Content analysis of corruption coverage: Cross-national differences and commonalities

Abstract

The article presents and summarizes some results from extensive cross-national content analysis of media coverage of corruption. The authors examined a sample containing 12,742 articles published in France, Hungary, Italy, Latvia, Romania, Slovakia and the United Kingdom from 2004 to 2013. A limited number of studies have been done thus far to reveal how the media deals with corruption cases in certain countries, and cross-national comparative analyses are exceedingly scarce. The core focus of the study is to reveal the significant differences in the corruption cases covered by the media according to the countries under analysis. We assume that some differences exist between the media coverage of corruption in the new and the old European Union member states and also that by classifying countries into groups based on their perception of the level of the corruption, some dissimilarities will be revealed between them. We conclude that the distinction between countries based on whether they are old or new European Union members does not wholly determine the nature of reporting on international or national corruption cases, for example, Italy was more similar to the old European Union member states in this sense. Considering the level of institutionalization of corruption cases, Italy appears to be more similar to the other old European Union members, but we should clarify that differences based on this feature of the cases are not clearly highlighted in the interaction model. However, if we use the variable perceptions of corruption to classify countries, we find that countries with a ‘cleaner’ environment (the United Kingdom, France) place more focus on reporting corruption cases in the international arena.

Keywords

Corruption media content analysis

Introduction

This article presents and summarizes some results of an extensive cross-national content analysis of media coverage of corruption. The study provides insights about differences between countries regarding their media system, political background and the representations of corruption. The authors examined a sample of articles published in France, Hungary, Italy, Latvia, Romania, Slovakia and the United Kingdom. The particular focus of this article is on describing the main characteristics of the represented corruption cases in the articles, including the event arena, the agent and the client of the corruption transaction and the type of interaction between them. We apply a classification method to examine the specific patterns, or interaction types, of these corruption cases by country. The typologies of corruption were established at the country level using the aforementioned characteristics; the interaction types are shaped by local specificities, just as news reporting practices are.

The main motivation for the analysis was to explore the differences and similarities between media representation practices in selected countries from a cross-national comparative perspective. Consequently, we did not formulate and test hypotheses, however, we had some expectations regarding the outcomes of the investigation. For instance, we expected that new European Union (EU) members would have different reporting patterns to those of old ones. Also, we made the assumption that there would be differences between the articles from different countries in terms of the extent to which they deal with institutionalized types of corruption.

This article is a result of the ANTICORRP research project. The extensive computer-assisted content analysis (CACA) and human-assisted content analysis (HACA) conducted therein investigated the media coverage of corruption in four news outlets in each of the seven countries involved in the research (France, the United Kingdom, Hungary, Italy, Latvia, Romania and Slovakia) over a period of 10 years, from 2004 to 2013. The resulting database included 183,491 articles¹ in total and was analysed using CACA, while a representative sample containing 12,742 articles was analysed with the assistance of human coders (HACA)² (Mancini, 2016b). HACA facilitates deeper understanding of the results of CACA, although the findings in this case were largely corroborative (Mancini, 2016b). HACA was used to identify in detail the main topics in the coverage of corruption, the nature of coverage, the causes and consequences of corruption, the type(s) of actors involved in corruption, the type of corruption interaction, the object of exchange and the point of view of the journalist, as could be reconstructed from the articles under analysis (Marchetti, 2016).

Thus far, a limited number of studies have been undertaken to examine how the media deals with corruption cases in certain countries, and there is a lack of cross-national comparative analyses. Moreover, the authors experienced that the topic of the media coverage of corruption in Western democracies was underrepresented in pre-existing research. One of the few counterexamples is the analysis of Chang Sup Park (2012) that investigates how the media frames political corruption in Illinois. Park found that, even if the number of reports about political corruption scandals is considerable, local newspapers do not engage in debate about corruption, media coverage is episodic and public discourse about scandals becomes trivialized, potentially leading to the inactivity of citizens in anti-corruption initiatives.

An interesting approach of some earlier content analyses was to focus on public media. Such an inquiry was undertaken in Bosnia and Herzegovina during a project funded by the European Union called ‘Curriculum for transparency, curriculum for accountability’. It concluded that the presence of content related to corruption was rather low on mainstream news and that a deep and educational approach to the topic of corruption was missing from public broadcasts. Nigerian content analysis also highlights the social responsibility of the media as concerns promoting vigorous resistance to corruption (Fadiaro et al., 2014) but raises inquiries about the distribution of reports about corruption in the newspapers under analysis; the authors found that even if there was an increase in the coverage of corruption articles, the articles were concentrated on the inside pages of Nigerian newspapers.

Another approach to media analysis is to investigate the impact of government influence on media coverage of corruption scandals. For instance, an investigation into governmental spending on advertising in Argentina found empirical evidence for a negative correlation between government advertising and reporting on government corruption (Di Tella and Franceschelli, 2011). Moreover, legislation that restricts media freedom also can result in a decline in the number of reports about political and bureaucratic corruption, as occurred in Mexico (Stanig, 2015). Szántó et al. (2009) examined articles from eight online news portals between 2001 and 2007 in Hungary with a particular focus on the reported agents of corruption and the nature of corruption transactions.

In our study, we refer to the principal agent–client model because it provides a conceptual framework of corruption for the analysis. This approach is based on rational choice theory, so the conceptualization of corruption is closely related to utility maximizing principles (Szántó, 1999). Our HACA dataset contains information about agents (e.g. tax collectors) who are employed by principals (e.g. chief officers of tax offices) and about clients (e.g. taxpayers) who may interact with agents, the key players in corruption transactions. The data contain cases of political corruption and of non-political corruption, although most recorded cases were political corruption committed by government or state representatives. The main hypothesis behind the model is that agents become corrupt if the net benefit from being honest is smaller than the net benefit from taking part in corrupt activities. However, a client will attempt bribery if the potential net benefit of this exceeds the net benefit of failing to bribe. The act of corruption is often a transaction between the agent and the client, but the initiator of the transaction varies. For example, when the act of corruption is bribery, the client usually bribes the agent to obtain extra benefits. But the agent can also extort the client, in this case, the agent is the initiator. In the case of embezzlement, it is typically (but not always) the agent who steals from the principal (Szántó et al., 2012).

Research questions

The core focus of the study described in this article was to reveal the significant differences in the corruption cases covered by the media according to the countries under analysis. First, we aimed to identify the main characteristics of corruption cases covered in the media that represented the widest distribution of countries. Second, we were interested in constructing a typology of the cases of corruption based on their main features. Finally, we intended to highlight which kinds of cases were the most and least characteristic in the media of the selected countries based on the typology.

Two primary assumptions guided the analysis. One was that old EU members (Italy, the United Kingdom and France) might have more a stable media market which was less likely to be a tool for direct political influence. Therefore, reports about corruption cases in these countries might significantly differ. While some publications provide information about issues that are more marketable and that are appealing to the audience, others write more about local and national level politicians in order to gain or to demolish political capital. We assume that newspapers in new EU member countries are more prone to political instrumentalization, therefore report on local and national level political corruption. Important international grand corruption cases presumably are mostly reported by news outlets from the old EU member states. Accordingly, the first research question and the related sub-questions³ were as follows:

RQ1. Are there differences between the new and the old EU member countries in terms of the media representation of corruption cases by level of reported corruption?

RQ1.1. Are newspapers in the old EU member states reporting more about grand corruption cases of international interest?

Accordingly, we predicted that Italy, France and the United Kingdom would be more similar to each other compared to the other member countries. In addition, we presumed that Hungary, Romania, Latvia and Slovakia would show similarities. These differences would not only emerge because the media products are different but because in new EU member states, we assumed that corruption was highly institutionalized. Institutionalization can be observed when the corruption cases are committed by multiple actors repeatedly (Szántó et al., 2012). Thus,

RQ1.2. Are newspapers in new EU member states reporting about more institutionalized cases of corruption?

The second assumption that perceptions of corruption might differ by country was used as an indicator of the level of corruption and not as evidence of corruption per se. In this sense, Italy was considered to be more similar to the new EU members than the United Kingdom or France. Based on this hypothesis, we expected to be able to classify countries based on the perceived level of corruption. Therefore, the second research question and the connected sub-questions were

RQ2. Are there differences between countries with higher and lower perceptions of corruption.⁴

RQ2.1. Is the media coverage of corruption in countries with high or low corruption perception similar?

As the perception on corruption is used as a proxy for the level of corruption on a country, we accept that the media representation of corruption might not be entirely dependent on the level of corruption. In this sense, the second research question is less well supported.

Methodology

Since we focus on reported cases of corruption in our research, we were able to analyse only 5212 relevant articles which dealt with particular cases from the total sample of the HACA.⁵ The HACA codebook covered many issues such as type of article (news article, satirical, editorial commentary, interview etc.), the impetus for the article, the arena of the corruption event, the viewpoint and framework of the article, type of corruption, type of transaction and type of actors involved in corruption.

The following section discusses the variables related to the representation of specific corruption cases in the seven examined countries. A case of corruption most frequently occurs between one or more agents and clients (Lambsdorff, 2007; Rose-Ackerman, 1978; Szántó, 1999). Such a transaction can be characterized by the nature of the agents, clients, exchanged goods and/or whether the transaction was repeated or was a one-off event. Articles were categorized as representations of a specific corruption case when they specified the actors who had engaged in a named act of corruption. The event arena of the corruption case was determined by the arena in which the event mainly took place. This included ‘international’ when the action took place in multiple countries or when international organizations (such as EU, United Nations, World Health Organization) were involved, ‘foreign’ when it took place in a foreign country, ‘national’ when it involved actors that were of national importance or held important positions in the country where the newspaper was based. ‘Local’ level corruption involves actors with regional, municipal or provincial importance. Another variable that was considered was whether the article focused on the client (in the position of briber or extortee) or on the agent (also, as bribee or extorter). We coded the type and position of both client and agent. Type refers to whether the act of corruption was committed by a single actor, a group of actors, an institution or company, or unknown. The position variables included many categories, such as government or high state representative, local representative, policeperson, citizen, business owner, lawyer and journalist. The category ‘type of transaction’ applies to whether the corruption involved a one-off transaction or was a repeated, stable series of interactions between actors (Marchetti, 2016).

Using these variables as applied to the corruption cases described in the articles, we classified types of interaction that were characteristic in our database. Interaction types represent the co-occurrence of different variables that are characteristic of a certain corruption case. To establish these interaction types, we used Chi-square Automatic Interaction Detector (CHAID) analysis.⁶

Results were aggregated at the country level. We aimed to investigate which interaction types were characteristic of which country or countries and which were not.

Results

First, we summarize the differences between country corruption cases as represented in the media using the main characteristics of the cases. Second, we introduce the CHAID model that demonstrates the types of corruption interactions. Our database indicates the considerable differences between the analysed countries in terms of the corruption cases (Czibik et al., 2016). There are differences in the representation of the type and position of actors, the exchanged objects and the type of transactions. General conclusions can also be made about the representation of corruption cases according to the news outlets.

Concerning the representation of the agents and clients in the articles, the main emphasis on corruption as a transaction is usually on the agent (see Figure 1), mainly because the agent is more likely to be associated with a public institution, while in many cases, it is hard to know who the client is. Also, agents are more likely to be politicians, who should have a higher level of social responsibility – and thereby they mostly become the key actors in corrupt transactions – so articles about them may have greater social impact. Clients, who are more likely to be businessmen involved in grand corruption, remain in the background because they are less interesting to the audience or because of a lack of deeper journalistic investigation.

Figure 1.

Agent or client focus.

Single actor agents are more likely to be named at the local level. This might be because news outlets tend to politically instrumentalize corruption cases involving opposition politicians. However, it may be that audiences are less interested in reports about international politicians.

Local-level action in specific corruption cases is significant in the case of Latvia and Italy. In Italy, actors involved in corruption often work in groups. This may be because corruption in Italy is traditionally more institutionalized, involving groups and institutions (Sberna et al., 2015).

Similarity with type of actor involved in the corruption cases did not always entail a similar situation. For example, articles published in the United Kingdom described similar types of agents being involved in corruption to the Hungarian ones: about 40% single and 30% group actors. However, British news outlets mostly described international corruption cases and, in many aspects, are similar to French news outlets.

The closer the event arena is to the local level, the fewer the not applicable cases and the more one-off corruption transactions were reported. The more international a transaction was, the more repeated transactions were reported by the news outlets. The explanation for this may be that on the international level, corruption is represented as a more general phenomenon. Also, when a case receives international attention, it has presumably been going on for a while. One-off transactions were more likely to be represented on the local level, but this style of representation did not emerge uniformly in every country. A distinction could be made between new and old EU member countries.

The media in new EU member states reports about more one-off transactions at a local and national level, while old EU members’ media reports more about repeated transactions (see Figure 2). In old EU member states’ news outlets, there is a stronger tradition of journalism and stronger competition between media products. While new EU members’ media outlets still consider one-off transactions to be important news, French, Italian and British media report more about ‘huge’ or ‘important’ cases that make their media products more marketable.

Figure 2.

Type of original transaction by country (national and local level) in percentage.

In order to understand the differences between countries better, we constructed a typology of interaction types using the variables related to the corruption case and the actors involved in it. Through doing this, we were able to tell what types of interaction are characteristic of which countries and which interaction types are less characteristic. This tool was also helpful for comparing countries and testing similarities and differences between the new and old EU members.

The following section describes how a chi-square automatic interaction detection (CHAID) classification method was applied to the specific cases. The model examined the specific interactions between the agent and the client by country. The outcome (dependent) variable in the analysis was the examined countries. The independent variables were initially all the variables – that affect and characterize the transactions between an agent and a client – and the context, but the analysis proved that only four of them significantly affect the outcome country.

After identifying the significant variables for the analysis, in order to simplify the process, we recoded them so as to reduce the number of values:

One important variable was the event arena or where the corruption case happened. We recoded the variable to have one of two values: international or foreign or national or local event arena.

The second important variable was the type of agent. After recoding, the agent was determined to be one of two variables: single agent or non-single agent (the latter category contains both group actors and institutional actors).

The third variable considered was the type of transaction: this was either one-off or repeated.

The fourth variable was the position of the agent: this was recoded into government or state representative or other type of agent.⁷

Using the aforementioned independent variables and the countries as a dependent variable, the CHAID method revealed 15 interaction types:

LOC/SING/ONE: Event arena is local/national or missing, committed by single actor, one-off transaction.

LOC/SING/REP: Event arena is local/national or missing, committed by single actor, repeated transaction.

LOC/SING/TTM: Event arena is local/national or missing, committed by single actor, type of transaction is unidentifiable.

LOC/NSING/ONE: Event arena is local/national or missing, committed by non-single actor, one-off transaction.

LOC/NSING/REP: Event arena is local/national or missing, committed by non-single actor, repeated transaction.

LOC/NSING/TTM: Event arena is local/national or missing, committed by non-single actor, type of transaction is unidentifiable.

LOC/ACTM/ONE: Event arena is local/national or missing, committed by not identified actor, one-off transaction.

LOC/ACTM/REP: Event arena is local/national or missing, committed by unidentified actor, repeated transaction.

INT/ONE/GOV: Event arena is international/foreign, type of transaction is one-off, committed by government or state representative as agent.

INT/ONE/NGOV: Event arena is international/foreign, type of transaction is one-off, committed by non-governmental agent.

INT/REP/GOV: Event arena is international/foreign, type of transaction is repeated, committed by government or state representative as agent.

INT/REP/NGOV: Event arena is international/foreign, type of transaction is repeated, agent was non-governmental representative or missing.

INT/TTM/GOV: Event arena is international/foreign, type of transaction is unidentifiable, committed by government or state representative as agent.

INT/TTM/NGOV: Event arena is international/foreign, type of transaction is unidentifiable, committed by non-government or state representative as agent.

INT/TTM/GOVM: Event arena is international/foreign, type of transaction is unidentifiable, position of agent was missing.

For example, for the first interaction type, the method first considered only local and national level cases and cases without clear national or international affiliation (the variability of the other independent variables was lowest after this process of separation), then the cases in which the type of actor was single and then the cases which only happened once (one-off transactions). Accordingly, each row of the table represents one type of interaction.

We assessed which countries are most and least characteristic regarding these interactions. Results are summarized in Table 1 which describes the most and the least representative two countries within the types of interactions based on the proportion of articles derived from them.

Table 1.

Most and least characteristic countries for the types of interactions revealed by the CHAID algorithm.

Type of interaction	Most characteristic countries	Least characteristic countries
1. LOC/SING/ONE	Italy/Hungary	France/the United Kingdom
2. LOC/SING/REP	Italy/Hungary	France/the United Kingdom
3. LOC/SING/TTM	Italy/Latvia	The United Kingdom/Romania
4. LOC/NSING/ONE	Hungary/Slovakia	Romania/France
5. LOC/NSING/REP	Hungary/Italy	Romania/France
6. LOC/NSING/TTM	Italy/Hungary	Romania/France
7. LOC/ACTM/ONE	Latvia/Italy	France/the United Kingdom
8. LOC/ACTM/REP	Hungary/the United Kingdom	France/Romania
9. INT/ONE/GOV	Hungary/Slovakia	The United Kingdom/Romania
10. INT/ONE/NGOV	Hungary/Slovakia	Romania/Latvia
11. INT/REP/GOV	UK/Hungary	Romania/Italy/Slovakia
12. INT/REP/NGOV	UK/France	Romania/Latvia/Italy
13. INT/TTM/GOV	UK/France	Romania/Latvia
14. INT/TTM/NGOV	UK/France	Romania/Latvia
15. INT/TTM/GOVM	UK/France	Slovakia/Italy

As can be seen, Italy is the most characteristic country in the case that the event arena is not international and the case is committed by single actors, no matter whether the transaction is one-off, repeated or unidentifiable. Italy is also characteristic in the case that the event arena is local/national or missing, the case is committed by a non-single actor and when the type of transaction is unidentifiable. Moreover, Italian news outlets are not likely to include foreign and international cases that involved repeated transactions or cases where both the transaction and type of agent was missing. The reason why international cases are relatively underrepresented in the Italian media may be related to the greater presence of mafia-type organizations with complex and extended networks, influence and power compared to other EU member states (Sberna et al., 2015).

Hungary is the country most characteristic in the case of five types of interactions, including when the event arena is local or national and the case is committed by non-single (group or institution) or unidentifiable actors, regardless of whether the transaction is repeated or not. Hungary is also the most characteristic country in the case of the international or foreign event arena, when the type of transaction is one-off, regardless of whether the agent is a governmental representative or non-governmental actor. Slovakia also matches this second pattern, being the second most characteristic country in these types of interactions. Slovakia is the least characteristic country in the case of foreign and international cases that have repeated transactions and cases where both the transaction and type of agent were missing.

Latvia is the most characteristic country regarding one type of interaction: when the event arena is local/national or missing, the case was committed by an unidentified actor and involved a one-off transaction. Latvia is the least characteristic country when the event area is international or foreign and the case was committed by a non-governmental actor as agent. It is also the least characteristic country in the case of the foreign or international event area, when the type of transaction was unidentifiable.

The United Kingdom is the most representative country in the case of five types of interactions. These types of interactions include cases when the event arena is international or foreign and the type of transaction was either repeated or not possible to identify, regardless of whether the agent was a governmental or non-governmental actor. France is the second most characteristic country in four of these five types of interactions, with very similar corruption transactions. The United Kingdom is the least characteristic country when the event area is local or national and the act was committed by a single actor, regardless of the transaction type. It is also the least characteristic country when the event arena was international/foreign, where the type of transaction was one-off and the case was committed by the government or a state representative as the agent, or the agent is missing.

France is among the least characteristic countries in almost every case when the event area is local or national. The focus on foreign and international corruption in the United Kingdom and France is also pointed out by Mancini (2016a) who emphasizes that the representation of corruption is completely different between the so-called established democracies and new democracies. Also, English and the French media may attract greater interest from foreigners because these languages are well known by people around the world. In addition, using the level of the perception of corruption as a classifying factor for the selected countries, it can be concluded the media in the United Kingdom and France, countries with a relatively low perception of corruption, focus on international cases. This phenomenon highlights the special situation of Italy, a country that is an old EU member state but whose media has similar characteristics to the media of new member states. Also, the perception of corruption in Italy is similar to that of the new EU member countries.⁸

Due to the low number of cases, Romania remains an insignificant country in this regard in almost all of the types of interactions.

Table 2 demonstrates how the country-specific cases are divided between the types of interactions. Some similarities can be seen between the United Kingdom and France and also between Italy, Hungary, Slovakia and Latvia. A main difference between countries is resulting from the event arena of the articles. Regarding the dissimilarities in the representation of institutionalized corruption between old and new EU member states – captured by the type of transaction – the interaction model does not offer clear evidence, however, descriptive analyses suggest that some differences also exist here.

Table 2.

Countries and cases according to the types of interactions revealed by the CHAID algorithm.

Type of interaction	Italy	United Kingdom	France	Slovakia	Hungary	Latvia	Romania
1. LOC/SING/ONE	11.5% 158	2.1% 16	2.9% 12	15.4% 96	10.2% 123	12.5% 79	21.7% 48
2. LOC/SING/REP	13.8% 189	3.5% 26	3.9% 16	6.9% 43	8.4% 102	6.2% 39	12.2% 27
3. LOC/SING/TTM	29.3% 401	3.9% 29	9.7% 40	15.1% 94	6.4% 77	22.7% 143	17.2% 38
4. LOC/NSING/ONE	2.8% 38	(0.7%) 5	(1.0%) 4	9.3% 58	11.1% 134	6.0% 38	(1.4%) 3
5. LOC/NSING/REP	11.6% 159	5.1% 38	2.4% 10	5.3% 33	21.8% 263	8.9% 56	(2.3%) 5
6. LOC/NSING/TTM	15.1% 207	2.0% 15	3.9% 16	7.7% 48	7.5% 91	7.9% 50	(3.2%) 7
7. LOC/ACTM/ONE	5.0% 68	4.3% 32	2.9% 12	6.9% 43	4.8% 58	16.0% 101	27.1% 60
8. LOC/ACTM/REP	1.0% 14	2.8% 21	(0.5%) 2	(0.2%) 1	2.2% 27	(1.4%) 9	(2.7%) 6
9. INT/ONE/GOV	0.9% 12	(1.1%) 8	5.3% 22	5.6% 35	6.0% 72	2.7% 17	(3.6%) 8
10. INT/ONE/NGOV	(0.6%) 8	2.3% 17	2.7% 11	3.7% 23	2.0% 24	(0.3%) 2	(0.4%) 1
11. INT/REP/GOV	1.4% 19	15.7% 117	9.0% 37	3.0% 19	7.1% 86	3.5% 22	(0.4%) 1
12. INT/REP/NGOV	0.7% 10	19.3% 144	16.2% 67	4.2% 26	5.3% 64	1.6% 10	(0.4%) 1
13. INT/TTM/GOV	4.1% 56	14.6% 109	22.8% 94	9.8% 61	3.7% 45	6.3% 40	(1.4%) 3
14. INT/TTM/NGOV	1.8% 24	13.0% 97	11.9% 49	6.1% 38	2.0% 24	1.7% 11	(1.4%) 3
15. INT/TTM/GOVM	(0.4%) 6	9.6% 72	5.9% 21	(1.0%) 6	1.4% 17	2.1% 12	4.5% 10
Total	100.0% 1369	100.0% 746	100.0% 413	100.0% 624	100.0% 1207	100.0% 630	100.0% 221

Percentages and numbers of cases represent the share and number of articles according to the different types of interactions by countries; percentages in cells with low case numbers (below 10) are in parentheses; the percentages and case numbers are in bold in the cases of the three most frequent categories for each country.

Conclusion

Our first research question aimed to clarify the differences between the coverage of corruption by media outlets in our examined countries. There emerged a clear difference as to the arena in which corruption took place (local, national, international arena) but also as to the involved actor and the type of transaction.

We defined interaction types in our observed corruption cases to see what kind of interactions was characteristic of the media representations of corruption for each country. From the CHAID analysis, the following conclusions can be drawn: Italy was the most characteristic country in the case when the event arena was local or national and the case was committed by single actors or non-single actors, no matter whether the transaction was one-off, repeated or not possible to identify. The fact that corruption in Italy takes place mostly at local level involving groups and institutions confirms previous studies (Sberna et al., 2015). Hungary was most characteristic when the event arena was local or national and the case was committed by groups or institutions and unidentified actors, regardless of whether the transaction was repeated or not. Hungary was also the most characteristic country in the case of the international or foreign event arena when the type of transaction was one-off, regardless of whether the agent was a governmental representative or non-governmental actor. Slovakia was similar to Hungary in this sense, being the second most characteristic country in these types of interactions. Latvia was the most characteristic country regarding one type of interaction: when the event arena was local/national or missing, the case was committed by an unidentified actor and in the case of one-off transactions. The United Kingdom was the most representative country in all the foreign/international case-dominated types of interactions, while French news media took a very similar position to that of the British.

Our first research question proved to be partially true. Old EU members do report more on relevant, international corruption cases. Italy regarding news reporting on corruption shows more similar pattern to new EU members more than to old EU members. Italy was a characteristic country in four outcome categories where the event arena was local or national together with Hungary and Latvia. The corruption transactions presented on the Italian articles were committed either once, repeatedly or the transaction was unidentifiable. When the corruption was committed by non-single actor, the transaction was likely to be unidentifiable. Italy was not a characteristic country in any of the outcome categories that were containing international corruption cases. All those categories were dominated by articles from France and from the United Kingdom, regardless of the type of actor, transaction type and whether they were committed by governmental or non-governmental actor.

This observation also leads up to answering our research question regarding institutionalization of the corruption cases (RQ1.2). Considering the institutionalization of the corruption cases that were examined, Italy appears to be more similar to the other old EU members. Although the results for this research question were mixed, as there were no countries where the news outlets clearly focused on institutionalized cases. News outlets from new EU members (mainly from Hungary) and Italy reported more repeated transactions on a local and national level of corruption committed by both single and non-single actors as groups or institutions. At the same time, news outlets from the United Kingdom and France also often report on repeated cases but on an international level, committed by governmental or non-governmental actors.

As for the second research question and the third research sub-question (RQ2.1.), we can conclude that, taking the perception of corruption as a factor for classifying countries into consideration, the ones with a ‘cleaner’ environment (the United Kingdom and France – see Table 3) focus more on reporting corruption in the international arena. In contrast, the countries with higher corruption perception levels tend to pay more attention to local cases. In addition, this group is more mixed regarding the further observed characteristics of the corruption cases apart from the event arena than the United Kingdom and France.

Table 3.

Corruption perception in the selected countries in 2013.

Country	TI CPI^a	Control of Corruption indicator of the World Bank^b
United Kingdom	76	1.69
France	71	1.31
Latvia	53	0.27
Slovakia	47	0.06
Hungary	54	0.29
Romania	43	−0.18
Italy	43	−0.04

TI CPI: Transparency International Corruption Perception Index, where 100 means the ‘cleanest’ country and 0 means the most corrupt. ‘The Corruption Perceptions Index ranks countries and territories based on how corrupt their public sector is perceived to be. The country or territory score indicates the perceived level of public sector corruption on a scale of 0–100, where 0 means that a country is perceived as highly corrupt and 100 means it is perceived as very clean’. (Corruption Perception Index, 2013)

Control of Corruption captures perceptions of the extent to which public power is exercised for private gain, the scores are ranging from approximately −2.5 to 2.5. For more details, see: http://data.worldbank.org/data-catalog/worldwide-governance-indicators.

In addition, a notable result of the analysis is the empirical evidence for the agent-focused nature of the media in all of the examined countries. Lambsdorff (2007) pointed out that the criminal law and the penalties usually punish the agents of the corrupt transaction, while clients are less affected.

Regarding possible further research directions, the assessment of the potential effects of professional investigative journalism on reporting corruption in different countries would be a rewarding topic. In addition, the results presented in this study should be put in a broader context of cross-national media analyses (e.g. Hallin and Mancini, 2004) in order to investigate whether the differences between the practices of the analysed countries on reporting corruption could be correlated with other properties of their media.

Footnotes

Acknowledgements

We would like to thank Paolo Mancini and the reviewer of this Issue for their comments on the article.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The ANTICORRP project which this publication is based on was funded by the European Commission’s Seventh Framework Programme, Grant Agreement No. 290529.

Notes

References

Corruption Perception Index (2013) Corruption Perception Index 2013. Available at: http://www.transparency.org/cpi2013/results (accessed 26 January 2017).

Czibik

Hajdu

Pápay

et al . (2016) Type of actors, objects of exchange, type of original transactions in the coverage of corruption. Available at: http://unipub.lib.uni-corvinus.hu/2473/1/D6_1_13.pdf.

Di Tella

Franceschelli

(2011) Government advertising and media coverage of corruption scandals. American Economic Journal: Applied Economics 3(4): 119–151.

Fadiaro

Aminu

(2014) Coverage of corruption news by major newspapers in Nigeria. New Media and Mass Communication 24: 53–59.

Hallin

Mancini

(2004) Comparing Media Systems: Three Models of Media and Politics. Cambridge: Cambridge University Press.

Hámori

(2001) A CHAID alapú döntési fák jellemzői [The Characteristics of the CHAID-based Decision Trees]. Statisztikai Szemle 79(8): 703–710.

Kass

(1980) An exploratory technique for investigating large quantities of categorical data. Applied Statistics 29: 119–127.

Lambsdorff

(2007) The Institutional Economics of Corruption and Reform: Theory, Evidence and Policy. Cambridge: Cambridge University Press.

Mancini

(2016a) A comparative research on the print press coverage of corruption. Available at: http://anticorrp.eu/wp-content/uploads/2016/09/D6.1.2.pdf.

10.

Mancini

(2016b) Extensive content analysis study on the coverage of stories on corruption: Executive summary. Available at: http://anticorrp.eu/wp-content/uploads/2016/09/D6.1.1-Executive-Summary.pdf.

11.

Mancini

Mazzoni

Cornia

et al . (2016) Representations of corruption in the British, French and Italian press. International Journal of Press/Politics 22: 67–91.

12.

Marchetti

(2016) Methodology of the human-assisted content analysis. Available at: http://anticorrp.eu/wp-content/uploads/2016/09/D6.1.11.pdf.

13.

Park

(2012) How the media frame political corruption: Episodic and thematic frame stories found in Illinois newspapers. Available at: http://paulsimoninstitute.siu.edu/_common/documents/whats-in-the-water/water-illinois/park.pdf.

14.

Rose-Ackerman

(1978) Corruption: A Study in Political Economy. New Haven, CT: Academic Press.

15.

Sberna

Vannucci

Szántó (2015) Integrated report anti-corruption policies revisited (Vol. Work Package: WP9 Organised crime and impact on vulnerable groups). European University Institute. Available at: http://anticorrp.eu/wp-content/uploads/2015/04/D9.1_Integrated-Report_final1.pdf.

16.

Stanig

(2015) Regulation of speech and media coverage of corruption: An empirical analysis of the Mexican press. American Journal of Political Science 59: 175–193.

17.

Szántó

(1999) Principals, agents, and clients: Review of the modern concept of corruption. Innovation: The European Journal of Social Science Research 12(4): 629–632.

18.

Szántó

Tóth

Varga

(2009) Korrupciógyanús esetek a magyar médiában 2001–2007 között on-line hírforrások tartalomelemzése [Corruption Cases in the Hungarian Media between 2001–2007 Content Analysis of On-line News Outlets]. Available at: http://unipub.lib.uni-corvinus.hu/95/1/korrupcio_2009_mediaelemzes_2001_07_090310.pdf.

19.

Szántó

Tóth

Varga

(2012) The social and institutional structure of corruption: Some typical network configurations of corruption transactions in Hungary. In: Vedres

Scotti

(eds) Networks in Social Policy Problems. Cambridge: Cambridge University Press, pp. 156–176.

20.

Wedel

Kamakura

(2012) Market Segmentation: Conceptual and Methodological Foundations (Vol. 8). London: Springer.

21.

Yap

Ong

Husain

NHM

(2011) Using data mining to improve assessment of credit worthiness via credit scoring models. Expert Systems with Applications 38: 13274–13283.