09:00 - 10:30
Parallel sessions 1
09:00 - 10:30
Room: C-Building - N14
Chair/s:
Gabriella Daroczy, Freideriki Tselekidou
Narrative is not just a way to package information; it is cognitive infrastructure with its specific architecture. This symposium, spanning five perspectives, shows how narrative architecture shapes thinking across contexts.

The first talk investigates how narrative competence supports mathematics, especially word problems. An online adult study linked narrative skills to performance (reaction time and accuracy) across problem types, including carry/borrow operations and tasks that varied number relevance. This points to cross-domain links between narrative skill and math problem-solving.
The second talk shows how early oral narrative macrostructure predicts later reading comprehension, with strong evidence in Greek–German bilingual learners. The results align with models like the Simple View of Reading and stress the value of early narrative abilities for later literacy.
The third talk explores how coherence relations connect parts of a text and guide meaning making. Interpretations of the same relation can shift with language and cultural perspective. Using an annotation-based approach across originals and translations, the work maps these differences and explores computational models to capture them.
The fourth talk shows how literary reading may be shaped not only by text-internal features (language use, themes) but also by extrinsic cues, i.e. paratextual information such as signals of a work’s canonical status. An online study shows how paratextual cues about a novel excerpt’s literary quality (Booker Prize nomination vs. none displayed on the cover) influence story perception, reading experience, and text processing.
The fifth talk examines the effect of narrative structure on revising mental models after corrections (debunking effect) versus sticking to misinformation (continued influence effect). Study 1 varied psychological distance (Germany vs. another continent) and emotional valence (positive vs. negative); Study 2 varied correction design, testing whether including and ordering explanations (why the misinformation is false) impacts debunking.

To sum up, across these studies, narrative appears as a basic mental tool across domains that selects what matters, links ideas, and guides belief updating. From math to bilingual literacy, from cross-linguistic interpretation to paratext effects and misinformation correction, the common message is clear: shaping narrative structures and cues can meaningfully steer learning, comprehension, and reasoning.
Submission 257
Correction of Misinformation as a Function of Narrative Text Structure
SymposiumTalk-05
Presented by: Benedikt T. Seger
Benedikt T. SegerGerhild Nieding
University of Würzburg, Germany
We examined the extent to which recipients of narrative text revise their mental models (debunking effect) or stick to misinformation (continued influence effect, CIE) after its correction. These effects were measured using a rating scale on inference statements (= items) that are either compatible or incompatible with the corrected misinformation.

Study 1 (229 participants, 32 items) tested the CIE following the structural integration of corrections in narrative text (immediate refutation of misinformation), compared to its non-integration (delayed correction message). As expected, the CIE is higher in delayed compared to immediate corrections. In Experiment 1, this effect is moderated by emotional valence, with negative, but not positive misinformation increasing the CIE in immediate, but less so in delayed correction. When misinformation is emotionally neutral (Experiment 2), the psychological distance of the described events acts as a moderator. Accordingly, there is a relative increase in CIE when psychological distance is high and correction is immediate.

In Study 2 (302 participants, 60 items), we investigated how the structure of correction messages (as suggested in the Debunking Handbook; Lewandowsky et al., 2020; doi:10.17910/b7.1182) modulates the debunking effect. The debunking effect is higher when the misinformation is presented first, compared to a “sandwich” structure with the misinformation placed between facts. Debunking effects of longer correction messages are higher than those of shorter ones that summarize these, but lower than those of shorter messages that leave out the explanation why the misinformation is wrong. These findings are in conflict with some of the Debunking Handbook’s suggestions.