The focus of this doctoral thesis is political communication process in terms of citizen's media and information practices and the effects these practices have on their willingness to engage in various forms of political participation. However, these effects are not conceptualized as linear and uniform for all citizens. The transforming nature of media audiences in the „new media environment“ (Cardoso, 2011; Castells, 2009; Prior, 2007), makes information practices, communication and participation of citizens socially stratified in specific ways, which reproduces inequalities of citizens in democratic process (Hargittai i Walejko, 2008; Lindell 2015; Lutz, 2016; Napoli, 2012; van Dijk, 2012). The theoretical approaches used in the thesis are media malaise (Delli Carpini, 2004; Putnam, 2000) and virtuous circle (Norris, 2000), which postulate the possible nature of the effects of media (information) use on political participation; and Bourdieu's (1986) concepts of habitus, economic, social and cultural capital, which are used to analyse the social stratification of media and information practices. According to media malaise theory, greater exposure to media, and especially to television, leads towards alienation and less trust in political and social institutions (Delli Carpini, 2004: 400). This theory is described in Putnam's (2000) book Bowling Alone, in which he claims that television privatises leisure time, in a way citizens spend more time indoors, have fewer social interactions, are less keen to spend time in activites that foster social capital and therefore less civically engaged. Research which supports this theory showed that negative media coverage of politics and media framing of politics as a strategic game makes citizens more cynical about politics and makes them feel less adequate to make changes through their participation (Delli Carpini, 2004: 399). On the other hand, authors that believe that media can mobilize people to political participation, argue that exposure to political information offered by media and political participation are in virtuous circle – citizens that are more engaged in politics will follow news media more, which will keep their political interest and make them further engaged (Norris, 2000; Strömbäck i Shehata, 2010). According to virtuous circle theory, citizens that follow news media the most are those who have most political knowledge, trust in institutions and are more likely to be politically active. However, media environment and media audiences in such environment are profoundly changed and political communication processes should be examined in the light of such changes. With multiplication and fragmentation of different forms of media in the post-broadcast era, citizens should have more information choices in a „high choice“ media environment (Prior, 2005, 2007). However, this may not be true- the increasing choice provided by new media may even increase inequalities in political knowledge, as those with less political interest will avoid some information sources (Prior, 2005, 2007). The new communication models of „mass-self communication“ (Castells, 2009) or „networked communication“ (Cardoso, 2011), provided by digital networked media, adds to the media choice, and enables new media practices. Audiences can search, avoid, share or create information, be autonomous, selective and participative, but at the same time might become fragmented, polarized, or divided by media repertoires (Ahlers, 2006; Hasebrink and Domeyer, 2012; Ksizek and Tewksbury, 2003; Napoli, 2012; Yuan, 2011; Webster, Ksiazek, 2012). The fragmentation of media audiences, the differences in media choices they make and in ther media practices could be traced following their social stratification (Lindell, 2015). In such environment it is much harder to analyse uniqe and linear media effects. Political communication research should rather adapt to these changes and embrace more flexible approaches from media audiences research (Kleis Nielsen, 2014). Therefore, to adress such changes and the stratification of media audiences, in this thesis the non-mediacentric approach is followed. Media audiences are studied from the sociological perspective, analysing how do media users, as social agents in certain social position use media (Lindell, 2015). The concept of habitus (Bourdieu, 1979/2011) is taken as a key concept to analyse different information and media practices, which are structured by the social position of the media user, the resurces media users have and environment and circumstances in which they act, but which also enables the analysis of media users as social agents, or their agency, enabled or limited by structures. Bourdieu's approach is seen by communication scholars as useful to address cross-media or trans-media use in contemporary media environment and „systematically describe the interrelation between the individual's media use and his/her societal position in a structured society“ (Weiss, 2000/2001, in Hasebrink and Domeyer, 2012: 763). Bourdieu (1986, 1979/2011) claimed that social relations are impossible to analyse without employing conepts of capital in all of its forms, not just the one used in economic theory. Depending on the field in which it operates, capital exists as an economic capital, cultural capital and social capital. Economic capital could be directly transformed in money or ownership rights and cultural capital can be transformed into economic capital through institutionalization to educational qualifications (Bourdieu, 1986). Bourdieu's matrix of social space is constructed through dimensions of economic and cultural capital. However, this analysis should also include social capital, especially when it regards post-socialist societies in which political and social capital were far more important for the social position then economic capital (Bourdieu, 1998; Cvetičanin, 2007, 2012; Cvetičanin and Popescu, 2011). Similarly as Bourdieu (1979/2011) approached the analysis of cultural consumption of consumers with specific habituses on specific social positions, media consumers, or media audiences will be analysed in this thesis. In order to distinguish between different groups of media users defined by social stratification of various media practices, the concept of media/ information repertoires is introduced (Hasebrink and Domeyer, 2012). The concept of media repertoires is useful to analyse media use in cross-media environment, and refers to „the entirety of media he or she regularly uses“ or „relatively stable trans-media patterns of media use“ (Hasebrink and Domeyer, 2012: 759). As stratification of media audiences is analysed in the thesis, or inequalities which produce different information repertoires and practices, besides Bourdieu's concept of economic, cultural and social capital, the analysis also draws from the digital divide and digital inequalities theories (Hargittai, 2003; Hargittai and Walejko, 2008; van Dijk, 2012) which offer explanations to unequal access to and use of (digital) media. In this thesis four research questions were answered: 1) Do different media/ information practices of citizens have mobilization effect (virtuous circle theory) for differnet forms of political participation? 2) What kind of media audiences typologies in terms of their media/information repertoires exist? 3) In what way does the social position of the media user (economic, cultural and social capital) influence the formation of media/ information repertoires? 4) What kind of media practices and preferences and attitudes towards politics are specific to types of media audiences defined by their media/information repertoires? For the purpose of determining patterns of socially stratified media use and political participation, a combination of quantitative and qualitative methods will be used. The integration of the quantitative and qualitative data will be done through explanatory sequential mixed methods design (Creswell, Plano Clark, 2011: 82). This kind of data integration in mixed methods design is most useful when the research aim is to assess trends and relationship between variables and explain the deeper mechanisms or reasons behind these trends (Creswell, Plano Clark, 2011: 82). Mixed methods explanatory sequential design is done in two phases. In the first phase, the quantitative strand is designed and implemented by creating instruments for quantitative analysis, collecting and analysing data. In the second phase, qualitative analysis follows quantitative analysis design and these data are analysed keeping in mind the results of the first phase. First and second phase are “integrated” through the analysis of both quantitative and qualitative data (Creswell, Plano Clark, 2011: 83). Research results don't confirm the hypothesis that more exposure to information has a significant effect on political participation, besides on online participation. On the other hand, commercial media have media malaise effect on some forms of political participation. Therefore, media in Croatia seem not to have a significant beneficial effect on political participation, as virtuous circle theory would suggest. Croatian media audiences are fragmented based on their information repertoires in five ideal types, with uninterested users being the most frequent type. These different groups of media users differ because of their social position, the economic, cultural, social capital and other resources they hold, and are informing themselves in different ways and with different pursposes. Analysis also resulted in five ways these media audiences groups frame politics. Politics is framed mostly negatively, audiences being apathetic to it, framing it as a „dirty game“, seing politics as being mediatized (which is also perceived as negative), having populist approaches to politics, or more rarely expressing interest. However, noone expressed positive frames of politics as a democratic process.