CHAPTER 3
Understanding Political Communications
Although there is an abundance of heated rhetoric and conjecture,
and everyone who watches television seems to have a view about
the issue, it has been surprisingly difficult to find systematic evidence
that proves the media malaise hypothesis. There is a broad consensus
that the process of political communication has changed, but it remains
questionable whether this has had a major impact on the contents of
election news, still less influenced public attitudes and behavior. Unfortunately,
discussions of the perceived problems of the news media often
fail to distinguish between criticisms based on unsystematic observations
and those based on more solid ground. Many recent books on the
news media, in discussing phenomena such as trends towards 'soft' or
'infotainment' news, have simply assumed that the content of news
coverage must influence the public, in a simple 'hypodermic-syringe
model', with no prior evidence. But this model has been largely abandoned
in communications research as we have come to realize that the
public actively react to, deconstruct, and interpret what they watch and
read, rather than simply absorbing messages like passive sponges.1
The
attempt to understand the political influence of the news media raises
difficult theoretical and methodological challenges. Previous studies
exploring whether political coverage in the news media contributes
towards civic malaise have generally employed trend analysis, experimental
designs, or cross-sectional surveys, and each of those methods
has certain advantages and disadvantages.
TREND ANALYSIS: DIFFUSE THEORIES OF
MEDIA MALAISE
One approach has been to compare trends in the content of news
coverage with trends in public opinion. Popular accounts often assume
36
UNDERSTANDING POLITICAL COMMUNICATIONS
a causal connection if negative news about government has grown in
recent years along with public cynicism about political institutions.
Content analysis provides a systematic description of the media landscape,
and monthly polls monitor the pulse of public opinion.2
The
media are then believed to exert a diffuse, long-term, and cumulative
influence on the political culture. It is the steady repetition of messages
over and over again, not individual exposure, that is thought to
entrench mainstream orientations in most viewers. Robinson and
Sheehan first suggested the linkage between declining trust in American
government institutions and the rise of television news.3
As encapsulated
by Austin Ranney: 'It is hard not to put two facts side by side:
one is the fact that the age of television began in the 1950s and reached
its presen dominance by the mid-1960s; the other is the fact that the
rise in public cynicism has been continuous through the same period.
. . . These two facts do not prove that television portrayal of politics
explains all the decline in confidence, but it is not unreasonable ... to
conclude that television has made a major contribution to that decline.'4
Similar studies in Germany, Sweden, and Japan have shown that
increased coverage of scandals and negative news has accompanied
declining confidence in political leaders.5
That approach was exemplified in an influential study in which
Thomas Patterson argued that there had been a shift in the culture and
values of American television journalism in the post-Vietnam, postWatergate
era that had gradually infected the rest of the news media.6
For evidence, that study examined the evaluative tone of coverage of
American presidential elections in Time and Newsweek since the 1960s,
and it found increasing negativity in election news: The proportion of
'bad news' in news magazines accounted for about one-quarter of campaign
coverage in 1960. That grew to about 40% in presidential elections
from 1964 to 1976, and then rose to about 50-60% in elections
from 1980 to 1992. The data followed a pattern of stepped plateaus,
rather than a steady linear rise. Although we lack direct evidence monitoring
the culture of journalism in this period, Patterson argued that
Vietnam and Watergate were seminal events that transformed American
news, as the press turned against politicians.
While intuitively plausible, time-series analysis faces two main challenges
before it can be accepted as fully convincing. First, can we assume
that there has been a substantial change in the content of news over
time, with the growth of'negative news' or 'infotainment'? The evidence
available from the United States is limited, and Patterson's data from
37
THE NEWS MEDIA AND CIVIC MALAISE
those sources may be unrepresentative of the broad range of news
media.7
Dalton, Beck, and Huckfeldt's analysis of campaign coverage
in a nationwide sample from the U.S. press in the 1992 elections
found that the contents were fairly neutral, with local papers presenting
their readers with multiple messages about each party and
candidate, or a diverse set of evaluative viewpoints, rather than predominantly
negative coverage.8
Systematic content analysis that could
compare long-term trends in typical news coverage across many countries
is lacking, and, as discussed in the next chapter, the existing
evidence from Britain and Germany challenges the common assumption
of growing tabloidization in these countries.9
The declinism thesis
may be falling into the trap of assuming a 'golden age' of journalism
that in fact, proves mythical. 'Tabloidization' refers simultaneously, and
thereby ambiguously, both to news formats and to subjects. As discussed
in detail in the next two chapters, one possible interpretation of developments
in recent decades is that perhaps the news may have diversified
into both more popular and more serious formats, rather than
simply having moved down-market in terms of the types of stories
covered.
Even if we accept the presumed changes in the content of news, with
the growth of tabloid or negative news in the United States and Europe,
as a working assumption, we still face a large inferential leap before
we can establish the impact of news coverage on public opinion. The
evidence in the macro-level studies is open to many alternative
interpretations.
Any parallel trends over time may in fact be independent. There may
be no systematic linkage between the type of coverage and the public's
response. Studies have found that even when political news on American
networks has used a conflict frame, for example in covering the
debate between the president and Congress over the issue of gays in the
military, the public tended to discount such framing, instead interpreting
the story in terms of the underlying events or the merits of particular
policy proposals.10
Even if news proves negative or conflictual,
therefore, content analysis may provide a misleading picture of how the
public respond and construct their understandings from the messages
they see.
Or the association may prove spurious, as the result of other causal
factors: An increase in the incidence of government corruption, for
example, might logically produce both more negative media coverage
and greater public cynicism about politicians. In the cultural account,
38
UNDERSTANDING POLITICAL COMMUNICATIONS
it is particularly important that the timing of events in Vietnam and
Watergate should relate systematically to changes over time in political
coverage; otherwise, many factors could be said to have driven trends
in the news in recent decades.11
The increased focus on strategy in
American campaign news during the 1970s and 1980s, for example,
might plausibly be explained by actual changes in electioneering, such
as the rising importance of primaries, the lengthening of the campaign
season, the declining salience of many of the hot-button issues of the
1960s and the accompanying generational and cultural conflict, and
above all the growth of professional political marketing. Campaign
news may have changed to reflect the fact that election strategy has
become more important, and substantive policy issues have become less
important, in determining election outcomes.
Equally plausibly, the direction of causality might be reversed: If
political news seems more cynical, that might be the result rather than
the cause of cynicism in the wider political culture. As Robinson and
Sheehan note, television journalists are part of a broader set of norms
and values in society:12
'To some degree the entire process must be circular,
with the networks affecting the public and the public affecting the
networks in return.' They argue that the media influences the public,
because in several instances, such as civil rights, the networks have been
ahead of the prevailing view.13
But this argument fails to explain certain
apparent major anomalies, situations in which the news media charged
ahead like cavalry, but the poor bloody foot soldiers failed to follow. For
example, media malaise theories need to account for how the endless
onslaught of 'scandal' coverage in the news frenzy that afflicted the
second term of the Clinton presidency failed to damage his long-term
public popularity, and indeed probably boosted it. If this prolonged saturation
coverage did not erode support for the president, then it seems
implausible to expect that more transient 'scandals' would have major
impacts on public opinion. As John Zaller suggests, we need to understand
the conditionality of media effects, both when coverage of scandal
matters for public opinion and when it does not.14
As Lance Bennett
concluded, after a lengthy critique of the time-series data presented by
Putnam, a circular process may be at work: 'The well documented political
uses and abuses of television are as much a response to, as primary
causes of, societal breakdown, individual isolation, and generalized discontent
with politics.'15
Given these potential problems, the best that can be said about trend
analysis is that it generates interesting hypotheses that deserve further
39
THE NEWS MEDIA AND CIVIC MALAISE
examination, but essentially the case remains unproven, more faith
than fact.
EXPERIMENTAL STUDIES
Another approach is through experiments that monitor the process of
short-term individual-level opinion changes in response to specific
media messages. Such experiments take the form of 'if X, then Y', and,
in principle, if people are randomly allocated to groups, and the analysis
compares differences between groups, the prior backgrounds, attitudes,
and values of subjects should not influence the results.16
The logic
of such experiments is disarmingly simple: If some are shown negative
news, for example news highlighting political scandals, government
waste, or policy failures, while others watch clips featuring positive
news, how do both groups react? This method potentially should
provide some of the most convincing and rigorous evidence, evidence
that might settle the media malaise debate.
Such experiments have long been used to examine the media malaise
perspective. In the mid-1970s, Michael Robinson showed 212 subjects
a single controversial documentary, The Selling of the Pentagon, and he
found differences in internal political efficacy after exposure to the
program.17
Cappella and Jamieson conducted perhaps the most thorough
experimental work on political cynicism. Their study argues that
strategic coverage of policy debates has come to predominate; winning
and losing become the central concerns; the language of wars, games,
and competition predominates; there is discussion of performers,
critics, and voters; there is much emphasis on the performances and
styles of candidates; and great weight is given to polls and their latest
rankings in evaluating candidates. Of course, there is little that is new
in all this; after all, elections are primarily about who wins and forms
the government, not simply a civics debate to educate the public. But
their study argues that over the years this framing has come to predominate
in campaign coverage.18
To test for the effects of such developments, Cappella and Jamieson
conducted experiments involving 350 subjects in six media markets.
One group was exposed to news in the print and broadcast media
framed strategically, where winning or losing was the predominant way
of characterizing the motivations of the candidates. Another group was
shown substantive news about health care framed in terms of issues,
where the stories concerned problems facing society and proposed solu-
40
UNDERSTANDING POLITICAL COMMUNICATIONS
tions. The study found that those who saw the strategic frame were
more likely to have cynical responses, meaning that they saw selfinterest
as the primary motivation of politicians:C
A story can be framed
in terms of the advantages and disadvantages for the candidate's chances
of election or in terms of the advantages and disadvantages for the
constituency. Mistrust of politicians and their campaigns arises when
strategy framing dominates.'19
The study concluded that American
network news was guilty of sensationalizing and oversimplifying
complex policy issues like health care, emphasizing the political game
over substantive debate, contributing towards a 'spiral of cynicism'
among the public.
Such experiments certainly come closer than many other methods to
nailing down causal effects in a rigorous manner, but they face the
common problem of how far one can generalize from experimental
results to the real world. Experiments may involve a large number of
participants who are allocated to stimulus and control groups wholly at
random. Yet the findings can be strongly influenced by the particular
methodology used, including the stimulus messages that are presented,
the means used to measure political attitudes like 'cynicism', and other
artifactual elements in the design. For example, in the Cappella and
Jamieson study, nonstandard measures of political cynicism limited
replicability with other research, and the operationalization of these
items may also have been subject to problems of circularity.
The problem of excessive coverage of strategy represents one dimension
of the media malaise case. Another important aspect concerns the
impact of 'negative' news, which can be regarded as critical or damaging
from the perspective of one particular actor. In an influential study,
Ansolabehere and Iyengar demonstrated that watching negative or
'attack' television advertising discouraged voter turnout and decreased
political efficacy in the United States.20
Yet parallel studies in Britain
came to a different conclusion. Experiments on the impact of negative
and positive television news, conducted in the 1997 British general election
campaign among 1,125 subjects, found that negative news failed
to damage, while positive news served to boost, levels of party support.21
As will be discussed further in the concluding chapter, that pattern
proved significant even after the use of a wide range of controls. The
contrasting findings from the experimental studies of the effects of negative
television news in Britain, and Ansolabehere and Iyengar's study
of negative TV ads in the United States, may be attributable to any of
three reasons: variations in the specific conditions under which they
41
THE NEWS MEDIA AND CIVIC MALAISE
were conducted (such as their operationalization and measurements of
negativity and party support); the repetitive 30-second TV ads and the
longer TV news stories may have influenced viewers in different ways;
people may have reacted differently in the U.S. and British media,
electoral, and political contexts. Experimental studies may be able to
provide precise findings that can address the issue of causality in media
effects, but it can be difficult to generalize from the necessarily artificial
conditions of an experiment to the real world.
SURVEYS: SPECIFIC THEORIES OF MEDIA MALAISE
Perhaps the most common approach has been to look for individuallevel
evidence from cross-sectional national surveys. Behavioral
research has focused on understanding the conditions of media exposure
believed to produce certain individual-level effects, including variations
in source, content, channel, receiver, and destination. Several
studies have compared the attitudes and behaviours of regular users of
different types of media, such as newspapers and television news, or
viewers of television debates and campaign ads.
This approach was exemplified by Michael Robinson, who used
American NES survey data from the sixties to show that those who
relied on television news had lower political efficacy, greater social distrust
and cynicism, and weaker party loyalties than those who relied on
newspapers, radio, and magazines for their political news.22
Experimental
data from 212 subjects were used to confirm the direction of
causality. For Robinson, the media malaise story runs as follows: In the
1950s and 1960s television news developed a mass audience, reaching
an 'inadvertent audience' who watched the news although they were
otherwise inattentive to political information. The inadvertent audience
is theorized to be particularly vulnerable to the messages in what they
watch and prone to believe in the credibility of the networks. American
television journalism is said to have certain characteristics, namely, a
tendency to present interpretive, negativist, and anti-institutional news.
The result is that viewers, particularly those of the inattentive audience
who lack other forms of political information, respond to such content
by growing more cynical, frustrated, and despairing about public affairs
and more disenchanted with social and political institutions. The main
evidence that Robinson presented, in addition to the experimental findings
already mentioned, were simple cross-tabulations of the 1968 NES
data on internal political efficacy scores, subdivided into those relying
42
UNDERSTANDING POLITICAL COMMUNICATIONS
solely on TV for information, those relying mainly on TV, and those
relying on some other news medium. Robinson concluded that those
who relied solely on TV had less confidence that they could have an
effect in the political system: 'Those who rely upon television in following
politics are more confused and more cynical than those who do
not. And those who rely totally upon television are the most confused
and cynical of all.'23
Robinson theorized that five factors are involved in the explanation
of those relationships: the size of the television news audience; public
perceptions of the credibility of the networks; the interpretive character
of television news; the emphasis on conflict and violence; and the
anti-institutional theme in network news. In later work, he suggested
that network television news was strongly influenced by the prestige
press, notably the New York Times, the Washington Post, and the Wall
StreetJournal, with greater focus on 'inside-the-beltway' strategic analysis,
rather than on the traditional coverage of politics by the regional
press and wire services.24
In the post-Watergate era, after the standard
NES indicators of trust in American politicians experienced free-fall,
the thesis that television news was responsible for civic malaise seemed
plausible to many. Others broadened the critique: An early study by
Miller and associates linked the content of newspapers, particularly critical
political coverage, with feelings of political disaffection experienced
by their readers.25
More recently, Robert Putnam analyzed American survey data and
reported that the heaviest users of television entertainment were least
socially trusting and least willing to join community groups.26
Putnam
related the dramatic transformations in our leisure patterns associated
with the rise of TV to broader trends in civic engagement: As television
began to saturate American homes in the 1950s, that produced a postcivic
generation. This could help to explain the new cohort patterns of
political mobilization and why generational cohorts raised in this new
cultural environment are less likely than their parents to trust others,
to join voluntary associations, and to vote. It has been shown that social
participation, such as belonging to clubs, attending church, or working
on community projects, can be strongly and consistently predicted by
TV use, and such participation is down among those who say that they
habitually depend upon television as their primary form of entertainment.
In short: 'More television-watching means virtually less of virtually
every form of civic participation and social involvement.'27
The
reasons for that pattern are not entirely clear, though Putnam suggests
43
THE NEWS MEDIA AND CIVIC MALAISE
that time spent on television may displace other recreational activities
and community involvement outside the home, and watching primetime
entertainment television may also foster passivity.28
Whatever the
reason, television entertainment (which does account for the vast bulk
of TV watching) although not TV news, is thereby indicted for the dramatic
erosion of civic engagement and social capital in America.
Analysts of cross-sectional surveys face four major challenges in
interpreting the available evidence. The most important is that crosssectional
surveys carried out at only one point in time make it difficult
to resolve the classic chicken-and-egg direction of causality. Is there a
selection bias? Does political interest cause us to turn on Meet the Press
or Nightline7
. Or is there a media effect? Does watching these programs
make us more politically interested? In the same way, does watching
television sitcoms and prime-time dramas produce less social trust and
less community involvement? Or do people who don't trust others and
are not engaged in their community simply prefer, as a matter of personal
choice, to stay home and watch TV? We cannot tell from crosssectional
survey data. The uses-and-gratifications approach argues that
we choose to watch programs that are most in tune with our prior predispositions
and tastes.29
In this view, our exposure to the new media
may tend to reinforce our political views (which is still an important
effect) rather than change our political attitudes.
Second, people often generalize about 'newspaper readers', 'television
viewers', or even 'Internet users' as though we all had a single experience
of these media. In practice, with the modern proliferation of
television channels, my TV experience (Nightline, C-SPAN, and ER)
probably is far removed from your TV experience (Monday Night Football,
MTV, and Oprah). Ideally, we need to compare the effects of variance
in the media messages so that we can see whether people who
consistently use one distinctive source (such as crime-focused local TV
news) differ from those who use others (such as right-wing talk radio).
Unfortunately, in practice it is often difficult to disentangle news sources
through survey research: Our measures of media habits are often diffuse
and imprecise. (How many hours per day do I usually watch the news?)
Often there is little variation in the content of mainstream sources like
television news on different channels, so we cannot easily contrast the
effects of watching ABC or NBC. We usually have multiple and overlapping
uses of different media. For example, tabloid readers often are
also fans of popular TV; broadsheet readers often listen to currentaffairs
programs; people who watch TV news often are newspaper
44
UNDERSTANDING POLITICAL COMMUNICATIONS
readers; and so on. One way to monitor media use is to ask the standard
question long employed in American polls: Where do you get most
of your news - from the newspapers or radio or television or magazines
or talking to people, or where? But that question is poorly designed, for
it is akin to asking electors what influenced their votes, rather than
analyzing this process indirectly. That question seeks a simple trade-off
answer, but given our multiple uses, most of us are unable to provide a
sensible answer. I get most of my news from Internet newspapers and
online TV bulletins, from National Public Radio (NPR) and the BBC
World Service, and, depending upon the topic, from occasional programs
like Nightline and Meet the Press. What reply should I give? The
alternative is to ask about habitual reliance on a series of different
sources, such as how many days per week one usually listens to the
radio, or watches the TV evening news, or reads a paper. Such selfreports
of media exposure are also unsatisfactory, because they take no
account of one's degree of attention, but they provide a more reliable
indicator than a simple trade-off question. Our case can be strengthened
if we can establish a significant and consistently positive relationship
between this weak measure of media use and the indicators of civic
engagement.
Moreover, there is no consensus in the literature, for other studies
based on survey analysis have challenged the media malaise claims.
Earlier studies strongly indicated that heavy use of television was associated
with certain indicators of political apathy, as Putnam suggested,
both in America and in other postindustrial societies.30
But that was not
a problem of television news per se: People who often watched TV news
and current-affairs programs were among those most involved in a wide
range of civic activities, such as voting, campaigning, and organizational
membership.31
Recently, Stephen Bennett has also challenged the theory
of the pernicious effects of American TV news, concluding that mediaexposure
measures are not significant predictors of trust in govern-
ment.32
In series of studies involving several countries it has been found
that regular viewers of television news and readers of broadsheet papers
have higher-than-average levels of political information, interest, and
engagement.33
Lastly, individual-level survey analysis is concerned with monitoring
the specific influence of media malaise on particular groups of news
media users. But that does not address the diffuse version of the media
malaise thesis. If the whole country has been affected by similar trends,
for example if American journalists are collectively overtaken by
45
THE NEWS MEDIA AND CIVIC MALAISE
Monica madness, then it becomes almost impossible to disentangle the
effects of different media sources on the public. Like the air we breathe,
if daily news about political scandals or government failures is all
around us, from the New York Times to the New York Post, from the
Drudge Report to Larry King Live, we cannot tell if the public is cynical
because of this endless diet from the media or whether journalists are
simply feeding the voracious public appetite for such headline news, or
both. Only stringent comparative designs for studies across countries
can allow us to explore cross-cultural differences.
COMPARATIVE STUDIES
This leads to the conclusion that the impact of the news media ideally
needs to be studied using a triangulated research design within a crossnational
setting. As Blumler, McLeod, and Rosengren argue, comparative
research can allow us to overcome national and time-bound
limitations on the generalizability of our theories, assumptions, and
propositions.34
At present, the bulk of the existing research has been
conducted within the United States, and it remains unclear to what
extent the patterns found in these studies are evident in other coun-
tries.35
As discussed in Chapter 13, many features of the news environment
in the United States may be products of 'American
exceptionalism'. Despite the formidable problems facing comparative
research, and the serious limitations of data, such a strategy is worthwhile
because it can begin to counteract both 'naive universalism'
(assuming everywhere is the same as us) and 'unwitting parochialism'
(assuming everywhere is different to us).36
COMPARING POSTINDUSTRIAL SOCIETIES
The comparative framework adopted for this book focuses on postindustrial
societies, defined as the twenty-nine member states of the
Organization for Economic Cooperation and Development. This comparison
includes most of the major developed economies and established
democracies in the world, including all G7 and European Union
(EU) states. The advantage of this design is that it allows us to compare
a wide range of advanced industrialized societies and democratic
states that are reasonably similar in terms of their levels of economic,
social, and political development. This follows the classic logic of the
'most-similar-system' design that assumes that the factors common to
46
UNDERSTANDING POLITICAL COMMUNICATIONS
relatively homogeneous societies are irrelevant to explaining their
differences.37
The common levels of literacy, education, and affluence in
postindustrial societies mean that we can discount these factors in
searching for explanations for civic participation. At the same time,
there remain significant contrasts in the news environments, in the
political systems, and in the dependent variables concerning levels of
political knowledge, interest, and civic engagement among citizens in
these states. At the broadest level, we are seeking to move from an analysis
of nations towards an analysis of types of political communication
systems. The analysis of newspaper-centric and 'television-centric'
media systems presented in Chapter 4 is one example of this approach.38
The basic economic and social indicators for the countries in this
comparison are summarized in Table 3.1. The OECD includes more
than a billion people in large and small states, ranging from the United
States, Japan, Mexico, and Germany at one end of the population spectrum
down to Luxembourg and Iceland at the other. Many of the most
affluent societies in the world, characterized by a GDP per capita of over
$30,000, are members of the OECD, such as Switzerland, Japan, the
United States, and the Scandinavian states, although at the lower level
of economic development the OECD has countries with GDP per capita
below $10,000, including member states in southern, central, and
eastern Europe, as well as Mexico.39
All these post-industrial economies
are overwhelmingly based on the service sector, which accounts for twothirds
of civilian employment and roughly the same proportion of contribution
to GDP. Just over one-quarter of jobs in the OECD states
remain in manufacturing industries, and less than one-tenth are in agriculture.
The only countries with more than one-fifth of the work force
in agriculture are Greece, Mexico, Poland, and Turkey. The size of the
public sector varies substantially between countries, whether measured
by government expenditure as a percentage of GDP or by the size of
public sector employment. The largest public sectors are found in the
countries of Scandinavia and northern Europe, especially in small
welfare states with a strong social-democratic tradition, such as Sweden
and The Netherlands. In contrast, the levels of public sector spending
are far lower in Japan and South Korea. Lastly, the indicators show
that OECD societies are among the most literate and best educated in
the world, with, on average, over one-fifth of their working-age populations
attaining some higher education. Thus comparisons among
OECD member states should allow us to detect any significant differences
in their news environments, for example between countries with
47
Table 3.1. Social and economic indicators, OECD countries, mid-1990s
Country
Australia
Austria
Belgium
Canada
Czech Rep
Denmark
Finland
France
Germany
Greece
Hungary
Iceland
Ireland
Italy
Japan
Korea, S.
Luxembourg
Mexico
Netherlands
NZ
Norway
Poland
Portugal
Spain
Sweden
Switzerland
Turkey
UK
US
Area
(square
miles)
7687
84
31
9976
79
43
338
549
357
132
93
103
70
301
378
98
3
1973
41
269
324
313
92
505
450
41
781
245
9373
Pop.
(lOOO's)
1996
18,289
8,060
10,127
29,964
10,316
5,262
5,125
58,380
81,877
10,465
10,195
270
3,621
57,473
125,864
45,545
418
96,582
15,494
3,640
4,370
38,618
9,935
39,270
8,901
7,085
62,695
58,782
265,557
GDP per
capita
1996 ($)
21,375
28,384
25,409
19,330
5,445
33,230
24,420
26,323
28,738
11,684
27,076
19,525
21,127
36,509
10,644
40,791
3,411
25,511
18,093
36,020
10,425
14,894
28,283
41,411
2,894
19,621
27,821
Service Sector
Contribution
to GDP
% Services
1996
69.5
67.9
70.2
72.1
58.4
72.1
64.9
71.7
68.4
67.9
68.5
54.7
65.5
60.0
50.9
74.9
69.5
69.8
66.6
65.5
62.9
64.8
70.5
63.5
52.5
70.8
71.9
Size of Public Sector
General
Government
Expenditure
% of GDP
Mid-1990s
35.6
48.6
51.7
45.8
40.5
59.6
55.9
51.6
46.6
52.1
35.1
36.9
49.5
28.5
15.7
45.0
50.0
45.8
42.5
41.2
63.8
47.7
42.3
34.3
Government
Employment %
of Total
Employment
1996
16.0
22.8
19.0
19.6
30.7
25.2
24.9
15.4
19.9
13.4
16.1
6.0
12.0
11.9
22.1
30.8
15.3
15.7
31.2
14.0
14.1
13.4
Educational Indicators
% Pop. with at
Least UpperSecondary
Educ
(25-64-year-
olds)
1996
52.8
69.5
53.5
75.2
83.4
62.0
65.4
68.4
83.7
42.5
47.2
34.9
59.8
29.3
61.2
59.1
81.2
73.7
20.1
28.0
74.7
82.2
23.0
75.9
85.8
% Pop. with at
Least Higher
Educ (25-64-
year-olds)
1996
24.3
7.9
24.6
46.9
20.4
20.5
18.6
22.6
17.4
19.9
25.3
28.6
13.1
11.0
16.1
28.3
21.1
21.5
33.3
Source. OECD.
THE NEWS MEDIA AND CIVIC MALAISE
predominantly commercial or public sector television, or between those
with high and low levels of newspaper circulation, controlling for reasonably
common levels of social and economic development.
The basic features of the political systems are listed in Table 3.2. The
OECD contains most of the world's major established democracies, as
well as three newer democracies that have joined the organization more
recently: Hungary, Poland, and the Czech Republic. The only exceptions
to this generalization are Mexico and Turkey, which can best be classified
as semi-democracies. In Mexico, the Partido Revolucionario Institucional
(PRI) has held power at the federal level since 1929, although
under increasing electoral challenge in recent years, and it has certain
authoritarian characteristics. Turkey currently lacks important political
rights and civil liberties and has had a mixed and unstable record of
democratic development.40
The Gastil index, provided by Freedom
House every year since 1973, has monitored worldwide levels of political
rights and civil liberties on two scales, ranging from 1 (most free)
to 7 (least free). By the mid-1990s, most of the nations in our comparison
fell into the 'free' column, with the exceptions of Mexico and
Turkey, which were classified as 'partly free'.
As shown in Table 3.2, the countries compared in this book, though
all democracies, feature a wide range of different types of political institutions.
The electoral system is one of the most important aspects of
any constitution, and the OECD countries range from proportional
representation using national or regional party lists, as in The Netherlands
and Sweden, through mixed systems, such as those in Germany
and Italy, to plurality and majoritarian systems, like the first-past-thepost
systems in the United Kingdom and the United States. Their party
systems also vary substantially, and these are classified on the effective
number of parliamentary parties elected to the lower house (ENPP) in
the early and middle 1990s. This allows us to distinguish among predominantly
one-party systems (characteristic of Mexico and Japan),
two-party or two-and-a-half-party systems (found in Australia and
the United States), the moderate multiparty pluralism (with between
2.5 and 4.5 ENPP) common in many European systems, and fragmented
multiparty pluralism (with ENPP greater than 4.6).41
The
remaining columns in Table 3.2 indicate the opportunities for electoral
participation within each system, including popular contests for the
upper house in bicameral legislatures, direct votes for the presidency,
and the frequency of national referendums. The comparison of OECD
nations also reveals important differences in the news environments
50
Table 3.2. Political systems, OECD countries, mid-1990s
Country
Australia
Austria
Belgium
Canada
Czech Rep
Denmark
Finland
France
Germany
Greece
Hungary
Iceland
Ireland
Italy
Japan
Korea, S.
Luxembourg
Mexico
Netherlands
NZ
Norway
Poland
Portugal
Spain
Sweden
Switzerland
Turkey
UK
US
Electoral
System
for Lower
House
1996
AV
PR
PR
Plurality
PR
PR
PR
Majority-
runoff
Mixed
PR
Mixed
PR
STV
Mixed
Mixed
Plurality
PR
Mixed
PR
Mixed
PR
PR
PR
PR
PR
PR
PR
Plurality
Plurality
Number
of
Members
Lower
House
1996
148
183
150
295
200
179
200
577
656
300
386
166
630
500
299
500
150
120
165
460
230
350
349
200
550
659
435
Number of
Effective
Parliamentary
Parties
1991-95
2.42
3.40
7.95
2.35
4.85
4.70
4.88
2.96
2.78
2.17
2.89
3.48
7.45
3.95
2.70
2.28
5.38
2.16
4.15
3.85
2.55
2.67
3.51
5.60
4.40
2.26
2.00
Type of
Party
System
1991-95
Two party
Moderate
pluralism
Fragmented
pluralism
Two party
Fragmented
pluralism
Fragmented
pluralism
Fragmented
pluralism
Moderate
pluralism
Moderate
pluralism
Two party
Moderate
pluralism
Moderate
pluralism
Fragmented
pluralism
Moderate
pluralism
Moderate
i I '
One party
predominant
Fragmented
pluralism
Two party
Moderate
pluralism
Moderate
pluralism
Moderate
pluralism
Moderate
pluralism
Moderate
pluralism
Polarized
pluralism
Moderate
pluralism
Two party
Two party
Popular
Election
for Upper
House
1996
Yes
No
Indirect
No
Yes
Unicameral
Unicameral
Indirect
No
Unicameral
Unicameral
Mixed
Yes
Yes
Unicameral
Yes
No
Unicameral
Unicameral
Yes
Unicameral
Yes
Unicameral
Yes
Unicameral
No
Yes
Popular
Election
for
President
1996
No
Majority-
runoff
No
No
No
No
Majority-
runoff
Majority-
runoff
No
No
No
AV
No
No
Plurality
Plurality
No
No
No
Majority-
runoff
Majority-
runoff
No
No
No
No
No
Elec.
college/
plurality
National
Referendums
N.1945-95
23
1
1
1
13
1
12
0
4
5
20
29
6
0
10
1
5
4
3
275
4
1
0
Political
Rights
Index
1997
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2
1
4
1
1
1
1
1
1
1
1
4
1
1
Civil
Liberties
Index
1997
1
1
2
1
2
1
1
2
2
3
2
1
1
2
2
2
1
3
1
1
1
2
1
2
1
1
5
2
1
Type of
Democracy
1997
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Free
Partly
free
Free
Free
Free
Free
Free
Free
Free
Free
Partly
free
Free
Free
Sources: Political rights, civil liberties, and type of democracy: Freedom Review, 'Index of Freedom', January 1998,28(1). Electoral system: Lawrence LeDuc, Richard
Niemi, and Pippa Norris. 1996. Comparing Democracies. London: Sage. Plurality systems, first past the post; AV, alternative vote; STV, single transferable vote; PR,
party list; mixed, combination of plurality and party list systems. Number of effective parliamentary parties defined as those with at least 3% of seats in the lower
house in the most recent election. Type of party system based on ENPP in the latest election available: 0-2.5, two party; 2.6-4.5, moderate pluralism; 4.6+, fragmented
pluralism.
UNDERSTANDING POLITICAL COMMUNICATIONS
within this wide range of advanced postindustrialized economies and
democratic states.
CONTENT ANALYSIS OF EU NEWSPAPERS
AND TELEVISION
For aggregate trends, UNESCO is the most authoritative source for official
statistics worldwide, such as data on the numbers of television sets
and the circulation of newspapers. When we turn to content analysis of
the news media, however, we focus on the available data from the fifteen
member states of the European Union. The content analysis used in this
book is derived from Monitoring Euromedia, a monthly report published
by the European Commission from January 1995 to September
1997.42
The company that carried out the research, Report International,
used quantitative and qualitative methods to study coverage of
the EU in newspapers in all 15 member states, and television in six
member states, providing the most comprehensive cross-national
content-analysis data set that is currently available. Monitoring Euromedia
examined the contents of 189 newspapers every month, including
all the national papers and the most important regional papers in
all member states. The detailed list of sources is provided in the Technical
Appendix at the end of the book. The survey included heavyweight
broadsheets like Le Monde, the Financial Times, and the Frankfurter
Allgemeine Zeitung, as well as large-circulation tabloids like Der Bild
and the Sun. The weekly magazine press was also included, such as the
Economist and L'Express, as well as dailies and Sunday papers. The
average monthly report identified some 11,000-12,000 articles selected
as containing information on the EU and its policies. The study then
selected a random sample of 50% of the articles to be analyzed every
month, or some 5,000-6,000 articles. Over the whole thirty-threemonth
period the study therefore analyzed the contents of just under
200,000 articles.
Each article was coded according to the source, country, date, and
type of information contained (facts, opinions, or comment).
Articles were also assigned two or three different 'topic' codes,
because most covered more than one subject. These categorized topics
such as foreign policy, monetary policy, EU institutions, and enlargement
of the EU.
A selection of stories was also coded on whether the topic was evaluated
positively or negatively. This can be termed the 'directional' code
53
THE NEWS MEDIA AND CIVIC MALAISE
or tendency, which was scaled from 1 (very negative) through 2 (slightly
negative), 3 (neutral), and 4 (slightly positive) to 5 (very positive). When
the positive and negative evaluations balanced, stories were given a
neutral code. Supervisors checked for inter-coder reliability and consistency
of coding practices.
Most research on news balance has concentrated on the extent to
which election news has been evenhanded in terms of partisanship or
ideology, such as in the amount of coverage of different candidates or
issues.43
But elections are special cases: Practices in broadcasting often
are strictly regulated by explicit fair-treatment regulations, as in allocating
equal time to all sides in leadership debates and equal airtime for
party broadcasts; in contrast, newspaper partisanship often increases
during campaigns. While it is particularly important that campaign
coverage be balanced, it is difficult to generalize from patterns found in
this context to the daily editorial practices in newsrooms. Other comparative
research has focused on how a particular dramatic event was
reported in different countries, such as a positive or negative frame
when reporting the Persian Gulf War or the downing of the Korean airliner
over Russia. Only a couple of studies have attempted to compare
typical daily news coverage across different countries.44
This analysis of
routine coverage of the EU over a thirty-three-month period provides
a unique look at how the concept of directional balance operates in
newspapers and television outside of election campaigns. By comparing
the amount and balance of EU coverage in different member states
over time, we can analyze whether or not the news media have provided
an effective civic forum, as discussed in Chapter 2, encouraging public
debate about the EU.
The monthly content analysis for television was based on daily
news and current-affairs programs in six countries (Belgium, France,
Germany, Italy, Spain, and the United Kingdom). The study recorded
the main news programs in each of those countries, with details as
given in the Technical Appendix, analyzing 500-600 programs per
month. During the course of the thirty-three-month period the study
therefore analyzed some 16,000-20,000 programs in total. Monitoring
Euromedia included the main news bulletins in each country, such as
the 6 O'clock News and 9 O'clock News on BBC1 and Newsnight on
BBC2 in Britain, the 6:30 A.M., 8:00 P.M., and midnight Telegiornale on
RAH in Italy, and the Desayunos de RN and Telediario on TVE in Spain.
The study examined the extent to which news stories in these programs
contained information on the European Union and its policies, and
54
UNDERSTANDING POLITICAL COMMUNICATIONS
around 300 EU-related stories were coded every month, following the
same process used for newspaper articles. Monitoring Euromedia also
compared coverage of EU special events (notably the Turin InterGovernmental
Conference, IGC, and the Florence and Dublin councils)
in all member states by 73 public and commercial television stations.
These sources therefore allow us to compare the topics and directions
of coverage of the EU by a wide range of newspaper and television
outlets for each month. We can monitor how news coverage changed
during these years in response to political developments, such as the
1996 Turin IGC, the process of moving towards Economic and Monetary
Union, and events like the British beef crisis caused by 'mad-cow'
disease. Using this data set, we can examine the amount and tone of the
news about the EU, making comparisons between countries and over
time. If coverage of EU policies and institutions became more negative
in some member states than in others, for example with the European
ban on British beef generating splash Euroskeptic headlines in the
British tabloids, while generating more popular support for the EU in
the French press, we can assess whether or not that led to changes in
public opinion about the EU that differed across member states.
Inevitably, having to rely on secondary data limits our ability to
examine certain important questions raised by theories of media
malaise. Most importantly, we lack any direct evidence whether or not
there has been a long-term change in the news culture since the 1970s,
as some suggest, for example whether there has been an increase in negativity,
or more frequent disdainful commentary by reporters, or a shift
from a substantive to a strategic frame in news stories. Nor can we use
this data set to examine many of the subtle nuances of news coverage
of the EU, such as the extent to which the media present national political
leaders, EU officials, and members of the European Parliament
(MEPs) speaking at length, in context, rather than having journalists
provide their interpretations of events, or the extent to which national
frames, as one would expect, dominate stories about Europe. Such features
of news stories may well play key roles in shaping the content of
news coverage about Europe.
What we can do in this study, however, is examine certain long-term
effects. If we assume that this typical content of news has changed over
time and that this has contributed to public disenchantment with the
political process, as media malaise theories hypothesize, then we should
see changes in public opinion as monitored by the long-term series of
surveys we examine. This study analyzes the American public using the
55
THE NEWS MEDIA AND CIVIC MALAISE
NES series from 1948 to 1998, and European public opinion using the
Eurobarometer series from 1970 to 1999. Media malaise theories
suggest two alternative hypotheses that we shall discuss in detail and
test in subsequent chapters. One possibility is that if changes in news
coverage have increasingly turned off all the public or a major segment
of the public, then that should be evident in a shrinkage in the size of
the news audience. In that scenario, many people disgusted with negative
journalism would be expected to turn to other channels or to cease
buying newspapers. Another possibility is that people may have continued
to watch TV news and read newspapers, perhaps because of
habitual leisure patterns, but over time those who have paid the most
attention to the news have become increasingly cynical and disenchanted
with government institutions and political leaders. If we
assume that news did become more negative in the United States in the
1960s and early 1970s, as many assert, and that this fuelled public disenchantment
with politics, then we should see a changed relationship
between attention to the American news media and a range of indicators
of civic malaise.
A second potential problem is that coverage of European affairs,
involving distant, complex, and low-salience issues, may differ in certain
important respects from news about domestic politics. The latter may
well provide more coverage of the drama, personalities, and salient
issues more relevant and immediate to the lives of citizens. This is true,
but in principle it can be argued that these conditions should maximize
the potential impact of the news media's coverage. For domestic politics,
the public has multiple sources of information. For example,
people can evaluate the economic performance of the government on
the basis of their own pocketbooks, the economic conditions of their
friends, colleagues, and neighbours, and news reports of the trade gap,
the rate of inflation, and the growth of jobs. Given all these sources,
people may choose to discount some of the information provided by
the news media. In contrast, in regard to European affairs, though some
EU policies may have direct and visible impacts, most are conducted at
such an abstract and technical level that citizens have to rely almost
wholly on the news media for their information, along with cues from
opinion-leaders and personal discussions. In this regard, coverage of the
EU can be regarded as comparable to how public opinion is shaped
towards other foreign-policy issues, such as conflict in Kosovo, trade
with China, or the Gulf War. Plausibly, if we find few systematic effects
56
UNDERSTANDING POLITICAL COMMUNICATIONS
on public opinion from coverage of the EU, then we might expect to
find even weaker effects from news coverage of domestic politics.
Lastly, the content analysis provides no direct evidence regarding
coverage in the 'new' news, meaning television magazines, live discussion
programs, and talk radio, let alone the flourishing sources of news
on the Internet. This is a valid criticism, up to a point. The content
analysis we use draws heavily on the mainstream evening news programs
on television. But it does also include leading current-affairs
magazines like the BBC's Sunday Breakfast with Frost and BBC2's Newsnight,
as well as German ARD's Europa-magazin and Pressedub and
France 2's Sunday Revue de Presse. These can be seen as roughly equivalent
to American television magazine programs like Meet the Press,
Nightline, 20/20, and Dateline. If the 'new' news has infected traditional
standards of mainstream journalism, as some suggest, then this should
be picked up by our analysis. In addition, it remains unclear whether
there is a distinct 'new' news sector in Europe. In Britain, for example,
one of the oldest BBC radio programs, Any Questions, now forty
years old, and the direct descendant of television's Question Time, has
always involved live discussion of public questions and debate between
political leaders. News magazines, in different formats, have been
popular since the 1960s. Certainly there are some equivalents to the
American 'new' news in Europe, such as the Spanish tertulias, twentyfour-hour
radio talk channels, and Internet magazines, but their audience
currently remains limited. The content analysis of television and
newspapers used in this study, while less than ideal in terms of longterm
time-series data, and while limited to the 15 European OECD
countries, therefore does provide a suitable basis for a comparative
study of typical news coverage of European affairs across the EU.
PUBLIC OPINION
Content analysis, no matter how comprehensive, remains silent about
the effects of coverage. To understand the impact of attention to the
news media on the public, we draw on two decades of Eurobarometer
surveys ranging from the first European Community Study in 1970 to
the most recently available survey in March-April 1999. Surveys were
conducted two to five times per year, with about 1,000 face-to-face
interviews in each member state, with reports published on a biannual
basis for the European Commission. These studies have been supple-
57
THE NEWS MEDIA AND CIVIC MALAISE
mented since January 1996 by Europinion surveys (European
continuous-tracking surveys) that have sought to monitor public
opinion about key issues and institutions via telephone interviews each
week, with results released on a monthly basis. These rich data sets allow
us to monitor whether people who are most attentive to newspapers,
television and the Internet differ in any significant ways in their political
attitudes, opinions, and behaviours towards the European Union,
in terms of its institutions and its policies in the fifteen member states.
When there are key events - such as the Maastricht agreement, the
introduction of the euro, and the resignation of the Santer European
Commission - these sources allow us to compare coverage in the news
media with public opinion. Because EU policies often involve fairly
complex and technical issues with which ordinary people have had no
direct experience, this provides a strong test of the learning effects of
the news media. It is difficult for European citizens to know much about
these issues except via the news media, so if journalism fails in its informational
role, then that may have important implications for European
Union governance, raising widespread concern about a 'democratic
deficit'. Within this context we can explore the role of the news media
system as a mobilizing agent and the effects of media use on political
knowledge, interest, and activism in different European member states.
One potential criticism of using European data is the 'American
exceptionalism' argument. Much of the media malaise literature originated
in the United States, and many of the claims about changes in
news journalism may relate to specific historical and cultural factors
peculiar to America. As we shall see, the news environment in the
United States is more television-centric (and with far more commercially
oriented TV) than those in most European countries. The predominant
liberal political culture in America may also be more
mistrustful of government than is the more social-democratic tradition
in the smaller European welfare states. To test whether or not patterns
found in European public opinion were also evident in the United
States, we drew on half a century of data from the American National
Election Studies (NES) from 1948 to 1998. Obviously there are some
important differences between the NES and the Eurobarometer, including
the specific questions that monitor media use and civic engagement,
as well as the electoral context of the NES. Nevertheless, by matching
functionally equivalent, if not identical, measures, we can examine the
impact of attention to the American news media on similar indicators
of political knowledge, trust, and participation. As mentioned earlier,
58
UNDERSTANDING POLITICAL COMMUNICATIONS
the NES also has the advantage of allowing us to monitor trends from
1952, before the television age became established, until 1998, representing
the early years of the Internet age.
The next chapter goes on to use official statistics from UNESCO and
other international bodies to examine some of the most important
structural trends in the news environment in postindustrial societies
since the end of World War II,45
including the following: press diversity,
such as the number of national daily newspapers being published,
changes in circulation and sales figures, and concentration of ownership;
television diversity, including the structure of competition between
public-service and commercial channels, regulation of broadcasting,
and the availability and penetration of cable, satellite, and other new
communications technologies; and Internet usage, a development that
has proceeded far faster in some countries than in others. Given an
understanding of those matters, we can then start to consider what
impact these differences might have on the content of the news and its
potential effects on the public.
59