Table of Contents
List of Figures and Tables
Preface and Acknowledgements
2 Theory and Preliminary Analysis
2.1 The Rationale of Talking Intervention
2.1.1 The Discourse on Intervention as Frame Contest
2.1.2 Framing and the Banality of the Body-Bag Hypothesis
2.1.3 Rethinking the Role of the Media
2.1.4 The Need for a Psychology of Discourse
2.2 Towards an Interactionist Perspective of Framing
2.2.1 Framing – Still a Messy Paradigm
2.2.2 Framing as Manipulation
2.2.3 Framing as Interaction
2.2.4 The Notion of Frame Congruence
2.2.5 The Salience of Political Sophistication
2.2.6 Testing Foreign Policy Frames – So Far
2.3 The German Experience of Talking Intervention
2.3.1 The Changing Conception of German Foreign Policy
2.3.2 A Case in a Point: The Debate on the Lebanon Mission
2.3.3 Risk versus Responsibility Framing
2.3.4 Testing the Frame-Attitude Connection: Hypotheses
3 Data and Methods
3.1 Survey Design and Variables
3.1.1 Structure and Procedure of the Survey Experiment
3.1.2 First Part: Measuring of Independent Variables
3.1.3 Second Part: Newspaper Article
3.1.4 Third Part: Measuring of Dependent Variables
3.2 Data Screening and Sample Characteristics
3.2.1 Cleaning up the Act
3.2.2 Sample Characteristics
3.3 Statistical Analysis
3.3.1 Estimation Strategies
3.3.2 Analysis of Interactions
4.1 Findings for Support
4.1.1 Descriptive Statistics
4.1.2 Inferential Statistics
4.2 Findings for Overall Opinion
4.2.1 Descriptive Statistics
4.2.2 Inferential Statistics
4.3.1 Control Variables – in Brief
4.3.2 Risk Outperforming Responsibility
4.3.3 Surprises for Frame Congruence and Political Sophistication
4.3.4 The Competent Citizen?
4.3.5 Practical Relevance of the Findings
4.3.6 A Very Last Word on Theory
Appendix A: Computer Content Analysis
Appendix B: Student Survey and Article Versions
Appendix C: Principal Component Analysis
Appendix D: Latent Class Analysis
Appendix E: Data Screen and Sample
Appendix F: Linear Regression Analysis
Appendix G: Multinomial Logit Analysis
Diese Diplomarbeit entwickelt und prüft Hypothesen über den Einfluss von politischen Begründungsmustern auf individuelle Einstellungen zu Auslandseinsätzen der Bundeswehr. Das zentrale Argument der Arbeit lautet, dass solch ein Einfluss in der Tat feststellbar ist, seine Stärke jedoch maßgeblich von individuellen Merkmalen der Bürger abhängt.
Das praktische Interesse an dieser Untersuchung ergibt sich aus der zunehmenden Beteiligung Deutschlands an militärischen Stabilisierungs- und Friedensmissionen. Die Aversion der deutschen Bevölkerung gegenüber militärischen Todesfällen sowie eine zunehmende Medialisierung der Weltpolitik erfordern in der Abwesenheit eines inhärent „nationalen Interesses“ eine plausible Rechtfertigung von Interventionen und deren rhetorische Untermauerung durch politische Entscheidungsträger. Als exemplarischer Fall wird die Bundestagsdebatte zur Beteiligung an der Friedensmission im Libanon inhaltsanalytisch untersucht. Hierbei werden zwei diametral entgegengesetzte Argumentationslinien oder „Frames“ festgestellt. Der erste Frame betont vor allem die mit einem Einsatz verbundenen Risiken, wohingegen der zweite Frame die historische und moralische Verpflichtung einer deutschen Beteiligung hervorhebt.
Die Ergebnisse der Inhaltsanalyse bilden die Grundlage für die Konzipierung eines Survey-Experiments an dem Studierende der Universität Konstanz teilnehmen. Das Survey-Experiment besteht aus drei Teilen: Der erste Teil erhebt relevante individuelle Merkmale, insbesondere den Grad des spezifischen politischen Vorwissens sowie Umfang und Art politischer Prädispositionen. Der zweite Teil umfasst das eigentliche Treatment. Es besteht aus einem fiktiven Zeitungsartikel, der eine mögliche Ausweitung des Libanon-Einsatzes auf die syrische Grenze beschreibt. Dieses Außenpolitikszenario wird in drei sich jeweils nur geringfügig unterscheidenden Versionen entworfen. Eine Version betont die Verantwortung der deutschen Politik eine Mandatserweiterung zu bejahen (Verantwortungsframe). Die andere Version stellt eine deutsche Beteiligung als mit zu hohen Risiken verbunden dar (Risikoframe). Eine Kontrollversion bemüht sich um Neutralität und spricht sich für keines der beiden Argumente aus. Nachdem die Studierenden jeweils eine Artikelversion gelesen haben, werden sie im dritten Teil nach ihrer generellen Meinung und spezifischen Einstellungen zur Erweiterung des Bundeswehreinsatzes befragt.
Die Analyse und Auswertung der Antworten mit Hilfe von linearen und logistischen Regressionstechniken bestätigt die zentrale Annahme dieser Arbeit: Der Einfluss von politischer Argumentation kann nur unter Einbeziehung individueller Merkmale der Bürger vollständig nachvollzogen werden. Im Vergleich zur neutralen Version reduziert der Risikoframe die Zustimmung zum Bundeswehreinsatz, wohingegen der Verantwortungsframe wirkungslos bleibt. Der Einfluss des Risikoframes ist jedoch abhängig vom individuellen Vorwissen und den politischen Prädispositionen der Probanden. Je besser sie über deutsche Außenpolitik informiert sind, umso eher erweist sich der Risikoframe als wirkungslos. Inkongruenz zu politischen Prädispositionen erhöht hingegen die Wirkungsweise des Risikoframes. Darüber hinaus ergibt die Modellierung eines dreifachen Interaktionseffekts, dass die Wirksamkeit des Verantwortungsframes bei gegebener dispositiver Kongruenz positiv vom Grad des politischen Vorwissens abhängt.
All diese Effekte sind statistisch signifikant und verdeutlichen den Charakter politischen Überzeugungs- und Urteilsvermögens als Ergebnis eines Zusammenspiels von Kommunikationsinhalt und individueller Merkmale der Bürger. In inhaltlicher Hinsicht liefert die Diplomarbeit eine experimentelle Bestätigung der militärischen Opfer-Aversion in Deutschland und knüpft an die wachsende Literatur des „Body-Bag-Effect“ an. Jedoch zeigt sich, dass die Sichtweise einer manipulierbaren Bevölkerung zu vereinfachend ist, da sie unter bestimmten Umständen die ihr offerierten Begründungsmuster mit ihren Prädispositionen in Einklang bringen und sie als Heuristiken zur politischen Willensbildung benutzen kann. Indem diese Diplomarbeit das Framingkonzept vor dem Hintergrund einer interaktiven Sichtweise verfeinert, setzt sie in theoretischer und methodologischer Hinsicht jüngste Vorschläge der politischen Kommunikationsliteratur um.
List of Figures and Tables
Figure 3.1: Design of the Survey Experiment
Figure 3.2: Scree Plot of Eigenvalues and Component Numbers
Figure 3.3: Student Numbers for Revised Sample, Sorted for Sex and Affiliation
Figure 3.4: Political Sophistication Split by Faculties
Figure 3.5: Political Sophistication Split by Predisposition
Figure 4.1: Support as a Function of Framing and Political Sophistication
Figure 4.2: Support as a Function of Framing and Political Predisposition
Figure 4.3: Mean Differences for Different Levels of Political Sophistication
Figure 4.4: Mean Differences for Different Levels of Political Predisposition
Figure 4.5: Agreement as a Function of Framing and Political Sophistication
Figure 4.6: Agreement as a Function of Framing and Political Predisposition
Figure 4.7: Change in Odds to Agree if Exposed to Responsibility Framing for Different Levels of Sophistication and Predisposition
Table 2.1: Tabulation of Category Frequencies
Table 3.1: Rotated Component Matrix
Table 3.2: Latent Classes of Student Population
Table 4.1: Support as a Function of Framing, Predisposition, and Sophistication
Table 4.2: Results from OLS-Estimation for Support
Table 4.3: Results of the Hierarchical F-Test
Table 4.4: Agreement as a Function of Framing, Predisposition and Sophistication
Table 4.5: Results from Multinomial Logit Estimation for Overall Opinion
Table 4.6: Results of the Likelihood Ratio Test
Preface and Acknowledgements
Der Gedanke zu dieser Arbeit entstand im Juli 2006. Nach dem „Sommermärchen“ in Deutschland wütete der „Sommerkrieg“ zwischen Israel und dem Libanon. Die ganze Welt konnte im Internet oder im Fernsehen die verschiedenen Interpretationen des Kriegsgeschehens verfolgen. Nachdem die Waffen schwiegen ging es um die Frage ob sich die Deutschen an einer Friedensmission beteiligen sollten. Sowohl Gegner als auch Befürworter eines Engagements betonten die „besondere Verantwortung“ Deutschlands und eine öffentliche Debatte um die richtige Auslegung dieses Begriffes entbrannte. Ich fragte mich ob der rhetorische Schlagabtausch von Relevanz für die eigentliche Interventionsentscheidung sein könnte. Die Idee von „Talking Intervention“ war geboren.
Diese Arbeit wäre in dieser Form nicht möglich gewesen hätten mir nicht viele Menschen zur Seite gestanden. Erwähnen möchte ich vor allem die großzügige Bereitschaft von 16 Konstanzer Professoren und Professorinnen unterschiedlichster Fachbereiche mir jeweils 20 Minuten Ihrer wertvollen Vorlesungszeit zu gewähren: Anja Achtziger, Thomas Dekorsy, Frank Janning, Daniel Keim, Wilhelm Kempf, Rüdiger Klimecki, Alfred Leitenstorfer, Gerhard Müller, Jürgen Osterhammel, Alexander Prestel, Monika Reif-Hülser, Rudolf Rengier, Hartmut Riehle, Gerald Schneider, Marc Scholl und Gereon Wolters haben mit ihrem Entgegenkommen den interdisziplinären und unkonventionellen Charakter der Universität Konstanz unterstrichen. Insgesamt haben 860 Studenten am Survey-Experiment teilgenommen und damit fast zwei Wochen an Arbeitszeit erbracht. Meine beiden „Assistentinnen“ Babette Luckert and Michaela Rentl sorgten dafür, dass vor allem Naturwissenschaftler weniger konzentriert, aber dafür umso bereitwilliger die Fragebögen ausfüllten. Ohne Simon Liepold hätte ich 50 Kilogramm Versuchsmaterial nicht so problemlos über das Universitätsgelände bringen können (das nächste Mal doch lieber ein Online-Survey!). Rutger Hagen weihte mich in die Geheimnisse von SPSS ein. Sein „Survival Manual“ hat mir im wahrsten Sinne des Wortes das Leben mit Statistik leichter gemacht. Florian Kunze munterte mich 2 Wochen vor Abgabe mit Bildern von der Haute Route auf und machte Lust auf neue Abenteuer. Er war mir ein unerreichtes Vorbild im Bereich Zeitmanagment, als dass er seine Diplomarbeit einen Monat vor Abgabetermin fertig stellte. Martin Brunner erwies sich als wichtigster Berater und war der Ruhepol bei gemeinsamen Mensaausflügen. Muttersprachlerin Kathleen Claussen zeigte sich als ebenso kompetente wie strenge Korrekturleserin von Ihrer besten Seite. Ich bedanke mich für das Feedback des Konstanzer Kolloquiums Friedensforschung, vor allem bei Monika Spohrs und Jan Böhnke für die Hilfestellung zur Latent Class Analysis. Annika Olofsson gab mir während der Pausen im Campus-Café wertvolle Ratschläge für (inzwischen) fortgeschrittenere SPSS-Anwendungen. Rainer Schnell nahm sich die Zeit meinen Fragebogen kritisch zu würdigen und ein „könnte klappen“ zu murmeln. Willi Nagl machte mich auf einen unansehnlichen „Todesanzeigenrand“ aufmerksam und erklärte mir geduldig den Nichtunterschied von ANOVA und Regressionsanalyse. Der Konstanzer Polizeisportverein fungierte während eines Fußball-Länderspiels als nichtakademische Validierungsstichprobe, ebenso wie etliche Arbeitskollegen, Freunde und Bekannte meiner Eltern. Meiner Familie gilt ein besonderer Dank für Ihre Motivation und vielfältigen Beistand. Ich bedanke mich bei der Universität Konstanz und der Hans-Jäckh Stiftung für ihre finanzielle Unterstützung. Mein Dank gilt auch Kurt Hübner, der für mich im Mai 2005 an der York University ein Inhaltsanalyseprogramm anschaffen ließ, auf das ich nun wieder zurückgreifen konnte. Vor allem bedanke ich mich bei meinen beiden Betreuern Gerald Schneider und „Zwockel“ Wilhelm Kempf. Sie gaben mir wertvolle Tipps und schlugen wichtige Richtungskorrekturen vor. Ich entschuldige mich bei ihnen für knapp 115 Seiten Text und das Bombardement an Fußnoten. Da mir die Betreuer jeweils 90 Seiten, bzw. bis zu 150 Seiten zustanden, hoffe ich dass ich mit dem Erreichen des Durchschnitts auf großzügige Nachsicht des ersteren Gutachters stoße. Zu guter Letzt gebührt Kerstin Faßbender große Anerkennung. Sie hat die Arbeit in der gesamten Zeit begleitet, mich immer wieder aufgebaut, den Verzicht auf einen gemeinsamen Urlaub in Kauf genommen und mich bedingungslos unterstützt. Sie ist die Beste.
After the assassination of the soldier the political climate might change and the role of German troops abroad might be questioned again. However, it would be better if – however painful this process will be – the death of the sergeant would open the eyes for reality: If the Federal Republic does not want to claim a special position in the world, if it wants to keep on helping the community of nations with its medical soldiers, doctors and logistics troops, there will be more victims to mourn.
(Süddeutsche Zeitung, 16 October 1993)
In his autobiography and subsequent media campaign, former German Chancellor Gerhard Schröder remarked that during his chancellorship he had been kept from getting a good night’s sleep just three times: confronting the decision to call for federal re-elections in May 2005 and settling on German troop assignments in Kosovo and Afghanistan in 1999 and 2001 (Schröder 2007, 374). We cannot say (yet) if his successor is experiencing similar exceptions to peaceful nights, but as a matter of fact, Germany currently deploys roughly 7.300 soldiers spread over ten different crisis areas that range from neighbors such as Bosnia-Herzegovina to such distant places as Afghanistan (Bundeswehr 2007). In a November 18, 2006, article, The Economist notes that whether Germany “wants to be or not, the country is a Mittelmacht, or middle power” being in a “good position to take responsibility in cases where it can bring something to the table.”
The paradigm shift from self-appointed post-war reluctance to relatively unscrutinized engagement in multilateral military commitments would have been unthinkable ten years ago. Since the 1994 constitutional court’s decision on the Parlamentsvorbehalt, however, international conflict prevention and global crisis management have increasingly become a core function of the Bundeswehr, complementing its traditional mandate of maintaining national security and defense. Several authors argue that the country’s growing international involvement mirrors the “normalization” of its foreign policy and an emerging self-conception as a “grown-up” nation, and consider the embarking on the “German way” as a long-overdue and desirable development (e.g. Bahr 2003; Schöllgen 2003). Politically, the rush towards international commitment is primarily justified by humanitarian arguments and the need to react to challenges of an increasing interconnectedness in world affairs, believed to require new strategies of national security. Peter Struck, former German minister of defense, coined the telling phrase that Germany’s safety had to be defended at the Hindukush as well. The official lingo is “Networked Security”, declared as a key concept in the recently released White Paper on German security policy and the future of the Bundeswehr (Federal Ministry of Defense 2006).
Whether German security concerns are effectively tackled at the Hindukush or along the Lebanese waterside is debatable, but is not up for discussion in my thesis. Rather, I deal with the political discourse on those contested issues and their impact on individual attitudes. In fact, out-of-area-missions have frequently been coupled with significant levels of public debate. A current example is the prevailing disagreement about the surveillance mission of German Tornado war planes in Afghanistan, expectant the constitutional court’s rule on an Organstreit proceeding. Notwithstanding those controversies, the intense German foreign policy year 2006 saw overall public support of multilateral army assignments, standing at an impressive 81 percent according to a recent opinion poll (EMNID 2006).
A thorough understanding of the determinants of individual attitudes towards interventions may be crucial to understanding foreign policy decisions in democratic states since aggregate public opinion influences electoral choice (Aldrich, Sullivan, and Borgida 1989; Aldrich et al. 2006), public policy (Page and Shapiro 1992) and elite dispute (Entman 2004; Nacos, Shapiro, and Isernia 2000), among other things. In this thesis, I argue that, for the German case, domestic backing of multilateral foreign assignments has been temporarily won by discursive persuasion or specific foreign policy frames, making the public accept the need for intervention and peacekeeping. In general terms, a frame is a particular interpretation and problem definition of an issue (Gamson 1992). Foreign policy debates dealing with German interventionism can be understood as frame contests between politicians and other special interest groups (see Aldrich and Griffin 2003; Entman 2004; Norris 1998), all of them trying to get their favored message across. This competition for “issue ownership” (Petrocik 1996) usually takes place in the mass media, for individuals the main (and often only) source about foreign policy events (Entman 2000, 2004).
Frequently, interventionist talking is based on the concept of The Responsibility to Protect which claims to represent the emergence of a new understanding of state sovereignty, global security and international law (ICISS 2001). According to this principle, states are not only obliged to observe and protect human rights in their own domestic realm but also to guarantee human security in all the rest of the states. In the case of grave human rights violations (i.e. genocide) the international community ought to interfere, also using military force. Politicians increasingly make use of this rationale and legitimize forcible humanitarian interventions in terms of a genuine ethical commitment of the Western world. For instance, using strong moral imperatives (“Never again Auschwitz!”), former German foreign minister Joschka Fischer provided many citizens with a compelling legitimation for Germany’s participation in the NATO-led operation in Kosovo.
It is far from certain, however, if a moral justification of risk and cost-intensive military assignments will remain uncontested. Most important, public consensus usually dwindles when confronted with significant numbers of casualties. Combat casualties are central to the understanding of the opinion-policy nexus because the unwillingness to pay the costs of war is one of the central mechanisms by which public opinion might affect foreign policy. This “body-bag effect” is believed to have contributed to US withdrawals from Vietnam, Beirut and Somalia, for instance (see Robinson 2002). As a consequence, the riskier foreign endeavors happen to be, the more the rhetorical straitjacket justifying continuing participation in world affairs becomes important.
Therefore, even if the German public seems fairly comfortable with the current foreign policy approach at the moment, ongoing support for sending troops abroad is anything but certain. The countrywide uproar in reaction to the circulation of some unsavory skull pictures, taken in Afghanistan is a case in point. The clearly inappropriate conduct of some soldiers not only ruined the defense secretary’s simultaneous presentation of the White Paper but also generated widespread skepticism towards the army’s mandate in that area. In consequence, approval rates of the mission plummeted (see the November 20, 2006, edition of Der Spiegel). Most recently, a suicide bombing in Afghanistan left three German soldiers dead and several wounded, spawning an intense debate about the purpose of military assignments in general.
In this thesis, I aim to scrutinize the German discourse on intervention and provide some evidence of the impact of interventionist rhetoric on individual attitudes. I argue that the Cold War explanation of an ideological stalemate has been gradually displaced by a rationale conceiving world affairs as the joint responsibility of an international community. This reasoning, I believe, is also mirrored in the German debate on foreign assignments. A computer content analysis of a major plenary debate of the German Bundestag reveals that two themes or storylines achieved predominance in the foreign policy discourse. On the one hand, proponents of interventions frequently stress the country’s moral duty to engage in foreign commitments. In essence, the argument goes that global challenges call for global responsibility, obliging the once-so-reluctant Germany to act as well. On the other hand, opponents argue that interventionist endeavors are prone to unforeseeable risks, above all, the loss of human lives. They consequently conclude that foreign assignments would constitute an inappropriate foreign policy instrument, and recommend non-coercive options.
Throughout this thesis, I refer to the former rationale as responsibility framing, while I characterize the latter as risk framing. My central argument states that both frames can bear substantial influence on citizens’ beliefs and ideas about the usefulness and desirability of interventions. The strength of the framing effect, however, is conditionalized by individual-level factors, such as political knowledge and prior dispositions. That is to say, successful frames predispose some understanding and competence of the subject matter and must match the target audience’s already existing preference structure and basic political orientations in order to be effective. Traditional framing research neglects such considerations, often assuming that citizens are uniformly susceptible to framing attempts (Druckman 2004; Hiscox 2006; Sniderman and Theriault 2004). Citizens are usually found to be incapable of political judgment, blown from one side or the other of a subject matter depending on what kind of frame they are exposed to at the moment (Nelson 2004). In line with recent advancements in the political psychology and political communication literature, I propose that these results can largely be attributed to deficiencies in the predominant concept of framing and the prevalence of unsophisticated research designs. Consequently, I abandon in this thesis the classical perspective of framing effects, instead conceptualizing them as conditional and interactive constructs.
I present evidence for this assertion from a survey experiment, aimed at measuring university students’ sensitivity to foreign policy frames. The treatment manipulation consisted of a mock newspaper article integrated in a larger investigation on foreign policy attitudes. The article feigned a possible extension of the UNIFIL mandate to the Lebanese-Syrian border, indicating the deployment of German ground troops. Subjects either received a version emphasizing the overall riskiness of the mission (echoing the risk frame), a version stressing the country’s moral and historical duty to participate in the mission (reflecting the responsibility frame), or a neutral text version. A pre-test assessed the subjects’ foreign policy knowledge, as well as other theoretically relevant individual-level characteristics. The post-test measures captured the overall opinion and support of the Lebanon intervention. Results were subsequently analyzed using linear and logistic regression procedures.
The analysis of the results generates some counter-intuitive yet instructive insights. Consistent with expectations, effects for risk framing are highly significant and considerably reduce approval rates in both analyses. Its impact, however, is conditional on the level of political knowledge and congruence to prior dispositions. Surprisingly, political knowledge and congruent dispositions dampens the risk framing effect. Responsibility framing, in contrast, seems largely ineffective. The hierarchical modeling of higher-order terms reveals that responsibility framing nevertheless amplifies the odds to agree with the Lebanon intervention, given high knowledge levels and congruent predispositions.
Taken together, the thesis makes a substantive and theoretical contribution to the literature. First, the substantive issue concerns the experimental evidence of risk framing having a significant impact on peoples’ attitudes towards foreign assignments, thus confirming the literature on public causality shyness (see e.g. Aldrich et al. 2006; Boettcher and Cobb 2006). Second, and constituting in part the precondition for the substantive findings, I offer some theoretical and methodological improvements for framing research. The authentic simulation of a hypothetical intervention resembles much better “real” politics than does the artificial situation of answering one-sided survey questions. Most importantly, by explicitly modeling the conditional nature of framing effects, I accommodate for the fact that citizens are not equally susceptible to persuasion attempts. Indeed, some citizens might not be manipulated by frames but use them to reach conclusions that are consistent with their underlying principles. In short, this thesis’ findings demonstrate that the impact of interventionist talking can only be fully understood if applying more sophisticated and theory-driven research designs.
The thesis comprehends five sections which are organized as follows. Subsequent to this introduction, the second section deals with theoretical aspects of the thesis. I clarify my central argument and its underlying rationale, work on major shortfalls of the classical “framing as manipulation” literature, and offer an alternative theoretical framework explaining framing effects as the product of both message content and personal attributes. I present some preliminary finding for the German discourse on intervention and derive my research hypotheses. The third section describes the methodology. I explain the design and variables of the survey experiment and report sample characteristics and results of the initial data-screening procedure. Next, I elucidate the logic of interaction analysis and substantiate the chosen estimation strategies. The fourth section presents findings from regression analyses and hierarchical testing procedures and discusses them with respect to their practical relevance and broader generalizibility. The thesis concludes with a summary of central insights and caveats, and gives some recommendations for future research.
Sie machen sich keine Vorstellung, welch ein ungeheurer Druck auf der Regierung lastet, wenn irgendwo auf der Welt eine humanitäre Krise ausbricht, und zwar Druck insofern, als gefragt wird: „Und was macht die deutsche Bundesregierung? Warum seid ihr nicht schon längst da unten und helft?“ (Michael Gerdts)
Changing the terms of the debate from ‘right to intervene’ to ‘responsibility to protect’ helps to shift the focus of discussion where it belongs – on the requirements of those who need or seek assistance. (ICISS 2001, 18)
2 Theory and Preliminary Analysis
This section has three goals. First, I outline the inspiration and underlying assumptions of my general argument and show why and how political discourse on intervention is relevant for foreign policy-making. Second, I describe major premises of the standard account of framing and highlight important deficiencies. An alternative approach is suggested, emphasizing the conditional character of framing effects. Third, I flesh out the notion of competitive framing by investigating the German Bundestag debate on the Lebanon mission and deduce hypotheses about the impact of interventionist talking and its counter-rationale on individual attitudes.
Note that I deliberately depart from the common practice to dedicate a separate section to a literature review. The character of this thesis is rather unusual insofar as the substantive argument (i.e. the German discourse on intervention influences individual attitudes towards foreign assignments) comes with a theoretical and methodological argument (i.e. the framing literature has to resort to an interactionist perspective and make use of more sophisticated research designs). In my view, the in this chapter offered tripartite breakdown in the background of the general argument, motivation for the theoretical/methodological refinement, and explication of the hypotheses is the most feasible and intelligible approach. Of course, I incorporate relevant literature throughout the section.
2.1 The Rationale of Talking Intervention
Making sense of foreign affairs used to be simple with the Cold War paradigm providing ideological schemata and a powerful national-interest frame for policymakers and the public alike (Entman 2004; Nacos, Shapiro, and Isernia 2000; Risse 2000). The unexpected and sudden collapse of the Soviet Union, however, eviscerated the demonization of communism and other familiar clichés, and opened the space for new interpretations of global politics. The mounting complexities of regional conflicts, multilateral peace-keeping missions and other kind of interventions demanded a convincing and comprehensive substitute for the traditional tenets of realpolitik.
I suggest that in the absence of palpable national interests or security concerns, democratic decision-makers increasingly resorted to the power of value-based language. In order to justify foreign assignments and tackle public concerns and resistance, they have legitimized interventions by evoking cosmopolitan and humanitarian principles. Referring to this thesis’ title, political leaders “talk intervention” which I conceptualize as the intentional creation and dissemination of rhetorical justifications of a particular humanitarian military deployment.
The central objective of this thesis is to assess the impact of interventionist talking and its rhetorical refutation on individual attitudes towards foreign assignments. In that way, I aim to provide a microfoundation of the broader argument that political discourse can exert some influence on a country’s foreign policy decisions. Instead of viewing framing as the manipulation of large parts of the population, I argue that the impact of framing is very much dependent on the characteristics of the audience itself. Before I turn to the individual level of analysis, however, I lay out in this Chapter 2.1 the broader rationale or interventionist talking, claiming that political discourse on intervention can be understood as a contest between two opposing frames.
2.1.1 The Discourse on Intervention as Frame Contest
I substantiate and objectify the contention that discourse matters by means of the framing concept, borrowed from the psychological and political communication literature. Framing refers to the definition and remedy of a decision problem and mirrors the strategic construction of political discourse. To be precise, I employ a classical framing definition put forward by Entman: “To frame is to select some aspects of a perceived reality and make them more salient in a communicating text, in such a way as to promote a particular problem definition, causal interpretation, moral evaluation and/or treatment recommendation” (1993, 52). This definition satisfies the political debate on foreign assignments as it captures changes in the problem definition (we witness a humanitarian disaster), causal interpretation (something or somebody can be blamed), moral evaluation (we are responsible to act), and/or treatment recommendation (we ought to intervene).
In most democracies political debates about military assignments can be conceived as a frame contest where proponents and adversaries of intervention battle for audience attention by disseminating two alternative narratives or moral evaluations of a particular military assignment. On the one hand, proponents of interventions highlight the moral obligation to act on behalf of the “international community” of “civilized” states to receive public support. The invocation of moral values is a powerful weapon in the persuader’s arsenal and has been found to constitute a strong and pervasive factor on public opinion (Zaller, 1992, 23) and provide recipients an important cue to attach their general attitudes and beliefs to a more specific policy (i.e. intervention). On the other hand, adversaries of intervention will put together an alternative narrative that potentially possesses as much strength and resonance with the target audience as the humanitarian frame. Instead of appealing to cosmopolitan principles, they will emphasize specific risks and uncertainties associated with foreign assignments, aimed at intensifying public skepticism and concerns. This is especially achieved by rhetorically disseminating negative information or negative emotions as they attract more attention than objectively equivalent positive information or positive emotions (e.g. Marcus 2003; Marcus, Neuman, and MacKuen 2000; Soroka 2006).
The central goal of the frame contest is simply to generate support or opposition for a specific foreign policy decision (see Entman 2004, 47). Two factors are central to understanding the impact of framing on foreign policy making. The first factor relates to the so-called body-bag effect which assumes a negative relationship between growing numbers of casualties and public support of interventions. The second factor concerns the role of the mass media. The dissemination of global news and real-time coverage amplifies the public’s attention towards casualties, somewhat multiplying the body-bag effect. Yet, at the same time the media works in the opposite direction by confronting the public with vibrant pictures of the suffering, in turn producing humanitarian sentiments. Taken together, growing media influence and looming public casualty shyness increase the need for, and potential gain from, framing and interventionist talking in particular.
2.1.2 Framing and the Banality of the Body-Bag Hypothesis
Public opinion was once believed to have little if any impact on the foreign policy process. The “Almond-Lippmann consensus” maintained that public opinion was volatile, so poorly structured as to constitute “non-attitudes” (Converse 1964), and, in the final analysis became largely irrelevant to the conduct of foreign affairs (Holsti 1996). Almond famously summarized this conventional wisdom by stating that foreign policy leaders have “to shout loudly to be heard even a little” (1950, 71).
The sustained popular opposition to the Vietnam War, however, noticeably contradicted the image of the public as unpredictable and irrational. Over the following years, a series of important studies (Milstein 1969, 1973, 1974; Mueller 1971, 1973; Verba et al. 1967) found that public support for U.S. military commitment dropped in proportion to the increase in casualties, indicating the presence of a fairly stable reluctance to pay the human cost of war.
The contention of a causal link between rising military casualties and declining public opinion support is known as the body-bag hypothesis. Casualty shyness is said to be a significant feature of post-Cold War politics in most democracies, severely restricting the conduct of foreign affairs (Holsti 2000, 215). It draws on the notion that favorable public opinion is an important resource that political leaders rely on for achieving their political goals, such as staying in office (Downs 1957). Given that domestic audience costs jeopardize a leader’s political survival (Fearon 1994) casualty reluctance serves as an important constraint to engage in powerful military interventions (i.e. the use of ground forces). The body-bag hypothesis is also mirrored in structural theories of the democratic peace that contend that democracies are less likely to engage in wars because of the public’s aversion to military conflicts (Doyle 1986; Maoz and Russett 1992). The findings on public casualty shyness have served as the watershed for a reconsideration of the opinion-policy nexus. By now, the evidence that public opinion is relevant to foreign affairs is quite robust (see e.g. Aldrich at al. 2006; Holsti 1996; Nacos, Shapiro, and Isernia 2000; Page and Shapiro 1992; Russett 1990).
The assumption of a mechanistic and negative public reaction to casualties certainly is too simplistic, however. Recent research has put forward a more nuanced view and argues that the public’s response to operations that suffer military casualties is not automatic but context dependent (Klarevas 2002). For instance, Larson (1996, 2000) contends that public casualty tolerance closely follows casualty tolerance of domestic elites. Kull and colleagues, in contrast, find that public support for a military mission is more robust if the public sees that other countries also support the mission (Kull, Destler, and Ramsey 1997; Kull and Destler 1999; Kull and Ramsey 2000). Nincic (1992) ascertains that there are far more instances in which casualties and citizen support appear uncorrelated. Grounded on individual-level data, research by Feaver and Gelpi (2004) suggests that casualty aversion is largely a product of expectations of success and judgments about the “rightness” of the decision to go to war. Eichenberg (2005) reaches a similar conclusion in an analysis of aggregate public support for U.S. military operations. Boettcher and Cobb (2006) demonstrate that the negative effect of information about American casualties is softened by the provision of causality ratios, i.e. placing the American casualties in relation to body counts of enemy dead.
This stream of research demonstrates that the prima facie compelling evidence for the body-bag effect is largely dependent on context. The public may be sensitive but not intolerant to casualties. Jentleson (1992, 53) suggests that casualty tolerance is based on “the principal policy objective” envisioned by the military operation. He distinguishes three such purposes: first, foreign policy restraint, that is, “to coerce […] an adversary engaged in aggressive actions against the United States or its interests”; second, internal political change, i.e. to influence “the domestic political authority structure of another state”; and third, humanitarian intervention, the “provision of emergency relief […] to people suffering from famine or other gross and widespread humanitarian disasters” (Jentleson and Britton 1998, 399). Jentleson shows that the impact of these objectives is more important than other factors such as the risk of casualties, the multilateral character of the mission, and the level of perceived interests. Moreover, he discovers that public support is higher for purposes of foreign policy restraint and humanitarian intervention than for purposes of internal political change (1998).
It is quite evident that in order to evaluate the magnitude of the body-bag effect and the mediating role of framing one has to analytically incorporate a mission’s perceived principal policy objective and its expected success. Switching outlooks regarding the actual purpose and effectiveness of an intervention is likely to affect casualty shyness and approval of the mission (though special in many aspects, Iraq is a case in point). For instance, Jakobsen (1996) finds that interventions are driven by a combination of the CNN effect (see below) and good chances of success. Investigating the peacekeeping missions in Lebanon and Somalia, Burk (1999) discovers that public support starts to decline when the missions are increasingly perceived as interference into the internal affairs of another state (see Bardes 1997 for a similar interpretation on Somalia and Haiti). Popular support of a certain intervention appears to be largely dependent on the subjective appraisal of the intervention’s very purpose.
In this context, framing becomes crucial as it provides the public a “mental recipe” (Nelson and Kinder 1996, 1073) for preparing a judgment about the nature of a particular military commitment. Political leaders will try to steer these perceptions and engage in competitive framing to either neutralize or to take advantage of the impact of casualty shyness. The outcome of the frame contest determines if the logic of the body-bag effect will turn into a germane factor in the decision to intervene. If “talking intervention” comes out on top, casualty shyness is unlikely to be an issue or will be attenuated at least. As soon as its rhetorical counterpart prevails, however, the body-bag effect will become an important domestic constraint for the independent conduct of foreign affairs.
2.1.3 Rethinking the Role of the Media
The second factor that is relevant for the understanding of the frame contest is the role of the mass media. Not surprisingly, the public relies on the media as the principal source for information on foreign policy events as these events are beyond the realm of personal experience for most individuals, constituting “high-threshold issues” (Lang and Lang 1981). There is some disagreement, however, on how to understand the role of the media in the opinion-policy nexus and in international politics more generally. One group of scholars conceptualizes the media as an active and independent actor in the formulation of global politics, asserting that the advent of real-time television coverage has both heightened public casualty reluctance and humanitarian claims by giving the deaths a vividness and immediacy that makes them more shocking (Livingston 1997; Neuman 1996). Most famously, the hypothesis of the CNN effect has suggested that global television has become an important or even dominant factor in the formulation of foreign policies (Robinson 2002). In contrast, a second group perceives the media as the passive abettor of the state and dominant interests. The manufacturing consent literature (Herman and Chomsky 2002) has argued that powerful economic and political elites use the media to mobilize public support for governmental policies. The indexing hypothesis suggests that reporters index their coverage to reflect the range of opinion within the government or elites (Bennett 1990).
There are some shortcomings with these approaches to media influence. First, both perspectives comprehend the media as a single and coherent actor operating alongside states and nonstate actors. Clearly, this view is deficient in its understanding of political complexity. Second, most studies are subject to a selection bias as they restrict analysis to media coverage of interventions that actually took place, therefore neglecting instances of non-occurrences (Robinson 2002; Seybolt 2007). Third, scholars often rely on case studies that apply different forms of content analyses to scrutinize the relationship between media coverage, public opinion and foreign policy decisions. Although correlational analyses can make useful contributions to the understanding of the media’s role in the opinion-policy nexus, they are not an entirely satisfactory substitute for the detection of the causal mechanisms underlying such relationships (see Geddes 2003; McDermott 2002).
Recently, researchers have begun to address the first and second deficiency of the literature (conceiving the media as monolithic actor; omitting counterfactuals) and have developed more sophisticated theories of media influence that are informed by framing theory. Robinson’s (2002, chap. 2) “Policy-Media Interaction Model” predicts that under conditions of policy uncertainty and critical and empathy-framed coverage, the media can be an important factor in convincing political leaders to pursue humanitarian objectives. Case studies of humanitarian crises in Kosovo, Rwanda, Bosnia and Iraq support his hypotheses, yet reveal that media coverage is not a major factor for ground troop intervention. Entman (2004) provides an integrative theory of the triangular relationship between the media, the public and the U.S. government (“Cascading Activation Model”). He suggests that several actors engage in combat to shape frames that reach the public through the media and influence public opinion formation. The model argues that some actors have more power than others to push frames down the road to the public and, therefore, could assist in identifying when and how the media affects foreign policy-making (see Gilboa 2006).
Without doubt, the mass media has gained importance: It constrains political leaders, yet at the same time provides them with opportunities to mobilize public support to advance their goals. Entman (2004) convincingly demonstrates that media coverage increasingly represents a powerful vehicle for political framing effects. Trained as political scientists, however, Entman (2004) as well as Robinson (2002) fail to decisively depart from classical approaches to media influence as both continue to conceive global news networks as some kind of genuine actor in international politics and foreign policy making.
I suggest that in the context of political debates on foreign interventions it might be useful to conceptualize the media not as an actor but as the forum of the framing contest and to shift attention to the characteristics of frames themselves. The frame that prevails in the framing contest and manages to dominate the interpretation of a foreign policy issue or the principal policy objective of an intervention is more likely to affect public opinion and ultimately foreign policy itself. The central question, then, is not only which actors use which frames (as Entman 2004 asks), but also what explains the overall persuasiveness of a successful frame. This moves the analytical focus to a microfoundation of framing effects, addressing the third deficiency of the literature on media influence.
2.1.4 The Need for a Psychology of Discourse
In the preceding paragraphs I have argued that the increasing importance of the global media along with looming public casualty shyness have heightened the importance of political discourse to legitimize foreign interventions, especially when no clear national interest is at stake. I substantiated the somewhat abstract notion of political discourse by means of the framing concept, introducing the idea of a rhetorical competition between proponents and adversaries of military commitments. Most importantly, I assumed that frames can exert some influence on public opinion towards interventions.
Clearly, the applied logic in this chapter relates to macro-effects of discourse, that is, effects of discourse at the societal level (see Johnson-Cartee 2004, chap. 1). Those macro-effects, however, must be understood as aggregations of effects at the micro-level, that is, effects related to the individual. In this view, political discourse can be understood as an important source of over-time change in individuals’ foreign policy preferences and attitudes. The emphasis on the individual as the unit of analysis is useful as it allows a much better analysis of the impact of interventionist talking and tackles the important questions of whether and, if so, why some individuals are more sensitive to framing attempts. Therefore, the next chapter shifts from the macro-level to the micro-level of analysis and elaborates the framing concept in more detail. Echoing Almond’s remark, I will show that foreign policy leaders have not to “shout loudly to be heard only a little”, but just need to speak clearly and in familiar terms.
2.2 Towards an Interactionist Perspective of Framing
Chapter 2.1 has introduced the general background of my argument and has argued that to definitely confirm the impact of political discourse on public opinion formation one has to shift the focus from the aggregate to the individual level of analysis. Chapter 2.2 satisfies this requirement in that it explicates the psychological foundations of the framing concept and highlights some important shortcomings of the standard account of framing.
2.2.1 Framing – Still a Messy Paradigm
In 1993, Robert Entman bemoaned the lack of conceptual clarity and consistency had led to a “fractured paradigm” of framing, inhibiting the promising development of a new theory of mass communication. Other scholars stated that there appeared to be almost as many definitions of “frames” and ideas regarding how they influence public opinion as there were researchers using the term (Nelson, Oxley, and Clawson 1997). Even worse, researchers on agenda-setting (McCombs, Shaw, and Weaver 1997) disputed the conceptual autonomy of framing, viewing it as a mere extension of the agenda-setting process.
Today, working within the fractured paradigm appears to be a great success, at least judged by the vast and increasing quantity of framing studies. A Web of Science keyword search for “framing”, for instance, yielded over 500 publications for the year 2006, compared to just 200 studies in 1998. This is due to the fact that scholars of social movements, bargaining behavior, decision making, media and communications effects, political psychology, public opinion and many others increasingly make use of the framing concept (see Caragee and Roefs 2004). However, there is rarely a precise and consistent analytical categorization of the concept’s usage (Kühberger 1998; Levin, Schneider, and Gaeth 1998). Disciplinary synonyms like schema, script, heuristic, image, and so forth abound (see Dahinden 2006, chap. 2). Nearly 15 years after Entman’s famous assertion it thus still seems correct to state that framing has a scattered conceptualization, missing a unifying theory or framework to organize a wide variety of results in framing research.
Given the precocious state of the art (see Sniderman and Theriault 2004 for a similar assessment), I provide a concise definition and treatment of the subject matter and a brief clarification of what I believe are major inconsistencies and shortfalls of the literature. To begin with, there is some ambiguity with respect to communication frames and individual frames. Some scholars use the terms “frame” and “framing” when referring to the portrayal or interpretation of an issue in a communication source, whereas others use the term to refer to an individual’s understanding and perception of the issue. Druckman proposes to distinguish between “frames in communication” and “frames in thought” (2001a, 227-8). This suggestion resembles Kinder and Sanders’ (1996, 164) comparison between frames “embedded in political discourse” and frames that are “internal structures of the mind” as well as Scheufele’s (1999) discussion of “media frames” and “individual frames”. Without doubt, frames in communication play a central role in shaping frames in thought. As I will show below, the effect of frames in communication, however, cannot be adequately understood without considering prior political knowledge and predispositions, that is, frames in thought. In fact, this assertion is the thesis’ central argument and constitutes the very essence of framing.
Second, another crucial suggestion relates to the analytical distinction between equivalency framing effect and emphasis framing effect (Druckman 2001a). The former term describes an effect that occurs when different, but logically equivalent phrases cause individuals to alter their preferences. Applications in prospect theory can be subsumed under this category (see McDermott 1998, 2004; Tversky and Kahneman 1981). In contrast to equivalency framing effects, emphasis framing effects refer to a situation where, “by emphasizing a subset of potentially relevant considerations” a communication source leads individuals “to focus on these considerations when constructing their opinions” (Druckman 2004, 672). Hence, emphasis frames focus on different yet potentially relevant considerations of an issue. It is evident that political choices usually involve a deliberative process where it is difficult to assume ex ante equivalent gains or losses of a particular course of action (Sniderman and Theriault 2004, 136). Referring in those contexts to the relatively strict concept of equivalency framing seems rather inappropriate (see O’Neill 2001 for other important problems of equivalency frame applications in the field of international relations). The more relaxed conception of emphasis framing appears to be the appropriate concept for research on media influence, political persuasion and public opinion formation. The definition suggested by Entman (see Chapter 2.1.1) implicitly corresponds in its understanding to emphasis framing.
Third, there is no consensus on how to understand the underlying causal mechanism of framing (Johnson-Cartee 2004, chap. 1; Nelson, Clawson, and Oxley 1997). As I have argued in the preceding chapter, an understanding of effects at the macro-level (e.g. a change in public opinion towards a particular intervention) requires a thorough explanation of micro-level processes. The disagreement about the very nature of these processes, I argue, constitutes a major obstacle to the emergence of a more coherent conception of framing and the advancement of the field of political communication more general.
In the following, I devise a dichotomous classification of the literature, taking into account the clarifications I just made, in particular the distinction between “frames in communication” and “frames in thought”, and delineate what I believe it the standard position of framing. This view regards framing as the making of particular considerations or beliefs temporarily more accessible and comprehends the public as a rather passive and manipulable audience. In contrast, an opposing position interprets framing as the weighting of already existent beliefs, claiming that a thoughtful public may use frames to reach political judgments that are consistent with their underlying principles. I contend that the second position captures much better the nature of political debate and interventionist talking in particular.
2.2.2 Framing as Manipulation
The standard interpretation of framing builds on basic research in cognition and emphasizes automatic information processing (Cappella and Jamieson 1997; Price and Tewksbury 1997). More specifically, it is grounded in the work on survey response and attitude change by John Zaller. Zaller (1992, 41) presumes that the essential elements of opinion are considerations, defined to include any reasons for favoring one side of a dispute over another. In contrast to Converse (1964) who believes that large numbers of the public do not hold any considerations of major issues, Zaller proposes that most people possess opposing considerations on most issues all the time. Therefore, the problem is not that people have too few ideas to draw on, but that they have too many and contradictory ones (Zaller 1992, chap. 2; see also Zaller and Feldman 1992).
Zaller argues that in order to evaluate a certain argument people pick up from their memory the information that is most easy to get to at a particular moment. For instance, when they have to answer a survey question, these answers reflect the considerations that happen to be salient to the person at the moment the question is posed. In this perspective, framing occurs by simply manipulating the accessibility of considerations (Cappella and Jamieson 1997; Sniderman, Brody, and Tetlock 1991). Considerations that have been primed by recent mention by a communication source will be at the “top of the head” (Taylor and Fiske 1978), that is, easily accessed from memory, and will therefore be more likely to influence opinion than unprimed, inaccessible considerations. This idea of framing effects is also known as the passive-receiver model, because it assumes that people respond to framing more or less indiscriminately (see Brewer 2001). In fact, when viewing framing as the result of a largely unconscious accessibility process, framing and priming become conceptually indistinguishable.
The idea that the majority of persons on the majority of issues are ambivalent -coupled with the logic of the framing-as-priming hypothesis- constitutes the standard account of framing. Zaller states that this notion is “extremely well supported both by experimental psychology and by research on political attitudes” (1992, 276). In fact, when a study tries to demonstrate a framing effect, it usually succeeds, that is, the public is found to be moved from one side of an issue to the other depending on how the issue is framed (Sniderman and Theriault 2004; see Druckman 2004 for a possible publication bias).
I refer to this “classical” view as the framing-as-manipulation perspective . Ironically, this scenario reminds us of the Almond-Lippmann consensus where public opinion was viewed as unstable, ignorant and unimportant for the conduct of foreign affairs (Holsti 1996). As has been shown, political realities (e.g. in form of popular opposition to the Vietnam War) clearly contradicted this conventional wisdom. By the same token, the underpinning presumptions of the manipulation perspective are at odds with the complexities of human beings and politics itself.
2.2.3 Framing as Interaction
The standard account of framing implies that citizens (1) base their political preferences on arbitrary information that lacks a secure anchoring in underlying, enduring convictions, and (2) that framing works through a passive accessibility process rendering citizens’ political judgment prone to manipulation. This interpretation has recently been criticized (e.g. Chong and Druckman 2007; Druckman 2001, 2001a, 2004; Druckman and Nelson 2003; Nelson 2004; Shen 2004; Shen and Edwards 2005; Sniderman and Theriault 2004). In line with this development, I argue that the major alternative to accessibility models of framing entails a more deliberate and self-conscious process with the citizen as the “final arbiter, who chooses which of the available considerations are relevant and who decides how important each consideration should be” (Kinder 2003, 378; emphasis in the original).
Proponents of this view refer to several deficiencies of the standard account of framing. One major deficiency of the manipulation approach is the neglect of political realities. In democracies, political debates are characterized by a clash of arguments that expose citizens to different suggestions how to think about a problem and its remedy. Yet, classical framing studies have tended to overlook the fact of competitive framing (Chong and Druckman 2007). They have restricted attention to experimental designs where citizens are artificially exposed to only one way of thinking about a political issue and not to contesting views of the issue at stake. Sniderman and Theriault illustrate this point in stating that “framing studies have neglected the fact that frames are themselves contestable” (2004, 141-142).
Most important, the assumption that framing has a uniform and unconditioned effect for all individuals is doubtful. Several studies have in fact found that framing effects are not straightforward product of accessibility (Nelson, Clawson, and Oxley 1997). Rather, it appears that framing effects may depend on how favorably citizens respond to a particular argument. They may engage in some deliberate processing in response to a particular frame, responding readily to frames that produce positive reactions and rejecting frames that produce negative reactions (Druckman 2001). Evidence of such active processing on the part of citizens suggests that they sometimes act as thoughtful, rather than passive, receivers of information (Brewer 2001). It also suggests that framing is something different than priming. Through framing, political leaders seek to establish a dominant definition and interpretation of an issue, whereas priming refers to the accessibility of information at a particular moment. Whereas priming says only, “Think of this consideration,” framing says, “Think of it this way” (Aldrich et al. 2006, 486).
Accordingly, Nelson and colleagues argue that changes in the accessibility of considerations are not a necessary mediator of framing effects (Nelson, Clawson, and Oxley 1997; Nelson, Oxley, and Clawson 1997; Nelson and Oxley 1999; Nelson 2004; Nelson and Garst 2005). Framing might instead operate through the accentuation of already existing beliefs without necessarily introducing new information. In this sense, frames tell people how to weigh competing values, beliefs and emotional attachments and influence opinion by suggesting which of the many and possibly conflicting consideration should predominate. Considerations that are more important to individuals are more likely to be stored in long-term memory and to be chronically accessible. It is evident that in this perspective responsiveness to framing should be related to individual-level characteristics and that these characteristics might serve as important constraints to the manipulation by framing.
More sophisticated studies have provided evidence that citizens can resist a frame under a variety of circumstances, e.g., when they do not trust its source (Druckman 2001), when they receive two different frames (Sniderman and Theriault 2004) or when they are allowed to discuss the issue with other people exposed to an alternative frame (Druckman and Nelson 2003). In addition, other studies have incorporated individual-level characteristics such as education (Hiscox 2006) or motivation (Chong and Druckman 2007) and demonstrated that those characteristics can have a profound impact on the overall effectiveness of framing. What is common to all these studies is that they refrain from regarding framing effects as the result of the unconditional imposition of some political will on a gullible public. Instead, they are interpreted as the joint product of message content, political context and receiver characteristics (Brewer 2001, Druckman 2001, Nelson 2004). I refer to this view as the framing-as-interaction perspective, claiming that this account presents the conceptual antipode to the manipulation approach.
In this thesis, I restrict myself to potential individual-level moderators and do not deal with contextual moderators, such as source credibility (Druckman 2001) or interpersonal communication (Druckman and Nelson 2003). Though not testing political context explicitly, I mirror “realistic” political debate, in that all treatment versions in the survey experiment encompass alternative considerations about the potential usefulness and dangers of foreign assignments (see Chapter 3.1.3). With regard to individual-level characteristics, I suggest that framing effects will depend on (1) the degree to which individuals possess already existent dispositions that match the cues inherent in the frame and (2) the degree to which individuals are able to recognize those cues and connect them to their dispositions. I explicate these assumptions below.
2.2.4 The Notion of Frame Congruence
The first individual-level characteristic I focus on is the individuals’ political predisposition. Political predisposition can be broadly understood as any favorable attitude towards a particular political communication (Zaller 1992, 28). In the framing-as-interaction perspective, the impact of any frame should depend on how the frame interacts with such predispositions. In the case of favorable or positive predispositions it is more likely that the framing invokes the intended attitude change than in the case of unfavorable or negative predispositions (Popkin 1994; Sniderman, Brody, and Tetlock 1991). Quite simply, in order to exert influence, “frames in communication” must resonate with “frames in thought”. I refer to this general notion as frame congruence: A frame must match the political predispositions of its target audience in order to become effective. This also explains why successful frames frequently invoke core values: Because citizens already endorse those values and therefore might attach higher attention to the frame (see Chapter 2.3).
The idea of framing effects depending on the degree of the frame-predisposition match takes after cognitive consistency theories. These theories explain how people organize their opinion to avoid inconsistencies that are presumed to be psychologically painful (O’Keefe 2002, chap. 4). Of the various theories proposing a motive to maintain cognitive consistency the most prominent is the theory of cognitive dissonance (Festinger 1957). Festinger proposes that the psychological state of dissonance is aversive and that people will be motivated to engage in cognitive activity in order to reduce it. One solution to dissonance is to minimize the importance of one of the elements that causes dissonance (O’Keefe 2002, 79; see also Simon, Greenberg, and Brehm 1995). This nicely resembles the suggestion of Nelson and colleagues who conceive framing as the deliberate weighting of chronically accessible beliefs.
2.2.5 The Salience of Political Sophistication
The second important individual-level moderator is the amount of political information an individual possesses, also known as the person’s political awareness, expertise, knowledge or involvement (Zaller 1992, Krosnick and Bannon 1993, Popkin and Dimock 2000). Notwithstanding operational differences, I unanimously refer to these constructs as political sophistication. Political sophistication can be roughly defined as the quantity and organization of a person’s political cognitions towards an object (Luskin 1987). To put it differently, political sophistication relates to the prior familiarity with a political issue.
Finding that political sophistication makes a difference is a “standard result” in interactionist framing studies (Sniderman and Bullock 2004, 347). The direction of this influence, however, remains unclear. Some studies find that frames have a greater impact on less sophisticated people (e.g. Haider-Markel and Joslyn 2001). Echoing the tenets of the manipulation approach, these authors argue that people with low political sophistication possess fewer strongly held attitudes and thus exhibit increased susceptibility to framing (i.e. priming). In contrast, highly politically sophisticated individuals are believed to be less manipulable because they possess a broad range of political attitudes that can serve as counterweights when confronted with contesting arguments. In contrast, Nelson, Clawson, and Oxley (1997) find that framing is much stronger among respondents who are familiar with the argument. As explained in Chapter 2.2.3, this should be due to the fact that framing works through the weighting of different considerations held in long-term memory and the assumption that only better-informed individuals will store such relevant considerations. Thus, sophistication is not believed to serve as a counterweight to manipulation attempts but as an indispensable prerequisite for the reception of political messages.
Zaller (1992) demonstrates that the two views are not mutually exclusive. In his Reception-Acceptance-Model (a kind of subtype of the more general RAS-Model) he conceptualizes attitude change as a two-step process (1992, chap. 7). The Reception Axiom states that in a first step people’s likelihood to receive a frame is a direct function of their general level of political sophistication. Given the reception of a message, the Acceptance Axiom claims that in a second step people resist information that is inconsistent with their predispositions, but only if they are sufficiently sophisticated to notice such inconsistencies. As a consequence, the most sophisticated are at the same time more likely to receive new political information and less likely to be changed by such information. Zaller (1992, 124) suggests that people who are moderately sophisticated are the most likely to change their beliefs (see also Converse 1964). These peoples’ beliefs are thought to be more susceptible to change because they are more likely to receive new information than less sophisticated persons and are more likely to be persuaded by new information than more sophisticated persons.
In brief, theoretical explanations and empirical results for the impact of political sophistication are mixed. Zaller’s “Reception-Acceptance Model” makes plain that the issue of political sophistication is complicated at best, and likely to be intertwined with the concept of frame congruence.
2.2.6 Testing Foreign Policy Frames – So Far
The suggested breakdown of the framing literature into two broad perspectives is largely informed by studies dealing with attitudes towards domestic (that is, mostly U.S. American) policy issues. Recently, however, scholars have conducted framing experiments that explicitly addressed foreign policy issues and to some extent incorporated recommendations of the interactionist approach to framing.
In the realm of trade policy, for instance, Hiscox (2006) assesses the impact of framing on individuals’ attitudes towards globalization and international trade. He finds that individuals that are given a pro-trade introduction in a survey question are not more likely to express support for trade than those who receives no introduction. In contrast, when given an anti-trade introduction, individuals’ support significantly drops. Hiscox controls for political sophistication which is measured by education levels. It turns out that less educated individuals are more sensitive to framing than highly educated individuals. A disadvantage of his study relates to the fact that his treatment condition is made up of a very brief and one-sided paragraph, thus disregarding genuine political debate on globalization.
Other studies have made use of adopting different versions of news articles. For instance, one study discovers that depicting a foreign country as competitor to the United States leads to less favorable opinions regarding this country, whereas depicting common interest of both countries leads to more favorable opinions (Brewer 2006). The study sticks to the standard account of framing, as it does not control for possible interactions between framing condition and individual-level characteristics. Schuck and de Vreese (2006) examine the impact of a risk and opportunity frame on individual support for EU enlargement. Participants in the opportunity frame condition show higher levels of support compared to those in the risk frame condition. Schuck and de Vreese check for the moderating impact of political sophistication and find it to exert an attenuating influence on the framing effect, that is, individuals with low levels of political sophistication are more likely to be affected by framing. Most relevant for my thesis, Berinsky and Kinder (2006) present individuals with news reports on the 1999 Kosovo crisis, framed either to promote or prevent U.S. intervention. They show that “seemingly subtle differences in the presentation of identical information in the news media can affect the organization and recall of information and ultimately influence political judgments” (Ibid 640). They find no evidence for interactions between framing conditions and political sophistication or need for cognition (Bizer et. al. 2000), however.
Although some studies on foreign policy frames incorporate political sophistication in their analysis, they frequently fail to address the issue of frame congruence, that is, to consider the match between the frame and the individuals’ political predispositions. Even if predispositions are accounted for, measures are rather crude and unspecific, such as political ideology, trust in government, and partisanship. For instance, Liberman (2006) uses death penalty support as a proxy measure for moral predispositions (i.e., retributiveness and humanitarianism), which he believes should influence public support for the punitive use of “war against evildoers”. Death penalty support should be closely related to moral predispositions, yet Liberman fails to prove or explicate the assumed association in more detail. Besides the relative disregard of relevant predispositions, none of the studies analyzes the interaction between political predisposition and sophistication. This neglect is rather troublesome since Zaller’s Reception-Acceptance Model has shown that it might be imperative to approach political sophistication and predisposition in concert (see Chapter 2.2.5). The only instance where this is actually done (and that I am aware of) is the study of Hermann, Tetlock, and Visser.
To conclude, also more specific studies on the attitudinal impact of foreign policy frames are characterized by major shortcomings. First, even those studies that can be assigned to the interactionist perspective often fail to operationalize individual-level characteristics with more domain-specific measures. Second, given that predisposition is considered at all, its potential interaction with political sophistication stays unexamined, bar one exception (i.e. Hermann, Tetlock, and Visser 1999). Finally, many framing experiments suffer from a lack of external validity in that they artificially expose participants to one-sided frames that are merely represented by a single sentence or paragraph.
This Chapter 2.2 showed that central tenets of the standard account of framing are at odds with the complexities of the real world. More advanced approaches to framing such as the interactionist perspective should model framing effects as a function of message content and personal attributes and pay closer attention to individual-level characteristics, above all to prior dispositions and political sophistication. Some recent studies on foreign policy frames have met these demands, yet they fail to adequately incorporate predispositions and their connection with political sophistication. By way of analyzing the German discourse on foreign intervention, the next chapter fleshes out the argument of talking intervention I developed in Chapter 2.1 (the “macro-perspective”), at the same time adopting the theoretical and methodological recommendations I have put forward in Chapter 2.2 (the “micro-perspective”).
2.3 The German Experience of Talking Intervention
This chapter aims at substantiating the notion of interventionist talking and its counter-rationale by means of a case study and to derive testable hypotheses on their impact on political preferences. By means of analyzing the German Bundestag debate on the Lebanon assignment, I identify the incidence of two dominant frames which correspond in their logic to the justification of interventions in terms of humanitarian principles and the evocation of negative sentiments to avoid such interventions, respectively. Given the non-existence of other studies, this frame identification process constitutes the central theoretical underpinning for the design of the survey experiment (see Chapter 3.1). It therefore seems adequate to present empirical results in this section.
2.3.1 The Changing Conception of German Foreign Policy
It is quite evident that after its reunification and recovery to full sovereignty Germany has assumed a stronger role in world affairs (see Bahr 2003; Haftendorn 2006; Schöllgen 2003). In particular, the country is one of the largest providers of peacekeeping troops and deploys about 7.300 armed forces in 10 different assignments (Bundeswehr 2007). So far, only low-risk missions have been carried out, however, and taking hazards is strictly avoided (the holding off from Afghanistan’s south constitutes the most recent example). This is partly due to the fact that pacifism and casualty reluctance are deeply routed in the German society (Dalgaard-Nielson 2005). As implied by the logic of the body-bag effect (see Chapter 2.1.2), an unwary rush to more dangerous commitments would possibly jeopardize public support for interventions and international involvement in general. I believe that German policymakers are aware of this contiguity and therefore justify the increasing participation in foreign commitments by using value-based language.
A key to understanding the German foreign policy discourse and the discourse on intervention in particular is the concept of “responsibility”. Very roughly, this term can be defined as the commissioning of national interests under a higher value system such as the protection of human rights or democracy. At the same time a certain disposition and capability for foreign policy action is accentuated (von Bredow 2006, 201). The concept of responsibility consists of two dimensions which are mirrored in most governmental texts on German foreign policy (Ibid 200-205). The first dimension relates to Germany’s past and its accountability for Nazi crimes. This liability is said to persist to this day, implying the abandonment of power politics and reluctance in the conduct of foreign affairs. The second dimension is concerned with the stability of the international system. It presumes that growing threats to international peace and security ask for a higher commitment on the part of Germany. This commitment is to be guided by the principles of multilateralism and carried out in the context of the European Union or the United Nations.
Baumann (2002) argues that, prior reunification, the first dimension or interpretation of responsibility was predominant in the German discourse, whereas the second dimension played only a marginal role. He shows that especially foreign minister Hans-Dietrich Genscher was at pains to draw a clear line between Germany’s post-war understanding of responsibility and the more traditional concept of power-politics (Ibid 22). One decade later, Baumann asserts, the emphasis has been reversed. Now, the predominant demand is that Germany must assume more responsibility in world affairs, resulting from its economic and political weight. This position is nicely exemplified by a statement of Karl Kaiser:
„Die Destabilisierungstendenzen in der Weltpolitik nach dem Ende des Ost-West-Konflikts und das gewachsene Gewicht des vereinten Deutschlands mit seinem hohen Verflechtungsgrad mit der Weltökonomie und -politik stellen für die deutsche Politik eine unausweichliche Herausforderung dar. Dies gilt nicht nur für die globale Entwicklungen, sondern auch für die europäischen […] deshalb kann die größte Demokratie Westeuropas nicht der Verantwortung ausweichen, durch ökonomische, diplomatische und auch militärische Mittel die Durchsetzung der Völkerrechtsordnung mitzutragen …“ (1994, 8, quoted in von Bredow 2006, 202; empahsis added by the author).
It is remarkable that the notion of “to take responsibility” here becomes a synonym for “to exert influence” or “to participate in military assignments”, somewhat approximating the traditional concept of power politics (see Baumann and Hellmann 2001, 71-72). Despite the asserted “continuity in German foreign policy” (Harnisch 2001), a notable shift in both foreign policy discourse and action has occurred, today even more pronounced than at the time of Kaiser’s remark. In particular, the longing for a stronger world role has been most notably mirrored by Germany’s (hitherto unsuccessful) quest for a permanent seat on the UN Security Council.
The concept of responsibility has provided German policy-makers with a powerful cue to substantiate the need for foreign entanglements, possibly more so than in other Western democracies. For instance, in her speech at the 43rd Munich Conference on Security Policy on February 10, 2007, Chancellor Angela Merkel emphasized the shared “global responsibility for global crises”, requiring a new and comprehensive understanding of security (Bundesregierung 2007). German policy-makers sign up to the concept of The Responsibility to Protect originated by the International Commission on Intervention and State Sovereignty (ICISS 2001). In essence, this report offers guidelines for humanitarian military measures to protect people that suffer massive human rights violations (Seybolt 2007, 12). This reference to the responsibility concept seems to be a clever move since it resonates well German views on the civilizing role of law and norms and its self-understanding as a “civilian power” (Harnisch and Maull 2001).
2.3.2 A Case in a Point: The Debate on the Lebanon Mission
By way of analyzing the plenary debate on the Lebanon mandate, I aim to better grasp the specific discourse on foreign assignments and to obtain some guidelines on how to construct the treatment condition of my survey experiment. The debate took place from September 19 to September 20, 2006, and was probably one of the most intense and emotionally charged debates in that year. The 599 members of the German Bundestag had to decide about the deployment of (at most) 2,400 air and naval troops to join the United Nations peace-keeping mission in Lebanon (UNIFIL). A great majority (442 yeas to 152 nays, five abstentions) voted for the assignment.
The evocation of a moral or historical responsibility clearly was at the core of the discussion. Strikingly, both proponents and advocates of the assignment agreed on Germany’s inherent accountability for the Middle East and Israel in particular, but came to diametrically opposed conclusions. One side (CDU, SPD, Greens) argued that its unique responsibility obliged Germany to act on the ground, also militarily. This contribution was conceived as an indispensable condition for the securing of a durable peace and the stabilization of the whole region. In particular, intervention adherents emphasized the shared duty of Germany, the European Union and the international community to take action. A good example for this line of reasoning is an excerpt of the speech of foreign minister Frank-Walter Steinmeier (SPD):
„Das versetzt uns – mit uns meine ich Europa – in die Lage, jetzt auch im Nahen Osten mitzuhelfen, Frieden zu schaffen und zu sichern […] Weil das so ist und weil wir wussten, dass das Schweigen der Waffen nur durch internationale Präsenz, durch ein internationales Hilfeversprechen erreicht werden konnte, durften und dürfen wir nicht abseits stehen. Wir stehen in einer gemeinsamen Verantwortung. Für die Bundesregierung sage ich: Wir sind entschlossen uns dieser Verantwortung zu stellen“ (Plenarprotokoll Deutscher Bundestag 16/49; empahsis added by the author).
Overall, proponents stood out for their general optimism and appeared quite confident with respect to the mission’s chances of success. The opponents’ side (FDP, PDS/Left Party), in contrast, concluded that Germany’s unique responsibility implied the abstention of any military action in the Middle East. The leading thought in all justifications was the need to avoid a certain risk. Skeptics advanced the view that Germany lacked neutrality in the conflict and that a military engagement would unnecessarily imperil its hard-earned and well-respected status as the area’s mediator and conflict broker. Instead of deploying armed forces, diplomatic and political initiatives had to be pursued. As Florian Toncar (FDP) elaborated:
„Das für mich entscheidende Argument, gegen den vorliegenden Antrag zu stimmen, ist, dass Deutschland Gefahr läuft, das Vertrauen, das es bei allen Konfliktparteien der Region genießt, zu verlieren. Im Unterschied zu vielen anderen Staaten, die häufig von einer der Konfliktparteien als parteiisch wahrgenommen werden, verfügt Deutschland bisher über eine allseits hohe Akzeptanz und wird als Vermittler mit guten Kontakten zu allen Beteiligten allgemein anerkannt. Der von der Bundesregierung beantragte Einsatz bewaffneter Kräfte in der Region birgt zumindest die Gefahr, dieses politische Kapital zu verlieren. Deutschland hat mehr zur Beilegung dieses Konflikts beizutragen als Soldaten. Es sollte vom Einsatz bewaffneter Kräfte in der jetzigen Form absehen, um seinen politischen Handlungsspielraum zu bewahren. Vielmehr sollte Deutschland sich im Verhandlungsprozess für eine dauerhafte Friedenslösung als ehrlicher Vermittler anbieten“ (Plenarprotokoll Deutscher Bundestag 16/50).
Another prevalent rationale was the possibility of a direct confrontation between Israeli and German military. Such an encounter would be “unthinkable”, constituting nothing less than a moral disaster for the German psyche. Most important, however, was the preoccupation with the German soldiers themselves. This concern was not confined to the soldiers’ physical health alone but also to their mental integrity. Opponents argued that German forces would be completely overburdened by the mission’s tasks. Gregor Gysi (PDS/Left Party) made this point clear by claiming:
„In meiner Generation ist das alles kompliziert und wirr genug. Es gibt keine Klarheit. Es gibt Angst vor Diskussionen. Wir, Frau Bundeskanzlerin, sind nach meiner Auffassung keinesfalls berechtigt, diese völlig ungeklärte Gedanken- und Gefühlswelt, die in unserer Generation noch immer herrscht, die jungen Soldaten austragen zu lassen. Diese können das nicht. Sie sind überfordert. Wir alle haben nicht das Recht, sie in eine solche Situation zu bringen“ (Plenarprotokoll Deutscher Bundestag 16/49).
Despite the position’s somewhat larger heterogeneity in its political reasoning, most intervention adversaries agreed that a military assignment would be irresponsible and too dangerous to accomplish, jeopardizing some established status quo, either the mental and physical integrity of German soldiers, or Germany’s international credibility as an accepted and effective arbitrator. Compared to the interventionist position, the tone was more negative and the outlook more pessimistic.
In order to somewhat substantiate this interpretation I conducted a computer content analysis by means of the text analysis module WordStat (Peladeau 2003). I incorporated all plenary speeches (n = 27), as well as all amendments attached to the final protocol (n = 21) which together constituted a text corpus of roughly 40.000 words. 25 statements argued for, and 23 statements against the intervention. The single statements were designated as unit of analysis and coded with respect to party affiliation and represented argument. The analysis was effected by the specification of an inclusion dictionary that automatically removed variant forms of keywords from the text corpus. These expressions (n = 69) were predefined based on theoretical considerations and a minimum frequency criterion. They were then allocated to broader and mutually exclusive categories (n = 11) and checked for consistency in meaning using the Keyword-In-Context technique (see Ibid).
Table 2.1: Tabulation of Category Frequencies
illustration not visible in this excerpt
Figures correspond to column percent and raw counts (in parentheses); **p ≤ 0.01; ***p ≤ 0.001 for chi-square of category frequency; (*) p ≤ 0.05; (**) p ≤ 0.01 for chi-square of case occurrence (not shown in the table). Both tests refer to counts. Cluster solution is based on similarity index.
Overall, results confirmed the findings from the qualitative analysis. To keep things simple, I confine myself to reporting the outcomes of a univariate frequency analysis. As shown in Table 2.1, I used a chi-square test to assess the relationship between the frequencies of the main categories and the speakers’ opinion, either favoring (“Pro”) or rejecting the Lebanon mission (“Con”).
It is evident that there is a clear difference between the two positions. Intervention advocates emphasized the need for international cooperation (“Concerted Action”), especially in the European context (“Europe”), had a positive outlook (“Future”, “Peace”, and “Positivity”), and stressed the inherent responsibility to act (“Responsibility”). Adversaries, in contrast, resorted to a more pessimistic language (“Negativity”, “Risk”, and “War”), cautioned to draw the lessons of the German past (that is, to abstain form international affairs), and gave warning of the potential loss of the country’s bargaining position (“Mediator”). A cluster solution performed on the cross-tabulation table revealed that the distribution of keywords was consistent with expectations. Note that the clustering is not performed on individual documents but directly on the cross-tabulation. As a consequence, the similarity index computed for the keywords and used for clustering does not represent their co-occurrences within documents but measures the similarity of their distribution among the pro and con positions. The finding of the cluster solution was corroborated by a correspondence analysis that graphically explored the relationship between content categories and pro and con positions (see Appendix A7).
2.3.3 Risk versus Responsibility Framing
In conclusion, both qualitative and quantitative analysis substantiates the assumption of the existence of two general, yet very distinctive lines of reasoning. I argue that these two dominant narratives can be conceived as emphasis frames as defined and explicated in Chapter 2.2.1. Accordingly, I understand the plenary debate on the Lebanon intervention as frame contest, where one frame highlights Germany’s inherent “responsibility to protect” while the other frame stresses the risks associated with intervention. Throughout this thesis, I refer to the former interpretation as responsibility framing and to the latter as risk framing. Note that these two frames mirror the two dimensions of the responsibility concept in German foreign policy, only in the more specific realm of military deployments: the risk frame corresponds to the responsibility concept dominant prior reunification (that is, to stay out of world affairs), and the responsibility frame corresponds to the responsibility concept dominant after reunification (that is, to participate in world affairs). More general, responsibility and risk framing can be conceived of a subtype of interventionist talking and negative and emotional framing, respectively.
 In meiner Heimatstadt Mainz heißen die Österreicher Zwockel. Man erklärt diesen Namen so, dass ein österreichisches Regiment in Mainz stationiert war und die Soldaten Zweige an ihren Hüten hatten. Diese nannten sie in ihrem österreichischem Dialekt „Zwoargerl“ (Quelle: Wikipedia).
 The quote is part of a comment about the first Bundeswehr casualty. A sergeant of a medical corps stationed in Phnom Penh (Cambodia) was shot for unknown reasons by a motorcyclist (translation by Kümmel and Leonhard 2005).
 Gerdts is the former head of the foreign affairs department of the Bundespresseamt. He made this point at a conference on the media’s role in foreign policy, held from February 20 to February 21, 2003 at the Department of Foreign Affairs in Berlin (Auswärtiges Amt 2003).
 This is not to say that the (problematic) concept of the “national interest” or national security considerations have become irrelevant to foreign policy-making and the rhetorical justification thereof. In fact, the U.S. strategy of pre-emptive war is largely explained by national security interests (with 9/11 as the watershed). Likewise, security concerns play a major role in the formulation and justification of Bundeswehr assignments, exemplified by the “Hindukush-remark” of Peter Struck. I refrain from elaborating on the argument that the legitimation of interventions in terms of humanitarian principles merely constitutes a veiled and sophisticated version of realpolitik. In this thesis, I am more interested in the actual impact of cosmopolitan arguments and its resonance within the audience, not in their origin or motivation (see Hasenclever 2000 for the latter).
 I understand intervention rather narrowly as unilateral or multilateral military activity that encompasses the direct and overt commitment of combat-ready armed forces for humanitarian purposes, either with or without UN mandate (see Vertzberger 1998, 114). Seybolt (2007, chap. 2) develops a typology of humanitarian interventions distinguishing four types of engagement: (a) to assist aid delivery, (b) to protect aid operations, (c) to save victims and (d) to defeat perpetrators. I refrain from incorporating different types of interventions in my analysis, instead focusing on a “hard case”: the hypothetical deployment of ground troops at the Lebanese-Syrian border (see Chapter 3.1.3). Future studies, however, should consider the possibility that the impact of interventionist discourse is contingent on the degree of the intervention’s severity. See also my conclusion in Section 5.
 For instance, Feldman (1988) has given empirical support that citizens base their opinions on the connection that they draw between a particular issue and their core values. In subsequent work, Feldman and Steenbergen (2001) find that humanitarianism (understood as the desire to help the disadvantaged) can serve as a stable anchor for policy preferences. Hurwitz and Peffley (1987) propose a hierarchical model of foreign policy opinions in which core values determine individuals’ general postures which in turn determine opinions on particular foreign policy issues.
 The framing of a foreign intervention can be arrayed along a continuum from complete dominance by one frame to a complete equilibrium between competing frames. Entman (2004, 47) refers to the latter situation as frame parity.
 Given the relative lack of research about casualty shyness in the European context, I primarily focus on American studies. The basic mechanisms between public opinion, the media, and foreign policy-making should nevertheless also apply to other major democracies.
 This chapter on the body-bag effect is largely informed by reviews of Aldrich et al. (2006) and Boettcher and Cobb (2006)
 This thinking is reflected in the so-called zero-dead doctrine, assuming that “war without bloodshed” is the only form of military force that is acceptable in democracies, except when direct threats to immediate national interests are involved (Everts and Isernia 2002). Germany’s reluctance to deploy ground troops in the south of Afghanistan or NATO’s self-imposed restriction to air strikes in the Kosovo intervention are clear examples for the doctrine’s application.
 Smith (2003, quoted in Kümmel and Leonhard 2005, 37) argues that the casualty factor increases the further a particular war or military operation is removed from core national interests and proposes the following hierarchy of causes in descending importance: defense of the homeland, defence of allies, promotion of vital interests, punishment of aggression, prevention of genocide, and humanitarian assistance.
 In contrast, Kull and Ramsey (2000a) highlight the importance of elite misperceptions. They show that U.S. policy-makers have often underestimated the public’s willingness to accept the costs of international engagement, providing empirical support for Entman’s (2000) notion of “perceived public opinion”.
 The “rally around the flag” hypothesis constitutes the antipode to the body-bag hypothesis. Here, the use of military force is said to cause a patriotic public to support their leader during a crisis (Russett 1990). In my view, this phenomenon is likely to arise in the case of immediate threats against national sovereignty or core national interests and is rather unimportant for the case of humanitarian interventions and/or peace-keeping missions.
 There is a plethora of different typologies and theories on the role of the media in foreign affairs. In my view, the proposed distinction between active and passive functions seems sufficient. See Robinson (2002) for an excellent treatment of various theories of media influence.
 Though intuitively compelling, studies have yet to present sufficient evidence for the CNN effect (Gilboa 2005). Empirical support for the indexing hypothesis has been considerable (see e.g. Mermin 1999; Zaller and Chiu 2000), whereas in my view support for the notion of manufacturing consent rests rather on circumstantial evidence.
 This practice is generally known as “selecting on the dependent variable” and causes inference problems. Quite simply, one cannot learn anything about a causal effect when the dependent variable does not vary (King, Keohane, and Verba 1994, 147).
 This instrumental view of the media is similar to the concept of media diplomacy, understood as the proposition that the conducting of classical diplomacy is extended by the emergence of a global media system (see Ebo 1996; Gilboa 1998). Of course, the downside of this approach is the neglect of the differing interest groups in the media itself (e.g. journalists and other media professionals) and their engagement in framing activities. Johnson-Cartee (2004) describes these kinds of aspects and many more.
 In this sense the media is a necessary but not a sufficient condition for successful political persuasion. In the absence of any media coverage even the most powerful interventionist frame will have a hard time to reach and influence its target audience (the relative indifference towards the situation in Darfur is a recent example). Given a certain degree of media coverage, however, framing and the frame contest gain in importance.
 Of course, the success of a frame is likely to depend on source credibility (see Druckman 2001). For analytical clarity, however, I refrain from addressing source credibility, instead focusing on message content and the recipients’ individual-level characteristics (see Chapter 2.2).
 See D’Angelo (2002) for a rebuttal.
 Agenda setting is generally understood as the ability of the news media to define the significant issues of the day (McCombs and Shaw 1972). McCombs, Shaw, and Weaver (1997) used the term second-order agenda setting to describe the media’s ability to define and interpret those issues. It is interesting to see that agenda-setting researchers claim the framing and priming concepts as part of their territory, while others argue that agenda-setting is “best conceived as variant of priming and framing” (Price and Tewksbury 1995, 2). Other studies have referred to agenda setting, priming and framing without differentiation (e.g., Popkin 1994). From my point of view, the concepts are distinguishable in that the notion of agenda setting is largely set up and tested at the aggregate level, whereas the concepts of framing and priming are also grounded on the micro-level (e.g. the individual processing of information).
 The search included the Science Citation Index Expanded, the Social Sciences Citation Index and the Arts and Humanities Citation Index and was conducted on May 11, 2007.
 Note that emphasis frames do not violate the invariance property of expected utility theory (see Druckman 2004). This is important because the argument that (equivalency) framing effects render central normative assumptions of rational choice theory descriptively inaccurate has spawned considerable controversy in the social sciences (see Bueno de Mesquita and McDermott 2004; Mercer 2005).
 For the sake of convenience I subsequently use the expression “framing/framing effect” when referring to “emphasis framing/emphasis framing effects”. Note that besides the concept of emphasis frames the term “issue frame” is also very common in the literature (e.g. Jacoby 2000; Sniderman and Theriault 2004). Another frequent expression is “value frame” (e.g. Barker 2005; Brewer and Gross 2005). Both are conceptually similar to the understanding of emphasis frames in that they highlight the importance of a particular argument versus others. I prefer to use the term emphasis frame as it points at the theoretical notion of framing at the micro-level, understood as the unequal weighting of considerations stored in long-term memory.
 It is very important to stress here that Zaller provided one of the most comprehensive accounts of mass communication and public opinion (Kinder 2003). His “Receive-Accept-Sample (RAS) Model” consists of four axioms that describe how individuals respond to political information they encounter (1992, chap. 3). Two of these axioms (the Reception Axiom and the Resistance Axiom) emphasize the pivotal role of political knowledge and predispositions for information processing, an issue I will come back to below. The point I want to make here is that the manipulation perspective of framing mainly draws on only one of his axioms, namely, the Accessibility Axiom (see Ibid 48). The understanding of framing as priming (framing makes considerations temporarily more accessible in short-term memory) and the view of the citizenry as being easily manipulable somewhat downplayed the two other axioms’ application to the field of political persuasion resulting in the incidence of relatively unsophisticated research designs.
 Nelson and colleagues base their argumentation on expectancy-value models of attitudes. These models regard attitudes as summary evaluations based on a weighted average of a sample of beliefs about an attitude object (Ajzen and Fishbein 1980; Anderson 1981). Nelson and colleagues distinguish priming (the accessibility of beliefs) and framing (the unequal weighting of beliefs) from persuasion. Persuasion is understood as the adding or (Bayesian) updating of new information. This distinction, however, should be less relevant here than the evidence that framing is conceptually distinguishable from priming and that framing effects should be contingent on individual-level characteristics (see also Hiscox 2006).
 Both the accessibility and the belief importance model are memory-based accounts of framing effects. Lodge, Steenbergen, and Brau (1995) provide a major alternative, the so-called on-line model. According to this model, individuals access and report a previously formed opinion based on a steady stream of information over time instead of constructing the opinion based on information stored in memory.
 In a similar vein, constructivist scholars working on norm diffusion define the notion of congruence as an “ideational affinity to other already accepted normative frameworks” (Keck and Sikkink 1998, 204) and argue that political leaders try to connect emerging norms to established ones when striving to persuade the public (Finnemore and Sikkink 1998, 906).
 Note that dissonance may occur in the affective as well as in the cognitive domain (Keller and Block 1999).
 Another significant finding is the discovery that incongruence between communication frames and the individual diminish comprehension because the new information does not match past learning stored in memory (Graber 2005).
 The “perceptual-readiness” hypothesis of Sniderman, Tetlock, and Elms (1999) reaches a similar conclusion. Higher sophisticated people are supposed to be more receptive to conditions under which particular political principles apply, and therefore are more affected by framing.
 The applied number of axioms is inconsistent in Zaller’s work. He suggests four axioms in the RAS Model, two axioms in the Reception-Acceptance Model. In turn, Zaller and Feldman (1992) propose three axioms in their theory of response instability.
 I will satisfy this requirement by modeling a three-way interaction between frame condition, political sophistication, and predisposition (see Chapter 3.3).
 Hermann, Tetlock, and Visser (1999) embedded several experiments in a national survey in order to find out if public decisions about the use of force rest either on dispositional or situational factors. They identify several interactions between situational factors and political sophistication and argue that sophistication is reflected by the systematic and deliberate matching of situational and dispositional factors, allowing more knowledgeable people to “consistently translate their general ideological dispositions into specific policy attitudes” (1999, 569). This assertion will be mirrored in my fourth hypothesis (see Chapter 2.3.4).
 The review is clearly not exhaustive but aims at clarifying my central argument here, regarding the overall lack of theoretical and methodological sophistication, even in very recent studies on foreign policy frames. A considerable amount of experimental studies has explored the relationship between equivalency framing and the populations’ sensitivity to casualties (e.g. Boettcher 2004; Boettcher and Cobb 2006; see also Vertzberger 1998). As I have argued in Chapter 2.2.1, however, the concept of equivalency framing seems rather inappropriate for the analysis of political debate.
 Current missions are: Active Endeavour (maritime surveillance in the Mediterranean), Enduring Freedom (maritime surveillance at the Horn of Africa), EUFOR (Bosnia and Herzegovina), ISAF (Afghanistan), KFOR (Kosovo), UNIFIL (maritime surveillance of the Lebanon coast), UNOMIG (surveillance mission in Georgia), UNMEE (surveillance mission in Ethiopia), UNMIS (surveillance mission in Sudan), UNAMA (capacity building in Afghanistan).
 The power of the cue responsibility is demonstrated by Iyengar (1991) who shows that attitudes and actions are altered by the manner by which individuals attribute responsibility. He finds that frames that attribute responsibility to collective units (thematic frames) tend to engender support for governmental or societal intervention, in contrast to frames that attribute responsibility to the individuals (episodic frames), in turn preserving the status quo.
 In mid-October, German naval forces took command of the Maritime Task Force and will keep it at least until end of August 2007. The mandate’s main objective consists in the surveillance of the Lebanese waterside in order to prevent the clandestine shipment of arms and other illicit goods.
 WordStat is specifically designed to study textual information such as interviews, journal articles, public speeches etc. The program is not a stand-alone application, however, and must be run from the statistical software package SimStat or CodeMiner (see Peladeau 2003). See Weber (1990) for several advantages of content analysis. Most important, computer-assisted content analysis leads to “perfect coder reliability” (Ibid 41).
 There are many shortcomings with cluster analysis. Most important, it does not offer any real goodness of fit test which makes it difficult to choose the exact number of clusters on an empirical basis (however, see Dahinden 2006, chap. 10 for a scree test based on amount of explained variance). Factor analysis avoids this shortcoming, yet comes with the requirement of interval scaled data. König (2004) suggests introducing latent class analysis to the frame identification process. Latent class analysis can be performed on ordinally and nominally scaled variables and allows for a number of goodness of fit measures (Ibid; see also section 3). Relating to my case, a latent class analysis can be accomplished by extracting the data on keyword or category frequencies from WordStat. Then, the data is further analyzed using a statistical package for latent class analysis, e.g. the WINMIRA program (Von Davier 1997). Future studies should elaborate on this point.
 Of course, I am aware of the risk of a self-confirmation bias here. I therefore made a considerable effort to develop a categorization schema that was informed by the data (in form of the minimum frequency criterion), as well as by theoretical considerations (the keywords’ assignment to the categories). See Appendix A for a list of the categories with attendant keywords and a diagram of the keyword validation process. Furthermore, I provide figures and test results for case occurrences (i.e., number of speakers), as well as statistics for the clustering process and the correspondence analysis. Refer to the CD-Appendix for an expanded categorization dictionary, a break-down of the analysis to political parties (instead of pro and con position), the protocol of the debate on the Lebanon intervention, and all SimStat and WordStat data files.