In the polarised context that we live in, citizens’ conceptions of the media are ideologically shaped, and they identify political positions in different media outlets. The political nature of media outlets reduces citizens’ trust in the media, or at least in media outlets that are not aligned with their own ideology. At the same time, fake news and disinformation are commonplace.

This scenario raises key questions for social scientists: do people tend to believe some news more than others? Are they prone to believing fake news? What factors can help us understand which news items are believed? This article explores these questions through an experiment conducted in Spain with news stories based on real and fake information and with real and invented media outlets.

The fieldwork was carried out between May 9 and May 25, 2023. The research firm GESOP was responsible for overseeing the execution of the fieldwork. The methods used in this study consisted of a survey experiment conducted with 2,500 residents of Spain aged 18 to 74. 

The questionnaire was online, self-administered, using a panel of Internet users. It included five sections: a) a set on media consumption and perception; b) a set on political core beliefs; c) a 10-question version of the IPIP personality test (named TIPI); d) a set of demographic questions, and e) the experimental part. 

The experiment consisted of showing each respondent six pieces of news on the studied topics (gender, immigration, climate change, Ukraine war, covid-19 vaccines, and sexual violence), consecutively. Each piece of news was randomly varied in its combination of the following elements: name of the outlet, tone of the headline (positive, neutral, or negative), gender of the journalist, (apparent) nationality of the journalist, whether the piece included a picture or not, and the tone of the picture (positive, neutral, or negative).  

The news stories used in the experiment were presented as being from different Spanish media outlets ‒ i.e., CTXT, El Mundo, El País, Newtral, Okdiario, RTVE, and Telecinco ‒ and an invented media outlet: Noticias Hoy. Unlike the real media outlets, that had “lovers” and “haters” among the participants, the Noticias Hoy invented one tended to get more medium scores.

The data derived from the experiment show that the level of trust by individuals in the selected Spanish media outlets and the invented one was relevant to whether they believed the news presented in the article. Specifically, previously expressed trust in a media outlet significantly increased the likelihood of a participant believing the news presented under its news media brand. During the experiment, this occurred whether the news story in question was based on real or fake information. 

A comparison among three headline versions was run during the experiment. For each news item, three versions were created: one with a positive headline, one with a neutral headline, and one with a negative headline. For example, for the anti-vaccine news story, the following three headlines were presented: a) “Recent studies suggest that covid-19 vaccines may involve mild adverse effects” (positive); b) “Recent studies point to possible adverse effects of covid-19 vaccines” (neutral), and c) “Recent studies call into question covid-19 vaccines due to their adverse effects” (negative). Articles with the positive headline received higher average scores for perceived truthfulness and likelihood of sharing than those with neutral or negative headlines. This finding held for both real and fake news.

The news stories used in the experiment dealt with several topics: the anti-vaccine movement, climate change, gender, the war in Ukraine, migration, and sex crimes. Only the news stories about sex crimes were real. Participants who reported previous knowledge about a news story (real or fake) were more likely to share it. However, higher levels of self-assessed previous knowledge on the topic did not affect the truthfulness attributed to fake news.

On average, respondents tended to report being moderately or highly knowledgeable about the different topics. Nevertheless, for perceived truthfulness, there was a large concentration of values at 0 (that is, respondents attributed no credibility to the narratives). Values in perceived truthfulness above 0 and below 4 were infrequent for all self-reported knowledge levels. A value of 5 was rather frequent, which can be interpreted as a middle-of-the-road positioning on the controversial topics covered. Perceived truthfulness levels above 7 were more frequently given when self-reported knowledge levels were also equal to or above 7.  

The immediate conclusion that can be drawn is that the respondents’ self-reported level of knowledge on a particular topic was not a good predictor of their capacity to identify fake news. On the contrary, respondents who gave high truthfulness scores to these fake stories tended to self-report higher knowledge levels than the rest of respondents. This finding could inform policies for prevention and intervention: citizens may tend to assess themselves as more knowledgeable than they really are, making them more likely to believe fake news. Hence, media literacy policies as well as fact-checking strategies are crucial. 

Political self-identification was measured in two ways. On the one hand, participants rated their political orientation on a scale ranging from extreme left to extreme right. An “apolitical” option was also available. On the other hand, participants responded to a battery of questions indicating their degree of agreement with statements about their core beliefs. 

These questions allow setting individuals along two axes, one representing the individualism-collectivism continuum and another representing the authoritarianism-liberalism continuum. This way of measuring political ideology offers a much finer grained portrait of political positions than the classical left-right scale. 

Both measures point in the same direction when it comes to results. Specifically, they show that identifying with the right or agreeing with the conservative core beliefs (which tend more toward authoritarianism and individualism) increased the likelihood of believing the news stories and of sharing them, regardless of whether they were real or fake.

On the individualism-collectivism axis, core beliefs on a) the unacceptability of abortion and b) the fitness of homosexual parents tend to exclude each other, i.e., respondents who agreed with one of them, tended to reject the other, indicated by a negative correlation (-0.36). 

On the authoritarianism-liberalism axis, the highest correlations are observed between “F. Convictions about the beliefs of conservatives” and “G. Friendship with a person of a contrary ideology”, with a correlation of 0.45, and between “E. Distrust of laws”, and “C. Distrust of the system” (0.37). Further research would be needed to explore how widespread these beliefs are among different social groups.

This research has identified several factors associated with trust in fake news stories: previous trust in the news media brand, a positive statement in the headline, identifying oneself as on the political right, and having core beliefs also aligned with the political right (closer to the authoritarian and individualist leanings). Similar factors are associated with participants’ intention to share fake news content, although the news media brand was less important, while self-report of previous knowledge was more important. In the polarised context that we live in, fake news is commonplace. Identifying the factors that lead citizens to trust fake content and spread it can help us to counter disinformation.

MASIP, P.; J. SUAU; C. RUIZ-CABALLERO (2020): «Percepciones sobre medios de comunicación y desinformación: ideología y polarización en el sistema mediático español», Profesional de la Información, 29(5).

SNIDERMAN, P. M.; J. N. Druckman (ed.) (2011): «The Logic and Design of the Survey Experiment», Cambridge Handbook of Experimental Political Science, 102-114.