprivileged Minority Groups 7051: Minorities Abroad 706: Non-Economic Demographic Social Groups 7061: Youth 7062: Old People 7063: Women 7064: Ethnic minorities New 7065: Families with children New 707: Rehabilitation and Compensation New 000: no one of codes can be applied Full results of expert analysis of major Lithuanian party manifestos are expected to appear in late 1999 and early 2000, but still it will depend on financing opportunities of the project in 1999. Major technical trouble was to obtain electronic copies of all party programs. It became into very significant issue after starting preparations for the computer- based analysis of party programs in 1998. As an exemplary model of this analysis was chosen a frame developed by Michael Laver and John Garry in 1997 16. In particular a revised manifesto coding scheme and content analysis dictionary were taken as basic components into an initial frame in the NUD*IST program to analyze Lithuanian party programs. Despite that it was very much expected, the technical problems were most important at the start of using of the NUD*IST program in the analysis. First of all, we needed to find out how to convert some specific Lithuanian Latin signs as ė, č, ą, į, ž etc. into English Latin letters. Next question was about formatting of all the texts under analysis for NUD*IST4 program. Because even electronic party programs versions (if it was possible to get such copy at all without a re-typing of printed text) were made on different computers and, more important, in different styles of paragraphs and sentences. All this preparatory work became into highly time-consuming effort. Nevertheless, a challenge of efficient content analysis through the NUD*IST program was even greater. From substantive point of view, the revised manifesto coding scheme and content analysis dictionary developed from British and Irish party manifestos issued during the 1992 parliamentary elections by Laver and Garry needed very careful updating in order to reflect realities of Lithuania as a country of the consolidating democracy, which restored national independence in a parallel. Of course, political discourses, party rhetoric of advanced and consolidating democracies are substantially different. In this situation an adoption of the political dictionary in a one-to-one scale based on British and Irish party discourse into research of Lithuanian party programs is highly risky deal. On the other hand, mentioned differences are not so huge that it would be impossible to use a general study framework of advanced democracies in new democratic countries, especially, in those, whose values, habits and traditions were very much oriented to West European cultural influences during the centuries. Meanwhile, now there is a need to step aside from our mainstream analysis in party manifestos in order not only to introduce briefly another trend in content research which 16 Laver M., Garry J. (1997), Estimating Policy Positions from Party Manifestos. Paper prepared for the symposium Actors and Institutions in West European Parliamentary Democracies, Umea, June13-15.

10 is of direct importance but to some extent has had a guiding impact on party manifestos investigation by using the NUD*IST program. Almost by a mere accident in the mid-1997 I met Vaidas Morkevicius, a senior and very knowledgeable student, who expressed his interest to study Lithuanian political discourse using tools of content analysis. In a course of one academic year we were able to develop not only a coding frame of such analysis, but to prepare a final paper almost of one hundred pages and to formulate a separate research project on Lithuanian political discourse. A primary object of expert analysis was to investigate Lithuanian political symbols in printed media. According to Charles D. Elder and Roger W. Cobb, a symbol is any object used by human being to index meanings that are not inherent in, discernible from, the object itself 17. But political symbols might be defined as those symbols of relevance to the exercise of political authority and to the management of social conflict 18 As a general frame of political discourse research was based on a list of basic political subject proposed by David Easton in his study A Systems Analysis of Political Life (1965): political community, political regime, authorities of the government. However, as a direct operational basis of the political discourse analysis was chosen more empirically applicable typology of political symbols developed by Charles D. Elder and Roger W. Cobb structured into 3 groups: 1. Symbols of political community; 2. Regime norms, structures and roles; 3. Situational symbols based on : A. Current authorities; B. Nongovernmental political actors C. Policies and policy issues 19. As an empirical data basis of the political discourse study were selected only editorials of two Lithuanian dailies: Lietuvos Rytas and Lietuvos Aidas. Both dailies have had a nation-wide circulation, but there were some additional factors driving a choice of mentioned papers. Lietuvos Rytas was a daily with widest circulation in the country, as well as from an ideological viewpoint it was rather liberal and most objective paper in Lithuania in 1993-1997. On the other hand, Lietuvos Aidas was re-established by Supreme Council - the Constituent Seimas in 1990 in order to provide an information hand for the Right-wing political leadership dominated at the constituent parliament at that time. Since 1993 Lietuvos Aidas is a paper closely related to the Homeland Union 17 Elder, Cobb (1983), The Political Uses of Symbols, p. 28. 18 Elder, Cobb (1983), The Political Uses of Symbols, p.30. 19 Elder, Cobb (1983), The Political Uses of Symbols, pp.35-36

11 (Lithuanian Conservatives)- most politically relevant the Right- wing party in the country. The choice of two ideologically different dailies was based on an idea to reflect political discourse changes in most objective way and using a limited number of texts, because all the coding was done as an expert coding and in manual way. The investigation covered only a period of two years, i.e. 1993 and 1997. This time frame was chosen because of a need to limit an overall amount of information as well as taking into account that 1993 was the first full calendar year in power by the Lithuanian Democratic Labour party, the ex-communist political organization; and 1997 respectively was the first full calendar year in power by the Right- wing coalition made up by the Homeland Union (Lithuanian Conservatives) and the Lithuanian Christian Democratic party. On the other hand, 1993 and 1997 represent the same political period, i.e. a consolidation of democracy if the first multi-party elections (in 1992 in Lithuania) are considered as an entry point to the democratic consolidation in majority of the Central East European countries. Certainly, this frame of time allowed to observe not only situational changes of political discourse, but as well as to get some information about more prolonged changes in political culture. The selection of particular editorials was based on a systemic sampling, i.e. only editorials of every full second week from each paper were selected for the analysis. A formula of selection of particular editorials was as following: A l number of editorials of each paper per year, round off to the bigger whole number; S number of full weeks per year. Al S 2 Table 3. Number of selected editorials Lietuvos Aidas Lietuvos Rytas Year 1993 1997 1993 1997 Number 26 26 26 a 26 b Total 104 c a but including personified comments on Mondays, the total number of selected editorials was 33 b but including personified comments on Mondays, the total number of selected editorials was 31 c total numbers of editorials was equal to 116, including personified comments on Mondays in Lietuvos Rytas in 1993 and 1997. The total number of editorials was equal to 116 for a period of two years taken from two dailies. All the political symbols were classified into separate groups and subgroups (see: Table 4). Table 4. Classification of political symbols in the political discourse analysis (an example) 1. Political community 2. Structures, norms and roles of regime: 2 1 Structures of regime:

12 2 1 1 Internal structures of regime 2 1 2 Foreign and international structures. 2 2 Norms and roles of regime: 2 2 1 Empirical conditions: 2 2 1 1 Nature of the physical world 2 2 1 2 Human nature 2 2 1 3 Nature of society and social organization. This part covers the following problems: a) Is priority given to individual or collectivity; who is more important individual or collectivity; b) are social borders fixed, natural and immutable; or are they arbitrary, artificial and subject to change? c) is society fragile and vulnerable, and in constant threat of imminent breakdown; or is it strong and resilient 2 2 2 Normative premises. Symbols of normative premises about the ends and the means of social behavior: 2 2 2 1 Social obligations. Symbols regarding the responsibilities and duties of the individual to others and to society as a whole, and how these obligations to be fulfilled. 2 2 2 2 Conceptions of authority. This group of symbols linked to the bases of authority and the limits of the legitimate exercise of power. 2 2 2 3 Premises of distributive justice. Symbols of fairness and equity, as well as how are these promoted. 2 2 2 4 Conceptions of personal virtue. Identification of personal virtues which are to be admired, revered and encouraged. 2 2 2 5 Life values. Symbols of the goals of individual and social activity, what kind of means are acceptable in seeking these goals. 2 2 3 Political prescriptions: 2 2 3 1 Political values. Symbols identified with: a) what values need to be protected and encouraged in the political process, b) what values have a priority. 2 2 3 2 Scope of politics: a) normal subject of governmental authority, b) where is a border line between private and public concerns. 2 2 3 3 Decision protocols. Symbols about organizational procedures, institutions, norms of conflict management and decision-making. 2 2 3 4 Standards of political conduct. Symbols of evaluation the behavior of political leaders and institutions. 2 2 3 5 Conceptions of the citizen s role. Symbols which identify: a) rights and duties of citizen,

13 b) properties of responsible citizenship, c) legitimate identities of citizen. 3. Situational symbols. Symbols of current authorities as well as non-governmental political actors, current policies and policy issues 20. Along the adoption of the classificatory scheme of political symbols, a dictionary of political terms, what contained 493 words, was developed after few phases of selection. But an initial version of the dictionary has had about 900 words. Of course, we were not fully satisfied with a quality of political symbols dictionary. But it was obvious that serious updating and revision might be taken only after completing the first round of study. On the other hand, the expert coding gave some advantage as researcher was able to decide to which particular group of codes to put one or another political symbol looking on the context. Table 5. A dictionary of political symbols (an excerpt) 2 2 2 3 Distributive justice 2 2 2 3 01 Planned economy 2 2 2 3 02 Market economy 2 2 2 3 03 Monopoly 2 2 2 3 04 Competition 2 2 2 3 06 License 2 2 2 3 07 Privilege 2 2 2 3 08 Bonus/ premium 2 2 2 3 10 Profiteer 2 2 2 3 11 Social guarantees 2 2 2 3 12 Burden of taxes 2 2 2 3 13 Shadow economy 2 2 2 3 15 Economic differentiation/social inequality 2 2 2 3 16 Protectionism 2 2 2 3 17 Subsidies 2 2 2 3 18 Minimal life standard 2 2 2 3 19 Line of poverty 2 2 2 3 20 Bankruptcy 2 2 2 3 21 Pocket of authority 2 2 2 3 22 Smuggling 2 2 2 3 23 Interest of consumer 2 2 2 3 24 Plundering of state property 2 2 2 3 25 Purchasing power 2 2 2 3 26 Equal opportunities What are some findings (rather still preliminary than final) of the project in political discourse analysis? There is certain stability of political discourse in Lithuania, in sense, that an intensity and shares of main political symbol groups changed only slightly in 1997 compared to 1993. Dominating item is a functioning of political regime, performance of political institutions, and its share was equal to 22,4 per cent in 1993, and to 22,6 per cent in 1997 taking into account a data only from Lietuvos Rytas (see: figure below). A political 20 Elder, Cobb (1983), The Political Uses of Symbols, pp.44-46.

14 community category has taken a permanent second position in 1993 and 1997 respectively with 10,6 and 11,8 per cent. Only few categories show relative increase of their weight in political discourse of the country, i.e. origins of society - from 5,4 to 7,0, international and foreign structures - from 0,5 to 2,3, or efficiency of decisions - from 4,7 to 7,1 in 1993 and 1997. Few other findings are even more important than the empirical data about political discourse in Lithuania. First of all, a dictionary of political symbols is expected to be used in party manifests research using the NUD*IST program. The project on political discourse needs to be updated in two ways: the classification of political symbols need to be revised in order to achieve more comparative perspective, especially, to party manifests study framework, as well as data basis must be wider in sense of time frame and contents. It is clear enough that the project needs to include all major dailies (4 5) of Lithuania, and not only the editorials but it is necessary to analyze letters to the editor therefore. Figure 1. Intensity of Political symbols (%) in "Lietuvos Rytas" 25,0% 22,4% 22,6% 20,0% 15,0% 1993 11,8% 10,6% 1997 10,0% 8,4% 7,9% 6,6% 7,0% 7,1% 7,1% 5,7% 7,0% 5,4% 5,3% 5,1% 5,1% 5,8% 5,0% 4,9% 4,7% 3,8% 4,0% 4,5% 5,0% 3,6% 2,3% 2,8% 3,0% 1,9% 2,5% 2,3% 0,5% 1,2% 1,5% 0,8% 0,0% 1 Politinė bendruomenė 2 1 1 Režimo vidaus struktūros 2 2 1 1 Fizinė pasaulio prigimtis 2 1 2 Tarptautinės ir užsienio struktūros 2 2 1 2 Žmogiškoji prigimtis 2 2 2 2 Valdžios supratimas 2 2 1 3 Visuomenės prigimtis 2 2 2 1 Socialiniai individo įsipareigojimai 2 2 2 4 Asmeninės dorybės 2 2 2 3 Perskirstymo teisingumas 2 2 3 1 Politinės vertybės 2 2 2 5 Gyvenimo vertybės 2 2 3 2 Politinė sfera 2 2 3 3 Sprendimų priėmimas 2 2 3 4 Politinio elgesio standartai 2 2 3 5 Piliečio vaidmuo 3 Situacijos simboliai Due to the expected increase of information, the political discourse analysis could hardly be managed by the expert coding, and the only way to come to definite results is through using of the computer-based programs. III.

15 What are perspective plans in using content analysis in estimates of policy actors positions in Lithuania? Of course, mainstream research will be still focussed on party manifests and the individual presidential candidates programs. However, there is a serious challenge to find out and to analyze a relationship and a degree of congruence between party manifestos and actual policy implementation by different political parties. A methodology of such investigation will be based on instruments (analysis of national budget expenditures) described by H-D. Klingemann R. Hofferbert and Ian Budge in 1994 21. Political discourse analysis is intended to be continued through the NUD*IST program. Nevertheless, here is one permanent problem: how to update discourse analysis dictionary? As for now, we see a possible solution in a selective (relatively narrow up to 10 editorials per year from each paper under investigation) expert analysis of printed media sources in order to update the dictionary. More important, that a general basis of political discourse research will be radically expanded, and it will cover more papers under investigation. It seems to be appropriate include not only all editorials but readers letters to an editor as an object of study, even titles of articles from politically relevant pages of daily papers. Both projects, i.e. on party manifests and political discourse need to be linked through developing similar (or even same) dictionaries and coding frames. Technologically both projects will be based exclusively on the NUD*IST program. This paper was concentrated mostly on the first part of content analysis, i.e. coding and getting statistical data from political texts and comparing percentages of different coded items. Indeed, it is real picture because of the novelty of content analysis as such was no physical time to apply more sophisticated methods and formulas in the analysis of data received through the content research. Nevertheless, the most simple evaluation scheme- construction of the Left- Right scale is underway from the data of party manifests. Next stop and more advanced way to go through the data from coded programs- to use some statistical-modeling procedures as the agenda and the mandate model as well as their variations already elaborated by Western scholars 22. Proximity of party policy positions and opportunity of coalition-building is another intriguing question which needs some analytical attention. A correlation between party system structure and proximity of party ideological as well policy positions might be evaluated well on the basis of party manifests research. And the list of questions what might be seen in a new light looking through the data of content analysis is extremely long. Consciously I am leaving aside some small and short-term projects (again related to political parties studies) which are under implementation at the moment because it is expected that relatively soon we will be able to present our results to international academic community in one or another way. Summarizing all that was mentioned before, it seems that we, in Lithuania, have interesting experience in developing of content analysis from the scratch. But due to 21 Klingemann H-D., Hofferbert R.I., Budge I. (1994) Parties, Policies and Democracy, pp.41-42. 22 Klingemann H-D., Hofferbert R.I., Budge I. (1994) Parties, Policies and Democracy, pp. 44-50.

16 technologically as well as methodologically sophisticated character of content analysis, it would be important to emphasize a role of international cooperation and necessity too establish more or less permanent networking in order to be more efficient and to escape serious mistakes in our investigations. Moreover, I still believe that contemporary science is developing by efforts of individuals but who are members of cooperating academic communities. Bibliography Alantas V. (1931), Antano Smetonos raštų stilius Vairas, No. 10, pp.184-194. Ambrazevičius J.(1934) Stiliaus kultūra. A. Smetonos stilius.-židinys, No.8-9. Bielinis L. (1995) Ivardziai politiniame tekste ir politiko nuostatos. Politologija, vol. 6 (1), pp. 73-82. Bielinis L. (1996) Kalbejimo strategijos politiniame tekste (Lietuvos Respublikos Prezidento kalbu pavyzdziu). Vilnius: TSPMI. Bielinis L. (1998) Prezidento rinkimai 97: Kandidatu i Lietuvos Respublikos Prezidentus rinkiminiu nuostatu ir stereotipu lyginamoji analize.- Politologija, vol. 12 (2), pp. 23-37. Elder, C. D, Cobb R.W. (1983), The Political Uses of Symbols. New York & London: Longman, Evenšteinas M. (1934) Dvasinis Antano Smetonos veidas jo raštų šviesoje.in: Antanas Smetona. Kaunas. Klingemann H-D., Hofferbert R.I., Budge I. (1994) Parties, Policies and Democracy. London. Krupavicius A. (1997) The development of political science in Lithuania: Years of the breakthrough. European Journal of Political Research, 31: 499-517. Lassan E. (1995).Valdžios ir disidentijos diskursas Tarybų Sąjungoje: kognityvinė-retorinė analizė. (in russian). Vilnius: Vilniaus Universitetas. Laver M., Garry J. (1997), Estimating Policy Positions from Party Manifestos. Paper prepared for the symposium Actors and Institutions in West European Parliamentary Democracies, Umea, June13-15. Mackevičius D. (1997) Tevynes Sajungos (Lietuvos Konservatoriai) programu turinio analize. Kaunas: VDU. Salys A. (1931) Smetona mus rašto moko (pastabos apie jo raštų kalbą).- Jaunoji Karta, No.23. Volkens A. (1992) Content Analysis of Party Programmes in Comparative Perspective: Handbook and Coding Instructions. Berlin: WZB.