Revista Latina de Comunicación Social # 071 – Pages 160 to 178 [Research] [Funded] | DOI: 10.4185/RLCS-2016-1089en | ISSN 1138-5820 | Year 2016 How to cite this article in bibliographies / References MR Berganza Conde, C Arcila Calderón, R de Miguel Pascual (2016): “Negativity in Political News of Spanish News Media”. Revista Latina de Comunicación Social, 71, pp. 160 to 178. http://www.revistalatinacs.org/071/paper/1089/09en.html DOI: 10.4185/RLCS-2016-1089en Negativity in the Political News in the Spanish News Media RM Berganza Conde [CV] Professor. Rey Juan Carlos University (Spain). rosa.berganza@urjc.es [ORCID http://orcid.org/0000-0001-9862-2550] [ https://scholar.google.es/citations?user=1I26TO8AAAAJ&hl=es] C Arcila Calderón [CV] Lecturer. University of Salamnca (Spain). carlos.arcila@urjc.es [ORCID http://orcid.org/0000-0002-2636-2849] [ https://scholar.google.es/citations?user=5IrxYaAAAAAJ&hl=es] R Miguel Pascual [CV] Associate Professor. Rey Juan Carlos University (Spain). roberto.demiguel@urjs.es [ORCID http://orcid.org/0000-0001-8911-8272] [ https://scholar.google.es/citations?user=bXKkrZMAAAAJ&hl=es] Abstracts Introduction: Previous studies link higher degree of negativity in the news to cynicism and indifference of citizens towards politics and politicians (theory of Spiral of Cynicism). Methods: We content-analyzed 563 news pieces of 10 Spanish media (3 newspapers, 2 TV stations and 5 online media). The degree of negativity was measured through a negativity index, conformed by indicators of tonality, conflict, incapability and unfavorability, proposed by Leanguer, Esser and Berganza. (2012). Results and conclusions: Data confirmed that Spanish media depict political news in a negative way, although levels of negativity were lower than those of other European countries. A higher degree of negativity was found in items related to controversial topics, produced in formats of higher implication, and published by private, print or traditional media. Keywords Political news; citizens/citizenship; participation; indifference; negativity. http://www.revistalatinacs.org/071/paper/1089/09en.html Página 160 Revista Latina de Comunicación Social # 071 – Pages 160 to 178 [Research] [Funded] | DOI: 10.4185/RLCS-2016-1089en | ISSN 1138-5820 | Year 2016 Contents 1. Introduction. 2. The influence of political news on citizenship. 3. Negativity in the news. 4. Method. 5. Findings. 6. Discussion and conclusion. 7. References. Translation by Frutos Miranda Traductores 1. Introduction The main sources through which citizens access the events occurring within their political environment are the media and conversations held with people around them. There is an extensive bibliography showing a media tendency to present political events and political actors in negative terms in Spain ( umg rtner h qu s on font 2005) and many other Western countries (Schudson, 1999; Farnsworth & Lichter, 2011; Patterson, 1994; Sabato, 1991; Zaller, 1999; de Vreese, Esser & Hopmann, 2016). In this regard, information from the media acts as an intermediary, and it is essential for the health of democracy. Hence, the way political information is depicted can have an impact on how citizens perceive democratic institutions, thus contributing to citizens‟ indifference tow rds politics nd the politic l cl ss. Indeed, within Western democracies such as Spain (CIS, 2013, 2009) and generally in Europe (Eurobarometer, 2010. 2009; Mair, 2006), there is a growing cynicism among citizens towards politics and political institutions. This translates into a decline in political participation (and therefore into a lesser desire to vote, amongst other manifestations), an increasing distrust of politicians and political institutions as well as into a low evaluation of the political class and the institutions. This issue poses a significant challenge for democracy, which cannot be deprived of part of its essence: citizens‟ p rticip tion and trust in the political institutions (Mishler & Rose, 2001; Moy, Pfau, & Kahlor, 1999; Norris, 1999; Torcal & Montero, 2006). There is no general consensus on how to measure negativity in the news (Kleinnijenhuis, 2008; Ridout & Franz, 2008) or on which specific conditioning factors lead to variations in the degrees of negativity. However, there is theoretical (Cappella & Jamieson, 1997) and empirical evidence (Bennett, 2009; Patterson, 1994, 2002; Sabato, 1991; Moy & Pfau, 2000; Rozell, 1996) linking the presence of neg tive inform tion with decline in citizens‟ politic l involvement. The m in purpose of this work is to ascertain to what extent the characteristics of the medium and of the very news piece influence the levels of negativity with which political information is depicted. To that end, we use as reference the Spanish media publications. 2. The influence of political news on citizenship The influence of journalistic information on citizens has been a subject of study since the 1970s. For instance, works on agenda-setting by McCombs & Shaw (1972) have become classics in their field, showing a correlation between the most mentioned issues in the media and those that citizens worry most about. Works on framing (see, for instance, De Vreese, 2004b; and Berganza, 2009), establishing significant relationships between how information is depicted and how people think about it, are also noteworthy. http://www.revistalatinacs.org/071/paper/1089/09en.html Página 161 Revista Latina de Comunicación Social # 071 – Pages 160 to 178 [Research] [Funded] | DOI: 10.4185/RLCS-2016-1089en | ISSN 1138-5820 | Year 2016 In the last years there is an increase in empirical research focusing on the effects of political information. There are many studies which have stemmed from the now classic work by Cappella & Jamieson (1997) on the spiral of cynicism (Berganza, 2008; Adriaansen et al., 2011 y 2012; Brants et al., 2010; De Vreese, 2004a, 2005; and Elenbaas et al., 2008, among many others). Therein the uthors point out th t the decline in citizens‟ politic l p rticip tion nd their neg tive ev lu tion of the political class and political institutions is due to the persistent negativity in political news depicted by the media. This tendency towards negative coverage of politics has been proven to exist both in the United States and in Western Europe (Brandenberg, 2005; Lengauer, Esser, and Berganza, 2012; Patterson, 1994). Furthermore, another research conducted by journalists from four different countries (Denmark, Germany, Great Britain and Spain) revealed that relationships between political actors and journalists (political pressures or the way political communication departments work) also lead to “cynicism” ( p thy and indifference) among media professionals. Additionally, the forementioned rese rch evidenced th t such “cynicism” c n show through the politic l inform tion elaborated by these media professionals, which is subsequently delivered to the citizenship (van Dalen et al., 2011). The issue whether other personal factors play a role in the said process takes us to the following central tenet: motivated political reasoning, giving rise to political disaffection or cynical approaches to politics, is a dual process (Lodge & Taber, 2000; Taber & Lodge, 2001; Taber, Lodge, & Glather, 2001). In this regard, it is worth mentioning, first, that motivated political reasoning theoretically depends on the so-c lled “hot cognition” hypothesis (Abelson 1963) which holds that all sociopolitical concepts are affectively charged (Bargh, 1994, 1997; Fazio & Williams, 1986; Fazio, Sanbonmatsu, Powell & Kardes, 1986; Fiske, 1982; Lodge & Stroh, 1993; Lodge, McGraw & Stroh, 1993; McGraw, Lodge & Stroh, 1990; Morris, Squires Taber & Lodge, 2003). Hence, political leaders, social groups, topics, symbols and ideas about which we think, which have been cognitively appraised in the past, carry an affective charge (either positive or negative), and this feeling is directly tied to the evaluated concept in the long-term memory. In this regard, the attitudinal process involves understanding the relationship between these cognitive circumstances and the feelings triggered by exposure to political information currently disseminated in Spain. 3. Negativity in the news The research carried out so far shows that there is no single conclusion to be drawn about the impact of negative news stories on citizens. In fact, there are various studies concluding that controversial or confrontational news can stimulate mobilization and contribute to healthy skepticism (De Vreese & Tobiasen, 2007; Freedman & Goldstein, 1999; Norris, 2000; Schuck, Vliegenthart & De Vreese, 2014; Weintraub & Pinkleton, 1995). The absence of agreement among scholarly works can stem, to some extent, from the lack of consensus in the measurement of negativity levels in the news. However, the ambiguity of the concept (Kleinnijenhuis, 2008) has been replaced by a greater operationalization of the components that make up the construct (Lengauer, Esser & Berganza, 2012; Esser, Engesser, Matthes & Berganza, 2016). http://www.revistalatinacs.org/071/paper/1089/09en.html Página 162 Revista Latina de Comunicación Social # 071 – Pages 160 to 178 [Research] [Funded] | DOI: 10.4185/RLCS-2016-1089en | ISSN 1138-5820 | Year 2016 Based on a systematic review of the studies in this field, Lengauer et al. (2012) specifically bring forward that there are 5 different components within negativity in the news: overall tone, conflictcenteredness, incapability, actor-related negativity and pessimistic outlook. These key components allow for creating a valid and reliable index to estimate the level of negativity of a news story, thus enabling comparisons between studies, countries and cases under the same measure. With a common measurement framework, and concerning political news, the question that arises is the following: RQ1. What are the levels of negativity in the political news in the Spanish media and what are the conditioning factors leading to variations in the negativity index? Esser et al. (2016) point out that these factors can be found at three different levels: a micro-level, related to the characteristics of journalistic information itself; a meso-level, referred to the medium characteristics; and a macro-level t which the country‟s medi nd politic l system re t ken into account. Provided that our study addresses a national case, we will only deal with the first two levels, leaving the third level for comparative studies. On the one hand, at the micro-level we have found empirical evidence showing that journalistic messages are event-driven. Therefore, they are contingent on real-world conditions (Peter, 2003). For instance, we see that there are topics that tend to be more negative, such as unemployment (Garz, 2014) or generally topics such as insecurity, immigration or the functioning of democracy, which since they are inherently attached to social polarization and confrontation, tend to be depicted with higher levels of negativity (Esser et al., 2016). Consequently, we put forward the following hypothesis: H1. Negativity in political news is related with the characteristics of the news piece itself Moreover, we suggest that in political news in the Spanish media: H1.1 Negativity levels are higher in political news that include inherently negative topics (crime, immigration, partisan conflicts, political controversies or corruption) H1.2 Negativity levels are lower in political news elaborated in a news story format compared to other genres such as reports, interviews and opinion journalism On the other hand, at the meso-level, i.e. having regard to the types of media organization, we found that the medium characteristics can also be conditioning factors of the negativity levels (Takens, Van Atteveldt, Van Hoof, & Kleinnijenhuis, 2013; Albæk, van Dalen, Jebril, & De Vreese, 2014; Strömbäck, 2008, Casero-Ripollés et al., 2014). For example, in the online media we have found higher levels of negativity compared to those found in print media (Quandt, 2008). Furthermore, previous studies have confirmed that journalists who work for public media are less critical of political institutions (Hanitzsch & Berganza, 2012). As a result, we suppose there is a difference between the levels of negativity compared to private commercial media (de Vreese, Banducci, Semetko, & Boomgaarden, 2006). The foregoing leads to the following hypothesis: H2. Negativity levels in political news are conditioned by the medium characteristics Moreover, we suggest that in political news in the Spanish media: H2.1 Negativity levels are higher in the news published by the press than in television http://www.revistalatinacs.org/071/paper/1089/09en.html Página 163 Revista Latina de Comunicación Social # 071 – Pages 160 to 178 [Research] [Funded] | DOI: 10.4185/RLCS-2016-1089en | ISSN 1138-5820 | Year 2016 H2.2 Negativity levels are lower in news published in traditional sources (press or TV) than in news published in the online version of the relevant medium H2.3 Negativity levels are higher in news published in private media than in public media 4. Method 4.1. Sample and procedure The news published in the Spanish media were examined by analyzing quantitative information (Krippendorff, 2004; Neuendorf, 2002). This analysis was conducted with the support of a codebook that had been previously discussed and standardized by the Network of European Political Communication Scholars (NEPOCS), which had conducted a similar study in 15 other countries (De Vreese et al., Eds., in press). This codebook sets forth some of the most frequent key concepts in political communication research (see, in this regard, the special issue of the scientific journal Journalism 13 (2) of 2012 published by NEPOCS). Additionally, the codebook provides for the procedure to sample the units of analysis, which ensures both structural homogeneity and reliability of the data from the Spanish context, so they can be compared to the remaining countries in the study. Moreover, after some previous training, two independent coders analyzed 563 news pieces connected with national politics from three newspapers, two daily TV news and their respective online editions (chart 1). Regarding the newspapers, the sample gathered units of analysis from the two national up market headers, one from the center-left of the political spectrum and the other from the center-right (El País and El Mundo), and from a third mass market header (20 Minutos). The units of analysis from the TV information were obtained both from the public television show with the highest audience ratings (TV news, Telediario, from TVE) and from the daily TV news with the highest audience among mainstream TV stations (Tele 5). As for the online media, the sample was m de up of units selected from those newsp pers‟ nd TV st tions‟ websites. A composite fortnight sampling method was used for a total of 14 days between 15 April and 15 July 2012. As opposed to the composite week, the composite fortnight has the advantage of not distorting the sample when special events occur (Riffe et al., 2005). An ideal number of 5 news stories in the press and on TV was randomly chosen, and 3 news stories were chosen from the online versions of these media. Where a specific medium had not published or broadcasted any contents on the chosen date, these were replaced through a simple random sampling of the contents published by the same medium on the previous or on the following day. Type of medium Medium n % El País 71 20Minutos 71 12.6% Press 12.6% Press elpais.com 69 Tele5, TV news, 9 pm-9:30 pm 67 El Mundo 67 Up-market press Mass-market press) 12.3% Press online version Up-market press Mass-market TV 11.9% TV TVE1, Evening TV news, 9 pm-9:30 pm 61 11.9% Press 10.8% TV online version www.telecinco.es/informativos 9.1% 51 Medium sector TV online version http://www.revistalatinacs.org/071/paper/1089/09en.html Up-market press Public TV Commercial TV Página 164 Revista Latina de Comunicación Social # 071 – Pages 160 to 178 [Research] [Funded] | DOI: 10.4185/RLCS-2016-1089en | ISSN 1138-5820 | Year 2016 www.20minutos.es 37 6.6% Press online version Commercial TV www.rtve.es/alacarta/videos/telediario 35 6.2% TV online version www.elmundo.es 34 6% Press online version Up-market press TOTAL Public TV 563 100% Chart 1. Media and analyzed information Moreover, the websites of the chosen media were downloaded between 2 pm and 4 pm on the relevant dates with the support of the HTTrack tool.[1] Particular attention was paid to the correct home page (or start page), since this is the most specific site regarding the offline format. For inst nce reg rding the TV st tion‟s website the chosen st rting point w s not the st tion‟s website but rather the website that takes you directly to the news show. We only considered the full news stories or those th t were summ rized within the medium‟s homep ge. 4.2. Measures In order to determine the dependent variable, we designed a negativity index in the news (-1= positive; 0= neutral/balanced; 1= negative), based on the work by Lengauer et al. (2012). This index was created on the basis of the following indicators: 1. Tonality (-1= positive; 0= neutral/balanced; 1= negative). Indications of negative tonality are the framing of the story as failure, crisis, frustration or disappointment; whereas indications of positive tonality refer to depictions of political success, achievements, improvements, hopes or gains. 2. Conflict (-1= consensus centered; 0= neutral/balanced; 1= conflict centered). Conflictcenteredness refers to depictions of disputes, disagreements or controversy; whereas the consensus dimension refers to depictions of accordance, agreement and willingness to cooperate. 3. Incapability (-1= capability; 0= neutral/balanced; 1= incapability). Incapability refers to the depiction of critique or accusation of incapability or incompetence; whereas the competence dimension refers to depictions of capability or commendation in the news piece. 4. Unfavorability (-1= favorable; 0= neutral/balanced; 1= unfavorable). Unfavorability refers to the opinions expressed towards the actors depicted in the news piece from their perspective; whereas favorability refers to the positive or favorable impressions about those actors from their point of view. We have used this variable to characterize the 5 most important actors included in the piece. In order to verify the unidimensionality of the construct, an Exploratory Factor Analysis (EFA) was conducted. The indicators were analyzed using Varimax orthogonal rotation and the EFA (KMO= 0.701 p<0.000) suggested l tent structure of one component (λ = 2.06) expl ining 51.43% of the variance for the set of 4 items. The EFA allows for identifying underlying factors in a number of indicators, thus allowing for verifying the validity of the construct in a measurement (Pérez-Gil, h cón Moreno 2000; M cí 2010; Ig rtu 2006). In ddition the construct‟s internal consistency w s c lcul ted (it w s found α of ronb ch= 0.667) which suggested n dequ te reliability (Cronbach, 1951; Hayes, 2005). http://www.revistalatinacs.org/071/paper/1089/09en.html Página 165 Revista Latina de Comunicación Social # 071 – Pages 160 to 178 [Research] [Funded] | DOI: 10.4185/RLCS-2016-1089en | ISSN 1138-5820 | Year 2016 Furthermore, the following independent variables were measured: Micro-level:  Topic of the news story 1= Economics (macro and micro); 2= Functioning of democracy, civil rights and freedoms; 3= Employment, immigration and social affairs; 4= Education, culture and media; 5= Justice (insecurity and crime); 6= National security, defense and international affairs; 7= Politics; 8= Others  Journalistic genre 1= News story; 2= Report; 3= Opinion journalism and other formats Meso-level:  Type of medium 1=TV ; 2=Press; 3=TV online version; 4=Press online version  Medium sector 1=Reference press (up market); 2=Commercial press (mass market); 3=Public TV; 4=Commercial TV Indicator Negativity Tonality Conflict Incapability Unfavorability Micro-level Topic of the news story Journalistic genre Meso-level Type of medium Medium sector Krippendorff’s alpha Percentage of agreement 0.97 0.95 0.93 0.95 97.3 94.5 92.4 94 0.97 0.99 96.8 98.3 1.00 1.00 100 100 Chart 2. Results of the intercoder reliability test A prelimin ry test w s performed to determine the me surement‟s instrument reli bility using n intercoder reliability test. To that end, a random subsample (n=30) of redactional units from the final study s mple w s t ken. The degree of greement w s ssessed using Krippendorff‟s lph on the basis of the recordings performed by the two coders. A well-accepted interpretation of this coefficient is that values of at least 0.60 (Neuendorf, 2002) or 0.70 (Hayes & Krippendorff, 2007) are reliable. Chart 2 shows the results of this test for the study categories, as well as the percentage of agreement between the coders. 4.3 Analysis A preliminary analysis of the resulting data file was carried out following the recording and codification process. The aim of this analysis was to find possible errors in the codification process. Once the data file had been cleaned, we conducted an inductive-exploratory statistical analysis using http://www.revistalatinacs.org/071/paper/1089/09en.html Página 166 Revista Latina de Comunicación Social # 071 – Pages 160 to 178 [Research] [Funded] | DOI: 10.4185/RLCS-2016-1089en | ISSN 1138-5820 | Year 2016 SPSS (version 21). To the end of answering the main research question, we carried out a descriptive analysis of the data (central parameters). To test the hypotheses contained herein, bivariate and multivariate inferential analyses were conducted, including correlations, average difference tests and regressions. 5. Results Out of the total analyzed news pieces (n=563), most of them were news stories (85.2%), followed by reports (6.3%) and finally there was opinion journalism or other formats (8.5%), which shows that most political contents in the Spanish media are released through information genres. The specific topics covered include the following: economics (macro and micro) (44.4%), politics (10.15%), justice (insecurity and crime) (89.6%), employment, immigration and social affairs (8.7%), national security, defense and international affairs (6.7%), functioning of democracy, civil rights and freedoms (6.2%), education, culture and media (4.6%) and others (9.6%). Negativity index Tonality Conflict Incapability Unfavorability n M SD M SD M M El País 71 0.31 0.32 0.48 0.56 0.48 0.58 0.11 0.36 0.15 0.32 El Mundo 67 0.30 0.40 0.39 0.65 0.55 0.50 0.21 0.51 0.06 0.45 El País (online) 69 0.25 0.31 0.38 0.55 0.35 0.59 0.14 0.39 0.14 0.26 20Minutos 71 0.24 0.30 0.25 0.50 0.48 0.58 0.10 0.30 0.12 0.32 Tele5 67 0.16 0.34 0.25 0.47 0.16 0.67 0.12 0.44 0.12 0.26 TVE1 61 0.11 0.32 0.13 0.46 0.20 0.54 0.07 0.48 0.06 0.22 Tele5 (online) 51 0.07 0.31 0.25 0.52 0.04 0.60 -0.02 0.37 0.00 0.16 20Minutos (online) 37 0.05 0.30 -0.03 0.44 0.16 0.73 0.05 0.23 0.03 0.30 El Mundo (online) 34 -0.01 0.30 0.00 0.60 0.03 0.67 0.00 0.25 -0.05 0.30 TVE1 (online) 35 -0.03 0.36 0.00 0.54 -0.03 0.71 -0.11 0.40 0.04 0.19 Total 563 0.18 0.35 0.25 0.55 0.28 0.30 SD 0.63 0.09 SD M 0.40 0.08 SD Chart 3. Negativity indicators per medium Note: -1=Positive, 0=Neutral/Balanced, 1=Negative As can be seen in chart 3, the studied news pieces tend to be more negative than positive, although the negativity index is still weak (M=0.18, SD=0.35). In fact, the media depicted their political information in a neutral manner (Tele5 online, 20Minutos online, El Mundo online and TV1 online) or in a negative manner (El País, El Mundo, El País online, 20Minutos, Tele5 and TVE1). The components with the greatest weight in the negativity assessments were the presence of conflict within the piece (M=0.28, SD=0.63) and the use of a negative tone (M=0.25, SD=0.55), i.e. those elements related to the usage of language to frame a story and the mise-en-scène of the lack of consensus in the news, respectively. With the aim of verifying whether the negativity indexes were related with the news piece characteristics (micro-level), one-factor ANOVA (ONWEWAY) tests were performed to compare any existing differences between the averages for each category (H1). Moreover, we found http://www.revistalatinacs.org/071/paper/1089/09en.html Página 167 Revista Latina de Comunicación Social # 071 – Pages 160 to 178 [Research] [Funded] | DOI: 10.4185/RLCS-2016-1089en | ISSN 1138-5820 | Year 2016 significant differences between the negativity indexes in each of the topics of the analyzed pieces; Welch’s F (7, 147. 818) = 2.601, p < 0.05 (figure 1). We can see that news pieces on functioning of democracy, civil rights and freedoms (M=0.25, SD=0.25), employment, immigration and social affairs (M=0.23, SD=0.33), economics (M=0.20. SD=0.39), justice (M=0.18, SD=0.25) and politics (M=0.16, SD=0.39), obtained higher negativity scores (>0.1). However, news pieces on education, culture and media (M=0.09, SD=0.22), national security, defense and international affairs (M=0.07, SD=0.28) and, in general, on other topics (M=0.1, SD=0.27), had more neutral scores or scores close to zero. When contrasting these two topic groups, we noticed that the existing differences were significant, t(179.683)=3.579, p=0.000. This evidences that discussions in the media on topics such as immigration, economics and employment (inherently more negative) tend to be depicted in the media more negatively than those which are not inherently negative within society. This finding validates hypothesis H1.1. Figure 1. Negativity index depending on the topic of the political news story. Regarding the genre of the news piece (figure 2), we noticed that news stories (M=0.16, SD=0.33) obtained lower levels of negativity than reports (M=0.29, SD=0.32) or opinion journalism pieces and other pieces (M=0.25, SD=0.5). These differences were statistically significant: Welch’s F (2, 58.365) = 3.373, p < 0.05. Moreover, when comparing the news stories negativity scores to those of other genres and opinion journalism, we noticed that news stories reached significantly lower levels of negativity, t(94.564)=-2.378, p<0.05. The foregoing empirically supports H1.2, and validates our initial assertion, relating to the micro-level, where we assured that negativity in political information was related with the characteristics of the piece itself. The rules of journalistic writing allow for higher levels of interpretation and personal appreciations in opinion journalism than in news stories. This can explain the lower levels of negativity found in the latter genre. http://www.revistalatinacs.org/071/paper/1089/09en.html Página 168 Revista Latina de Comunicación Social # 071 – Pages 160 to 178 [Research] [Funded] | DOI: 10.4185/RLCS-2016-1089en | ISSN 1138-5820 | Year 2016 Figure 2. Negativity index depending on the genre of the political news story. Similarly, in order to find out if these negativity levels were related with the characteristics of the medium (H2), i.e. at the meso-level, we performed one-factor ANOVA (ONWEWAY) tests to verify if the existing differences between the averages in the different types of media were significant. Concerning the type of medium (figure 3), we noticed that the existing differences in the negativity levels in the press (M=0.28, SD=0.34), TV (M=0.14, SD=0.33), the online version of the newspaper (M=0.14, SD=0.32) and the online version of the TV show (M=0.03, SD=0.34) are statistically significant, F(3, 559) = 13.524, p < 0.000. Moreover, we found that the levels of negativity in the press and in the website version thereof are significantly higher than those of TV and its website version, t(559)=4.183, p<0.000. The foregoing empirically supports H2.1. Figure 3. Negativity index, depending on the type of medium. http://www.revistalatinacs.org/071/paper/1089/09en.html Página 169 Revista Latina de Comunicación Social # 071 – Pages 160 to 178 [Research] [Funded] | DOI: 10.4185/RLCS-2016-1089en | ISSN 1138-5820 | Year 2016 When comparing the negativity scores of the traditional media (press and TV) to those of their online versions, we found that the negativity levels were significantly lower in the website, t(559)=-4.308, p<0.000. H2.2 is not empirically supported by this finding. In fact, it confirms the complete opposite of what is proposed in such hypothesis, according to which the online media had higher levels of negativity. Notwithstanding the foregoing, we can assure that the type of medium is a variable that conditions the politic l news‟ levels of neg tivity in the Spanish media. The online versions of the media tend to be more concise. This could be the reason why they leave less room for interpretation and personal appreciations. Figure 4. Negativity index depending on the medium sector. Furthermore, when inquiring about the medium sector (chart 4), data show that negativity levels in the press traditionally regarded as reference press or up market (M=0.25, SD=0.35), the commercial or mass market press (M=0.18, SD=0.31), commercial TV (M=0.12, SD=0.33) and Public TV (M=0.07, SD=0.34) differ significantly F(3, 559) = 7,718, p < 0.000. Moreover, we verified that private media showed higher levels of negativity than public media, t(559)=-2.861, p=0.000, which supports H2.3. In sum, the validation of H2.1 and H2.3, as well as the refutation of H2.2, allow for asserting that the levels of negativity of the news pieces are indeed conditioned by the characteristics of the medium (meso-level). 6. Discussion and Conclusions The results show that most of the political information published by Spanish media scores low in negativity (>0.1 and <0.5) or at best it tends to be neutral or balanced (<0.1 and >-0.1), but on http://www.revistalatinacs.org/071/paper/1089/09en.html Página 170 Revista Latina de Comunicación Social # 071 – Pages 160 to 178 [Research] [Funded] | DOI: 10.4185/RLCS-2016-1089en | ISSN 1138-5820 | Year 2016 average it is not depicted in a positive manner. This trend is consistent with a large part of the previous research (Schudson, 1999; Farnsworth & Lichter, 2011; Patterson, 1994; Sabato, 1991; Zaller, 1999; de Vreese, Esser & Hopmann, 2016), which evidences the media preference for depicting political contents in a negative way. This preference makes sense if we understand that people react more strongly and pay more attention to negative information than to positive information (Soroka & McAdams, 2015). In fact, having regard to the NEPOCS results, the Spanish negativity average in political information (M=0.17, SD= 0.35) is way below that of other European countries such as Austria (M=0.33, SD= 0.44), Sweden (M=0.32, SD= 0.34) Portugal (M=0.29, SD= 0.50), France (M=0.25, SD= 0.40), United Kingdom (M=0.23, SD= 0.32) and Greece (M=0.22, SD= 0.41), where the negativity indexes are even higher (Esser et al., 2016). Data from the Spanish media provide empirical support to assert that both the characteristics of the news story itself (H1) and those of the medium (H2) condition the negativity levels in news pieces. In the first case, our data are consistent with previous works (Peter, 2003; Garz, 2014; Esser et al., 2016) showing a relation between how the journalistic messages are produced and the facts to which such messages refer. In the Spanish media we notice that the topics depicted with the higher levels of negativity are democracy, employment, immigration, the economy, justice and politics, which are precisely the topics that in the last few years have given rise to greater social unrest (increased by the economic crisis). However, news pieces on topics with less social controversy attached to them, such as culture, education, national security and international affairs, obtained more neutral or balanced levels. This suggests that the nature of the topics is actually linked with the way they are presented to citizens. From the data we also found out that news stories show lower levels of negativity, whereas genres where there is more room for interpretation (such as reports or opinion journalism) score higher. The foregoing could evidence that, in a given genre, the greater the elaboration, interpretation or personal appreciations, the greater the negativity. Regarding the characteristics of the medium, our data also confirm that these are also related with the way the pieces are produced and presented. We noticed that there is a higher negativity level in the press (compared to TV) and in the traditional media (compared to their online versions). However, this last finding refutes the results of previous studies, which had found exactly the opposite, i.e. that the online media showed the highest levels of negativity (Quandt, 2008). Regarding the differences based on the sector to which the medium belongs, our data confirm that public media (such as TVE) tend to be less critical (Hanitzsch & Berganza, 2012), thus having lower negativity levels (Vreese, Banducci, Semetko, & Boomgaarden, 2006). In sum, our study on Spanish media shows that both press and TV (and the online versions thereof) tend to depict political information in more negative terms, or at least neutral, rather than in positive terms. Our analysis reveals that the variability of negativity indexes is conditioned by the characteristics of the actual news story and the medium. Moreover, pieces on inherently negative topics, produced in formats in which there is a greater involvement by the journalist (by adding interpretations or personal appreciations) and published in reference (up market) print media, show more negativity. The prior conclusions have theoretical as well as practical implications. First, although our results are fairly consistent with previous studies in the field regarding the existence of conditioning factors (micro and meso), more consensus is needed on the hypotheses (for instance, on whether there is more negativity in online or in traditional media) to be able to draw a theoretical framework to guide http://www.revistalatinacs.org/071/paper/1089/09en.html Página 171 Revista Latina de Comunicación Social # 071 – Pages 160 to 178 [Research] [Funded] | DOI: 10.4185/RLCS-2016-1089en | ISSN 1138-5820 | Year 2016 future research. Second, the results may suggest that further work is required to consolidate style manuals to address certain topics (for instance, employment, immigration, economics, etc.) which although they might give rise to greater social unrest, they are not inherently negative. Additionally, on the one hand, from the standpoint of the journalistic profession we must continue discussing the limitations that journalists might have in public media. On the other, we must analyze the sensationalist strategies of certain private media to attract more audience with more negative pieces. In this research we only purposefully chose a number of media that can provide general guidance on the Spanish media landscape, yet that do not allow for generalizing the results. Therefore, it is necessary for future studies to include other more representative samples in order to build a more accurate map of negativity levels in the media (even using automated tools). However, the purpose of this study, rather than accurately describing the levels of negativity in the media, was to explore the conditioning factors (at the micro and meso levels) for the variation of such negativity index upon the analysis of certain reference media. In this regard, we must continue to inquire about how the macro-level conditioning factors can affect the levels of negativity in political information (Strömbäck & Kaid, 2008; Iyengar, 2011; Blumler & Gurevitch, 2001). Similarly, to show the actual effects of the conditioning factors it is advisable, on the one hand, to make progress in longitudinal studies that allow for monitoring the evolution of negativity in the same media throughout time. On the other hand, it is also advisable to verify the causality through experimental designs in which it is possible to control other variables not contained in this study. The addition of empirical evidence, particularly using reliable negativity measures and scales, can also be subject to meta-analyses that would enable a more consistent and overarching approach to the advancements in the field. *Funded research. This study has been conducted within the scope of the research project “El impacto de la calidad informativa en el desinterés y la apatía de la ciudadanía por la política en 15 países europeos: el caso de España” (The imp ct of inform tion qu lity on political cynicism and indifference among citizens in 15 European countries: the Spanish case) (ref. CSO2013-44874-R), funded by the Ministry of Economy and Competitiveness of Spain. Dates: - Beginning of research: 2014 - End of research: 2016 Nota 1. http://www.httrack.com/page/2/ http://www.revistalatinacs.org/071/paper/1089/09en.html Página 172 Revista Latina de Comunicación Social # 071 – Pages 160 to 178 [Research] [Funded] | DOI: 10.4185/RLCS-2016-1089en | ISSN 1138-5820 | Year 2016 7. References E Albæk, A Van Dalen, N Jebril & C De Vreese (2014). Political Journalism in Comparative Perspective. New York: Cambridge University Press. R Abelson (1963). “ omputer simul tion of “hot” cognition. En S. Tomkins Computer Simulation of Personality (pp. 277–298). New York: Wiley. D. Messick (Eds.), M Adriaansen, P Van Praag & C De Vreese (2011). “Subst nce m tters. How news content c n reduce politic l cynicism.” International Journal of Public Opinion Research, 22(4), 433-457. M Adri nsen P V n Pr g de Vreese (2012). “A mixed report: The effects of str tegic nd subst ntive news content on politic l cynicism nd voting.” Communications, 37(2), 153-172. JA rgh (1994). “The four horsem n of utom ticity: Aw reness, intention, efficiency, and control in soci l cognition.” In R. Wyer T. Srull (Eds.) Handbook of social cognition: Basic processes, vol. I (pp. 1–40). Hillsdale: Lawrence Erlbaum Associates. FR umg rtner L h qu s on font (2015). “All News is Bad News: Newspaper Coverage of Politic l P rties in Sp in”. Political Communication, 32(2), pp. 268-291. LW Bennett (2009). News: The politics of illusion (8th edition). New York: Longman. R erg nz (2008). “Medios de comunic ción „espir l del cinismo‟ y desconfi nz polític : Estudio de c so de l cobertur mediátic de los comicios elector les europeos.” Zer: Revista De Estudios De Comunicación, 25, pp. 121-139. R erg nz (2009). “Fr ming the Europe n Union nd building the medi gend : the 2004 Europe n P rli ment ry elections in the Sp nish d ily press.” Journal of Political Marketing, 8(1), pp. 59-69. H r ndenberg (2005). “Politic l bi s in the Irish medi : A qu ntit tive study of c mp ign cover ge during the 2002 gener l election.” Irish Political Studies, 20, pp. 297–322. K Brants, C De Vreese, J Moller & P Van Praag (2010). “The Re l Spir l of ynicism? Symbiosis nd mistrust between politici ns nd journ lists.” International Journal of Press/ Politics, 15(1), pp. 25-40. L Cappella & K Jamieson (1997). Spiral of cynicism. The press and the public good. Oxford: Oxford University Press. CIS (2013). 2013. Avance de resultados (Estudio No. 2.993). Available at: http://datos.cis.es/pdf/Es2993mar_A.pdf http://www.revistalatinacs.org/071/paper/1089/09en.html Página 173 Revista Latina de Comunicación Social # 071 – Pages 160 to 178 [Research] [Funded] | DOI: 10.4185/RLCS-2016-1089en | ISSN 1138-5820 | Year 2016 CIS (2009). July 2013 Barometer (Spain). Ética pública y corrupción (Study No. 2.826). Available at: http://www.cis.es/cis/export/sites/default/-Archivos/Marginales/2820_2839/2826/es2826.pdf A Casero-Ripollés, J Izquierdo-Castillo, & H Doménech-Fabregat (2014). “From W tchdog to Watched Dog: Oversight and Pressures between Journalists and Politicians in the Context of Medi tiz tion”. Trípodos, 1(34), pp. 23-40. LJ ronb ch (1951). “ oefficient lph 297-334. nd the intern l structure of tests.” Psychometrika, 16, pp. De Vreese (2005). “The spir l of cynicism reconsidered: the mobilizing function of news.” European Journal of Communication, 20(3), pp. 283-301. C De Vreese (2004a). The effects of strategic news on political cynicism, issue evaluations and policy support: A two-wave experiment. Mass Communication & Society, 7(2), 191-215. De Vreese (2004b). “The effects of fr mes in politic l television news on udience perceptions of routine politic l news.” Journalism and Mass Communication Quarterly, 81(1), pp. 36-52. C De Vreese, F Esser & D Hopmann (Eds.) (in press, 2016). Where’s the Good News? Comparing Political Journalism in 16 Countries. Routledge. De Vreese M Tobi sen (2007). “ onflict nd identity. Expl ining turnout nd nti-integrationist voting in the Danish 2004 elections for the Europe n P rli ment.” Scandinavian Political Studies, 30. pp. 87-111. De Vreese SA nducci H Semetko 2004 Europe n P rli ment ry Election pp. 477–504. HG oomg rden (2006). “The News over ge of the mp ign in 25 ountries.” European Union Politics, 7(4), M Elenbaas & C De Vreese (2008). “The effects of str tegic news on politic l cynicism nd vote choice mong young voters.” Journal of Communication, 58(3), pp. 550-567. F Esser, S Engesser, J Matthes & R Berganza (in press, 2016). “Neg tivity in the news. Exploring Drivers of Bas-News Tendency.” In De Vreese F Esser D Hopm nn (Eds.) Where’s the Good News? Comparing Political Journalism in 16 Countries. Routledge. Eurobarometer (2009). Special Eurobarometer. Attitudes of Europeans towards Corruption (Estudio 325) [Report] Available at: http://ec.europa.eu/public_opinion/archives/ebs/ebs_325_en.pdf Eurobarometer (2010). Standard Eurobarometer. Public Opinion in the European Union (Estudio 73) [Report] Available at: http://ec.europa.eu/public_opinion/archives/eb/eb73/eb73_vol1_en.pdf SJ Farnsworth & RS Lichter (2011). The nightly news nightmare: Television’s coverage of US presidential elections, 1988-2008. Lanham, MD: Rowman and Littlefield. http://www.revistalatinacs.org/071/paper/1089/09en.html Página 174 Revista Latina de Comunicación Social # 071 – Pages 160 to 178 [Research] [Funded] | DOI: 10.4185/RLCS-2016-1089en | ISSN 1138-5820 | Year 2016 RH F zio Willi ms (1986). “Attitude accessibility as a moderator of the attitude-perception and attitude-beh vior rel tions: An investig tion of the 1984 presidenti l election.” Journal of Personality and Social Psychology, 51, pp. 505–514. RH Fazio, D Sanbonmatsu, M Powell & F Kardes (1986). “On the utom tic ctiv tion of ttitudes.” Journal of Personality and Social Psychology, 50. pp. 229–238. ST Fiske (1981). “Soci l cognition nd ffect.” In H rvey J. (Ed.) Cognition, social behavior, and the environment (pp. 227–264). Hillsdale: Lawrence Erlbaum Associates. P Freedm n K Goldstein (1999). “Me suring medi exposure nd the effects of neg tive c mp ign ds.” American Journal of Political Science, 43(4), pp. 1189-1208. M G rz (2014). “Good news nd b d news: evidence of medi bi s in unemployment reports.” Public Choice, 161(3-4), pp. 499-515. A Hayes (2005). Statistical Methods for Communication Science. Mahwah, NJ: Lawrence Erlbaum Associates. A H yes K Krippendorff (2007). “Answering the ll for St nd rd Reli bility Measure for oding D t .” Communication Methods and Measures, 1(1). JJ Igartua (2006). Métodos cuantitativos de investigación en comunicación. Barcelona: Bosch. J Kleinnijenhuis (2008). “Neg tivity.” In W. Donsb ch (Ed.) The International Encyclopedia of communication (pp. 3188-3192). Malden: Blackwell Publishing. K Krippendorff (2004). Content Analysis. Thousand Oaks, CA: Sage. G Lengauer, F Esser y R Berganza (2012). “Neg tivity in politic l news: A review of concepts operationalizations and key findings.” Journalism, 13(2), pp. 179-202. M Lodge S T ber (2000). “Three steps tow rd theory of motiv ted politic l re soning.” En A. Lupia, M. McCubbins, & S. Popkin (Eds.), Elements of reason: Cognition, choice, and the bounds of rationality (pp. 183–213). Cambridge: Cambridge University Press. M Lodge K McGr w P Stroh (1993). “An impression-driven model of c ndid te ev lu tion.” American Political Science Review, 87, pp. 399–419. F M cí (2010). “V lidez de los Tests y el Análisis F ctori l: Nociones Gener les.” Ciencia y Trabajo, 12(35), pp. 276-280. ME Mc ombs DL Sh w (1972). “The Agend -Setting Function of M ss Medi .” The Public Opinion Quarterly, 2, pp. 176-187. http://www.revistalatinacs.org/071/paper/1089/09en.html Página 175 Revista Latina de Comunicación Social # 071 – Pages 160 to 178 [Research] [Funded] | DOI: 10.4185/RLCS-2016-1089en | ISSN 1138-5820 | Year 2016 K McGr w M Lodge P Stroh (1990). “ ndid te ev lu tion: The effect of s lience, order, and sophistic tion.” Political Behavior, 12, pp. 41–58. P M ir (2006). “Ruling the void: The hollowing of western democr cy.” New Left Review, 42, pp. 25-51. J Morris N Squires T ber M Lodge (2003). “The ctiv tion of politic l ttitudes: Psychophysiologic l evidence for the hot cognition hypothesis.” Political Psychology, 24, pp. 727–745. W Mishler R Rose (2001). “Wh t re the origins of politic l trust? Testing institution l nd cultural theories in post-communist societies.” Comparative Political Studies, 34, pp. 30–62. P Moy & M Pfau (2000). With malice toward all? The media and public confidence in democratic institutions. Westport: Praeger. P Moy M Pf u L K hlor (1999). “Medi use nd public confidence in democr tic institutions.” Journal of Broadcasting & Electronic Media, 43(2), pp. 137-158. KA Neuendorf (2002). The Content Analysis Guidebook. Thousand Oaks, CA: Sage. Norris, P. (2000) A virtuous circle. Political communication in postindustrial societies. Cambridge: Cambridge University Press. P Norris (1999). “Introduction: The growth of critic l citizens?” In P. Norris (Ed.) Critical citizens: Global support for democratic governance (pp. 1–27). Oxford, England: Oxford, University Press. L Pérez-Gil, S Chacón & R Moreno (2000). “V lidez de constructo: el uso de nálisis f ctori l exploratorio-confirm torio p r obtener evidenci s de v lidez.” Psicothema, 12(2), pp. 442-446. LJ Sabato (1991). Feeding frenzy. New York: Free Press. A Schuck, R Vliegenthart & C De Vreese (2014). “Who's fr id of conflict? The mobilizing effect of conflict fr ming in c mp ign news.” British Journal of Political Science, 44(4). M Schudson (1999). “Soci l origins of press cynicism in portr ying politics.” American Behavioral Scientist, 42(6), pp. 998-1008. S Sorok pp. 1-22. S McAd ms (2015). “News Politics nd Neg tivity.” Political Communication, 32(1), TE Patterson (2000). Doing well and doing good: How soft news and critical journalism are shrinking the news audience and weakening democracy. Cambridge, MA: Harvard University, Joan Shorenstein Center on the Press, Politics and Public Policy. TE Patterson (1994). Out of order. New York: Vintage. http://www.revistalatinacs.org/071/paper/1089/09en.html Página 176 Revista Latina de Comunicación Social # 071 – Pages 160 to 178 [Research] [Funded] | DOI: 10.4185/RLCS-2016-1089en | ISSN 1138-5820 | Year 2016 J Peter (2003). “ ountry ch r cteristics s contingent conditions of gend setting: The moderating influence of pol rized elite opinion.” Communication Research, 30(6), pp. 683–712. T Qu ndt (2008). “No news on the world wide web? A comp r tive n lysis of online news in Europe nd the United St tes.” Journalism Studies, 9(5), pp. 717-738. T Ridout M Fr nz (2008). “Ev lu ting me sures of c mp ign tone.” Political Communication, 2, pp. 158-179. D Riffe, S Lacy & FG Fico(2005) Analyzing Media Messages: Using Quantitative Content Analysis in Research. Segunda edición. Londres: Lawrence Erlbaum. MJ Rozell (1996). The press and the Bush presidency. Westport: Praeger. CS Taber & M Lodge (2001). “Motiv ted skepticism in politic l inform tion processing.” Communication submitted to the Annual Congress of the Midwest Political Science Association. S T ber M Lodge J Gl th r (2001). “The motiv ted construction of politic l judgments.” En JH Kuklinski (Ed.), Citizens and politics: Perspectives from political psychology. London: Cambridge University Press. J Strömbäck (2008). “Swedish election news cover ge: Tow rds incre sing medi tiz tion.” In J Strömbäck & LL Kaid (Eds.), The handbook of election news coverage around the world (pp. 160174). Nueva York: Routledge. J T kens W V n Atteveldt A V n Hoof J Kleinnijenhuis (2013). “Medi logic in election c mp ign cover ge.” European Journal of Communication, 28, pp. 277-293. M Torc l JR Montero (2006). “Politic l dis ffection in comp r tive perspective.” En M Torc l JR Montero (Eds.), Political disaffection in contemporary democracies: Social capital, institutions, and politics (pp. 3–19). Abingdon, England: Routledge. A Van Dalen, E Albaek & C De Vreese (2011). Suspicious minds: Explaining political cynicism among political journalists in Europe. European Journal of Communication, 26(2), pp. 147-162. AE Weintr ub E Pinkleton (1995). “Positive nd neg tive effects of politic l dis ffection on the less experienced voter.” Journal of Broadcasting & Electronic Media, 39(2), pp. 215-235. JR Z ller (1999). “A theory of medi politics. How the interests of politicians, journalists, and citizens sh pe the news.” [Unpublished m nuscript] Av il ble t http://www.sscnet.ucla.edu/polisci/faculty/zaller/media%20politics%20book%20.pdf ______________________________________________________ How to cite this article in bibliographies / References http://www.revistalatinacs.org/071/paper/1089/09en.html Página 177 Revista Latina de Comunicación Social # 071 – Pages 160 to 178 [Research] [Funded] | DOI: 10.4185/RLCS-2016-1089en | ISSN 1138-5820 | Year 2016 M R Berganza Conde, C Arcila Calderón, R de Miguel Pascual (2016): “Negativity in Political News of Spanish News Media”. Revista Latina de Comunicación Social, 71, pp. 160 to 178. http://www.revistalatinacs.org/071/paper/1089/09en.html DOI: 10.4185/RLCS-2016-1089en Article received on 24 November 2016. Accepted on 7 February. Published on 11 February 2016. http://www.revistalatinacs.org/071/paper/1089/09en.html Página 178