The Influence of Political Events on Attitudes Towards the European Union

Abstract

This two-wave panel study was designed to investigate the effects of the media coverage leading up to and including an important European Union event (a summit meeting of EU leaders) on citizens’ attitudes towards the EU and European integration. A random sample of 817 citizens in the Netherlands was surveyed one month before the Amsterdam Summit in June 1997 and three days after it had ended. Two types of attitudes towards Europe were distinguished by scaling analysis: (a) national–pragmatic attitudes towards the EU and (b) supranational–idealistic attitudes towards the EU. Results indicated that supranational–idealistic attitudes were influenced positively as a result of the media coverage related to the summit, whereas national–pragmatic attitudes did not show a significant change. A control variable, attitudes towards immigrants, which was included to detect possible testing effects, showed no change. Effects of the summit’s media coverage were in the same direction across all levels of political knowledge and political attentiveness. When predicting change in supranational–idealistic attitudes, controlling for the original attitude and political knowledge, those who were most attentive to politics were more strongly influenced. These findings challenge traditional views of the impact of knowledge, attention and interest on attitude change.

Full text

British Journal of Political Science
http://journals.cambridge.org/JPS
Additional services for British Journal of Political Science:
Email alerts: Click here
Subscriptions: Click here
Commercial reprints: Click here
Terms of use : Click here
The Inuence of Political Events on Attitudes Towards the
European Union
HOLLI A. SEMETKO, WOUTER VAN DER BRUG and PATTI M. VALKENBURG
British Journal of Political Science / Volume 33 / Issue 04 / October 2003, pp 621 - 634
DOI: 10.1017/S0007123403000280, Published online: 17 September 2003
Link to this article: http://journals.cambridge.org/abstract_S0007123403000280
How to cite this article:
HOLLI A. SEMETKO, WOUTER VAN DER BRUG and PATTI M. VALKENBURG (2003). The
Inuence of Political Events on Attitudes Towards the European Union. British Journal of Political
Science, 33, pp 621-634 doi:10.1017/S0007123403000280
Request Permissions : Click here
Downloaded from http://journals.cambridge.org/JPS, IP address: 145.18.72.97 on 11 Dec 2014

B.J.Pol.S. 33, 621–634 Copyright 2003 Cambridge University Press
DOI: 10.1017/S0007123403000280 Printed in the United Kingdom
The Influence of Political Events on Attitudes
Towards the European Union
HOLLI A. SEMETKO, WOUTER VAN DER BRUG AND
PATTI M. VALKENBURG*
This two-wave panel study was designed to investigate the effects of the media coverage leading up to and
including an important European Union event (a summit meeting of EU leaders) on citizens’ attitudes towards
the EU and European integration. A random sample of 817 citizens in the Netherlands was surveyed one month
before the Amsterdam Summit in June 1997 and three days after it had ended. Two types of attitudes towards
Europe were distinguished by scaling analysis: (a) national–pragmatic attitudes towards the EU and (b)
supranational–idealistic attitudes towards the EU. Results indicated that supranational–idealistic attitudes were
influenced positively as a result of the media coverage related to the summit, whereas national–pragmatic
attitudes did not show a significant change. A control variable, attitudes towards immigrants, which was
included to detect possible testing effects, showed no change. Effects of the summit’s media coverage were
in the same direction across all levels of political knowledge and political attentiveness. When predicting
change in supranational–idealistic attitudes, controlling for the original attitude and political knowledge, those
who were most attentive to politics were more strongly influenced. These findings challenge traditional views
of the impact of knowledge, attention and interest on attitude change.
One of the European Union’s routine institutional practices is the gathering together of
heads of state to negotiate agreements and sign treaties. Since the location of EU summit
meetings has rotated with the EU presidency, a different EU country has played host to
one or more of these events. These events usually generate considerable press coverage
in the host country and other EU countries, and represent the high points in the generally
low attention paid to EU matters in the news in most countries.
1
This study explores
changes in attitudes towards European unification as a result of the media coverage of the
summit meeting that took place in Amsterdam in June 1997.
With the political leaders from EU countries visiting Amsterdam, EU politics dominated
the Dutch national news stream for a brief period of time. This setting thus provided a
unique opportunity to conduct a quasi-experiment to assess the extent to which this highly
visible event about EU leadership and common issues affected citizens’ attitudes towards
European integration. Moreover, it enables us to study whether changes in attitudes are
* Semetko: Halle Institute of Global Learning and Department of Political Science at Emory University; Van
der Brug and Valkenburg: Department of Communication Science at the Amsterdam School of Communications
Research (ASCoR) at the University of Amsterdam. The authors would like to thank the Journal’s anonymous
reviewers for their helpful comments, and also the Royal Netherlands Academy of Sciences (KNAW) and ASCoR
at the University of Amsterdam for providing funding for this research. An earlier version of this article was
presented at the Annual Meetings of the American Political Science Association in Washington, D.C., 2000.
1Pippa Norris, A Virtuous Circle (Cambridge: Cambridge University Press, 2000).

622 SEMETKO,VAN DER BRUG AND VALKENBURG
related to measures of political involvement. These questions have so far been studied
primarily with survey data in an American context.
2
Our contribution is important because
it allows us to assess whether previous findings are also valid in a non-US context.
Moreover, our study goes beyond previous research by combining the advantages of a
panel-survey design and a quasi-experimental design.
Our study is also important because of the potential impact these summits may have on
elite perceptions and public opinion about the European integration process. Turning
points in EU history are often marked by reference to these summits; the Treaty of Nice
in 2000 is one of the most recent examples. Yet we know little about whether these events,
which are important symbolically as well as practically, have any effect on public opinion
about the ongoing processes of European integration. Recent developments make
particularly relevant the question of whether such events have an impact on opinion about
EU institutions and plans. In September 2000, with a turnout of 85 per cent, a majority
of Danes voted in a referendum against joining the euro, the common EU currency that
is now standard in eleven countries. Britain and Sweden are expected to hold referendums
on this issue within the terms of the next parliaments. The Treaty of Nice signed in
December 2000 promised to further integration and to open the EU to a number of central
European countries within the next few years. Each concrete step to further integration
made by political elites is questioned by vocal and sometimes sizeable publics in EU
countries. It is important to study the effects of these events on public opinion because the
legitimacy of elite actions in the direction of furthering European integration depends
partially upon the level of public support for such moves.
3
2See, for example, Philip Converse, ‘Attitudes and Non-attitudes: Continuation of a Dialogue’, in E. R. Tufte,
ed., The Quantitative Analysis of Social Problems (Reading, Mass.: Addison-Wesley, 1970), pp. 206–61; David
H. Weaver, Doris A. Graber, Maxwell E. McCombs and C. E. Eyal, Media Agenda-Setting in a Presidential
Election: Issues Images and Interest (New York: Praeger, 1981); Jon A. Krosnick and Donald R. Kinder, ‘Altering
the Foundations of Support for the President through Priming’, American Political Science Review,84(1990),
499–512; Jon A. Krosnick and Laura A. Brannon, ‘The Impact of the Gulf War on the Ingredients of Presidential
Evaluations: Multidimensional Effects of Political Involvement’, American Political Science Review,87(1993),
963–75; Larry M. Bartels, ‘Messages Received: The Political Impact of Media Exposure’, American Political
Science Review,87(1993), 267–85; Benjamin I. Page and Robert Y. Shapiro, The Rational Public: Fifty Years
of Trends in Americans’ Policy Preferences (Chicago: University of Chicago Press, 1992); John R. Zaller, The
Nature and Origin of Mass Opinion (Cambridge: Cambridge University Press, 1992); John R. Zaller and M. Hunt,
‘The Rise and Fall of Candidate Perot: The Outsider Versus the Political System – Part II’, Political
Communication,12(1995), 97–123; Donald R. Shaw, ‘A Study of Presidential Campaign Event Effects from 1952
to 1992’, Journal of Politics,61(1999), 387–422; Pippa Norris, John Curtice, David Sanders, Margaret Scammell
and Holli A. Semetko, On Message: Communicating the Campaign (London: Sage, 1999). And in experimental
settings, see Shanto Iyengar, Donald R. Kinder, M. D. Peters and Jon A. Krosnick, ‘The Evening News and
Presidential Evaluations’, Journal of Personality and Social Psychology,46(1984), 778–87; Shanto Iyengar,
Mark D. Peters and Donald R. Kinder, ‘Experimental Demonstrations of the “Not-So-Minimal” Consequences
of Television News Programs’, American Political Science Review,76(1982), 848–58; Shanto Iyengar and
Donald R. Kinder, News That Matters (Chicago: University of Chicago Press, 1987); Steve Ansolabehere and
Shanto Iyengar, Going Negative: How Political Advertisements Shrink and Polarize the Electorate (New York:
Free Press, 1995).
3See, for example, Richard Sinnott, ‘Bringing Public Opinion Back In’, in Oskar Niedermayer and Richard
Sinnott, eds, Public Opinion and Internationalized Governance (Oxford: Oxford University Press, 1998),
pp. 11–32; Jacques Thomassen and Hermann Schmitt, ‘In Conclusion: Political Representation and Legitimacy
in the European Union’, in Hermann Schmitt and Jacques Thomassen, eds, Political Representation and
Legitimacy in the European Union (Oxford: Oxford University Press, 1999), pp. 255–68; Robert Rohrschneider,
‘The Democracy Deficit and Mass Support for an EU-wide Government’, American Journal of Political Science,
46 (2000), 463–75.

Influence of EU Political Events on Attitudes to the EU 623
ATTITUDE CHANGE
The field of public opinion and attitude research has generated considerable interest in
uncovering what lies behind changes in public opinion. Philip Converse’s (1964) seminal
article argued that most Americans have no consistent pattern to their opinions or beliefs,
and in particular that opinions did not depend on ideological principles.
4
His analysis of
NES panel data from 1956, 1958 and 1960 revealed that so many people change their
opinions on policy issues that it appeared as if people were answering at random.
His general conclusion was that people’s policy attitudes were for the most part
non-attitudes.
5
From this perspective then, movement in public opinion is largely
explained by a lack of knowledge, interest and ideology that results in a certain randomness
of opinion or non-attitudes.
6
Of the many contrasting views on the state of public opinion and what moves it, the work
by Page and Shapiro stands out because it describes a ‘rational public’ whose opinions are
moved by information in a way that reveals consistency between policy preferences and
basic values.
7
They focus on aggregate public opinion in the United States and show that
it is largely stable over time on many aspects of public policy. This does not necessarily
contradict Converse, who described individual-level opinion change as largely random.
These random changes at the individual level could appear stable at the aggregate level.
Page and Shapiro show, however, that public opinion moves in response to events or new
information about an issue, and they identified parallel publics (cohorts or demographic
groups) who used similar standards to assess issues and whose opinions largely moved in
the same direction. This suggests that opinion is not random and that attitudes exhibit
consistency. More recently, other survey-based studies have also found that public opinion
moves in response to (campaign) events and as a result of new information that becomes
available.
8
The literature also does not provide a simple or clear-cut answer about what to expect
concerning the relationship between attitude change, on the one hand, and political
involvement, on the other. In research on media effects, political involvement is usually
defined as having three components: knowledge, interest and exposure to information.
These individual components are usually positively correlated with one another.
According to the most basic model, we would expect those citizens with the lowest levels
of exposure, interest and knowledge to change only randomly, while the attitudes of those
who were better informed and more aware would remain stable or would display a
4Philip Converse, ‘The Nature of Belief Systems in Mass Publics’, in David E. Apter, ed., Ideology and
Discontent (New York: Free Press, 1964), pp. 206–61.
5See also Converse, ‘Attitudes and Non-attitudes’.
6Subsequent research by Pomper (1972) and Nie et al. (1976) suggested that Converse’s findings were
time-bound and that ideological awareness, issue awareness and attitude consistency had increased over the 1960s.
See Gerald M. Pomper ‘From Confusion to Clarity: Issues and American Voters 1956–1968’, American Political
Science Review,66(1972), 415–28; Norman H. Nie, Sidney Verba and John R. Petrocik, The Changing American
Voter (Cambridge, Mass.: Harvard University Press, 1979 [1976]). Disputes over the evidence on attitude
consistency, however, persisted. A general discussion of this debate is found in Carol J. Glynn, Susan Herbst,
Gerald J. O’Keefe and Robert Y. Shapiro, Public Opinion (Boulder, Colo.: Westview, 1999), at pp. 249–98.
7Page and Shapiro, The Rational Public.
8See, for example, Bartels, ‘Messages Received’; Zaller and Hunt, ‘The Rise and Fall of Candidate Perot’;
Shaw, ‘A Study of Presidential Campaign Event Effects from 1952 to 1992’; Wouter van der Brug, ‘Perceptions,
Opinions, and Party Preferences in the Face of a Real World Event: Chernobyl as a Natural Experiment in Political
Psychology’, Journal of Theoretical Politics,13(2001), 253–80.

624 SEMETKO,VAN DER BRUG AND VALKENBURG
systematic pattern of change.
9
This led experimental researchers to distinguish between
political ‘novices’ and political ‘experts’ when studying the effects of watching television
news, for example, on public perceptions of the importance of issues.
10
In The Nature and
Origin of Mass Opinion, John Zaller argues that a more detailed distinction in levels of
attentiveness is necessary, because attitude change may be non-linearly related to political
attention. He expects citizens with intermediate levels of attentiveness to be most likely
to be affected by news. This argument is empirically supported by Ansolabehere and
Iyengar, who find that the effects of (negative as well as positive) political advertisements
are strongest among intermediately informed citizens.
11
Since the operationalization of this
distinction was not always consistent across studies and usually based on only one aspect
of involvement, such as knowledge, we have to agree with Krosnick and Brannon’s
conclusion: ‘the empirical evidence on the moderating role of political expertise is actually
rather inconsistent.’
12
That led them to investigate the separate and combined roles of
knowledge, interest and exposure in the ‘priming’ of evaluations of US President George
Bush, over the period surrounding the Gulf War when his approval ratings skyrocketed
more than 20 percentage points. Krosnick and Brannon concluded that the evidence points
towards a social psychological model of information processing:
Greater knowledge constitutes a greater ability to interpret, encode, store, and retrieve new
information. And higher levels of exposure and interest are associated with a greater likelihood
of forming on-line political evaluations and a dilution of priming effects due to a wider range
of knowledge being activated by media coverage.
13
They further argued that future research should examine these three dimensions –
knowledge, interest and exposure – in studying attitude change. We investigate all three
aspects of involvement in this study.
ATTITUDES ABOUT EUROPEAN INTEGRATION
There are various possible attitudes concerning some aspects of the European integration
process that may or may not change in response to a major event such as the summit. Some
of these attitudes can be characterized by taking into account the three aspects of the
legitimacy of the European Union – when viewed from the perspective of European
citizens – that were distinguished by Jacques Thomassen and Hermann Schmitt.
14
The first
is the extent to which a European demos exists, or how far citizens feel that they belong
to a European community.
15
They refer to this aspect as the domain of a political system.
The second is the scope of a political system, and concerns the question of what level of
government (regional, national or supranational) should be mainly responsible for policies.
The legitimacy of the European Union depends partially upon the willingness of citizens
to attribute responsibilities for policies or problem solving to it. Research shows that this
9Converse, ‘Attitudes and Non-attitudes’; Converse, ‘The Nature of Belief Systems in Mass Publics’.
10 Iyengar and Kinder, News That Matters; Iyengar, Kinder, Peters and Krosnick, ‘The Evening News and
Presidential Evaluations’.
11 Ansolebehere and Iyengar, Going Negative.
12 Krosnick and Brannon, ‘The Impact of the Gulf War’, p. 965.
13 Krosnick and Brannon, ‘The Impact of the Gulf War’, p. 972.
14 Thomassen and Schmitt, ‘In Conclusion’.
15 Angelika Scheuer, ‘A Political Community?’ in Hermann and Thomassen, eds, Political Representation and
Legitimacy in the European Union, pp. 25–46.

Influence of EU Political Events on Attitudes to the EU 625
willingness depends to a large extent upon the kinds of issues in question.
16
A third aspect
of legitimacy concerns institutions and processes of government.
In addition to these aspects of the democratic legitimacy of the European Union, other
kinds of attitudes may exist, which view European integration from a pragmatic or
self-interested perspective. This perspective takes into account one’s opinion about
whether one’s country benefited from EU membership and whether EU membership has
more advantages than disadvantages.
17
Our panel study contains measures of various
conceivable attitudes towards European integration, so that we can explore to what extent
opinions on these matters are structured by latent attitudes.
Since political elites in the EU countries move towards further European integration,
it is important to study the public support for such moves.
18
The Amsterdam Summit was
conducted in an optimistic spirit; therefore, it allows us to assess whether such an event
helped to increase the public support for European unification.
OPINION DEVELOPMENT AND CHANGE
Previous research on attitudes towards the European Union in the Netherlands has shown
that ‘opinion on European issues is not very well developed’.
19
We therefore might expect
to find that an important event in the news such as the summit would serve as a catalyst
for changes in or development of opinions. In this study, we focus on four kinds of changes
in opinions that might emerge as a result of media coverage of the summit. Given the lack
of strong opinions on EU issues, one possible change is that individuals who do not hold
opinions about Europe actually form opinions as a result of exposure to information about
the summit. A second possible development is that a common attitude structure towards
Europe and EU institutions emerges or becomes stronger as a consequence of media
attention to EU affairs. A third possible development is that citizens become more
favourable or unfavourable in their attitudes towards the European Union and European
integration. Finally, we are interested in whether attitude changes, to the extent that they
can be found, are different for citizens with different levels of political involvement.
The literatures on public opinion formation and on EU attitudes do not provide clear-cut
answers on what we should expect to find. We therefore do not offer hypotheses about the
kinds of changes we might expect to find as a result of the media attention given to this
important EU event.
METHODS
Data and Sample
The data were collected before and after the June 1997 Amsterdam Summit, known in the
Dutch press as the ‘Eurotop’ summit. The nationally representative panel study was
16 Richard Sinnott, ‘Policy, Subsidiarity, and Legitimacy’, in Niermayer and Sinnott, eds, Public Opinion and
Internationalized Governance, pp. 246–76; Lieven De Winter and Marc Swyngedouw, ‘The Scope of EU
Government’, in Schmitt and Thomassen, eds, Political Representation and Legitimacy in the European Union
pp. 47–73.
17 See, for example, Michael Marsh, ‘Policy Performance’, in Schmitt and Thomassen, eds, Political
Representation and Legitimacy in the European Union, pp. 90–110.
18 Sinnott, ‘Bringing the Public Opinion Back In’; Thomassen and Schmitt, ‘In Conclusion’.
19 Willem E. Saris, ‘The Public Opinion about the EU Can Easily Be Swayed in Different Directions’, Acta
Politica: International Journal of Political Science,32(1997), 406–35, at p. 428.

626 SEMETKO,VAN DER BRUG AND VALKENBURG
specifically designed to measure the impact of this major event on Dutch attitudes towards
the European Union. While the total sample size had 1,025 respondents in both waves,
panel mortality meant that 817 respondents participated in both waves. All analyses below
are based on the 817 respondents from the two-wave panel.
20
Design
This study is a quasi-experiment based on a nationally representative two-wave panel
study. The first wave took place over one week beginning 15 May 1997, the summit took
place on 16–17 June, and the second wave began on 20 June. The summit and the
considerable coverage it received in the news is therefore the experimental intervention.
While attitudes towards the European Union might be expected to change as a result
of the special attention given to this local high profile event, attitudes towards immigrants
could be expected to remain stable during this same period. Attitudes towards immigrants
were thus used as a control measure. This control measure was included to detect possible
testing effects.
Political Involvement
The panel data contain fifteen indicators of the three dimensions of involvement
distinguished in the literature: exposure, interest and knowledge. Scaling analysis
demonstrated clearly that these fifteen indicators formed two strong scales: political
knowledge and political attentiveness.
21
Political knowledge is measured with items about
politics or political leaders in Europe and in the country. Political attentiveness is composed
of both the interest and exposure items. Since these items form a strong unidimensional
scale and because exposure and interest were found to have highly similar effects on
attitude change (as in Krosnick and Brannon’s study), we interpret these items as indicators
of the same latent trait: attentiveness.
Attitudes Towards European Integration and Immigrants
The survey included eight items measuring opinions concerning European integration. The
results of a Mokken scaling analysis revealed that these formed two scales, and these are
discussed further in the results section. We also used four items to measure attitudes
towards immigrants, and the scale that these items formed is also discussed in our
presentation of the results.
20 The data were collected by the Netherlands Institute for Public Opinion (NIPO), using individual respondents
in their telepanel. The NIPO telepanel is a representative sample of the Dutch public and has been the basis for
numerous academic studies. See, for example, Jan Kleinnijenhuis and Jan de Ridder, ‘Issue News and Electoral
Volatility: A Comparative Analysis of Media Effects during the 1994 Election Campaigns in Germany and the
Netherlands’, European Journal of Political Research,33(1998), 413–37; Jan Kleinnijenhuis and David P. Fan,
‘Media Coverage and the Flow of Voters in Mutiparty Systems: The 1994 National Elections in Holland and
Germany’, International Journal of Public Opinion Research, 11 (1999), 233–56. We found no evidence in the
subsample that would lead us to be concerned about panel bias. The analyses presented in Table 1 and Table 2
were also done for the full sample of 1,025, with the same substantive results.
21 See Appendix attached to the website version of this article.

Influence of EU Political Events on Attitudes to the EU 627
RESULTS
Opinion Formation
In our introduction, we speculated that the first possible change in attitudes resulting from
the summit was that individuals who do not hold opinions about Europe actually form
opinions as a result of the extensive media coverage. Did the number of people with
attitudes towards European issues increase as a result of the summit? Table 1 presents the
percentage of respondents in both waves who answered ‘Don’t know’ to any of the
respective items. In order to control whether any changes are due to testing effects,
22
four
items concerning attitudes towards immigrants were also analysed. Since there are no
reasons to expect that attitudes towards immigrants would change during the month in
between the two waves of interviews, any changes in the percentages of ‘Don’t know’
answers would lead us suspect that testing effects occurred.
Table 1 shows that people are more likely to have opinions about immigrants than they
are about European matters. The number of ‘Don’t knows’ is considerably larger on the
European attitude items, which is consistent with earlier research on Dutch attitudes on
European issues.
23
A second point is that there is stability at the aggregate level in the
TABLE
1Percentage of ‘Don’t Know’ Answers in Each Wave
% ‘Don’t % ‘Don’t % ‘Don’t
know’ know’ know’
first wave second wave both waves
Attitudes towards Europe
Advantages EU membership larger than
disadvantages 38.1 35.6 23.7
European unification is moving too rapidly 22.3 21.9 13.5
European unification is a good thing 26.7 24.1 14.1
Only the big countries have a say in
European affairs 12.7 13.0 7.0
EU is a threat to small countries 16.6 14.6 8.3
Unemployment should be dealt with at
European level 16.4 16.8 8.2
Crime should be dealt with at European level 10.9 9.2 5.0
Willing to make personal sacrifices to help
other EU countries 9.1 8.2 3.4
Attitudes towards immigrants
We should welcome people from abroad 7.0 6.7 2.8
Too many people from Turkey and Morocco
in the Netherlands 8.4 7.6 3.1
Islamic cultures enrich the Dutch culture 7.1 7.1 3.5
Too many people from Surinam in the
Netherlands 11.5 9.9 4.8
22 It is a common finding in panels that changes in attitudes may result from participating in the test. We may
well expect that the sheer fact that someone is interviewed about European matters for more than half an hour
increases their interest. If that were to be the case we could erroneously attribute such a test effect to the summit
meeting.
23 Saris, ‘The Public Opinion about the EU Can Easily Be Swayed in Different Directions’.

628 SEMETKO,VAN DER BRUG AND VALKENBURG
percentages of ‘Don’t know’ answers. For none of these twelve items (European attitude
items, as well as attitudes towards immigrants) is there a significant difference between
the percentage ‘Don’t know’ answers in the first and second wave. Moreover, the small
differences in percentages of non-response are not systematically lower or higher in either
of the two waves. The third column in Table 1 shows the number of those individuals who
said ‘Don’t know’ at both waves. Since this number is considerably lower than in the other
two columns, it indicates that changes did occur at the individual level. However, in the
European attitude items as well as in the immigrant attitude items approximately one-third
to one-half of those with no opinion at the first wave changed to having an opinion at the
second wave. Since this rate of change is common to both the European attitude items and
the immigrant attitude items, we cannot conclude that the summit had any effect on the
number of citizens who held opinions on European issues.
Attitude Structures
A second possible development is that a common attitude structure towards Europe and
EU institutions emerges or becomes stronger as a consequence of media attention to EU
affairs. In order to assess the extent to which this was the case, we carried out a large series
of scaling analyses with the Mokken scaling model.
24
The Mokken scaling model is a
stochastic variant of the deterministic Guttman scale. It is designed to test whether a set
of items that theoretically measure the same latent trait meet the criteria of unidimension-
ality and cumulativity.
25
One of its main statistics is the H-coefficient, which denotes the
homogeneity of the items. According to Mokken, H-coefficients below 0.3 indicate that
the items do not form a cumulative scale, H-coefficients between 0.3 and 0.4 indicate a
weak scale, between 0.4 and 0.5 a moderately strong scale and values above 0.5 indicate
a strong scale.
The results of the scaling analyses for attitudes towards the European Union and attitudes
to immigrants in the two waves of the panel are summarized in Table 2. The four items
measuring attitudes towards immigrants, our control measure, form a strong Mokken scale
in both waves of the panel, and the magnitude of the scale coefficient Hdoes not, as
expected, change notably between the two waves of interviews.
The eight items that measure attitudes towards Europe were not structured by one latent
trait. Instead, two latent dimensions were found to underlie the opinions on these eight
items. The five items of the first scale in Table 2 refer to various aspects of European
unification, whether it is generally a good thing, what the benefits are and to the position
of the smaller and larger states in the European Union. Essentially, these items tap attitudes
towards European integration from a national perspective. From this national perspective,
the contents of these items are of a rather pragmatic or self-interested nature: what does
my country get out of the European Union?
The items that constitute the second scale capture attitudes towards Europe from a
supranational perspective. The first item – are you prepared to make personal sacrifices
to help other European peoples? – captures the extent to which citizens are willing to
24 See Appendix attached to the website version of this article.
25 Rob J. Mokken, A Theory and Procedure of Scale Analysis: With Applications in Political Research (The
Hague: Mouton, 1971). See also Kees Niemoller and W. Van Schuur, ‘Stochastic Modes for Unidimensional
Scaling’, in David Mackay, Norman Schofield and Paul Whiteley, eds, Data Analysis and the Social Sciences
(London: Pinter, 1983); William Jacoby, Data Theory and Dimensional Analysis, Sage University Paper Series
on Quantitative Applications in the Social Sciences, Series No. 07–078 (Newbury Park, Calif.: Sage, 1991).

Influence of EU Political Events on Attitudes to the EU 629
take personal responsibility for Europe. The other two items capture the extent to which
citizens are willing to allocate responsibility to European institutions for domains of policy
making. The second scale combined two aspects of the legitimacy of the European Union
from the perspective of European citizens distinguished by Thomassen and Schmitt, its
domain and its scope.
26
In view of these considerations we assigned labels to the scales. The first scale was called
national–pragmatic attitudes towards the European Union, the second one was called
supranational–idealistic attitudes towards the European Union.
Table 2 shows very modest increases in the magnitudes of the scale coefficients between
the two waves of the panel. As could be expected, the increase in the H-coefficient is the
smallest for attitudes to immigrants, the control variable. But the increases in strength of
the two scales measuring attitudes towards European integration, are also marginal. This
implies that the structure underneath these three attitude scales is highly stable, and that
the media coverage surrounding the summit hardly increased the extent to which these
attitudes are structured.
Attitude Change
A third and final research question investigates whether attitudes towards Europe changed
as a result of the summit, and, if so, whether these changes are dependent upon citizens’
political involvement as measured by political attentiveness and political knowledge. To
assess whether attitudes changed significantly, three t-tests were conducted, one for each
of the three scales. The first scale concerns attitudes towards immigrants, our control
variable in the quasi-experiment. Since this was our control variable, we did not expect
any changes over time in these attitudes. The corresponding t-value is 0.212 (p⫽0.832),
which indeed shows that there was no significant change in attitudes towards immigrants.
The second scale concerns what we have called national–pragmatic attitudes towards the
European Union in Table 2. Here too the corresponding t-value 0.134 (p⫽0.893) shows
that there was no significant change in these attitudes over the period of the summit. Thus,
the intervention of the summit did not lead to people believing that their country or they
themselves were better off in the European Union.
Things are different, however, for the supranational–idealistic attitudes towards the
European Union. The paired samples t-test for this scale revealed that citizens became
significantly more willing to hand over to the European Union responsibility for key issues
such as unemployment and crime (t⫽3.934, p⫽0.000). Now that we have seen that the
summit influenced supranational–idealistic attitudes, we are interested in whether these
changes are dependent upon citizens’ political involvement as measured by political
attentiveness and political knowledge. We therefore conducted a regression analysis in
which change over time was the dependent variable. To do so, we created a new variable
to measure the change in supranational attitudes over time by subtracting the attitude at
the first wave from the attitude at the second wave.
When estimating the effects of political attentiveness and political knowledge on change
over time, we also controlled for the original attitude. Citizens with higher levels of
knowledge and attentiveness had more positive attitudes towards the European Union
before the summit than less politically involved citizens. To the extent that attitude change
depends on the original attitudes, the relationship between levels of involvement and
26 Thomassen and Schmitt, ‘In Conclusion’.

630 SEMETKO,VAN DER BRUG AND VALKENBURG
TABLE
2Main Results of Scaling Analyses*
First wave Second wave
% agree H-coefficient % agree H-coefficient
National pragmatic attitudes towards the EU
Advantages EU membership are larger
than disadvantages (agree) 20 0.51 22 0.51
European unification is moving too rapidly
(disagree) 24 0.47 24 0.47
EU is a threat to small countries (disagree) 29 0.52 29 0.51
European unification is a good thing
(agree) 33 0.53 32 0.54
Only the big countries have a say in
European affairs (disagree) 37 0.44 34 0.42
Scale 0.49 0.49
N⫽445 N⫽573
Supranational attitudes towards the EU
Willing to make personal sacrifices to help
other EU countries 11 0.43 11 0.40
Unemployment should be dealt with at
European level 21 0.58 28 0.59
Crime should be dealt with at European
level 54 0.65 59 0.68
Scale 0.56 0.57
N⫽636 N⫽650
Attitudes towards immigrants
Too many people from Surinam in the
Netherlands 38 0.54 37 0.54
We should welcome people from abroad 40 0.52 40 0.55
Islamic cultures enrich the Dutch culture 55 0.56 53 0.56
Too many people from Turkey and
Morocco in the Netherlands 60 0.69 59 0.70
Scale 0.58 0.59
N⫽669 N⫽684
*Data were deleted listwise

Influence of EU Political Events on Attitudes to the EU 631
attitude change may be obfuscated. There are two ways in which attitude change may
depend on previously held positions. The first one is the consequence of the use of bounded
scales. For someone with the most positive position towards the European Union, it is
impossible to become even more positive. So, even news that reinforces his positive
attitude will not lead to an observable attitude change, since the position is already at (or
near) the maximum. Secondly, it is feasible that, due to selective perceptions of the news,
previously held beliefs are reinforced, so that negative attitudes become more negative,
and positive ones become more positive. To control for these possibilities, regression
analysis was conducted with attitude change as the dependent variable, and with three
predictors: the original attitudes, political knowledge and political attentiveness. The result
is presented in Table 3.
There is a particularly strong negative effect of originally held attitudes on changes in
supranational–idealistic attitudes. Further inspection of the direction of change demon-
strated that those who were originally most critical towards the supranational ideals of the
European Union changed their attitudes the most, thus becoming less critical. Once this
effect is accounted for, we find no effect of political knowledge but a significant effect of
political attentiveness on attitude change.
27
The most attentive citizens changed their
attitudes significantly more than those who were less attentive. A comparison of the means
of the different groups reveals that all levels of political attentiveness change in the same
TABLE
3Results of Regression to Predict Change in Supranational Attitudes
Towards the EU
B(s.e.) P(two-sided)
Intercept 0.100 (0.026) 0.000
Political knowledge (7-point scale) 0.010 (0.007) 0.127
Political attentiveness (10-point scale) 0.005 (0.002) 0.049
Original attitude ⫺0.468 (0.036) 0.000
R
2
adjusted 0.229
N568
27 In addition to the t-tests and the regression analysis, we also conducted three MANOVA analyses, one for
each of the three scales. Attitudes of each respondent measured at both time points are treated in these MANOVAs
as within-subject factors. Political attentiveness and political knowledge are between-subject factors. In order to
check for possible nonlinear relationships, the various measures of political attentiveness and political knowledge
were both coded into three categories. These MANOVA analyses confirmed the findings of our t-tests and
regression analysis. No significant over-time change was found in attitudes towards immigrants and national
pragmatic attitudes towards the European Union. In addition, no interactions were found between change over
time and levels of attentiveness and knowledge.
The MANOVA shows a significant over-time effect, F(1,559) ⫽5.58, p⬍0.05, 2⫽0.01 for supranational
attitudes towards the European Union, a finding which is consistent with the t-tests reported here. The only
difference between the MANOVA and the regression analysis is that in the MANOVA the interaction effects
between the rate of change and levels of attentiveness (F(2,559) ⫽0.100, p⫽0.905) or knowledge
(F(2,559) ⫽1.062, p⫽0.347) are not significant. In the regression analysis we found a significant effect of
attentiveness on attitude change. The difference in findings may be caused by the fact that regression is a more
powerful test. In the regression we estimated a linear effect, in the MANOVAs we estimate means for various
subgroups.

632 SEMETKO,VAN DER BRUG AND VALKENBURG
direction, becoming more positive towards the European Union.
28
This general trend is also
indicated by the positive and significant value for the intercept.
In summary, the effect of the summit on supranational attitudes towards the European
Union was a general one that influenced inattentive as well as ill-informed citizens in the
same way as attentive and well-informed citizens. In contrast to previous research,
however, we find that attitude change was greatest among the most attentive citizens.
CONCLUSIONS
We identified two types of attitudes towards the European Union. National–pragmatic
attitudes were measured by a number of questions to determine whether respondents
believed they or their country would be better off in the European Union. Supranational–
idealistic attitudes were measured by a number of questions to determine whether
respondents were supportive of the supranational ideals of the European Union. While the
former did not change significantly from wave one to wave two, the latter did. Thus,
because of the summit, citizens became stronger supporters of EU supranational ideals.
How durable or persistent are these effects? Our results show that public attitudes
towards the European Union changed and in a positive direction. People became more
favourable to the European Union taking greater responsibility for social problems. But
we cannot be certain about the durability or persistence of these effects. Previous
experimental research involving media effects also provides no clear answer on the
durability question. Iyengar, Peters and Kinder, for example, found support for television
news having an agenda-setting effect and concluded that these effects can persist ‘up to
twenty-four hours’ because the post-test questionnaire was administered one day after the
final news broadcast was aired in their experimental laboratory.
29
Studies of attitude change that use survey data usually do not address the matter of the
durability of the change. Van der Brug, however, shows that the Chernobyl accident of
1986 generated a large aggregate change in attitudes towards nuclear energy among Dutch
citizens, which lasted at least until 1998.
30
An EU summit is not comparable to a nuclear
accident and it did not have a similar lasting impact on public opinion. The effect we found,
however, did last longer than the time span of existing experimental studies. Our post-test
began three days after the summit had ended, and it was completed within three days. From
this, we can conclude that the effect persisted for up to at least six days.
The attitudinal effect remained even after the summit had ended, though we do not know
how other post-summit political news may have modified this effect. News production
cycles are largely on a daily basis. What have been called ‘issue attention cycles’ in the
news may be somewhat longer, with news focusing on one subject or story for an extended
period of time.
31
Given that the balance of EU news versus domestic political news as well
as the tone of such political news (consensus-oriented or conflict-oriented) changes
regularly in the national European news outlets, our research would lead us to expect some
effect on European public opinion about the European Union assuming the EU news was
highly visible. But we cannot say anything further about the direction of the expected
28 We distinguish three levels of attentiveness. The mean for the lowest level group changed from 0.23 to 0.26,
for the middle group the mean changed from 0.27 to 0.32, and for the highest level group the mean changed from
0.35 to 0.41.
29 Iyengar, Peters and Kinder, ‘Experimental Demonstrations’, p. 855.
30 Van der Brug, ‘Perceptions, Opinions, and Party Preferences in the Face of a Real World Event’, p. 60.
31 Anthony Downs, An Economic Theory of Democracy (Boston, Mass.: Addison-Wesley, (1990[1957]).

Influence of EU Political Events on Attitudes to the EU 633
effect (whether it would depend upon the context of the event and its coverage in the news),
or the durability of the effect.
The media coverage of the summit had very different effects on the two European
attitude scales distinguished here. Supranational–idealistic attitudes became significantly
more favourable as a result of the summit, while national–pragmatic attitudes did not
change significantly. The Treaty of Amsterdam agreed upon by the leaders during the
summit further involved the European Union in the process of solving social problems in
EU member countries. Provisions concerning employment guarantees and improved
security including greater co-operation between police forces in EU countries were a part
of the treaty. This might explain why we found that the media coverage of the summit had
an influence on what we have called supranational–idealistic attitudes, which specifically
refer to citizens’ willingness to have Europe take more responsibility for solving social
problems. The attitude change found in our study was at least a sign of increased confidence
in the European Union.
In much of the research on news effects, the dependent variable is support for political
parties or candidates, and therefore partisanship and ideology is an important part of the
explanation. Zaller distinguished between consensus-based and polarizing issues and their
potential consequences for public opinion, with reference to US cases and the role of
political ideology and partisanship in shaping effects on public opinion.
32
Our study goes
beyond these US examples and refers to a European event or issue, in the form of the Treaty
of Amsterdam, which is above the level of national political ideologies or partisanship. The
Amsterdam Summit was a consensus-based event in which there was agreement on issues
among European elites who represented the broad range of political ideology. Our findings
suggest that further research is needed on how different types of attitudes may be influenced
by different types of political events. Consensus-based events may produce very different
conditions for the development of and change in attitudes than those produced by
conflict-based events.
This study suggests that the regular summit of European leaders is a potentially
important channel for influencing public opinion about European integration. But the
context of each summit (and whether they reach consensus or conclude with conflict) may
mean different consequences for public opinion. The June 1997 Amsterdam Summit
generated a considerable amount of news coverage in the national media not only about
the substantive issues under discussion in the signing of the Treaty of Amsterdam.
33
Among the most memorable images are those of Tony Blair, Wim Kok and the other EU
prime ministers riding bicycles on the bridges over the canals.
Our panel study and quasi-experimental design – the combination of a panel and control
variables – provides a strong structure for analysing the effects of a major political event
on attitude change, and it is unique in the literature. Including attitudes to immigrants as
a control variable means that the internal validity of the design is guaranteed, which is a
bonus compared to most studies that have used survey data. The fact that we estimate the
effect of a real event in natural circumstances, rather than manipulated news effects in the
context of a laboratory experiment, greatly increases the external validity of our results.
Our research raises questions about the commonly assumed variation with knowledge
32 Zaller, The Nature and Origin of Mass Opinion.
33 Holli A. Semetko and Patti M. Valkenburg, ‘Framing European Politics: A Content Analysis of Press and
Television News’, Journal of Communication,50(2000), 93–109. See also Holli A. Semetko, Claes de Vreese
and Jochen Peter, ‘Europeanised Politics – Europeanised Media?’ West European Politics,23(2000), 121–41.

634 SEMETKO,VAN DER BRUG AND VALKENBURG
and attentiveness. Studies of the US public have shown that effects varied with levels of
political knowledge or attentiveness, with those who are less attentive or knowledgeable
found to be open to greater influence. Our findings show that those who were more
politically attentive were more likely to have been influenced by the information they
received. Whereas Krosnick and Brannon found that the ‘highly knowledgeable about
politics but relatively inattentive’ were the most influenced, we found that when controlling
for the original attitude, political attentiveness but not political knowledge emerged as a
significant influence on attitude change.
34
A possible explanation for the differences in findings stem from the fact that the summit
was a consensus-based event generating considerable news coverage, whereas the events
in American studies generally pertained to election campaigning or to scandals. As a case
in point, Krosnick and Kinder’s study focuses on the effect of the Iran–Contra scandal,
and they find that the least attentive are affected the most.
35
A possible explanation for the
differences in findings is that less attentive citizens are affected most by such scandalous
news or events, whereas consensus-based issues or events are more likely to affect citizens
who are more attentive or politically involved.
36
Our findings suggest that more research
is needed on the kind of events and issues (consensus or polarizing) and the kind of news
that affects the politically involved and the uninvolved.
Our finding that supranational–idealistic attitudes towards the European Union changed
significantly therefore may be in part a consequence of this national favourable or
consensus-based context in which the 1997 summit decisions were being made. Recent
summits were more conflict-ridden, and it may well be that the harmonious Amsterdam
Summit will later turn out to have been quite exceptional. Further research should be done
to determine how the differing national conditions and information provisions surrounding
EU summits enhance or diminish public opinion about the legitimacy of the European
Union and the integration process.
34 Krosnick and Brannon, ‘The Impact of the Gulf War on the Ingredients of Presidential Evaluations’, p. 974.
35 See Krosnick and Kinder, ‘Altering the Foundations of Support for the President Through Priming’.
36 See also Zaller, The Nature and Origin of Mass Opinion.

Influence of EU Political Events on Attitudes to the EU 634a
APPENDIX
This appendix gives an overview of the results of Mokken scaling analyses for the indicators of political
attentiveness and political knowledge (Table A.1). According to Mokken the minimum value for the
H-coefficient for items to form a unidimensional cumulative scale is 0.3.37 Values higher than 0.5 indicate
a strong scale. The political attentiveness scale presented in Table A.1 combines all items from the three
aspects of attentiveness: interested in news, reading about news and discussing news. These analyses
demonstrate that there is no need to analyse these aspects of attentiveness separately, because they measure
the same trait.
TABLE
A.1 Scales of Political Knowledge and Political Attentiveness*
First wave Second wave
Mean H-coefficient Mean H-coefficient
Political knowledge
Knows function of Santer 0.16 0.68 0.22 0.67
Knows fraction leader PvdA 0.38 0.51 0.44 0.50
Knows value of euro 0.44 0.39 0.53 0.39
Knows minister of social affairs 0.49 0.53 0.52 0.55
Knows minister of finance 0.58 0.59 0.67 0.61
Knows subject of Schengen Agreement 0.65 0.61 0.69 0.56
Scale 0.53 0.53
Political attentiveness
Reads about European unification 0.75 0.71 0.75 0.72
Reads about foreign news 0.92 0.64 0.87 0.67
Reads about domestic news 0.94 0.65 0.91 0.67
Talks about foreign news 1.01 0.65 0.97 0.69
Talks about European unification 1.03 0.61 1.01 0.69
Talks about domestic news 1.05 0.64 1.02 0.67
Interested in news about European unification 1.16 0.62 1.12 0.69
Interested in foreign news 1.21 0.67 1.12 0.67
Interested in domestic news 1.23 0.70 1.19 0.75
Scale 0.65 0.69
*All items measuring political knowledge are dichotomized (wrong/right answer). All political attentiveness
items are trichotomized.
37 Mokken, A Theory and Procedure of Scale Analysis.