EyePoint: Investigating Trust Factors in Misleading Information Messages Involving Oculomotor Data (Comprehensive Project)

Project team leaders: Olessia Koltsova, Elena Artemenko.

The spread of misinformation in social media is now becoming evident, which makes it necessary to look for reliable trust predictors in unreliable information. A complex project “Investigating Trust Factors in Misleading Information Messages Involving Oculomotor Data” unites a group of projects, aimed at comprehensive research of various predictors of reader's perceived message credibility. A distinctive feature of the group of projects is the combination of data from large online experiments and laboratory experiments involving behavioral oculomotor data.

1. The role of cognitive distortion of confirmation bias and social approval signals in informational messages trust
Participants: Elena Artemenko, Taisiia Uliianova, Maksim Terpilovskii, Reinhold Kliegl.

This project focuses on investigating two main predictors:

·        cognitive bias of prevalence of subconsciousness (confirmation bias)

·        social (un)approval signals (user comments).

The cognitive confirmation bias effect states that messages are perceived as more credible if they support the reader's personal beliefs. The confirmation bias effect is very persistent and almost uncorrectable.

A significant amount of research also shows that social (non)approval signals (e.g., user comments) influence human behavior, including perceptions of the credibility of an information message.

In our study, we investigate whether the presence of a comment can affect the effect of confirmation bias. We also use eye-tracking to identify key social media reading strategies and their impact on perceptions of credibility and accuracy in detecting fake news.

Results are presented at the following conferences:

1. ICA Regional Conference 2022 Computational Communication Research in Central and Eastern Europe (Helsinki, Finland, June 27-29, 2022)
2. ECEM2022: 21st European Conference on Eye Movements (Leicester, UK, August 21-25, 2022).
3. Experimental studies of Language and Speech (Russia, Tomsk, October, 2022)
4. ESCOP 2023: 23rd Conference of the European Society for Cognitive Psychology (Porto, Portugal, September 06-09, 2023)

Publications:

1. Confirmation bias and social endorsement cues in fake news detection accuracy and credibility perception: an eye tracking approach (Elena Artemenko, Olessia Koltsova, Maksim Terpilowskii, Taisia Uliyanova) - in progress
2. “Naive media literacy”: strategies for distinguishing credible and untrustworthy information in social media (Elena Artemenko, Natalia, Olessia Koltsova, Maksim Terpilowskii, Taisia Uliyanova) - in progress

2. The Influence of Interface Elements on Trust in Online News (student's project)

Participants: Liliia Faizova, Elena Artemenko.

The project is aimed to study the influence of interfaces on the assessment of participants' trust in credible and untrustworthy news messages. The study includes 3 stages:

1. Identification of the most preferred way of consuming news content. This step was accomplished through an online survey (N = 153). The results of the survey allowed us to determine the type of interface that is developed in the second stage.
2. Identification of the most and least preferred combinations of elements in terms of credibility. At this stage it is planned to use conjoint analysis (N = 1200).
3. Online experiment to study the influence of the interface on the message credibility, including the eye-tracking methodology.

Results presented at the following conferences:

1. Carl Dunker Memorial Summer School in Cognitive Psychology, Faculty of Psychology, ION RANCHI
2. International Summer School on Neurolinguistic Studies “Eye-tracking for Science and Life”, Center for Language and Brain, National Research University Higher School of Economics, Moscow
3. Applied Cognitive Science Summer School, Laboratory for Cognitive Psychology of Digital Interfaces Users

Publications:

1. The Influence of Interface Elements on Trust in Online News (in progress)

3. Behavioral correlates of the confirmation bias effect.

Participants: Elena Artemenko, Anastasia Zhitkova, Sergei Koltcov.

The research has two directions:

Direction 1. Search for behavioral (oculomotor) correlates of the confirmation bias effect.

The search for behavioral and neurophysiological correlates of mental processes is one of the key steps in automating their attribution by machine learning methods (when creating classifiers). The confirmation bias effect belongs to the class of cognitive bias of decision-making and is widely represented in the literature: information is perceived as more reliable if it confirms personal beliefs, and vice versa. A number of studies have demonstrated increased cognitive load under the influence of confirmation bias. The frequency and duration of blinks have also been shown to be related to various aspects of cognitive load, one of the causes of which is the receipt of information that does not correspond to the recipient's beliefs. The present study aims to find a connection between cognitive load and oculomotor activity (as its possible behavioral correlate).

Direction 2. Automatic attribution of texts on the basis of oculomotor activity.

In this direction, the main goal is to extract, using machine learning methods, the feature base from oculomotor activity in order to create classifiers of a model that allows to attribute cognitive dissonance caused by the action of cognitive confirmation bias (i.e., the situation when contexts do not correspond to user beliefs)

Results are presented at the following conferences:

1. XXIV Congress of the I.P. Pavlov Physiological Society (St. Petersburg, Russia, September 11-15, 2023)