Crime news and its antecedents. A comparative analysis of crime

coverage on TV in 11 countries

Stefaan Walgrave (stefaan.walgrave@ua.ac.be)

Danielle Sadicaris (danielle.sadicaris@ua.ac.be)

Paper prepared for the ECPR Joint Workshop Sessions, 14-18 April 2009, Lisbon

Portugal

ABSTRACT

The paper studies crime coverage in TV news in a comparative perspective. Our aim

is to tap the amount and type of crime coverage across countries and TV stations

and to explore the factors leading to differences in the level and type of crime

coverage. Based on a sample of 11 countries and 24 TV stations—we code the main

TV news during four weeks for each channel—we analyze and compare (1) the

amount of crime coverage and the types of crime that get exposure, (2) the framing

of the crime stories by assessing which actors are quoted, and (3) the sensationalist

character of the formats used to report about crime. We find that real crime rates

matter for the amount of crime coverage and that private channels more than

public channels use sensational formats to tell crime stories in the news. The media-

politics system matters for the framing and format of crime coverage; stations in

Polarized Pluralist systems are interpreting crime less structurally and bring their

crime stories in a more sensationalist way. The best predictor of crime coverage is

market competition. Fragmented media markets lead to an increase in crime

coverage, to more human interest and less structural framing and to more

sensationalist crime coverage.

1

INTRODUCTION

For students of political communication studying crime coverage in the media

is relevant for at least two reasons. First, probably more than other kinds of coverage,

crime coverage has considerable political consequences. On the individual level,

previous studies showed that crime coverage (depending on the type and framing of

coverage) has strong a effect on the public agenda—that are the topics people care

about politically (Lowry, Nio, and Leitner 2003; Soroka 2002). Crime coverage

strongly affects how people think about the real-world occurrence of crime. The more

the media report about crime events the more the audience thinks that crime rates

are high in their society (Davis 1952). People depend on the media for assessing the

crime level in their environment (Lowry, Nio, and Leitner 2003). Also, research has

substantiated that there is a strong association between crime coverage and feelings

of insecurity. Of course, crime news and feelings of insecurity mutually reinforce each

other—people who feel unsafe ‘demand’ more crime news on TV—but there is enough

empirical evidence that crime coverage as such has an effect on feelings of insecurity

and fear (Gordon and Heath 1991; O'Keefe and Reid-Nash 1987; Romer, Aday, and

Hall Jamieson 2003). These feelings of insecurity contribute to feelings of general

malaise and to social (in people) and political (in political institutions) distrust

(Elchardus and Smits 2002; Gerbner, Gross, Morgan, and Singorielli 1994). These

feelings, in turn, lead to changing voting behavior in favour of extremist and populist

parties.

On the aggregate level, studies showed that rising amounts of crime coverage

elicit political reactions and subsequent policy changes in crime policy. More than

coverage about social affairs or the environment, for example, crime coverage affects

what politicians consider to be a priority. If crime coverage goes up, attention for

crime in parliament and government goes up too (Walgrave, Soroka, and Nuytemans

2008). So, crime coverage fosters, probably more than other types of coverage,

political attention and policy making. Also on an aggregate level, research has found

that there is an association between the level of crime coverage and the electoral

success of right-wing populist parties like the Belgian Vlaams Belang (Walgrave and

Deswert 2004).

A second reason for the relevance of studying crime coverage from a political

communication perspective, is that crime coverage can be considered as a useful

indicator of the ‘quality’ of TV news and the informative value TV news has for

citizens. Many scholars have been worried about the declining quality of the news in

2

general and TV news in particular (see for example: Patterson 1993). Tabloidization

and sensationalization are processes that have been described and documented over

time (for an overview see Norris 2000). Especially political news and foreign news

are said to be the victims of the gradual shift towards soft and sensational news. In

these accounts, crime coverage has often served as a prime indicator: more crime

coverage indicates less quality (Ericson, Baranek, and Chan 1991; Hauttekeete 2004).

So, evolutions in the amount and type of crime coverage are strongly associated with

other aspects of TV coverage and can serve as an indicator of declining or changing

attention to political news, withering foreign news attention, and the tendency

towards soft news.

The study engages in studying crime coverage in TV news in a comparative

perspective. Our aim is to tap the amount and type of crime coverage across

countries and TV stations and to explore the factors leading to differences in the

amount and type of crime coverage. Based on a sample of 11 countries and 24 TV

stations—we coded the main TV news during four weeks for each channel—we

analyze and compare (1) the amount of crime coverage and the types of crime that get

exposure, (2), the framing of the crime stories by assessing which actors are quoted,

and (3) the sensationalist character of the formats used to report about crime. The

amount, sensationalism, and framing of crime stories form the dependent variables

of the study. The paper not only describes the crime coverage situation in 11 countries

but also explores why crime is covered different, or not, in these countries. We expect

to find differences between TV stations depending on the real crime rates, the media-

politics system, the type of station (public-private), and the media market situation.

CRIME IN THE NEWS: HYPOTHESES

Communication scholars have spent some time investigating crime coverage

on TV. Most of this work has dealt with the effects of crime coverage on the public or

on political elites (see above). Far fewer empirical studies have explored where crime

coverage comes from and why some channels devote more, and different, attention to

crime than others. This study focuses on four possible explanations.

A first and obvious explanation holds that crime coverage rates are simply

being driven by real crime rates. As crime goes up or down coverage follows suit and

goes up or down too. This base-line hypothesis stating that the real world drives the

news is not entirely in line with extant literature on crime coverage. Many studies

3

showed that crime coverage is not, or hardly, connected to real world crime levels.

Some studies even found, over a longer period of time, that coverage and reality were

negatively correlated (Antunes and Hurley 1977; Ghanem 1996; Windhauser, Seiter,

and Winfree 1991). Most studies find no correlation whatsoever (Graber 1980; Sheley

and Ashkins 1981; Walgrave and Uce 2007). When focusing on more specific types of

crime and their coverage, scholars came to more mixed conclusions. Sheley en

Ashkins (1981) state that in the US especially the coverage of murders is completely

unrelated to the real number of murders. Graber arrives at similar conclusions for

what she defines as street crime (Graber 1979). Marsh finds a similar

overrepresentation of news about violent offences and an underrepresentation of

property crime (Marsh 1991). Sherizen concluded that the relation between crime

rates and crime news is generally negative but that it was positive when it related to

murder. The less a kind of offense occurred, the more coverage there was but with

murders the relation was positive (Sherizen 1978). Still, all these studies examined

crime rates and coverage through time; none of them, except for the study by Marsch,

focused on comparing crime coverage between countries. It may be the case that

crime coverage in a country or region does not follow the real crime evolution but

that the share of crime news in a country does reflect the overall level of crime in that

country compared to other countries. All previous studies are longitudinal, the

present study is comparative. Thus, we hypothesize that real crime rates matter and

that countries with much crime would have more crime in the news. We have no clue

whether the real crime rate has an effect on our two other dependent variables: the

sensational format of crime coverage and the framing of crime in the news. So, we

only formulate one hypothesis: There is a positive relation between the real crime

rate and the amount of crime coverage (H1).

A second series of hypotheses is related to the distinction between public

service and private business broadcasting. Public service and private channels have

essentially different goals (Holz-Bacha and Norris 2001). Crime coverage is

essentially commercial coverage fitting into the tabloidization of news (Esser 1999).

Crime often has a high news value and attracts a wide audience (Lotz 1991).

Moreover, crime news is ‘easy’ news. It is often very accessible, there are a lot of

sources, it is easy to narrate, there are clearly defined good (victims) and bad guys

(offenders) involved, the subjects of crime news are weak and cannot challenge wrong

or exaggerated news coverage… (Lowry, Nio, and Leitner 2003). Hence, the economic

and journalistic cost of crime news is low, while its impact in terms of audience may

be high. This makes crime news attractive for a commercially driven media outlet.

4

Market-driven journalism resorts more to sensationalism to reach the audience

(Hendriks Vettehen, Nuijiten, and Beentjes 2006). Many students of broadcasting

have examined, or speculated about, the differences in news between commercial and

public broadcasters (for an overview see Norris 2000). Some studies found that

private channels aired more soft and sensational news than public broadcasters

(Canninga 1994; Hvitfelt 1994; Norris 2000: 110-111; Powers, Kristkjansdottir, and

Sutton 1994). Therefore, we expect that also in terms of crime news we will find more

of it, and in different formats and framing, on private newscasts than on public

channels. Our three hypotheses are: Private channels cover more crime than public

broadcasters (H2a); Private channels use the human interest frames more often

than public TV stations and they use the structural and institutional frames less

often (H2b); Private channels rely more than public broadcasters on sensational

formats when bringing crime news (H2c).

Students of comparative communication finally got their ‘bible’ when Hallin

and Mancini (2004) published their Comparing Media Systems in 2004. The book is

widely cited by almost any study engaging in comparing media and political systems.

The authors’ main claim is that the relation between media and politics differs

systematically across countries leading to different media systems each with their

specific features and dynamics. The typology put forward by Hallin & Mancini is

situated at the system-level. They distinguish three types of systems: the Liberal, the

Democratic Corporatist and the Polarized Pluralist type. The book does not produce

predictions about the specific news content we may expect as a consequence of the

typology. Yet, if Hallin & Mancini’s seminal work really wants to be useful for

comparative studies it needs to be translated in testable hypotheses about media

content. What is the interest of devising a typology of media systems when that

typology is unrelated to and does not make a difference for what media actually do?

Therefore, bridging system and content, we put forward some hypotheses on how

media systems would lead to differences in crime coverage. What dimensions of the

typology are relevant for crime news? Especially the professionalization of journalism

may bear consequences for crime news. Professionalization refers to the fact that

journalists are autonomous (also from commercial pressures), that they have their

own distinct professional norms and that they have a public service orientation

(Hallin and Mancini 2004: 33-41). Both the Liberal model and the Democratic

Corporatist model are characterized by a strong professionalization, the first in a non-

institutionalized and the latter in a institutionalized way. The Polarized Pluralist

model, in contrast, is less professionalized and journalism is more ‘instrumentalized’

5

for political and commercial goals. We anticipate that strong professionalization leads

to less, less sensational and more structurally framed news. Hence three hypotheses:

There is more crime coverage in Polarized Pluralist countries than in Liberal or

Democratic Corporatist countries (H3a); Crime coverage in Polarized Pluralist

countries relies less on structural and institutional frames than in Liberal or

Democratic Corporatist countries (H3b); Crime coverage features more sensational

formats in Polarized Pluralist countries than in Liberal or Democratic Corporatist

countries (H3c).

Finally, we expect that the media market competition affects the amount,

format and frame of crime coverage. In some countries, the competition among the

main TV stations is mitigated as there are only a handful of stations amongst which,

most of the time, a strong public broadcaster (Picard 2002). In other countries, the

media market is strongly fragmented with fierce competition between many players.

As crime has the capacity to attract a large audience, we expect that crime coverage

would be affected by the market situation. That media market competition can affect

the content of news has been substantiated (Atwater 1984; Powers, Kristkjansdottir,

and Sutton 1994). Competition forces media outlets to cater the prejudices of their

readers; the more competition the more aggressive catering to such prejudices there

will be (Mullainathan and Shleifer 2005). Studies have pointed out that increasing

competition among news providers leads to more sensational and human interest

news (Slattery, Doremus, and Marcus 2001). We coin three final hypotheses: Strong

competition on the media market leads to a larger amount of crime coverage (H4a);

Strong competition on the media market leads to a less structural and institutional

framing and a more human interest framing of crime news (H4b); Strong

competition on the media market leads to a more frequent use of sensational

formats (H4c).

DATA AND METHODS

We rely on a comparative dataset containing evidence regarding four weeks of

TV coverage in eleven Western countries. The sample of TV news was taken between

December 2006 and April 2007 and contains 28 news casts per country and TV

station. In each country, the exact same days were selected so as to avoid that

international crime or events would bias the data and make them non-comparable

between countries. In total, the evidence covers news from 24 different TV stations,

each time the largest commercial channel and the public broadcaster. In each country

6

and for each station, we analyzed the main TV news show in prime time. In total 672

entire broadcasts and 9.514 news items were watched, analyzed and coded. Table 1

contains a description of the dataset.

Table 1: Sample of 28 days of coverage on 24 TV stations (December 2006-April 2007)

TV Station Country Channel type # news items VRT Belgium-Flanders Public 595 VTM Belgium-Flanders Private 594

RTL-tvi Belgium-French Public 559 RTBF Belgium-French Private 512 CTV Canada Public 347 CBC Canada Private 208 Fr2 France Public 574 TF1 France Private 712 ARD Germany Public 262 ZDF Germany Public 293 RTL Germany Private 336 RTE Ireland Public 484 RAI Italy Public 487 NOS Netherlands Public 334 RTL4 Netherlands Private 286 NRK Norway Public 481 TV2 Norway Private 400 TRT Turkey Public 598 Star Turkey Private 496 BBC UK Public 234 ITV UK Private 155 ABC US Private 232 NBC US Private 232 CBS US Private 103 Total 9,514

Although the number of TV news broadcasts is the same throughout the

sample, we took 28 days for each country, the total number of news items in the

sample differs extensively. This is due to two factors: the length of the main

newscasts differs and also the length of the news items differs across channels. In

France, the private broadcaster TF1 has long newscasts with on average short items

leading to a large amount of analyzed items (N=712) while the other extreme, CBS in

the US, has short newscasts leading to a much smaller sample of news items (N=103).

The issue here is whether we can compare these different samples. We believe this to

be the case as we took for each country the main news show issued in prime time.

Whether stations make long or short news shows or use long or short items does not

7

really matter; we are interested in TV stations’ structural policies regarding crime

news. In the bulk of our analyses we do not take the number of items but rather the

number of seconds devoted to a certain issue as point of departure. Also, we most of

the time work with percentages and not with the absolute figures.

Apart from the different format, there are of course other differences between

the stations in the sample than them being public or private channels. In most

countries the national TV stations have the largest audience. They generate the

newscasts that are watched by most people. Yet, this does not apply to the US. The

three American stations in the sample belong to the national networks. Many

Americans, though, tend to rely on local TV stations to get their daily news.

Moreover, it is probably through their local news that most Americans get their daily

portion of crime (Romer, Aday, and Hall Jamieson 2003). So, our selection of

national networks most likely biases the data in favour of less and different crime

coverage in the US. We acknowledge that the US data are probably not entirely

comparable with the European data.

A final remark about the sample relates to its size. Although the total number

of items is very large, the amount of items per country and per station is not

extremely big. With 28 days of news chances exist that our sample is too small and

that random differences due to local events and temporary situations conceal

structural differences between countries and stations. For example, serial killings in a

certain country during the research period may seriously affect the amount and type

of crime news in that country without pointing towards a systematic difference in

coverage. We expect crime coverage to be highly event-driven. That is why we spread

out our sample over a period of five months to avoid single events to dominate the

sample. As we will show later, the variance between days in terms of the amount of

crime coverage is extensive. We are aware of those problems but they do not impede

our undertaking. If the sample is too small this will lead to conservative errors: we

will decide that there is no pattern while there might be one if we had more data.

Moreover, we are not aware of any other international comparative news study that

draws upon a larger sample non-electoral sample; most studies focus on election

periods (Van Aelst and De Swert 2009: 152).

Trained coders, most of them native speakers, watched the recorded news

broadcast and analyzed them item per item. Per item a large amount of variables is

coded. The crucial variable is the issue code. Drawing upon a detailed issue-

codebook, coders could attribute up to three issue codes to a single news item. If one

of these three possible codes refers to the broad crime code, we consider the item to

8

be about crime. In many cases, coders only gave one code; the average issue code per

item is 1.47.

Within the broad crime code, we apply a fine-grained codebook that details

with what type of crime the story deals—the codebook distinguished 18 different

types of crime. We regroup them into five major categories: (1) violent crime

(homicide, rape, paedophilia, robbery, assault, parental or conjugal violence, human

trafficking), (2) property crime (theft, burglary), (3) drugs & co (drugs, vandalism,

hooliganism, street violence, rioting), (4) white collar crime (fraud, corruption) and

(5) others (prison escapes, abduction). Apart from the type of crime in the item, two

other series of variables are coded: the actors that are interviewed in the item and the

amount of sensational elements in the item.

First, in terms of the actors, we only record an actor if he/she is actually

interviewed or quoted in the news item or when he/she forms the subject of the news

item. Many different people are shown in crime news items. We distinguish four

groups of actors: (1) involved persons (standers-by, witnesses, victims,

perpetrators…), (2) police/justice (police officers, advocates, judges…), (3)

commentators (observers, commentators, experts…), (4) politics (politicians,

spokespersons of social movements…) and (5) others. The idea is that the distinction

between types of actors grasps the frame that is used to tell the crime story: from

what perspective, from which angle are the stories about crime brought? Crime

stories can be told from the perspective of the involved people which implies that the

main angle would be human interest: How are people reacting when they are

confronted with crime? Who are the victims? Who are the perpetrators? The frame

adopted is probably also an episodic frame, it does not focus on the structural causes

but rather on the personal perception and consequences of crime (Iyengar and

Kinder 1987). Taking the perspective of politicians, in contrast, most likely leads to a

thematic framing where structural causes, interpretations and consequences are put

centre stage. Police/justice actors in a story suggest that the story gets an institutional

wrapping: it deals with how the police discovered and dealt with the crime, how the

justice system is reacting and how the crime is processed by the repressive

institutions of the state. In sum, we distinguish a human interest, a structural and an

institutional frame.

Second, in terms of the ‘colour’ and format of the coverage, each news item

was scored on a number of dummy variables tapping the sensational and shocking

character of the news. Three types of sensationalism are being distinguished:

sensationalism in images, in sounds and in the emotions of people shown.

9

Sensational images are images showing actual violence, bodies of dead people, or

injuries of people. The sound of a crime news item too can be sensational: it can have

a music score, contain dramatic sound (e.g. gun shot), contain dramatic sound

produced by people (e.g. crying, screaming…) and can incorporate sounds of joy.

Finally, the people shown in the news items can display strong emotions of sadness,

fear, fury or joy (for a similar operationalization of sensationalism see: Hendriks

Vettehen, Nuijiten, and Beentjes 2006). We use a very simple aggregated dichotomy:

as soon as a news item contains one of the things above, we consider it to be a

sensational item.

In summary, the study has three dependent variables: (1) the amount of crime

news; (2) the actors quoted in the news item indicating the framing of the item; (3)

the sensational character of the news item.

The independent variables are fourfold. As put forward above, we hypothesize

that the type of TV station determines the crime news. This variable is easy to

operationalize: TV stations are scored ‘1’ when they are private and ‘0’ when they are

public channels. The second independent variable is the media and political system.

We simply follow Hallin & Mancini’s typology and code their Liberal systems with ‘1’,

the Democratic Corporatist systems ‘2’ and the Polarized Pluralist systems ‘3’. The

third independent variable is the actual amount of crime occurring in the countries

under study. Here we rely on crime rate data produced by Eurostat, the Canadian

Centre for Justice Statistics, and the Criminal Justice Statistics (US). All figures are

from 20061 Crime statistics are notoriously tricky to compare. The data we use are all

based on police statistics and thus largely based on crime reporting by citizens. The

figure used in the analyses is the percentage citizens that has reported a violent or a

property crime to the police in 2006. The fourth independent variable is the media

market share/fragmentation for the different countries and TV stations. No

centralized data are available so we construct three separate variables based on

disparate sources2 each grasping another aspect of the media competition. The first

measure is simple market share: how large (in %) is the audience of the TV station.

Note that we take general market share of a channel as we cannot distinguish

between TV news and general market share. Second, we tap market fragmentation by

adding up the market shares of the public broadcaster and its largest private

1 Eurostat (http://epp.eurostat.ec.europa.eu/ ); The Canadian Centre for Justice Statistics

(http://www.statcan.gc.ca); Sourcebook of criminal justice Statistics (http://www.albany.edu/sourcebook/ )

2 We mainly used the EU Mavise database (http://mavise.obs.coe.int/) and the data from the European Journalism Centre (http://www.ejc.net/media_landscape/).

10

11

competitor and then reversing the value. The larger the market fragmentation

coefficient the more fragmented the market is. Finally, we create a third market

variable coined media competition. This variable is defined as on ordinal variable

with ‘1’ pointing towards a situation of duopoly, ‘2’ indicates that there are three main

players in the TV-market, ‘3’ stands for more than three players, and ‘4’ means that

there are many players and that the market is very fragmented.

ASSESSING THE AMOUNT AND TYPE OF CRIME COVERAGE

How much crime is there on the news? Our data allow us, according to our

knowledge for the first time (but see: Heinderyckx 1993), to assess the amount of

crime news in eleven countries and on 24 TV stations. Table 2 gathers the data on

crime and compares them with all other issues in the news. The table is rank-ordered

on the last column. As mentioned earlier, coders could assign several issue codes to a

news item. If one of these codes is a crime code, we consider the story to be dealing

(also) with crime.

Table 2: Coverage (in %) per issue on 24 TV stations in 11 countries (December 2006-April 2007) Country BEL-FL BEL-FL BEL-F BEL-F CA CA FR FR GE GE GETV Station VRT VTM RTBF RTL-tvi CTV CBC Fr2 TF1 ARD ZDF RTL Crime 15.5 22.2 17.9 21.8 15.9 19.1 12.8 11.2 12.8 14.5 20.2 Politics 11.0 6.1 16.4 10.5 10.8 12.9 19.0 16.1 13.5 9.2 4.6 Economics/Finances

11.4 8.5 9.9 14.1 10.9 9.3 10.8 12.1 16.7 15.6 15.8Social Issues

7.5 9.2 7.2 7.7 16.1 15.9 14.1 13.0 4.6 6.2 12.5

Justice 11.8 13.4 12.8 12.0 10.7 8.4 7.7 6.7 13.5 9.9 9.1Environment/Energy 9.4 12.0 10.2 13.1 10.7 11.7 10.3 13.1 6.2 12.9 10.6War & Peace

7.1 3.5 7.0 3.2 9.0 12.8 4.1 5.5 14.8 14.1 5.1

Sports 14.8 10.5 5.3 8.0 4.6 5.9 5.6 5.3 8.0 12.8 17.9Disasters 8.3 10.3 5.2 7.7 11.0 5.6 8.1 10.6 6.4 3.8 12.7Celebrities/Royals

6.0 6.1 3.6 6.9 13.3 9.0 9.1 9.0 1.1 2.7 7.0

Defence 5.6 3.6 6.0 2.5 8.2 14.7 2.4 2.9 12.1 10.9 3.8Culture 4.9 4.9 6.0 5.4 4.8 2.7 10.9 11.9 4.6 3.9 2.3Mobility

1.4 1.0 1.4 3.9 1.4 1.9 0.2 1.8 12.1 0.9 0.5

Work 5.4 2.7 6.7 4.8 3.4 2.1 5.0 7.2 7.5 9.3 4.5International

3.5 0.8 6.0 1.2 4.2 3.0 4.5 3.6 6.4 3.0 1.8

Rights 2.7 2.7 4.0 2.5 3.7 2.5 3.8 2.9 7.8 3.2 1.1Tourism 3.8 4.6 3.0 3.2 3.3 1.1 5.3 5.2 1.3 5.0 4.8Media 2.8 3.5 4.2 3.9 1.9 5.2 0.5 2.0 2.3 2.3 4.6Education 0.8 2.2

3.1 1.7 0.5 1.2 4.0 4.1 0.8 0.5 1.5

Europe 2.9 1.4 5.5 2.2 0.0 0.6 1.5 0.7 6.2 5.0 1.7Religion 1.9 1.7 1.1 0.9 0.5 1.8 2.8 1.3 0.8 0.8 1.2Science 2.6 0.7 0.5 1.2 1.8 2.2 1.9 2.1 1.6 2.5 1.9Migration 1.9 0.8 1.0 0.5 0.4 1.6 1.0 1.4 1.5 0.6 0.5Total % (double codes) 143.0 132.5 143.9 139.0 147.1 151.4 145.5 149.7 162.7 149.7 145.7 Total seconds 51,019 57,461 48,715 52,121 33,492 18,870 55,374 59,145 17,801 21,089 24,477

12

13

Table 2 (continued) Country IT NL NL NO NO TU TU UK UK US US US TOTTV Station RAI NOS RTL4 NRK TV2 TRT Star BBC ITV ABC NBC CBS Crime 20.0 20.7 19.3 14.4 18.4 14.0 32.6 21.8 21.3 6.3 8.7 10.5 17.8Politics 14.7 17.7 13.3 10.7 15.3 11.8

5.4 11.6 12.3 9.9 17.4 24.4 12.3Economics/Finances 8.7 12.6 12.4 8.7 7.3 6.9 3.3 12.7 6.6 13.5 10.9 13.4 10.5Social Issues 11.6 7.8 11.9 11.4 7.0 5.2 8.1 11.2 7.6 11.2 16.4 11.1 10.3 Justice 9.1 11.6 13.9 10.1 11.7 5.7 5.1 11.9 16.6 4.2 9.4 12.8 10.2Environment/Energy 4.3 7.9 10.3 11.3 7.6 9.2 5.9 8.3 13.8 5.6 5.2 7.8 9.5War & Peace 12.1 15.1 13.3 10.5 8.9 14.2 2.9 14.0 22.0 25.6 19.7 32.4 9.5 Sports 5.4 1.7 2.1 10.2 19.9 9.3 3.7 6.5 10.3

6.0 1.8 0.0 8.3

Disasters 7.1 5.9 9.2 6.9 7.3 7.5 8.5 6.2 8.8 4.6 5.6 12.8

7.8Celebrities/Royals

10.1 3.7 5.9 9.8 4.6 2.8

15.4

5.4 4.7 14.7 11.3 5.3 7.4

Defence 8.9 6.9 7.7 7.1 4.1 6.7 2.5 10.9 13.4 21.7 30.7 30.3 7.1Culture 5.6 6.1 3.1 4.6 2.2 1.2 5.8 3.0 2.4 15.8 3.5 0.4 5.4Mobility

5.1 1.8 8.6 4.5 5.1 9.3 2.0 12.2 15.2 16.5 19.7 5.0 4.4

Work 2.3 5.8 4.2 4.6 6.0 1.2 0.2 2.9 0.2 2.8 4.9 1.2 4.2International 4.2 5.6 6.7 7.4 4.9 7.0 1.5 3.2 3.2 2.4 4.4 0.3 3.8Rights 6.2 4.1 3.7 4.6 3.6 4.1 6.8 4.2 2.2 3.0 4.1 4.1 3.8Tourism 2.7 2.7 6.4 2.1 3.3 3.1 2.5 2.7 0.6 5.0 2.6 2.6 3.5Media 2.9 3.1 4.3 1.4 2.5 4.5 9.2 3.0 1.3 6.0 0.2 1.2 3.3Education 0.4 2.7 3.0 1.7 0.3 3.1 0.9 4.0 1.0 2.6 0.0 0.0 1.9Europe 3.3 3.6 4.1 1.4 0.9 1.0 0.8 0.7 0.0 0.0 0.0 0.0 1.8Religion 3.9 1.5 2.9 0.8 1.0 0.9 3.9 1.2 1.4 2.0 3.3 0.0 1.8Science 1.1 1.0 0.1 1.5 1.9 1.9 1.4 1.2 0.0 6.2 3.0 2.9 1.6Migration 0.6 1.0 1.6 1.3 1.4 0.5 0.4 0.7 2.5 0.0 1.4 1.2 1.1Total % (double codes) 150.3 150.5 167.9 146.9 145.4 131.1 128.7 159.6 167.3 185.5 184.1 179.5 147.2 Total seconds 44,072 32,444 25,290 44,179 25,989 43,423 65,994 27,111 18,890 18,584 23,420 11,191 868,320

The first thing that catches the eye is that crime coverage is the single most

important topic in the news. On average over 11 countries and 24 TV stations, 17.8

percent of the news is devoted to crime stories. The second most popular topic

following at a distance is politics with 12.3 percent, then follow economics, social

issues etc. Interestingly, the stories dealing with the justice issue strongly affiliated

with crime get another substantial piece of the news cake with 10.2 percent. We

defined justice stories here as stories that deal with the organization of the police or

the justice department, stories that focus on crime policy and prison policy, stories

that relate to crime prevention, and, most importantly, all stories dealing with trials.

This final trial subcategory is by far the largest one in the justice category (5,54%).

When we would use a broader definition of crime stories, including the law suits that

are the consequences of crime and that do contain a good deal of information about

the crime itself, the total share of crime stories would have been well above one fifth

of al stories. The point is simple: there is a lot of crime on TV.

Comparing the different TV stations shows that the high crime average is not

due to one or a few crime stations. For 13 of the 24 channels crime is the most

important topic in the news; on 19 channels crime figures among the top two topics.

Interestingly, US network news is much less crime focused than most European

national TV news. On ABC, NBC and CBS crime comes only at place 8-9 in the

ranking. Interestingly, in the 1990s crime news still was the number one topic on

these same networks (Lowry, Nio, and Leitner 2003: 62). This American anomaly

anno 2006-2007 confirms what we said earlier; our sample of stations is probably

not well-suited to compare European with American channels. Apart from the US,

also in France (FR2 and TF1) and in Germany (ARD and ZDF) crime is challenged by

runners-up (mainly politics). The good news for our analyses is that, notwithstanding

the substantial average amount of crime news, the variance between stations is

large—there is something to be explained. The champion of crime is the Turkish

commercial channel Star that systematically devotes one third of its news to crime

coverage. The American ABC is least prone to broadcast crime stories with only 6.3

percent of the stories.

Crime coverage is event driven. There is not much continuity in the news

about crime. On some days a large chunk of the daily news goes to crime, the next day

hardly any news covers crime events. This is documented in Table 3 containing the

average amount and the standard deviation of crime per station. For some TV-

stations the standard deviation is larger than the mean which implies that crime

coverage is extremely volatile and goes up and down oscillating wildly.

14

Table 3: Average amount (%) and standard deviation of daily crime news per station

Channel Country Mean Std. Deviation Star Turkey 32.6 19.1 VTM Belgium-Flanders 22.2 12.7 BBC UK 21.9 19.4 RTL-tvi Belgium-French 21.8 13.2 ITV UK 21.3 19.1 NOS Netherlands 20.7 12.6 RTL Germany 20.3 10.4 RAI Italy 20.0 12.8 RTL4 Netherlands 19.3 14.7 CBC Canada 19.1 16.6 TV2 Norway 18.4 12.7 RTBF Belgium-French 17.9 8.7 CTV Canada 15.9 13.9 VRT Belgium-Flanders 15.5 9.2 RTE Ireland 14.7 9.1 ZDF Germany 14.5 12.5 NRK Norway 14.4 11.2 TRT Turkey 14.0 10.0 ARD Germany 12.8 12.8 France2 France 12.8 7.3 TF1 France 11.2 8.5 CBS US 10.5 14.8 NBC US 8.7 9.6 ABC US 6.3 9.1 Total 17.8 13.9

That crime news is strongly event driven can also be demonstrated by

correlating the daily crime share between the TV stations in the same country.

Correlations (Pearson) are mostly high and range from .76 in Turkey to .33 in

Canada. These coefficients indicate that crime coverage in a country reacts to events

and that most TV stations react to some extent similarly to the same events or

absence of events. That crime is so volatile challenges to some extent the

representativity of the sample. Our 28 days may be biased due to sheer (un)luck

picking the ‘wrong’ days in some countries. It remains to be seen whether we find a

consistent pattern in crime coverage.

Regarding the first dependent variable, Table 3 contains the most important

information: the total share of crime coverage on each TV station. Yet, not only the

amount of crime but also the type of crime that is being reported matters. Therefore,

Table 4 documents the types of crime that are covered.

15

Table 4: Types of crime coverage (in %) Violent Property Drugs &

co White Collar

Others Total % Total seconds

Star 55.9 17.3 14.5 8.9 3.4 100 21,514 VTM 51.0 10.8 17.8 17.8 2.5 100 12,752 BBC 57.4 5.6 7.4 9.3 20.4 100 5,923 RTL-tvi 52.0 6.0 15.3 18.7 8.0 100 11,350 ITV 65.6 0.0 9.4 12.5 12.5 100 4,023 NOS 48.1 6.2 17.3 21.0 7.4 100 6,726 RTL 55.7 5.7 18.6 8.6 11.4 100 4,956 RAI 48.8 3.9 14.2 6.3 26.8 100 8,812 RTL4 56.2 9.6 11.0 13.7 9.6 100 4,875 CBC 43.2 13.6 9.1 18.2 15.9 100 3,596 TV2 48.1 13.0 24.7 6.5 7.8 100 4,791 RTBF 43.6 6.9 20.8 23.8 5.0 100 8,711 CTV 59.3 1.7 6.8 20.3 11.9 100 5,314 VRT 57.7 8.1 12.6 18.0 3.6 100 7,909 RTE 42.4 8.7 17.4 17.4 14.1 100 7,085 ZDF 45.7 2.2 17.4 17.4 17.4 100 3,056 NRK 42.7 9.8 24.4 22.0 1.2 100 6,362 TRT 44.8 7.3 20.8 22.9 4.2 100 6,078 ARD 30.0 5.0 27.5 17.5 20.0 100 2,280 France2 47.4 5.2 18.6 8.2 20.6 100 7,079 TF1 45.1 7.8 16.7 12.7 17.6 100 6,626 CBS 54.5 0.0 0.0 27.3 18.2 100 1,179 NBC 43.5 0.0 4.3 34.8 17.4 100 2,032 ABC 30.4 8.7 0.0 43.5 17.4 100 1,180 Total 49.6 8.0 16.1 15.8 10.5 100 154,209 Note: As coders multiple-coded, percentages are calculated based on a summation of all codes.

As expected, the largest share of crime coverage deals with violent crimes

(49.9%). Although these crimes are in reality less frequent than property crimes, they

largely outweigh property crime (8.0%). Among the TV stations with a lot of crime

coverage, the share of violent crime is larger then among the stations that show less

crime in the news. Drugs, vandalism, hooliganism etc. (16.1%) get more attention

than property crime and the same applies to white collar crime (15.8%). Overall the

variation between countries and stations is considerable.

The second dependent variable is the framing of crime news as measured by

the type of actors quoted and dealt with in the news stories. If a story contains people

that are somehow personally involved in the event we suppose a story gets a human

interest frame. If a story includes political actors we suppose that the crime event is

covered according to a structural frame. If the story contains police and judiciary

16

sources it is told according to a institutional frame. Table 5 contains the data per TV

station.

Table 5: Actors quoted and mentioned in crime stories (in %) Involved =

Human Interest

Political = Structural

Judiciary = Institutional

Commentators N actors

ABC 14 50 0 9 22 ARD 27 48 12 2 66 BBC 40 28 10 10 116 CBC 30 22 11 5 121 CBS 50 17 17 17 6 CTV 38 14 13 7 137 France2 30 19 7 2 211 ITV 42 18 11 6 62 NBC 21 36 9 13 92 NOS 28 34 12 3 190 NRK 20 23 16 5 151 RAI 42 9 5 7 168 RTBF 33 27 9 6 234 RTE 39 26 11 2 186 RTL 39 26 16 4 115 RTL4 32 21 19 3 128 RTL-tvi 44 14 10 0 261 Star 51 20 7 4 398 TF1 29 25 12 6 154 TRT 28 45 3 4 118 TV2 23 23 15 8 158 VRT 36 25 15 3 185 VTM 44 13 13 2 231 ZDF 27 39 10 6 83 Total 35.5 23.4 10.7 4.6 3,593 Note: As several actors may be quoted in the same item the percentages are calculated based on a summation of all actors types.

Some stations only broadcasted a very limited amount of crime stories

featuring specific actors. The American CBS, for example, only referred to six actors

in their crime coverage. This is definitely not sufficient to determine the typical frame

CBS employs to tell the crime stories. Other stations, in contrast, abundantly use

speaking actors to help telling the audience about crime. The Turkish Star, for

example, showed almost 400 actors in its crime news during the 28 day research

period. In general, TV news seems to rely a lot on involved actors; one third of all

actors speaking are victims, offenders, eye witnesses, friends, family etc. suggesting

17

Table 6: Actors quoted in crime stories per crime type (in %) Involved =

Human Interest

Political = Structural

Judiciary = Institutional

Comment N actors

Violent 45.8 20.6 10.4 4.4 1876 Property 29.1 15.4 18.9 2.0 254 Drugs & co 20.9 28.5 11.2 2.4 545 White Collar 18.3 31.8 9.2 6.7 761 Other 43.5 22.1 9.2 6.8 294 Note: As several actors may be quoted in the same item the percentages are calculated based on a summation of all actors types.

that the human interest frame is widely employed. The structural frame ranks

second: one fourth of the performing actors are politicians or representatives of civil

society. The institutional frame seems to be less attractive for the news, only one in

ten actors belongs to the state’s repressive apparatus. The table documents that

differences between stations are large. The Turkish Star, again, stands out with in

half of its stories a human interest frame while the American ABC seems to discard

that frame almost entirely (14%). ABC, in contrast, heavily relies on structural

framing (50%) while the Italian RAI hardly ever gives the stage to politicians when

covering crime (9%). Institutional framing varies less between 0 and 19 percent of the

actors.

The third dependent variable is the news format, which we defined as the

degree of sensationalism of the news. We explained earlier how we operationalized

this dimension in images, sounds and emotions. Table 7 contains the data per station.

There is a clear association between the type of crime reported and the frame

used. Table 6 crosses crime type and framing. Violent crimes are, more than other

crime types, most covered emphasizing the personal perceptions and experiences of

the people directly involved. White collar crime is most interpreted as a political and

structural problem. Property crimes are most framed as being issues the judiciary

deals with.

18

19

Table 7: percentage news items with sensational features and overall sensationalism index

Images Emotions sounds Total Violence Dead Injured Sadness Fear Fury Joy Music Drama

sounds Drama

Persons Joy Sensationalism

index N

Star 25,7 15,6 10,1 19,6 3,4 16,2 6,1 70,4 15,6 14,5 10,1 88,8 179TRT 17,7 2,1 6,3 7,3 1,0 6,3 0,0 50,0 4,2 6,3 0,0 70,8 96NBC 8,7 4,3 8,7 8,7 0,0 8,7 4,3 4,3 4,3 8,7 8,7 47,8 23RTL4 20,5 11,0 4,1 4,1 5,5 4,1 2,7 0,0 1,4 1,4 4,1 41,1 73CTV 10,2 15,3 3,4 13,6 1,7 3,4 8,5 1,7 0,0 5,1 5,1 39,0 59RTL 11,4 10,0 1,4 8,6 2,9 1,4 5,7 2,9 4,3 5,7 4,3 38,6 70CBS 9,1 9,1 0,0 18,2 0,0 0,0 9,1 0,0 0,0 0,0 0,0 36,4 11NOS 19,8 6,2 2,5 1,2 1,2 4,9 3,7 3,7 3,7 1,2 6,2 35,8 81TF1 8,8 7,8 5,9 7,8 3,9 4,9 2,9 0,0 5,9 7,8 4,9 27,5 102RTL-tvi 4,7 1,3 6,0 11,3 3,3 4,7 1,3 2,0 1,3 4,0 1,3 26,7 150VTM 7,0 1,9 3,2 7,0 1,3 2,5 2,5 5,7 1,9 5,1 2,5 26,1 157RTBF 9,9 4,0 5,9 7,9 0,0 8,9 1,0 0,0 3,0 5,0 2,0 25,7 101RAI 11,0 3,9 3,9 4,7 2,4 1,6 3,9 3,1 5,5 5,5 7,9 25,2 127ARD 7,5 7,5 5,0 5,0 0,0 7,5 7,5 0,0 7,5 7,5 5,0 25,0 40Fr2 9,3 3,1 2,1 5,2 4,1 8,2 6,2 2,1 2,1 3,1 3,1 24,7 97CBC 6,8 4,5 6,8 2,3 0,0 2,3 2,3 0,0 0,0 0,0 0,0 22,7 44BBC 11,1 1,9 3,7 3,7 1,9 7,4 9,3 1,9 1,9 3,7 7,4 22,2 54ITV 6,3 6,3 3,1 0,0 3,1 9,4 3,1 6,3 3,1 0,0 6,3 21,9 32ABC 4,3 4,3 8,7 0,0 0,0 0,0 4,3 4,3 0,0 0,0 0,0 21,7 23TV2 13,0 2,6 2,6 5,2 1,3 9,1 2,6 0,0 5,2 2,6 0,0 20,8 77NRK 13,4 3,7 4,9 4,9 1,2 6,1 3,7 0,0 6,1 6,1 4,9 19,5 82ZDF 2,2 6,5 0,0 6,5 0,0 0,0 4,3 2,2 2,2 2,2 2,2 15,2 46RTE 9,8 2,2 2,2 1,1 1,1 3,3 2,2 0,0 1,1 2,2 0,0 13,0 92VRT 3,6 1,8 0,9 0,9 0,9 2,7 3,6 0,0 0,9 0,9 1,8 8,1 111Total 11,5 5,6 4,5 7,1 2,0 5,8 3,7 10,6 4,2 5,0 3,9 33,5 1927Note: the sensationalization index refers to the total amount (%) crime stories that has at least one of the sensationalism features

The table reveals that there are huge differences between stations in the extent

that they use sensational formats to support their crime stories. The index (last

column) summarizes the table and varies between 88.8 (Star, private channel in

Turkey) and 8.1 (VRT, public broadcaster in Belgium-Flanders). This indicates that

almost all crime stories aired in Star have at least one sensationalist feature and that

almost no stories aired on VRT have sensational formatting. Both Turkish channels

really stand out as also the Turkish public TRT uses a lot of sensational formats. On

closer inspection, the Turkish sensationalism is partly due to the abundant use of a

music score when reporting about crime. Yet, even without taking music into

account, Star would still be the champion of sensationalism (68.3) while TRT would

score slightly above the median (26.1). Looking at the different sensational formats

shows that all elements are used to some extent and not only in Turkey. Crime stories

often contain images of actual violence, of dead or injured people. People featuring in

crime news regularly show extreme emotions of sadness, fear, fury or joy. TV news

often employs dramatic sounds to increase the effect of the news.

Crime type and sensationalism are associated. Table 8 crosses both variables.

In contrast to what one may have expected, it are not the violent crime stories that

are told most sensationally (34.7) but rather the items about drugs, vandalism,

hooliganism etc (49.7). Property crimes are least sensationally formatted (4.5) which

makes sense as these crimes are not violent and thus no violence or injured people

can be shown and as people’s reactions to these sorts of crimes are probably less

emotional.

20

21

Table 8: Type of crime and sensationalism index per TV station

Violent Property Drugs & co White

Collar Others Total

Star 85.0 0.0 96.2 81.3 66.7 88.8 TRT 62.8 50.0 85.0 72.7 25.0 70.8 NBC 80.0 0.0 100.0 12.5 25.0 47.8 RTL4 48.8 3.2 75.0 10.0 28.6 41.1 CTV 37.1 0.0 75.0 16.7 28.6 39.0 RTL 30.8 11.1 61.5 0.0 50.0 38.6 CBS 33.3 — — 33.3 50.0 36.4 NOS 30.8 25.0 57.1 35.3 0.0 35.8 TF1 34.8 14.3 41.2 7.7 11.1 27.5 RTL-tvi 28.2 14.3 52.2 3.6 8.3 26.7 VTM 26.3 0.0 32.1 17.9 25.0 26.1 RTBF 22.7 12.5 47.6 0.0 20.0 25.7 RAI 27.4 0.0 50.0 12.5 14.7 25.2 ARD 41.7 0.0 45.5 0.0 0.0 25.0 France2 26.1 0.0 33.3 0.0 20.0 24.7 CBC 21.1 16.7 50.0 37.5 0.0 22.7 BBC 32.3 0.0 25.0 0.0 9.1 22.2 ITV 28.6 — 33.3 0.0 0.0 21.9 ABC 14.3 0.0 — 20.0 25.0 21.7 TV2 21.6 0.0 36.8 0.0 16.7 20.8 NRK 20.0 0.0 40.0 5.6 0.0 19.5 ZDF 19.0 0.0 12.5 0.0 25.0 15.2 RTE 7.7 25.0 31.3 6.3 15.4 13.0 VRT 9.4 0.0 21.4 0.0 0.0 8.1 Total 34.7 4.5 49.7 18.1 17.7 33.5

EXPLAINING CRIME COVERAGE

Our explanatory variables are all situated on the aggregate TV station level.

This implies that the statistical power of our analyses is confined since we only have

24 cases at our disposal. For some analyses the N is even smaller since some data are

missing. This limits our statistical tools and compels us to rely on simple correlations.

Yet, our independent variables are structural and clearly ‘preceding’ crime coverage

both in time and theoretically. Therefore, we believe we can cautiously make causal

inferences based on simple correlations. Correlations between the dependent and

independent variables can be found in Table 9. As the number of observations is

small, we apply slightly more relaxed significance thresholds (p<.1).

22

Table 9: Dependent and independent variables, correlations (Pearson) and significance (p-value)

Amount Framing Format Human interest Structural InstitutionalReal crime rate

Violent crime (N=24) .533 (.007) .184 (.194) -.190 (.187) .157 (.233) -.343 (.101) Property crime (N=19) -.219 (.368) -.099 (.343) .043 (.431) .036 (.442) -.527 (.020)

Channel type Public-private channel (N=24) .178 (.204) .064 (.384) -.194 (.181) .085 (.346) .282 (.091)

Media & politics system Liberal (N=24) -.280 (.092) .047 (.413) .030 (.445) -.115 (.297) -.118 (.291)

Democratic Corporatist(N=24) .175 (.207) -.151 (.240) .061 (.388) .507 (.006) -.280 (.092) Polarized Pluralist (N=24) .111 (.303) .130 (.272) -.110 (.304) -.489 (.008) .481 (.009)

Market Market share (N=18) -.211 (.400) -.281 (.130) -.457 (.028) .243 (.165) -.555 (.017)

Fragmentation (N=15) .488 (.094) .480 (.035) -.420 (.060) -.532 (.021) .769 (.001) Competition (N=24) -.155 (.470) .062 (.387) .392 (.029) -.521 (.005) .558 (.005)

For the impact of real crime rates on the amount, framing and format of crime

news we distinguish between violent crime and property crime rates. The evidence

seems to confirm H1 as there is a positive and significant relation between the

number of violent crimes recorded per 1000 inhabitants in the 11 countries under

study and the amount of coverage of these types of crime in the same countries. This

finding contradicts many available studies that found no link, or a negative one,

between reality and coverage. When comparing internationally, though, this

association seems to hold and in countries with more violent crime more violent

crime is reported in the news. We did not expect to find a relation between crime

coverage framing and real crime rates and this is indeed what we find. We did not

expect to find a relation between real crime rates and crime coverage formats either

but our evidence suggests such a relation does exist. The property crime statistics are

negatively correlated with the frequency of sensational formats in the news.

Apparently, when property crime is omnipresent, the need to frame it dramatically

decreases.

One of the main variables of the study is the distinction between public and

private broadcasters. The hypotheses based on the literature we put forward was that

commercial TV would feature more, less structurally framed, and more sensational

crime coverage. Only one of these hypotheses receives support: private channels

generate more sensational crime stories that are on average more underscored by

music, contain dramatic sounds or emotional and shocking images etc. This confirms

H2c. But H2a en H2b are falsified. Private channels do not bring more crime news

nor do they frame their crime news differently than public broadcasters. By and large,

the effect of channel type on crime coverage is on the low side.

The media-politics typology from Hallin & Mancini does a better job in

explaining differences between TV stations. We expected that the Polarized Pluralist

model, due to the low professionalization of journalism in these countries, would

feature more, differently framed and more sensational news. These expectations are

warranted by the facts. In terms of the crime coverage level, the Polarized Pluralist

type does not cover more crime news than both other types but the Liberal type does

bring significantly less crime news (than the Polarized Pluralist type). So, we can

cautiously maintain H3a. In terms of framing, there is evidence that the institutional

frame, as predicted in H3b, is less used by the Polarized Pluralist type while stations

operating in a Democratic Corporatist environment rely more on this frame. The

second part of H3b, stating that the Polarized Pluralist environment would inspire TV

stations to adopt a human interest frame, does not get support. Polarized Pluralist

23

media contexts do produce, as expected in H3c, more sensational crime coverage—

the correlation is quite strong. Stations operating in a Democratic Corporatist

context, in contrast, cover crime in a less sensational way than stations in other

contexts.

Only a superficial look at the table suffices to notice that market features are

by far the strongest predictors of the amount, framing, and format of crime coverage

on the 24 TV stations in our sample. All our hypotheses regarding the effect of media

market competition on crime coverage are corroborated. Of all three market

variables, especially the fragmentation of the TV market seems to grasp the crucial

competition factor that drives the crime news. Fragmentation refers to the reversed

aggregated viewer rates (in%) of both the main public and commercial station. A high

fragmentation means that both major news channels together only have a small share

of the total market and that, thus, there are many other stations each getting a good

piece of the news cake. In fragmented markets, TV stations devote more time to crime

in their news (H4a), they use the human interest frame more often while they employ

the structural and institutional frame less frequently (H4b), and they much more

wrap up their crime stories in a sensational format (H4c). Especially when

competition is fierce TV stations seem to adopt the old adagio: ‘When it bleeds it

leads’. Regarding format, both the market share and the competition variables yield

further support for H4c: large stations use less sensational formats and heavy

competition leads to more sensation.

We do not dispose of enough cases to fully test all hypotheses multivariately.

So we have no clue which of the independent variables matters most. Conducting a

number of exploratory regression analyses in which only the significant variables in

Table 9 were incorporated reveals a clear structure, though: the amount of crime is

most affected by market fragmentation and the same applies to the human interest

frame. The occurrence of structural framing is primarily caused by the market

variables as well. Both market fragmentation and the Polarized Pluralist country type

significantly explain the use of institutional frames. Finally, sensational news

formatting can be explained best (adj. R² .846) as a combined effect of all four main

variables: real world property crime, channel type, Polarized Pluralist country type,

and market fragmentation.

In sum, comparing the four potential drivers of the amount, framing and

format of crime news, we find that real crime rates matter to some extent. The

distinction between public and private broadcasters produces less powerful results

than expected. Private broadcasters do format their crime news in a more sensational

24

way, but they do not cover crime more often nor do they systematically use different

frames. Media and politics systems matter to some extent. We find systematic

differences with media in Polarized Pluralist country types using less institutional

frames and relying on more sensational formats; media in Democratic Corporatist

countries do exactly the opposite. The strongest predictor by far is the TV market

situation. Especially market fragmentation strongly affects how much crime is

covered in the news, it affects the frames employed to tell the story with more human

interest and less institutional and structural interpretations, and it affects the use of

sensational formats.

CONCLUSION AND DISCUSSION

The empirical results of the study can be summarized in a few points:

(1) Crime is the single most covered issue across countries and TV

stations.

(2) Crime news is event-driven and volatile.

(3) Violent crime is covered most, followed by drugs and white collar

crime.

(4) Most of the time, crime is framed in a human interest frame,

especially when it regards violent crime.

(5) A large amount of crime stories contain sensationalist format

elements.

(6) Real crime rates affect the amount of crime coverage but not its

framing or format.

(7) Private channels use more sensationalist formats to cover crime,

but they do not cover more crime or frame their crime stories

differently.

(8) In Polarized Pluralist media-politics systems, TV stations’ crime

reports are more sensational and less institutionally framed than in

both other country types.

(9) Market competition and especially market fragmentation strongly

affect the amount, framing, and format of crime coverage.

Our study including 24 stations in 11 countries revealed large differences

between the channels. Interestingly, we found the largest differences in story

formats rather than in the amount of crime coverage. Most of the broadcasters in

25

the study devoted a considerable share of their main evening news to crime

events, few stations abstained from reporting about crime. The way crime stories

are told, however, and whether they are formatted in a sensational, shocking and

emotional wrapping differs extensively; some channels use an extremely

sensational language, others channels stay close to the facts and dedramatize

crime. Also the perspective used to tell crime stories varies considerably: some

channels adhere to a human interest perspective, others adopt a more structural

(politics) or institutional (justice) framing. Framing and format vary much more

than amount.

The most important finding is that the specific market context is the very

best predictor of the amount, frame and format of crime reporting in the news. Of

course, we do find differences between public and private channels and between

media situated in different country types, but these differences are modest

compared to the much larger differences we find between channels that operate in

a relatively non-competitive and non-fragmented markets with a few large players

compared to those operating in highly competitive, fragmented markets with

many players. If TV markets are highly fragmented stations are engaged in a

ruthless fight for audiences and crime news, especially sensational crime news, is

considered as a primary arena in the battle for the public. Even public

broadcasters in fragmented markets go along and compete with their private

challengers by sensationalizing their crime reporting. This result supports earlier

studies that focused on market context to explain content differences between

media outlets.

The study set off by claiming that studying crime news is important for two

reasons: crime coverage bears political consequences and it is a valuable indicator

of news quality. In terms of the political consequences of crime coverage, we

argued that crime coverage contributes to making people worried about crime,

that it affects their perception of the incidence of crime, that is may lead to

feelings of insecurity and malaise and, ultimately, to voting for extremist parties.

Also, we contended that political elites are sensitive for crime coverage and that

the level of attention for crime in the political system depends on the level of

crime coverage in the news. If all this is true, we would expect there to be

substantial differences between the countries at stake in how their citizens think

about crime (and politics) and how political elites deal with crime. The present

study focused on comparatively assessing the antecedents of crime coverage, we

leave it to others studies to tap the political consequences of crime coverage.

26

Finally, regarding news quality, the study suggests that the quality of the

news is in peril. Crime coverage is ubiquitous in the news and often packed in a

sensational and typical soft news framing. Public broadcasting makes a small

direct difference by restraining from frequently employing sensationalist formats.

In some countries with a different media-politics system rooted in different

historical trajectories, the quality of the news seems to be safeguarded for the time

being as long as professional journalism standards are holding firm. The main

threat for news quality does not seem to be commercialization per se but rather

the fragmentation of media markets. The more fragmented the markets are, the

less they care about quality and the more they compete down-market while

catering to the fears and prejudices of the public at large. The presence of a large

public broadcaster may suppress trends towards market fragmentation and

maintain quality in the news. Thus, while public broadcasters do not strongly

differ in their crime coverage from their commercial competitors their presence

may structure the competition in a market and thus, indirectly, contribute to

better news quality.

REFERENCES

Antunes, G.E. and P.A. Hurley. 1977. "The Representation of Criminal Events in

Houston's Two Daily Newspapers." Journalism Quarterly 54:750-760.

Atwater, T. 1984. "Product Differentiation in Local Tv News." Journalism Quarterly

61:757-762.

Canninga, L. 1994. "Een Vergelijkende Analyse Van De Nieuwscultuur in De

Buitenlandse Berichtgeving Op Brtn, Vtm, Nos En Rtl4." Pp. 91-114 in Media

En Maatschappij, edited by H. Verstraeten and P. Perceval. Brussel:

VUBPress.

Davis, F.J. 1952. "Crime News in Colorado Newspapers." American Journal of

Sociology 57:325-330.

Elchardus, Mark and Wendy Smits. 2002. Anatomie En Oorzaken Van Het

Wantrouwen. Brussel: VUBPress.

Ericson, R.V., P.M. Baranek, and J. Chan. 1991. Representing Order: Crime, Law,

and Justice in the News Media. Toronto: University of Toronto Press.

Esser, Frank. 1999. "'Tabloidization' of News. A Comparative Analysis of Anglo-

American and German Press Journalism." European Journal of

Communication 14:291-324.

27

Gerbner, G., L. Gross, M. Morgan, and N. Singorielli. 1994. "Growing up with

Television: The Cultivation Perspective." in Perspectives on Media Effects:

Advances in Theory and Research, edited by J. Bryant and D. Zillmann.

Hillsdale NJ: Erlsbaum.

Ghanem, Salma. 1996. "Media Coverage of Crime and Public Opinion: An

Exploration of the Second Level of Agenda Setting." PhD. Thesis, University of

Texas at Austin, Austin.

Gordon, M. and L. Heath. 1991. "The News Business, Crime, and Fear." in Agenda-

Setting: Readings on Media, Public Opinion, and Policy Making, edited by D.

L. Protess and M. McCombs. Hillsdale NJ: Erlbaum.

Graber, Doris. 1979. "Is Crime News Coverage Excessive?" Journal of communication

29:81-92.

—. 1980. Crime News and the Public. New York: Praeger.

Hallin, Daniel and Paolo Mancini. 2004. Comparing Media Systems. Three Model of

Media and Politics. New York: Cambridge University Press.

Hauttekeete, Laurence. 2004. "De Tabloidisering Van Kranten: Mythe of Feit? De

Ontwikkeling Van Een Meetinstrument En Een Onderzoek Naar De

Tabloidisering Van Vlaamse Kranten. Onuitgegeven Doctoraalscriptie."

Vakgroep communicatiewetenschappen, UGent, Gent.

Heinderyckx, Francois. 1993. "Television News Programmes in Western Europe: A

Comparative Study." European Journal of Communication 8:425-450.

Hendriks Vettehen, Paul, Koos Nuijiten, and Johannes Beentjes. 2006. "Research

Note: Sensationalism in Dutch Current Affairs Programmes 1992-2001."

European Journal of Communication 21:227-237.

Holz-Bacha, Christina and Pippa Norris. 2001. ""To Entertain, Inform, and

Educate": Still the Role of Public Television?" Political Communication

18:123-140.

Hvitfelt, H. 1994. "The Commercialization of the Evening News: Changes in Narrative

Techniques in in Swedish Tv News " Nordicom review 2:33-41.

Iyengar, Shanto and Donald R. Kinder. 1987. News That Matters. Television and

American Opinion. Chicago: University of Chicago Press.

Lotz, R. 1991. Crime and the American Press. New York: Praeger.

Lowry, Dennis T., Tarn Ching Josephine Nio, and Dennis Leitner. 2003. "Setting the

Public Fear Agenda: A Longitudinal Analysis of Network Tv Crime Reporting,

Public Perceptions, and Fbi Crime Statistics." Journal of Communication

53:61-73.

28

Marsh, H. 1991. "A Comparative Analysis of Crime Coverage in Newspapers in the

United States and Other Countries from 1960-1989: A Review of the

Literature." Journal of Criminal Justice 19:67-79.

Mullainathan, Sendhil and Andrei Shleifer. 2005. "The Market for News." The

American Economic Review 95:1031-1053.

Norris, P. 2000. A Virtuous Circle. Political Communications in Postindustrial

Societies. Cambridge: Cambridge University Press.

O'Keefe, G.J. and K. Reid-Nash. 1987. "Crime and Real-World Blues."

Communication Research 14:147-163.

Patterson, Thomas. 1993. Out of Order. New York: Knopf.

Picard, Robert G. 2002. "Research Note: Assessing Audience Performance of Public

Service Broadcasters." European Journal of Communication 17:227-235.

Powers, A., H. Kristkjansdottir, and H. Sutton. 1994. "Competition in Danish

Television News." Journal of Media Economics 7:21-30.

Romer, Daniel, Sean Aday, and Kathleen Hall Jamieson. 2003. "Television News and

the Cultivation of Fear of Crime." Journal of communication 53:88-104.

Sheley, J.F. and C.D. Ashkins. 1981. "Crime, Crime News, and Crime Views." Public

Opinion Quarterly 45.

Sherizen, S. 1978. "Social Creation of Crime News. All the News Fitted to Print." in

Deviance and Mass Media, edited by C. Wininck. BevrlyHills / London: Sage

Publications.

Slattery, Karen, Marc Doremus, and Linda Marcus. 2001. "Shift in Public Affairs

Reporting on the Network Evening News: A Move Towards the Sensational."

Journal of Broadcasting and Electronic Media 45:290-302.

Soroka, Stuart N. 2002. "Issue Attributes and Agenda-Setting by Media, the Public,

and Policymakers in Canada." International Journal of Public Opinion

Research 14:264-285.

Van Aelst, Peter and Knut De Swert. 2009 "Politics in the News. Do Campaigns

Matter? A Comparison between Political News During Campaign Periods and

Routine Periods in Flanders (Belgium)." Communications 34:149-168.

Walgrave, Stefaan and Knut Deswert. 2004. "The Making of the (Issues of the)

Vlaams Blok. The Media and the Success of the Belgian Extreme-Right Party."

Political Communication 21:479-500.

Walgrave, Stefaan, Stuart Soroka, and Michiel Nuytemans. 2008. "The Mass Media's

Political Agenda-Setting Power. A Longitudinal Analysis of Media, Parliament

29

and Government in Belgium (1993-2000)." Comparative Political Studies

41:814-836.

Walgrave, Stefaan and Volkan Uce. 2007. "Nieuws over Criminaliteit Op Het Tv-

Nieuws: Ruim Overdreven?" Pp. 189-210 in De Kwaliteit Van Het Nieuws.

Kwaliteitsindicatoren Voor Televisienieuws, edited by M. Hooghe, K. De

Swert, and S. Walgrave. Leuven: Acco.

Windhauser, J., J. Seiter, and T. Winfree. 1991. "Crime News in the Louisiana Press."

Journalism Quarterly 67.

30