European public opinion and the European Union : the knowledge gap

European Public Opinion and the European Union: The Knowledge Gap Richard SINNOTT University College Dublin Working Paper n.126 Barcelona 1997 Introduction In contrast to the situation just a few years ago, it is now well established, both in theory and in practice, that public opinion is a significant factor in the process of European integration1. This significance will almost certainly continue; it may even increase. Although there has also been considerable growth in research in this area, most of the research has focused on evaluations of European integration and has tended to pay less attention to how much people actually know about the European Community or European Union. In fairness, the neglect of the knowledge factor has been due in the main to a dearth of data. Thus, Wessels, in a comprehensive analysis of the development of support for integration, attempted to deal with the cognitive dimension but had to recognize that, given the available data, this could only be tackled indirectly: “One possibility might be to use direct measures of knowledge or perception. Another possibility is to tackle the question indirectly by asking whether people have developed an orientation towards the EC. For practical purposes we choose the second method”. (Wessels, 1995, p.110). In the same volume Niedermayer and Sinnott analysed measures of knowledge but only in relation to the one European institution for which such measures were then available, namely, the European Parliament. In short, data availability has greatly hampered research on the cognitive dimension of attitudes to the European Union2. It can be argued that the lack of detailed research of this kind is more serious in this instance than it might be in other contexts. The probability is that, given the remoteness and the embryonic nature of the European political system, attitudes to it are likely to be less well informed and less well structured than attitudes to national politics. Consequently, a clearer and more comprehensive picture of the knowledge or lack of knowledge underlying attitudes to European integration is essential in any attempt at a fuller understanding of the nature of European public opinion. Against this background, the paper begins with a brief consideration of the quality of the data on how much the public knows about the European Community or European Union3. It then explores the criteria that ought to be applied in assessing the adequacy of whatever level of knowledge is found to exist. This is followed by a description of how much people actually know about the European Union, focusing mainly on the situation in the spring of 1993 and paying particular attention to perceptions of decision-making and representation. This account is supplemented by consideration of some evidence from the spring of 1995 on knowledge of European and national decision-making in three particular policy areas. The paper then turns to the “so what?” issue, examining some empirical evidence regarding the effects that variations in levels of knowledge appear to have on attitudes to integration. Having shown that knowledge matters, the paper attempts to determine why people have different levels of knowledge. Some indication of this can be found in the public’s perception and evaluation of the available sources of EU information; if people do not know where to find the information or if they tend to seek it from sub-optimal sources or if they lack confidence in the sources, the end result is likely to be low levels of information. It may well be, however, that the public’s knowledge is conditioned by more fundamental socio-demographic and political factors; it may also be conditioned by the domestic political and institutional context within which people learn about and react to European developments. Accordingly, the paper presents a multivariate analysis of the sources of differences in levels of knowledge, taking into account not just communication variables and sociodemographic factors but also variables that seek to capture some of the variation in the way in which the European Union is experienced across the member states. A brief concluding section provides a summary and some reflections on the findings. The available data The main body of data in this area is the series of Eurobarometer surveys conducted for the European Commission. Over a long number of years these have asked repeatedly about levels of awareness of various Community institutions. With rare exceptions, the examination of the cognitive aspect of orientations towards the Union in the Eurobarometer series did not in the past go beyond this rather vague and intangible notion. More recently, the quality of Eurobarometer data on knowledge of the European Union has improved substantially. Specifically, Eurobarometer 39 and 40 (spring and autumn 1993) have considerable data on this topic and subsequent Eurobarometers have extended these detailed investigations by examining knowledge of additional aspects of EU institutions or policies. The analysis that follows focuses mainly on the data from Eurobarometer 39 and introduces evidence from later Eurobarometers where this throws further light on the matter. Before proceeding to examine this evidence, however, it is necessary to consider the criteria by which the levels of knowledge which prevail might be assessed. How much is enough? One way of attempting to answer this question is to take account of how European citizens themselves judge their level of knowledge. According to the spring 1993 Eurobarometer, a very substantial majority (71 per cent) of Europeans felt that they were not well informed about the European Community as compared with 27 per cent who felt they were well informed. The range runs from the 54 per cent feeling not well informed in the Netherlands to 79 per cent in Italy (Eurobarometer, 1993a, p. A30). Before leaping to the conclusion that there is a great hunger out there among the public for more Euro-information, however, it should be noted from Eurobarometer 40 that only 23 per cent agreed with the statement “I really need to know a lot more about the European Community”; whereas 38 per cent chose the response “I would like to have some more information about the European Community” and 28 per cent opted for “As far as I am concerned, I am happy with what I already know” (Eurobarometer, 1993b, p. A37). A more important yardstick in assessing levels of EU knowledge is the level of knowledge of national politics and the national political system. This is a useful corrective to the danger of taking low levels of EU knowledge out of context and drawing dire conclusions. Perhaps knowledge of the national political system is just as low and the problem really lies with knowledge of politics and of political institutions as such. Accordingly, the presentation of the data on levels of knowledge of the European Union that follows will be accompanied by consideration of some evidence of knowledge of national politics. Finally, if one were to take a normative perspective, either one derived from democratic theory or from a commitment to European integration, one could perhaps establish a yardstick by which to evaluate the existing state of knowledge among the public. For example, from an integrationist perspective it could be argued that, given the nature of the issues raised by the current phase of integration, the successful completion of the European project requires a high level of public knowledge and understanding of the process and of the issues involved on the part of the European public. How else is support for complex and far-reaching developments such as European Monetary Union to be assured? The problem is that it would be extremely difficult to quantify what either the minimum or the desirable level of knowledge should be. Accordingly, this paper will confine itself to examining this aspect of the matter from an empirical perspective by investigating the relationship between levels of knowledge and support for integration. Knowledge of European affairs A comprehensive account of prevailing levels of knowledge would require a survey dedicated to this topic and any small set of questions will have certain limitations. The knowledge questions in Eurobarometer 39 have a strong institutional bias: they ask about the membership of the Community, the location of the Commission, the name of its president, the names of incumbent Commissioners, which is its most powerful institution (in the sense of having the final say on legislation) and who elects the MEPs4. Despite this institutional orientation, however, these questions can probably be treated as indicative of more general levels of knowledge and understanding. In any event, knowledge of how the Union functions is one of the key aspects of an adequate public understanding of European affairs and has a particularly important bearing on the problem of the democratic deficit and on the legitimacy of the Union. A series of eight questions dealing with knowledge of the above topics can be used to form an index of knowledge of the EU; scores on the index cover a thirty point range (from 0 to 29)5. Dividing such a scale up into discrete levels evaluating the significance of each level is difficult and inevitably somewhat arbitrary. Two considerations should be borne in mind in this regard. First, most of the questions are very simple, indeed it might be argued that some are simplistic, and they might be regarded as no more than the minimum that a reasonably well informed European citizen ought to know about the Union (the exception to this is the question about the membership of the Commission). The second point to bear in mind in interpreting scores on such an index is that, with several of the questions offering only two alternatives, guessing is easy and relatively rewarding. This results in an inflation of the scores relative to the real level of knowledge. Bearing this in mind, the scores are divided as follows in Figure 1: 0-5: “no knowledge”; 6-11: “very little knowledge”; 12-17: “some but not much knowledge”; 18-23: “moderately well-informed” and 24-29: “very well informed”. On this interpretation of the scale, 10 per cent of the citizens of the European Union are very well informed and a further 24 per cent are moderately well informed. This means, however, that two-thirds (65 per cent) emerge, with “some but not much knowledge” or less. This less well informed majority of the European citizenry comprises three groups: 26 per cent with “some but not much knowledge”, 24 per cent with “very little knowledge” and 15 per cent with “no knowledge” at all. If the countries are ranked on the basis of the combination of the proportions who are very well informed and moderately well informed, Denmark and Luxembourg have the most knowledgeable publics, followed by Belgium, Greece and France. The least knowledgeable are Italy and the former East Germany, followed by Northern Ireland, Britain and West Germany6. It was suggested above that, before any hard and fast conclusions are drawn about levels of knowledge of the EU, such levels need to be seen in the context of how much people know about the national political system. It is extremely difficult to devise precisely comparable scales of knowledge at the national and supranational levels and this difficulty is reflected in the fact that the scale used in the Eurobarometer 39 report to measure national knowledge is based on only four questions, compared to seven questions in the European knowledge scale in the same report. The four items cover the name of the capital of the country, the name of the prime minister, the institution having the final say on legislation and the identity and role of the head of state. Accepting that a scale based on such questions provides only an approximate basis for comparison of national and European levels of knowledge, the contrasts are still striking. Across the twelve member states in spring 1993, 78 per cent showed high to very high levels of knowledge of the national political system in comparison to only 38 per cent showing such levels in relation to the European system7 (Eurobarometer, 1993a, p. 55-59). More detailed examination of the responses to two of the individual questions that go to make up the index of European knowledge reveals particular problems in regard to knowledge of key aspects of the functioning of the institutions of the European Union. The first of these two questions deals with which of the European institutions is seen to be most powerful «in terms of having the final say on European Community legislation». Across the Union, one quarter of the citizens have no opinion on this matter. A plurality (35 per cent) attribute decisive power (erroneously) to the European Parliament (see Table 1). A further 20 per cent express the stereotypical view that the bureaucrats in Brussels (i.e. the Commission) decide everything; the correct answer (the Council of Ministers) was chosen by one-in-five. The only countries with a significantly above average correct response are Denmark, Luxembourg and, to a somewhat lesser extent, the Netherlands. Given the novelty of their accession to the Community and indeed the novelty of their experience of democracy, it is perhaps not surprising that it is the people of former East Germany who are most prone to exaggerate the role of the European Parliament. Likewise, given the stereotypes prevalent in the British debate about European integration, it comes as no surprise that the greatest tendency to exaggerate the role of the Commission is found in the United Kingdom (see Table 1). The most important feature of all of this, however, is the lack of awareness of the role of the Council of Ministers: 79 per cent either don’t know who decides or don’t realise that the power of decision lies with the representatives of the national governments. This suggests that the subtleties of elite debates about intergovernmentalism versus supranationalism and about the intricacies of qualified majority voting in the Council of Ministers do not have much resonance in public opinion. It might be argued that, erroneous though the notion that the European Parliament has the final say in legislation is, it establishes some degree of legitimacy for the European policy process. However, other considerations aside, the degree of legitimacy that might be regarded as being conferred in this way must be seen as highly attenuated in virtue of answers to the second of the questions singled out above. Asked the apparently simple question «who elects the members of the European Parliament?» and presented with a card showing two incorrect and one correct option, only two out of five respondents chose the correct response (see Table 2); this was after three rounds of direct elections and a mere year or so before a fourth round. Again quite understandably, former East Germany shows the lowest rate of correct response (27 per cent). It is also notable that three of the four countries with reasonably high levels of correct response (Belgium, Luxembourg and Greece) have compulsory or at least quasicompulsory voting which results in very large proportions of the adult populations in these countries participating in the very process referred to in the question “who elects the members of the European Parliament?”. On the other hand, Italy, which had compulsory voting until 1993 and continues to have very high turnout in European elections, has a lower than average level of correct response (35 per cent) to the question and Denmark, which has relatively low levels of turnout in European Parliament elections, is among the countries with the highest levels of knowledge of who elects the European Parliament. Mention of Denmark draws attention to another institutional or contextual variable that may affect the level of popular involvement with European issues and that may therefore lead to increased levels of knowledge, namely, the holding of referendums on EU issues. Of course, knowledge could also be increased by institutional factors having little or nothing to do with participation, for example, the mere presence of a major European institution in a country. These and other contextual variables will be considered in exploring the origins of differences in levels of knowledge later in this paper. In order to fill out this account of the state of knowledge of European affairs among the European public, it is worth noting some data from a more recent Eurobarometer which provide evidence on knowledge in relation to certain key policy sectors. In Eurobarometer 43 respondents were asked about their perception of the allocation of decision-making power over a wide range of policies 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 listed in the question were three which provide a particularly telling test of public awareness of what is arguably the key to the whole integration process, i.e. the sharing of decision-making power between the national and supra-national levels. The three issues were: ‘agriculture’, ‘foreign policy towards countries outside the EU’ and ‘defence’. Since it became fully established in the late 1960s, the Common Agricultural Policy has been the preeminent common policy of the Community and one of the clearest examples of the transfer of decision-making to the European level. One would presume that after some twenty five years of the CAP, much of it marked by controversy and frequently punctuated by all-night negotiating sessions in Brussels, the fact that that agricultural policy is “at least to some extent decided jointly within the European Union/decided at the European Union level” would be widely recognised by the mass public. Given that two treaty changes and a good deal of recent debate about European integration have focused on the development of a common foreign policy, one would also expect a reasonably widespread public perception of the actual decision-making process in regard to foreign and defence policy, i.e. that the former is decided “at least to some extent” at EU level and the latter is not8. Despite the assumed salience of the CAP, in only five of the fifteen member states do majorities of the citizens realize that agricultural policy is decided “at least to some extent” at European level. The realization is most widespread in Finland (69 per cent) and Denmark (64 per cent), followed by Sweden (58 per cent), the Netherlands (56 per cent) and Luxembourg (56 per cent). At the other end of the scale one finds the four Mediterranean countries (Portugal 26 per cent; Italy 27 per cent; and Greece and Spain 34 per cent). While some of the dearth of knowledge in these latter countries might be put down to the northern European orientation of the CAP over the years, the fact remains that only 35 per cent of people in the main paymaster-country (Germany) realize that agricultural policy is made at European level and the realization remains a minority one in two of the main beneficiaries -Ireland (42 per cent) and France (48 per cent) (see Figure 2). Turning to perceptions of foreign and defence policy, the bar-chart in Figure 3 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 Union. 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 Union level; on foreign policy, however, only 46 per cent of Finns get it right. After Finland, the next most informed in regard to defence policy are Sweden, Ireland, Austria and Greece but, even more than Finland, all of these are marked by a serious under-estimation of the European input into foreign policy-making. Denmark is the best informed in regard to the European dimension of foreign policy formulation but a very substantial minority of Danes erroneously attribute decision-making on defence to the European Union. 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; furthermore, the publics in these countries see no difference in European Union involvement in decision-making on foreign policy and decision-making on defence9. In summary, whether measured by means of some composite index of knowledge or by individual items such as those relating to legislative power in the European Union or to the mode of election to the European Parliament or to the decision-making process in key areas such as agricultural policy, and foreign and defence policy, the level of knowledge of Europe and European affairs among the mass public seems lamentable. But lamentable in what sense? In other words, do low levels of knowledge have any practical political consequences? The implications of ignorance There is indeed no guarantee that increased knowledge of the workings of the European Union would lead to increased support for it. Eurosceptics would no doubt argue that the contrary outcome is the more likely and the evidence does indicate that there are some very well informed opponents of European integration. However, an examination of the relationship between levels of sense has a very substantial positive effect on levels of knowledge (beta 0.194). But it is by no means the end of the story: even among people with the same high or low degree of interest in European politics, level of knowledge is affected by their access to a whole range channels of communication. The most obvious and most effective channel of communication is overall media consumption: whatever the level of interest, the greater the media consumption, the higher the level of knowledge (beta 0.123). Discussion of politics reinforces this effect (0.081), as does a sense of attachment to a political party (0.034). It is true that party attachment (“feeling close to a political party”) does not appear to have as strong an effect as media consumption or political discussion; nonetheless it does seem to act to some degree as a channel of communication which links the individual citizen to the wider political world and enhances knowledge of that world. Apart from these general channels of political communication, level of knowledge is also affected by orientation to specific sources of knowledge about the EU. These orientations or tendencies to resort to and to trust various sources of information have been described in some detail above. A difficulty in dealing with the data in Tables 3 and 4 was that there were so many discreet items of information; this difficulty is compounded if one is seeking to assess the effects of these various orientations. In an attempt to deal with this problem, factor analyses of the responses to the questions on source of information and trust in sources were carried out. These showed that there are indeed clear patterns underlying the individual items. Four factors emerge from the factor analysis of responses to the question “when you are looking for more information about the European Community, its policies, its institutions, where do you look first?”. The four factors are: a current affairs source-orientation, an official source-orientation, a collective/corporatist source-orientation and, finally, the absence of any sourceorientation. Three factors (trust-distrust of the media, trust-distrust of collective or corporatist sources and trust-distrust of official sources) emerge from the analysis of the trust-in-sources items15. The question is: do these sourceorientations and source-evaluations affect knowledge of the EU? The situation seems to be that having an orientation to consult various sources of information can make a difference and that the degree of difference depends on the nature of the orientation. A current affairs source-orientation (looking both to newspapers and magazines and to discussions with family, friends and colleagues as sources of information) has a substantial positive effect on knowledge (beta 0.115); a tendency to look to official sources has a slightly less but still quite positive impact (0.093); a collective/corporatist source orientation (expecting to acquire information at meetings, or from trade union, professional or consumer associations) has a much smaller impact on level of knowledge. Not surprisingly, the absence of any source-orientation (literally not knowing where to look for information) has a significant negative effect. The effects of trust in various sources is more complex. Distrust of official sources has no effect one way or the other. Distrust in the media, however, is associated with lower levels of knowledge, though the effect is slight (beta -0.015 with a significance level (0.062) just above the conventional cut-off point). In the case of collective or corporatist sources, there seems in fact to be such a thing as a healthy scepticism: a syndrome of distrust in trade unions or professional associations, churches, business, and schools and universities as sources of information about the European Union16 actually has a small but significant positive effect on the respondent’s level of knowledge (beta 0.033). Overall then the evidence suggests that considerable progress can be made in identifying the sources of low levels of knowledge of European affairs. We know that, irrespective of individual-level characteristics, if a person comes from a country which has high turnout in European Parliament elections, which hosts a major EU institution, which has joined the Community in one of the later waves of accession, and, or which holds referendums on EU issues, that person’s knowledge of the EU is likely to be higher than it would otherwise be. In addition, knowledge is adversely affected by a range of socio-demographic variables: low levels of education, a manual occupation, and advancing years. Less predictable but quite striking is the impact of gender: other things being equal, women are less informed than men17. The tendency for farming as an occupation to reduce levels of knowledge may also be regarded as surprising in view the dependence of most farmers on the European Union for much of their income. In terms of attitudes and orientations, interest in EU affairs plays an important but not an overriding role. Media use, political discussion and party attachment also help; their impact is consistent with the fact that a tendency to look to current affairs sources has a greater positive effect on knowledge than a tendency to look to collective or corporatist sources. Finally, scepticism as to the trustworthiness of various sources is either neutral in the case of official sources, negative if it is directed at media sources and actually positive, that is associated with a higher level of knowledge, if it relates to collective or corporatist sources. Conclusion Impending political developments and the overall challenges facing the European Union suggest that a fairly rigorous yardstick should be applied in assessing the adequacy of the public’s knowledge of European affairs. By any standards, however, the level of European knowledge is low. Lack of knowledge relates to a range of institutional and policy areas and involves in particular an overestimation of the power of the Commission and of the Parliament and an underestimation of the role of the Council of Ministers. Furthermore, a large segment of public opinion seems unaware of the Parliament’s representative character. All of this may contribute to a feeling that ‘Europe’ is something out there that imposes its decision on ‘us’ and to an underestimation of the degree of accountability that actually exists. The evidence certainly suggests that low levels of knowledge diminish support for the European Union, for the Maastricht Treaty and for further developments in integration. For the most part this is because the ill-informed are indecisive or ambivalent but there is also some evidence that ignorance can generate opposition. The findings regarding the origins of ignorance of the European Union and European affairs have been summarized in some detail in the section immediately preceding this conclusion and need not be rehearsed again here. What is perhaps worth dwelling on, however, is not so much this or that particular finding as an overall impression that emerges from a number of the findings: knowledge is lower if one lives in a country in which accession to the Community took place a long time ago and at a stage when there was little or no controversy about membership; conversely, as well as being higher in countries that have joined more recently, it is higher in countries which have publicly debated integration issues and decided on them in referendums; it tends to be higher also in countries which have high turnout in European Parliament elections; at the individual level, knowledge increases with interest in European politics, frequency of political discussion, feeling close to a political party and tending to look to current affairs coverage and discussions as sources of EU information. This adds up to an overall picture in which politics and the politicization of European issues enhances knowledge. Since European issues are constantly becoming more politicised, this may be taken as an encouraging sign from an integrationist perspective. It also, however, implies a dilemma: the politicization of issues may increase knowledge, which may in turn lead to greater levels of support for integration; on the other hand, politicization may also generate opposition. The question is which tendency will predominate? The answer to this question will determine whether the outcome of the politicization of European issues will be a public opinion that is increasingly informed and supportive or one that remains illinformed and increasingly opposed. Appendix1 Appendix2 Table1 Table2 Table3 Table4 Table5