News Coverage of the Enlargement of the European Union and Public Opinion:
A Case Study of Agenda-Setting Effects in the United Kingdom
Oya Dursun
University of Texas at Austin
odursun@mail.la.utexas.edu
Abstract
While the British government has largely been in favor of the recent round of European
Union (EU) enlargement, the British public opinion was mostly against it. To account for
the gap between public opinion and official policy on enlargement in the United
Kingdom, this project scrutinizes the way the British media treats the issue of
enlargement following the Laeken Council of December 2001 up until the day of
enlargement – 1st of May 2004.
The research is contextualized in agenda-setting theory. This project tests both the first
and second-level agenda-setting effects, and the consequences of agenda-setting for
public understanding and evaluations of the enlargement of the EU. To what extent the
frequency of the coverage of enlargement influences how important people consider it to
be? Do the frames the media apply translate into patterns the public uses to interpret
those affairs? What are the effects of frames on public opinion toward enlargement? To
provide answers to these questions, this paper combines quantitative content analysis and
survey data through a comparison of the trends in the Eurobarometer surveys with
enlargement-related news coverage in The Times, The Guardian and The Daily Mail.
Keywords: agenda-setting, EU enlargement, British media, public opinion
Paper prepared for presentation at the European Union Studies Association’s Ninth
Biennial International Conference, 31 March-2 April 2005, Austin, TX, U.S.A.
Introduction
What is the relationship between policy, public opinion and the media? This
question has been investigated extensively in the U.S. setting, yet still remains a
relatively overlooked one in the European context. The knowledge about how the media
covers European political affairs and how citizens in different Member States react to
news about European affairs and processes of integration is very limited. Also there is a
lack of integrating theoretical frameworks in studies on the visibility of European issues
in the media (Semetko et al. 2000). This paper attempts to address these deficiencies in
the political communication literature through an examination of the agenda-setting
functions of print media in the United Kingdom (UK) on the issue of the enlargement of
the European Union (EU) in May 2004.
Besides being a political process, accession of ten new Member States is also an
important socio-cultural enterprise with widespread repercussions. The enlargement issue
stopped to be purely technical, and started to affect the daily lives of individuals.
Consequently, Europe has become more “visible” and as such more liable to arouse
public disquiet. While the British government has been in favor of the latest eastern
enlargement of the EU, the majority of both the British public were against enlargement.
As the legitimacy of elite actions depends upon the level of public support for European
integration processes, it is important to study the effects of the coverage of these issues
on public opinion (Rohrschneider 2000, Meyer 1999).
The media are the principal means by which majorities of people receive
information about policy issues in general, and the EU in particular (Eurobarometer 61,
Page and Shapiro 1992, De Vreese 2003). According to Standard Eurobarometer (EB)
1
61, 50 percent of the British people prefer to get information about enlargement from
daily newspapers. “Which aspects of an issue are covered in the news – and the relative
emphasis on these various aspects of an issue – makes a considerable difference in how
people view that issue” (McCombs 2000, p. 7). Hence, this paper looks in the way media
treats the issue of enlargement, and scrutinizes media’s effects on public opinion in order
to account for the gap between British public opinion and official policy.
Although EU issues generally receive low attention in the news in most European
countries; EU’s latest round of enlargement has been one of the most highly visible issues
concerning Europe. As such, enlargement provides an opportunity for an interesting case
study of the effects of news coverage on public opinion of a historic European event. The
research is contextualized in agenda-setting theory. It tests both the first- and secondlevel agenda-setting effects of media in the British context through an exclusive focus on
the news coverage of enlargement in The Times, The Guardian and The Daily Mail from
January 1, 2002 to May 1, 2004.
Moreover, enlargement is an “unobtrusive issue”: personal experience on
enlargement during the time frame of this study was greatly limited, if not non-existent.
For such issues, need for orientation is largely satisfied through the use of the mass media
and that the degree of media influence increases with greater exposure (McCombs, 2005,
p. 64).
Greater insight into the effects of media on the issue of enlargement which will be
gained with the results of this paper has significant implications for possible future
referenda on the subsequent rounds of enlargement. This study also provides a test for the
existence of democratic deficit, a widely debated concept in the EU studies literature.
2
Theoretical Background
Research in this paper is contextualized in agenda-setting theory. Agenda-setting
theory posits that elements that are prominent on the press agenda, both “objects” and
their “attributes”, frequently become prominent on the public agenda. This idea of an
agenda-setting role of the press has its origins in Walter Lippmann’s Public Opinion,
which begins with a chapter titled “The World Outside and the Pictures in Our Heads.”
He argues that the press is a major contributor to those pictures in our heads (Lippmann
1922).
The term “agenda setting” is first coined by Maxwell McCombs and Donald
Shaw to describe the correspondence between the ranking of major issues on the press
and public agendas in their pioneering work (McCombs and Shaw 1972). Since the
1970s, the agenda-setting influence of the press has been widely replicated.
Traditional (or first-level) agenda-setting theory emphasizes how mass media,
elites, and the public interact and influence one another to affect issue salience
(McCombs and Shaw 1972, Rogers, Dearing, and Bregman 1993, McCombs and
Reynolds 2002). Traditional agenda-setting effects point to the visibility and perceived
importance of a problem or an issue due to its visibility or salience in the media. As
Bernard Cohen noted in his seminal work The Press and Foreign Policy, the press “may
not be successful much of the time in telling people what to think, but it is stunningly
successful in telling its readers what to think about” (Cohen 1963, p. 13). Cues repeated
day after day effectively communicate the importance of each topic (McCombs 2000, p.
1). The more coverage an issue receives, the more concern individuals have with the
3
issue. In other words, individuals learn how concerned they should be through the
amount of coverage the issue receives (Wanta 1997, McCombs 2000).
For each “object” there also is an agenda of “attributes” because when the media
and the public think and talk about an object, some attributes are emphasized, others are
given less attention, and many receive no attention at all (McCombs 2000, p. 6). Since
the agenda-setting functions of mass media are vital at both the “object” (the things on
which the attention of the media and the public are focused) and “attribute”
(characteristics and traits that describe the object) levels, agenda-setting effects are not
only limited to affecting issue salience (i.e. first-level agenda-setting). The second-level
of agenda setting adds to the previous agenda-setting research by examining a much more
complex digestion of content: it examines how media organizations select and present
certain characteristics and properties of an object and how that selection and presentation
influence the public’s perception of an object and its attributes.1 The aspects of issues
selected for attention by the media influence the public’s perception of these issues
(McCombs 2000, p. 7). This effect is also known as framing.
Framing is most commonly defined as “select[ing] some aspects of a perceived
reality and mak[ing] them more salient in a communicating text, in such a way as to
promote a particular problem definition, causal interpretation, moral evaluation and/or
treatment recommendation for the item described” (Entman 1993, p. 52, original
emphasis). It constitutes a second-level agenda-setting effect (McCombs 1997). Framing
effects refer to changes in evaluations and interpretations as the result of aspects made
salient through selection, organization and emphasis. “Thinking of frames as attributes of
1
Ghanem 1997, p. 3-14; McCombs and Reynolds, 2002. Quoted in Hsiang and McCombs 2004, p.22.
4
an object provides the theoretical link between agenda-setting and framing research”
(Hsiang and McCombs 2004, p. 24).
The United Kingdom is renowned for its skepticism and reluctance with respect to
the EU issues. Various studies on British political communication literature point out to
the pessimism of the public on EU issues and Euro-skeptic nature of the British press.2
Carey and Burton (2004) point out that research in the British context has found that the
media asserts influence on attitudes and behavior and that these effects are significant,
although relatively small.3 Looking at the impact of the press on attitudes towards
monetary union, for instance, Curtice (1999) finds that there was no significant evidence
that newspapers influenced their readers on this issue either in the period between
elections or in the campaign leading up to the 1997 election.4
A number of recent studies have identified the importance of certain frames in the
news by focusing on their consequences for the public’s interpretation of events and
issues.5 Pippa Norris (2000), for instance, dealing with the influence of the media on
attitudes towards European integration finds that, when an attentive public receives
extensive media coverage of an issue that displays a consistent directional bias, the media
are likely to have an impact sufficient enough to change public attitudes at the aggregate
level. She finds a strong relationship between negative press coverage of monetary union
and decreasing levels of aggregate support both for the euro and the EU in general. 6 But
2
Dalton and Duval; Carey and Burton, 2004.
Norris et al., 1999; Newton and Brynin, 2001, quoted in Carey and Burton 2004, p. 625.
4
Curtice 1999, p. 26-27.
5
Capella and Jamieson, 1997; Graber, 1988, 1993; Iyengar, 1987, 1991; Iyengar and Kinder, 1987;
Neuman et al., 1992; Norris, 1995; Patterson, 1993.
6
Norris 2000, p. 206.
3
5
she warns that the direction of causality in the relationship between media coverage and
aggregate public attitudes on monetary union cannot be established.
Anderson and Weymouth (1999) conduct a detailed analysis of the discourse of
the British press on issues relating to European integration during the general election
campaign of 1997 and during the British presidency of the EU in 1998. They find that
articles within the Euro-skeptic press during this time (Sun, Mail, Express, Times and
Telegraph) explicitly aimed to influence attitudes on European issues, and the purpose of
this discourse was to influence and to persuade their readers of the non-acceptability of
monetary union.7
Research Questions and Hypotheses
This study aims at testing both the first- and second-level agenda-setting effects of
media and the consequences of agenda-setting in terms of opinions and attitudes towards
enlargement in the British context, focusing exclusively on the coverage of The Times,
The Guardian, and The Daily Mail.
To assess media effects on public opinion, this paper draws on two different
sources of data. First, it utilizes several EB – both Standard and Flash – survey data to
measure people’s attitudes towards the latest round of EU enlargement. Secondly, it
conducts a quantitative content analysis on the above-mentioned major print media
outlets in the UK.
a. First-Level Agenda-Setting Effects
To test for the first-level agenda-setting functions of the British media, this study
asks to what extent the frequency of the coverage of enlargement influences how
important people consider it to be. Following the theory, one would expect that the higher
7
Anderson and Weymouth 1999, quoted in Carey and Burton (2004), p. 625.
6
the frequency of enlargement coverage is on the media agenda, the higher the perceived
importance of this issue will be on the public agenda. Since name recognition is a
necessary condition for salience,8 this study focuses on the name recognition of
enlargement in the British and public, which is available in the Eurobarometer survey
data.
Hypothesis 1: Media salience of enlargement will be positively correlated with its
public salience – i.e. the proportion of the public who recognize enlargement.
The public agenda is measured through survey research. Both the Standard
Eurobarometer surveys and the Flash Eurobarometer surveys on enlargement ask the
following question to determine the salience of enlargement for people: “Before this
interview, had you already heard of the enlargement of the European Union?” The
available answers for this question are: yes (1) and no (0).
As a test of first-level agenda setting effects, this research plots the average
amount of enlargement-related coverage per month against the percentage of
Eurobarometer respondents who indicated their awareness of enlargement. It relied on
simple correlation to test the statistical significance of the relationship between
enlargement coverage and the proportion of respondents who expressed that they are
aware of the latest round of enlargement.
b. Second-Level Agenda-Setting Effects
Second-level agenda-setting theory holds that media not only shape the salience
of “objects” in public opinion, but “attributes” of these objects as well. The prominence
of an issue’s attributes in the “pictures in our heads” is influenced by the pattern of
attributes in the press coverage for that issue.
8
McCombs and Kiousis, 2004.
7
Regarding the second-level agenda-setting effects, this paper investigates if the
media coverage of enlargement tells people “how to think about it”, besides telling them
“what to think about”. In this study, “object” is the enlargement of the EU and the
“attributes” are the various traits that define the images of enlargement. To investigate the
second-level agenda-setting effects, a content analysis of substantive “attributes” of
enlargement is conducted. It is assumed here that the public support for enlargement is
dependent on the perceived consequences of enlargement by people. As such, this study
employs “political,” “economic,” and “socio-cultural” consequences of enlargement as
attributes. Media frames (or attributes) analyzed in this study are directly extracted from
the Flash Eurobarometer survey questionnaires.
The main research questions at this level are: What are the effects of these news
frames on attitudes toward enlargement? Do the frames the media apply to European
affairs translate into patterns the public uses to interpret those affairs?
Hypothesis 2: The more salient an image attribute is in the press coverage of
enlargement, the more likely the people are to describe enlargement in terms of that
attribute (i.e. attributes of enlargement emphasized by the media will correlate with
enlargement attributes salient to public).
By manipulating the relative saliencies of various consequences of enlargement,
media has a great potential to impact public opinion on the issue. Linking the framing
concept to public opinion and enlargement requires an identification of the predominant
frames in the coverage of enlargement.
The results of content analysis are then compared with the survey data on public’s
attitudes on the consequences of enlargement. Predominant frames in the coverage of
8
enlargement over time are examined in order to see whether the frames the media apply
translate into frames the public uses to evaluate enlargement. Is news about enlargement
excessively concerned with political consequences, economic consequences, or sociocultural consequences of the enlargement? Then the same question is asked for the
public’s framing of enlargement.
For the analysis of second-level agenda-setting effects, rank order correlation is
conducted. After creating dichotomous variables for “peace,” “moralduty,” “stronger,”
“expensive for the country,” etc., I conduced a rank ordering between the public and the
press agendas.
Method
Time Frame
Time frame for this study is from January 2002 to May 2004. January 2002 is
chosen to be the starting date of the time frame since on 14th and 15th December 2001, the
Laeken European Council declared that the accession process was “irreversible” and
stressed the EU’s determination to bring the negotiations with those countries ready to
join to a close by the end of 2002. In October 2002, the Commission recommended
concluding the accession negotiations with Cyprus, the Czech Republic, Estonia,
Hungary, Latvia, Lithuania, Malta, Poland, Slovakia and Slovenia by the end of 2002.
The Copenhagen European Council, held on December 2002, concluded negotiations
with the ten candidate countries listed above, and set 1st of May 2004 as the accession
date. On April 9, 2003, the European Parliament gave its assent to the accession of the
ten acceding states. The Treaty of Accession was signed on April 16, 2003. May 1, 2004
9
marked a historic moment as Cyprus, Czech Republic, Estonia, Hungary, Latvia,
Lithuania, Malta, Poland, Slovakia and Slovenia joined the EU.
Besides its chronological importance, another rationale for selecting this time
frame is the availability of more frequently conducted public opinion surveys by
Eurobarometer (Standard, Special, and Flash Eurobarometer Surveys on Enlargement).9
Standard Eurobarometer surveys are usually conducted twice a year. During the time
frame of this study, ten public opinion surveys were conducted on enlargement. This
enables a closer comparison of news coverage and survey results and makes the results of
this study more valid. Having the polling data from surveys conducted frequently during
the period of approximately two-and-a-half years is appropriate to account for changes in
people’s attitudes towards enlargement in time.10
Research Design
This study combines content analysis and survey data. It compares the trends in
Eurobarometer polls to enlargement-related news coverage in the UK, and examines the
computerized media contents of leading British press outlets using the Lexis-Nexis
database. According to the Standard Eurobarometer surveys analyzed in this study, daily
newspapers comes second to the TV in being the most popular sources of information on
the EU, chosen by 50 percent of the UK sample (Eurobarometer 61).
The British newspapers that are content analyzed in this study include elite/
quality newspapers as well as a major tabloid, with the highest circulation rates in
9
Please refer to Annex I for the dates of fieldwork of EB surveys used in this analysis.
Under ideal conditions, this study would also examine the relationship between the media coverage and
public opinion after the enlargement day. But the latest of a series of Eurobarometer surveys assessing
public opinion on enlargement has been conducted on March 2004. Hence this limits the time frame of this
study.
10
10
Britain: The Times, The Guardian and The Daily Mail. In the original research design of
this study, The Sun was planned to be included in our content analysis; nevertheless,
since only the Irish version of this paper is available through the Lexis-Nexis database; it
is replaced by another popular British tabloid – the Daily Mail. The Guardian is one of
Britain’s oldest newspapers with a well-respected national and international reputation. It
has maintained a radical, left of center editorial stance ever since its foundation. The
Times – Britain’s newspaper of record – represents the conservative voice in the UK.
Including The Daily Mail in analysis enables us to test for the homogeneity of press
agendas between tabloids and broadsheets.
Data
This analysis draws on two sources of data: polls taken by the Eurobarometer
between 2002 and 2004; and a content analysis of press coverage of enlargement of the
EU in Britain. Eurobarometer survey polls all used a probability sampling procedure and
were based on approximately 1,000 respondents.
The unit of analysis is the entire text of articles. This study content analyzes the
population of all newspaper articles that concentrate on enlargement in the media under
focus in Britain. A census is deemed more appropriate for this study since it covers a
shorter period of time.
In the selection of our sample of media content, I aimed both for elite newspapers
and tabloid press from both sides of the left–right divide with a high circulation and
considerable influence on other types of media, as well as on the political system. These
papers can be considered as influential sources of information for other media, as well as
a good cross-section of national press opinion on socio-economic and political issues.
11
Regarding the first-level effects, a keyword search is conducted to identify the
volume of media attention devoted to enlargement. As has been the case in most agendasetting studies, the salience in the media of the issue is defined in terms of the number of
stories (e.g. McCombs and Kiousis 2004; King, 1997; McCombs and Shaw, 1972;
Rogers and Chang, 1991). Articles that contain keywords – European Union enlargement,
EU enlargement and accession – in the headline or the lead paragraphs are included in
analysis.
In a second step I eliminated all the articles in our selection which were either
duplicates (due to the different editions of newspapers), or were not sufficiently relevant:
for example, if the keyword appeared just as a passing remark in one sentence or in an
irrelevant context. The selection included editorials and opinion pieces as well as news
stories of more than two paragraphs. The final dataset contained 342 newspaper articles.
Coding is conducted manually by two native speakers of British English.11 Coders
are trained before the coding and supervised during the coding stage. The frequencies of
these articles are counted, and then matched with the level of public salience. A team of
coders is trained and supervised. Written coding instructions, a copy of the codebook,
copies of newspaper articles, and coding spreadsheet are provided to coders. The training
sessions familiarized coders with the purpose of the research, the content that will be
coded, and the codebook that will be used. Ethical standards of content analysis research
are also emphasized in the training session. Each article was coded for the following
variables: “peace,” “moralduty,” “stronger,” “difftotake,” “histnatural,” “countless,”
Two undergraduate students – Timothy Sheffield of the University of Manchester and Jack de Glanville
of Umist University – conducted the coding. The author wants to thank them for their assistance.
11
12
“unemployment,” “expensivecountry,” “expandmarkets,” “CAP,” “welfare system,”
“cultricher,” “immigration,” “drugsmug,” “illegalimmig,” “remotecitizens.”
After repeated pretesting of the codebook to determine the reliability of coding
decisions, a sample of articles from each newspaper is coded. Intercoder reliability tests
are conducted on a randomly selected subsample of 10% of the news stories (34 articles).
To ensure intercoder reliability, several pretests were conducted till Scott’s pi reached 86
percent across the variables of content analysis.12 Once the final version of the codebook
is drafted, coders started coding each article directly onto SPSS spreadsheet.
The average monthly coverage of enlargement-related news in the Guardian and
The Times were several times larger than in The Daily Mail. Because of The Mail’s
limited coverage of enlargement, the analysis is corrected to adjust for the low reliability
of Mail scores in months with very few sampled articles. To correct for sampling method,
the weight variable was utilized to adjust for the differing probabilities that cases have of
being selected in a sample.
For measuring public salience, this study employs an indirect measure – name
recognition of enlargement – that is available in Eurobarometer survey data. The
assumption is that people must be able to recognize enlargement in order to consider it
salient. Name recognition is a necessary condition for and can serve as an indirect
measure of salience when other options are unavailable (McCombs and Kiousis 2004).
For exploring the second-level agenda-setting effects, the agenda of attributes to
be analyzed include three main categories:13 1) Political consequences of enlargement
Scott’s pi as a reliability index takes into account the extent of inter-coder agreement that results from
chance. See Holsti, 1969, p. 140-141.
13
Regarding the analysis of major frames, Semetko and Valkenburg (2000) employ “economic
consequences” frame, which they argue reports an event, problem, or issue in terms of the consequences it
12
13
(i.e. effects of enlargement on the EU decision-making process; EU’s standing in the
international political arena; impact on peace and stability in Europe; historically natural;
effect on the weight of Britain within the EU, etc.); 2) economic consequences of
enlargement (i.e. effects of enlargement on un/employment; the cost of funding of new
Member States; investment opportunities in a bigger market; impact on the Common
Agricultural Policy; and its impact on the social welfare system); and 3) socio-cultural
consequences of enlargement (i.e. impacts of enlargement on cultural life; immigration
from new EU member states; the fight against drug smuggling and organized crime;
illegal immigration; EU institutions’ relations to local populations).
To what extent concerns about enlargement’s political consequences in the media
are mirrored in public opinion? To test whether agenda setting or second-level agenda
setting resulted from the newspapers’ framing of enlargement, the content analysis data
were compared with the questions asked by the Eurobarometer polls that relate directly to
the second-level of agenda setting.
Using the Flash EB survey results, the correspondence between the news agenda
of attributes and the public agenda of attributes on enlargement is calculated to test for
attribute agenda-setting effects. The results of content analysis are compared with the
survey data on public’s attitudes on the consequences of enlargement (using Spearman’s
Rho).
Results
will have economically on an individual, group, institution, region, or country. Neuman et al. (1992) and
Graber (1993), and Eichenberg and Dalton (1993) also identify it as a common frame in the news. To the
economic consequences frame, this study adds the political and socio-cultural consequences of enlargement
to investigate the extent to which these frames are mirrored in public support for enlargement.
14
To better account for the first-level agenda-setting effects, this paper at first
focused on a time period under which three surveys were conducted in two months: Flash
EB 132.1 - September 2002; Standard EB 58.1 - October – November 2002; and, Flash
EB 132.2 - November 2002. Hence it analyzed the media coverage between 05/07/02
and 11/26/02. As can be seen from the figure below, the first-level effects by the media
coverage have become prevalent among public with only a time-lag of one month from
September to October 2002.
TSPLOT (from May 2002 to November 2002)
.7
.6
.5
.4
.3
.2
.1
COVPERCE
0.0
AWARENES
MAY 02
JUN 02
JUL 02
AUG 02
SEP 02
OCT 02
DA TE
TSPLOT (from January 2002 to May 2004)
15
NOV 02
.7
.6
.5
.4
.3
.2
.1
COVERPER
0.0
AWA RENES
JAN 02 MAY 02 SEP 02
MAR 02
JUL 02
JAN 03 MAY 03
NOV 02 MAR 03
SEP 03
JUL 03
JAN 04 MAY 04
NOV 03 MAR 04
DATE
Correl ation be twe en Enla rge me nt Cove rper and Aw are ness
COVE RPE R
Pearson Correlation
Sig. (1-tailed)
N
AW ARENE S Pearson Correlation
Sig. (1-tailed)
N
COVE RPE R
1
.
29
.261
.086
29
AW ARENE S
.261
.086
29
1
.
29
*Correlation is significant at the 0.10 level (one-tailed)
We can observe that the Pearson correlation between public awareness and press
coverage of enlargement is .26. So, we would reject the null hypothesis that there is no
statistically significant relationship between the amount of press coverage and the public
awareness of enlargement of the EU. Hence the results of the first-level agenda-setting
analysis solidly concurs the existence of a transfer of issue salience from the media to the
public. As the issue of enlargement received more coverage by the newspapers, the
British public became more familiar with it.
16
Of the 342 articles, the political consequences and economic consequences frames
are adopted equally – approximately 57 percent each (57.3 percent, and 56.5 percent,
respectively); and 53.1 percent adopted the socio-cultural consequences frame. Hence we
can conclude that each general consequence of enlargement was covered almost equally.
In the cross-tabulation procedure, the chi-square tests conducted on each frame and each
newspaper revealed statistically significant results for the social consequences frame.14
While The Daily Mail has a positive relationship; the other two newspapers have negative
relationship with socio-cultural consequences frame. This emphasizes how a tabloid deals
more closely with the social consequences of an important European event.
The EB survey data suggest that although the percentage of people who have
already heard of enlargement in the UK is well below the EU15 average, issue salience
increases among people with an increase in the enlargement-related press coverage as the
actual date of enlargement approaches. Hence, the British public learned how concerned
they should be through the amount of coverage enlargement received.
As long as the content analysis of the three papers goes, we can conclude that the
coverage of the Times and the Guardian was almost identical. The coverage of the Daily
Mail was, on the other hand, quite different from the two broadsheets. First of all, the
number of articles allotted to enlargement was significantly lesser than the number of
articles assigned in the Times and the Guardian. Secondly, the Daily Mail’s coverage was
more negative on the effects of enlargement, and was more critical of the government’s
policies on enlargement. This was especially apparent on the issue of immigration. The
tabloid has criticized the way the government has treated the issue of free movement of
14
Since content analysis examines census instead of a random sample, the differences that are found are
real, which makes it unnecessary to run statistics that report significant differences or project to the
population. Even so, the statistical tests are conducted.
17
labor, and warned with a heavy language that this issue will cause lots of problems
ranging from its effects on the welfare system, the education system, and unemployment
in Britain.
SOCCODIC * TIMES
Crosstab
Count
TIMES
.000
SOCCODIC
.000
1.000
1.000
63
51
114
98
131
229
Total
Total
161
182
343
Chi-Square Tests
Pearson Chi-Square
Continuity Correction a
Likelihood Ratio
Fis her's Exact Test
Linear-by-Linear
As sociation
N of Valid Cases
Value
4.751b
4.263
4.751
As ymp. Sig.
(2-sided)
.029
.039
.029
df
1
1
1
4.737
1
Exact Sig.
(2-sided)
Exact Sig.
(1-sided)
.038
.019
.030
343
a. Computed only for a 2x2 table
b. 0 cells (.0%) have expected count less than 5. The minimum expected count is
53.51.
Symm etri c Me asures
Int erval by Interval
Ordinal by Ordinal
N of Valid Cases
Pearson's R
Spearman Correlat ion
Value
-.118
-.118
343
As ymp.
a
St d. Error
.054
.054
a. Not as suming the null hypothes is.
b. Us ing the asymptotic s tandard error ass uming the null hypot hesis.
c. Based on normal approximation.
SOCCODIC * GUARDIAN
18
b
Approx . T
-2. 188
-2. 188
Approx . Sig.
.029c
.029c
Crosstab
Count
SOCCODIC
GUARDIAN
.000
1.000
98
62
130
52
228
114
.000
1.000
Total
Total
160
182
342
Chi-Square Tests
Pearson Chi-Square
Continuity Correction a
Likelihood Ratio
Fis her's Exact Test
Linear-by-Linear
As sociation
N of Valid Cases
Value
3.970b
3.525
3.968
As ymp. Sig.
(2-sided)
.046
.060
.046
df
1
1
1
3.958
1
Exact Sig.
(2-sided)
Exact Sig.
(1-sided)
.051
.030
.047
342
a. Computed only for a 2x2 table
b. 0 cells (.0%) have expected count less than 5. The minimum expected count is
53.33.
Symm etri c Me asures
Int erval by Interval
Ordinal by Ordinal
N of Valid Cases
Pearson's R
Spearman Correlat ion
Value
-.108
-.108
342
As ymp.
a
St d. Error
.054
.054
a. Not as suming the null hypothes is.
b. Us ing the asymptotic s tandard error ass uming the null hypot hesis.
c. Based on normal approximation.
SOCCODIC * DAILYMAI
Crosstab
Count
SOCCODIC
Total
.000
1.000
DAILYMAI
.000
1.000
125
35
103
79
228
114
Total
160
182
342
19
b
Approx . T
-1. 998
-1. 998
Approx . Sig.
.046c
.046c
Chi-Square Tests
Pearson Chi-Square
Continuity Correction a
Likelihood Ratio
Fis her's Exact Test
Linear-by-Linear
As sociation
N of Valid Cases
Value
17.764 b
16.808
18.141
As ymp. Sig.
(2-sided)
.000
.000
.000
df
1
1
1
17.712
1
Exact Sig.
(2-sided)
Exact Sig.
(1-sided)
.000
.000
.000
342
a. Computed only for a 2x2 table
b. 0 cells (.0%) have expected count less than 5. The minimum expected count is
53.33.
Symm etri c Me asures
Int erval by Interval
Ordinal by Ordinal
N of Valid Cases
Pearson's R
Spearman Correlat ion
Value
.228
.228
342
As ymp.
a
St d. Error
.051
.051
b
Approx . T
4.316
4.316
Approx . Sig.
.000c
.000c
a. Not as suming the null hypothes is.
b. Us ing the asymptotic s tandard error ass uming the null hypot hesis.
c. Based on normal approximation.
Results of a series of chi-square tests between the newspaper coverage and three
main consequences of enlargement revealed that there is a statistically significant
relationship between the socio-cultural consequences of enlargement and the way three
newspapers under focus of this study covers this issue. The most significant relationship
was observed between the coverage of the Daily Mail and the socio-cultural
consequences of enlargement. This is probably due to the tabloid format and the populist
framing of this particular paper.
Analysis of how enlargement is described in the press and by the public
established a ranking of its attributes. In turn, these rankings on each agenda were
compared to determine the strength of the press’ attribute agenda-setting influence on
British public. When doing that the “affective attributes” of enlargement were taken into
20
account. The articles were previously coded according to the negativity/positivity scale
(ranging from -2 to 0 to +2). But due to low levels of inter-coder reliability, the -2 and -1,
and +2 and +1 were merged to achieve greater level of reliability. After that, a rank
ordering of the press coverage of various aspects of enlargement and the public’s rating
of the importance of these aspects was undertaken. Below are the results of this analysis
for the coverage from December 2002 to March, 26 2003, and the results of Flash EB 140
survey which was executed from 21st to 23rd of March 2003.
Nonparametric Correlations
Correlations
Spearman's rho
Newscoverage
Correlation
Coefficient
Sig. (2-tailed)
N
Flash140
Correlation
Coefficient
Sig. (2-tailed)
N
Newscover
Flash140
1.000
.188
.
.455
16
16
.188
1.000
.455
.
16
16
Although the results of this rank order correlation was not statistically significant,
the relationship between the news coverage and the Flash EB 140 public opinion data
points to a positive correlation on the various consequences of enlargement. The reason
for its insignificance might be the small N, which is 39 newspaper articles.
TSPLOT of Newspaper Coverage One Month before each Flash
EB surveys
21
1.0
.8
JUNSEP02
.6
FLEB1321
.4
OCTNOV02
FLEB1322
.2
DECMAR03
d
re
tu
ul
C red
fa
e l oy
W pl
m
ne
U
nd
ro
v i ug
Engsm
ru
D
i
al
eg
Il l i n
e a
t tl m
Se nd
pa
Ex
u
at
tn
is ci
H ote
em
R
ta
to
iff s
D tl e
n
ou c
C ndi
pe
Ex gd i
n
ro
St dic
al
or c
M edi
ac
Pe
0.0
FLEB142
CONSEQUE
As can be seen from the above figure, the largest difference in coverage has been
observed on the issue of both illegal immigration and immigration. Increasingly, more
and more newspaper articles started to cover the possibility that illegal immigration will
become a problem in Britain. Most of the newspaper articles criticized government’s
“soft touch” on asylum and immigration. This effect was most apparent in Daily Mail’s
coverage.
22
1.0
.8
.6
.4
.2
JUNSEP02
d
re
tu
ul
C red
fa
e l oy
W pl
m
ne
U
nd
ro
v i ug
En sm
g
ru
D
i
al
eg
Il l i n
e a
t tl m
Se nd
pa
Ex
u
at
tn
is ci
H te
o
em
R
ta
to
iff s
D tl e
n
ou c
C ndi
pe
Ex d i
ng
ro
St
c
di
al
or c
M edi
ac
Pe
d
re
tu
ul
C red
fa
e l loy
W p
m
ne
U
nd
ro
vi ug
En sm
g
ru
D
i
al
eg
Il l i n
e a
t tl m
Seand
p
Ex tu
a
tn
is ci
H ote
em
R
ta
to
iff s
D tl e
n
ou c
C ndi
pe
Ex d i
ng
ro
c
St
di
al
or c
M edi
ac
Pe
FLEB1322
0.0
FLEB1321
0.0
CONSEQUE
1.0
.8
.6
.4
.2
OCTNOV02
CONSEQUE
23
If we take a closer look on the above graph which compares public attitudes on
the specific consequences of enlargement as depicted in the Flash Eurobarometer 132.2
survey data and media’s treatment of these attributes one month before the fieldwork of
the survey; we can see that regarding the second-level effects in most attributes, the
public opinion and the media has been at odds.
1.0
.8
.6
.4
.2
DECMAR03
d
re
tu
ul
C red
fa
e l oy
W pl
m
ne d
U
n
ro
v i ug
En sm
g
ru
D
i
al
eg
Ill i n
e a
t tl m
Se nd
pa
Ex
u
at
tn
is ci
H ote
em
R
ta
to
iff s
D tl e
n
ou c
C ndi
pe
Ex gd i
n
ro
St dic
al
or c
M edi
ac
Pe
0.0
FLEB142
CONSEQUE
Also, as we look at the above graph comparing the public opinion survey and
news coverage from December to March 2003, we can see that both in the media and in
the public there are rising concerns on increase in the number of Eastern European
citizens coming to Britain both legally and illegally. As a side effect, there is a parallel
concern that immigration would lead to a rise in the unemployment level.
24
The second-level emphasizes the attributes of public issues that should concern
the public. So far, the research does not point out to strong second-level effects on public
opinion. But more in-depth studies should be conducted. The non-significant
relationships may be due to the small population sizes of the study. Since the coverage of
enlargement was limited, this study could not get a hold of a large N. The coverage of
enlargement reached its peak in April 2004, and even then only a total of 45 newspaper
articles have been published a month.15
50
45
40
35
30
25
20
15
10
5
04
Y
A 04
MR 4
P 0
A R4
A
M B0
E 04
FN 3
0
JA C 03
E
DV
O 03
N T 3
C
O P 03
E 0
SG
U 3
A L 03
0
JUN 03
JUA Y 3
0
MR 3
P 0
A R3
A
M B0
E 03
FN 2
0
JA C 02
E
DV 2
O
N T 02
C 0
O P 02
E
SG2
U
A L 02
0
JUN 2
0
JU Y 2
A 0
M R 02
P
AA R 2
M B 02
E 0
FN
JA
0
As can be seen from the above graph, the number of enlargement-related articles
have fluctuated over time. It has especially reached peaks during the European Council
meetings. And as the date of enlargement has approached it increased considerably.
15
Further exploration of the second-level analysis is attached at the end of this article in the Annexes
section.
25
One methodological issue needs to be addressed here. There is no guarantee that
the media coverage was actually seen and processed by the respondents to the
Eurobarometer polls. Exposure is an important variable in the agenda-setting process.16
However, the three newspapers here have the largest circulations of any newspaper in the
region. Previous research also suggests that large newspapers in a region can set the
agenda of smaller papers. Thus, the agenda of attributes in these three papers could have
been filtered down to smaller papers and other media, making the media agenda
accessible to all respondents throughout the country.
Discussion
The 1972 referendums in Ireland, Denmark, Norway and France marked the
public’s first direct participation in Community affairs. Britain’s referendum on
continued EC membership in 1975 broadened the public’s role in deciding the future
course of integration. The direct elections of the European Parliament have
institutionalized this trend.17 And now that there are increasing signs that referenda will
be held in several member states of the EU for the Turkish membership, the importance
of convincing the public becomes more visible.
Enlargement issue brought with it a number of controversies raised in the British
public. Take, for example, the economic consequences of enlargement: since the funding
of enlargement is mostly dependent upon the former EU15 Member States, informing
public opinion about enlargement and gaining public support in these countries is
essential.
16
17
Wanta, 1997.
Dalton and Duval 1986, p. 119.
26
Some opponents of enlargement fear the arrival of cheap labor from the new
Member States, or that enlargement will increase drug trafficking and organized crime in
their countries. On the other hand, some proponents of enlargement think that an enlarged
European Union will be better off economically, and play a stronger political role on the
international scene despite increased difficulties in the decision-making process.
Public opinion has forced a change on government positions on the free
movement of labor and British immigration policy in overall. As a result of the
heightened alarm in the British media, especially in the tabloids, there emerged growing
concerns among the British public about more and more Eastern Europeans flying to
Britain with a “one-way ticket”. While the British government was pursuing a liberal
policy towards asylum and immigration; as a result of the public pressure, the
government at the final hour, changed its position and applied a number of limitations to
the free movement of labor coming from Eastern Europe.
Regarding the first-level agenda-setting effects, this research revealed supporting
evidence on the transfer of issue salience from the media to the public. Although the
current study presents some valuable findings on the second-level agenda-setting effects,
this topic requires further research to better account for the impact of the media on
shaping the public’s framing of the enlargement.
Bibliography
Anderson, P. J. and Weymouth, A. (1999) Insulting the Public? The British Press and the
European Union. Harlow: Longman
Capella, J. and Jamieson, K. (1997). Spiral of Cynicism. New York: Oxford University
Press.
Carey and Burton, “Research Note: The Influence of the Press in Shaping Public Opinion
towards the European Union in Britain” Political Studies (2004) vol 52, pp. 623640.
27
Cohen, Bernard C. (1963). The Press and Foreign Policy, Princeton, NJ: Princeton
University Press.
Curtice, J. (1999) “Was It The Sun Wot Won It Again? The Influence of Newspapers in
the 1997 Election Campaign”, Working Paper 75, Oxford: Centre for Research
into Elections and Social Trends.
Dalton, Russell J. and Robert Duval, “The Environment and Foreign Policy Options:
British Attitudes toward European Integration, 1972-1979”, British Journal of
Political Science, Vol. 16, No. 1 (January 1986), 113-134.
De Vreese, C. H. (2003). Framing Europe: Television News and European Integration.
Amsterdam: Aksant Academic Publishers.
Eichenberg, Robert C. and Russell J. Dalton, “Europeans and the European Community:
The Dynamics of Public Support for European Integration,” International
Organization 47, 4, Autumn 1993, pp. 507-534.
Entman, R. (1993). “Framing: Toward Clarification of a Fractured Paradigm,” Journal of
Communication, vol. 43, i. 4, p. 51-58.
---- (1991) “Framing US Coverage of International News: Contrasts in narratives
of the KAL and Iran Air Incidents”, Journal of Communication, Vol. 41, i. 4, p. 627.
Ghanem, Salma I. “Filling in the Tapestry: The Second Level of Agenda Setting,” in
Communication and Democracy, ed. Maxwell E. McCombs, Donald L. Shaw, and
David H. Weaver (NJ: Lawrence Erlbaum, 1997), 3-14
Graber, D. (1988). Processing the News: How People Tame the Information Tide. New
York: Longman.
Graber, D. (1993). Mass Media and American Politics. Washington, D.C.: CQ Press.
Hsiang Iris Chyi and Maxwell McCombs, “Media Salience and the Process of Framing:
Coverage of the Columbine School Shootings”, Journalism and Mass
Communication Quarterly, Vol. 81, No. 1, Spring 2004, pp. 22-35.
Holsti, Ole R.(1969) Content Analysis for the Social Sciences and Humanities, MA:
Addison-Wesley.
Iyengar, S. and D. Kinder. (1987a). News That Matters. Chicago: University of Chicago
Press.
Iyengar, S. (1987b). Television News and Citizens’ Explanation of National Affairs.
American Political Science Review, 81, p. 815-831.
Iyengar, S. (1991). Is Anyone Responsible? How Television Frames Political Issues.
Chicago: University of Chicago Press.
Jones, Bryan D., and Frank R. Baumgartner. (2004). “Representation and Agenda
Setting,” Policy Studies Journal, February 2004, v. 32, i. 1, p. 1- 24.
King, P. (1997). “The Press, Candidate Images, and Voter Perceptions,” in McCombs, M.
E., D. L. Shaw, and D. Weaver (Eds.), Communication and Democracy, Mahwah
NJ: Erlbaum.
Kiousis S. and M. E. McCombs. (2004) “Agenda-Setting Effects and Attitude Strength:
Political Figures during the 1996 Presidential Election,” Communication
Research, vol. 31, p. 36-57.
Lippmann, W. (1922). Public Opinion. New York: Macmillan.
McCombs, M. E. and D.L. Shaw. (1972). “The Agenda-setting Function of Mass Media,”
Public Opinion Quarterly, vol. 36, p.176-184.
28
McCombs, M. E. (2000). “The Agenda-Setting Role of the Mass Media in the Shaping of
Public Opinion”, paper.
McCombs, M. E. and A. Reynolds. (2002). “News Influence on Our Pictures of the
World,” in J. Bryant and D. Zillmann (Eds.), Media effects: Advances in theory
and research (2nd Ed.) (pp. 1-18). Mahwah, NJ: Lawrence Erlbaum.
Mendelsohn, M. (1993). “Television Frames in the 1988 Canadian Elections”, Canadian
Journal of Communication, vol. 18, p. 149-171.
Meyer, Christopher. (1999). “Political Legitimacy and the Invisibility of Politics:
Exploring the European Union’s Communication Deficit,” Journal of Common
Market Studies, December 1999, vol. 37, i. 4, p. 617-638.
Nelson, T. E. Clawson, R. A. and Oxley Z. M. (1997). “Media Framing of a Civil
Liberties Conflict and Its Effect on Tolerance”, American Political Science
Review, vol. 91, p. 567-583.
Newton, K. and Brynin, M. (2001) “The National Press and Party Voting in the UK”,
Political Studies, 49 (2), pp. 265-285.
Norris, P. (1995). The Restless Searchlight: Network News Framing of the Post-Cold
War World. Political Communication, 12, 357-370.
Norris, P., Curtice, J., Sanders D., Scammell, M. and Semetko, H. (1999) On Message:
Communicating the Campaign. London: Sage.
Page, B. I. and R. Y. Shapiro. (1992). The Rational Public: Fifty Years of Trends in
Americans’ Policy Preferences. Chicago: Chicago University Press.
Patterson, T. (1993). Out of Order. New York: Knopf.
Rogers, E. M. and Chang S. (1991). “Media Coverage of Technology Issues: Ethiopian
Drought of 1984, AIDS, Challenger, and Chernobyl,” in Wilkens, L. and P.
Patterson (Eds.) Risky Business: Communicating the Issues of Science, Risk, and
Public Policy, New York: Greenwood.
Rogers, E. M., J. W. Dearing, and D. Bregman. (1993). “The Anatomy of Agenda-Setting
Research,” Journal of Communication, vol. 43 i. 2, p. 68-84.
Rohrschneider, R. (2000). “The Democracy Deficit and Mass Support for an EU-wide
Government,” American Journal of Political Science, vol. 46, p. 463-75
Saris, Willem E. (1997). “The Public Opinion about the EU Can Easily Be Swayed in
Different Directions,” Acta Politica: International Journal of Political Science,
vol. 32, 406-35.
Schoenbach, K. and Semetko Holli A. (1997). “Agenda setting, Agenda Reinforcing or
Agenda Deflating? A Study of the 1990 German National Election,” Journalism
Quarterly, vol 68, p. 837-846.
Semetko, H., J. G. Blumler, M. Gurevitch, and D. Weaver. (1991).The Formation of
Campaign Agendas: A Comparative Analysis of Party and Media Roles in Recent
American and British Elections. Hillsdale, NJ: Lawrence Erlbaum Associates.
Semetko, H. A., Claes H. de Vreese, and Peter Jochen. (2000). “Europeanised Politics –
Europeanised Media? European Integration and Political Communication,” West
European Politics, vol. 23, i. 4, p.121.
Semetko H. A., W. Van Der Brug, and Patti M. Valkenburg. (2003). “The Influence of
Political Events on Attitudes towards the European Union,” British Journal of
Political Science, October 2003, vol. 33 i. 4, p. 621-635.
29
Seymour-Ure, C. (1997) “New Labour and the Media”, in A. King (ed.), Britain at the
Polls 2001. Chatham NJ: Chatham House, pp. 117-142.
Wanta, Wayne. (1997). The Public and the National Agenda: How People Learn About
Important Issues. Mahwah, NJ: Lawrence Erlbaum.
Weaver, D. (1991). “Issue Salience and Public Opinion: Are There Consequences of
Agenda-Setting?,” International Journal of Public Opinion Research, vol. 3, p.
53-68.
Zaller, J. (1991). “Information, Values, and Opinion,” American Political Science
Review, vol. 85, p. 1215-1237.
ANNEX I: The Names and the Field Research Dates of Eurobarometer Surveys to
be used in this Research
Standard Eurobarometer (EB) 56.3 - January – February 2002
Standard EB 57.1 - March – May 2002
Flash EB 132.1 - September 2002
Standard EB 58.1 - October – November 2002
Flash EB 132.2 - November 2002
Flash EB 140 - March 2003
Standard EB 59.1 - March – April 2003
Standard EB 60.1 - October – November 2003
Standard EB 61.0 - February – March 2004
* Standard Eurobarometer public opinion data are available through the ICPRS website.
* Flash Eurobarometer on Enlargement data are gathered through the University of
Cologne’s Central Archive for Empirical Social Research Data Service.
ANNEX II: Chi-Square Tests
When I conducted chi-square tests on the specific consequences of enlargement and each
of the newspapers under study, I have found several statistically significant relationships.
For instance, the Guardian mostly emphasizes the peaceful consequences of enlargement
in the European continent.
PEACEDIC * GUARDIAN
Crosstab
Count
PEACEDIC
Total
.000
1.000
GUARDIAN
.000
1.000
190
85
38
30
228
115
Total
275
68
343
30
Chi-Square Tests
Pearson Chi-Square
Continuity Correction a
Likelihood Ratio
Fis her's Exact Test
Linear-by-Linear
As sociation
N of Valid Cases
Value
4.268b
3.696
4.138
As ymp. Sig.
(2-sided)
.039
.055
.042
df
1
1
1
4.255
1
Exact Sig.
(2-sided)
Exact Sig.
(1-sided)
.045
.029
.039
343
a. Computed only for a 2x2 table
b. 0 cells (.0%) have expected count less than 5. The minimum expected count is
22.80.
Symm etri c Me asures
Int erval by Interval
Ordinal by Ordinal
N of Valid Cases
Pearson's R
Spearman Correlat ion
Value
.112
.112
343
As ymp.
a
St d. Error
.056
.056
b
Approx . T
2.073
2.073
Approx . Sig.
.039c
.039c
a. Not as suming the null hypothes is.
b. Us ing the asymptotic s tandard error ass uming the null hypot hesis.
c. Based on normal approximation.
Or when we analyze the crosstabulation of peaceful consequences of enlargement
and the coverage of Daily Mail, we can see that there is a significant negative relationship
between these two variables. This may signify that the Daily Mail has been more
skeptical about the peace and stability implications of enlargement.
PEACEDIC * DAILYMAIL
Crosstab
Count
PEACEDIC
Total
.000
1.000
DAILYMAI
.000
1.000
174
101
55
13
229
114
Total
275
68
343
31
Chi-Square Tests
Pearson Chi-Square
Continuity Correction a
Likelihood Ratio
Fis her's Exact Test
Linear-by-Linear
As sociation
N of Valid Cases
Value
7.619b
6.846
8.209
As ymp. Sig.
(2-sided)
.006
.009
.004
df
1
1
1
7.597
1
Exact Sig.
(2-sided)
Exact Sig.
(1-sided)
.006
.004
.006
343
a. Computed only for a 2x2 table
b. 0 cells (.0%) have expected count less than 5. The minimum expected count is
22.60.
Symm etri c Me asures
Int erval by Interval
Ordinal by Ordinal
N of Valid Cases
Pearson's R
Spearman Correlat ion
Value
-.149
-.149
343
As ymp.
a
St d. Error
.047
.047
b
Approx . T
-2. 783
-2. 783
Approx . Sig.
.006c
.006c
a. Not as suming the null hypothes is.
b. Us ing the asymptotic s tandard error ass uming the null hypot hesis.
c. Based on normal approximation.
Also, while there is a significant positive relationship between the coverage of the
Guardian and the moral implications of enlargement in terms of the reuniting of
European continent for the first time after the end of the Cold War; there is a negative
significant relationship between the coverage of the Daily Mail and the moral duties of
reuniting Europe. Here again, the skeptical attitude of the Daily Mail is obvious.
MORALDIC * GUARDIAN
Crosstab
Count
MORALDIC
Total
.000
1.000
GUARDIAN
.000
1.000
203
88
25
26
228
114
Total
291
51
342
32
Chi-Square Tests
Pearson Chi-Square
Continuity Correction a
Likelihood Ratio
Fis her's Exact Test
Linear-by-Linear
As sociation
N of Valid Cases
Value
8.400b
7.492
7.993
As ymp. Sig.
(2-sided)
.004
.006
.005
df
1
1
1
8.375
1
Exact Sig.
(2-sided)
Exact Sig.
(1-sided)
.006
.004
.004
342
a. Computed only for a 2x2 table
b. 0 cells (.0%) have expected count less than 5. The minimum expected count is
17.00.
Symm etri c Me asures
Int erval by Interval
Ordinal by Ordinal
N of Valid Cases
Pearson's R
Spearman Correlat ion
Value
.157
.157
342
As ymp.
a
St d. Error
.057
.057
a. Not as suming the null hypothes is.
b. Us ing the asymptotic s tandard error ass uming the null hypot hesis.
c. Based on normal approximation.
MORALDIC * DAILYMAIL
Crosstab
Count
MORALDIC
Total
.000
1.000
DAILYMAI
.000
1.000
183
109
46
5
229
114
Total
292
51
343
33
b
Approx . T
2.926
2.926
Approx . Sig.
.004c
.004c
Chi-Square Tests
Pearson Chi-Square
Continuity Correction a
Likelihood Ratio
Fis her's Exact Test
Linear-by-Linear
As sociation
N of Valid Cases
Value
14.824 b
13.609
17.630
As ymp. Sig.
(2-sided)
.000
.000
.000
df
1
1
1
14.780
1
Exact Sig.
(2-sided)
Exact Sig.
(1-sided)
.000
.000
.000
343
a. Computed only for a 2x2 table
b. 0 cells (.0%) have expected count less than 5. The minimum expected count is
16.95.
Symm etri c Me asures
Int erval by Interval
Ordinal by Ordinal
N of Valid Cases
Pearson's R
Spearman Correlat ion
Value
-.208
-.208
343
As ymp.
a
St d. Error
.039
.039
a. Not as suming the null hypothes is.
b. Us ing the asymptotic s tandard error ass uming the null hypot hesis.
c. Based on normal approximation.
EXPENSIVEDIC * DAILYMAIL
Crosstab
Count
EXPENDIC
Total
.000
1.000
DAILYMAI
.000
1.000
193
87
35
27
228
114
Total
280
62
342
34
b
Approx . T
-3. 925
-3. 925
Approx . Sig.
.000c
.000c
Chi-Square Tests
Pearson Chi-Square
Continuity Correction a
Likelihood Ratio
Fis her's Exact Test
Linear-by-Linear
As sociation
N of Valid Cases
Value
3.556b
3.017
3.445
As ymp. Sig.
(2-sided)
.059
.082
.063
df
1
1
1
3.546
1
Exact Sig.
(2-sided)
Exact Sig.
(1-sided)
.074
.043
.060
342
a. Computed only for a 2x2 table
b. 0 cells (.0%) have expected count less than 5. The minimum expected count is
20.67.
Symm etri c Me asures
Int erval by Interval
Ordinal by Ordinal
N of Valid Cases
Pearson's R
Spearman Correlat ion
Value
.102
.102
342
As ymp.
a
St d. Error
.056
.056
b
Approx . T
1.890
1.890
Approx . Sig.
.060c
.060c
a. Not as suming the null hypothes is.
b. Us ing the asymptotic s tandard error ass uming the null hypot hesis.
c. Based on normal approximation.
As can be observed from the above statistics, the Daily Mail covered the issue of
enlargement as expensive for the UK. This is another point supporting the observation
that the Daily Mail treats the issue of enlargement with caution.
COUNT-LESSDIC * GUARDIAN
Crosstab
Count
COLESDIC
Total
.000
1.000
GUARDIAN
.000
1.000
175
100
53
15
228
115
Total
275
68
343
35
Chi-Square Tests
Pearson Chi-Square
Continuity Correction a
Likelihood Ratio
Fis her's Exact Test
Linear-by-Linear
As sociation
N of Valid Cases
Value
5.006b
4.384
5.292
As ymp. Sig.
(2-sided)
.025
.036
.021
df
1
1
1
4.991
1
Exact Sig.
(2-sided)
Exact Sig.
(1-sided)
.031
.016
.025
343
a. Computed only for a 2x2 table
b. 0 cells (.0%) have expected count less than 5. The minimum expected count is
22.80.
Symm etri c Me asures
Int erval by Interval
Ordinal by Ordinal
N of Valid Cases
Pearson's R
Spearman Correlat ion
Value
-.121
-.121
343
As ymp.
a
St d. Error
.049
.049
b
Approx . T
-2. 247
-2. 247
Approx . Sig.
.025c
.025c
a. Not as suming the null hypothes is.
b. Us ing the asymptotic s tandard error ass uming the null hypot hesis.
c. Based on normal approximation.
Surprisingly, the Guardian has framed enlargement as creating a situation in
which the UK will count less within the 25-membered Union; while the Daily Mail has
framed it as an opportunity in which Britain will count more within an enlarged EU.
COUNT-LESSDIC * DAILYMAIL
Crosstab
Count
COLESDIC
Total
.000
1.000
DAILYMAI
.000
1.000
191
83
38
30
229
113
Total
274
68
342
36
Chi-Square Tests
Pearson Chi-Square
Continuity Correction a
Likelihood Ratio
Fis her's Exact Test
Linear-by-Linear
As sociation
N of Valid Cases
Value
4.707b
4.103
4.553
As ymp. Sig.
(2-sided)
.030
.043
.033
df
1
1
1
4.693
1
Exact Sig.
(2-sided)
Exact Sig.
(1-sided)
.043
.023
.030
342
a. Computed only for a 2x2 table
b. 0 cells (.0%) have expected count less than 5. The minimum expected count is
22.47.
Symm etri c Me asures
Int erval by Interval
Ordinal by Ordinal
N of Valid Cases
Pearson's R
Spearman Correlat ion
Value
.117
.117
342
As ymp.
a
St d. Error
.056
.056
a. Not as suming the null hypothes is.
b. Us ing the asymptotic s tandard error ass uming the null hypot hesis.
c. Based on normal approximation.
HISNADIC * GUARDIAN
Crosstab
Count
HISNADIC
Total
.000
1.000
GUARDIAN
.000
1.000
203
90
25
24
228
114
Total
293
49
342
37
b
Approx . T
2.178
2.178
Approx . Sig.
.030c
.030c
Chi-Square Tests
Pearson Chi-Square
Continuity Correction a
Likelihood Ratio
Fis her's Exact Test
Linear-by-Linear
As sociation
N of Valid Cases
Value
6.301b
5.506
6.014
As ymp. Sig.
(2-sided)
.012
.019
.014
df
1
1
1
6.282
1
Exact Sig.
(2-sided)
Exact Sig.
(1-sided)
.014
.011
.012
342
a. Computed only for a 2x2 table
b. 0 cells (.0%) have expected count less than 5. The minimum expected count is
16.33.
Symm etri c Me asures
Int erval by Interval
Ordinal by Ordinal
N of Valid Cases
Pearson's R
Spearman Correlat ion
Value
.136
.136
342
As ymp.
a
St d. Error
.057
.057
a. Not as suming the null hypothes is.
b. Us ing the asymptotic s tandard error ass uming the null hypot hesis.
c. Based on normal approximation.
HISTORICALLYNATURALDIC * DAILYMAIL
Crosstab
Count
HISNADIC
Total
.000
1.000
DAILYMAI
.000
1.000
185
108
43
6
228
114
Total
293
49
342
38
b
Approx . T
2.526
2.526
Approx . Sig.
.012c
.012c
Chi-Square Tests
Pearson Chi-Square
Continuity Correction a
Likelihood Ratio
Fis her's Exact Test
Linear-by-Linear
As sociation
N of Valid Cases
Value
11.446 b
10.365
13.232
As ymp. Sig.
(2-sided)
.001
.001
.000
df
1
1
1
11.413
1
Exact Sig.
(2-sided)
Exact Sig.
(1-sided)
.000
.000
.001
342
a. Computed only for a 2x2 table
b. 0 cells (.0%) have expected count less than 5. The minimum expected count is
16.33.
Symm etri c Me asures
Int erval by Interval
Ordinal by Ordinal
N of Valid Cases
Pearson's R
Spearman Correlat ion
Value
-.183
-.183
342
As ymp.
a
St d. Error
.041
.041
b
Approx . T
-3. 431
-3. 431
Approx . Sig.
.001c
.001c
a. Not as suming the null hypothes is.
b. Us ing the asymptotic s tandard error ass uming the null hypot hesis.
c. Based on normal approximation.
The framing of enlargement as a historically natural process again points us to
some expected results. Previous two chi-square tests reveal that the Daily Mail does not
frame enlargement as a historically natural process, while the Guardian frames it as a
justified and natural development.
SETTLEINOURCOUNTRYDIC * GUARDIAN
Crosstab
Count
SETTLDIC
Total
.000
1.000
GUARDIAN
.000
1.000
149
86
78
29
227
115
Total
235
107
342
39
Chi-Square Tests
Pearson Chi-Square
Continuity Correction a
Likelihood Ratio
Fis her's Exact Test
Linear-by-Linear
As sociation
N of Valid Cases
Value
2.969b
2.559
3.032
As ymp. Sig.
(2-sided)
.085
.110
.082
df
1
1
1
2.960
1
Exact Sig.
(2-sided)
Exact Sig.
(1-sided)
.108
.054
.085
342
a. Computed only for a 2x2 table
b. 0 cells (.0%) have expected count less than 5. The minimum expected count is
35.98.
Symm etri c Me asures
Int erval by Interval
Ordinal by Ordinal
N of Valid Cases
Pearson's R
Spearman Correlat ion
Value
-.093
-.093
342
As ymp.
a
St d. Error
.052
.052
a. Not as suming the null hypothes is.
b. Us ing the asymptotic s tandard error ass uming the null hypot hesis.
c. Based on normal approximation.
SETTLDIC * DAILYMAI
Crosstab
Count
SETTLDIC
Total
.000
1.000
DAILYMAI
.000
1.000
168
67
61
47
229
114
Total
235
108
343
40
b
Approx . T
-1. 725
-1. 725
Approx . Sig.
.085c
.085c
Chi-Square Tests
Pearson Chi-Square
Continuity Correction a
Likelihood Ratio
Fis her's Exact Test
Linear-by-Linear
As sociation
N of Valid Cases
Value
7.511b
6.850
7.361
As ymp. Sig.
(2-sided)
.006
.009
.007
df
1
1
1
7.489
1
Exact Sig.
(2-sided)
Exact Sig.
(1-sided)
.007
.005
.006
343
a. Computed only for a 2x2 table
b. 0 cells (.0%) have expected count less than 5. The minimum expected count is
35.90.
Symm etri c Me asures
Int erval by Interval
Ordinal by Ordinal
N of Valid Cases
Pearson's R
Spearman Correlat ion
Value
.148
.148
343
As ymp.
a
St d. Error
.055
.055
b
Approx . T
2.763
2.763
Approx . Sig.
.006c
.006c
a. Not as suming the null hypothes is.
b. Us ing the asymptotic s tandard error ass uming the null hypot hesis.
c. Based on normal approximation.
On the issue of immigrants settling in Britain, both the Guardian and the Daily
Mail shows similar patterns with differing intensities. They both framed enlargement as
leading to a large number of immigrants settling in Britain, but the Mail has underlined it
relatively more vigorously. Again, the same observation is valid for the issue of illegal
immigration as illustrated by the below chi-square statistics.
ILLIMDIC * GUARDIAN
Crosstab
Count
ILLIMDIC
Total
.000
1.000
GUARDIAN
.000
1.000
164
99
64
16
228
115
Total
263
80
343
41
Chi-Square Tests
Pearson Chi-Square
Continuity Correction a
Likelihood Ratio
Fis her's Exact Test
Linear-by-Linear
As sociation
N of Valid Cases
Value
8.567b
7.794
9.140
As ymp. Sig.
(2-sided)
.003
.005
.003
df
1
1
1
8.542
1
Exact Sig.
(2-sided)
Exact Sig.
(1-sided)
.004
.002
.003
343
a. Computed only for a 2x2 table
b. 0 cells (.0%) have expected count less than 5. The minimum expected count is
26.82.
Symm etri c Me asures
Int erval by Interval
Ordinal by Ordinal
N of Valid Cases
Pearson's R
Spearman Correlat ion
Value
-.158
-.158
343
As ymp.
a
St d. Error
.048
.048
a. Not as suming the null hypothes is.
b. Us ing the asymptotic s tandard error ass uming the null hypot hesis.
c. Based on normal approximation.
ILLIMDIC * DAILYMAI
Crosstab
Count
ILLIMDIC
Total
.000
1.000
DAILYMAI
.000
1.000
191
72
38
42
229
114
Total
263
80
343
42
b
Approx . T
-2. 956
-2. 956
Approx . Sig.
.003c
.003c
Chi-Square Tests
Pearson Chi-Square
Continuity Correction a
Likelihood Ratio
Fis her's Exact Test
Linear-by-Linear
As sociation
N of Valid Cases
Value
17.449 b
16.335
16.737
As ymp. Sig.
(2-sided)
.000
.000
.000
df
1
1
1
17.398
1
Exact Sig.
(2-sided)
Exact Sig.
(1-sided)
.000
.000
.000
343
a. Computed only for a 2x2 table
b. 0 cells (.0%) have expected count less than 5. The minimum expected count is
26.59.
Symm etri c Me asures
Int erval by Interval
Ordinal by Ordinal
N of Valid Cases
Pearson's R
Spearman Correlat ion
Value
.226
.226
343
As ymp.
a
St d. Error
.056
.056
a. Not as suming the null hypothes is.
b. Us ing the asymptotic s tandard error ass uming the null hypot hesis.
c. Based on normal approximation.
DRUGDIC * GUARDIAN
Crosstab
Count
DRUGDIC
Total
.000
1.000
GUARDIAN
.000
1.000
209
113
19
2
228
115
Total
322
21
343
43
b
Approx . T
4.275
4.275
Approx . Sig.
.000c
.000c
Chi-Square Tests
Pearson Chi-Square
Continuity Correction a
Likelihood Ratio
Fis her's Exact Test
Linear-by-Linear
As sociation
N of Valid Cases
Value
5.783b
4.693
7.033
As ymp. Sig.
(2-sided)
.016
.030
.008
df
1
1
1
5.766
1
Exact Sig.
(2-sided)
Exact Sig.
(1-sided)
.016
.010
.016
343
a. Computed only for a 2x2 table
b. 0 cells (.0%) have expected count less than 5. The minimum expected count is
7.04.
Symm etri c Me asures
Int erval by Interval
Ordinal by Ordinal
N of Valid Cases
Pearson's R
Spearman Correlat ion
Value
-.130
-.130
343
As ymp.
a
St d. Error
.037
.037
a. Not as suming the null hypothes is.
b. Us ing the asymptotic s tandard error ass uming the null hypot hesis.
c. Based on normal approximation.
ENVIRDIC * GUARDIAN
Crosstab
Count
ENVIRDIC
Total
.000
1.000
GUARDIAN
.000
1.000
221
105
7
9
228
114
Total
326
16
342
44
b
Approx . T
-2. 418
-2. 418
Approx . Sig.
.016c
.016c
Chi-Square Tests
Pearson Chi-Square
Continuity Correction a
Likelihood Ratio
Fis her's Exact Test
Linear-by-Linear
As sociation
N of Valid Cases
Value
3.967b
2.959
3.708
As ymp. Sig.
(2-sided)
.046
.085
.054
df
1
1
1
3.955
1
Exact Sig.
(2-sided)
Exact Sig.
(1-sided)
.058
.046
.047
342
a. Computed only for a 2x2 table
b. 0 cells (.0%) have expected count less than 5. The minimum expected count is
5.33.
Symmetric Measures
Value
Asymp.
Std.
Error(a)
Approx.
T(b)
Approx. Sig.
Interval by Interval
Pearson's R
.108
.058
1.997
.047(c)
Ordinal by Ordinal
Spearman
Correlation
.108
.058
1.997
.047(c)
N of Valid Cases
342
a Not assuming the null hypothesis.
b Using the asymptotic standard error assuming the null hypothesis.
c Based on normal approximation.
The Guardian’s coverage on enlargement also emphasized the good effects of
enlargement on cooperation on environment.
UNEMPDIC * TIMES
Crosstab
Count
TIMES
UNEMPDIC
Total
.000
1.000
.000
172
57
229
1.000
97
17
114
Total
269
74
343
45
Chi-Square Tests
Pearson Chi-Square
Continuity Correction a
Likelihood Ratio
Fis her's Exact Test
Linear-by-Linear
As sociation
N of Valid Cases
Value
4.479b
3.909
4.698
As ymp. Sig.
(2-sided)
.034
.048
.030
df
1
1
1
4.466
1
Exact Sig.
(2-sided)
Exact Sig.
(1-sided)
.037
.022
.035
343
a. Computed only for a 2x2 table
b. 0 cells (.0%) have expected count less than 5. The minimum expected count is
24.59.
Symm etri c Me asures
Int erval by Interval
Ordinal by Ordinal
N of Valid Cases
Pearson's R
Spearman Correlat ion
Value
-.114
-.114
343
As ymp.
a
St d. Error
.050
.050
b
Approx . T
-2. 124
-2. 124
Approx . Sig.
.034c
.034c
a. Not as suming the null hypothes is.
b. Us ing the asymptotic s tandard error ass uming the null hypot hesis.
c. Based on normal approximation.
The Times draw an optimistic picture of the issue of enlargement’s impact on
unemployment. The Daily Mail, on the other hand, following its usual Euro-skeptic line,
emphasized the possibility of a rise in unemployment after the enlargement.
UNEMPLOYMENTDIC * DAILYMAIL
Crosstab
Count
UNEMPDIC
Total
.000
1.000
DAILYMAI
.000
1.000
190
79
38
35
228
114
Total
269
73
342
46
Chi-Square Tests
Pearson Chi-Square
Continuity Correction a
Likelihood Ratio
Fis her's Exact Test
Linear-by-Linear
As sociation
N of Valid Cases
Value
8.917b
8.101
8.587
As ymp. Sig.
(2-sided)
.003
.004
.003
df
1
1
1
8.891
1
Exact Sig.
(2-sided)
Exact Sig.
(1-sided)
.005
.003
.003
342
a. Computed only for a 2x2 table
b. 0 cells (.0%) have expected count less than 5. The minimum expected count is
24.33.
Symm etri c Me asures
Int erval by Interval
Ordinal by Ordinal
N of Valid Cases
Pearson's R
Spearman Correlat ion
Value
.161
.161
342
As ymp.
a
St d. Error
.056
.056
a. Not as suming the null hypothes is.
b. Us ing the asymptotic s tandard error ass uming the null hypot hesis.
c. Based on normal approximation.
WELFADIC * GUARDIAN
Crosstab
Count
W ELFADIC
Total
.000
1.000
GUARDIAN
.000
1.000
167
98
61
17
228
115
Total
265
78
343
47
b
Approx . T
3.017
3.017
Approx . Sig.
.003c
.003c
Chi-Square Tests
Pearson Chi-Square
Continuity Correction a
Likelihood Ratio
Fis her's Exact Test
Linear-by-Linear
As sociation
N of Valid Cases
Value
6.236b
5.573
6.583
As ymp. Sig.
(2-sided)
.013
.018
.010
df
1
1
1
6.218
1
Exact Sig.
(2-sided)
Exact Sig.
(1-sided)
.014
.008
.013
343
a. Computed only for a 2x2 table
b. 0 cells (.0%) have expected count less than 5. The minimum expected count is
26.15.
Symm etri c Me asures
Int erval by Interval
Ordinal by Ordinal
N of Valid Cases
Pearson's R
Spearman Correlat ion
Value
-.135
-.135
343
As ymp.
a
St d. Error
.049
.049
a. Not as suming the null hypothes is.
b. Us ing the asymptotic s tandard error ass uming the null hypot hesis.
c. Based on normal approximation.
WELFADIC * DAILYMAIL
Crosstab
Count
W ELFADIC
Total
.000
1.000
DAILYMAI
.000
1.000
190
74
38
40
228
114
Total
264
78
342
48
b
Approx . T
-2. 513
-2. 513
Approx . Sig.
.012c
.012c
Chi-Square Tests
Pearson Chi-Square
Continuity Correction a
Likelihood Ratio
Fis her's Exact Test
Linear-by-Linear
As sociation
N of Valid Cases
Value
14.649 b
13.621
14.066
As ymp. Sig.
(2-sided)
.000
.000
.000
df
1
1
1
14.606
1
Exact Sig.
(2-sided)
Exact Sig.
(1-sided)
.000
.000
.000
342
a. Computed only for a 2x2 table
b. 0 cells (.0%) have expected count less than 5. The minimum expected count is
26.00.
Symm etri c Me asures
Int erval by Interval
Ordinal by Ordinal
N of Valid Cases
Pearson's R
Spearman Correlat ion
Value
.207
.207
342
As ymp.
a
St d. Error
.056
.056
b
Approx . T
3.901
3.901
Approx . Sig.
.000c
.000c
a. Not as suming the null hypothes is.
b. Us ing the asymptotic s tandard error ass uming the null hypot hesis.
c. Based on normal approximation.
Finally, as expected, the Daily Mail emphasized the negative impacts of
enlargement on British welfare system, while the Guardian covered the issue quite
positively.
49