layout: true --- class: left, middle, title-slide background-image: url("https://cdn.pixabay.com/photo/2017/09/10/18/25/question-2736480_960_720.jpg") background-size: cover # Emotions in <br> misinformation studies: ### Distinguishing emotional state <br> from response & misinformation <br> recognition from susceptibility ### Hannah Metzler #### Complexity Science Hub Vienna & <br> Medical University of Vienna #### Workshop Järfälla, 15 Sep 2023 <br> Slides: www.hannahmetzler.eu/measuring_emo_misinfo --- <!-- layout:true --> <!-- <div class="my-footer"><span> --> <!-- <a href="https://doi.org/10.1126/science.aao2998">Lazer et al. (2018) </a> --> <!-- </span></div> --> <!-- --- --> <!-- # Misinformation as umbrella term --> <!-- ### False or misleading information --> <!-- .pull-left[ --> <!-- <br><br> --> <!-- ```{r, echo=FALSE, out.width=500, fig.align='center'} --> <!-- knitr::include_graphics("figures/fake-news-3801637.svg") --> <!-- ``` --> <!-- ] --> <!-- .pull-right[ --> <!-- * Independent of intent --> <!-- * Intentional: Disinformation <br> --> <!-- * False News <br> --> <!-- * Biased information <br> --> <!-- * Conspiracies & myths --> <!-- ] --> --- layout: true <div class="my-footer"><span> <a href="https://doi.org/10.1016/j.tics.2021.02.007"> Pennycook & Rand (2021); </a> <a href="https://doi.org/10.1186/s41235-020-00252-3"> Martel et al. (2020); </a> <a href="https://doi.org/10.1177/0956797611413294"> Berger et al. (2011); </a> <a href="https://doi.org/10.1027/1618-3169/a000489"> Greenstein & Franklin (2011); </a> <a href="https://oxford.universitypressscholarship.com/view/10.1093/oso/9780190900250.001.0001/oso-9780190900250-chapter-16"> Weeks & Garrett (2019); </a> <a href="https://www.pexels.com/photo/annoyed-black-woman-having-video-chat-on-smartphone-at-home-5699823/"> Photo by Alex Green </a> </span></div> --- ## Emotions & belief in misinformation - 2 theories ### 1) Emotions make us open to manipulation .broad-left[ * Emotions attract attention * Trigger actions * Hinder reflection * High arousal emotions * Sensational news & emotional manipulation ] .narrow-right[ <img src="figures/pexels-alex-green-5699823.jpg" width="400" style="display: block; margin: auto 0 auto auto;" /> ] **Increased emotions => misinformation belief & sharing** --- layout: true <div class="my-footer"><span> <a href="https://press.princeton.edu/books/hardcover/9780691178707/not-born-yesterday"> Mercier (2020); </a> <a href="https://doi.org/10.1177/20563051221150412"> Altay et al. (2022); </a> <a href="https://www.sciencedirect.com/science/article/pii/S2352250X22001440"> Robertson et al. (2022); </a> <a href="https://doi.org/10.1073/pnas.2013464118"> Guess et al. (2021);</a> <a href="https://doi.org/10.1007/978-1-4939-0867-7_10"> Steckler et al.(2014) ;</a> <a href="https://psycnet.apa.org/doiLanding?doi=10.1037%2Fa0013503"> Nesse & Ellsworth (2009)</a> </span></div> --- ## Emotions & belief in misinformation - 2 theories ### 2) Evolutionary theory of communication: Emotions are adaptive .broad-left[ * Communication must benefit sender & receiver to be evolutionarily stable * At default, people are skeptical, not gullible. * Function of emotions depends on context: * Prior beliefs * Social motives: belonging & reputation * Source of emotion: state vs. response ] .narrow-right[ <img src="figures/N8.PNG" height="300" /> ] ??? - Stable: if people could be easily deceived, others would take advantage of that. - Pretend to be angry to get something from us. - If this happened often, people would realize they only pretend, they don't follow up on their threats - so people would stop attending to the emotional signals - the signal would die out - skeptical: hard to convince people of things they don't already believe (everyday life experience) - Dependent on specific context - Scientifically minded person: angry about lies being told that vaccination during pregnancy kills babies - Vaccine skeptics angry that babies die - Emotions as signals for social opportunities & threats - In this view, what causes an emotion is crucial to predict it's influence --- layout: true <div class="my-footer"><span> <a href="https://journals.sagepub.com/doi/pdf/10.1177/0963721417718261"> Douglas et al. (2017); </a> <a href="https://press.princeton.edu/books/hardcover/9780691178707/not-born-yesterday"> Mercier (2020) </a> </span></div> --- # Example: Misinformation in crisis times .pull-left[ **1) Uncertainty = susceptibility** * Anxiety & loss of control * Sense making to reduce uncertainty * Simple cause-effect & evil-good stories ] .pull-right[ **2) Adaptive emotions & behavior** * Openness to potentially relevant information is adaptive * Not all beliefs influence behavior * Gaining status by sharing socially relevant (=surprising) news * Extremist: cutting bridges to prove loyalty ] ??? Cutting bridges by making extreme statements and proving loyalty to a fringe group - statements need to be crazy Difference between reflective and intuitive beliefs: beliefs that are relevant for our actions --- layout:true --- ## Methods overview * Pre-registered replication of correlation study 1 from [Martel, Pennycook & Rand (2020)](https://doi.org/10.1186/s41235-020-00252-3) * Online survey, Austria, Dec 2021, N=422, left-leaning & educated sample * Actual false & real COVID-19 news items (n=24): accuracy rating task <center> <img src="figures/N15.PNG" height="270" /> <img src="figures/NR14.PNG" height="270" /> </center> --- layout:true <div class= "my-footer"><span> <a href="https://doi.org/10.1186/s41235-020-00252-3"> Martel et al. (2020) </a> </span></div> --- ## Replication: Emotion & discernment don't correlate .pull-left[.center-left[ Martel et al. (2020) <!-- --> Replication <img src="figures/replication_fake.png" width="500" /> ]] .pull-right[.center-right[ **Emotional state prior to exposure** * Pre-registered * Time of reference <br> "past days" vs. "right now" * Other differences: Topic, education, political orientation * Neglected factors: * the source * the person's interpretation ]] --- layout:true --- ## Immediate emotional response to news items .broad-left[ <!-- --> ] .narrow-right[ * Exploratory * Most people reported an emotional response * More anger after false than real news * b = 0.44*** (LMEM) * Happy/excited after real news * b = -0.46** (LMEM) ] --- ## Angry responses driven by people recognizing lies <img src="figures/curvi-linear.svg" width="800" style="display: block; margin: auto;" /> * Higher anger in people **good & bad** at recognizing false news * In our sample: mostly in people with good performance * Angry responses because they **recognize** false news? <div style="color:grey"><small><left> Robust polynomial curves for news discernment & emotional responses. <br> Black dotted line = linear model. </small> </div> --- ## People were angry for different reasons .broad-left[ <img src="figures/WordShiftGraph_full.png" height="430" /><img src="figures/WordShiftGraph.png" height="430" /> ] .narrow-right[ * Wordshift graph * False (N = 2,873) * Real (N = 2,470) * Anger & disbelief for false items: <br> Recognition * Additional analyses: informed people are angry ] --- # Take aways * Prior beliefs determine the meaning of emotion: <br> * **Susceptibility vs. Recognition** * How and when we measure emotion in studies matters * **Immediate response** rather than prior state (source) * Stimulus: content matters * Forced exposure: low emotions * Social media data: Comments to posts & prior beliefs ### Emotions `\(\neq\)` gullibility. Emotions are adaptive. --- # Thank you! Questions or comments? .pull-left[.center-left[ Thanks to the project team! <img src="figures/collaborators.svg" width="600" style="display: block; margin: auto;" /> <br> <br> ... and our funder! <br> <br> <img src="figures/WWTF_Text_RGB.svg" width="320" style="display: block; margin: auto 0 auto auto;" /> ]] .pull-right[ * Project: [hannahmetzler.eu/emomis](hannahmetzler.eu/emomis) * Pre-print: [psyarxiv.com/udqms/](https://psyarxiv.com/udqms/) * Staying in touch: * metzler@csh.ac.at * [hannahmetzler.eu](hannahmetzler.eu) * [@hannametzler1](https://twitter.com/hannahmetzler1) ] <img src="figures/csh.png" width="210" /> <img src="figures/uni_wien_logo_blue.jpg" height="90" /> <img src="figures/Meduni-wien.svg" height="90" /> --- layout:true class: title-slide, middle background-image: url("https://cdn.pixabay.com/photo/2017/09/10/18/25/question-2736480_960_720.jpg") background-size: cover --- # Additional information --- layout:true --- # Links * [OSF repository](https://osf.io/tgzxr/): materials, data, code * [Pre-print](https://psyarxiv.com/udqms/) * [OSF preregistration](https://osf.io/2r6bj) --- # Immediate emotional response * Pre-registered emotional responses to each news item: “When reading this news, I feel…” <img src="figures/Screenshot_emotional_responses.png" width="800" style="display: block; margin: auto;" /> * Free text: "Which impulse or thought does this news article elicit?"