© Institute for Security Studies of WEU 1997. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photo-copying, recording or otherwise without the prior permission of the Institute for Security Studies of WEU.
ISSN 1017-7566

Published by the Institute for Security Studies of Western European Union and printed in Alençon, France, by the Imprimerie Alençonnaise.
 

CONTENTS

Preface

Summary

Public opinion and the new security environment

Methodological issues
    What is public opinion?
    What is an attitude?
    Salience
    Knowledge, information and non-attitudes
    Issue politicization and issue publics

Aspects of contemporary European attitudes to some security policy issues
    The data
    Institutionalized security cooperation
    NATO
    The enlargement of NATO and EU
    CFSP
    Conflict intervention
    Nuclear issues
    The cognitive dimension of attitudes to a common European policy

Conclusions

Figures and tables
 

PREFACE
 
With the end of the Cold War, security has acquired very different components. In Europe, it has broadened to include conflict prevention and crisis management, in an attempt to substitute persuasion for enforcement. It has essentially become a political rather than a military concept whose features are foresight, transparency and accountability, and which combines political and economic as well as military measures.

The change of emphasis from collective territorial defence to cooperative security is already producing a disaggregation of the security reflexes that prevailed during the Cold War. Some degree of 'renationalization' of defence postures according to specific geostrategic interests is appropriate and beneficial, provided it generates the national consensus necessary to reconstitute adequate multinational forces. The community of nations, reassured by the disappearance of a clear and present danger to vital national interests, may otherwise become over-exposed to the many international ills and revert to a more egotistical outlook.

Furthermore, like monetary matters, security issues are no longer the preserve of high office but must increasingly involve public opinion, for reasons that pertain to both democratic consensus and international legitimation. Intervention in the internal affairs of collapsed states, interposition in civil wars and monitoring of democratic processes have been acquired as developments of international common law. But the right to intervene, albeit only under certain conditions, has not become an automatic duty: the motives and arrangements have to be argued through and explained to the public, on the specific merits of each case. Politicians and policy-makers must therefore involve the public in the new security culture that circumstances suggest, inter alia in order to steer cooperative projects through the log-jam of structural reforms, budgetary adjustments and institutional enlargements that is looming ahead.

The complexities that make up the views expressed in opinion polls, the difficulty in motivating and stimulating the public in matters that have traditionally been beyond its concern, and the methodology of polling itself, are analysed in detail by Richard Sinnott, an authority on the subject, in this chaillot Paper, an earlier version of which was discussed during a seminar last autumn. Its publication comes, we hope, in time to assist those who may be involved in the debate, ratification, referendums and other democratic and constitutional measures that will give added substance to the current processes of institutional reform and enlargement of the European security structures.
 
Guido Lenzi
Paris, July 1997
 
SUMMARY

Political decision-makers have always had to take account of public opinion. In the post-Cold War period of re-evaluation of national and institutional roles in questions of security, and given the complexity and uncertainty of challenges and the immediacy of modern communications, paying attention to public opinion has become an imperative.
Having argued why this is so, the author devotes a section of the paper to detailed discussion of the nature of public opinion. He shows that the reality of public opinion, as opposed to the indicators of it that are reflected in the polls, has to be discerned and a critical attitude taken when interpreting the data provided by polls.
Recent evidence is presented of European attitudes to security issues, based on a selection of data from European and US surveys. The particular 'issue sectors' discussed are institutionalized security cooperation (covering attitudes to NATO, NATO and EU enlargement, CFSP), conflict intervention and nuclear issues; the data from polls on these demonstrate clearly that levels of awareness and knowledge of foreign and defence policy are much lower than is frequently assumed.
The paper's conclusions begin with indications of the state of European public opinion. These suggest that, whereas there may, for instance, be consistent support for NATO in certain countries, public opinion is generally complex and has weak cognitive foundations, so that further critical evaluation of polls is needed if the public's real attitudes are to be perceived. The implications of the paper's findings are that policy-makers ignore public opinion at their peril; that they should be highly critical of data from polls; that there is not just one public but a series of publics; that the stratification of public opinion is not rigid; and that attitude changes are likely to occur as a result not only of public information campaigns, but also politicizing events which, like the process of revision of the Treaty on European Union, may be predictable. Finally, the author recommends that the European security institutions, and in the first place WEU, should develop a common approach to polling on security policy.
 
 PUBLIC OPINION AND THE NEW SECURITY ENVIRONMENT

While there may be much debate about the meaning and implications of concepts such as 'cooperative security', 'interdependence' and 'globalization', there is no doubting the major change in the security environment that has occurred, both gradually over the last quarter century and dramatically in response to the fall of the Berlin Wall, the relinquishment of Soviet control over Eastern Europe and the breakup of the Soviet Union. Though the overall effect has been a vast improvement in the international security environment, this has been accompanied by an equally vast increase in the complexity and uncertainty attached to security issues and security policy. Martin and Roper summarize this paradox from a European Union perspective: 'As a paradoxical result, although the defence problems of the members of the European Union are less pressing and less mortally dangerous than they recently were, they are likely to prove more difficult to manage and more divisive precisely because there is more latitude available in deciding what to do.'⁽¹⁾

Beyond the immediate problems of conflict prevention and crisis management, the transition has raised fundamental questions as to the nature and scope of security policy and security cooperation. The latter has led to re-evaluations of the purpose and organization of existing international security institutions, to the search for new institutions with new purposes and to the formulation of security objectives for international institutions that had previously avoided the security issue.

The most explicit and far-reaching reorientation is that which pertains to the European Union; it is embodied in Article J.4 of the Maastricht Treaty: 'The common foreign and security policy shall include all questions related to the security of the Union, including the eventual framing of a common defence policy, which might in time lead to a common defence.' Although institutional issues may sometimes appear remote and abstract, these issues have urgency and momentum and can induce controversy. This is particularly evident in the case of Article J.4 of the Maastricht Treaty, in its reference to 'the eventual framing of a common defence policy, which might in time lead to a common defence' and which, until the Amsterdam Treaty's wording has been finalized and the Treaty has been signed in October 1997 (and subsequently ratified), will remain the legally valid text concerning the relationship between the EU and WEU on defence issues. Nor has this difficulty been totally overcome in the Amsterdam text, which refers to 'the progressive framing of a common defence policy . . . which might lead to a common defence'. The negotiations and drafting of the Amsterdam Treaty have shown that it was not possible to reach a consensus on the progressive integration of WEU in the EU, even though the new article (J.7) of this draft Treaty includes the explicit mention of future EU/WEU integration; cooperation in the field of armaments as part of the CFSP; implicit mention of WEU's Petersberg tasks; the possibility for the EU to issue political guidelines to WEU whenever it 'avails itself' of WEU; and the possibility of closer cooperation between two or more member countries in the spirit of the Treaty.

European Union institutions are not the only ones undergoing potentially controversial re-evaluation and redesign. Since 1989, NATO has been systematically re-examining its role and its relationship to other European security institutions. The crucial issues in this re-examination are how to maintain a continued US involvement in European security and how the European pillar of the Atlantic Alliance is to be developed. These issues are inseparable from the discussions in the IGC, and are also intimately bound up with how individual member states see their role, and with the question of the full reintegration of France into the Alliance. When we bear in mind that all of these discussions, redefinitions and repositionings are taking place against the background of the prospective enlargement of the various institutions, it becomes apparent that the institutional agenda is nothing if not complex, sensitive and pressing.

From a public opinion perspective, what all of this adds up to is that, just at the time when it is facing new and uncertain challenges and is confronted with major choices, foreign and security policy is less insulated from a more engaged public opinion. The engagement of public opinion arises partly in response to the policy developments and the policy challenges just described, partly in response to changing attitudes to sovereignty and the role of the state, which can be attributed to growing interdependence, and partly in response to a revolution in education and communications. changes in the latter are related to the process of globalization and their significance lies in the speed with which 'foreign' crises and major events in other political systems are relayed into the immediate living environment of individuals and households.⁽²⁾

But do policy-makers really need to pay attention to this new and unpredictable public opinion? 'The public be damned' was William Henry Vanderbildt's forthright response when asked whether the public should be consulted about a proposal to introduce luxury trains. That approach to public opinion may have been all right for a businessman, as long as he was sure there was a market for his product and before the days of consumer activism and political correctness. Political decision-makers may share this sentiment but have always had to be more circumspect; as a participant in a 1980 colloquium on 'La France face aux dangers de guerre' in Paris is reported to have said: 'To have courage means above all not to be discouraged by opinion polls.'⁽³⁾ Not being discouraged does not mean, however, that public opinion can be ignored. Given the nature of the security policy agenda and the nature of both contemporary public opinion and contemporary communications, political decision-makers must take account of public opinion, however discouraging or frustrating it may be; indeed, the imperative is not just a recent one - such a realist view was articulated as long ago as 1780 by Edmund Burke when he referred to 'The coquetry of public opinion, which has her caprices, and must have her way'.⁽⁴⁾

Of course, attention to public opinion becomes absolutely unavoidable when institutional change requires treaty ratification, whether that ratification is based on the representative democratic politics of parliamentary debate and decision or on the direct democratic politics of the referendum. In neither case will the governments of the member states of the European Union want to face again the difficulties and uncertainties that dogged the ratification of the Maastricht Treaty. But, it must be emphasized, the constraints and demands of public opinion are not confined to situations of treaty change; they are now an integral part of the background of all foreign and security policy-making.⁽⁵⁾

The purposes of the paper are: (1) to clarify certain aspects of the nature of public opinion as it relates to security policy in the new security environment, (2) to establish a framework for and an approach to the realistic interpretation of public opinion poll evidence in this area and (3) against this background, to consider some aspects of contemporary European attitudes to security policy. The concluding section of the paper will offer some reflections on the implications of these three themes for the policy-maker.
 
METHODOLOGICAL ISSUES
What is public opinion?
The simple answer to this question, and the most frequently implied when public opinion figures in political discussion, is that public opinion is the opinion polls. This is a very unsatisfactory answer: it confuses measurement and the object being measured. Without wishing to enter into abstruse methodological or epistemological discussion, this distinction is crucial. Political science begins by defining certain concepts, for example authority, democracy, legitimacy, participation, public opinion and so on. In order to observe these, however, it must go a step beyond conceptual definition to operational definition, that is, to defining the concepts in terms of how they are to be measured.

The importance of the distinction between conceptual and operational definition in the present context is that any given opinion poll is merely a particular 'operationalization' of, or, in other words, a particular attempt to measure, something larger and more elusive, namely public opinion as it actually is. Bearing this in mind gives one a healthily critical perspective on opinion polls. French, German, Spanish, or Hungarian public opinion is neither the latest opinion poll from the country in question nor the summation of all such opinion polls: it is an underlying complex reality which is partially and imperfectly captured by a given set of operationalizations or measuring instruments (consisting of the sampling methodology and the set of questions asked in an opinion poll or series of opinion polls). It is frequently said that an opinion poll is a snapshot of public opinion. The problem is that the metaphor is not pushed at all far enough. What needs to be added is that the snapshot may have been under-exposed or over-exposed; the lens may have been telescopic where a wide-angle lens was required or vice versa; the focal length may have been inappropriate; the camera may have been shaken, and so on. In short, the reality of public opinion must not be confused with any set of indicators of it, and we need to maintain a highly critical attitude, not to public opinion itself (the 'public opinion be damned' school of thought), but to the evidence presented and its interpretation. As Donsbach et al. put it: 'Unfortunately, the problem today is not the calculation of statistical significance or the reduction of mass data, so as to back up the findings: it is the subtle search for indicators of the climate of opinion which is at stake . . . However, a great of deal of doubt remains whether the results presented here - or elsewhere, by other scholars and poll institutes - measure "true" public opinion.'⁽⁶⁾

All of this only says what public opinion is not: it is not simply the results of the latest poll. In positive terms, public opinion can be defined as the aggregate of attitudes in a given population towards politically relevant objects (values, goals, institutions, organizations, groups, persons, problems, policy instruments and policy options). Public opinion is not, however, an undifferentiated aggregate; it cannot be estimated by just adding up all the replies in a particular set of response categories. There is no single public: there are various publics, differentiated from one another in terms of degrees of knowledge, interest and salience, of political involvement and of structured or unstructured response to the issues or policies or institutions in question. But to pose the problem of aggregation is to anticipate the argument in several of the sections that follow. Before pursuing these aspects, it is necessary to say something further about the concept of attitudes.
 
What is an attitude?
In raising this question, we run a very real risk of getting bogged down in complex debates whose proper place is in a seminar on social psychology. Accordingly, the question is best answered briefly and peremptorily by quoting a classic definition: 'An attitude is a mental and neural state of readiness, organized through experience, exerting a directive or dynamic influence upon the individual's response to all objects and situations with which it is related.'⁽⁷⁾ The most important step in developing the concept of attitudes is the next one, which is to specify that attitude is a multidimensional concept comprising four components: (1) cognition, (2) salience, (3) affect or evaluation and (4) behavioural intention. In short, an individual can have an awareness or knowledge of an object, can rate its importance, can view the object positively or negatively,⁽⁸⁾ and can be disposed to act towards the object in a particular way.

Much discussion of public opinion tends to focus on the third of these components, that is, on whether the public responds positively or negatively to a policy, or a leader or an institution or whatever. This is perfectly understandable; the affective or evaluative dimension is what comes to mind first of all when we think of attitudes. In order to interpret the significance of attitudes in this narrow sense, however, we need to know whether they are likely to have any behavioural consequences; hence the need to look at the fourth component (behavioural intentions). Equally, if not more importantly, in order to make sense of the affective, evaluative and behavioural responses, it is absolutely vital to examine the cognitive dimension of the attitude and the matter of salience.

Behavioural intentions receive fairly sustained treatment in the context of studies of electoral behaviour, since from the point of view of political decision-makers how people vote is ultimately what matters. While there is a need for much more interaction between those who study elections and those who study particular aspects of public opinion, it should also be noted that politically relevant behaviour extends well beyond voting. Voting is the mass political phenomenon par excellence but, particularly in an age of new social movements, individuals may join demonstrations, sign petitions, write to their public representative or express their views through a wide variety of more or less formal or informal groups to which they are attached in varying degrees.⁽⁹⁾ Beyond such political actions, individuals may discuss politics, issues and leaders; in the context of such discussions, attitudes such as satisfaction with incumbent office-holders may have behavioural consequences.

Unlike behavioural intentions, the cognitive and salience dimensions of attitudes have tended to be neglected. A lot of research makes a ritual nod in their direction but they are rarely taken into account in a systematic way. Yet their significance in the interpretation of public opinion, especially public opinion on foreign and security policy, can hardly be overstated. The next two sections spell out this significance, beginning with the question of salience.
 
 Salience
One would expect that, for a number of reasons, the salience of foreign and security policy would be quite low. In general, élites tend to overestimate the salience of any political issue; for the mass of people, their political problems, or at least political definitions of their problems, rank in importance and attention well below more immediate and personal preoccupations of economic security, health, family concerns and, for many, leisure interests and pursuits. On the limited scale of their political concerns, foreign and security issues tend to occupy the lower rungs. Foreign and security specialists should not express regret or even shock at this, since, traditionally, the direct impact of the vast majority of a country's foreign policy decisions on the lives of its own citizens has been marginal.

However, there are two qualifications to be made to this argument. Normally, foreign policy has limited impact on the lives of a country's individual citizens but, if things go seriously wrong, it can have the most immediate and comprehensive impact. Even without things going seriously wrong, a foreign or security issue can come to loom quite large due to a combination of events and the political response to them. Thus, the salience of foreign policy and security issues is variable, according to circumstances and according to the nature of the issue. This of course makes the measurement of salience all the more important.

The second qualification to the basic thesis regarding the low salience of this issue area has already been mentioned: the dividing line between 'domestic' and 'foreign' is beginning to break down; foreign affairs are becoming domestic affairs and aspects of domestic affairs are becoming internationalized.⁽¹⁰⁾ This phenomenon, which is a product of growing economic and political interdependence, should not be exaggerated; neither should it be neglected. It makes the 1990s different from preceding decades, and it requires a new and more penetrating analysis of the salience of 'foreign' policy issues for public opinion.

Responding to this requirement is, in methodological terms, not exceptionally difficult. While there may be debate about the best measure of salience to use in survey research (e.g., importance for oneself versus importance for the country; or a rank ordering of the top two or three issues from a given list of items versus a scale judgement applied to each item versus an open-ended question), there is extensive evidence on the utility of a wide range of such measures and incorporating them into a research design is not difficult. What is difficult, and very rarely done, is to incorporate evidence of salience into the analysis of the data. What we sometimes find is a juxtapositioning of evidence on the substance of the public's policy preferences with evidence on the overall importance of foreign affairs; what we need is a breakdown of the distribution of preferences by levels of interest, not just by levels of interest in foreign affairs in general but by level of interest in the particular issue. If we had this we would be some way along the road towards taking adequate account of the stratification of public opinion in this area; but only some way, because we would still have to tackle the more complex and difficult problem of knowledge and information.
 
Knowledge, information and non-attitudes
Turning to the matter of the public's information about and knowledge of international politics and foreign policy brings us face to face with a major current debate about the quality of public opinion on matters of foreign policy. This issue is so central that a brief consideration of the terms of the debate is essential.⁽¹¹⁾ What has been dubbed 'the conventional wisdom' takes a dim view of public opinion on foreign and security policy. It argues four propositions: that public attitudes in this area lack an adequate knowledge base, that they are unstructured, that they are unstable and that they have no relevance for policy-making. The tone was set as early as 1950 by Gabriel Almond, who concluded that '. . . foreign policy attitudes among most Americans lack intellectual structure and factual content. Such superficial psychic states are bound to be unstable since they are not anchored in a set of explicit values and means ends calculations or traditional compulsions.'⁽¹²⁾

This conventional wisdom has come under sustained attack in recent years, to the extent that it might be said that the revisionist view is now in the ascendant. However, the revisionist critique is not as comprehensive or as telling as might appear at first sight. For one thing, it does not contest the first proposition of the conventional wisdom - that levels of knowledge of international affairs are woefully inadequate.⁽¹³⁾ Moreover, the revisionist critique of the fourth proposition - regarding the effect that public opinion has on the making of foreign policy - can be endorsed; it does not in fact have any major implications for one's understanding of the nature of public opinion. This means that the debate centres on two issues: the question of structure or constraint in foreign policy attitudes and the question of stability.

In regard to the proposition that attitudes are unstructured or unconstrained, the main criticism of the conventional wisdom makes a distinction between general foreign policy orientations and abstract beliefs, on the one hand, and specific foreign policy attitudes and opinion on immediate issues, on the other. It then argues that, while constraint might be lacking in relation to views on particular matters,⁽¹⁴⁾ it does obtain at the level of general orientations and beliefs. It turns out, however, that much of the attitude constraint in question is based on assessments of the Soviet Union.⁽¹⁵⁾ This gives rise to three problems. First, is this constraint simply a by-product of the Cold War? Secondly, what happens to such attitude constraint with the demise of the Soviet Union? Thirdly, is assessment of the Soviet Union really the basis of an intellectual structure or a belief system in the sense in which this term was used by both Almond and Converse?⁽¹⁶⁾

The main evidence against the conventional wisdom on the third issue (the instability of attitudes) focuses on aggregate or collective public opinion.⁽¹⁷⁾ The argument is that 'many of the familiar deficiencies of individuals' opinions - weak informational bases, lack of structure, instability over time, and the like . . . are overcome in the aggregation process, so that collective opinion is highly stable, well structured, and responsive to the best available information.'⁽¹⁸⁾ The evidence of the relative stability of collective public opinion during the last fifty years, and of the predictability of such changes in opinion as have occurred, is a convincing refutation of the view that public opinion as a whole is capricious. It does not, however, solve the problem of the quality of opinion at the individual level or negate the evidence that public opinion at this level can be highly unstable. Shapiro and Page are at pains to point out that their argument only applies to collective public opinion and they acknowledge that 'much individual-level instability cancels out across the population'.⁽¹⁹⁾ Moreover, their conclusion that 'The quality of public opinion tends to reflect the quality of the information and the choices with which the public is presented' is an argument in favour of investigating the quality of public opinion at the individual level and, in particular, the cognitive dimension of attitudes.

Reflecting on the implications of this debate as it arises in the context of research on the link between foreign policy issues and voting behaviour, Rattinger provides a useful summary: 'The debate between conventional wisdom and "revisionists" has its European counterpart as well. It often appears in the form of disagreement over the thesis of a "democratization" of foreign and security affairs . . . The lesson to be learned here is that both the conventional wisdom and the revisionist school can be part right . . . The theoretical and empirical task . . . cannot be to establish that all is black or all is white. The task rather is to investigate which portions of the electorate are "black" and which are "white", which are all shades of gray, how this comes about, and what consequences flow from this assortment of views for the overall distribution of political choices.'⁽²⁰⁾ In short, the public's awareness and knowledge of foreign policy issues, the possibility of the occurrence of 'non-attitudes' and the structure or lack of structure underlying attitudes in this area remain important considerations in the interpretation of public opinion on foreign and security policy.

In an attempt to bring together the implications of the foregoing discussion of salience, knowledge and non-attitudes, Figure 1 presents a typology of attitudes and non-attitudes. In examining the typology, it should be borne in mind that, in practice, public opinion poll responses are usually treated as belonging to just two types, i.e. all those in the right hand column (the negative or positive response to the items in question) and all those in the left hand column (the 'don't know' responses). Attitudes are then reported on this basis; indeed they are frequently reported on an even more simplified basis, the 'don't knows' being excluded or ignored. The reality is much more complex.

In the first place, responses given in an opinion poll are grounded in varying degrees of knowledge of the attitude object (an issue, problem, institution, person or policy action). In Figure 1, three degrees are assumed: none, weak and strong. Responses will also vary in terms of the salience the respondent attaches to the attitude object. While this is also a matter of degree, in this case it is treated as a dichotomy: the object is either salient or it is not. Combining these two distinctions with the positive/negative versus 'don't know' responses that constitute the actual answers given to the opinion pollster yields no less than twelve types of attitude or non-attitude (they are numbered in Figure 1 for ease of reference). Cell one contains those who don't know anything about the object, who know they don't know and who acknowledge that they don't have an attitude. Unfortunately, this explicit acknowledgment of lack of awareness or lack of knowledge is much rarer than it ought to be. The polar opposite is found in cell 12: responses that express an attitude based on reasonably comprehensive knowledge of an object that is regarded as salient by the respondent; the problem is that responses that are not explicit 'don't knows' are frequently treated as if they were of this type.

The 'non-attitude' problem comes in three shapes (cells 2, 3, and 4 of Figure 1). Respondents in cell 2 give a positive or negative response to an object without any knowledge of the object but also without attaching any degree of importance to it. Although Cell 3 does not include a positive or negative response, it is a form of non-attitude in so far as it attaches salience to an attitude object that is unknown to the respondent. Cell 4 is the most serious case of non-attitudes: the respondent both makes a positive or negative judgment and attributes salience to the object despite knowing nothing about it.

The other four cells in the left-hand column of Figure 1 (cells 5, 7, 9, and 11) can be dealt with briefly. They are all explicit 'don't knows' on the question or item in the poll; they differ in their nature and implications depending on whether they are combined with weak or strong knowledge of the object and with a sense of salience or non-salience. The differences between them underline the fact that the 'don't knows' as reported in surveys can gloss over the differences between genuinely neutral responses, responses expressing indifference and responses reflecting confusion.

The three remaining types (cells 6, 8 and 10) comprise capricious attitudes, ill-informed attitudes and latent attitudes. Cell 6 is where Edmund Burke's public opinion caprices are found - expressions of positive or negative attitudes that are based on poor knowledge and are non-salient and which, if measured again the next day, could well be different. Ill-informed attitudes (cell 8) are likely to be more stable because of their salience for the respondent. They are not, however, immune to change, the most likely source of change being new information. Finally, latent attitudes are positive or negative responses that are based on accurate knowledge but accompanied by a low level of salience. In this case the most likely source of change is the mobilization of attention and interest, leading to the conversion of latent attitudes into committed attitudes.

These observations about attitude change and its sources are systematized in Figure 1 by the arrows and their accompanying letters and numbers: A, B₁, C₁ etc. Attitude change is often thought of as the conversion of negative to positive responses or vice versa. Such changes would occur within the cells in the right hand column of Figure 1 and so do not appear in the figure. In fact, political campaigns or public information campaigns or unplanned disturbances in the political environment are much more likely to involve one or more of the three types of change illustrated by the lettered/numbered arrows in the Figure. If we assume that a process of mobilization occurs in response to political campaigning or to the course of events, respondents may move from the left-hand to the right-hand column of Figure 1 (arrow A). This is seen in the reduction of the rate of 'don't knows' in opinion polls, a phenomenon that is regularly observable in the transition from pre-campaign to campaign polls as elections approach but also occurs in non-electoral contexts in times of immediate crisis or heightened political debate. As an alternative to or in combination with move A, campaigns or unfolding events may increase the salience of an attitude object (these changes are indicated in Figure 1 by the arrows B₁, B₂ and B₃). The third kind of attitude change involves increases in knowledge and is probably the most difficult transition. It is denoted by the arrows C₁ and C₂ in Figure 1. Of course attitude change may involve combinations of all of these moves; on the chessboard of attitude types, individuals may move as pawns, knights, bishops or queens or as combinations of any or all of them. In short, attitude change is highly complex and represents much more than a conversion from 'against' to 'for'. Politicization is as important as persuasion.

Issue politicization and issue publics
Attitude formation and attitude change and any impact that attitudes may have on policy occur in the context of particular political structures and processes. In regard to the impact of attitudes on policy, Risse-Kappen has argued that 'Differences in political institutions, the structure of society and coalition-building processes in the policy networks (state domination, societal control and democratic corporatism) largely account for the variations in the impact of public opinion on foreign and security policy.'⁽²¹⁾ In regard to attitude formation and attitude change, events and processes politicize issues. Politicization heightens the salience of an issue, leads people to form attitudes to it and may increase levels of knowledge regarding it. Politicizing processes may be exogenous or endogenous. In the former case they spring from outside the political system as major problems or challenges arising in the international environment. Many politicizing events are, however, endogenous to the policy process, i.e. they originate in major initiatives or decisions taken by political leaders or in crises arising in the course of implementation of a policy. Alternatively, they may be endogenous to the political process, as when an existing or a new party or movement succeeds in putting a new issue on the agenda or in raising the profile of an old issue. On its own, this is a less likely source of politicization because, without the stimulus of either external shocks or policy developments that have politicizing consequences, it is difficult for parties or groups to politicize an issue.

Because it is uneven, rather than creating a homogenous public opinion, the process of politicization creates 'issue publics'; as Converse put it: '. . . we come a step closer to reality when we recognize the fragmentation of the mass public into a plethora of narrower issue publics.'⁽²²⁾ But what constitutes an issue and an issue public? From a policy-maker's point of view, it may seem that the policies are the issues. What he or she wants to know is what the public thinks of what the government or the international organization is doing or proposes to do. At the most general level, security policy as a whole could perhaps be thought of as an attitude having its own issue public. The problem is that it is a very general attitude object that may encompass diverse issues and different issue publics. In practice, public opinion polling frequently focuses on highly specific policies and on the preferences the public may have in regard to individual policy options (e.g., for or against sending troops to Bosnia). Attitudes to issues in this very specific sense give rise to the opposite problem: while important in themselves, they are too specific to define an issue public. They do not exist in isolation but rather are embedded in a nexus of attitudes to a wider range of objects. Consequently, rather than thinking of issues in terms of security policy in general or of particular policy options, it is better to focus on something in between these two poles, i.e. on issue sectors, defined as conflicting combinations of attitudes to values, actors, problems, instruments and policies relating to some aspect of, in this case, security policy. The presentation of the public opinion evidence in the following section of this paper focuses on three such issue sectors: attitudes to the institutionalization of security cooperation, attitudes to conflict intervention and attitudes to nuclear weapons.
 
ASPECTS OF CONTEMPORARY EUROPEAN ATTITUDES
TO SOME SECURITY POLICY ISSUES

The data
Before embarking on an examination of the evidence on European public opinion and security policy, it is necessary to make some preliminary remarks about the nature and sources of the data used. First, it is important to note that this is not an exhaustive compilation of data in this area. Rather, data have been selected by reference to the three topics indicated and in terms of the emphasis in the foregoing discussion on the need to pay special attention to the salience of attitude objects and to the knowledge or lack of knowledge that the public has of them. An important additional criterion in selecting the data to present was comparability: because the concern is with public opinion across a range of European countries, and because of the extra interpretative leverage that can be obtained from comparative analysis, priority was given to data that were either precisely or at least approximately comparable. Where the degree of comparability is only approximate, this is noted both as a necessary qualification of any conclusions drawn and as a reminder of the need for the coordination of research efforts in this area.

The main sources of the comparative data presented are the Eurobarometer and United States Information Agency (USIA) surveys. Eurobarometer surveys, involving representative samples of the adult population in each country surveyed, have been conducted twice-yearly in every member state of the European Community/Union since the mid-1970s. The surveys focus on attitudes to European integration, paying some attention to foreign and security policy matters on a regular basis and with more extensive analysis of these issues from time to time. USIA surveys are also polls of representative samples of the public in a wide range of countries. They differ from Eurobarometer surveys in that the choice of countries is both more selective and broader: the focus tends to be on the main West European countries and on the countries of Central and Eastern Europe, though some data on Russian and American public opinion are also collected. The USIA series of surveys is much the longer of the two, dating back to the early 1950s; however, only very recent data from either the Eurobarometer or USIA surveys are used in the present paper. In addition to these comparative surveys, the paper draws on national surveys or opinion polls, especially where these provide repeated measures of the same variables (all opinion polls used are cited in the relevant figure or table, and in footnotes).
 
 Institutionalized security cooperation
Institutional development is chosen as one of the key issue sectors to be considered in this report for two reasons. First, cooperative inter-state behaviour that is facilitated by institutions or, at a minimum, by 'regimes'⁽²³⁾ is one of the hallmarks of the contemporary international system. Having emerged at first on the basis of a functionalist logic and confined to economic or technical matters, such behaviour is now increasingly manifest in the security field. Given the traditional monopoly held in this area by the nation-state, this development raises issues of the legitimacy and role of the nation-state, issues that are intimately linked to public opinion. The second reason for focusing on this issue area is that it has its own politicizing effect, resulting from the various recurring efforts to reform or strengthen institutions of international cooperation. These have occurred most regularly within the European Community/Union (the Single European Act, the Maastricht Treaty, the Intergovernmental Conference negotiations for a revision of Maastricht, etc.). They are not, however, confined to the European integration process; there is, for example, the on-going process of NATO reform and particularly the issue of NATO enlargement; there are also the implications for the Brussels Treaty, which established the WEU. Such processes give rise to domestic debate and ratification requirements, in some instances amounting to direct consultation of the public in a referendum. Even without the latter step, these bring the issue of institutional cooperation before the bar of public opinion in a direct and immediate way.

Despite a tendency in analyses of security policy to assume that everything changed utterly in 1989, in fact the continuities have been substantial, and an awareness of them is essential in attempting to examine current attitudes to security cooperation. changes in threat perception, at least as far as public opinion is concerned, not only pre-date the fall of the Berlin Wall but pre-date the end of the Cold War. Study after study of this period has shown that the perception of a threat from the Soviet Union declined in European public opinion at what some security analysts would have regarded as a surprisingly early stage. This led Flynn and Rattinger to conclude that, as far as public opinion is concerned, 'One can say with confidence that the concept of security that we assume dominated the early postwar years - primarily military, primarily East-West - no longer exists.'⁽²⁴⁾ The significance of this statement is that it was made in 1985.
 
 NATO
If some 'post-Cold War' aspects of public opinion seem to pre-date the end of the Cold War, other, apparently quintessentially Cold War attitudes have survived into the new period. This is particularly true in this area of institutional development, at least in terms of support for NATO, and is illustrated in Figure 2. In Britain favourable attitudes to NATO have remained almost entirely stable over the last 15 years, rarely moving much above or below 70 per cent.⁽²⁵⁾ German attitudes to NATO have fluctuated somewhat but the fluctuations have been at a relatively high level and have never reduced support for the Alliance below 60 per cent. In Germany, the period since 1991 has seen a fairly stable level of support in the mid to high sixties. Italy also shows a fluctuating pattern, in this case in a range from the mid fifties to the low sixties. In terms of support for NATO, France has shown a more clear-cut trend than any of these three countries. Support has grown substantially over the period since 1980, the growth in favourable attitudes to NATO in France being remarkably steady up to 1991; it then encountered a significant check, falling from over seventy per cent to just fifty per cent between 1991 and 1992. That check was temporary and French support for NATO recovered to a level of about 60 per cent in the years 1993-95. The 1996 USIA survey shows a slight decline from this level to 54 per cent. Spain, where this USIA question has only been asked since 1991, is the clearest exception in the group of five countries considered in Figure 2: support for NATO in Spain has certainly been steady, but at a relatively low level, ranging between 40 and 46 per cent over the period 1991-96. The evenly balanced division of Spanish opinion on NATO presumably reflects both the recency of Spain's accession to NATO and the politicization of the issue in the Spanish debate on membership.

Leaving Spain to one side, the most striking thing about the evidence on support for NATO in Figure 2 is the continuity in the high level of support for the institution through the period of transition. Analysing data from a much wider range of countries, Everts points to a similar continuity: 'However we measure commitment, it is evident that NATO has enjoyed a great deal of loyalty and commitment in most member countries throughout its existence. This commitment appears to have survived the momentous international changes in recent years.'⁽²⁶⁾

What explains this combination of declining threat perceptions and continuing support for an organization whose raison d'être was to counter the perceived threat? Both Rattinger and Risse-Kappen argue that this has occurred because support for particular security arrangements or institutions is only weakly related or may even be de-coupled from threat perceptions.⁽²⁷⁾ More generally, it can be argued that over time an institution may attract a level of support which is independent of the particular purposes for which the institution was founded. As Everts and Sinnott, commenting on the persistence of pro-NATO attitudes put it, 'This suggests that an international regime can undergo transformation, even transformation related to the very purposes for which it was founded, and still maintain a high level of public support.'⁽²⁸⁾

The institutional legacy of the Cold War and the apparently successful handing on of that legacy in terms of support for NATO is further illustrated in Table 1. In both 1993 and in 1996, substantial majorities in Germany and France, and especially in Britain, expressed confidence in NATO to 'deal effectively' with European problems.⁽²⁹⁾ With the notable exception of France vis-à-vis the EU, the publics of these countries look to NATO rather than to other institutions (OSCE, WEU or EU) for this purpose. Furthermore, NATO and the EU are better known, or, at least, more familiar: the OSCE and the WEU tend to elicit higher levels of 'don't know', although it should also be noted that there are significant decreases in the level of 'don't knows' in relation to both institutions in all three countries between 1993 and 1996.⁽³⁰⁾

Comparable data are available for five countries in Central and Eastern but only for 1993. In four of the five countries, that is in the Czech Republic, Hungary, Slovakia and Bulgaria, confidence in NATO was either substantially lower than it was in the three West European countries or was matched by confidence in OSCE or in WEU or in both. In these four countries also one finds levels of 'don't know' response in regard to NATO that are much higher than in the three West European countries and are not substantially different from the levels relating to OSCE and WEU. A partial exception on both counts among the countries of Central and Eastern Europe is Poland; in 1993 Polish public opinion seemed to distinguish between attitude to NATO and attitude to either OSCE or WEU and showed more familiarity with NATO than with either of the other two institutions (see Table 1); on the other hand, in 1993 the level of NATO 'don't know' responses in Poland was still substantially higher than it was in Britain, France or Germany.

Attitudes to the Atlantic Alliance encompass more than belief that NATO is essential, or more than confidence in NATO as an institution. When, in 1995, respondents were asked whether they believed that the United States cared about the security of Europe or (in the case of the five Central and East European countries) about the security of Central and Eastern Europe, very substantial majorities (two-thirds or more) in Britain, France and Germany expressed the view that their leading NATO ally was concerned (see Table 2). In 1995, the contrast with several of the countries in Central and Eastern Europe in this regard was striking. The proportions believing that the US cared about the security of Central and Eastern Europe varied from 42 or 43 per cent in the Czech Republic and Slovakia to 29 and 28 per cent in Hungary and Poland. Nor was the lower level of perceived US concern a function of general attitudes to the United States - overwhelming majorities (in excess of three-quarters) in all five Central and Eastern European countries expressed favourable attitudes to the United States in the same surveys. One year later, however, confidence in US concern in all of the four Central European countries in question had improved significantly; at the same time it had dropped slightly in each of the West European countries, so that the gap between Western and Eastern Europe, while still significant, was much less than it had been just one year earlier.

Positive attitudes to NATO and positive assessment of the United States' concern for European security in Britain, France and Germany do not, however, entail acceptance of the status quo within the Atlantic Alliance. Indeed, at first sight, the evidence suggests considerable demand for change. Table 3 shows that, over the last three years, sizeable proportions of the populations of Germany, France and Britain wanted the Europeans to 'assume more responsibility and control of [European] security'. In assessing the implications of these responses, the first step is to examine in detail the wording of the question:

'One way some have proposed to deal with post-cold war security issues is to reduce the role of the US in NATO and strengthen the role of western Europe. This would mean Europeans would have much greater say on issues of western security. However, they would have to spend more money on their defence and assume more of the responsibility for the security of western Europe. Which do you think would be the best for the security of western Europe - keeping the same security relationship with the US or assuming more responsibility and control of our security?'

The opening part of the question is straightforward enough. Thereafter the question becomes complicated as it seeks to introduce a balance between 'the Europeans having much greater say on issues of western security', on the one hand, and 'spending more money on their defence and assuming more of the responsibility for the security of western Europe' on the other. It seems doubtful that this attempt at a balanced question succeeds. It is true that taxpayers tend not to want to spend more money; it is also, however, a feature of survey research that taxpayers will not, as opinion poll respondents, acknowledge the connection between government spending and taxation. Furthermore, the notions of Europeans having a 'much greater say' and 'assuming more responsibility' are affectively loaded. The lack of balance is repeated in the final segment of the question, which counterpoises 'keeping the same security relationship with the US' to 'assuming more responsibility and control of our security'.

If this critique of the question-wording is accepted, the data reported in Table 3 may risk exaggerating the extent of the desire for change in the Atlantic Alliance. This conclusion is supported by a closer look at the evidence on French attitudes in this area. In the Service d'Information et de Relations Publiques des Armées (SIRPA) series of surveys, French respondents over the years have been asked a question which offers four alternatives: an alliance between Western Europe and the United States, a European Alliance independent of the United States, non-participation in any alliance, that is, a position of absolute neutrality and, finally, the option of an alliance with Russia.⁽³¹⁾ These data show growing French support for the Atlantic Alliance between 1985 and 1995 (see Figure 3). In 1985, 34 per cent opted for a Western European/United States alliance, 26 per cent for an independent European alliance and 24 per cent for neutrality. By 1995, support for the Atlantic Alliance in France had risen to 46 per cent.⁽³²⁾ The point is that the evidence in Table 3 for 1995 indicated only 27 per cent support for keeping the same security relationship with the US, while 70 per cent wished to assume more responsibility for and control of European security. One should not, therefore, read too much into the figures indicating the level of support for 'assuming more responsibility and control of our security' shown in Table 3; they certainly do not imply a rejection of the Atlantic dimension of the Alliance, and may not even imply deep-rooted support for greater 'burden-sharing'. As a corollary, one should probably not draw any major implications from the appearance of a shift in French attitudes towards maintaining the current level of US involvement in European security in the data in Table 3 for 1996; the SIRPA data for the same year show, if anything, a slight fall in French support for American involvement in European security.

An Italian survey from January 1996 throws further light on the matter. In this case the choice was a four-way one: NATO as it is at present, a greater role for European countries, an exclusively European defence force and neutrality. It is notable that the proportion choosing neutrality (19 per cent - see Table 4) is rather similar to that in France. However, the idea of an exclusively European defence force receives virtually no support (5 per cent) and the bulk of opinion is in favour of the Atlantic Alliance, 39 per cent supporting a greater role for the European countries and 32 per cent favouring NATO as it is at present. By and large, this bears out the critique of the data on Britain, Germany and France in Table 3, that is, if the attractive sounding option of a greater role for European countries is counterpoised to the status quo, it is likely to win considerable support but that this support stops well short of rejecting American involvement.⁽³³⁾
 
 The enlargement of NATO and EU
The greatest potential change facing NATO is enlargement. The evidence suggests that public opinion on this issue does not impose stringent constraints on or make insistent demands for action. Table 5 displays attitudes to possible NATO membership in seven Central and East European countries in 1996. Majorities in almost all of the countries concerned favoured NATO membership but the demand was low-key: those 'somewhat in favour' tended to outnumber those 'strongly in favour'; the notable exception was Romania, where 56 per cent declared themselves to be strongly in favour of NATO membership. Further evidence of the weakness of the public demand for NATO membership in the countries of Central and Eastern Europe is found in attitudes to possible requirements of NATO membership (Table 6). Consistent and substantial support across the four most obvious requirements (a mutual defence commitment, stationing of NATO troops, NATO exercises in the country and NATO overflights) is found only in Poland: two-thirds of the Poles support the first and third requirements listed, while just over fifty per cent support the second and fourth requirements. In almost all the other countries, on almost all the requirements, opposition outweighs support, usually by a very substantial margin (see Table 6).⁽³⁴⁾

Turning to opinion on NATO enlargement in West European countries, we find majorities taking an optimistic attitude in Britain, Germany and France (Table 7). In Italy and Spain, however, the proportion taking an optimistic view (that 'admitting central and east European countries into NATO will benefit the overall security of Europe') dips just below 50 per cent. It is not that pessimism regarding enlargement is rife in these two countries; the difference is that the level of 'don't knows' (almost one-third of respondents) is much higher than in Britain, Germany or France. Even in these three countries, however, the positive attitude to NATO enlargement is seen to be qualified when further questions probe a bit more deeply. Thus, majorities in Germany and France take the side of caution when faced with the choice between immediate NATO expansion and 'not moving too quickly on expanding NATO because Russia feels threatened by NATO expansion and the West's relations with Russia could worsen as a result' (Table 8). Only the British opt (by a very narrow margin) to proceed immediately. However, the difficulty of achieving balance in a complex question which seeks to spell out the rationale for the options offered is again evident in this case. 'To address the security vacuum in Eastern Europe' is pitted against 'because Russia feels threatened by NATO expansion and the West's relations with Russia could worsen as a result'. The specificity of the negative consequences of rapid expansion as compared with the abstractness of the notion of 'addressing the security vacuum' may well have drawn those with little interest or knowledge to the non-committal ('not move too quickly') option.⁽³⁵⁾ In this context it is worth noting the minuscule and clearly unrealistic level of 'don't know' responses to the question (see Table 8).

The enlargement of the European Union is frequently put forward as either an alternative or a supplementary means of solving the security dilemmas of the countries of Central and Eastern Europe. Support for enlargement of the Union to include countries of Central and Eastern Europe is quite widespread in the Union as a whole (65 per cent).⁽³⁶⁾ As Table 9 shows, however, such support varies considerably - from levels well in excess of 80 per cent in Finland and the Netherlands to something in the region of 50 per cent or less in Ireland, Portugal and Spain. From the point of view of our interest in public opinion and security, the key question is whether this support for enlargement is based on security considerations or on economic or moral considerations.

Table 9 shows that mutual security interests constitute the predominant rationale for enlargement. This response is chosen by 49 per cent across the EU as a whole, compared with 23 per cent citing moral duty and 20 per cent economic interest. The most frequent reference to the security factor is to be found in the three Scandinavian member states and in Germany, Greece and the United Kingdom. Overall, the data tend to confirm the expectation that geographical proximity to Central and Eastern Europe would lead to greater concern with the security dimension, though the two exceptions should be noted: Austria (relatively low concern with the security issue despite geographical proximity) and the United Kingdom (relatively high concern despite geographical distance).

The task of creating a zone of peace and security in Europe is not a new one for the European Union. Indeed, many would argue that this was the driving force behind the integration project in its early stages, and that it received renewed impetus following the fall of the Berlin Wall and the unification of Germany. In this perspective, enlargement becomes the next step in reinforcing and extending the European 'security-community', with the latter term bearing the very specific meaning set out by Deutsch in 1957: 'the attainment within a given territory, of a "sense of community" and of institutions and practices strong enough and widespread enough to assure, for a long time, dependable expectations of peaceful change.'⁽³⁷⁾ Over and above the traditional security rationale for enlargement already considered, there is some evidence of support for this more specific notion of a 'security-community' in European public opinion. Faced with a choice of two statements, one of which endorses the security-community concept and the other of which sees it as irrelevant (arguably because it has already been achieved), 49 per cent of respondents in the EU as a whole choose the former while 26 per cent choose the latter (14 per cent disagree with both statements and a further 11 per cent give a 'don't know' response (see Table 10)). The idea appears to have least resonance in Ireland, Portugal and Spain (see the 'don't know' column in Table 10).
 
 CFSP
While long-term developments in integration may lead to a 'security-community' in the sense discussed in the preceding paragraph, in the short term specific decisions have to be made about institutionalizing a Common Foreign and Security Policy. USIA data show substantial support in Britain, France, Germany, Italy and Spain for the development of 'a common European defence force'. Significant pockets of opposition are, however, found in Britain and Germany and relatively high levels of 'don't know' occur in France and especially in Spain (see Table 11). Unfortunately, although the intensity of this support was measured in the survey (see the question in Table 11), the data are not reported in this form and we do not have even this indirect measure of the salience of the attitude in question. Whatever the general attitude to a common defence force may be, widespread support in the countries just considered for a common defence policy is confirmed and the picture extended to all 15 EU member states in Figure 4. The question underlying the data in Figure 4 dealt with a range of European Union policies and was posed as follows: 'What is your opinion on each of the following proposals? Please tell me for each proposal whether you are for it or against it.' Among the proposals was: 'The European Union member states should have a common defence and military policy'. Such a proposal receives substantial majority support (73 per cent) across the European Union as a whole. Figure 4 also shows however, that the range of support varies very considerably from above or close to 80 per cent in the Netherlands, Germany, Italy, and Greece to less than 50 per cent in three of the four neutral member states (Ireland 48 per cent; Sweden 41 per cent and Finland 40 per cent). In the other neutral member state, Austria, public support for a common defence and military policy, at 64 per cent, is not far below the European average.

When the issue is posed in the starker terms of national versus joint EU decision-making on defence, a greater division of opinion becomes apparent (see Figure 5). The wording of the question in this case is: 'Some people believe that certain areas of policy should be decided by the [national] government while other areas of policy should be decided jointly within the European Union. Which of the following areas of policy do you think should be decided by the [national] government and which should be decided jointly within the European Union?' Among the twelve areas of policy included in the question was 'defence'. On the basis of this measure, the balance of opinion in the Union is still in favour of policy integration, but only by a small margin (52 per cent to 44 per cent).⁽³⁸⁾ Support for a common policy remains as high as on the measure in Figure 4 only in the case of the Netherlands. In all other member states there is a noticeable drop in support. Denmark, the United Kingdom, Austria, Greece and Ireland join Sweden and Finland in having majorities against a common defence policy. Opinion in the latter two states is overwhelmingly negative, 81 per cent favouring national decision-making in Sweden and 90 per cent doing likewise in Finland.

What accounts for the rather different pictures portrayed in Figure 4 and Figure 5? Is this further evidence of the unreliability and waywardness of opinion polling and of the truth of the allegation that the answer one gets depends on the question one asks? The fact is that the answer does depend on the question, but this does not mean that the evidence is arbitrary or unreliable. Leaving aside unsatisfactory question wording, different questions measure different variables. In the present case, the question in Figure 4 measures support for a broadly defined 'common defence and military policy' that might not amount to much more than general cooperation in defence matters. As noted above, the question in Figure 5 is a tougher one because it poses the choice between national autonomy and shared sovereignty. Both questions are valid within their own terms of reference. The discrepancies in the results they produce are a telling illustration of the point made at the outset of this paper, i.e., opinion poll evidence is an imperfect and partial measure of the underlying complex reality that is public opinion. Rather than throwing up our hands when differently worded questions regarding the same topic produce different results, we should take advantage of the additional insight into the complexities of public opinion that the discrepancies provide. As well as revealing different attitudes, differences in question wording are likely to magnify variations in response among those who do not have an attitude to the issue in question but are determined to give a response out of politeness, a misplaced desire to be helpful or embarrassment at having to confess ignorance. The response variations are magnified because such individuals are responding simply and solely to the stimuli presented, i.e. to the specific words and cues contained in the question.⁽³⁹⁾
 
 Conflict intervention
Precisely because intervention of this sort is, almost by definition, contemplated or undertaken in crisis situations, the collection of data on public opinion relating either to the crisis or to the intervention itself tends to be ad hoc, unsystematic and productive of evidence which is comparable neither across time nor between countries. This point has been argued trenchantly in a recent article which assesses the measurement of American public opinion in foreign policy crises and proposes a standard set of questions for such situations.⁽⁴⁰⁾ Sobel argues that 'The inconsistent and limited survey research in this area makes it extremely difficult to analyse what the public actually thinks during flare-ups in foreign affairs . . . Generally low consistency and comparability within and between survey organizations often hinder the understanding of the complexities of events or trends over time.'⁽⁴¹⁾ As an illustration of the problem, he notes that only four time-series questions can be found in more than 600 Bosnia-related items in a major polling database covering the period 1991 to 1995. The practical implications of different question wordings can be quite dramatic, as when differences among 50 questions asked between 1992 and 1995 regarding the use of air strikes showed 'support for allied bombing ranging from 30 to 85 per cent'.⁽⁴²⁾ In an attempt to overcome the diversity of question wording, Sobel proposes nine key areas and sets out a list of 33 questions which could be used to measure these nine areas in the case of any foreign intervention situation. In line with the argument advanced at the outset of this paper, he places issue salience and issue 'attention' and knowledge to the forefront of his list of nine key topics.

Not surprisingly, the problems of comparability and consistency of question wording increase exponentially when one turns from US to European data on public opinion regarding conflict intervention Accordingly, it is difficult to draw substantive conclusions from the available data, and the main lesson to be learned from the analysis is the imperative need for comparability and coordination in future research in this area. As argued above, and as Sobel notes in his discussion of opinion in the United States, the first requirement is an assessment of the salience of the particular crisis in national public opinion. Table 12 takes four European countries and illustrates alternative approaches to the measurement of the salience of actual or potential flashpoints.

The French question seeks to pinpoint the source of the greatest perceived threat to France, comparing areas of actual or potential conflict (the countries of Central Europe and the Balkans, and of the Middle East and Africa) with the traditional source of perceived threats to West European security, namely the former Soviet Union (the 'countries of the Far East' are also included). In this comparison, it is striking that the 'greatest threat to France' is seen by a plurality of respondents throughout the early 1990s as coming from Central Europe and the Balkans.

The question posed in Germany shown in Table 12 focused mainly but not exclusively on individual countries as the source of threat but in this instance the threat was to world peace and the respondent was presented with a long list of countries. Allowing for the different wording of the question, there are some commonalities and some contrasts with French perceptions. Like the French, German respondents were most likely to be anxious about the threat from the Balkans but less likely to mention the countries of the Commonwealth of Independent States. Substantial minorities were concerned about a threat from the Middle East and from Islam or Iran but perhaps the main finding was that two in five Germans saw no country posing a threat to world peace. If, as appears to be the case, this response category was actually presented to respondents along with the list of countries, the size of the group choosing it illustrates yet again the importance of presenting respondents with, as it were, an escape hatch; if such a device is not provided in a question like this, the probability of generating 'non-attitudes' is considerable.

The measurement of salience in the Italian survey shown in Table 12 is very different. Posing a much broader range of potential threats (from economic inequalities to terrorism and from Islamic fundamentalism to immigration from the Third World), the Italian surveys found that 'the Balkans' rated very low as a source of perceived threat. As measured in this way, the key threats for Italians are economic inequalities and terrorism. Finally, the Spanish question shown in Table 12 takes a much more specific approach, asking how much of a danger the current conflict in Algeria poses for Spain. On this assessment, this particular potential flashpoint is seen by only a minority of the Spanish public as representing a danger (6 per cent a great danger and 25 per cent somewhat of a danger).

Each of these four approaches has advantages and disadvantages and, in combination with other questions on this topic, they could be used as a basis for designing a cross-nationally comparable measure of the salience of local or regional conflicts or of threat perception more generally for use in future surveys. In the meantime, however, we are left a long way from a genuinely comparable measure of this key variable of issue salience, let alone of the matter of the public's level of knowledge or understanding of the issues or threats involved.⁽⁴³⁾

Bearing in mind that the data on the salience of the conflict in ex-Yugoslavia or in other potential flashpoints are inadequate, it is still worth considering attitudes to military involvement in the situation in Bosnia. Table 13 presents relevant but not identical measures from Britain, France, Germany, Italy and Spain. The basic picture to emerge from Table 13 is that majorities in all four countries support their country's military involvement in Bosnia. The level of support is most consistent and highest in Britain, but this may need to be discounted in view of the reference in the British questions for 1993, 1994 and 1995 to the purpose of 'protecting humanitarian convoys'. When the question was changed in January 1996 to 'do you approve or disapprove of the use of British troops in Bosnia for peacekeeping purposes?', there was some decrease in support. The question posed in Germany is also of questionable comparability: first, it refers to UN forces without any reference to the domestically controversial issue of German involvement in such forces; secondly, it envisages massive military intervention. Despite the scale of the intervention, the data suggest considerable German support - only 29 per cent reject the proposal.

Within this overall picture there is evidence of a small but significant decline in support for military involvement in Bosnia up to mid-1995. This occurred in France, Italy and Spain, the decline leaving Spanish public support short of a majority (see Table 13). The rise in Italian support in January 1996 may have been due to the IFOR intervention or to the fact that the Italian question drew attention to the peace agreement and to the Italian Government's decision to support the agreement with a contingent of troops.

In the absence of more systematic data, especially comparable data on salience and knowledge, it is difficult to draw any firm conclusions from these data on attitudes to conflict intervention. One can say that intervention receives majority support but the reliability of the support, the likely effect on it of adverse developments in conflict intervention and the lessons to be learned regarding public opinion in future conflict situations can, on the basis of present evidence, only be guessed at.⁽⁴⁴⁾ It is clear that this particular issue sector is one that requires extensive further research.
 
 Nuclear issues
As with certain aspects of public opinion already considered, attitudes to nuclear weapons are likely to have been considerably affected by the political events and controversies of the early to mid-1980s. Risse-Kappen suggests, for example, that the politics of this period led to a 'legitimacy crisis of nuclear deterrence in the Federal Republic'.⁽⁴⁵⁾ One would anticipate that events since 1989 would also have had significant effects on attitudes, though the precise extent and direction of these effects might be a matter of debate. Events also affect research agendas, however, and the declining salience of the nuclear issue has led to a drying up of the flow of data in this area, making any such hypotheses difficult to test. Despite these difficulties, this paper treats issues relating to nuclear weapons as one of the key issue sectors because of the sensitivity of public opinion in this area and because of the potential for politicization that lies in policy events such as nuclear testing and in instances or threatened instances of nuclear proliferation.

One would anticipate differences between public opinion in nuclear weapons states and non-nuclear weapons states on issues in this sector. The data in Table 14 provide a limited test of this hypothesis, the test being limited to three countries and qualified by the usual problem of having to compare responses to different questions. Bearing these limitations in mind, the data suggest that there is majority support for nuclear deterrent strategies in both Britain and France. In addition to the minority 15 per cent who believe that Britain should 'always have nuclear weapons', 58 per cent believe in keeping British nuclear weapons 'until others get rid of theirs'. The implicitly unilateralist option to simply 'get rid of its nuclear weapons' is endorsed by only 23 per cent. The nearest comparable French question offers only two options - for or against French nuclear deterrence and, on this measure, support for French policy hovers around the 60 per cent mark. This represents a gradual but significant evolution in French attitudes over the last decade or so, as shown in Figure 6. In 1984, 72 per cent supported the French nuclear deterrent as against 14 per cent who opposed it. Over the years, the gap between these two positions has narrowed, reaching the 61 to 28 per cent outcome in 1996 shown in Table 14. These majorities are more or less reversed in Germany; in this case the data are from 1991 and relate to NATO nuclear weapons - 57 per cent supported unilateral abandonment of the NATO nuclear deterrent compared to 43 per cent who felt that NATO nuclear weapons should be retained as long as the (then) Soviet Union possessed such weapons.

Public opinion in both Britain and France, while supportive of the status quo, takes a minimalist view of the nuclear option and is specifically opposed to new developments or new deployments. Thus, between 1993 and 1995, French opposition to the resumption of nuclear tests in the Pacific was as substantial and as consistent as French support for maintaining an independent nuclear deterrent (compare Tables 15 and 16). The British evidence is more indirect since is relates to a hypothetical situation of the siting of American nuclear missiles in Britain. Sixty per cent would be opposed to such a deployment, taking the view that American nuclear missiles would make 'Britain a less safe place [in which] to live'. When the issue of Britain 'having its own independent nuclear missiles' is considered, opposition drops to 37 per cent; a majority believes either that such missiles make Britain a safer place to live in (45 per cent) or that they make no difference (12 per cent).

The German data on specific nuclear policy issues relate to the inclusion of British and French nuclear forces in the process of nuclear disarmament (see Table 17). Not surprisingly, there is overwhelming support for such a proposal: 82 per cent agree, 59 per cent strongly, and a mere seven per cent disagree. This is in line with the findings for France and Britain: such support as there is for nuclear weapons policies is limited, at most, to support for the status quo; any change which enhances the role of nuclear weapons is likely to meet widespread resistance, while any prospect of diminishing that role is likely to be widely welcomed. Having said that, it must be emphasized that the data on preferences in this area are sparse and the data on salience, knowledge and understanding sparser still.
 
 The cognitive dimension of attitudes to a common European policy
The argument in the first part of this paper has been that non-attitudes, non-salient attitudes and attitudes based on low levels of awareness and information are much more serious problems, and much more serious obstacles to our understanding of public opinion, especially in the area of security policy, than is frequently assumed to be the case. Ideally, this theme would be systematically treated in dealing with each of the issue sectors examined above. Unfortunately, in most instances the unavailability of appropriate data makes such treatment impossible. This part of the empirical section of the paper seeks to advance our understanding of this problem by focusing on just one of the topics dealt with - attitudes to a common foreign and defence policy.

The discussion above noted the extremely low level of 'don't knows' produced by the question dealing with preferences for national government versus joint European Union decision-making on defence policy (see Figure 5). It is striking indeed that the level of 'don't knows' across the European Union as a whole on this question is less than half what it is on the question regarding a common defence and military policy (compare Figures 4 and 5 above). This reinforces the suspicion that there may be a considerable volume of non-attitudes in the data on attribution of defence policy to the European Union level.

This suspicion, or hypothesis if you like, is confirmed by two further pieces of evidence. The question which generates the data shown in Figure 5 is a standard Eurobarometer one that is repeated at regular intervals but always in the same format. The level of 'don't know' responses fluctuates slightly but always at a very low level. Thus, although we have abundant data over time on this issue, the data are of little or no use in attempting to come to grips with the non-attitude problem. In 1994, however, a research project on turn-out in the European Parliament elections provided the opportunity to insert some new questions in the Eurobarometer and, in this context, to use a more exploratory question on policy attribution. The new question was: 'There has been a lot of discussion recently about the European Union (European Community). Some people say that too many issues are decided on by the European Union (European Community), others say that more issues should be decided on by the European Union (European Community). Which of the following statements comes closest to your view?' The response categories included the statement: 'I haven't really thought about it'. This was chosen by 25 per cent of respondents which, when added to the nine per cent 'don't knows' also elicited by the question, yields fully one-third of the sample who have no view on the question of the attribution of policy competence to the European Union.⁽⁴⁶⁾ At a minimum, this strongly reinforces the suspicion that the data in Figure 5, which showed a very considerable polarization of attitudes across the European Union on an issue that is quite central to the current discussions of institutional development, contain a substantial proportion of either non-attitudes or, at best, capricious or ill-informed attitudes.

A recent Eurobarometer report provides a second and more direct means of testing the matter.⁽⁴⁷⁾ Respondents were asked specifically about their perception of the allocation of decision-making power between national governments and 'the European Union level'. The question wording was as follows: 'In fact the [nationality] government, together with those of the other countries in the European Union, have agreed that a number of policy areas will be decided jointly within the European Union, and not by each country separately. Can you tell me which areas of policy are already, at least to some extent, decided at the European Union level?' Among the 22 issues mentioned were 'foreign policy towards countries outside the EU' and 'defence'. Given that two treaty changes and a good deal of recent debate about European integration have focused on precisely these issues, one would expect a reasonably widespread public perception of the actual situation, which is that foreign policy is decided 'at least to some extent' at EU level and defence is not.⁽⁴⁸⁾ The bar-chart in Figure 7 presents two bars for each member state, the first one showing the perceived attribution of defence matters and the second showing the perceived attribution of foreign policy to the European level. In a well informed public, the first bar should be quite low and the second should be quite high. This correct configuration of perceptions is not in fact found in any member state. Finland comes closest: there, only 12 per cent see defence decisions being made at European level; on foreign policy, however, only 46 per cent of Finns get it right. After Finland, the next most informed are Sweden and Denmark and, to a lesser extent, the United Kingdom - at least some proportion of the public in each of these states see the distinction between foreign and defence decision-making. Three other countries (Ireland, Austria and Greece) have reasonably accurate perceptions in the defence area but all three much less so when it comes to joint decision-making in foreign policy. At the other end of the scale in terms of perception of joint decision-making on defence, the outcome is equally if not more surprising: over 40 per cent of people in France, Germany, Belgium, the Netherlands and Luxembourg believe that defence issues are decided 'at least to some extent' at EU level; moreover, the public in these countries see no difference between decision-making in foreign policy and decision-making in defence.⁽⁴⁹⁾ One should note that this is also the situation in the fifteen member states taken as a whole: while only 38 per cent of the European public believe that foreign policy is, at least to some extent, decided at EU level, 38 per cent of the European public also believe that defence matters are decided in this fashion.

This discussion has dwelt at some length on public perceptions in this area for two reasons. The first is that this evidence is vitally important as we attempt to understand preferences in regard to defence-decision making; our interpretation of the rather polarized picture in Figure 5 is considerably altered when we realize the highly inaccurate nature of much public perception in this area. The second reason for dwelling on these perceptions is that they offer strong confirmation of one of the fundamental arguments in the first half of this paper, namely, that levels of knowledge and awareness in regard to foreign and defence policy are much lower than frequently assumed, with all the consequences this has for the expression of non-attitudes and ill-formed attitudes in responses to opinion poll questions.
 
CONCLUSIONS

In the light of the methodological caveats issued throughout this paper, it is not surprising that the substantive conclusions that can be drawn regarding European public opinion and security issues are partial and tentative. The data which would allow a reasonably complete and definitive interpretation simply do not exist. On the other hand, certain indications of the state of European public opinion do emerge and can be briefly summarized.

There is a striking continuity in support for NATO in Britain, France, Germany and Italy. This is accompanied by some desire for a greater role for European countries in the Alliance but support stops well short of a negative view of American involvement in European security. On NATO enlargement, the evidence suggests that public opinion in the West does not either impose stringent constraints on or make insistent demands for government action. With the exception of Poland, lack of insistence and lack of strong commitment regarding enlargement is also the main characteristic of opinion on NATO membership in Central and Eastern Europe. On the issue of enlargement of the European Union to include the countries of Central and Eastern Europe, support is quite widespread in the Union as a whole, and it appears that mutual security considerations rather than any moral imperative or perceived economic interests form the main basis of this support.

Data on support for the defence dimension of the CFSP illustrate the point made above about the complexity of public opinion. In terms of the strong version of a common defence policy (that defence policy should be decided jointly within the European Union), Denmark, the United Kingdom, Austria, Greece and Ireland join Sweden and Finland in having majorities against. Opinion in the latter two states is overwhelmingly negative. In all EU-member states, however, it is essential to take account of the cognitive dimension of attitudes on this issue. The evidence suggests that on this very fundamental question of the level of governance at which defence policy is decided, and, indeed, in regard also to the level at which foreign policy is decided, European public opinion encompasses vast areas of ignorance or incomprehension: an accurate view of foreign and defence decision-making is not in fact found in any member state of the Union.

This paper has also reviewed some evidence on public opinion on conflict intervention. The collection of data in this area tends to be ad hoc, unsystematic and productive of evidence which is comparable neither across time nor between countries. Acknowledging the difficulties in finding comparable measures of salience (measures of knowledge seem to be almost entirely lacking), one can say that intervention receives majority support but the reliability of the support, the likely effect on it of adverse developments in conflict intervention and the lessons to be learned regarding public opinion in future conflict situations can, on the basis of present evidence, only be guessed at. It is clear that this particular issue sector is one that requires extensive further research.

The third issue sector considered in the paper is that of nuclear weapons. This is obviously a highly sensitive area and one with some potential for politicization arising from policy events such as nuclear testing, instances or threatened instances of nuclear proliferation and processes of treaty negotiation or revision. The evidence from Britain, France and Germany suggests that such support as there is for nuclear weapons policies is limited, at most, to support for the status quo. Perhaps not surprisingly, given the nuclear weapons status of Britain and France, support for the nuclear status quo in these countries is in the majority. In 1991, and explicitly taking into account the possession of nuclear weapons by the then Soviet Union, German support for a NATO nuclear deterrent was in a minority, albeit a substantial one (two out of every five German respondents giving an opinion on the issue felt that NATO should retain its nuclear weapons in such circumstances). In short, it seems that, in the three countries considered, any active pursuit of a nuclear defence policy is likely to meet widespread resistance, while any prospect of diminishing the role of nuclear weapons is likely to be widely welcomed.

The foregoing findings are presented as illustrations of the basic theses advanced in the first half of the paper rather than as a definitive account of the state of European public opinion on security policy. The findings highlight certain aspects of public opinion in this area, the most important being its complexity and its weak cognitive foundations. They underline the need for critical evaluation of opinion poll evidence and, hopefully, the findings and the discussion provide some guidelines for the conduct of such critical reflection. Not being discouraged by opinion polls may mean exercising one's critical faculties as much as having courage. This brings us to the implications of the findings for policy-makers.

In considering the implications, the first point to make is that the new security environment makes public opinion on security matters an important but potentially volatile factor that policy-makers ignore at their peril. The second implication that policy-makers might draw from the discussion of European public opinion in this paper is that they should adopt a highly critical attitude to the public opinion poll evidence presented to them. Public opinion is not the undigested results of the latest poll: it is an elusive and complex phenomenon requiring careful interpretation. A third implication is that, while public opinion can be defined as the aggregate of attitudes in a given population towards politically relevant objects, it is not an undifferentiated aggregate. Rather, there is a segmented and stratified series of publics, and any assessment of the policy implications of any set of poll findings must bear this in mind. The fourth implication of the discussion is a further specification of this point: probably the most important factors giving rise to a fragmented public opinion are the different degrees of salience or importance that individuals attach to security policy issues and the different degrees of knowledge they bring to these matters. Consequently, the stratification of public opinion is not rigid. Indeed, attitude change is as likely to consist in people moving between different attitudinal strata as it is to consist in conversion from one position to another.

The fifth implication to be drawn from the discussion is that attitude change of either kind is likely to be brought about by a combination of politicizing events and public information campaigns rather than by the latter alone. Politicizing events or processes may arise unexpectedly (e.g., the sudden eruption or escalation of ethnic conflict); they may also, however, occur on a more predictable basis, as in the politicization that accompanies treaty revision in the European Union. Political leaders and policy-makers have a much better chance of hitching their campaigning and opinion-forming efforts to the latter than to the former; the evidence in this paper on the low level of knowledge and understanding of security policy issues underlines the need to provide opinion leadership in this regard.

A prerequisite for any public information or leadership effort in relation to security policy is a substantial improvement in the coordinated monitoring and interpretation of public opinion in this field. This leads to a number of policy recommendations. The first relates to the national level: relevant policy-makers in each country should reconsider the priorities, objectives and methodologies of the public opinion research they commission or otherwise use or monitor (where no such commissioning, monitoring or use occurs, it is respectfully suggested that remedial action be taken). The second recommendation relates to the international level: joint action should be undertaken to assist and reinforce national efforts, and to meet a necessary condition if a fuller understanding of European public opinion on security policy is to be attained, namely that the research and evidence should be comparable across national boundaries. With this in mind, the full range of European security institutions should come together to develop a common approach to polling on security policy. In the short term, this would involve the creation of a comprehensive data bank of public opinion on security issues. In the medium term, it should lead to the initiation of a coordinated and comparative research project that could be implemented, for example, through the Eurobarometer. Pending such a development, it is recommended, finally, that the Institute for Security Studies of the WEU establish a task force to ensure a full exchange of evidence and of interpretations of public opinion in this field among its member and partner states.

1. Laurence Martin and John Roper (eds.), Towards a Common Defence Policy (Paris: The Institute for Security Studies of WEU, 1995), p. 1.

2. In the light of these and other changes associated with globalization and interdependence, Rosenau argues that '. . . the micro level of individuals has to be integrated into the analysis [of the emerging global order] because structures at the macro level seem increasingly vulnerable to shifts in the skills and orientations of the publics they encompass', and that we must proceed 'as if citizens at the micro level are variables relevant to the emergent global order'. 'Citizenship in a changing Global Order', in J.N. Rosenau and E.O. Czempiel (eds.), Governance without Government: Order and change in World Politics (Cambridge: Cambridge University Press, 1992), p. 274.

3. Renata Fritsch-Bournazel, 'France: attachment to a non-binding relationship', in Gregory Flynn and Hans Rattinger, The Public and Atlantic Defence (Totowa, NJ: Rowman and Allanheld, 1985), p. 70.

4. Edmund Burke, The Works of the Right Honourable Edmund Burke, vol. V (London: George Bell, 1889), p. 492.

5. For a discussion of the relevance of public opinion to international conflict and cooperation generally, see Richard Sinnott, 'Bringing Public Opinion Back In', in Oskar Niedermayer and Richard Sinnott (eds.), Public Opinion and Internationalized Governance (Oxford: Oxford University Press, 1995).

6. Wolfgang Donsbach, Hans Mathias Kepplinger and Elizabeth Noel-Neumann, 'West German Perceptions of NATO and the Warsaw Pact: long-term content analysis of Der Spiegel and trends in public opinion', in Don Munton and Hans Rattinger (eds.), Debating National Security: the Public Dimension (Frankfurt: Verlag Peter Lang GmbH, 1991), p. 250.

7. G.W. Allport, 'Attitudes', in C.M. Murchison (ed.), Handbook of Social Psychology (Worcester, MA: Clark University Press, 1935).

8. The distinction between affective and evaluative is sometimes used to differentiate a more emotional (affective) response from a more instrumental (evaluative) response. Distinguishing between them in practice can, however, be difficult; in the discussion that follows one or other term will be used as seems appropriate in the context.

9. For an overview, see Dieter Fuchs and Hans-Dieter Klingemann, 'Citizens and the State: A changing Relationship', in Hans-Dieter Klingemann and Dieter Fuchs (eds.), Citizens and the State (Oxford: Oxford University Press, 1995); see also Richard Topf, 'Beyond Electoral Participation', in Hans-Dieter Klingemann and Dieter Fuchs, ibid.

10. Richard Sinnott,'Integration theory, subsidiarity and the internationalization of issues: the implications for legitimacy', Working Papers of the Robert Schuman Centre, RSC No. 94/13, European University Institute, Florence, 1994.

11. Initiated mainly in an American context, the debate has also begun to be taken up in Europe (see below).

12. Gabriel A. Almond, The American People and Foreign Policy (New York: Harcourt Brace, 1950), p. 69. Converse later gave the most explicit and now classic statement of the position. He argued that 'as one moves from elite sources of belief systems downwards', one encounters a drop in levels of information, the disappearance of a belief system as such and of constraint between ideas that are in fact related and a shift in attention from the abstract and remote to the concrete and immediate. He emphasized that '. . . the changes in belief systems of which we speak are not a pathology limited to a thin and disoriented bottom layer of the lumpenproletariat; they are immediately relevant in understanding the bulk of mass political behaviour. It is this latter fact that seems to be consistently misunderstood by the sophisticated analysts who comment in one vein or another on the meaning of mass politics.' See Philip E. Converse, 'The nature of belief systems among mass publics', in D.A. Apter (ed.), Ideology and Discontent (New York: Free Press, 1964), p. 213. The problem lies in part with the nature of the opinion poll as a measuring instrument which can generate random responses that '. . . come from people with no real attitudes on the matter in question, but who for some reason felt obliged to try a response to the item despite our generous and repeated invitation to disavow any opinion when none was felt' See Philip E. Converse, 'Attitudes and non-attitudes: Continuation of a dialogue', in E.R. Tufte (ed.), Quantitative Analysis of Social Problems (Wesley, MA: Addison, 1970), p. 175. To drive the point home, he adds: 'Whatever our intentions, the attitude questionnaire is approached as though it were an intelligence test, with the 'don't know' and the 'can't decide' confessions of mental incapacity' (ibid., p. 177).

13. See Ole Holsti, 'Public Opinion and Foreign Policy: challenges to the Almond-Lippman Consensus', International Studies Quarterly, vol. 36, 1992, p. 447.

14. 'We take strong exception to the conventional wisdom that in the face of this uncertainty, the public is unable to connect its foreign policy attitudes together or connect them to its political behaviour . . . the common theme in this research is that while citizens' specific foreign policy attitudes may not be consistently related to one another or to political evaluations, more general foreign policy orientations play a stronger role in foreign policy decision-making.' Jon Hurwitz and Mark Peffley, 'American images of the Soviet Union and National Security Issues', in Munton and Rattinger, op. cit., pp 102-3. Similarly, Wittkopf notes: 'At the abstract level of foreign policy beliefs [the American people] responded with a remarkable degree of coherence . . . On the immediate issues of the day, however, American public opinion sometimes appeared coherent and consistent but at other times malleable and inconsistent.' Eugene R. Wittkopf, 'Public Attitudes Toward American Foreign and National Security Policies since Vietnam', in Munton and Rattinger, op. cit., p. 176.

15. Holsti, pp. 460-1.

16. This is related to a more general problem which is that attitude consistency can itself be a methodological artifact. Referring to the European evidence , Rattinger reports that: 'We have seen . . . some strong or even very strong indications of attitude constraint. However, many of these have occurred when the two survey items were substantively very close together or almost even multiple measurement (e.g., should US forces be withdrawn and do they enhance West German security). In many other instances there is an ominous lack of consistency . . . '(Rattinger, 1991, p. 332).

17. Holsti describes this line of research based on aggregate or collective opinion as the 'most comprehensive challenge to the Almond-Lippmann thesis . . .' (Holsti, p. 446).

18. Robert Y. Shapiro and Benjamin I. Page, 'Foreign Policy and the Rational Public', Journal of Conflict Resolution, vol. 32, no. 2, p. 213.

19. Ibid.

20. Hans Rattinger, 'Domestic and foreign policy issues in the 1988 Presidential Election', European Journal of Political Research, vol. 18, 1990, p. 635.

21. Thomas Risse-Kappen, 'Public opinion, domestic structure, and foreign policy in liberal democracies', World Politics, vol. 43, July 1992, pp. 479-512.

22. Philip E. Converse, 'The nature of belief systems among mass publics', in D.A. Apter, Ideology and Discontent (New York: Free Press, 1964), p. 245.

23. According to the now classic definition, regimes are 'sets of implicit or explicit principles, norms, rules, and decision-making procedures around which actors' expectations converge in a given area of international relations.' S.D. Krasner (ed.), International Regimes (Ithaca, NY: Cornell University Press, 1983), p. 2.

24. Op. cit. in note 3, p. 371.

25. Measured by responses to the question: 'Some people say that NATO is still essential to our country's security. Others say it is no longer essential. Which of these views is closer to your own?' For data derived from this question going back to the late 1960s and covering France, Britain, West Germany and the Netherlands, see Richard C. Eichenberg, Public Opinion and National Security in Western Europe (London: Macmillan, 1989), pp. 124-5.

26. Philip Everts, 'NATO, the European Community and the United Nations', in Oskar Niedermayer and Richard Sinnott (eds.), Public Opinion and Internationalized Governance (Oxford: Oxford University Press, 1995), p. 426.

27. See Rattinger, 1991; Risse-Kappen, 1991.

28. See Everts and Sinnott, 1995, pp. 451-2.

29. Although the nature of the European problems was not specified, the general context of the question suggests European security problems. Note that respondents in Central and East European countries were asked about confidence in institutions or organizations without the reference to European problems.

30. Unfamiliarity with or indecision regarding the WEU and OSCE was much less prominent in Germany than in either of the other two countries in 1993; it had virtually disappeared by 1996. In France and Britain, on the other hand, the 'don't knows' in relation to WEU and OSCE, while declining, remained significantly higher than 'don't knows' in relation to NATO or the EU; this was particularly so in France. However, the warnings given above regarding non-attitudes apply in this case as in every other: in order to obtain an adequate picture of public confidence in international institutions we would need evidence on the salience of the institution in question and of the problems it is supposed to deal with and, most importantly, we would need evidence on the public's awareness and knowledge of the institution. We do not have these data, and the different levels of 'don't know' shown in Table 1 may only be the tip of the iceberg in terms of non-attitudes to European security institutions.

31. The full question wording was: 'Pour assurer la sécurité de la France, laquelle de ces formules vous semble la meilleure? Participer à une alliance militaire entre les pays de l'Europe de l'Ouest et les Etats-Unis; participer à une alliance militaire entre les pays de l'Europe mais indépendante des Etats-Unis; participer à une alliance avec la Russie; ne participer à aucune alliance, prendre une position de neutralité absolue.'

32. Support for an independent European alliance was 27 per cent and support for neutrality was 17 per cent. The proportion supporting an alliance with Russia in 1995 was 2 per cent; there were also 8 per cent 'don't knows'. As the proportions supporting a Russian alliance and giving 'don't know' responses are only given for 1994 and 1995 in the SIRPA/SOFRES report, neither of them is included in Figure 3. The combined total for both these categories can be calculated for any particular year by subtracting the proportions in Figure 3 for that year from 100. Thus, for example, in 1985 the combined total of support for a Russian alliance plus the 'don't knows' was 16 per cent.

33. Minimal levels of support for European defence arrangements independent of the United States were also evident in responses to a somewhat similar question asked in several countries between 1979 and 1987; see Eichenberg, op. cit. in note 25, pp. 127-32.

34. Table 6 also shows attitudes to a potential increase in the national budget for military purposes as a consequence of joining NATO. In this case substantial majorities in all countries, including Poland, oppose the putative requirement. Given the economic and budgetary situations of the countries in question, this is not surprising. Also, it may be that supporters of NATO membership in the countries concerned make the calculation that NATO-based defence would be cheaper than either purely national defence or a regional defence pact.

35. Eichenberg rightly warns against 'survey questions that juxtapose positive and negative images in the abstract' (1989, p. 25). The above discussion suggests that it may be even worse to juxtapose an abstract positive image and a concrete negative image or vice versa.

36. The wording of the question was: 'Some say the countries of Central and Eastern Europe, such as the Czech Republic, Hungary, Poland and Slovakia, should become member states of the European Union. What is your opinion on this? Should they become members in less than 5 years, in the next 5 to 10 years, in over 10 years?' A negative response category was not included in the question but spontaneous negative responses were recorded. Perhaps because of the absence of a negative category, the level of 'don't knows' was quite high (21 per cent for the EU as a whole and 40 per cent or more in Ireland, Portugal and Spain)

37. Karl W. Deutsch, Political Community and the North Atlantic Area (Princeton, NJ: Princeton University Press, 1957), p. 5. The prominence given to the notion of 'sense of community' in Deutsch's definition of 'security-community' is a telling reminder of the extent to which public opinion and political culture can affect security issues in a way that goes far beyond the simple matter of support for or opposition to particular policies. Sense of community is defined as 'a matter of mutual sympathies and loyalty; of "we-feeling", trust, and mutual consideration; of partial identification in terms of self-images and interests; of mutually successful predictions of behaviour, and of co-operative action in accordance with it . . .' (ibid., p. 36). For a discussion of the implications of the Deutsch theory for the development of international cooperation and integration, see Sinnott, op. cit. in note 5.

38. Note that the level of 'don't knows' in response to this question is extraordinarily low (4 per cent for the European Union as a whole). The point is examined in some detail in a later section of the paper.

39. Evidence on perceptions and knowledge of the level at which foreign policy and defence policy are decided is considered below.

40. Richard Sobel, 'Polling on Foreign Policy Crises: Creating a Standard Set of Questions', in The Public Perspective, February/March 1996, pp. 13-16.

41. Ibid., p. 13.

42. Ibid.

43. On the question of knowledge of conflict situations abroad, Wybrow notes that, while 16 per cent of the British public in 1992 claimed to be following events in Yugoslavia very closely and a further 52 per cent said they were following events fairly closely, 'this claimed interest in events . . . was not . . . matched by knowledge of the country: 60 per cent, for example, did not know which of the three countries that made up Yugoslavia the United Nations had an economic embargo against.' See Robert J. Wybrow, 'British Attitudes towards the Bosnian Situation.' Paper prepared for presentation at the joint meeting of the American Association of Public Opinion Research and the World Association of Public Opinion Research, Salt Lake City, Utah, May 1996.

44. For a detailed analysis of the reaction of Dutch public opinion to Dutch military involvement in peacekeeping efforts in Bosnia, see Philip Everts, 'The "body-bag" hypothesis as alibi: public support for military UN operations in the Netherlands - the case of Bosnia-Herzegovina', paper presented at the International Conference on 'Public Opinion, Democracy and Security Policy', Siena, 7-10 October, 1996.

45. Thomas Risse-Kappen, 'Anti-nuclear and Pro-Detente? the transformation of the West German security debate', in Don Munton and Hans Rattinger (eds.), Debating National Security: the Public Dimension (Frankfurt: Verlag Peter Lang GmbH, 1991), p. 288.

46. For a preliminary report on this research see Jean Blondel, Richard Sinnott and Palle Svensson, 'Institutions and Attitudes: Towards an Understanding of the Problems of Low Turnout in the European Parliament Elections of 1994', EUI Working Paper, RSC no. 96/19 (Florence: European University Institute, 1996).

47. EB 43, Spring 1995.

48. As specified in the Maastricht Treaty, joint decision-making on defence is a matter for the future; see Article J.4 quoted in the introduction above.

49. The German public does see a difference but in the wrong direction: 45 per cent see defence as a matter subject to joint decision and only 38 per cent say the same for foreign policy. In general, the accuracy of perceptions seems to be related to the degree of controversy over the issue in the particular country. This is further confirmation of the importance of the notions of the salience and politicization of issues discussed in the first part of this paper.

10/15/97
SDS&EGH,Paris for WEU/ISS