« Back to all events

Believing the news: How individual ideology and source credibility affect judgements of informational truth

Dates:
  • Mon 30 Nov 2020 10.30 - 12.30
  Add to Calendar 2020-11-30 10:30 2020-11-30 12:30 Europe/Paris Believing the news: How individual ideology and source credibility affect judgements of informational truth

This thesis explores what makes people believe (and share) news about politics. Two independent variables play a particular role throughout the four papers: The congruence of information with people’s attitudes or ideology, and the credibility of sources. In Chapter 1, a first study with a German sample tests how congruence with pre-existing attitudes and the type of source (real versus fake) affect believing news reports. Contrary to predictions, credible sources cannot mitigate the tendency for assimilation bias, i.e. the tendency to give more credit to congruent information. A second study confirms this finding for sharing intentions. Chapter 2 presents a study from Germany that explores drivers of source credibility. As predicted, subjects are more likely to believe a source that has previously—i.e. through experimental manipulation—provided congruent information. Machine learning methods reveal that people with low media trust and voters of the populist right are most susceptible to this effect. The last two chapters shift the focus to the US, and to asymmetries in information processing between liberals and conservatives. Chapter 3 examines three previous peer-reviewed studies that test which side is more prone to assimilation bias and is more able to tell true from false. Various robustness checks show that results of these studies do not survive variations in the selection of informational stimuli. This points to the need to ensure externally valid representation of a clearly defined target population of information. Chapter 4 tries to meet this requirement by, firstly, collecting a large set of real-world news items (both true and false), and by secondly randomly sampling stimuli from this collection. In a subsequent survey, liberals turn out to be more truth-discerning, but also more biased than conservatives. These asymmetries can be partly attributed to the real-world news supply, partly to psychological differences per se.

via zoom - DD/MM/YYYY
  via zoom -

This thesis explores what makes people believe (and share) news about politics. Two independent variables play a particular role throughout the four papers: The congruence of information with people’s attitudes or ideology, and the credibility of sources. In Chapter 1, a first study with a German sample tests how congruence with pre-existing attitudes and the type of source (real versus fake) affect believing news reports. Contrary to predictions, credible sources cannot mitigate the tendency for assimilation bias, i.e. the tendency to give more credit to congruent information. A second study confirms this finding for sharing intentions. Chapter 2 presents a study from Germany that explores drivers of source credibility. As predicted, subjects are more likely to believe a source that has previously—i.e. through experimental manipulation—provided congruent information. Machine learning methods reveal that people with low media trust and voters of the populist right are most susceptible to this effect. The last two chapters shift the focus to the US, and to asymmetries in information processing between liberals and conservatives. Chapter 3 examines three previous peer-reviewed studies that test which side is more prone to assimilation bias and is more able to tell true from false. Various robustness checks show that results of these studies do not survive variations in the selection of informational stimuli. This points to the need to ensure externally valid representation of a clearly defined target population of information. Chapter 4 tries to meet this requirement by, firstly, collecting a large set of real-world news items (both true and false), and by secondly randomly sampling stimuli from this collection. In a subsequent survey, liberals turn out to be more truth-discerning, but also more biased than conservatives. These asymmetries can be partly attributed to the real-world news supply, partly to psychological differences per se.


Location:
via zoom -

Affiliation:
Department of Political and Social Sciences

Type:
Thesis defence

Contact:
Monika Rzemieniecka (EUI - Department of Political and Social Sciences) - Send a mail

Defendant:
Bernhard Clemm von Hohenberg

Examiner:
Prof. Hanspeter Kriesi (EUI - Department of Political and Social Sciences)
Prof. Andrew Guess (Princeton University)
Prof. Heike Klüver (Humboldt University Berlin)

Supervisor:
Prof. Diego Gambetta (EUI - Department of Political and Social Sciences)

Similar events

 

Page last updated on 18 August 2017