FAKENEWS-m (ADVANCED MODELING OF FAKE NEWS PERCEPTION)

Supervisor: Reinhold Kliegl

This project is a spin-off of the earlier research supported by the grant from the Russian Science Foundation No 19-18- 00206 in 2019-2021. The project focuses on the in-depth analysis of the data collected in the previous research with advanced innovative methods. The project consists of three parts.

1. Modeling the effects of narrative and source type on news perception across three countries and two social networking sites with LMMs (completed)

In this part we perform advanced modeling and theoretical re-assessment of our online experiment that involved 8559 social media users from Russia, Ukraine and Kazakhstan who were presented true and false news items about neighboring countries that also varied by the type of source (domestic / foreign), type of narrative (dominant / alternative) and language (Russian / Ukrainian, for Ukraine only). We estimate a linear mixed-model with individual-level and news-item-level random effects and a system of contrasts of interest to account for complex interactions in our data.

Participants: Olessia Koltsova, Kirill Bryanov, Tetyana Lokot, Alex Miltsov, Sergei Pashakhin, Alexander Porshnev, Yadviga Sinyavskaya, Maksim Terpilovskii & Victoria Vziatysheva  

Publication:

Bryanov K., Kliegl R., Koltsova O., Lokot T., Miltsov A., Pashakhin S., Porshnev A., Sinyavskaya Y., Terpilovskii M., & Vziatysheva V. (2023). What Drives Perceptions of Foreign News Coverage Credibility? A Cross-National Experiment Including Kazakhstan, Russia, and Ukraine. Political Communication,  40:2, 115-146, DOI: 10.1080/10584609.2023.2172492  Download (PDF, 898 Kb)  

2. Modeling cognitive factors of news perception across three countries with signal detection theory (SDT) (in progress)

In this part we apply signal detection theory (SDT) as a theoretical and statistical approach to the same dataset of 8599 experiment participants, but with a focus on trust (reinterpreted as bias), but also on the sensitivity (resolution) of individuals. Accordingly, cognitive factors affecting news perception are studied using political factors as control variables. The analysis of the results is based on cumulative regression modeling with mixed effects, with bias and sensitivity as central SDT concepts modelled jointly.

Participants: Alexander Porshnev, Maksim Terpilowski, Maximilian Rabe

3. Modeling confirmation bias in divisive news perception by media professionals and lay readers with signal detection theory (SDT) (completed)

Participants: Victoria Vziatysheva, Olessia Koltsova, Maksim Terpilovskii

This part of the project is based on another online experiment (N=1946) conducted in Russia in 2021. Here we test how confirmation bias – that is, an inclination to believe news concordant with reader’s views –  affects trust in fake and true news about three socially divisive topics (including abortions and death penalty) among two groups of participants: media professionals and ordinary social media users. SDT is used as a theoretical and statistical approach.

Publication:

Victoria Vziatysheva, Olessia Koltsova, Maksim Terpilovskii, Reinhold Kliegl. Are media professionals better at fake news recognition and less susceptible to confirmation bias? A signal-detection approach (under review)