Peer Threat Evaluations Shape One’s Own Threat Perceptions and Feelings of Distress

Peer Threat Evaluations Shape One’s Own Threat Perceptions and Feelings of Distress | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Peer Threat Evaluations Shape One’s Own Threat Perceptions and Feelings of Distress Lisa Espinosa, Erik C. Nook, Martin Asperholm, Therese Collins, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-3875288/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract We are continuously exposed to what others think and feel about content online. How do others’ evaluations shared in this medium influence our own beliefs and emotional responses? In two pre-registered studies, we investigated the social transmission of threat and safety evaluations in a paradigm that mimicked online social media platforms. In Study 1 (N=103), participants viewed images and indicated how distressed they made them feel. Participants then categorized these images as threatening or safe for others to see, while seeing how “previous participants” ostensibly categorized these images (these values were actually manipulated across images). We found that participants incorporated both peers’ categorizations of the images and their own distress ratings when categorizing images as threatening or safe. Study 2 (N=115) replicated these findings and further demonstrated that peers’ categorizations shifted how distressed these images made them feel. Taken together, our results indicate that people integrate their own and others’ experiences when exposed to emotional content and that social information can influence both our perceptions of things as threatening or safe, as well as our own emotional responses to them. Our findings provide replicable experimental evidence that social information is a powerful conduit for the transmission of affective evaluations and experiences. Psychology Cognitive Neuroscience social learning emotional influence threat and safety learning online study Figures Figure 1 Introduction Seven out of ten Americans report using social media platforms in their daily lives (Anderson & Auxier, 2021), making online content a central source of information exchange. Although social media offers opportunities to connect with others remotely, it also provides a medium through which emotional evaluations of events in the world around us can become amplified. For example, travelers in 2016 panicked and hid after reading online that there was a gunman in the JFK airport, even though no gunman existed (Wilson & Joseph, 2016). What psychological processes explain such rapid online transmission of threat evaluations and their concomitant emotional and behavioral responses? Here, we develop an experimental paradigm for testing how social information shapes (i) evaluations of stimuli as threatening and safe, and (ii) actual emotional reactions to those stimuli. A growing body of work has shown that positive and negative emotions spread between people, both in face-to-face interactions (Parkinson, 2011; Peters & Kashima, 2015) and online (Fan et al., 2020; Ferrara & Yang, 2015; Kramer et al., 2014). One experimental study investigating online emotion contagion manipulated the amount of emotional content Facebook users saw in their newsfeeds and showed that merely manipulating the number of positive and negative posts one viewed influences, respectively, the number of positive and negative posts those users then make (Kramer et al., 2014). Interestingly, it has been shown that people predominantly share negative (vs. positive) content online, regardless of whether the content is related to a negative or positive event, such as losing or winning a political election (Schöne et al., 2021). This is consistent with classic research showing a “negativity bias” in attention, memory, and conformity (Rozin & Royzman, 2001). Furthermore, anger seems to spread faster than joy on social media, as anger tweets are more likely to spread through even weak social ties than joyful tweets (Fan et al., 2020). Importantly, exposure to such negative content appears to lead to more online engagement (Bellovary et al., 2021) and spreads more widely (Brady et al., 2019; Wang & Inbar, 2022). Similarly, scientists have shown that the expression of emotions, such as moral outrage, spreads across users through design features of social media platforms (Crockett, 2017). Features, such as using notifications that capture and hold their attention, and using algorithms that learn from users’ behavior, drive users to share moral-emotional content online by increasing motivation to share (Brady et al., 2020; Brady et al. 2023). As such, growing literature shows that specific emotional responses (e.g., outrage) and broader affective states (i.e., positivity and negativity) can flow through online social networks. More research is needed to move the focus beyond the momentary contamination of affective responses towards an understanding of how online information shapes emotions and more durable evaluations through learning. Here, we examined the processes underlying such learning by asking how information of others’ evaluations of threat and safety are integrated with our own evaluations and decisions. The online transmission of threat evaluations has become central to our current experiences, as we have seen the power of online information in shaping evaluations of COVID-19 and other major events as either dangerous or harmless (Fuentes & Peterson, 2021; Haman, 2020). Anecdotal observations and a growing literature on social learning show that it is psychologically efficient for people to learn to evaluate stimuli as threatening/safe by observing other people’s reactions, rather than directly engaging with potentially threatening situations. Consistent with this, research has indicated that we learn similarly well when we observe others’ threat responses or when told by others that something is dangerous, as when we learn through direct experience (Lindström et al., 2019a; Olsson & Phelps, 2004). These forms of social learning also involve similar neural and computational mechanisms as direct experience (Olsson et al., 2020). For example, people will show physiological responses to images that are associated with someone else receiving a shock (Haaker et al., 2017), and verbal instructions that a certain image will be followed by a shock produce similar levels of learning as either via direct experience or observation (Olsson et al., 2020; Phelps et al., 2001). To our knowledge however, it is unknown if threat/safety cues can be transferred through observation in an online setting and, if so, the impact of this social transmission on people’s actual emotional responses and decisions. For instance, research on conformity has shown that exposure to online peer norms can shift both behavior and deeper affective or neural responses to affective stimuli (Martin et al., 2018; Nook et al., 2016; Nook & Zaki, 2015; Prehn et al., 2013). These studies support the hypothesis that social perceptions of stimuli as threatening might influence both one’s own evaluations of those stimuli as threatening as well as actual feelings of distress. Here we unite these bodies of work through a novel experimental paradigm. In this study we investigate whether choosing to label situations as threatening or safe is influenced by seeing peer threat/safety evaluations, and whether peer evaluations influence our feelings of distress. Though we are interested in online spread of information, we developed an online paradigm with high degree of experimental control rather than using social media data to specifically manipulate peer threat and safety evaluations. The two current studies investigated if people conform to, propagate, and are emotionally influenced by peers’ threat/safety evaluations in an online setting. In Study 1, we examined if receiving information about peers’ evaluation of images as threatening or safe influences an individual’s decision to label and share these images as threatening or safe. This first study established a suitable paradigm, and Study 2 replicated Study 1 and extended it to test if peers’ evaluation of images as threatening or safe affects individual’s emotional response to that image by shifting their feelings of distress. Investigating how threat/safety information spreads through online networks can clarify the psychological processes explaining how social groups come to share evaluations of major events (e.g. pandemics, natural disasters and political events) as dangerous or safe, laying a foundation for interventions that can shape the online spread of threat evaluations. Study 1 Method Participants Online data collection took place on the MTurk platform. A total of 109 individuals completed the online experiment. Given widespread issues with “bot-like” activity on MTurk (Mason & Suri, 2011), we preregistered several exclusion criteria to ensure data were of high quality. This led us to exclude six participants who excessively clicked away from the window in which the task was administered (> 20 switches). Participants passed all other exclusion criteria (i.e., completing the task too quickly [2 SD under the mean] or failing “captcha” trials in which objects presented as images had to be named). Consequently, a total of 103 participants (44 females, age range = 21 - 71, M age = 39.77, SD age = 12.78) were included in the analyses. Sample size was established using power analysis based on estimated effect sizes produced by a pilot study with 43 individuals (39 usable following data quality exclusions). The pilot study, which was identical to the current study, showed that obtaining 80% power to replicate the smallest hypothesized effect ( d = 0.39) would require 54 participants. However, because of the risk for unusable data collected on MTurk (10% in our pilot study), we decided, prior to initiating data collection, to increase sample size to about 100 individuals, which would give us a power of 95% for that effect size. Approval for this study was obtained by the Regional Ethical Review Board in Stockholm (Dnr: 2018/2511-31/5). Experimental Paradigm and Procedure After providing informed consent, participants completed a study that included two phases ( Figure 1A & 1B ). During phase 1, participants were asked to rate how distressed 60 negative images made them feel on a sliding scale (0 = not at all distressed , 100 = very distressed ). Participants completed two practice trials before rating the 60 images used in the analyses. This task provided a measure of participants’ initial affective response to the images. Images were presented one at a time in a randomized order for each participant. Images were drawn from the Open Affective Standardized Image Set (OASIS; Kurdi, Lozano, & Banaji, 2017) and contained aversive scenarios with a diverse set of themes including humans, animals, objects and scenes. Normed valence means ranged from mildly negative to very negative (see Table S1 ). During phase 2, participants were informed that they were part of a large-scale study on creating rating norms for a set of images, and that 100 other MTurk workers had previously categorized the images they just saw as “threatening” or “safe”. We instructed participants to categorize an image as “threatening” if it “is likely to cause emotional distress to others” and to categorize it as “safe” if it “is not likely to cause emotional distress to others”. Participants were also led to believe that their own categorizations would add to the full set of answers that would be used in subsequent studies. Participants then saw the same images a second time along with the number of previous MTurk workers who had ostensibly categorized these pictures as either threatening or safe, respectively. However, these numbers were randomly generated on each trial. Participants categorized each image as threatening or safe by clicking on either a red X or a green checkmark, respectively. When clicking on either symbol, participants saw the number of previous categorizations increase by 1 for their chosen category. In order to mimic social media platforms in which people watch peers’ responses to online content, express their own evaluation of the content, and share this with peers (i.e. others can see how they evaluate the content), participants were asked to click a “share” button to share their categorization for future participants to view. At the end of the study, participants reported their age, gender, and completed a set of questionnaires, including the balanced emotional empathy test (BEES; Mehrabian, 1996), the support for free speech scale (Alvarez & Kemmelmeier, 2018), the generalized anxiety disorder scale (Spitzer et al., 2006), the posttraumatic stress disorder checklist (PCL-5) and the patient-health questionnaire (Kroenke et al., 2001). For transparency, we report that these questionnaires were collected but we do not include them in the analyses. Finally, participants were debriefed on the study (including an explanation of the study’s deception) and were paid for their time. Statistical Analyses We preregistered a series of analyses involving computational modelling, however giving our concerns regarding the suitability of the computational modelling approach to this paradigm, for the main text we present a much simpler but conceptually identical analytic plan using mixed effect model. We present results from our preregistered analyses in the Online Resource, see section II and II.1. To test the hypothesis that peers’ threat categorizations influence participants’ decision to categorize an image as threatening, we conducted a logistic mixed-effect model at the trial level testing the relationship between peers’ threat categorizations and participants’ own categorization for each image, controlling for participants’ initial distress ratings for each image. Participant ID was included as a random effect to nest trials within participants. A significant relationship between the number of peers who categorized an image as threatening and participants’ decisions to categorize images as threatening provides evidence of conformity to these group norms. All analyses used an alpha threshold of 0.05 and include measures of effect size. Specifically, we report odds ratios of each predictor, produced using the “model_parameters” function in the parameters package (Lüdecke et al., 2020). Results In line with preregistered hypotheses, we found that participants’ decision to categorize images as threatening was positively related to both their initial distress ratings (OR = 1.01, 95% CI = [1.01, 1.02], p < .001) and the number of their peers who categorized the image as threatening (OR = 1.06, 95% CI = [1.05, 1.06], p < .001). Thus, participants were more likely to categorize an image as threatening if a greater number of their peers did so, even after controlling for participants’ own distress ratings. Study 2 In Study 1 we showed that participants’ decisions to categorize images as threatening or safe were influenced by peer threat/safety evaluations, even when controlling for participants’ own feelings of distress. Because peer information was randomized on each trial, this indicates that threat evaluations can spread from person to person online. Our results however do not tell us whether conforming to peers’ threat/safety evaluation influences individuals’ emotional state. In Study 2, we sought to both replicate the results of Study 1 and extend them by investigating if conforming to peers’ categorization of images would lead participants to update their emotions regarding these images. If so, this will demonstrate that exposure to online peer norms influences both one’s behavior and emotional state. Method Participants Online data collection again took place on the MTurk platform. A total of 127 individuals completed the online experiment. We excluded 12 participants who failed to pass specific pre-determined data quality measures. Nine participants excessively clicked away from the window in which the task was administered (> 20 switches). Three participants were excluded because they displayed no variability in their data; They categorized all images as either “threatening” or “safe”, making it impossible to model their choice behavior. All remaining participants passed pre-determined quality measures. No participants completed the task too quickly (2 SD under the mean) or failed any of the “captcha” trials in which objects presented as images had to be named. Consequently, a total of 115 participants were included in analyses (55 females, age range = 21 - 71, M age = 37.92, SD age = 12.21). We aimed to recruit the same number of participants as in Study 1 (i.e. 100 mTurk workers), and this sample size was established using power analyses based on estimated effect sizes produced by a pilot study with 43 mTurk workers (see preregistration: osf.io/sjq2y). Slightly more participants than this target number completed the study while it was active on MTurk. Approval for this study was obtained by the Regional Ethical Review Board in Stockholm (Dnr: 2018/2511-31/5). Experimental Paradigm and Procedure After providing informed consent, participants completed the same Phase 1 (initial distress rating) and 2 (categorization task) as in Study 1. However, Study 2 implemented four minor changes to Study 1’s procedures: (i) Eight images were estimated as too neutral (i.e. OASIS normed valence ratings) and were replaced in Study 2 (see Table S1 ). (ii) The numbers of ostensible previous categorizations as threatening or safe were randomly selected from a set of pre-determined numbers (and not randomly generated on each trial) to reduce noise across participants. (iii) The position of the threat and safe information on the left or right side of each image during Phase 2 was counterbalanced to remove potential confounds, i.e. 50 % of the participants saw the threat information on the left and right, respectively. (iv) Following the categorization phase, participants completed a third study phase in which they provided a follow-up distress rating for each image. This phase was identical to the initial distress rating (e.g. participants rated how distressed the negative images made them feel on a sliding scale from 0 = not at all distressed to 100 = very distressed ). This additional task provides a measure of participants’ affective response to the images after being exposed to peers’ categorizations. At the end of the study, participants reported their demographic information (i.e., age, gender), completed a set of questionnaires, were debriefed on the study (including an explanation of the study’s deception), and were paid for their time. Statistical Analyses Replicating Study 1. We first sought to replicate results of Study 1. As in Study 1, we ran a trial-level logistic mixed-effect model to assess the relationship between peers’ threat categorizations on participants’ own categorization, controlling for participants’ initial distress ratings. Participant ID was included as a random effect to nest trials within participants, and we again report odd’s ratios as effect sizes. Testing influence of peer evaluations on follow-up distress ratings . We also conducted a trial-level linear mixed-effect model to assess the relationship between peers’ threat categorizations and participants’ follow-up distress ratings. This regression included both participant ID as a random effect and participants’ initial distress ratings. We hypothesized that peers’ threat categorizations would influence follow-up distress ratings, over and above participants’ initial distress ratings. Here, we report standardized betas as effect sizes, using the “standardize_parameters” function in the parameters package (Lüdecke et al., 2020). As an additional test of our hypotheses, we examined whether participants who more strongly take on peer threat categorizations are also more likely to update their distress ratings to match group norms. Specifically, w e computed a conformity score for each participant (Klucharev et al., 2009; Nook & Zaki, 2015) in order to estimate the strength of the relationship between the number of peers who categorized the image as threatening and participants’ decision to categorize the image as threatening controlling for participants’ initial distress ratings. For each participant, we performed a logistic regression (due to binary outcome variable) that quantified the strength of the relationship between the number of peers who categorized the images as threatening and the participant’s own threat categorizations (i.e. if they categorized the image as threatening or safe themselves), controlling for participant’s initial distress ratings. In effect, we conducted the following logistic regression for each participant: logit(Y)=β 0 + β 1 Χ 1 + β 2 Χ 2 Where, Y = a vector containing the participant’s decisions to categorize images as threatening (coded as 1) or safe (coded as 0), β 0 = intercept of the logistic regression (not extracted or used for further analyses), β 1 = the coefficient that quantifies the strength of the relationship between the number of peers who categorized images as threatening and the participant’s tendency to categorize images as threatening (i.e., the participant’s conformity score ), X 1 = a vector containing the number of prior participants who ostensibly categorized images as threatening, β 2 = the coefficient that quantifies the strength of the relationship between the participant’s initial distress ratings and the participant’s tendency to categorize images as threatening (not extracted or used for further analyses), and X 2 = a vector containing the participant’s initial distress ratings for each image. Although this equation measures how strongly peers’ threat evaluations relate to the participant’s threat evaluations, the fact that threat/safety categorizations are perfectly anticorrelated (because they sum to 100 on every trial) means that replacing X 1 with peers’ safety evaluations would produce identical conformity scores, only with the sign reversed. High conformity scores (β 1 ) indicate that peers’ categorization behavior had a strong influence on participants’ categorizations at trial level, after controlling for their initial distress ratings. Low conformity scores (i.e., scores close to 0) indicate that peers’ categorization behavior had little influence on participants’ categorizations after controlling for their initial distress ratings. Negative conformity scores indicate that participants tended to provide categorizations that were opposed to peers’ categorizations, after controlling for their initial distress ratings. Additionally, we produced a measure of “emotional influence,” which uses a regression modeling approach to assess how strongly participants’ follow-up distress ratings were influenced by other participants’ categorization of the image as threatening or safe to share. For each subject, a multiple linear regression was conducted (due to continuous outcome variable) to compute the beta estimate assessing how strongly group categorizations related to participants’ follow-up distress ratings, after controlling for their initial distress ratings. In effect, we conducted the following regression for each participant: Y = β 0 + β 1 X 1 + β 2 X 2 Where, Y = a vector containing the participant’s follow-up distress rating, β 0 = intercept of the multiple linear regression (not extracted or used for further analyses), β 1 = the coefficient that quantifies the strength of the relationship between the number of peers who ostensibly categorized images as threatening and the participant’s follow-up distress rating (i.e., the participant’s emotion influence score ), X 1 = a vector containing the number of prior participants who ostensibly categorized images as threatening, β 2 = the coefficient that quantifies the strength of the relationship between the participant’s initial distress ratings and the participant’s follow-up distress rating (not extracted or used for further analyses), and X 2 = a vector containing the participant’s initial distress ratings for each image. We hypothesized that the degree to which participants incorporate peers’ threat information in the categorization of images (i.e. conformity score) relates to how strongly peers’ information influences their emotional responses to the images (i.e. emotional influence score). Computational approach. We again preregistered analyses in which parameters from computational models were used to test (i) whether peer evaluations shifted participants’ categorizations, (ii) people differed in their sensitivity to group threat or safety categorizations, and (iii) whether peer threat or safety categorizations influenced follow-up distress ratings. These findings again corroborated those below, although peer safety evaluations may be driving follow-up distress ratings. However, again due to concerns regarding the suitability of the computational model when peer threat and safety categorizations are perfectly anticorrelated, we warn against overinterpreting them and present these analyses in section II of the Online Resource, Section II.2. Results Replicating Study 1 In line with preregistered hypotheses, we found that participants’ decision to categorize an image as threatening was related to the number of peers who categorized it as threatening (OR = 1.01, 95% CI = [1.01, 1.02], p < .001), even after controlling for participants’ initial distress ratings (OR = 1.05, 95% CI = [1.05, 1.06], p < .001). Testing Influence of Peer Categorization on Distress Ratings In line with preregistered hypotheses, we observed that participants’ follow-up distress ratings were significantly related to the number of peers who categorized images as threatening ( β = .05, 95% CI = [.04, .07], p < .001), even after controlling for participants’ own initial distress ratings ( β = .94, 95% CI = [.92, .96], p < .001). This means that across the sample, participants’ follow-up distress rating was positively influenced by the number of peers who ostensibly categorized the image as threatening. Finally, Pearson correlation tests showed that emotional influence scores were significantly correlated with conformity scores ( r (114) = .38, p < .001) indicating that the extent to which a one adopts others’ categorizations is related to the extent to which they take on their emotional reactions. General Discussion Here we present two studies examining whether evaluations of situations as threatening or safe are influenced by social information and whether peer evaluations influence our feelings of distress in simulated online settings. We conducted two studies in which participants categorized negative images as threatening or safe for others to see while exposed to peers’ evaluations of these images. Results showed that individuals integrated peers’ evaluations with their own, and that doing so shifted their feelings of distress. All hypotheses were preregistered and the key result replicated across studies. These findings extend our current understanding of how people learn what is threatening or safe in their online environment by showing that peers’ threat evaluation of online content can propagate and emotionally influence others. People use online platforms to share images, videos and texts with others. This results in a massive, world-wide network of inter-connected data influencing users’ online as well as offline emotions and behaviors (Althoff et al., 2017 ). By determining that threat/safety information can be transmitted via observation of peers’ evaluations, our results are in line with the literature showing that threat/safety signals can be learned through social inputs (Golkar et al., 2013 ; Haaker et al., 2017 ) and indicate that such learning can also take place through observation of online behavior (e.g. clicking or liking). This opens up a novel experimental paradigm for investigating observational learning of threat/safety cues in a digital setting, with implications for research on threat/safety learning, social influence, and the spread of anxiety. Here we demonstrate online peer influence at two levels: evaluations of images as threatening/safe and one’s own emotional response to these images. Prior work has also shown that learning what is threatening or safe through observation of others can influence emotions (Higgins & Rholes, 1978 ; Nook et al., 2016 ; Prehn et al., 2013 ) and down-stream behaviors like the decision to approach or avoid stimuli ((Lindström et al., 2019b ). Our study extends this body of work by showing that how much individuals are emotionally influenced is correlated to how much they incorporated peers’ categorizations. More specifically, our results show that seeing that others evaluate something as threatening leads individuals to feel more distressed. While previous work indicates that exposure to social safety cues can immunize against observational fear learning (Golkar & Olsson, 2016 ), this study suggests that the observation of social safety cues online could as well prevent the maintenance of negative emotions due to exposure to negative content online. We also provide initial steps in developing a novel computational model for these processes, although we refrain from overinterpretation of modeling results due to the perfect collinearity of peer threat and safety categorizations. Nonetheless, we hope these initial innovations will be of use to future research on this topic. Our study has several strengths. The use of preregistration and within-paper replication demonstrate the reliability of our results. Moreover, the pretest/posttest design of Study 2 provided us with a measure of the change in individuals’ distress to the images and gave an insight on how emotional responses can shift. One challenge of the investigation of emotion influence online is that it is indeed difficult to estimate a change of emotion compared to a baseline and ensure that this change is due to the exposure to others’ behaviors or emotions (Goldenberg & Gross, 2020 ). One limitation is that this study did not unveil the underlying processes for changes in individuals’ emotional responses from pre- to posttest. Future work could investigate whether this emotional change reflects changes in individuals’ appraisals of the images based on peers’ evaluations (Gross, 1998 ) and/or whether these shifts are reflected in neurophysiological processing of images (e.g. late-positive potential (LPP); (Willroth et al., 2017 ). Finally, it is possible that our randomization procedures made some categorizations difficult to believe (e.g., a fairly neutral image being categorized as very threatening by peers). This may have led participants to doubt the categorizations were rated by others, which would affect their tendency to conform. However, participants’ behavior systematically changed to resemble peers’ behavior, suggesting that participants believed the ratings enough to be influenced by them. Taken together, the present research suggests that seeing how others evaluate threat/safety information influences not only how individuals evaluate this information, but also their emotional responses. These findings and experimental paradigms lay the foundation for several future lines of research. For example, it can be used to (i) identify the mediating mechanisms that explain the social transmission of threat, (ii) test interventions that block such transfer, and (iii) identify individual difference factors that exacerbate social threat transmission. Clarifying how we incorporate threat/safety information may assist in the development of interventions to prevent the disproportional magnification and the maladaptive effects of online dissemination of threatening information after major events (e.g., natural, disasters, political events, or terrorist acts). As such, the current findings contribute to scholars working on affective, social, digital, and clinical areas of research. Statements and Declarations This research was supported by the Knut and Alice Wallenberg Foundation (KAW 495 2014.0237) from KI Development (KID) grant (2-3591/2014) and a Consolidator Grant (2018-00877) from the Swedish Research Council (Vetenskapsrådet) to A. Olsson. E.C.Nook reports grants from a National Science Foundation Graduate Research Fellowship (DGE1144152) and Graduate Research Opportunities Worldwide (GROW) Fellowship. The authors report no conflict of interest. Ethical Approval Approval for this study was obtained by the Regional Ethical Review Board in Stockholm (Dnr: 2018/2511-31/5). Informed Consent: All participants provided written informed consent before their participation in both study 1 and 2. Funding This research was supported by the Knut and Alice Wallenberg Foundation (KAW 495 2014.0237) from KI Development (KID) grant (2-3591/2014) and a Consolidator Grant (2018-00877) from the Swedish Research Council (Vetenskapsrådet) to A. Olsson. E.C.Nook reports grants from a National Science Foundation Graduate Research Fellowship (DGE1144152) and Graduate Research Opportunities Worldwide (GROW) Fellowship. Conflict of Interest The authors have no relevant financial or non-financial interests to disclose. Availability of data and materials Sample sizes, data exclusions, hypotheses and analyses were preregistered for both Study 1 (https://osf.io/sjq2y) and Study 2 (https://osf.io/x6tna) prior to beginning data collection. Data, instructions to participants and analytic code are accessible at https://osf.io/vh46q/. The studies were collected on Amazon Mechanical Turk (MTurk) using custom code that is no longer compatible with the MTurk platform, but code can be shared upon request. References Althoff, T., Jindal, P., & Leskovec, J. (2017). Online Actions with Offline Impact. Proceedings of the Tenth ACM International Conference on Web Search and Data Mining , 537–546. https://doi.org/10.1145/3018661.3018672 Alvarez, M. J., & Kemmelmeier, M. (2018). Free speech as a cultural value in the United States. Journal of Social and Political Psychology, 5 (2), 707–735. https://doi.org/10.5964/jspp.v5i2.590 Anderson, M., & Auxier, B. (2021). Social Media Use in 2021. Pew Research Center, April , 1–6. Bellovary, A. K., Young, N. A., & Goldenberg, A. (2021). Left- and Right-Leaning News Organizations Use Negative Emotional Content and Elicit User Engagement Similarly. Affective Science. https://doi.org/10.1007/s42761-021-00046-w Brady, W. J., Crockett, M. J., & Van Bavel, J. J. (2020). The MAD Model of Moral Contagion: The Role of Motivation, Attention, and Design in the Spread of Moralized Content Online. Perspectives on Psychological Science, 15 (4), 978–1010. https://doi.org/10.1177/1745691620917336 Brady, W. J., Wills, J. A., Burkart, D., Jost, J. T., & Bavel, J. J. Van. (2019). An Ideological Asymmetry in the Diffusion of Moralized Content on Social Media Among Political Leaders. Journal of Experimental Psychology: General 2019, 148 (10), 1802–1813. Crockett, M. J. (2017). Moral outrage in the digital age. Nature Human Behaviour, 1 (11), 769–771. https://doi.org/10.1038/s41562-017-0213-3 Fan, R., Xu, K., & Zhao, J. (2020). Weak ties strengthen anger contagion in social media . 1–19. Ferrara, E., & Yang, Z. (2015). Measuring Emotional Contagion in Social Media. PLOS ONE, 10 (11), e0142390. https://doi.org/10.1371/journal.pone.0142390 Fuentes, A., & Peterson, J. V. (2021). Social media and public perception as core aspect of public health: The cautionary case of @realdonaldtrum and COVID-19. PLoS ONE, 16 (5), 1–9. https://doi.org/https://doi.org/10.1371/journal. pone.0251179 Goldenberg, A., & Gross, J. J. (2020). Digital Emotion Contagion. Trends in Cognitive Sciences, 24 (4), 316–328. https://doi.org/10.1016/j.tics.2020.01.009 Golkar, A., & Olsson, A. (2016). Immunization against social fear learning. Journal of Experimental Psychology: General, 145 (6), 665–671. https://doi.org/10.1037/xge0000173 Golkar, A., Selbing, I., Flygare, O., Öhman, A., & Olsson, A. (2013). Other People as Means to a Safe End. Psychological Science, 24 (11), 2182–2190. https://doi.org/10.1177/0956797613489890 Gross, J. J. (1998). The emerging field of emotion regulation: An integrative review. Review of General Psychology, 2 (3), 271–299. https://doi.org/10.1037/1089-2680.2.3.271 Haaker, J., Golkar, A., Selbing, I., & Olsson, A. (2017). Assessment of social transmission of threats in humans using observational fear conditioning. Nature Protocols, 12 (7), 1378–1386. https://doi.org/10.1038/nprot.2017.027 Haman, M. (2020). The use of Twitter by state leaders and its impact on the public during the COVID-19 pandemic. Heliyon. https://doi.org/doi.org/10.1016/j.heliyon.2020.e05540 Higgins, E. T., & Rholes, W. S. (1978). “Saying is believing”: Effects of message modification on memory and liking for the person described. Journal of Experimental Social Psychology, 14 (4), 363–378. https://doi.org/10.1016/0022-1031(78)90032-X Klucharev, V., Hytönen, K., Rijpkema, M., Smidts, A., & Fernández, G. (2009). Reinforcement Learning Signal Predicts Social Conformity. Neuron, 61 (1), 140–151. https://doi.org/10.1016/j.neuron.2008.11.027 Kramer, A. D. I., Guillory, J. E., & Hancock, J. T. (2014). Experimental evidence of massive-scale emotional contagion through social networks. Proceedings of the National Academy of Sciences , 111 (24), 8788–8790. https://doi.org/10.1073/pnas.1320040111 Kroenke, K., Spitzer, R. L., & Williams, J. B. W. (2001). The PHQ-9. Journal of General Internal Medicine, 16 (9), 606–613. https://doi.org/10.1046/j.1525-1497.2001.016009606.x Lindström, B., Golkar, A., Jangard, S., Tobler, P. N., & Olsson, A. (2019a). Social threat learning transfers to decision making in humans. Proceedings of the National Academy of Sciences of the United States of America. https://doi.org/10.1073/pnas.1810180116 Lindström, B., Golkar, A., Jangard, S., Tobler, P. N., & Olsson, A. (2019b). Social threat learning transfers to decision making in humans. Proceedings of the National Academy of Sciences , 116 (10), 4732–4737. https://doi.org/10.1073/pnas.1810180116 Lüdecke, D., Ben-Shachar, M., Patil, I., & Makowski, D. (2020). Extracting, Computing and Exploring the Parameters of Statistical Models using R. Journal of Open Source Software, 5 (53), 2445. https://doi.org/10.21105/joss.02445 Martin, R. E., Villanueva, Y., Stephano, T., Franz, P. J., & Ochsner, K. N. (2018). Social influence shifts valuation of appetitive cues in early adolescence and adulthood. Journal of Experimental Psychology: General, 147 (10), 1521–1530. https://doi.org/10.1037/xge0000469 Mason, W., & Suri, S. (2011). Conducting behavioral research on Amazon’s Mechanical Turk . https://doi.org/10.3758/s13428-011-0124-6 Nook, E. C. C., Ong, D. C. C., Morelli, S. A. A., Mitchell, J. P. P., & Zaki, J. (2016). Prosocial conformity: Prosocial norms generalize across behavior and empathy. Personality and Social Psychology Bulletin, 42 (8), 1045–1062. https://doi.org/10.1177/0146167298248001 Nook, E. C., Ong, D. C., Morelli, S. A., Mitchell, J. P., & Zaki, J. (2016). Prosocial Conformity. Personality and Social Psychology Bulletin, 42 (8), 1045–1062. https://doi.org/10.1177/0146167216649932 Nook, E. C., & Zaki, J. (2015). Social Norms Shift Behavioral and Neural Responses to Foods. Journal of Cognitive Neuroscience, 27 (7), 1412–1426. https://doi.org/10.1162/jocn_a_00795 Olsson, A., Knapska, E., & Lindström, B. (2020). The neural systems of social learning. Nature Reviews NeuroscieNce. https://doi.org/10.1038/s41583-020-0276-4 Olsson, A., & Phelps, E. A. (2004). Learned Fear of “Unseen” Faces after Pavlovian, Observational, and Instructed Fear. Psychological Science, 15 (12), 822–828. https://doi.org/10.1111/j.0956-7976.2004.00762.x Parkinson, B. (2011). Interpersonal emotion transfer: Contagion and social appraisal. Social and Personality Psychology Compass, 5 (7), 428–439. https://doi.org/10.1111/j.1751-9004.2011.00365.x Peters, K., & Kashima, Y. (2015). A multimodal theory of affect diffusion. Psychological Bulletin, 141 (5), 966–992. https://doi.org/10.1037/bul0000020 Phelps, E. A., O’Connor, K. J., Gatenby, J. C., Gore, J. C., Grillon, C., & Davis, M. (2001). Activation of the left amygdala to a cognitive representation of fear. Nature Neuroscience, 4 (4), 437–441. https://doi.org/10.1038/86110 Prehn, K., Korn, C. W., Bajbouj, M., Klann-Delius, G., Menninghaus, W., Jacobs, A. M., & Heekeren, H. R. (2013). The neural correlates of emotion alignment in social interaction. Social Cognitive and Affective Neuroscience, 10 (3), 435–443. https://doi.org/10.1093/scan/nsu066 Rozin, P., & Royzman, E. B. (2001). Negativity Bias, Negativity Dominance, and Contagion. In Personality and Social Psychology Review (Vol. 5, Issue 4). Schöne, J. P., Parkinson, B., & Goldenberg, A. (2021). Negativity Spreads More than Positivity on Twitter After Both Positive and Negative Political Situations. Affective Science, 379–390. https://doi.org/10.1007/s42761-021-00057-7 Spitzer, R. L., Kroenke, K., Williams, J. B., & Löwe, B. (2006). Generalized Anxiety Disorder 7-item (GAD-7) scale. Archives of Internal Medicine, 166 , 1092–1097. Wang, S.-Y. N., & Inbar, Y. (2022). Re-Examining the Spread of Moralized Rhetoric From Political Elites: Effects of Valence and Ideology. Journal of Experimental Psychology: General, 1–36. https://doi.org/10.7910/DVN/FQ8MIL Willroth, E. C., Koban, L., & Hilimire, M. R. (2017). Social Information Influences Emotional Experience and Late Positive Potential Response to Affective Pictures. Emotion, 17 (4), 572–576. Wilson, M., & Joseph, G. (2016). False Reports of Gunfire at J.F.K. Airport Offer a Real Case Study in Security. The New York Times, 13–15. Additional Declarations The authors declare no competing interests. Supplementary Files SocialTransmissionofThreatEvaluationsOnlineSupplResource.docx Online Supplementary Resource to accompany Peer Threat Evaluations Shape One’s Own Threat Perceptions and Feelings of Distress Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-3875288","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":267794267,"identity":"30f469d3-dfce-4c1b-8938-eaac5bcf1a96","order_by":0,"name":"Lisa Espinosa","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA6klEQVRIiWNgGAWjYBACPmYIzQOhKoCYmbkBrxY2VC1nQFoYCWhB4TG2gUkCWtjZH3748OeOjDl7+8MPP+fVRvO3A7X8qNiGx2E8xpIz257xWPacMZbs3XY8d8ZhxgbGnjO38WlhkOZtOMxjcCOHQYJ327HcBqAWZsY2fFrYH//+8wekJf3xz79zjuXOJ6yFwUyagQ2kJcEMaF1N7gbCWnjMLHuBfjE4c8bMWubYgdyNQC0H8fmFn//44xs//tyxNzje/vjmm5q63HnnDx988KMCtxYoOABjHEblEqOljgjFo2AUjIJRMNIAACUpWgWCy4Y2AAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0001-7327-6815","institution":"Karolinska Institutet, Stockholm, Sweden","correspondingAuthor":true,"prefix":"","firstName":"Lisa","middleName":"","lastName":"Espinosa","suffix":""},{"id":267794268,"identity":"8d7bf2e9-10b1-44c2-a8f3-afbc4b481127","order_by":1,"name":"Erik C. Nook","email":"","orcid":"https://orcid.org/0000-0001-7967-0792","institution":"Princeton University, New Jersey, Unites States","correspondingAuthor":false,"prefix":"","firstName":"Erik","middleName":"C.","lastName":"Nook","suffix":""},{"id":267794269,"identity":"d11b0ef6-6ae5-476c-9a0b-7af598f73b4e","order_by":2,"name":"Martin Asperholm","email":"","orcid":"https://orcid.org/0000-0002-6340-3638","institution":"Karolinska Institutet, Stockholm, Sweden","correspondingAuthor":false,"prefix":"","firstName":"Martin","middleName":"","lastName":"Asperholm","suffix":""},{"id":267794270,"identity":"a7b56412-3221-466b-9d88-d0a0f526c559","order_by":3,"name":"Therese Collins","email":"","orcid":"","institution":"Monash University, Melbourne, Australia","correspondingAuthor":false,"prefix":"","firstName":"Therese","middleName":"","lastName":"Collins","suffix":""},{"id":267794271,"identity":"8cbea520-089b-4c83-84f0-172ea1ebc35b","order_by":4,"name":"Juliet Y. Davidow","email":"","orcid":"https://orcid.org/0000-0002-6857-3855","institution":"Northeastern University, Boston, Unites States","correspondingAuthor":false,"prefix":"","firstName":"Juliet","middleName":"Y.","lastName":"Davidow","suffix":""},{"id":267794272,"identity":"9f6701b1-9ef8-4d8d-8f0a-62f4ced9f6b3","order_by":5,"name":"Andreas Olsson","email":"","orcid":"https://orcid.org/0000-0001-5272-7744","institution":"Karolinska Institutet, Stockholm, Sweden","correspondingAuthor":false,"prefix":"","firstName":"Andreas","middleName":"","lastName":"Olsson","suffix":""}],"badges":[],"createdAt":"2024-01-18 09:00:55","currentVersionCode":1,"declarations":{"humanSubjects":true,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":true,"humanSubjectConsent":true,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-3875288/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-3875288/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":49893953,"identity":"c3cf7c19-cb93-4dfd-80ef-fa7cb28ccd80","added_by":"auto","created_at":"2024-01-19 21:23:45","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":103348,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eExperimental Procedures for Studies 1 and 2. \u003c/em\u003e(A) In phase 1 (included in both Study 1 and Study 2), participants rated how distressed they felt when watching each of the 60 negative images. (B) In phase 2 (also included in both studies), participants watched the same images again and categorized each image as “threatening” (by clicking the red X) or “safe” (by clicking the green checkmark) for others to view and then “shared” this information by clicking the “SHARE” button. Below each image, participants saw numbers that ostensibly indicated how many previous participants had marked the image as threatening or safe. These values were in fact manipulated across images. (C) In phase 3 (included in Study 2 only), participants rated each image a second time on how distressed they made them feel. Note that the illustrations used in this Figure do not represent the images used in the studies, see details information and examples of the images in \u003cstrong\u003eTable ESM_1 \u003c/strong\u003ein the \u003cstrong\u003eOnline Resource\u003c/strong\u003e.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-3875288/v1/4efbb98d0c5ca7bf463ae869.png"},{"id":49894395,"identity":"2b472f2f-4e0b-489e-a7df-c46c288554cf","added_by":"auto","created_at":"2024-01-19 21:31:45","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":427177,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-3875288/v1/cfd5df93-c91a-4e2a-a386-37b16aa40a55.pdf"},{"id":49893954,"identity":"8ac383ed-a3ba-4b59-951d-1da8f05e4052","added_by":"auto","created_at":"2024-01-19 21:23:45","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":608206,"visible":true,"origin":"","legend":"\u003cp\u003eOnline Supplementary Resource to accompany Peer Threat Evaluations Shape One’s Own Threat Perceptions and Feelings of Distress\u003c/p\u003e","description":"","filename":"SocialTransmissionofThreatEvaluationsOnlineSupplResource.docx","url":"https://assets-eu.researchsquare.com/files/rs-3875288/v1/582de94f1bea057face76cb4.docx"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003ePeer Threat Evaluations Shape One’s Own Threat Perceptions and Feelings of Distress\u003c/p\u003e","fulltext":[{"header":"Introduction","content":"\u003cp\u003eSeven out of ten Americans report using social media platforms in their daily lives (Anderson \u0026amp; Auxier, 2021), making online content a central source of information exchange. Although social media offers opportunities to connect with others remotely, it also provides a medium through which emotional evaluations of events in the world around us can become amplified. For example, travelers in 2016 panicked and hid after reading online that there was a gunman in the JFK airport, even though no gunman existed (Wilson \u0026amp; Joseph, 2016). What psychological processes explain such rapid online transmission of threat evaluations and their concomitant emotional and behavioral responses?\u0026nbsp;Here, we develop an experimental paradigm for testing how social information shapes (i) evaluations of stimuli as threatening and safe, and (ii) actual emotional reactions to those stimuli.\u003c/p\u003e\n\u003cp\u003eA growing body of work has shown that positive and negative emotions spread between people, both in face-to-face interactions (Parkinson, 2011; Peters \u0026amp; Kashima, 2015) and online (Fan et al., 2020; Ferrara \u0026amp; Yang, 2015; Kramer et al., 2014). One experimental study investigating online emotion contagion manipulated the amount of emotional content Facebook users saw in their newsfeeds and showed that merely manipulating the number of positive and negative posts one viewed influences, respectively, the number of positive and negative posts those users then make (Kramer et al., 2014). Interestingly, it has been shown that people predominantly share negative (vs. positive) content online, regardless of whether the content is related to a negative or positive event, such as losing or winning a political election (Sch\u0026ouml;ne et al., 2021). This is \u0026nbsp;consistent with classic research showing a \u0026ldquo;negativity bias\u0026rdquo; in attention, memory, and conformity (Rozin \u0026amp; Royzman, 2001). Furthermore, anger seems to spread faster than joy on social media, as anger tweets are more likely to spread through even weak social ties than joyful tweets (Fan et al., 2020). Importantly, exposure to such negative content appears to lead to more online engagement (Bellovary et al., 2021) and spreads more widely (Brady et al., 2019; Wang \u0026amp; Inbar, 2022). Similarly, scientists have shown that the expression of emotions, such as moral outrage, spreads across users through design features of social media platforms (Crockett, 2017). Features, such as using notifications that capture and hold their attention, and using algorithms that learn from users\u0026rsquo; behavior, drive users to share moral-emotional content online by increasing motivation to share (Brady et al., 2020; Brady et al. 2023). As such, growing literature shows that specific emotional responses (e.g., outrage) and broader affective states (i.e., positivity and negativity) can flow through online social networks. More research is needed to move the focus beyond the momentary contamination of affective responses towards an understanding of how online information shapes emotions and more durable evaluations through learning. Here, we examined the processes underlying such learning by asking how information of others\u0026rsquo; evaluations of threat and safety are integrated with our own evaluations and decisions.\u003c/p\u003e\n\u003cp\u003eThe online transmission of threat evaluations has become central to our current experiences, as we have seen the power of online information in shaping evaluations of COVID-19 and other major events as either dangerous or harmless (Fuentes \u0026amp; Peterson, 2021; Haman, 2020). Anecdotal observations and a growing literature on social learning show that it is psychologically efficient for people to learn to evaluate stimuli as threatening/safe by observing other people\u0026rsquo;s reactions, rather than directly engaging with potentially threatening situations. Consistent with this, research has indicated that we learn similarly well when we observe others\u0026rsquo; threat responses or when told by others that something is dangerous, as when we learn through direct experience (Lindstr\u0026ouml;m et al., 2019a; Olsson \u0026amp; Phelps, 2004). These forms of social learning also involve similar neural and computational mechanisms as direct experience (Olsson et al., 2020). For example, people will show physiological responses to images that are associated with someone else receiving a shock (Haaker et al., 2017), and verbal instructions that a certain image will be followed by a shock produce similar levels of learning as either via direct experience or observation (Olsson et al., 2020; Phelps et al., 2001).\u003c/p\u003e\n\u003cp\u003eTo our knowledge however, it is unknown if threat/safety cues can be transferred through observation in an online setting and, if so, the impact of this social transmission on people\u0026rsquo;s actual emotional responses and decisions. For instance, research on conformity has shown that exposure to online peer norms can shift both behavior and deeper affective or neural responses to affective stimuli (Martin et al., 2018; Nook et al., 2016; Nook \u0026amp; Zaki, 2015; Prehn et al., 2013). These studies support the hypothesis that social perceptions of stimuli as threatening might influence both one\u0026rsquo;s own evaluations of those stimuli as threatening as well as actual feelings of distress. Here we unite these bodies of work through a novel experimental paradigm.\u003c/p\u003e\n\u003cp\u003eIn this study we investigate whether choosing to label situations as threatening or safe is influenced by seeing peer threat/safety evaluations, and whether peer evaluations influence our feelings of distress. Though we are interested in online spread of information, we developed an online paradigm with high degree of experimental control rather than using social media data to specifically manipulate peer threat and safety evaluations. The two current studies investigated if people conform to, propagate, and are emotionally influenced by peers\u0026rsquo; threat/safety evaluations in an online setting. In Study 1, we examined if receiving information about peers\u0026rsquo; evaluation of images as threatening or safe influences an individual\u0026rsquo;s decision to label and share these images as threatening or safe. \u0026nbsp;This first study established a suitable paradigm, and Study 2 replicated Study 1 and extended it to test if peers\u0026rsquo; evaluation of images as threatening or safe affects individual\u0026rsquo;s emotional response to that image by shifting their feelings of distress. Investigating how threat/safety information spreads through online networks can clarify the psychological processes explaining how social groups come to share evaluations of major events (e.g. pandemics, natural disasters and political events) as dangerous or safe, laying a foundation for interventions that can shape the online spread of threat evaluations.\u003c/p\u003e"},{"header":"Study 1","content":"\u003cp\u003e\u003cstrong\u003eMethod\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eParticipants\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;\u003c/strong\u003eOnline data collection took place on the MTurk platform. A total of 109 individuals completed the online experiment. Given widespread issues with \u0026ldquo;bot-like\u0026rdquo; activity on MTurk (Mason \u0026amp; Suri, 2011), we preregistered several exclusion criteria to ensure data were of high quality. This led us to exclude six participants who excessively clicked away from the window in which the task was administered (\u0026gt; 20 switches). Participants passed all other exclusion criteria (i.e., completing the task too quickly [2 SD under the mean] or failing \u0026ldquo;captcha\u0026rdquo; trials in which objects presented as images had to be named). Consequently, a total of 103 participants (44 females, age range = 21 - 71, \u003cem\u003eM\u003c/em\u003e\u003csub\u003eage\u003c/sub\u003e = 39.77, \u003cem\u003eSD\u003c/em\u003e\u003csub\u003eage\u003c/sub\u003e = 12.78) were included in the analyses.\u003c/p\u003e\n\u003cp\u003eSample size was established using power analysis based on estimated effect sizes produced by a pilot study with 43 individuals (39 usable following data quality exclusions). The pilot study, which was identical to the current study, showed that obtaining 80% power to replicate the smallest hypothesized effect (\u003cem\u003ed\u003c/em\u003e = 0.39) would require 54 participants. However, because of the risk for unusable data collected on MTurk (10% in our pilot study), we decided, prior to initiating data collection, to increase sample size to about 100 individuals, which would give us a power of 95% for that effect size. Approval for this study was obtained by the Regional Ethical Review Board in Stockholm (Dnr: 2018/2511-31/5).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eExperimental Paradigm and Procedure\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAfter providing informed consent, participants completed a study that included two phases (\u003cstrong\u003eFigure 1A \u0026amp; 1B\u003c/strong\u003e). During phase 1, participants were asked to rate how distressed 60 negative images made them feel on a sliding scale (0 = \u003cem\u003enot at all distressed\u003c/em\u003e, 100 = \u003cem\u003every distressed\u003c/em\u003e). Participants completed two practice trials before rating the 60 images used in the analyses. This task provided a measure of participants\u0026rsquo; initial affective response to the images. Images were presented one at a time in a randomized order for each participant. Images were drawn from the Open Affective Standardized Image Set (OASIS; Kurdi, Lozano, \u0026amp; Banaji, 2017) and contained aversive scenarios with a diverse set of themes including humans, animals, objects and scenes. Normed valence means ranged from mildly negative to very negative (see \u003cstrong\u003eTable S1\u003c/strong\u003e). \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eDuring phase 2, participants were informed that they were part of a large-scale study on creating rating norms for a set of images, and that 100 other MTurk workers had previously categorized the images they just saw as \u0026ldquo;threatening\u0026rdquo; or \u0026ldquo;safe\u0026rdquo;. We instructed participants to categorize an image as \u0026ldquo;threatening\u0026rdquo; if it \u0026ldquo;is likely to cause emotional distress to others\u0026rdquo; and to categorize it as \u0026ldquo;safe\u0026rdquo; if it \u0026ldquo;is not likely to cause emotional distress to others\u0026rdquo;. Participants were also led to believe that their own categorizations would add to the full set of answers that would be used in subsequent studies. Participants then saw the same images a second time along with the number of previous MTurk workers who had ostensibly categorized these pictures as either threatening or safe, respectively. However, these numbers were randomly generated on each trial. Participants categorized each image as threatening or safe by clicking on either a red X or a green checkmark, respectively. When clicking on either symbol, participants saw the number of previous categorizations increase by 1 for their chosen category. In order to mimic social media platforms in which people watch peers\u0026rsquo; responses to online content, express their own evaluation of the content, and share this with peers (i.e. others can see how they evaluate the content), participants were asked to click a \u0026ldquo;share\u0026rdquo; button to share their categorization for future participants to view.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAt the end of the study, participants reported their age, gender, and completed a set of questionnaires, including the balanced emotional empathy test (BEES; Mehrabian, 1996), the support for free speech scale (Alvarez \u0026amp; Kemmelmeier, 2018), the generalized anxiety disorder scale (Spitzer et al., 2006), the posttraumatic stress disorder checklist (PCL-5) and the patient-health questionnaire (Kroenke et al., 2001). For transparency, we report that these questionnaires were collected but we do not include them in the analyses. Finally, participants were debriefed on the study (including an explanation of the study\u0026rsquo;s deception) and were paid for their time.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eStatistical Analyses\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe preregistered a series of analyses involving computational modelling, however giving our concerns regarding the suitability of the computational modelling approach to this paradigm, for the main text we present a much simpler but conceptually identical analytic plan using mixed effect model. We present results from our preregistered analyses in the Online Resource, see section II and II.1.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo test the hypothesis that peers\u0026rsquo; threat categorizations influence participants\u0026rsquo; decision to categorize an image as threatening, we conducted a logistic mixed-effect model at the trial level testing the relationship between peers\u0026rsquo; threat categorizations and participants\u0026rsquo; own categorization for each image, controlling for participants\u0026rsquo; initial distress ratings for each image. Participant ID was included as a random effect to nest trials within participants. A significant relationship between the number of peers who categorized an image as threatening and participants\u0026rsquo; decisions to categorize images as threatening provides evidence of conformity to these group norms. All analyses used an alpha threshold of 0.05 and include measures of effect size. Specifically, we report odds ratios of each predictor, produced using the \u0026ldquo;model_parameters\u0026rdquo; function in the \u003cem\u003eparameters\u003c/em\u003e package (L\u0026uuml;decke et al., 2020).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults \u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn line with preregistered hypotheses, we found that participants\u0026rsquo; decision to categorize images as threatening was positively related to both their initial distress ratings (OR = 1.01, 95% CI = [1.01, 1.02], \u003cem\u003ep\u003c/em\u003e \u0026lt; .001) and the number of their peers who categorized the image as threatening (OR = 1.06, 95% CI = [1.05, 1.06], \u003cem\u003ep\u003c/em\u003e \u0026lt; .001). Thus, participants were more likely to categorize an image as threatening if a greater number of their peers did so, even after controlling for participants\u0026rsquo; own distress ratings.\u0026nbsp;\u003c/p\u003e"},{"header":"Study 2","content":"\u003cp\u003eIn Study 1 we showed that participants\u0026rsquo; decisions to categorize images as threatening or safe were influenced by peer threat/safety evaluations, even when controlling for participants\u0026rsquo; own feelings of distress. Because peer information was randomized on each trial, this indicates that threat evaluations can spread from person to person online. Our results however do not tell us whether conforming to peers\u0026rsquo; threat/safety evaluation influences individuals\u0026rsquo; emotional state. In Study 2, we sought to both replicate the results of Study 1 and extend them by investigating if conforming to peers\u0026rsquo; categorization of images would lead participants to update their emotions regarding these images. If so, this will demonstrate that exposure to online peer norms influences both one\u0026rsquo;s behavior and emotional state.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethod\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eParticipants\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eOnline data collection again took place on the MTurk platform. A total of 127 individuals completed the online experiment. We excluded 12 participants who failed to pass specific pre-determined data quality measures. Nine participants excessively clicked away from the window in which the task was administered (\u0026gt; 20 switches). Three participants were excluded because they displayed no variability in their data; They categorized all images as either \u0026ldquo;threatening\u0026rdquo; or \u0026ldquo;safe\u0026rdquo;, making it impossible to model their choice behavior. All remaining participants passed pre-determined quality measures. No participants completed the task too quickly (2 SD under the mean) or failed any of the \u0026ldquo;captcha\u0026rdquo; trials in which objects presented as images had to be named. Consequently, a total of 115 participants were included in analyses (55 females, age range = 21 - 71, \u003cem\u003eM\u003c/em\u003e\u003csub\u003eage\u003c/sub\u003e = 37.92, \u003cem\u003eSD\u003c/em\u003e\u003csub\u003eage\u003c/sub\u003e = 12.21). \u0026nbsp;We aimed to recruit the same number of participants as in Study 1 (i.e. 100 mTurk workers), and this sample size was established using power analyses based on estimated effect sizes produced by a pilot study with 43 mTurk workers (see preregistration: osf.io/sjq2y). Slightly more participants than this target number completed the study while it was active on MTurk. Approval for this study was obtained by the Regional Ethical Review Board in Stockholm (Dnr: 2018/2511-31/5).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eExperimental Paradigm and Procedure\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAfter providing informed consent, participants completed the same Phase 1 (initial distress rating) and 2 (categorization task) as in Study 1. However, Study 2 implemented four minor changes to Study 1\u0026rsquo;s procedures: (i) Eight images were estimated as too neutral (i.e. OASIS normed valence ratings) and were replaced in Study 2 (see \u003cstrong\u003eTable S1\u003c/strong\u003e). (ii) The numbers of ostensible previous categorizations as threatening or safe were randomly selected from a set of pre-determined numbers (and not randomly generated on each trial) to reduce noise across participants. (iii) The position of the threat and safe information on the left or right side of each image during Phase 2 was counterbalanced to remove potential confounds, i.e. 50 % of the participants saw the threat information on the left and right, respectively. (iv) Following the categorization phase, participants completed a third study phase in which they provided a follow-up distress rating for each image. This phase was identical to the initial distress rating (e.g. participants rated how distressed the negative images made them feel on a sliding scale from 0 = \u003cem\u003enot at all distressed\u003c/em\u003e to 100 = \u003cem\u003every distressed\u003c/em\u003e). This additional task provides a measure of participants\u0026rsquo; affective response to the images \u003cem\u003eafter\u003c/em\u003e being exposed to peers\u0026rsquo; categorizations. At the end of the study, participants reported their demographic information (i.e., age, gender), completed a set of questionnaires, were debriefed on the study (including an explanation of the study\u0026rsquo;s deception), and were paid for their time.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eStatistical Analyses\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eReplicating Study 1.\u0026nbsp;\u003c/strong\u003eWe first sought to replicate results of Study 1. As in Study 1, we ran a\u0026nbsp;trial-level logistic mixed-effect model to assess the relationship between peers\u0026rsquo; threat categorizations on participants\u0026rsquo; own categorization, controlling for participants\u0026rsquo; initial distress ratings. Participant ID was included as a random effect to nest trials within participants, and we again report odd\u0026rsquo;s ratios as effect sizes.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTesting influence of peer evaluations on follow-up distress ratings\u003c/strong\u003e. We also conducted a trial-level linear mixed-effect model to assess the relationship between peers\u0026rsquo; threat categorizations and participants\u0026rsquo; follow-up distress ratings. This regression included both participant ID as a random effect and participants\u0026rsquo; initial distress ratings. We hypothesized that peers\u0026rsquo; threat categorizations would influence follow-up distress ratings, over and above participants\u0026rsquo; initial distress ratings. Here, we report standardized betas as effect sizes, using the \u0026ldquo;standardize_parameters\u0026rdquo; function in the \u003cem\u003eparameters\u003c/em\u003e package (L\u0026uuml;decke et al., 2020).\u003c/p\u003e\n\u003cp\u003eAs an additional test of our hypotheses, we examined whether participants who more strongly take on peer threat categorizations are also more likely to update their distress ratings to match group norms. Specifically,\u003cstrong\u003e\u0026nbsp;w\u003c/strong\u003ee computed a conformity score for each participant (Klucharev et al., 2009; Nook \u0026amp; Zaki, 2015) in order to estimate the strength of the relationship between the number of peers who categorized the image as threatening and participants\u0026rsquo; decision to categorize the image as threatening controlling for participants\u0026rsquo; initial distress ratings. For each participant, we performed a logistic regression (due to binary outcome variable) that quantified the strength of the relationship between the number of peers who categorized the images as threatening and the participant\u0026rsquo;s own threat categorizations (i.e. if they categorized the image as threatening or safe themselves), controlling for participant\u0026rsquo;s initial distress ratings. In effect, we conducted the following logistic regression for each participant:\u003c/p\u003e\n\u003cp\u003elogit(Y)=\u0026beta;\u003csub\u003e0\u003c/sub\u003e+ \u0026beta;\u003csub\u003e1\u003c/sub\u003e\u0026Chi;\u003csub\u003e1\u003c/sub\u003e+ \u0026beta;\u003csub\u003e2\u003c/sub\u003e\u0026Chi;\u003csub\u003e2\u003c/sub\u003e\u003c/p\u003e\n\u003cp\u003eWhere, Y = a vector containing the participant\u0026rsquo;s decisions to categorize images as threatening (coded as 1) or safe (coded as 0), \u0026beta;\u003csub\u003e0\u003c/sub\u003e = intercept of the logistic regression (not extracted or used for further analyses), \u0026beta;\u003csub\u003e1\u003c/sub\u003e= the coefficient that quantifies the strength of the relationship between the number of peers who categorized images as threatening and the participant\u0026rsquo;s tendency to categorize images as threatening (i.e., the participant\u0026rsquo;s \u003cem\u003econformity score\u003c/em\u003e), X\u003csub\u003e1\u003c/sub\u003e = a vector containing the number of prior participants who ostensibly categorized images as threatening, \u0026beta;\u003csub\u003e2\u003c/sub\u003e= the coefficient that quantifies the strength of the relationship between the participant\u0026rsquo;s initial distress ratings and the participant\u0026rsquo;s tendency to categorize images as threatening (not extracted or used for further analyses), and X\u003csub\u003e2\u003c/sub\u003e = a vector containing the participant\u0026rsquo;s initial distress ratings for each image. Although this equation measures how strongly peers\u0026rsquo; threat evaluations relate to the participant\u0026rsquo;s threat evaluations, the fact that threat/safety categorizations are perfectly anticorrelated (because they sum to 100 on every trial) means that replacing X\u003csub\u003e1\u003c/sub\u003e with peers\u0026rsquo; safety evaluations would produce identical conformity scores, only with the sign reversed. High conformity scores (\u0026beta;\u003csub\u003e1\u003c/sub\u003e) indicate that peers\u0026rsquo; categorization behavior had a strong influence on participants\u0026rsquo; categorizations at trial level, after controlling for their initial distress ratings. Low conformity scores (i.e., scores close to 0) indicate that peers\u0026rsquo; categorization behavior had little influence on participants\u0026rsquo; categorizations after controlling for their initial distress ratings. Negative conformity scores indicate that participants tended to provide categorizations that were opposed to peers\u0026rsquo; categorizations, after controlling for their initial distress ratings.\u003c/p\u003e\n\u003cp\u003eAdditionally, we produced a measure of \u0026ldquo;emotional influence,\u0026rdquo; which uses a regression modeling approach to assess how strongly participants\u0026rsquo; follow-up distress ratings were influenced by other participants\u0026rsquo; categorization of the image as threatening or safe to share. For each subject, a multiple linear regression was conducted (due to continuous outcome variable) to compute the beta estimate assessing how strongly group categorizations related to participants\u0026rsquo; follow-up distress ratings, after controlling for their initial distress ratings. In effect, we conducted the following regression for each participant:\u003c/p\u003e\n\u003cp\u003eY\u0026nbsp;=\u0026nbsp;\u0026beta;\u003csub\u003e0 +\u0026nbsp;\u003c/sub\u003e\u0026beta;\u003csub\u003e1\u003c/sub\u003eX\u003csub\u003e1\u003c/sub\u003e + \u0026beta;\u003csub\u003e2\u003c/sub\u003eX\u003csub\u003e2\u003c/sub\u003e\u003c/p\u003e\n\u003cp\u003eWhere,\u003csub\u003e\u0026nbsp;\u003c/sub\u003eY = a vector containing the participant\u0026rsquo;s follow-up distress rating, \u0026beta;\u003csub\u003e0\u003c/sub\u003e = intercept of the multiple linear regression (not extracted or used for further analyses), \u0026beta;\u003csub\u003e1\u003c/sub\u003e= the coefficient that quantifies the strength of the relationship between the number of peers who ostensibly categorized images as threatening and the participant\u0026rsquo;s follow-up distress rating (i.e., the participant\u0026rsquo;s \u003cem\u003eemotion influence score\u003c/em\u003e), X\u003csub\u003e1\u003c/sub\u003e = a vector containing the number of prior participants who ostensibly categorized images as threatening, \u0026beta;\u003csub\u003e2\u003c/sub\u003e= the coefficient that quantifies the strength of the relationship between the participant\u0026rsquo;s initial distress ratings and the participant\u0026rsquo;s follow-up distress rating (not extracted or used for further analyses), and X\u003csub\u003e2\u003c/sub\u003e = a vector containing the participant\u0026rsquo;s initial distress ratings for each image. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe hypothesized that the degree to which participants incorporate peers\u0026rsquo; threat information in the categorization of images (i.e. conformity score) relates to how strongly peers\u0026rsquo; information influences their emotional responses to the images (i.e. emotional influence score).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eComputational approach. \u0026nbsp;\u003c/strong\u003eWe again preregistered analyses in which parameters from computational models were used to test (i) whether peer evaluations shifted participants\u0026rsquo; categorizations, (ii) people differed in their sensitivity to group threat or safety categorizations, and (iii) whether peer threat or safety categorizations influenced follow-up distress ratings. These findings again corroborated those below, although peer safety evaluations may be driving follow-up distress ratings. However, again due to concerns regarding the suitability of the computational model when peer threat and safety categorizations are perfectly anticorrelated, we warn against overinterpreting them and present these analyses in section II of the Online Resource, Section II.2.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults \u0026nbsp;\u003c/strong\u003e \u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eReplicating Study 1 \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn line with preregistered hypotheses, we found that participants\u0026rsquo; decision to categorize an image as threatening was related to the number of peers who categorized it as threatening (OR = 1.01, 95% CI = [1.01, 1.02], \u003cem\u003ep\u003c/em\u003e \u0026lt; .001), even after controlling for participants\u0026rsquo; initial distress ratings (OR = 1.05, 95% CI = [1.05, 1.06], \u003cem\u003ep\u003c/em\u003e \u0026lt; .001).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eTesting Influence of Peer Categorization on Distress Ratings\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/strong\u003eIn line with preregistered hypotheses, we observed that participants\u0026rsquo; follow-up distress ratings were significantly related to the number of peers who categorized images as threatening (\u003cem\u003e\u0026beta;\u003c/em\u003e = .05, 95% CI = [.04, .07],\u003cem\u003e\u0026nbsp;p\u003c/em\u003e \u0026lt; .001), even after controlling for participants\u0026rsquo; own initial distress ratings (\u003cem\u003e\u0026beta;\u003c/em\u003e = .94, 95% CI = [.92, .96],\u003cem\u003e\u0026nbsp;p\u003c/em\u003e \u0026lt; .001). This means that across the sample, participants\u0026rsquo; follow-up distress rating was positively influenced by the number of peers who ostensibly categorized the image as threatening. Finally, Pearson correlation tests showed that emotional influence scores were significantly correlated with conformity scores (\u003cem\u003er\u003c/em\u003e(114) = .38, \u003cem\u003ep\u0026nbsp;\u003c/em\u003e\u0026lt; .001) indicating that the extent to which a one adopts others\u0026rsquo; categorizations is related to the extent to which they take on their emotional reactions.\u003c/p\u003e"},{"header":"General Discussion","content":"\u003cp\u003eHere we present two studies examining whether evaluations of situations as threatening or safe are influenced by social information and whether peer evaluations influence our feelings of distress in simulated online settings. We conducted two studies in which participants categorized negative images as threatening or safe for others to see while exposed to peers\u0026rsquo; evaluations of these images. Results showed that individuals integrated peers\u0026rsquo; evaluations with their own, and that doing so shifted their feelings of distress. All hypotheses were preregistered and the key result replicated across studies. These findings extend our current understanding of how people learn what is threatening or safe in their online environment by showing that peers\u0026rsquo; threat evaluation of online content can propagate and emotionally influence others.\u003c/p\u003e \u003cp\u003ePeople use online platforms to share images, videos and texts with others. This results in a massive, world-wide network of inter-connected data influencing users\u0026rsquo; online as well as offline emotions and behaviors (Althoff et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). By determining that threat/safety information can be transmitted via observation of peers\u0026rsquo; evaluations, our results are in line with the literature showing that threat/safety signals can be learned through social inputs (Golkar et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; Haaker et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2017\u003c/span\u003e) and indicate that such learning can also take place through observation of online behavior (e.g. clicking or liking). This opens up a novel experimental paradigm for investigating observational learning of threat/safety cues in a digital setting, with implications for research on threat/safety learning, social influence, and the spread of anxiety.\u003c/p\u003e \u003cp\u003eHere we demonstrate online peer influence at two levels: evaluations of images as threatening/safe and one\u0026rsquo;s own emotional response to these images. Prior work has also shown that learning what is threatening or safe through observation of others can influence emotions (Higgins \u0026amp; Rholes, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e1978\u003c/span\u003e; Nook et al., \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Prehn et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2013\u003c/span\u003e) and down-stream behaviors like the decision to approach or avoid stimuli ((Lindstr\u0026ouml;m et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2019b\u003c/span\u003e). Our study extends this body of work by showing that how much individuals are emotionally influenced is correlated to how much they incorporated peers\u0026rsquo; categorizations. More specifically, our results show that seeing that others evaluate something as threatening leads individuals to feel more distressed. While previous work indicates that exposure to social safety cues can immunize against observational fear learning (Golkar \u0026amp; Olsson, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2016\u003c/span\u003e), this study suggests that the observation of social safety cues online could as well prevent the maintenance of negative emotions due to exposure to negative content online. We also provide initial steps in developing a novel computational model for these processes, although we refrain from overinterpretation of modeling results due to the perfect collinearity of peer threat and safety categorizations. Nonetheless, we hope these initial innovations will be of use to future research on this topic.\u003c/p\u003e \u003cp\u003eOur study has several strengths. The use of preregistration and within-paper replication demonstrate the reliability of our results. Moreover, the pretest/posttest design of Study 2 provided us with a measure of the change in individuals\u0026rsquo; distress to the images and gave an insight on how emotional responses can shift. One challenge of the investigation of emotion influence online is that it is indeed difficult to estimate a change of emotion compared to a baseline and ensure that this change is due to the exposure to others\u0026rsquo; behaviors or emotions (Goldenberg \u0026amp; Gross, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). One limitation is that this study did not unveil the underlying processes for changes in individuals\u0026rsquo; emotional responses from pre- to posttest. Future work could investigate whether this emotional change reflects changes in individuals\u0026rsquo; appraisals of the images based on peers\u0026rsquo; evaluations (Gross, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e1998\u003c/span\u003e) and/or whether these shifts are reflected in neurophysiological processing of images (e.g. late-positive potential (LPP); (Willroth et al., \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). Finally, it is possible that our randomization procedures made some categorizations difficult to believe (e.g., a fairly neutral image being categorized as very threatening by peers). This may have led participants to doubt the categorizations were rated by others, which would affect their tendency to conform. However, participants\u0026rsquo; behavior systematically changed to resemble peers\u0026rsquo; behavior, suggesting that participants believed the ratings enough to be influenced by them.\u003c/p\u003e \u003cp\u003eTaken together, the present research suggests that seeing how others evaluate threat/safety information influences not only how individuals evaluate this information, but also their emotional responses. These findings and experimental paradigms lay the foundation for several future lines of research. For example, it can be used to (i) identify the mediating mechanisms that explain the social transmission of threat, (ii) test interventions that block such transfer, and (iii) identify individual difference factors that exacerbate social threat transmission. Clarifying how we incorporate threat/safety information may assist in the development of interventions to prevent the disproportional magnification and the maladaptive effects of online dissemination of threatening information after major events (e.g., natural, disasters, political events, or terrorist acts). As such, the current findings contribute to scholars working on affective, social, digital, and clinical areas of research.\u003c/p\u003e"},{"header":"Statements and Declarations","content":"\u003cp\u003eThis research was supported by the Knut and Alice Wallenberg Foundation (KAW 495 2014.0237) from\u003c/p\u003e\n\u003cp\u003eKI Development (KID) grant (2-3591/2014) and a Consolidator Grant (2018-00877) from the Swedish\u003c/p\u003e\n\u003cp\u003eResearch Council (Vetenskapsr\u0026aring;det) to A. Olsson. E.C.Nook reports grants from a National Science\u003c/p\u003e\n\u003cp\u003eFoundation Graduate Research Fellowship (DGE1144152) and Graduate Research Opportunities\u003c/p\u003e\n\u003cp\u003eWorldwide (GROW) Fellowship.\u003c/p\u003e\n\u003cp\u003eThe authors report no conflict of interest.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthical Approval\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eApproval for this study was obtained by the Regional Ethical Review Board in Stockholm (Dnr: 2018/2511-31/5).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eInformed Consent: \u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll participants provided written informed consent before their participation in both study 1 and 2.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research was supported by the Knut and Alice Wallenberg Foundation (KAW 495 2014.0237) from KI Development (KID) grant (2-3591/2014) and a Consolidator Grant (2018-00877) from the Swedish Research Council (Vetenskapsr\u0026aring;det) to A. Olsson. E.C.Nook reports grants from a National Science Foundation Graduate Research Fellowship (DGE1144152) and Graduate Research Opportunities Worldwide (GROW) Fellowship. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflict of Interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors have no relevant financial or non-financial interests to disclose.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSample sizes, data exclusions, hypotheses and analyses were preregistered for both Study 1 (https://osf.io/sjq2y) and Study 2 (https://osf.io/x6tna) prior to beginning data collection. Data, instructions to participants and analytic code are accessible at https://osf.io/vh46q/. The studies were collected on Amazon Mechanical Turk (MTurk) using custom code that is no longer compatible with the MTurk platform, but code can be shared upon request.\u003cstrong\u003e\u003cbr\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003e\u003cspan\u003eAlthoff, T., Jindal, P., \u0026amp; Leskovec, J. (2017). Online Actions with Offline Impact. \u003cem\u003eProceedings of the Tenth ACM International Conference on Web Search and Data Mining\u003c/em\u003e, 537\u0026ndash;546. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1145/3018661.3018672\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eAlvarez, M. J., \u0026amp; Kemmelmeier, M. (2018). Free speech as a cultural value in the United States. Journal of Social and Political Psychology, \u003cem\u003e5\u003c/em\u003e(2), 707\u0026ndash;735. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.5964/jspp.v5i2.590\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eAnderson, M., \u0026amp; Auxier, B. (2021). Social Media Use in 2021. Pew Research Center, \u003cem\u003eApril\u003c/em\u003e, 1\u0026ndash;6.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eBellovary, A. K., Young, N. A., \u0026amp; Goldenberg, A. (2021). Left- and Right-Leaning News Organizations Use Negative Emotional Content and Elicit User Engagement Similarly. Affective Science. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s42761-021-00046-w\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eBrady, W. J., Crockett, M. J., \u0026amp; Van Bavel, J. J. (2020). The MAD Model of Moral Contagion: The Role of Motivation, Attention, and Design in the Spread of Moralized Content Online. Perspectives on Psychological Science, \u003cem\u003e15\u003c/em\u003e(4), 978\u0026ndash;1010. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/1745691620917336\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eBrady, W. J., Wills, J. A., Burkart, D., Jost, J. T., \u0026amp; Bavel, J. J. Van. (2019). An Ideological Asymmetry in the Diffusion of Moralized Content on Social Media Among Political Leaders. \u003cem\u003eJournal of Experimental Psychology: General 2019, 148\u003c/em\u003e(10), 1802\u0026ndash;1813.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eCrockett, M. J. (2017). Moral outrage in the digital age. Nature Human Behaviour, \u003cem\u003e1\u003c/em\u003e(11), 769\u0026ndash;771. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41562-017-0213-3\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eFan, R., Xu, K., \u0026amp; Zhao, J. (2020). \u003cem\u003eWeak ties strengthen anger contagion in social media\u003c/em\u003e. 1\u0026ndash;19.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eFerrara, E., \u0026amp; Yang, Z. (2015). Measuring Emotional Contagion in Social Media. PLOS ONE, \u003cem\u003e10\u003c/em\u003e(11), e0142390. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1371/journal.pone.0142390\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eFuentes, A., \u0026amp; Peterson, J. V. (2021). Social media and public perception as core aspect of public health: The cautionary case of @realdonaldtrum and COVID-19. PLoS ONE, \u003cem\u003e16\u003c/em\u003e(5), 1\u0026ndash;9. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/https://doi.org/10.1371/journal. pone.0251179\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eGoldenberg, A., \u0026amp; Gross, J. J. (2020). Digital Emotion Contagion. Trends in Cognitive Sciences, \u003cem\u003e24\u003c/em\u003e(4), 316\u0026ndash;328. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.tics.2020.01.009\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eGolkar, A., \u0026amp; Olsson, A. (2016). Immunization against social fear learning. Journal of Experimental Psychology: General, \u003cem\u003e145\u003c/em\u003e(6), 665\u0026ndash;671. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1037/xge0000173\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eGolkar, A., Selbing, I., Flygare, O., \u0026Ouml;hman, A., \u0026amp; Olsson, A. (2013). Other People as Means to a Safe End. Psychological Science, \u003cem\u003e24\u003c/em\u003e(11), 2182\u0026ndash;2190. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/0956797613489890\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eGross, J. J. (1998). The emerging field of emotion regulation: An integrative review. Review of General Psychology, \u003cem\u003e2\u003c/em\u003e(3), 271\u0026ndash;299. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1037/1089-2680.2.3.271\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eHaaker, J., Golkar, A., Selbing, I., \u0026amp; Olsson, A. (2017). Assessment of social transmission of threats in humans using observational fear conditioning. Nature Protocols, \u003cem\u003e12\u003c/em\u003e(7), 1378\u0026ndash;1386. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/nprot.2017.027\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eHaman, M. (2020). The use of Twitter by state leaders and its impact on the public during the COVID-19 pandemic. Heliyon. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/doi.org/10.1016/j.heliyon.2020.e05540\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eHiggins, E. T., \u0026amp; Rholes, W. S. (1978). \u0026ldquo;Saying is believing\u0026rdquo;: Effects of message modification on memory and liking for the person described. Journal of Experimental Social Psychology, \u003cem\u003e14\u003c/em\u003e(4), 363\u0026ndash;378. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/0022-1031(78)90032-X\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eKlucharev, V., Hyt\u0026ouml;nen, K., Rijpkema, M., Smidts, A., \u0026amp; Fern\u0026aacute;ndez, G. (2009). Reinforcement Learning Signal Predicts Social Conformity. Neuron, \u003cem\u003e61\u003c/em\u003e(1), 140\u0026ndash;151. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.neuron.2008.11.027\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eKramer, A. D. I., Guillory, J. E., \u0026amp; Hancock, J. T. (2014). Experimental evidence of massive-scale emotional contagion through social networks. \u003cem\u003eProceedings of the National Academy of Sciences\u003c/em\u003e, \u003cem\u003e111\u003c/em\u003e(24), 8788\u0026ndash;8790. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1073/pnas.1320040111\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eKroenke, K., Spitzer, R. L., \u0026amp; Williams, J. B. W. (2001). The PHQ-9. Journal of General Internal Medicine, \u003cem\u003e16\u003c/em\u003e(9), 606\u0026ndash;613. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1046/j.1525-1497.2001.016009606.x\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eLindstr\u0026ouml;m, B., Golkar, A., Jangard, S., Tobler, P. N., \u0026amp; Olsson, A. (2019a). Social threat learning transfers to decision making in humans. Proceedings of the National Academy of Sciences of the United States of America. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1073/pnas.1810180116\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eLindstr\u0026ouml;m, B., Golkar, A., Jangard, S., Tobler, P. N., \u0026amp; Olsson, A. (2019b). Social threat learning transfers to decision making in humans. \u003cem\u003eProceedings of the National Academy of Sciences\u003c/em\u003e, \u003cem\u003e116\u003c/em\u003e(10), 4732\u0026ndash;4737. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1073/pnas.1810180116\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eL\u0026uuml;decke, D., Ben-Shachar, M., Patil, I., \u0026amp; Makowski, D. (2020). Extracting, Computing and Exploring the Parameters of Statistical Models using R. Journal of Open Source Software, \u003cem\u003e5\u003c/em\u003e(53), 2445. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.21105/joss.02445\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eMartin, R. E., Villanueva, Y., Stephano, T., Franz, P. J., \u0026amp; Ochsner, K. N. (2018). Social influence shifts valuation of appetitive cues in early adolescence and adulthood. Journal of Experimental Psychology: General, \u003cem\u003e147\u003c/em\u003e(10), 1521\u0026ndash;1530. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1037/xge0000469\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eMason, W., \u0026amp; Suri, S. (2011). \u003cem\u003eConducting behavioral research on Amazon\u0026rsquo;s Mechanical Turk\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3758/s13428-011-0124-6\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eNook, E. C. C., Ong, D. C. C., Morelli, S. A. A., Mitchell, J. P. P., \u0026amp; Zaki, J. (2016). Prosocial conformity: Prosocial norms generalize across behavior and empathy. Personality and Social Psychology Bulletin, \u003cem\u003e42\u003c/em\u003e(8), 1045\u0026ndash;1062. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/0146167298248001\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eNook, E. C., Ong, D. C., Morelli, S. A., Mitchell, J. P., \u0026amp; Zaki, J. (2016). Prosocial Conformity. Personality and Social Psychology Bulletin, \u003cem\u003e42\u003c/em\u003e(8), 1045\u0026ndash;1062. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/0146167216649932\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eNook, E. C., \u0026amp; Zaki, J. (2015). Social Norms Shift Behavioral and Neural Responses to Foods. Journal of Cognitive Neuroscience, \u003cem\u003e27\u003c/em\u003e(7), 1412\u0026ndash;1426. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1162/jocn_a_00795\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eOlsson, A., Knapska, E., \u0026amp; Lindstr\u0026ouml;m, B. (2020). The neural systems of social learning. Nature Reviews NeuroscieNce. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41583-020-0276-4\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eOlsson, A., \u0026amp; Phelps, E. A. (2004). Learned Fear of \u0026ldquo;Unseen\u0026rdquo; Faces after Pavlovian, Observational, and Instructed Fear. Psychological Science, \u003cem\u003e15\u003c/em\u003e(12), 822\u0026ndash;828. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/j.0956-7976.2004.00762.x\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eParkinson, B. (2011). Interpersonal emotion transfer: Contagion and social appraisal. Social and Personality Psychology Compass, \u003cem\u003e5\u003c/em\u003e(7), 428\u0026ndash;439. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/j.1751-9004.2011.00365.x\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003ePeters, K., \u0026amp; Kashima, Y. (2015). A multimodal theory of affect diffusion. Psychological Bulletin, \u003cem\u003e141\u003c/em\u003e(5), 966\u0026ndash;992. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1037/bul0000020\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003ePhelps, E. A., O\u0026rsquo;Connor, K. J., Gatenby, J. C., Gore, J. C., Grillon, C., \u0026amp; Davis, M. (2001). Activation of the left amygdala to a cognitive representation of fear. Nature Neuroscience, \u003cem\u003e4\u003c/em\u003e(4), 437\u0026ndash;441. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/86110\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003ePrehn, K., Korn, C. W., Bajbouj, M., Klann-Delius, G., Menninghaus, W., Jacobs, A. M., \u0026amp; Heekeren, H. R. (2013). The neural correlates of emotion alignment in social interaction. Social Cognitive and Affective Neuroscience, \u003cem\u003e10\u003c/em\u003e(3), 435\u0026ndash;443. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/scan/nsu066\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eRozin, P., \u0026amp; Royzman, E. B. (2001). Negativity Bias, Negativity Dominance, and Contagion. In \u003cem\u003ePersonality and Social Psychology Review\u003c/em\u003e (Vol. 5, Issue 4).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eSch\u0026ouml;ne, J. P., Parkinson, B., \u0026amp; Goldenberg, A. (2021). Negativity Spreads More than Positivity on Twitter After Both Positive and Negative Political Situations. Affective Science, 379\u0026ndash;390. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s42761-021-00057-7\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eSpitzer, R. L., Kroenke, K., Williams, J. B., \u0026amp; L\u0026ouml;we, B. (2006). Generalized Anxiety Disorder 7-item (GAD-7) scale. Archives of Internal Medicine, \u003cem\u003e166\u003c/em\u003e, 1092\u0026ndash;1097.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eWang, S.-Y. N., \u0026amp; Inbar, Y. (2022). Re-Examining the Spread of Moralized Rhetoric From Political Elites: Effects of Valence and Ideology. Journal of Experimental Psychology: General, 1\u0026ndash;36. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.7910/DVN/FQ8MIL\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eWillroth, E. C., Koban, L., \u0026amp; Hilimire, M. R. (2017). Social Information Influences Emotional Experience and Late Positive Potential Response to Affective Pictures. Emotion, \u003cem\u003e17\u003c/em\u003e(4), 572\u0026ndash;576.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eWilson, M., \u0026amp; Joseph, G. (2016). False Reports of Gunfire at J.F.K. Airport Offer a Real Case Study in Security. The New York Times, 13\u0026ndash;15.\u003c/span\u003e\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"social learning, emotional influence, threat and safety learning, online study","lastPublishedDoi":"10.21203/rs.3.rs-3875288/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-3875288/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eWe are continuously exposed to what others think and feel about content online. How do others’ evaluations shared in this medium influence our own beliefs and emotional responses? In two pre-registered studies, we investigated the social transmission of threat\u003cem\u003e \u003c/em\u003eand safety evaluations in a paradigm that mimicked online social media platforms. In Study 1 (N=103), participants viewed images and indicated how distressed they made them feel. Participants then categorized these images as threatening or safe for others to see, while seeing how “previous participants” ostensibly categorized these images (these values were actually manipulated across images). We found that participants incorporated both peers’ categorizations of the images and their own distress ratings when categorizing images as threatening or safe. Study 2 (N=115) replicated these findings and further demonstrated that peers’ categorizations shifted how distressed these images made them feel. Taken together, our results indicate that people integrate their own and others’ experiences when exposed to emotional content and that social information can influence both our perceptions of things as threatening or safe, as well as our own emotional responses to them. Our findings provide replicable experimental evidence that social information is a powerful conduit for the transmission of affective evaluations and experiences.\u003c/p\u003e","manuscriptTitle":"Peer Threat Evaluations Shape One’s Own Threat Perceptions and Feelings of Distress","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-01-19 21:23:40","doi":"10.21203/rs.3.rs-3875288/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"2725597c-dca0-447e-ab4b-c42e5a465f83","owner":[],"postedDate":"January 19th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":28216342,"name":"Psychology"},{"id":28216343,"name":"Cognitive Neuroscience"}],"tags":[],"updatedAt":"2024-01-19T21:23:40+00:00","versionOfRecord":[],"versionCreatedAt":"2024-01-19 21:23:40","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-3875288","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-3875288","identity":"rs-3875288","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}