YouTube remains one of the most popular platforms globally, with influencers attracting the largest subscriber bases across a wide range of genres and content types. Influencers consistently dominate rankings among content categories on YouTube, largely due to their strong focus on entertainment, including high-production value games, stunt and challenge videos, and fun and social experiments. Notably, recent trends in philanthropic initiatives often featuring large monetary giveaways, such as awareness campaigns and charity streams, have added to their increasing popularity and driven significant online engagement among emerging adults due to the portrayal of generosity blended with strong entertainment elements.

Content depicting such altruistic modelling behaviours, such as materialistic and emotional sharing, whether intentional or unintentional, presents a double-edged sword. On the one hand, it highlights altruistic efforts, such as helping those in need, addressing eudaimonic needs for audiences like learning and fostering a sense of purpose or moral elevation. On the other hand, it risks masking potential deception regarding the business model of those channels, promoting misconceptions about monetary spending and unrealistic standards of altruistic behaviours among emerging adults.

Despite the growing prevalence of such content, our understanding of its characteristics and the audience's reception remains limited. In response, this study aims to identify the characteristics of entertainment-based influencer content and explore how altruistic efforts are depicted—whether as genuinely others-focused (ultimate) or as self-gain-focused (instrumental). Furthermore, the study examines audience reactions through comments on YouTube videos to assess whether viewers perceive these behaviours as altruistic or instrumental.

Using quantitative content analysis, the study maps key characteristics of altruistic efforts on YouTube and analyzes audience responses to these portrayals. This approach aims to indirectly explore how such content influences audience perceptions of altruism by inspiring while providing entertainment.

Artificial intelligence (AI) can be a powerful tool in educational contexts, particularly since the advent of generative AI. For instance, AI chatbots were found to enhance students’ performance, motivation, and self-efficacy (Wu & Yu, 2024). Further, a survey of 6,311 students from universities in Germany indicates that almost two out of three students (63.4%) use AI tools (von Garrel & Mayer, 2023). Previous research suggested several constructs that may co-determine why students’ use generative AI tools (e.g., Abbas et al., 2024; Schei et al., 2024). However, the relative importance of personal and motivational constructs for students’ intention to use generative AI tools in different academic contexts is yet unclear. Hence, we investigated the relative importance of selected constructs from the technology acceptance model (Venkatesh & Bala, 2008) and self-determination theory (Ryan & Deci, 2020) for students’ intention to use generative AI tools regarding learning, oral exams, written exams, or practical exams (RQ1). Further, we examined the value added of self-determination theory in terms of motivational constructs (RQ2) and basic psychological needs (RQ3). Finally, we investigated students’ most important reasons for and against using generative AI tools in their studies (RQ4).

We conducted a preregistered mixed-methods study (https://osf.io/5nxy4/?view_only=0267ae9b15a34e05bb03c6767923d4f0) with 244 university students (Mage = 24.49, 187 female, 57 male). Scales for all constructs but experience were adapted from existing scales, resulting in acceptable internal consistency for all constructs (all Cronbach’s α >= 0.70) but trust in AI (α = 0.69) and result demonstrability (α = 0.60). Quantitative data were analyzed via multiple linear regression models (RQ1-RQ3) (bootstrapping with 5.000 iterations). Qualitative data were analyzed via inductive content analysis (RQ4) by using 20% of the data to create an initial category system that was iteratively revised. Test requirements were met, and the sample size exceeded the minimum sample size (n = 217, f2 = 0.15, β = 0.95).

Body image (BI) is a multifaceted construct involving the feelings, thoughts, and perceptions of individuals related to their own body appearance. The formation of BI begins in children around the age of five and becomes an essential part of their identity as they transition into young adulthood. However, concerns about BI are now widespread worldwide and are associated with mental health problems. One widely studied factor in influencing negative BI is media. Shifting from traditional media, current research focuses on the impact of social media applications. One of the most popular, highly visual media platforms is Instagram, which is gaining attention in BI research due to its variety of algorithm-based content, uses and features, and largest group of users (i.e., young people). Research has shown Instagram to be negatively related to the body satisfaction of women and men. Given this relationship, analysing potential protective factors for BI concerns is one key task for current research. This study aimed to assess the role of mindfulness as a protective factor against BI concerns of young people (18-34 years) in context of Instagram use. Mindfulness is the ability of being present in the moment and non-judgementally accepting one’s internal experiences. It has been positively related to well-being, self-esteem, and body satisfaction. We hypothesized that mindfulness along with its five facets (according to Baer et al., 2006) (a) is negatively related to Instagram use and (b) moderates the relationship between Instagram use and BI concerns. For this purpose, we measured multiple aspects of Instagram use (i.e., average time, amount of appearance content viewed, familiarity with content & application), trait mindfulness, and BI concerns (feeling “fat” and drive for muscularity) in an online survey. Multiple regression and moderation models (N=554; 66% female) were run to test our hypotheses.

In today's digital age, media consumption has become an integral part of daily life. While digital media offers numerous advantages, it also serves as a powerful source of distraction. The ability to manage cognitive resources and resist immediate rewards in favor of long-term goals has thus become increasingly relevant. This study examines the relationship between media usage and the ability to delay gratification.

Previous research indicates a potentially negative association between media use—particularly social media, online videos, media multitasking, and mobile device usage—and delay of gratification. Additionally, studies suggest that the motivation behind media consumption may influence self-regulation and impulsivity.

This study employs a survey-based quantitative approach to explore how different media types and usage motivations impact impulsivity and self-control. An online survey of N = 151 participants collected data on media usage habits and behaviors as related to delay of gratification. We applied four well-established self-report instruments: the Monetary Choice Questionnaire (MCQ), the Barratt Impulsiveness Scale (BIS), the Self-Control Scale (SCS), and the Delay of Gratification Inventory (DGI). Our findings contribute to a deeper understanding of the interplay between media consumption and self-regulatory capacities.

In a digitally driven world where social media plays an essential role in shaping body perceptions, previous research has revealed that exposure to idealized body portrayals often leads to reduced body satisfaction and self-esteem, further relating to mental health issues such as depression, anxiety, as well as eating disorders (for a meta-analysis, see de Valle et al., 2021). As a potential counter-trend, the body positivity movement (BoPo) advocates for body diversity and self-acceptance. Recently, virtual influencers (VIs)––highly human-like, computer-generated characters with their own social media accounts (e.g., @ lilmiquela)––have gained attention alongside humans for their potential to either reinforce or challenge dominant beauty standards, particularly thin- and fit-ideal body types. Mainly drawing on the Tripartite Influence Model (Thompson et al., 1999), the CASA paradigm (Computer Are Social Actors; Nass & Moon, 2000), and body image research, we conducted two preregistered experiments in the BoPo context to investigate the effects of highly realistic VIs’ body representations on young adult women’s body satisfaction and perceived body ideals (Experiment 1: https://aspredicted.org/5t7g-2f9z.pdf; Experiment 2: https://aspredicted.org/kqfc-8w6n.pdf). In a nuanced differentiation, Experiment 1 examined whether the depicted body types of VIs (thin-ideal vs. larger-sized) would elicit similar psychological effects on body satisfaction and perceived body ideals as those of HIs, additionally considering presented captions, perceived human-likeness, and previous exposure to BoPo content as moderators. Subsequently, Experiment 2 focused on identifying the differential effects of body-related images in athletic contexts, manipulating both the influencer type (VIs vs. HIs) and body type (larger-sized bodies vs. fit-ideal representations, e.g., toned bodies). It was assumed that the impact of VIs’ appearances would be weaker than those of HIs due to their artificial nature. Additionally, perceived similarity was considered as an individual psychological factor, reflecting the subjective sense of connection between the individual and the influencer.

Medical practitioners increasingly share sexual health knowledge on social media but balancing professional and personal self-disclosure is challenging. Personal self-disclosure can undermine perceived expertise, affecting influencers’ persuasiveness. This study conducted a 2 × 2 between-subject experiment (N = 322) to examine how medical social media influencers’ (SMIs) self-disclosure influences audience attitudes. Results showed that professional self-disclosure increased source credibility, leading to more positive attitudes toward knowledge posts. However, it did not enhance parasocial interactions. Personal self-disclosure had no interaction effects with professional self-disclosure on dependent variables. Findings suggested that source credibility is the primary driver of attitudinal change, as followers prioritize influencers’ expertise over relatability. While parasocial interactions can still influence attitudes, authoritative influencers may struggle to foster them due to their perceived social distance. This study offers insights into self-disclosure strategies in health knowledge distribution, particularly for reproductive health campaigns, highlighting the importance of credibility in medical SMIs’ knowledge dissemination.

Emojis have become an integral part of digital communication, yet their influence on interpersonal perceptions is still not fully understood. While previous research has examined their role in expressing emotions and enhancing message clarity, little is known about how emojis shape the perception of the sender’s likability in private messenger conversations. This study investigates whether and to what extent the use of emojis in one-to-one digital communication affects perceived likability and whether this effect is moderated by the recipient’s own emoji usage frequency. An online survey (N = 176, aged 13 to 78, 65.9% female) was conducted, in which participants rated chat messages and their authors based on various characteristics. Additionally, they provided insights into their personal emoji usage. The study aims to contribute to a deeper understanding of the social functions of emojis in digital communication and to fill a gap in research regarding their role in interpersonal impression formation.

The integration of technology into sex work has fundamentally reshaped the profession, with cam modeling emerging as a key example of how digital platforms mediate intimacy, autonomy, and identity. Cam modeling refers to the practice of performing live-streamed erotic or sexual content for an online audience, often involving real-time interaction with viewers. Despite its growing prevalence, cam modeling remains understudied, and discussions are often shaped by stereotypes rather than empirical insights. Understanding the psychological dimensions of this profession, including motivations, job satisfaction, and the impact of technological risks, is essential for gaining a more nuanced perspective on modern sex work and its socio-economic implications.

The present work analyzes a unique dataset of 331 cam models across 27 countries, representing one of the largest empirical investigations into this profession. Using a mixed-methods approach, this study explores key factors influencing entry into cam modeling, the role of money as both an initial motivator and a factor in long-term job satisfaction, and the impact of digital platforms on professional identity. Additionally, technological risks such as doxxing, content leaks, and online harassment are examined, highlighting the critical role of privacy in shaping cam models' well-being and career longevity.

By offering an data-driven perspective on cam modeling, the present work contributes to broader discussions on the intersection of technology, sex work, and psychological well-being. The findings have implications for digital platform governance, privacy policies, and the evolving nature of online intimacy, while also challenging prevailing assumptions about the financial, psychological, and technological realities of this profession.

Reflection is an essential step in the process of learning from experiences. To learn effectively, it is important to apply knowledge, reflect upon the experiences made, and in doing so make the connection between theory and experience. This knowledge can then be applied in future situations. Performance in reflective writing can be improved by applying reflective stages, such as Gibb’s Reflective Cycle, and questions guiding along the reflection process. This study investigates if deploying a large language model (LLM) that interactively guides the user along the reflection process improves the users’ reflective competencies. After developing a suitable prompt for the LLM, a user study was conducted, exploring user’s perception of the LLM. In a between-subjects online study, 240 participants were divided into four groups, all reflecting upon a previously written term paper. One group received written information about Gibb’s Reflective Cycle, while the second one reflected using written guided questions. The third group interacted with the LLM in form of a chatbot, the fourth in form of an animated, speech-based agent. All conditions were designed based on Gibb’s Reflective Cycle. Differences in performance were analyzed assessing inter-rater reliability as well as evaluating the use of cognitive words. Furthermore, the Technology Acceptance Model (perceived usefulness, ease of use), attitudes toward AI, trust, and intention to use were investigated as dependent variables in both LLM-conditions to further investigate the users’ perception of the LLM.

Background. The rapid global spread of smartphone technology has profoundly influenced daily life, including family dynamics. A growing body of research highlights the impact of smartphone use on social interactions, particularly when one person's attention to their device leaves their interaction partner feeling ignored—a phenomenon known as phubbing. This behavior is not limited to general social interactions but also occurs in parent-child relationships (parental phubbing).

Prior research suggests that children who experience parental phubbing exhibit higher rates of smartphone use themselves, have an increased risk of mental health problems such as depression, and lower satisfaction of their need for relatedness.

Most existing studies on parental phubbing have been cross-sectional, limiting insights into its dynamic effects over time. To address this gap, our study employs a longitudinal design that includes both mothers and fathers, allowing for a more comprehensive examination of how parental phubbing affects child well-being in daily life.

In this preregistered daily diary study, we investigate both intra- and interindividual associations between parental phubbing and children’s affective well-being in everyday family contexts.

Methods.N = 40 parental couples (i.e., 80 individuals) reported their phubbing behavior toward their child and the child's well-being (i.e., positive and negative affect, perceived fundamental need threat) over 10 consecutive evenings. Parental phubbing was assessed using a single self-report item rated separately by both parents on a 7-point scale. Positive affect was measured by averaging two items on happiness and relaxation, while negative affect was assessed via six items related to sadness, stress, and worry. Fundamental need threat was evaluated using five items adapted from the need threat scale by Williams et al. (2000). We conducted double intercept multilevel models to account for the dependencies in the dyadic data and to investigate the associations between parental phubbing and children's well-being at both the between- and within-person levels.

Previous studies have mainly dealt with positive parasocial relationships (PSRs; e.g. media personae as friends). In contrast, relationships with non-amicable media characters have only been studied sporadically (Click & Tukachinsky, 2023). As negative parasocial relationships (NPSRs) can also comprise strong emotional reactions of media recepients (e.g. hate), the study aimed to delve deeper into the prerequisites and consequences of NPSRs. Building on a pilot study (N=58), where we found that recipients cite in particular politicians as the most frequent category of disliked media figures and the Threat-based Hate Model (Martinez et al., 2022), it was hypothesized that higher perceptions of individual (real and symbolic) threat, that is triggered by the media persona will reinforce recipients' action tendency in the sense of avoidance (and not approach). It was also assumed, that this effect is mediated by the negative affects (extent of perceived aversion) and the quality and intensity of the parasocial relationship with the persona. A further aim of the study was to compare negative parasocial relationships towards politicians and other categories of media actors (e.g. presenters, actors, hosts). An online-survey (N=766) was conducted and participants were randomly assigned to name either a politician or another media actor (non-politician) they do not like or have negative feelings towards. Subjects completed scales to capture perceived threat (threat perception scale; Martinez et al., 2022), negative emotions towards the chosen persona (celebrity hate scale; Sternberg, 2008), negative parasocial relationship (NPSR-scale; Hartmann et al., 2008) and action tendencies (Click & Tukachinsky, 2023; Pretus et al., 2021). The effects of perceived threat on action tendencies towards the persona (approach versus avoidance) were tested in a mediation model.

The increasing use of AI tools in various tasks highlights the need to understand cognitive effort in AI interactions. While preliminary research suggests that AI tool use can negatively impact critical thinking through cognitive offloading, it may also enhance learning success (Gehrlich, 2025; Deng et al., 2025). The amount of invested mental effort (AIME) explains how students allocate cognitive resources based on perceived demand characteristics of the task, context, or stimulus and perceived self-efficacy, with higher AIME leading to better learning outcomes (Salomon, 1984). Although AIME has been studied in traditional and digital media (Beentjes, 1989; Schwab et al., 2018), little is known about its role in AI tool use. This exploratory study examines how students engage with AI tools in study/work and leisure contexts and how individual factors, such as personality traits, AI attitudes, and success/failure attributions, influence AIME. An online survey of university students (N = 214) assessed AI tool use (tool type, frequency, purpose), AIME, perceived self-efficacy, and success/failure attributions (Salomon, 1984; Beentjes, 1989). Additionally, Big Five personality traits (Rammstedt et al., 2012), Need for Cognition (Beißert et al., 2015), and AI attitudes (Stein et al., 2024) were measured to explore their relationship with AIME. By analyzing both quantitative and qualitative data, this study provides insights into the cognitive effort invested in AI tool use, highlighting the psychological factors that shape user engagement. Addressing the conference theme "Changing Media Environments as Facilitators for Learning?", our findings contribute to research on cognitive effort in AI interactions and inform AI-based educational design to optimize engagement and learning outcomes.

Despite being one of the most widely used types of media content, the use of (online) pornography, as well as its predictors and (potential) effects, remains an under-researched topic. However, studying pornography use can be challenging due to its stigmatized nature and resulting difficulties with measurement methods, especially if these are solely based on self-reports. In an earlier study, we used a combination of web tracking (only top-level domains, e.g., youporn.com) and survey data from German internet users to reassess key findings from previous research on (online) pornography research that were based on self-reports. For the present study, we were able to obtain full URLs for a subset of visits to the most popular online pornography sites (pornhub.com, xhamster.com, xnxx.com, xvideos.com, youporn.com) and apply automated procedures for classifying the visited URLs with keywords. The final classified dataset contains data from N = 873 users, spanning 452,149 pornography website visits classified with 106 distinct keywords. In the next step, we used keyword-assisted topic models to identify topics based on the keywords per URL. The proportion of these topics per user is our indicator of content preference. Based on this, we assess differences in and predictors of content preferences in online pornography use. In addition to descriptive analyses of content preference patterns, we investigate how (well) key sociodemographic attributes (esp. sex and age) as well as values and attitudes (sexism, social dominance orientation, religiosity) predict online pornography content preferences. At the conference, we will present the results of these analyses and discuss their implications for research on (online) pornography use.

Receiving sexually loaded messages is a common part of adolescents’ online experiences (Madigan et al., 2018). When such messages are unexpected, they are usually perceived as bothersome (Finkelhor et al., 2000). Nevertheless, after receiving them, adolescents rarely seek social support from parents (Mishna, 2023), despite evidence suggesting this can be an effective coping strategy (Kaiser et al., 2020; Manay, 2020).

Our mixed-method study has two aims. First, we use quantitative data to describe adolescents’ experiences with online sexual solicitations (i.e., unwanted requests for online sexual interactions, Schulz et al., 2016) and their disclosure of such experiences to others. Second, we use qualitative data to investigate the reasoning behind their choice not to disclose these experiences to their parents. Identifying the patterns of non-disclosure and barriers preventing adolescents from reaching out to their parents is essential for setting up effective prevention strategies and, subsequently, mitigating the potential harm of unsolicited sexual experiences online.

We use survey data from 1,500 Czech adolescents (ages 13-18, 52.3% female) collected online in 2022. Quotas for household income and region ensured the sample mirrors Czech families. Using closed items, adolescents were asked how many times they experienced online sexual solicitation, defined as receiving unwanted messages or requests of a sexual nature; how they felt after (very pleased-very upset); and how often they confided to parents, peers, professionals, or others (never-every time). Moreover, 871 adolescents shared their reasons for not disclosing such experiences to parents (open question with 5 text fields; data include in total 1,397 filled text fields with relevant answers). We analyze the answers using Structured Tabular Thematic Analysis (ST-TA) to identify occurring themes in adolescents’ reasons for non-disclosure and examine how these reasons differ based on age and gender.

Recently, body positivity (#BoPo) gained a lot of attention on social media, in media coverage, and in research. The social movement advocates for the love and acceptance of all bodies, sizes and shapes and fights against unrealistic beauty ideals (Cwynar-Horta, 2016; Cohen et al., 2019). It provides content that can, for instance, improve recipients’ body self-perceptions (e.g., Fardouly et al., 2023; Rodgers et al., 2022), though it is not without controversy (e.g., Sharp et al., 2023). This raises questions about how the movement and its messages are perceived, how they are assessed, and which factors influence this assessment. Following an explorative attempt, we asked:

RQ1) Who is aware of #BoPo?

RQ2) Who interacts with #BoPo-content on social media?

RQ3) What factors influence people’s support for (RQ3a) and identification with the movement (RQ3b)?

In cooperation with a professional survey institute, we conducted an online survey among German Internet users aged 18-64 (N=1426; age: M=43.4 years; males: 691, females: 726, diverse: 9). A quotation-procedure (criteria: age, gender, living region in Germany) was applied to increase the representativeness.

Participants were asked whether they were aware of the #BoPo-movement (yes/no). Further, scales from Gerson et al. (2017; passive/active use) and Feinberg et al. (2020; support for social movement & social identification) were adapted to #BoPo to measure interaction, support and identification.

Based on the existing literature and theoretical considerations, we selected established instruments to measure different self-concept components (feminist identity, political orientation) and psychosocial characteristics (body appreciation, eating disorder symptoms, self-esteem, self-objectification beliefs, social comparison orientation) that we considered predictors of engagement with #BoPo-content. We conducted t-tests to compare characteristics of participants who were aware vs. unaware of #BoPo (RQ1). Further, linear regressions were estimated to examine the relationships between sociodemographic and psychosocial factors and interaction, support, and identification with #BoPo (RQ2 and 3).

In recent years, countermovements like body positivity have surged on highly visual social media platforms like Instagram and TikTok in response to the edited, unrealistic appearance-focused content. The movement encourages to embrace self-acceptance, and appreciate diverse appearances, body sizes, and shapes. Research by Cohen et al. (2019) shows that online body-positive content fosters a more positive body image (e.g., increased body appreciation) compared to thin-ideal or neutral content. Yet, the impact of social media on individuals with clinical eating disorders (ED) remains unclear. While studies suggest that social media use can exacerbate ED behaviors (Dane & Bhatia, 2023), some research suggests that social media use among those with clinical ED may not always be harmful, depending on the willingness to change: Those less inclined to confront their ED thoughts and behaviors are more likely to use social media for validation and managing their reputation (Dingemans et al., 2023). Alternatively, the question is whether body-appreciating movements such as body positivity can aid recovery in people with clinical ED. Thereto, this study investigates how body positivity impacts their body image in contrast to the generally portrayed ideally slender appearances, and explores the directing role of social comparison motives (self-evaluation, self-improvement, self-evaluation) orientation herein.

Proposed relations are tested in an experimental design, with online body imagery (body positivity vs. thin-ideal vs. neutral) as between-subjects factor, and body appreciation and dissatisfaction as within-subjects factors (T0-T1). In total, 153 adult women with clinical ED symptomatology (Mage=25.03, SDage=7.89) viewed images showcasing either female bodies conforming to the body positivity trend (n=54), the thin equivalent (n=48), or neutral control images (n=51). Furthermore, we test the moderating effect of upward social comparison with other women's appearances encountered online. Our findings contribute to understanding the impact of social media on public health issues and clinical practices for ED.

Theory: Previous research has shown that viewers often form connections with characters in media productions, even desiring to develop friendships with them (Horton & Wohl, 1956). However, when it comes to villains, viewers’ attitudes become more complex. On one hand, viewers typically despise the immoral behavior of the villain. On the other hand, they are often drawn to these characters, valuing unique traits such as being unconventional (Greenwood et al., 2021). Villains are often crucial for creating a more in-depth and layered storytelling experience. In the context of the Joker, the most famous villain from the DC universe, viewers appear to highly value this character. However, the portrayal of the Joker varies widely. In some cases, the Joker wins over viewers’ hearts due to his skillful portrayal of evil and/or by a portrayal that is pretty depressed such as in the 2019 movie Joker. The portrayal of a character has huge impact on how individuals feel connected to a character, and thus, in turn on their media enjoyment and mood repair.

Research Question: To what extent do the different media portrayals of the Joker impact character engagement and hence, media enjoyment and mood repair?

Methods: Participants (N = 134) were randomly assigned to two groups. First, all participants indicated their affect (mood and arousal via the affect grid). Afterwards, all participants had to write a short essay about a moment in life that made them feel sad followed by another assessment of their affect. Next, participants either watched a crazy evil portrayal or a depressed one of the Joker. After watching the clips, participants’ affect was assessed for the third time and they answered questions on character engagement (parasocial relationships and recognizability) and media enjoyment (eudaimonic/hedonic entertainment experiences and mood repair). All scales were found to be reliable.

The research of emotions, thoughts or actions triggered by intense watching sessions of TV-series seems to be limited by two circumstances. First, there is a general scarcity of concepts and questionnaires that focus on the timeframe after a media use experience. Reception phenomena such as identification, engagement, transportation or enjoyment have been well-researched in the context of binge-watching. But all these media effects are framed to only occur during the exposure. Accordingly, there is a general need for a closer look on how the time-consuming exposure to TV-series impacts our everyday experiences away from screentime. Second, there is a strong focus on mental health outcomes. Currently, at least three distinct questionnaires exist for measuring addiction symptoms in the specific case of binge-watching. Depressive symptoms have been also frequently investigated as a potential outcome of binge-watching. After years of research into the mental health dangers of binge-watching, it is time to take a more holistic approach.

The aim of this study was to generate an authentic multidimensional instrument that does justice to the variety and duration of possible binge-watching effects. The items were developed inductively based on a qualitative study that accessed the hours and days after intense TV-series watching experiences of 24 young adults via semi-structured interviews. The final coding system and related key statements of the interviewees were used to generate 148 items. The original item-pool was then revised for clarity, simplicity, and readability until the final sample of 70 items was reached. The items were then tested in an online survey (N=102). All items referred to a specific TV-series participants were highly engaged with. The item-pool comprised three dimensions a priori: thoughts, affects and actions. In order to determine the psychometric properties as well as the dimensionality of the item pool factor analyses and correlational analyses were performed.

On social media, misinformation (i.e., “information that is false, inaccurate, or misleading”, (Pennycook & Rand, 2021, p. 389) is not only spread about political issues, climate change, and COVID-19 (Lewandowsky et al., 2023) but also about brands (Berthon & Pitt, 2018). For example, false posts have circulated claiming that Lamborghini faked Nürburgring lap times and that Apple deleted music collections when customers signed up for Apple Music. Misinformation about brands spreads widely on X (formerly Twitter): For instance, 42% of the 5,000 tweets in the FakeAds data corpus were annotated as misinformation (Alnazzawi et al., 2022). Such negative misinformation may have detrimental effects on word-of-mouth recommendations, attitudes toward the brand, and consumers’ purchase intentions (e.g., Mishra & Samu, 2021, Experiment 3). Ultimately, it may reduce a brand's sales and market share. In the context of political misinformation and health-related misinformation, interventions have been developed to counter false information (Ecker et al., 2022). In our preregistered experimental study (https://aspredicted.org/sfn5-wdmk.pdf), we investigated whether three interventions had an effect in the context of brand-related misinformation. Participants (N = 153) were presented with either (1) videos about misinformation (Roozenbeek et al., 2022), (2) infographics about misinformation, (3) a short training about informal fallacies (Hruschka & Appel, 2023), or a text about marketing (i.e., the control condition). We hypothesized that training on informal fallacies, presenting infographics, or showing videos would enhance discernment between misinformation and true information about brands, buffer against loss of trust in the brand, increase purchase intentions, and increase word-of-mouth recommendations. Trust in eight brands (Sheinin et al., 2011, adopted by Fârte & Obadă, 2021) was assessed before and after the intervention. Additionally, we measured discernment between eight false and eight true social media posts on X about the brands, word-of-mouth (Morhart et al., 2015; Price & Arnould, 1999), and purchase intention (two items from Spears & Singh, 2004) for each brand.

Users of social networking sites (SNS) are caught between the desire for authentic self-presentation and the pressure to present themselves in an idealized way (Uski & Lampinen, 2016). In response to this conflict, the SNS BeReal aims to promote authenticity by randomly triggering posting demands and restricting users’ control over self-presentation.

Recent research rooted in Self-Determination Theory (Ryan & Deci, 2000) suggests that digital autonomy consists of control (i.e., the technology-centered component) and authenticity (i.e., the user-centered experience; Dietrich et al., 2024). Authenticity in this context refers to the expression of one's self aligned with personal values, free from external pressures (Wood et al., 2008). This framework suggests that authenticity can still emerge even when SNS limit control (Deci & Ryan, 2000, p. 236). This raises the question of whether users can feel autonomous during their BeReal use despite the SNS’ self-presentational boundaries.

As perceived platform affordances (Fox & McEwan, 2017; Gibson, 1979) are situated at the interface between platform design and user perceptions, they can provide insights into how platform features shape individual experiences (Evans et al., 2017; Kreling et al., 2022). Thus, combining these perspectives, we investigate how the perceived affordances of limited editability, availability, and persistence influence users’ experiences of autonomy need satisfaction vs. frustration on BeReal. We hypothesize that limited editability and increased availability will positively impact autonomy need frustration and negatively satisfaction, as they limit self-presentation and control over the time of use (H1-H2). However, we expect the short persistence of posts to be positively related to autonomy need satisfaction and negatively to frustration, as the ephemerality of posts encourages more spontaneous self-presentation (H3). A two-week diary study with regular BeReal users (N = 55 participants, T = 502 daily surveys) investigates these hypotheses on a daily level, untangling within- and between-person relationships.

Theoretical Background

Communication with Generative AI (so-called “chatbots”) has become an almost omnipresent topic within university context. In this study we took a step back to examine underlying cognitive principles of communication. Taking another person's knowledge into account is referred to as perspective-taking. It is employed during communication, e.g. during speech production and reception (Clark, 1992). However, research has also pointed to a lack of perspective-taking and cognitive biases resulting in non-adapted (“mis”) communication (Horton & Keysar, 1996). The usage of privileged information is such a well-documented instance. Keysar (1994) used an experimental design, revealing “an illusory transparency of intention” in interpersonal communication. This means that people might mistakenly believe their conversation partner will understand the intention of a message, even when they received no clue to this intention.

Research question

Is the phenomenon “illusory transparency of intention” applied to communication with Large Language Models?

Method

We developed different scenarios describing typical situations in university students lives. Using a 1 x 2 within-subject design, privileged information about the underlying perspective of the student communicating with the chatbot was either positively or negatively framed. Participants read a total of 11 scenarios (three each within the negative and positive condition and five fillers). After each scenario they rated the addressees’ perception of sarcasm which indicates whether they (erroneously) included privileged information. In a second round they were asked about their own perception of sarcasm. Additionally, two open questions about the chatbots understanding and what it might answer aimed to collect deeper insights.

Due to technological changes, new forms of informal learning have emerged. Especially online videos such as YouTube play an important role, as studies have shown that they can have positive effects on variables such as motivation and learning performance (Shoufan & Mohamed, 2022).

Research has also shown that protagonist variables such as perceived trustworthiness and expertise or perceived human conversational voice predict positive evaluations of How-to videos on YouTube (Utz & Wolfers, 2022).

We aim to conceptually replicate this work in the context of educational videos. Since a person’s self-concept has a significant influence on their actual performance (Eccles & Wigfield, 2000), we want to investigate the relationship between educational video-consumption (operationalized as frequency of watching) and mathematical self-concept under the additional influence of said protagonist characteristics to evaluate a possible influence of YouTube could have on the viewers’ mathematical self-concept.

For this reason, we examined a YouTube Channel about mathematics on university level. 579 participants were recruited via several social media channels and completed an Online questionnaire. We assessed media consumption, mathematical self-concept, perceived trustworthiness, perceived expertise and perceived conversational human voice, among others. The study was preregistered prior to data collection.

Our hypotheses are:

(1) There is a positive correlation between the consumption of YouTube videos and viewers' mathematical self-concept.

(2) Perceived trustworthiness, perceived expertise and perceived conversational human voice have a mediating influence on this relationship.

The data is analyzed as a multiple mediator analysis with media consumption as the predictor, mathematical self-concept as the criterium and perceived trustworthiness, perceived expertise and perceived conversational human voice as parallel mediators. We believe that our work will contribute to a better understanding of the interplay of trustworthiness variables in the context of informal learning.

Theory

Learning diaries as formative assessments are promising to support the learning process and stimulate reflection. In an open learning diary, learners apply learning strategies to reflect on their learning and deepen their knowledge. To guide learners, learning diaries can be structured according to different learning strategies. However, grading and feedback on learning diaries is effortful for teachers. Artificial intelligence (AI) may assist teachers in the evaluation process. A prerequisite is that AI and teachers show high inter-rater agreement.

Research Questions

The current study aimed to analyze the agreement between teachers and ChatGPT-4o by examining four separately assessed learning strategy categories of a learning diary in adult education (organization, in-depth elaboration, transfer-supporting elaboration, and metacognition). The two research questions were:

• Q1: How accurate is the overall agreement between teachers and ChatGPT-4o, and are there differences across different learning strategy categories?

• Q2: Does the inter-rater agreement differ between teachers across the four learning strategy categories?

Method

Seven different adult education teachers and ChatGPT-4o evaluated a total of 540 learning diary entries. Each teacher assessed approximately 65 entries. Teachers were trained in criteria-based evaluation per learning strategy category. An engineered prompt supported the ChatGPT-4o model.

Teacher ratings served as the reference for the inter-rater agreement. Absolute accuracy and under-/overestimation (bias) were calculated for each learning strategy category as accuracy measures. Furthermore, overall accuracy values were calculated across the four categories for absolute accuracy and bias.

A doubly multivariate repeated measures ANOVA was conducted with the four learning strategy categories as repeated measures and absolute accuracy and bias as measures to test for accuracy differences between the learning strategy categories (Q1). Additionally, teacher was used as a between-subjects factor. Thus, the interaction of the learning strategy category and rating teacher regarding accuracy could be tested statistically (Q2).

Children should learn proper media handling early, as media are omnipresent and unavoidable. Since media literacy is not innate but develops through exposure, it is advisable to actively promote this skill starting in preschool. Media sign literacy (MSL) is the most fundamental aspect of media literacy developing at preschool age and focusing on the understanding and correct use of the signs and symbol systems that organize media. Research shows that MSL increases with age and that the media literacy of older children and adults can be improved through interventions. Therefore, it was assumed that this also applies to preschoolers, although this has not yet been examined. Based on this research foundation, a training program to promote MSL in preschoolers was developed and evaluated in this study. It was also explored whether MSL’s positive impact on academic precursors could be maximized through training. Furthermore, the MSL training group was compared with an active control group that completed a classic intelligence training, as intelligence has been found to be a relevant covariate before and shares a common variance with MSL An experimental pretest-posttest study with N = 152 preschoolers, aged 4 to 5 years, tested these assumptions. The experimental group (n = 49) participated in the self-developed MSL training, while the active control group (n = 53) completed Denkspiele mit Elfe und Mathis (a digital version of Klauer’s inductive reasoning training). Both training programs consisted of 10 sessions, implemented on tablets in preschools over approximately four weeks. A passive control group (n = 49) continued regular preschool activities. Participants' MSL and various intelligence subtests (matrices, classifications, similarities, information) were measured before and after the interventions. Additionally, academic precursor skills (mathematical abilities, letter knowledge, and phonological awareness) were assessed as dependent variables. Children's ages were controlled for.

Recent achievements in the field of artificial intelligence (AI) promise significant changes in many different areas, including education (Mohebi, 2024). Due to the effectiveness in increasing learning performance (Brown and Lara, 2024), researchers recommend fostering collaborative learning environments. For instance, AI-driven chatbots applicated in video-based collaborative learning environments could offer quick and diverse feedback (Colissi et al., 2021; David et al., 2019; Neto & Fernandes, 2019), possibly enhancing engagement (Orduño-Osuna et al., 2024) and critical thinking (Feigh, 2022). Building on the shared mental model theory (Converse and Cannon-Bowers, 1991) a multi-party interaction with AI-driven chatbots might lead to improved team performance (Mathieau et al., 2000), team-skills (Endsley, 2015; Morgan et al., 1993) and teamwork (Burtschner and Manser, 2012), enhancing the effectiveness of collaborative learning overall (Prichard et al., 2006; Ruiz-Rojas et al., 204). However, the utilization of multi-party AI-driven chatbots in video-based collaborative learning remains largely unexplored. Therefore, we conducted a field experiment to explore (1) how multiple users interact with multi-party AI-driven chatbots in a video-based collaborative learning environment, (2) what learner’s perceptions of these interactions are and (3) what challenges and benefits arise from its usage.

25 voluntary student participants from an ongoing lecture course engaged with a one-hour lecture video using a video-based collaborative learning tool with an AI chatbot integrated via the ChatGPT API, providing real-time assistance based on predefined learning topics linked to video timestamps. Afterward, participants were divided into three focus groups, answering research questions using Miro Boards. All sessions were recorded, anonymized, and transcribed, with Miro Board content digitized and analyzed qualitatively. Additionally, video log data and chatbot prompt content were examined to identify usage patterns. Our findings highlight the potential of multi-party AI-driven chatbots to enhance collaborative learning and lead to design recommendations summarized in our solution called Rainbow Bot.

Meeting face-to-face with online acquaintances (i.e., people known exclusively from the internet) is a relatively common adolescent behavior (Smahel et al., 2020) that offers benefits alongside risks. Prior research has examined its predictors, including individual traits, well-being, internet use, and parenting strategies (e.g., van den Heuvel, 2012; Mohan, 2023). However, the cognitive dimension—specifically, the role of adolescents' perceptions about the potential risks and benefits—remains underexplored.

Preventive programs and media portrayals usually depict face-to-face meetings as dangerous (Wurlete & Kenny, 2016; Dědková, 2015). Risk perceptions are reflected in adolescents’ discourse (Mascheroni et al., 2014) and have been linked cross-sectionally to a lower likelihood of meeting face-to-face (Mýlek et al., 2023). However, since most adolescents felt happy after such meetings (Smahel et al., 2020), participation itself may also lower risk perception. Clarifying this causal direction is therefore crucial.

Also important is the interplay of the perceived likelihood of benefits and the perceived likelihood and severity of risks. According to the Fuzzy Trace Theory, while adults tend to rely on categorical judgments and avoid severe risks, adolescents tend to weigh the probabilities of risks and benefits (White et al., 2018). A two-wave panel study found that perceived risks negatively predicted risky online sexual behavior (including with online acquaintances), while perceived benefits showed no effect (Baumgartner et al., 2010). Yet, these findings may not extend to face-to-face meetings.

Our study uses the random intercept cross-lagged panel model to examine the bidirectional effects between adolescents’ perceptions of risks and benefits and their participation in face-to-face meetings. Furthermore, we compare the relative impact of these perceptions on behavior, which is critical for effective prevention. Our data come from a three-wave online survey of 2,500 Czech adolescents (ages 11–16, 50% girls, collected 2021–2022), sampled to reflect Czech households in terms of socioeconomic status, region, and municipality size.

Students’ competences in counseling, communication, and interdisciplinary collaboration build the foundation of successful pedagogical and health-related practice (Greuel, 2016; Herter-Ehlers, 2021; Diefenbach & Höhle, 2018). The teaching of the aforementioned competences in higher education often challenges teachers and remains at a theoretical level due to for example limited teaching resources (Hempel et al., 2021). To address this challenge, the DigiFall project leverages technology to create a practice-oriented learning environment. The digital self-assessment follows the principles of problem-based and case-based learning (Wijnia et al., 2024).

Building on a comprehensive needs’ assessment with teachers, students, and practitioners, video-based branching scenarios were developed to simulate real-life challenges, which are followed by a predetermined set of communication and counselling activities. This approach provides students with the opportunity to self-assess (Andrade, 2019) and refine their skills. This study addresses three research questions: (a) How do students evaluate the usability of DigiFall? (b) Which learning outcomes are reported by students regarding the presentation of the practical work? (c) How do students rate their communication and counseling abilities and do they change after the use of the DigiFall self-assessment?

A total of 78 students enrolled in various academic institutions located within the German state of North Rhine Westphalia participated in the pre-post-study design. The participants used the DigiFall-tool in their pedagogical and health-related studies. The data was collected through the implementation of ad-hoc questionnaires which address the usability, nine areas of communication and counseling skills, skills for interdisciplinary collaboration and the perception of the practical work. The quantitative data obtained was subsequently subjected to descriptive statistical analysis in addition to a paired t-test analysis. Thematic coding and frequency analysis were employed to evaluate the qualitative data. This study provides insights into the effectiveness of digital learning environments, facilitating comprehension of their development and the students' learning gains.

Affective polarization—the tendency to prefer one’s own political group while disliking others—has increasingly characterized Indonesia’s political landscape. The literature highlights the significant role of social media in driving political polarization, prompting scholars to emphasize the need for social media literacy interventions to reduce the level of affective polarization. Existing social media literacy programs primarily focus on enhancing cognitive skills, such as critical thinking to counter misinformation, while polarization also stems from socio-emotional processes, including interactions within filter bubbles among like-minded individuals. A study on social media in Indonesia found that affective polarization is driven by two key factors: difficulty in identifying real news and engaging mostly within like-minded networks. This study aims to develop and test a social media literacy education program using an experiential learning approach, followed by structured discussions among individuals with opposing political views as an application of the intergroup contact approach. Pairing the intergroup contact approach with social media literacy education, according to the literature, is believed to enhance the effectiveness of the intervention’s objective. In this study, the objective of the intervention is increasing media bias awareness and promoting cross-cutting exposure. The emphasis on these aspects is based on the expectation that such interventions will mitigate key antecedents of affective polarization among Indonesian social media users. This study employs a randomized controlled trial with pre- and post-intervention measures, where participants are randomly assigned to one of three groups: a control group, a social media literacy-only group, and a combined social media literacy-intergroup contact group. Participants are social media users in Indonesia. Thus, the research question in this study is whether participants in the combined group will experience the lowest levels of affective polarization, mediated by an increase in media bias awareness and motivation for cross-cutting exposure.

The mental health of housewives is considered one of the most pressing societal concerns (Matinzadeh et al., 2020, p.12). As they navigate intimate relationships and parenthood, their shifting identity and reliance on family can heighten the risk of reactive depression. Meanwhile, the rise of social media has given women greater autonomy, increasing their digital self-representation and visibility. This study examines housewives’ digital emotional expressions using a supervised machine learning approach on 1,800 YouTube videos.

The research question is:

Do housewife-related videos on social media tend to express depression?

Hypotheses:

- H1: Compared to videos related to working women, housewife-related videos on social media express a higher level of depression.

- H2: Compared to daily vlogs, housewife-related videos on social media express a higher level of depression.

YouTube was chosen as the research platform. Videos were scraped using the YouTube API v3 with the hashtags #housewife, #careerwomen, and #vlogs. For each hashtag, 600 video URLs were collected, and their transcripts were extracted using the YouTube Transcript API, resulting in three text datasets for analysis.

Supervised machine learning was used in this study. The labelled dataset for depression detection was sourced from Irem et al. (2023), which contains over 27,000 texts annotated with labels 1 (depressed) or 0 (not depressed) in a nearly balanced distribution. Using the Bag of Words approach, multiple classifiers were trained on this dataset after preprocessing to classify emotions as either depressed or not. Subsequently, a one-way ANOVA test was conducted to determine whether there were statistically significant differences in depression expressions among the three video groups.

Anonymity is a crucial factor in computer-mediated communication (CMC), influencing communication and behavior online. Researchers investigating anonymous CMC have gradually stressed the need to focus on perceived anonymity (Scott & Rains, 2020), a subjective feeling of lack of identifiability related to oneself (i.e., how anonymous one feels) or others (i.e., how anonymous one perceives others to be). Perceived anonymity can correspond to technical one (e.g., invisibility), but they can even be in stark contrast (Kennedy, 2006). In our study, we focus on factors affecting perceived anonymity of others online. We investigate the effect of different modes of communication (textual vs. audio-video) and based on Media Richness Theory (Daft & Lengel, 1984), we expect that participants will perceive others less anonymously in the richer one. Further, we also consider the role of amount and types of disclosed personal information in terms of breadth, i.e., variety of topics disclosed (high vs. low), and depth, i.e., intimacy (high vs. low). We expect that perceived anonymity will decrease as the breadth and depth of self-disclosure increase (Nguen et al., 2012). A novel exploratory aspect of our study is its focus on the combined effects of modalities and two different types of self-disclosure on perceived anonymity. We will use 2 (textual vs. audio-video mode) x 2 (self-disclosure breadth: high vs. low) x 2 (self-disclosure depth: high vs. low) between-subject experimental design utilizing stimuli introduced as social media posts of another person. Data will be collected online via a professional agency that will recruit participants from its online panel of Czech young adults (18–⁠⁠⁠⁠⁠⁠38 years old). We will analyze the data with regression models, examining the main effect of the manipulated factors as well as their interactions in influencing the perceived anonymity of others.

Most studies report only limited socio-demographic details on their participants (e.g., age, ethnicity and gender distribution). Researchers often turn to “convenience samples”, with most participants sharing similar characteristics. Resulting samples are often WEIRD (Western, Educated, Industrial, Rich, and Democratic) and do not adequately reflect society’s diversity (Henrich et al., 2010; Seaborn et al., 2023). Creating more diverse samples, however, poses several – hitherto underexamined – questions: How diverse do participants have to be to be considered representative of a society? Does every study need to have participants representing all possible combinations of characteristics? How can the diversity of samples be reported in scientific publications? However, in order to make the results of user studies generalizable for citizens and to address the needs of everyone, a diverse sample is indispensable (Gidlund, 2015; May & Ross, 2017). While conducting studies within a research project, we repeatedly encountered the challenge of a non-diverse sample. Therefore, we compared the sample diversity of our recent studies with official statistics in the region surveyed. Merely individual characteristics are consistent with the data from the region, making the sample hardly representative and applicable to the general population. Examining our sample diversity yielded several learnings serving as a basis for further research and discussion, aimed at diversifying future studies in psychology and other disciplines. They also reveal challenges that impair the acquisition of diverse samples. Closer examination of what experiences are getting lost when focusing on more homogenous parts of the sample underlines the importance of this topic. It is of utmost importance to us to reflect these problems in a scientific discourse. The poster presentation is intended to serve as an exchange of experiences on sample diversity and representativeness. Therefore, we would like to position the sample diversity at MediaPsych to engage a discussion with experienced scientists.

Universities could provide an optimal setting to tackle the ongoing mental health crisis among university students suffering from mental health problems (Wiens et al., 2019). Simultaneously, promoting students’ well-being as well as preventing mental health issues is considered equally crucial (Baños et al., 2017). In this context, Positive Psychological Interventions (PPIs), i.e., intentional activities specifically designed to enhance well-being (Schueller, Kashdan, & Parks, 2014), constitute promising tools for the promotion of mental health. Specifically, Online PPIs that are delivered through digital platforms (Botella et al., 2011) could be utilized by educational institutions. In this context, integrating these interventions directly into university courses offers a great potential to enhance students’ well-being and thus ultimately contribute to a positive learning experience.

However, positive psychological activities based on the PERMA-model of well-being (Seligman, 2011), such as gratitude exercises and practicing kindness (Seligman et al., 2005), have been seldomly used in student populations (with the exception of Tagalidou, Baier, & Laireiter, 2019 and Yu, 2020) and, to the best of the authors’ knowledge, an approach of integrating these activities into teaching has not yet been applied.

Therefore, we aim at filling these gaps by 1. building a thorough collection of PPIs and related activities, 2. categorising them based on different criteria in terms of integratability into teaching settings which have been created with university students and lecturers using a participatory approach, 3. adjusting them for the teaching setting, and 4. having them used in university courses and evaluated by both university lecturers and students regarding their applicability and efficacy in increasing individual well-being, respectively. For the latter one, different measures, such as the PERMA-profiler (Butler & Kern, 2016), are used in a pre-post design.

Results presented at the conference contribute to an increased knowledge about effective measures to enhance students’ well-being in educational settings.

Previous research on narrative engagement has predominantly examined viewers’ subjective experiences, such as immersion and transportation, in the context of mainstream narratives with conventional linear structures. However, less attention has been given to how engagement operates in narrational complexity—characterized by structural disruptions to time, space, causality, or character perspective that require viewers’ active cognitive integration. This contrasts with thematic complexity, which relates to a story’s meaning or subject matter. To advance this understanding, we introduce a theoretical framework that maps the cognitive and affective mechanisms underlying the subjective experience of engagement with narrational complexity, drawing on qualitative interview data and existing literature.

We outline a mechanistic psychological approach that examines how executive functions—such as cognitive flexibility and goal-directed attention—foster flexible analytical dispositions essential for comprehending fragmented and ambiguous story structures. We further propose that this disposition, along with metacognitive self-regulation, helps sustain the cognitive effort required to maintain engagement through narrational complexity, particularly when viewers experience frustration due to confusion.

The proposed theory will be tested in a between-subjects experiment, in which participants are randomly assigned to view either a narrationally complex film or a conventionally structured film. Measures will be taken before, during, and after viewing to assess individual differences, baseline measures, engagement-related cognitive and affective processes, and real-world behavioural intentions. The study will examine how individual differences in cognitive flexibility influence responses to structurally complex narratives (RQ1) and explore the role of metacognitive self-regulation in sustaining engagement despite potential frustration (RQ2), as a potential mechanism underlying tolerance for ambiguity (RQ3). Additionally, it will investigate whether engaging with narrational complexity extends beyond media consumption, shaping real-life behavior such as willingness to engage with complex societal issues, including climate change (RQ4). This paper presents initial findings from a study exploring these relationships and their implications for future research.

Theories and models related to communication processes and stigma are manifold, and vary greatly with regard to their understanding of relevant concepts, causal links, and underlying communication processes. Even though most refer to stigma as a largely similar phenomenon, a comprehensive and widely accepted stigma communication theory is still lacking. Accordingly, we systematically examined (a) how stigma-related communication is conceptualized in the present research literature, (b) which models and theories of stigma-related communication exist, and how frequently they are referenced and empirically tested, (c) which aspects of communication are covered and which not, (d) which influencing factors are posited, as well as (e) which implications can be derived for stigma communication theory formation.

Following the PRISMA guidelines, a scoping review of reviews was conducted that included all English-language journal publications of literature reviews (i.e., systematic reviews, meta-analyses, scoping reviews, narrative reviews) with references to both stigma and communication that were published until end of May 2023. The databases PSYNDEX, APA PsychArticles, APA PsychInfo, MLA, and Web of Science were searched on June 1, 2023, using the English search terms stigma* AND (communicat* OR interact* OR language OR speech OR media) AND (review OR meta*). After the initial search and exclusion of duplicates, a title and abstract screening of the remaining records was performed by two coders independently. The full text screening of all resulting records was conducted to determine their eligibility regarding the type of literature review and the relevance concerning stigma-related communication processes. The subsequent sample was included for the full text coding of references to stigma and communication in theory, method, results, and implications made in the remaining reviews.

The result from the scoping review will provide the first systematic evaluation of stigma-related theories, models, and concepts in communication contexts to advance stigma communication theory and research.

This study focuses on profiling the digital media consumption habits of university students. Participants (n = 50) are asked to install StayFree, an application that tracks their smartphone usage over seven consecutive days, providing detailed data on time spent across various applications and usage categories. A brief questionnaire complements this data by collecting self-reported information on the use of other digital devices (e.g., computers, TVs, and gaming consoles) and participants’ perceptions of their media consumption.

The theoretical framework is grounded in Media Psychology, with a particular emphasis on the relationship between media consumption patterns and cognitive-emotional responses. Prior research has highlighted the role of digital overuse in altering attention, memory, and emotional regulation, as well as the potential for addictive behaviors driven by short-term gratification (Thomée et al., 2010; Vahedi & Saiphoo, 2018). This framework underscores the importance of profiling consumption habits to better understand how prolonged media exposure shapes psychological states and behaviors, and points to a gap in studies using monitoring apps such as StayFree to track daily usage (Harris et al., 2020).

The integration of StayFree data and questionnaire responses allows for a more comprehensive, less self-report-dependent profile of consumption habits, segmented by purpose (e.g., recreational, academic, or work-related). This dual-method approach ensures a more nuanced understanding of digital behavior patterns, providing valuable insights for future research.

In the influencer market actors compete not only in the classically understood quality of content, but also in the ability to show and present appropriate emotions and attract the audience with elements of private life. Since the appeal of influencers is based on building parasocial relationships, an influencer gains an advantage by skillfully doing emotional work and reinforcing the parasocial relationship with private life content.

However, such characteristics of influencers’ work create risks for both broadcasters and audiences. Influencers may burn out due to overexploitation of the emotion control center, and may also have disturbed characteristics of close relationships due to selling their own privacy. These risks are all the more serious because most influencers are young adults or adolescents. On the other hand, young audiences of influencers may derive faulty patterns of social relationships when the content presented to them contains, e.g., unnatural reactions and treating privacy as a sphere “for sale” and “always to be shown.”

All of this creates a space for the question: is the potentially threatening sale of intimacy and privacy frequented by some influencers, involving increased emotional labor, a “market mistake” or a response to audience expectations, giving something of a competitive advantage? This question is operationalized as: on YouTube, the world's most important video content portal, is there an increased interest in those influencer contents that contain references to private or romantic themes?

To answer this question, two methods will be used: (1) comparing the number of views and reactions of videos and (2) a novel, previously unused method of tracking interest in particular moments of a video based on a graph of the number of re-views, a functionality provided by YouTube's interface. Selected videos and profiles of the most popular Polish youtubers (500,000 or more subscribers) will be examined.

The aim of this study was to examine the influence ChatGPT exerts on university students' current learning motivation, as well as the impact of this motivation on the (technological) acceptance of ChatGPT in academic settings. Students were surveyed between the months of July and November 2023 using an online questionnaire. This questionnaire included scales measuring current motivation (FAM) and selected scales from the technological acceptance model 3 (TAM 3) to assess technological acceptance. A total of 111 complete datasets were analyzed. The results revealed a minor effect of ChatGPT on current learning motivation, which in turn influenced the acceptance of ChatGPT in academic settings.

In human-robot interactions (HRI), there is a plethora of errors that can occur, such as recognition or interpretation errors, slow response times, or connectivity issues. Many of these errors cannot be fixed in the foreseeable future, since their origin lies in external factors or the current maturity level of technology. Extant research has found contradictory results regarding the effects of errors. For example, Robinette et al. (2017) found a decrease in trust, while Ye et al. (2019) observed no such effects on trust. Additionally, while research has identified the users’ personality as a predictor for their reaction to errors (Hoffman et al., 2020), these personality differences remain largely under-investigated (Rothstein et al., 2022). Drawing on findings from human-to-human interaction, we hypothesize that emotional stability and agreeableness might predict the likelihood of forgiving errors (McCullough et al., 2001). In turn, the tendency to forgive errors might dampen negative effects on trust and robot evaluation.

To understand the effects of these errors on the perception of robotics interactions, we conducted a 3x1 between-subjects laboratory experiment with N = 62 participants, using virtual reality (VR). Participants interacted with a robot in a virtual library to receive personalized book recommendations, with one group experiencing a flawless robotic system, one second group experienced being misunderstood by the robot, and the third group facing long response times. Users’ personality traits comprised the Big Five constructs neuroticism and agreeableness, their technical affinity, capacity for empathy, and inclination to forgive errors. Robot evaluation was quantified by assessing trust in the robot after the interaction, as well as participants’ general attitudes towards robots (measured before and after the interaction, as well as one week after the experiment). The findings contribute to a better understanding of the individual handling of errors in HRI, allowing for more enjoyable use of robots.

Recent advancements in large language models and the increasing communicative sophistication of voice-based dialogue systems have given rise to a novel phenomenon: deep, intimate conversations with AI companions. A striking example is ChatGPT’s so-called ‘Do Anything Now (DAN)’ mode, which allows users to personalize interactions, shaping the AI into a companion with perceived emotions and personality through human-like language, emojis, and flirtation. By utilizing the voice function, the system displays social cues and is perceived as a social actor, fostering social norms, intimacy, and deeper engagement, which enhances self-disclosure. This, in turn, supports long-term interactions, allowing relationships with the system to develop similarly to human relationships. According to Social Penetration Theory (Altman & Taylor, 1973), self-disclosure is fundamental to both human and human-machine relationship formation. Thus, this study adopts a longitudinal approach to examine how personalization—specifically the ‘DAN’ mode in ChatGPT 4.o—impacts self-disclosure behavior, measured by word count and conversational depth, in intimate communication over time. To explore this, a four-week study is being conducted, where participants interact with ChatGPT’s voice function either with or without ‘DAN’ mode activated. A mixed-method approach integrates quantitative self-report measures and qualitative content analysis of disclosed narratives, based on the transcripts of the interactions. To further understand what affects self-disclosure over time, (a) system behavior, i.e., ChatGPT’s word count and output depth, (b) user reactions, i.e., interpersonal closeness, flattery, and fulfillment of communication preferences, and (c) between-subject differences, i.e., relationship status and gender, are also investigated. Consequently, this study explores the privacy relevance of disclosed content, and the extent of sensitive information shared. By investigating whether personalized voice-based interactions elicit deeper and more sustained self-disclosure, and by identifying key conversational themes, this research advances our understanding of AI-driven relational dynamics and challenges theoretical conceptualizations of self-disclosure in human-machine relationships.