THE COMPARATIVE STUDY OF POLITICAL COMMUNICATIONS

In many subfields, comparative politics has been advancing by leaps and bounds in recent years, as illustrated by the contemporary state of the art in The Oxford Handbook of Comparative Politics.Footnote ⁶ During previous decades, studies commonly drew upon a deep knowledge of specific languages, histories and contexts to illuminate particular countries or regions. This older tradition persists, but the last decade has also witnessed a torrential flood of large-N cross-national datasets, especially macro-level indicators of the quality of democratic governance.Footnote ⁷ Cross-national surveys of public opinion are now widely available, expanding their coverage to many developing societies.Footnote ⁸ Methodological advances include recognition of the value of mixed research designs. These combine the external validity, rigour and generalizability derived from large-N quantitative comparisons with the virtuous of specific qualitative case studies, describing the local processes underlying the broader patterns.Footnote ⁹ Fuzzy logic helps develop innovative typologies.Footnote ¹⁰ Innovative quasi-experimental designs, especially controlled studies of specific policy reform interventions, generate compelling results. The globalization of political science strengthens professional intellectual networks across national borders, especially within Europe. More open societies, linked with processes of democratization and development, facilitate the diffusion of political science worldwide. Funding councils prioritize cross-national collaboration. Educational mobility for graduate students within Europe encourages growing internationalization of the profession. In the subfield of communications, globalization has stimulated literatures examining the phenomenon of transnational broadcasting, cultural economicsFootnote ¹¹ and global media.Footnote ¹² For all these reasons, and many more, during the last decade contemporary theorizing about many central phenomena in comparative politics moved closer towards achieving the elusive goal of developing from the study of nations to the study of types.Footnote ¹³ At the same time, the impact of these changes has been uneven across the discipline. Certain subfields such as the study of democratization, political institutions, political economy, international relations and political culture have been transformed by these developments far more than others and arguably the comparative study of political communications continues to lag behind.

What counts as falling within the realm of ‘comparative’ political communications? And why is comparative work important? In the broadest sense, all normative and empirical analysis of political communications is comparative, if this is understood to mean contrasting different units, whether types of media channels (TV versus the internet), programme genres (news or soaps), ideal types (the roles of journalists as watchdogs or advocates), audiences (young versus old), media effects (framing and agenda-setting) or time periods (the golden age of network TV versus the cable era). In this platitudinous sense, all analysis in the social sciences needs comparison of observations falling into categories to make any sense of the world.

The heart of the comparative study of political communications, however, is conventionally understood to focus more narrowly upon contrasting spatial units, usually comparing nation-states but also, alternatively, local communities, media markets or global regions, such as Europe or Latin America. American scholars commonly take it for granted that all work on foreign countries is, somehow, automatically comparative (see, for example, the classification of book reviews in the American Political Science Review). Not so; just as there has to be two hands clapping, so there has to be more than one spatial unit to compare.

Moreover, in their chapter in the Esser and Pfestch volume, Gurevitch and Blumler suggest that a further essential distinction needs to be drawn between cross-national studies that use other places as a convenient way of testing propositions that would otherwise have been examined at home (for example, whether patterns of internet use in America hold, say, in Sweden or Germany), and more ambitious comparative research, which seek to understand how varying contexts (such as those generated by different types of state regimes and political institutions, cultural regions, levels of development or media systems) shape processes of political communications. Through systematic comparison, scientific research about political communications seeks to make descriptive or explanatory inferences based on empirical observations about the known facts, using transparent and public procedures, under conditions of uncertainty, to explain facts we do not know. To do this, comparative frameworks can either adopt what John Stuart Mill identified as the logic of the ‘most similar’ strategy (apples with apples), seeking to analyse the mass media while ‘controlling’ for certain shared cultural, social or regime characteristics; alternatively, the frameworks can follow the ‘most different’ strategy (apples with oranges), seeking to maximize contextual variations when identifying regularities in the phenomenon under examination.Footnote ¹⁴ Comparative research designs can also choose to focus in depth upon a few selected case studies, ideally illustrating broader theoretical frameworks and conceptual typologies. Or they can draw contrasts within a limited number of units (such as across EU member states). Or else they can adopt large-N comparisons seeking regularities over space and time around the world. Often convenience frameworks are commonly used, such as comparing the countries included within specific cross-national survey datasets, or the network of colleagues participating in a specific research project, without a great deal of thought about the consequences of such decisions for scientific inference. Not surprisingly, this practice often causes considerable confusion, since the ability to generalize more broadly from any research depends upon the way that the geographic units selected for comparison reflect a larger universe, just as survey research depends upon the rigour of selecting survey respondents drawn from a random sample of the population. Case studies with a limited number of cases, such as contrasts drawn between journalism in a few countries, are particularly problematic in this regard, if the findings are interpreted as representing the larger universe of, say, post-industrial societies. Given this understanding, the heart of the challenge facing comparative political communications, it is argued, lies in resolving complex conceptual, data and methodological issues, which continue to hinder theoretical and empirical progress.

If we can overcome these, the advantages of comparative research designs are many, as Esser and Pfetsch emphasize in their overview.Footnote ¹⁵ First, this process expands the contextual environments for observations, allowing broader generalizations to be established in the social sciences. This reduces national idiosyncrasies in the search for broader regularities over place and time. Through this process, theories couched in universal terms, based on observations derived from one or two cases, can be shown to reflect exceptional outliers. This process reveals parochialism and highlights the underlying causal factors taken for granted within any particular environment. Individual-level social psychological behavioural studies, in particular, commonly assume that certain general relationships exist – for example concerning the impact of negative news on voter turnout, or agenda-setting effects on policy priorities – when in fact these individual-level results may be heavily conditioned by the specific social, cultural and institutional context. Cross-national research is also potentially valuable for public policy-making, by highlighting alternative interventions, strategies and reforms that can inform the policy-making process, adapting best practices to local needs. The discussion about the comparative method by contributors in the Esser and Pfetch volume, notably chapters by Kleinsteuber, and by Wirth and Kolb, usefully summarizes these issues and also reviews the pros and cons of alternative research designs and methods used in the comparative study of political communications.

CLASSIFYING MEDIA SYSTEMS

One of the perennial issues in the comparative study of political communications has been the attempt to establish suitable theoretical models, conceptual tools and classificatory frameworks. Indeed one of the most important limits on the subfield has been the lack of a consensus about the core theoretical concepts and standardized operational measures. These are the essential building blocks and analytical tools for research, which help to reduce the cacophonous Babel. Ever since Aristotle, analytical typologies have always been a vital part of comparative politics. Other subfields established standard conceptual classifications that are widely used as the common basis for comparison, exemplified by Giovanni Sartori's categorization of party systems,Footnote ¹⁶ Maurice Duverger's classification of plurality and proportional representation electoral systemsFootnote ¹⁷ (subsequently developed by Douglas Rae and others)Footnote ¹⁸ and Arend Lijphart's distinction between majoritarian or consensus democracies.Footnote ¹⁹ Each of these paradigmatic frameworks has been amended and refined numerous times over the years, identifying subcategories and developing more effective operational measures, with the original ideas providing the foundation for subsequent intellectual development. A broad consensus has developed over the years, allowing common standardized measures to be used across diverse studies. By contrast, the conceptual and typological models used in political communications have usually been heavily normative, difficult to operationalize and measure, and with restricted applicability for worldwide comparisons.

But what should be included as an effective typology of ‘media systems’? One difficulty that is immediately encountered is to define which units should be compared, since the mass media includes multiple outlets – broadsheet and tabloid daily local and national newspapers, books and printed periodicals, radio and television broadcasts, as well as the complex and ever-growing range of newer information and communication technologies that are merging delivery platforms. Even the simple concept of ‘mass’ communications, which used to be exemplified by terrestrial broadcasts, has dissolved more recently with the expansion of interpersonal and group-to-group networks (are ‘blogs’ a form of mass communications if no one reads them?). The concept of a ‘system’ suggests joined-up interaction among disparate parts (the hip bone connected to the thigh bone etc.), so that it becomes confusing if a media system is defined by virtue of its separate components, such as the existence of public sector broadcasting or strong links between parties and newspapers, which may, in fact, be unconnected. Moreover, there may be one ‘system’ identified for national newspaper markets and another for national television broadcasting, so how can we define a ‘media system’ across such disparate phenomena? Indeed, many standard attempts to do so, by bringing in concepts such as ‘party parallelism’, in fact try to define a system of political communications, which is a very different animal. Misleading conceptual classifications can provide culturally stereotypical blinkers that hinder rather than help, by obscuring the real commonalities and contrasts in the cases under study (apples are misclassified as oranges).

The origins of comparative work on media systems can be traced to 1956's seminal Four Theories of the Press by Siebert, Peterson and Schramm.Footnote ²⁰ An early attempt to develop conceptual models describing how media systems functioned worldwide, the framework was heavily influenced by the Cold War era. The study theorized that media systems are embedded within their broader social and political context, although in practice they focused mainly upon the relationship between the state and the mass media, and on this basis they identified several distinct normative ideal types for how the press should function. The key contrast was drawn between the authoritarian model (including the Soviet communist variant), where the press was regarded as subservient to the state, versus the libertarian (and the subcategory of the social responsibility model), where the media functioned independently as a classic Fourth Estate, free of government control. The degree of freedom of expression is a critical aspect of media independence, with the libertarian model seen as exemplified by the free market of ideas among rival commercial newspapers and broadcasters in the United States, with public service broadcasting reflecting the social responsibility model, in stark contrast to state-owned and state-run Soviet media. Given the absence of rival frameworks, these conceptual models proved highly influential, even though the ideas came under sustained attack over the years. The fall of the Berlin Wall drove another nail in the coffin of the ‘Soviet’ model.Footnote ²¹ The absence of theorizing about the range of media systems found in the developing world became increasingly apparent.Footnote ²² Moreover the simple bright distinction between commercial versus state broadcasters became increasingly fuzzy with the growing commercialization of European public broadcasting, and indeed the expansion of PBS and NPR in the United States; many scholars cast doubt on the utility of the pure social responsibility and libertarian models.

In the light of the critiques, there have been many attempts over the years to modify and improve the original Siebert, Peterson and Schramm framework, such as that proposed by Blumler and Gurevitch.Footnote ²³ One of the most ambitious attempts, attracting widespread attention, has been developed more recently by Hallin and Mancini.Footnote ²⁴ The authors restrict the focus to classifying ‘media systems’ in 18 nations within North America and Western Europe, emphasizing that this facilitates comparison of like with like, covering many established democracies and post-industrial economies, without claiming that similar ideal types can necessarily be identified elsewhere. Hallin and Mancini suggest that media systems in the countries under comparison can be classified based on four major dimensions: (1) the degree of state intervention in the media system, especially via public service broadcasting, as well as by legal regulation and subsidies; (2) the extent of political parallelism, referring to how far news media outlets are partisan or more neutral, and how far media systems reflect party systems; (3) the historical development of media markets, especially the legacy of this process for contemporary newspaper circulation rates; and (4) the extent of journalistic professionalism. The authors believe that these criteria cluster together, at least loosely, into distinct types, suggesting a three-fold classification of media systems. They identify a liberal model, which they suggest prevails in Anglo-American countries (Britain, the USA, Canada and Ireland), characterized by commercial media and market mechanisms. By contrast, the democratic corporatist model, which is thought to prevail across the consensus democracies in northern Europe (Austria, Belgium, Denmark, Finland, Germany, Netherlands, Norway, Sweden and Switzerland), emphasizes the links between commercial media and organized social and political groups, within the context of an active but limited role of the state. Lastly, the polarized pluralist model, which they suggest typifies Mediterranean Europe (France, Greece, Italy, Portugal and Spain), integrates media into party politics, with a weaker commercial media and a stronger role for the state. The remainder of the book describes each of these models in more depth, to see how far the concepts fit the 18 cases under comparison. Hallin and Mancini conclude that, although the liberal model has dominated media studies, the polarized pluralist model probably provides a more accurate description of journalism in many parts of the world.

The book usefully provides an ambitious attempt to reformulate and classify media systems. Like recent work seeking to identify ‘varieties of capitalism’, Hallin and Mancini's study attempts to reformulate the distinctions in media systems among the universe of Western democracies. The typology has been used by other scholars both to understand contrasts within the countries included in the original study – just as in journalistic cultures and election coverage – as well as to consider its utility when analysing the media in other regions and types of regime, such as in Russia and China. Nevertheless, the Hallin and Mancini framework suffers from several major shortcomings that need to be addressed before we can conclude that this provides an appropriate conceptual typology for the subfield.

First, and perhaps most importantly, it is not apparent whether the four dimensions identified by Hallin and Mancini are indeed the critical ones that define the major contrasts today among contemporary media systems (or even, more accurately, political communication systems). There are two potential dangers to any classificatory schema, which have to steer successfully between the Scylla of minimalism and the Charybdis of maximalist concepts. Minimalist or ‘thinner’ notions focus attention upon just a few elements of political communications but they may thereby exclude certain potentially important aspects of the phenomenon under study. By focusing upon a narrow range of benchmarks, minimalist approaches reduce the risks of including theoretically irrelevant attributes and redundant elements into composite measures. They are more elegant and parsimonious. By contrast, maximalist or ‘thicker’ notions provide a richer and more comprehensive definition of the phenomenon, providing greater measurement validity, but at the same time certain aspects may be difficult to operationalize with any degree of consistency. This makes it more difficult to replicate studies, an essential test of the robustness of any results. Unfortunately, the Hallin and Mancini framework appears to suffer simultaneously from both problems.

For example, certain essential dimensions of any media systems are missing. The framework emphasizes the importance of the historical development of mass circulation newspapers, as one of the four dimensions used for classification. This clearly has an impact upon the contemporary structure of the news industry evident in different nations, and thus the reliance upon the print or broadcast media for news, just as the role of public broadcasting corporations established during the pre-war years of radio broadcasting continues to shape modern television. Nevertheless, if the larger theoretical point concerns national variations in access to different types of mass communications, then the major lacunae in the discussion, indeed the 800-pound gorilla in the room, concerns the role of new information and telecommunication technologies, which are not featured anywhere as part of the classification. It is well-established that there are major contrasts, even within post-industrial societies, in the spread and use of all sorts of electronic technologies, from the general use of computers and the internet to the diffusion of mobile (cell) phones, text messaging, online social networks, blogging and TV teletext. There are important contrasts in contemporary internet access even among closely related countries, for example, Eurostat estimates that in 2008 more than eight out of ten households had internet access in Sweden and the Netherlands, compared with just four out of ten in Italy and one-third in Greece.Footnote ²⁵ The contrasts worldwide are, of course, even broader. Surprisingly, none of these differences feature in any of the classifications proposed by Hallin and Mancini; indeed throughout the book, it is difficult to find even a passing mention of any of these important forms of information and communication. It appears as though journalism and the media systems are frozen in the mid-twentieth century. It could perhaps be suggested that the diffusion of new information and communication technologies often reflects other divisions in media systems, but this argument at least needs to be articulated for the exclusion to be justified.

Other important dimensions of media systems are also omitted from the Hallin and Mancini classification. For example, although the role of the state ownership and subsidies of the media is regarded as important, there is little explicit acknowledgement of the importance of press freedom, and minimal comparison of the legal framework guaranteeing freedom of expression. Perhaps this is not regarded as central to any study classifying media systems within established Western democracies, but as soon as the research travels further, even among the 25 member states of the European Union, this immediately becomes a critical dimension for comparing media systems. Moreover it seems perverse to exclude press freedom in any classification of media systems, since many institutions have developed a range of cross-national indicators, such as those provided by Freedom House, IREX, Reporters without Borders and CIRI. Issues such as political parallelism may indeed prove more important than press freedom for classifying media systems, but again this argument would need sustaining rather than being assumed by default.

Moreover, there are also problems about operationalizing the components of the proposed schema, and thus generating scientifically replicable standards and measures. The evidence presented by Hallin and Mancini remains descriptive, drawing upon their reading of selective historical examples, limiting how far their classification can be replicated by other scholars. There are also many impressionistic reasons to question the proposed categorizations and any potential misclassifications of specific cases sow doubt about the value of the overall schema. For anyone familiar with the UK media, for example, on many key aspects it seems utterly dissimilar to the media in the United States. As Scammell and Semetko note, ‘It [Britain] shares with the U.S. a commitment to free markets, freedom of speech, and self-regulation as the guiding principles for newspapers. It shares with northern Europe a history of highly partisan newspapers and regulated television markets, dominated by well-funded public service broadcasters.’Footnote ²⁶ Since the 1950s, Britain has had a dual system of broadcasting, dominated by the standards and ethos established by the British Broadcasting Corporation, with the commercial sector heavily regulated to maintain high standards of public broadcasting. American television, aside from minority and under-funded PBS and the anorak channels, C-Span, is predominately commercial. Election news TV and radio broadcasts in Britain display internal diversity, with stop-watch balance regulated and monitored across party coverage. Election news coverage on US TV channels such as Fox News and talk radio disregards all these principles. American presidential elections on TV are heavily ad driven. In Britain, no parties can buy any TV advertising and the free party political broadcasts are an opportunity for most voters to slip out and put on the kettle. Most US newspapers are mainly regional and metropolitan, as befits a federal state. Most British newspapers are national, as befits a unitary state. The overwhelming majority of US newspapers are respectable ‘quality’ broadsheets, apart from a handful of supermarket tabloids. The largest-circulation papers in the UK, such as the Sun, are popular down-market tabloids. US regional newspapers follow the tradition of balanced op eds, observing principles of internal pluralism, in the attempt to reflect a broad bipartisanship. British newspapers have become less overtly partisan over the years; nevertheless, anyone picking up leading papers such as the Guardian or the Daily Telegraph can usually identify quite easily which party they support, even outside election campaigns. The British press (but not television) has external pluralism; the idea that the press is neutral or information oriented cannot be seriously sustained. US reporters and broadcasters are often professionally trained with journalism and media studies degrees from long-standing university departments reflecting these disciplines. In Britain, until relatively recently, journalism was an apprenticeship with on-the-job vocational training, and even today there are relatively few journalism or media studies departments in the leading universities. In short, media systems in Britain and the United States seem, at first glance, to have almost nothing in common. And if the logic is faulty in these particular cases, then this raises doubts about other classifications, for example whether there are really closer similarities between Germany and Norway, or between Germany and France. Without any rigorous process for testing the classification independently, when by establishing certain standardized indicators or a set of explicit decision rules, the categorization proposed by Hallin and Mancini remains fuzzy, impressionistic and unscientific.

This leads naturally to the third point: the authors provide a general overview, rather than attempting to operationalize each of the four dimensions with empirical indicators. Yet it is important to measure the core concepts if we are to test rigorously how far the different dimensions actually cluster together in meaningful ways – and if we are to examine the systematic consequences arising from any differences in media systems, which should be the major point of the classificatory exercise. Many other attempts at comparing mass communications have also focused upon categorical approaches, suggesting that media systems can be classified into distinct types once they cross a specific threshold. But it seems preferable to consider utilizing continuous indicators, such as those concerning access to different forms of mass communications, measures of levels of press freedom or classifications of state ownership, implying subtle gradations in levels of mass communications. For either approach, we need to consider the major sources of random and non-random measurement error arising from these decisions that could potentially bias estimates of effects and generate misleading comparisons. Two criteria – validity and reliability – are particularly important for evaluating the construction of any empirical indicators in the social sciences. Valid empirical measures accurately reflect the analytical concepts to which they relate. Reliable empirical measures prove consistent across time and place, using data sources that can be easily replicated to allow scholars to build a cumulative body of research. Scientific research makes its procedures public and transparent, including the steps involved in selecting cases, gathering data and performing analysis.

It remains unclear whether the concepts proposed by Hallin and Mancini can be clearly related to broader theoretical concerns in mass communications, as well as operationalizable when applied to research. Potentially some of the four dimensions can be measured, at least imperfectly, through various indicators, but not all. The degree of state intervention in broadcasting has been of increasing interest to economists, in accounts about how media competition relates to government transparency and corruption. The degree of state intervention in major media firms has been gauged for 97 nations worldwide by Simeon Djankov and colleagues, monitoring the number of state and commercial television channels and their audience share, as well as the number of state-owned newspapers and their share of readership.Footnote ²⁷ Time-series data for newspaper circulation rates and newspaper titles, including for the major tabloids and broadsheet press, are also easily available from UNESCO for all the countries under comparison, although the series remains somewhat dated and UNESCO is in the process of establishing new datasets and collecting more comprehensive standardized cultural indicators. Therefore two of the Hallin and Mancini building blocks can be measured from the existing datasets. By contrast, it remains unclear how we can operationalize ‘political parallelism’, in terms of how far the partisanship of the mass media reflects the party system. This requires systematic content analysis of data that could classify the partisan bias in the media: a massive undertaking, not least because media partisanship varies even within particular countries by type of media outlet (local and national newspapers, radio, television and websites), by different newspaper sections (editorial, front-page) and among news and current affairs channels and programmes (such as Fox, CNN, PBS and C-Span), as well as over time. Since partisanship is also often in the eye of the beholder, any cross-national content analysis of a representative sample of media outlets is also open to problems of interpretation. Expert surveys would be another way to try to collect data, similar to expert scales of party systems, although the diversity of media outlets makes this more challenging. The available content analysis datasets also have problems of methodological standardization, transparency and replicability, so that each study commonly reinvents the wheel. We lack a common archive for content analysis datasets, such as the standardized resources available for the analysis of social surveys. As Strömbäck and Kaid conclude, after seeking to compare patterns such as the degree of negativity or personalization in campaign coverage: ‘There do not seem to be any standardized instruments and coding instructions. The unfortunate end result is that it is often difficult to compare the election news coverage across borders, and although the terminology used is often similar, the extent to which the empirical results are comparable is often uncertain.’Footnote ²⁸ Moreover, the extent of journalistic professionalism is also challenging to monitor with any degree of reliability, although surveys of news professionals are gradually expanding globally to fill the gap here. Some appropriate proxy aggregate indicators could also be collected, for example by examining the existence of journalism-training departments and accreditation processes used in different countries. Rather than trying to fit disparate phenomena into broader categories, where there are serious dangers of errors introduced by misidentifying cases, it may be preferable to utilize continuous measures that gauge scales for different types of media indicators, whether concerning access and use, press freedom, media ownership or related indices.

COMPARING ELECTION NEWS IN YOUNGER AND OLDER DEMOCRACIES

Turning to the other edited volumes under review, these illustrate the contemporary state of the art. The collection assembled by Strömbäck and Kaid focuses on how the structure, contents and effects of election news coverage vary across 22 countries. This is an important topic that illustrates both the strengths and weaknesses of the subfield. If there is any area where we should be able to develop some strong and robust cross-national findings, then this is it. Election news attracts considerable attention in political communications, as illustrated by the wide range of literature summarized in this volume, whether concerning the regulatory legal framework and structure of the mass media, the contents of election news coverage or the effects on the public. There is a great deal of useful descriptive information about these matters contained in the book and, moreover, the editors are to be applauded for including middle-income developing countries, such as South Africa and India, as well as the usual suspects. The contributors are all internationally well-known researchers, with expertise on each nation. Unfortunately the collection remains largely stuck in the older comparative tradition: separate chapters present a country-by-country Grand Tour of election news, rather than providing a synthetic synthesis organized by topic (such as comparisons of the role of opinion polls, the partisanship of the press, the legal regulation of television campaign broadcasts, the impact of election news on voting choice, and so on). As a result, although the volume presents a stimulating travelogue where you can visit and learn a great deal about each country, the net result for comparative political communications is unfortunately less than the sum of its parts. The reader needs to put together any general lessons derived from the cases.

In part this is because the loose theoretical and organizational framework proposed by the editors in the introduction remains too under-developed to generate interesting generalizations, imaginative propositions or potentially unexpected findings in the conclusions. Too often the results are of the ‘it depends’ variety. For instance, the editors suggest that the type of electoral and party system in each country might provide some potentially testable contrasts in election news media coverage, such as in the degree of partisan bias in the media, but they then suggest so many under-theorized linkages between the electoral system and the media system that by this logic almost anything could be explained or explained away. Does a majoritarian electoral system encourage a more partisan press – or a less partisan one? Does a multiparty system facilitate less negative news (because of the need for parties to work together in coalition governments after polling day) or more negative news (because of the degree of ideological fragmentation across the political spectrum)? Without developing a stronger and more rigorous set of theoretical propositions linking the type of political institutions to election news, it becomes difficult for contributors to test common propositions and for the editors to draw together any systematic lessons from the separate chapters. In the conclusion, the editors make an attempt to identify some common threads emerging from the chapters, but too often the generalizations are either at a level that is platitudinous and self-evident (‘media matters’), or a product of the specific research design and comparative framework used in the volume, or else simply of the weak ‘it depends’ variety. Hence, for example, the editors conclude that ‘Mediated communication clearly dominates the channels of information and persuasion in these countries’ (p. 421), but indeed the study never sought to look at low-income developing countries where this may not be the case, nor indeed to compare the role of the mass media with the impact of interpersonal communications, such as local party and candidate canvassing, community rallies, retail politics bypassing the news media and modern forms of social networking online. Without a direct comparison between mediated and interpersonal forms of communications, we cannot make any claims about which is most important for voting choice, campaign knowledge, partisan loyalties, political mobilization, and so on. The most ubiquitous channels of communications are not necessarily the most influential ones. We also learn from the conclusion that in almost every case under comparison, television is the ‘dominant’ channel for communication, but again this is derived from focusing comparison upon industrial and post-industrial economies; the more interesting question is the relative importance of different media channels upon different types of attitudes and behaviours, both among elites and citizens. Many of the other general conclusions in the volume, for example concerning the degree of media bias in campaign coverage or any agenda-setting effects on policy issues, rests heavily upon evidence and secondary literature derived from the US case, not from broader patterns and contexts.

Even the commonplace claim in the conclusion that election campaigns have become increasingly ‘mediatized’, or whether a process of ‘mediatization’ of politics has occurred in post-industrial societies, meaning that the mass media have become increasingly prominent actors in the political process, is not self-evident. To test the evidence for this thesis systematically we would need to gather evidence for the activity of all the major actors in this process, such as the direct channels of party–voter communications through traditional forms of local campaigning, as well as through the innovative linkages provided via new information communication technologies. Rather than a steady development involving progressive ‘mediatization’ of election campaigns, as proposed by the editors, instead there could be distinct stages of traditional, modern and postmodern campaigns. Understood in this light, some contemporary shifts in postmodern campaigns re-emphasize traditional face-to-face interpersonal communications, as exemplified by some of the contemporary techniques used to link together activists by the Obama presidential campaign. Multiple actors are engaged in campaigns, and communication scholars ideally need to integrate evidence derived from the activities of parties, candidates and activists, the mass media and the electorate for a comprehensive understanding of this process. Indeed, media-centric claims by media scholars are often framed at such a high level of abstraction that they are cannot be used to generate interesting and testable mid-level propositions that could be confirmed or refuted with empirical evidence derived from varying national contexts, thereby extending new knowledge in the comparative subfield.

Lastly, the edited volumes by Gunther and Mughan and by Katrin Voltmer present studies of political communications across a range of younger and older democracies. The Voltmer book, a product of an ECPR workshop, covers cases drawn from a couple of medium-income developing countries (South Africa and Taiwan), as well as Latin America and Eastern Europe. The book is loosely organized by arranging chapters in thematic sections, rather than around regional cultural areas. The volume focuses upon case studies of the role of the mass media in the process of democratization. The collection again contains useful information but most of the contributors present single-nation studies – almost one-third of the chapters are devoted to Russia – with some limited comparison of Latin America, and a four-nation comparison of newer democracies (Uruguay, Chile, Hungary and Bulgaria) by Voltmer and Schmitt-Beck. A tighter set of research issues, and the selection of a more representative set of cases of younger democracies from all world regions, would have bound the cases together more logically and allowed a more consistent comparative framework. The older volume by Gunther and Mughan considers similar issues of the changing relationship between the mass media and democracy, but it is more systematic in the thematic framework set by the editors in the introduction and conclusion. Chapters contrast developments in this relationship in ten countries, including older democracies such as The Netherlands, Britain and Germany, as well as newer democracies such as Hungary and Chile. But the conclusion suggests that the search for cross-national regularities across these cases more often reveals a complex reality that defies simple analysis. The main trends that emerge are again those that can be regarded as conventional; television has become the predominant source of national and international news across the cases under comparison; politicians have become increasingly aware of the power of television and sought to adapt its uses to their purposes; but the effects of the media remain highly conditional upon institutional and cultural contexts, such as legal regulations, technological developments and regime characteristics. In short, in some frustration the editors cautiously conclude with media effects that ‘it varies’.

The main lesson from the brief review of some of the recent work in the subfield is that these books do advance the subfield in certain important ways but we probably need substantial advances in research designs before we can establish a body of more systematic knowledge. The theoretical frameworks need revising with sharper, cleaner and more precise concepts. The cross-national aggregate indices that are emerging need to be utilized more fully, so that standardized measures become widely used and tested across a wide range of contexts and studies. The search for typological schema and categorical classifications of ‘media systems’ or ‘political communication’ systems should perhaps be abandoned, although we should still develop sharper typologies of the separate components within ‘systems’. Any cross-national studies need to think hard about the selection of cases, so that we move from ad hoc convenience comparisons towards a clearer sample of nations within a well-defined universe. And we need to establish archives of common datasets, for example using standardized measures to content-analyse election news, to accompany the growing range of cross-national social surveys that now include measures of media use. All of this is a daunting and substantial research agenda but this is necessary to make substantial progress in the subfield, to retreat from Babel, and to construct a more scientific footing for the whole enterprise.