# Publications ## 2026 ### [“Oh No, the Robots Are Here”: AI identity labels, social uptake and sanctioning in mixed human–AI communities](https://osf.io/preprints/psyarxiv/kwy5j_v2) Major, S., **Tomašević, A.**, Vranić, A., Cvetković, D., Vudragović, D., Stupovski, B., & Mitrović Dankulov, M. *PsyArXiv* · [Overview](https://atomasevic.xyz/papers/ai-identity-labels.html) · [Open access](https://doi.org/10.31234/osf.io/kwy5j_v2) · [Zenodo](https://doi.org/10.5281/zenodo.22964196) Abstract Social media users can now encounter autonomous AI accounts that post, reply and vote alongside other users. In a longitudinal field experiment, we examined how labelling these accounts relates to whom people answer and how they evaluate AI contributions. YSocial is a Reddit-like hybrid platform for social media simulation, where people interact with language-model agents. On it, 36 participants discussed popular culture with 72 AI accounts powered by an uncensored language model over 21 experiment days. The AI accounts kept persistent personas and remembered previous exchanges. They were visibly labelled in one community and unlabelled in the other. Among 2,143 human replies, Bayesian models estimated that the probability of replying to an AI-authored post or comment was 3.5 percentage points higher in the hidden-label community, given the same earlier posts and comments available in the thread (95% credible interval 0.5–6.5). In both communities, participants were more likely to reply to an AI account after it had replied to them and less likely after they had replied to it. The gap in criticism, including polite disagreement, between replies to AI and human accounts was 13.3 percentage points larger in the visible-label community (95% credible interval 5.1–21.4). The toxicity difference was uncertain. Visible labels were associated with less engagement with AI contributions and more criticism of them. ### [A Latent Oscillator Measurement Model to Simulate Emotional-Expression Score Dynamics in Video](https://arxiv.org/abs/2609.15273) **Tomašević, A.**, Golino, H., & Christensen, A. P. *arXiv* · [Overview](https://atomasevic.xyz/papers/latent-oscillator.html) · [Open access](https://doi.org/10.48550/arxiv.2609.15273) · [PsyArXiv](https://osf.io/preprints/psyarxiv/hf3g7) · [OSF](https://osf.io/y9jb3/) Abstract Facial-expression classifiers convert video into multivariate time series of scores with measurement error from classifiers, videos, and recording conditions. Empirical score series cannot establish whether the score channels reflect a smaller set of latent expressive processes or whether an analysis would recover those processes. We introduce the Latent Oscillator Measurement Model (LOMM), a data-generating model that separates latent dynamics, time-varying activity, and a factor-analytic observation model. The latent processes are damped, undamped, or amplifying linear oscillators. LOMM generates bounded scores or continuous indicators. Study 1 used four-fold cross-fitting with scores from 100 MAFW videos to calibrate LOMM and evaluate generated series on held-out videos. Median plausibility and coverage were 0.970 and 0.920 for LOMM, versus 0.510 and 0.370 for a calibrated static logistic-normal generator. Study 2 tested whether Dynamic Exploratory Graph Analysis (DynEGA), static EGA, GraphicalVAR, and GIMME recovered a known dimensional structure from continuous indicators generated by LOMM. At 100 observations per clip, with failed or timed-out fits counted as incorrect, correct-dimension recovery was 0.939 for DynEGA, 0.884 for static EGA, 0.777 for GraphicalVAR, and 0.176 for GIMME. Replacing the common fixed initialization with independent stationary starts for stable dimensions and bounded independent starts for amplifying dimensions reduced recovery for DynEGA, static EGA, and GraphicalVAR. LOMM provides a controlled test of whether an analysis recovers aspecified latent structure before score patterns are interpreted psychologically. ### [Reddit deplatforming and cohort mixing in generalist Voat communities](https://doi.org/10.1038/s41598-026-63627-4) **Tomašević, A.**, Vranić, A., Alorić, A., & Mitrović Dankulov, M. *Scientific Reports* · [Overview](https://atomasevic.xyz/papers/voat-deplatforming.html) · [Open access](https://www.nature.com/articles/s41598-026-63627-4_reference.pdf) · [arXiv preprint](https://arxiv.org/abs/2512.22348) Abstract Deplatforming, the permanent banning of entire communities, is a common tool for content moderation on mainstream platforms. Prior research largely examines banned communities and source-platform health, leaving receiving platforms less understood. We analyze four major Reddit community bans from 2015 to 2020 and their relationship to six generalist communities on Voat, asking how users appearing on Voat after these bans are associated with community structure and toxicity. Combining network analysis, ToxiGen classifier probabilities, and dynamic reputation modeling, we document distinct trajectories in cohort mixing, structural representation, toxicity, and repeated participation across Voat’s generalist communities. Breakpoint analyses provide the clearest evidence around the GreatAwakening period in the E-I index, while mean ToxiGen classifier probability rose from approximately 0.12 in late 2015 to above 0.24 by mid-2018 and then remained elevated. Fixed-cohort results show that high-activity arrivals were more likely than average arrivals to become structurally central, and survival analyses indicate longer persistence among high early-toxicity users in earlier post-ban cohorts. These findings identify receiving-platform changes that coincided with source-platform moderation. ### [Dynamic exploratory graph analysis (dynEGA) of affect and self-harm urge in non-suicidal self-injury (NSSI): Three illustrative case studies](https://doi.org/10.1016/j.psycr.2026.100342) Bruckbauer-Schwed, M., Roth, L. H. O., Plöderl, M., **Tomašević, A.**, Keglevic, M., & Laireiter, A.-R. *Psychiatry Research Case Reports*, 5(2), 100342 · [Open access](https://www.sciencedirect.com/science/article/pii/S2773021226000362/pdf) · [OSF preprint](https://doi.org/10.31234/osf.io/urfek_v1) Abstract **Objectives:** Non-suicidal self-injury (NSSI) is closely linked to negative emotions and self-harm urges. Understanding these processes may inform psychological interventions. This study presents three illustrative clinical cases to explore person-specific associations between affective states, self-harm urges, and NSSI behaviour. **Methods:** Experience-sampling methodology (ESM) was used to collect longitudinal data from three young adults with varying frequencies of NSSI over 14 days. Participants completed brief questionnaires four times daily, yielding 165 assessments. Dynamic Exploratory Graph Analysis (dynEGA) was applied to construct individual affective networks related to self-harm urges. **Results:** Across cases, NSSI urges were generally associated with negative affect, but substantial differences emerged in affective network structures. No consistent network pattern was observed across participants. Associations between specific affects and urge intensity varied considerably between individuals. In some instances, increases in negative affect appeared to precede stronger self-harm urges, although these patterns were not uniform. Urges were directly associated with subsequent NSSI behaviour in only a limited number of observations. **Conclusion:** These three illustrative clinical cases highlight potential associations between negative affect, self-harm urges, and NSSI behaviour while underscoring the heterogeneity of these processes. The findings support the importance of person-centred approaches in NSSI research and suggest that dynEGA may be a promising exploratory method for examining individual affective network structures and their relationship to self-harm urges and behaviour. ### [Between “we” and “well-being”: Do basic psychological needs mediate the relationship between interdependent self-construal and adolescent subjective well-being?](https://doi.org/10.1111/aphw.70172) Lazić, M., Joshanloo, M., **Tomašević, A.**, Šakan, D., Gavrilov-Jerković, V., Zotović-Kostić, M., Obradović, V., & Jovanović, V. *Applied Psychology: Health and Well-Being*, 18(4), e70172 · [Open access](https://pmc.ncbi.nlm.nih.gov/articles/PMC13428330/pdf/APHW-18-0.pdf) · [OSF preprint](https://osf.io/preprints/osf/r4vfp_v1) Abstract The role of identity content in adolescent subjective well‐being (SWB) is poorly understood. To address this gap, the present longitudinal study examined the relationships between two dimensions of interdependent self‐construal (relational and collective) and SWB (positive affect, negative affect, and life satisfaction). We also investigated whether basic psychological needs satisfaction could serve as a mechanism explaining the relationship between interdependent self‐construal and SWB. The sample consisted of 971 Serbian adolescents ( M age = 15.64, 64.5% female), who completed surveys at three time points over a 12‐month period. The results of random intercept cross‐lagged panel modeling indicated that: (1) adolescents reporting consistently higher levels of collective self‐construal also reported consistently higher levels of SWB, whereas adolescents reporting consistently higher levels of relational self‐construal showed consistently higher positive affect but neither higher life satisfaction nor lower negative affect; (2) within‐person deviations in interdependent self‐construal were not associated with subsequent deviations in SWB; and (3) the satisfaction of basic psychological needs did not mediate the association between interdependent self‐construal and SWB at the within‐person level. Overall, the findings suggest that the relationship between interdependent self‐construal and SWB operates primarily at the between‐person level rather than the within‐person level, indicating that stable patterns of self‐definition are more strongly related to adolescents' well‐being than within‐person changes in self‐definitions over the one‐year study period. ### [Towards Operational Validation of LLM-Agent Social Simulations: A Replicated Study of a Reddit-like Technology Forum](https://doi.org/10.1140/epjds/s13688-026-00674-x) **Tomašević, A.**, Cvetković, D., Major, S., Maletić, S., Anđelković, M., Vranić, A., Stupovski, B., Vudragović, D., Bogojević, A., & Mitrović Dankulov, M. *EPJ Data Science* · [Overview](https://atomasevic.xyz/papers/llm-simulation-validation.html) · [Cited by 1](https://openalex.org/works?filter=cites:W4415994324) · [GitHub](https://github.com/atomashevic/voat-simulation) · [YSocial client](https://github.com/atomashevic/YClient-Reddit) Abstract Validation of LLM-agent social simulations remains underdeveloped, with most studies relying on subjective assessments or single runs. We address this gap by running 30 independent 30-day simulations of a technology forum modeled on Voat’s v/technology, using stateless Dolphin Mistral 24B agents on the Y Social platform, and evaluating operational validity across five dimensions: activity patterns, network structure, toxicity, topical coverage, and stylistic convergence. Against 30 matched, non-overlapping 30-day Voat comparison windows, results show overlapping 99% confidence intervals for unique users, root posts, and daily active users, while comments, average thread length, and mean toxicity remain higher in simulation. Both simulated and empirical networks exhibit core–periphery structure, though simulated cores are larger and more diffuse and repeated interactions are less frequent. Topic alignment is near-complete, but toxicity is misallocated across content layers: simulated root posts are substantially more toxic than real submissions, while simulated comments are less toxic than Voat comments. These findings demonstrate that LLM agents in platform-faithful environments can reproduce familiar online regularities, while systematic divergences—particularly those linked to stateless agent design and content-layer calibration—point to concrete directions for future improvement. ### [Investigating the analytical robustness of the social and behavioural sciences](https://doi.org/10.1038/s41586-025-09844-9) Aczel, B., Szaszi, B., Clelland, H. T., & others, including **Tomašević, A.** *Nature*, 652(8108), 135–142 · [Open access](https://publications.aston.ac.uk/id/eprint/48959/1/B_Aczel_et_al_AAM_Investigating_the_analytical_robustness_of_the_social_and_behavioural_sciences.pdf) Abstract The same dataset can be analysed in different justifiable ways to answer the same research question, potentially challenging the robustness of empirical science. In this crowd initiative, we investigated the degree to which research findings in the social and behavioural sciences are contingent on analysts' choices. We examined a stratified random sample of 100 studies published between 2009 and 2018, in which, for one claim per study, at least five reanalysts independently reanalysed the original data. The statistical appropriateness of the reanalyses was assessed in peer evaluations, and the robustness indicators were inspected along a range of research characteristics and study designs. We found that 34% of the independent reanalyses yielded the same result (within a tolerance region of ±0.05 Cohen's d) as the original report; with a four times broader tolerance region, this indicator increased to 57%. Of the reanalyses conducted, 74% reached the same conclusion as the original investigation, 24% yielded no effects or inconclusive results and 2% reported the opposite effect. This exploratory study indicates that the common single-path analyses in social and behavioural research should not be simply assumed to be robust to alternative analyses. Therefore, we recommend the development and use of practices to explore and communicate this neglected source of uncertainty. ### [Identity Orientations: Personal, Relational, Public, and Collective Aspects of Identity](https://doi.org/10.1080/15283488.2026.2649028) Cheek, J. M., **Tomašević, A.**, Lazić, M., & Jovanović, V. *Identity* Abstract Identity orientations refer to the relative importance or subjective value that individuals place on various identity attributes or characteristics when constructing their self-definitions. In the first section of this paper, we review the twenty-year chronology of the development of the identity orientation measurement model and scales, from the two original scales to the AIQ-IV Tetrapartite Model of Personal, Relational, Public, and Collective Identity Orientation scales. Correlations between the four identity orientation scales and measures of the independent self, the interdependent self, and the three self-aspects of the tripartite RIC model are presented and discussed. In the second section, we present a detailed examination of existing psychometric analyses of the AIQ-IV, followed by an illustration of a possible line of research based on network psychometrics. Existing studies support the superiority of the four-factor exploratory structural equation modeling (ESEM) representation of the AIQ-IV responses over a traditional four-factor confirmatory factor analysis (CFA) model. Several studies examining measurement invariance of the identity orientation scales also are reviewed. We conclude with suggestions for future identity research using the AIQ-IV. ### [Making Sense of Student Feedback in Large Classes: A Practical Guide Using AI-Based Text Analysis](https://doi.org/10.1080/87567555.2026.2639369) Teles, M., & **Tomašević, A.** *College Teaching*, 1–11 · [Cited by 1](https://openalex.org/works?filter=cites:W7135058553) Abstract In large-enrollment courses, instructors often receive hundreds of student comments that are difficult to synthesize, and traditional Likert-scale evaluations rarely capture the nuances of students’ learning experiences. This study introduces an accessible, evidence-based approach that uses open-source artificial-intelligence text analysis to interpret qualitative feedback quickly and systematically. Using 678 comments from 270 students in lecture and active-learning sections of an undergraduate course, the analysis identified clear contrasts in how students described engagement, learning, and enthusiasm - differences that were not visible in the numerical ratings. The AI results extended the insights from standard evaluations, highlighting aspects of classroom climate and instructional design that instructors could address directly. Because the procedure runs on common computers with freely available software, it enables faculty to examine large sets of comments in minutes rather than hours. This paper offers a practical guide for applying this approach to course evaluations, helping instructors turn extensive qualitative feedback into concise summaries for teaching improvement. ### [Depression and anxiety symptoms in adolescents across 30 countries: Cross-national measurement invariance and relationships with subjective well-being](https://doi.org/10.1016/j.jad.2026.121693) Jovanović, V., Adams, S., Aritio-Solana, R., & others, including **Tomašević, A.** *Journal of Affective Disorders*, 406, 121693 · [Cited by 2](https://openalex.org/works?filter=cites:W7141463652) · [Open access](http://publicatio.bibl.u-szeged.hu/40128/1/Depression.pdf) Abstract Cross-cultural empirical evidence on the validity of measures of depression and anxiety symptoms in adolescents is scarce and inconclusive. The present study investigated the cross-national measurement invariance and convergent validity evidence of the Depression and Anxiety scales of the Depression Anxiety and Stress Scales (DASS-21). We sampled 16,802 adolescents (58.1% female) aged 14-19 years across 30 countries from five continents. Both exact and approximate measurement invariance were tested. The results provided evidence for the exploratory structural equation model (ESEM) as the best representation of the data in most countries. Full scalar invariance was not supported when tested across all countries, but only among two Southern Asian countries (Bangladesh, India), two Northern European countries (Estonia, Lithuania), the six former republics of Yugoslavia, and European and American countries grouped by UN subregions. Results of the alignment method supported the approximate invariance of the ESEM model across 29 countries. Associations between Depression and Anxiety ESEM factors and life satisfaction, positive affect, and negative affect conformed to the theoretical expectations in most countries. Our findings indicate that the Depression and Anxiety scales of the DASS-21 have potential for use in cross-national research on adolescent internalizing symptoms, but we highlight issues of using it outside of Western (European, North American) settings. ### [transforEmotion: An Open-Source R Package for Emotion Analysis Using Transformer-Based Generative AI Models](https://www.aup-online.com/content/journals/10.5117/CCR2026.2.2.TOMA) **Tomašević, A.**, Golino, H., & Christensen, A. *Computational Communication Research*, 8(2), 1 · [Overview](https://atomasevic.xyz/papers/transforemotion.html) · [Cited by 2](https://openalex.org/works?filter=cites:W7126181639) · [Open access](https://doi.org/10.5117/ccr2026.2.2.toma) Abstract This software demonstration article introduces transforEmotion, an open-source R package that addresses critical bottlenecks in communication research emotion analysis. Communication researchers currently face three key barriers: (1) modal fragmentation requiring separate tools for text/image/video analysis, (2) rigid emotion taxonomies that don't match theoretical frameworks, and (3) irreproducible workflows dependent on commercial APIs. transforEmotion removes these bottlenecks through unified multimodal processing, zero-shot classification with arbitrary labels, local inference capabilities, and seamless R integration. This enables systematic investigation of emotion dynamics across communication contexts and modalities that were previously technically difficult to analyze. ### [High-Performance Data Analytics for Analysis of Collective Trust Dynamics](https://doi.org/10.1007/978-3-032-22221-3_43) Cvetković, D., Major, S., **Tomašević, A.**, Maletić, S., Anđelković, M., Vranić, A., Stupovski, B., Vudragović, D., Alorić, A., Bogojević, A., & Mitrović Dankulov, M. In *Large-Scale Scientific Computations* (pp. 388–395). *Lecture Notes in Computer Science*, Springer ### [Psychometric Network Analysis of Online Behaviour](https://doi.org/10.1007/978-3-031-52643-5_28-1) **Tomašević, A.** *The Palgrave Encyclopedia of Cyberpsychology* ## 2025 ### [Multi-Platform Aggregated Dataset of Online Communities (MADOC)](https://doi.org/10.1609/icwsm.v19i1.35954) Dankulov, M. M., **Tomašević, A.**, Maletić, S., Anđelković, M., Vranić, A., Cvetković, D., Stupovski, B., Vudragović, D., Major, S., & Bogojević, A. *Proceedings of the International AAAI Conference on Web and Social Media*, 19(1), 2529–2538 · [Overview](https://atomasevic.xyz/papers/madoc.html) · [Cited by 3](https://openalex.org/works?filter=cites:W4411121016) · [Open access](https://ojs.aaai.org/index.php/ICWSM/article/download/35954/38108) · [arXiv preprint](https://arxiv.org/abs/2501.12886) · [Zenodo](https://zenodo.org/records/14637314) · [PyPI](https://pypi.org/project/pymadoc/) · [R package](https://github.com/atomashevic/rMADOC) Abstract The Multi-platform Aggregated Dataset of Online Communities (MADOC) is a comprehensive dataset that facilitates computational social science research by providing a unified, standardized dataset for cross-platform analysis of online social dynamics. MADOC aggregates and standardizes data from four distinct platforms: Bluesky, Koo, Reddit, and Voat, spanning from 2012 to 2024. The dataset includes 18.9 million posts, 236 million comments, and data from 23.1 million unique users across all platforms, with a particular focus on understanding community dynamics, user migration patterns, and the evolution of toxic behavior across platforms. By providing standardized data structures and FAIR-compliant access through Zenodo and corresponding Python and R packages, MADOC enables researchers to conduct comparative analyses of user behavior, interaction networks, and content sentiment across diverse social media environments. The unique value of the dataset lies in its cross-platform scope, standardized structure, and rich metadata, making it particularly suitable for studying societal phenomena such as community formation, toxic behavior propagation, and user migration patterns in response to platform moderation policies. ### [A network analysis of the associations between the Dark Triad traits and domains of emotional functioning](https://doi.org/10.1002/ijop.70011) Grabovac, B., Dinić, B. M., **Tomašević, A.**, Carroll, G., & Burke, T. *International Journal of Psychology*, 60(2), e70011 · [Open access](https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/ijop.70011) · [OSF](https://osf.io/zg7wd/) Abstract The aim of this study was to reexamine relationships between the Dark Triad traits and aspects of emotional functioning, such as emotional intelligence and empathy, as traits and abilities, by using a network analysis approach. The sample consisted of 359 Hungarians from Serbia. The results indicated that the optimal description of the network structure involved four communities: the Dark Triad traits, emotion recognition abilities as aspects of cognitive empathy, emotional intelligence traits and empathy and appraisal and expression of emotion in others as mainly traits of empathy focusing on others. Psychopathy and narcissism are central components connecting their Dark Triad community with emotional functioning communities. Psychopathy's main role stemmed from its negative relationship with emotional reactivity and recognising positive states, indicating emotional deficits in both trait and ability domains. On the other hand, narcissism had a positive relationship with the use of emotion to facilitate performance, but showed a negative relationship with recognising negative states, indicating a discrepancy between self-image and objective abilities. Machiavellianism showed weak relationships with emotional intelligence and empathy traits, but a non-significant one with the ability to recognise emotional states. Its strongest positive relationships were observed with cognitive empathy and the use of emotions to facilitate performance. ### [The (mis) measure of misbehavior: Cross-national invariance of the Youth Externalizing Problems Screener across 32 countries](https://doi.org/10.1017/S0954579425100473) Lazić, M., Adams, S., Aritio-Solana, R., Aryanto, C. B., Avsec, A., Bakhshi, A., Bender, M., Berjot, S., Zambrano, S. B., Brajša-Žganec, A., & others, including **Tomašević, A.** *Development and Psychopathology*, 37(5), 2827–2840 · [Cited by 3](https://openalex.org/works?filter=cites:W4414881851) · [Open access](https://research.tilburguniversity.edu/en/publications/ba789622-f194-4f3c-915b-9e65af49a2a3) Abstract The present study investigated the cross-national measurement invariance of a 10-item Youth Externalizing Problems Screener (YEPS) on a sample of 17,489 adolescents from 32 countries. The original one-factor and two-factor models of YEPS were found to provide a poor fit to the data in most countries. Following the removal of two semantically overlapping items and the inclusion of correlated error terms, adequate model fit was obtained in 31 of 32 countries. Measurement invariance testing of an abbreviated 8-item YEPS (YEPS-SF) supported configural invariance. Partial scalar invariance was achieved only after freely estimating numerous parameters. The alignment analysis revealed that 22% of parameters were non-invariant across countries. South Africa, Hungary, and India showed the largest number of non-invariant parameters, whereas the lowest number was detected in several European countries. These findings highlight the potential of the YEPS-SF for use within individual countries and the challenge of developing cross-culturally comparable measures, suggesting that cultural adaptations may be necessary. ### [Positive and negative emotional experiences in adolescents across 32 countries: Cross-national and gender differences](https://doi.org/10.1007/s11482-025-10486-2) Jovanović, V., Adams, S., Aritio-Solana, R., Aryanto, C. B., Avsec, A., Bakhshi, A., Bender, M., Berjot, S., Zambrano, S. B., Brajša-Žganec, A., & others, including **Tomašević, A.** *Applied Research in Quality of Life*, 20(4), 1731–1757 · [Cited by 8](https://openalex.org/works?filter=cites:W4413439424) · [Open access](http://publicatio.bibl.u-szeged.hu/40129/1/positive.pdf) Abstract Despite a growing interest in adolescent subjective well-being, cross-cultural research on positive and negative affect in adolescents has been surprisingly scarce. Moreover, a cross-cultural evaluation of affective well-being measures in adolescents has lagged behind research in adults. The present study evaluated the cross-national and gender invariance of the Scale of Positive and Negative Experience (SPANE), using adolescent samples from 32 countries (N = 17,489; age range = 14–19) spanning five continents. The original two-factor structure of the SPANE was supported in 29 of the 32 countries. We could establish partial metric invariance of the SPANE across countries and full or partial scalar invariance across gender in 18 and 11 countries, respectively. The alignment method showed evidence for approximate invariance across countries, allowing for cross-national means comparisons of positive and negative affect. In most countries, girls reported higher levels of negative affect, and no substantial gender differences in positive affect were observed. Cross-national differences in mean levels of positive and negative affect yielded complex findings. Our results suggest that the SPANE is a solid candidate for measuring emotional experiences among adolescents in large-scale cross-national studies. ### [Understanding democracy in Europe: Liberal, populist, and social-democratic dimensions](https://doi.org/10.2298/SOC2502171T) **Tomašević, A.** *Sociologija*, 67(2), 171–194 · [Open access](https://doi.org/10.2298/soc2502171t) Abstract How do citizens understand democracy, and how do these understandings vary across social groups and national contexts? Using data from Round 10 of the European Social Survey across 31 countries, this study develops and validates a three-dimensional model of democratic attitudes. Using latent variable analysis, we identify distinct Liberal, Populist, and Social-democratic dimensions of democratic understanding. Our findings demonstrate that these dimensions show partial measurement invariance across countries, allowing for meaningful cross-national comparisons of latent scores. While the dimensions are positively correlated, suggesting citizens want ?more of democracy? rather than trading off different aspects, they relate differently to social class, political behavior, and institutional trust. Social class emerges as a particularly important predictor, with workers showing stronger support for Populist and Social-democratic dimensions compared to service professionals. These findings contribute to our understanding of democratic legitimacy and stability by showing how democratic attitudes are embedded in social and political structures. ### [The Short Dark Triad across 14 cultures: A novel network-based invariance approach](https://doi.org/10.1016/j.paid.2025.113321) Dinić, B. M., Costantini, G., Papageorgiou, K. A., Bonfá-Araujo, B., Grabovac, B., Żemojtel-Piotrowska, M., Wertag, A., Kornienko, D., Tripathi, P., Kyriazos, T., & others, including **Tomašević, A.** *Personality and Individual Differences*, 246, 113321 · [Cited by 6](https://openalex.org/works?filter=cites:W4411177739) · [Open access](https://pure.qub.ac.uk/en/publications/626906f0-59af-43fd-a780-d47eb34dd98c) Abstract The aim of this study was to implement and extend exploratory graph analysis to examine the network-based invariance of the Short Dark Triad (SD3) across 14 nations/cultures worldwide. The sample included 15,690 (42.8% male) participants. First, the data were split into two halves: the first half was used for exploratory graph analysis to establish the three-dimensional configural measurement model, while the second half was analyzed using confirmatory factor analysis to test this configuration. Second, metric invariance was assessed within each culture based on community memberships established during configural invariance testing. Finally, network loadings were compared across cultures. The results showed that the network structure achieved both approximate and partial metric invariance, with 13 out of 27 items consistently demonstrating this invariance. The central and invariant item for Machiavellianism involved strategic information tracking, for narcissism–external validation of specialness, and for psychopathy–lack of control, reflecting their unique characteristics. Items related to a revengeful mindset and the demand for deserved respect exhibited the largest absolute differences in network loadings across cultures. The findings support the cross-cultural metric invariance of the SD3, highlighting both culturally universal and culture-specific indicators for each dark trait. ### [Aspects of Identity Questionnaire-IV: Longitudinal Invariance and Predictive Validity for Adolescent Subjective Well-Being and Internalizing/Externalizing Problems](https://doi.org/10.1007/s10826-025-03072-5) Jovanović, V., Lazić, M., **Tomašević, A.**, Šakan, D., Zotović-Kostić, M., Obradović, V., & Gavrilov-Jerković, V. *Journal of Child and Family Studies*, 1–12 · [Cited by 1](https://openalex.org/works?filter=cites:W4410188796) Abstract Self-construal plays an important role in mental health, but there has been only limited evidence from studies in adolescent samples. Research on adolescent self-construal has been hindered by a lack of psychometrically sound measures of self-construal. We investigated the longitudinal invariance and criterion-related validity of the Aspects of Identity Questionnaire-IV (AIQ-IV), a measure developed to assess four identity orientations (personal, relational, public, and collective) posited in the tetrapartite model of the self. The sample included 1340 adolescents (63.9% female, M age = 16.11) who completed measures at two time points 6 months apart. We found support for the four-factor model obtained using the exploratory structural equation modeling (ESEM), which showed evidence of full scalar invariance across time. The concurrent and prospective associations with subjective well-being and internalizing/externalizing problems provided evidence for criterion-related validity of the AIQ-IV. The four identity orientations showed moderate to high stability over 6 months. Our findings suggest that the AIQ-IV may be a useful tool for assessing adolescent self-construal. ### [Adolescent self-construal across cultures: Measurement invariance of the Aspects of Identity Questionnaire-IV in 30 countries](https://doi.org/10.1111/jora.70017) Jovanović, V., Adams, S., Al Banna, M. H., Aritio-Solana, R., Aryanto, C. B., Avsec, A., Bakhshi, A., Bender, M., Berjot, S., Zambrano, S. B., & others, including **Tomašević, A.** *Journal of Research on Adolescence*, 35(2), e70017 · [Cited by 11](https://openalex.org/works?filter=cites:W4409462890) · [Open access](https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/jora.70017) Abstract Despite the critical role of culture in understanding adolescent self and identity, there is a lack of cross-culturally validated measures of adolescent self-construal. The present study evaluated cross-national measurement invariance of the Aspects of Identity Questionnaire-IV (AIQ-IV), assessing four dimensions of self-construal: personal, relational, public, and collective. The sample included 16,795 adolescents aged 14-19 years from 30 countries across four continents. The four-factor structure of the AIQ-IV obtained using exploratory structural equation modeling (ESEM) was supported in the vast majority of countries. Exact invariance testing using multi-group ESEM supported configural invariance, indicating that the overall structure of the AIQ-IV was similar across countries. Full scalar invariance was supported only on a subset of countries (i.e., when tests were conducted using European countries grouped by UN geographical regions). An alignment approach provided evidence for the approximate invariance of the ESEM model, with 15.6% of parameters showing noninvariance and allowing for comparison of latent means. The largest number of noninvariant parameters was evident in Asian countries, with items assessing collective-interdependent aspects of identity showing the most variation across countries. A comparison of mean levels of identity orientations across countries revealed that culture-level dimensions of collectivism-individualism do not translate simply into individual-level dimensions of self-construal. ### [Aspects of Identity Questionnaire-IV: An examination of structural validity, gender invariance, and relationships with mental health and basic psychological needs among adolescents](https://doi.org/10.1080/00223891.2024.2367546) Jovanović, V., Lazić, M., Gavrilov-Jerković, V., Obradović, V., Šakan, D., **Tomašević, A.**, & Zotović-Kostić, M. *Journal of Personality Assessment*, 107(1), 28–40 · [Cited by 8](https://openalex.org/works?filter=cites:W4400032036) Abstract The present research evaluated evidence for structural and convergent validity and measurement invariance across gender of the Aspects of Identity Questionnaire-IV (AIQ-IV), which is designed to measure four identity orientations: Personal, Relational, Public, and Collective. We recruited two independent samples of Serbian adolescents aged 15–19 years (n1 = 436, n2 = 452). Both exploratory structural equation modeling (ESEM) and confirmatory factor analyses (CFA) were applied to examine the factor structure of the AIQ-IV. The ESEM model with four factors demonstrated a superior fit to the data in both samples and resulted in well-defined factors and deflated latent factor correlations compared to the CFA model. Multiple-group analysis supported the scalar invariance of the four-factor ESEM model across gender. The evidence for the convergent validity of the ESEM AIQ-IV factors was supported by meaningful correlations with various mental health indicators (life satisfaction, positive affect, negative affect, internalizing behaviors, and externalizing behaviors) and basic psychological needs (autonomy, relatedness, and competence) satisfaction and frustration. Our findings support the use of AIQ-IV as a multidimensional measure of identity orientations among adolescents and confirm the usefulness of applying the ESEM framework to evaluate the validity aspects of this scale. ### [Is "We" stronger than "I"? The role of identity orientations in predicting adolescent subjective well-being beyond basic psychological needs](https://doi.org/10.1080/15283488.2024.2404508) Lazić, M., Jovanović, V., Šakan, D., Gavrilov-Jerković, V., Zotović-Kostić, M., **Tomašević, A.**, & Obradović, V. *Identity*, 25(4), 501–510 · [Cited by 4](https://openalex.org/works?filter=cites:W4402731797) Abstract The role of self-construal in subjective well-being (SWB) among adolescents is poorly understood. Therefore, the main goal of the present study was to investigate the relationships between self-construal (i.e. four identity orientations: personal, relational, public, and collective) as conceptualized within the tetrapartite model of the self and adolescent SWB (i.e. life satisfaction, positive affect, and negative affect). Furthermore, we examined whether identity orientations have a unique role in predicting adolescent SWB over and above the satisfaction and frustration of basic psychological needs (autonomy, relatedness, and competence). A final sample consisted of 1,933 adolescents from Serbia (62.2% female), with a mean age of 15.66 years. The data were analyzed using the structural equation modeling. We found evidence that collective identity orientation was the most influential aspect of identity for adolescent SWB, while public identity orientation had a weak, but robust negative contribution to SWB. Personal and relational identity orientations had limited predictive value for SWB. Collective identity orientation had a unique contribution in explaining adolescent SWB over and above satisfaction and frustration of basic psychological needs. ### [The face of populism: examining differences in facial emotional expressions of political leaders using machine learning](https://doi.org/10.1007/s42001-025-00392-w) Major, S., & **Tomašević, A.** *Journal of Computational Social Science*, 8(3), 62 · [Overview](https://atomasevic.xyz/papers/face-of-populism.html) · [Cited by 5](https://openalex.org/works?filter=cites:W4411018573) · [Open access](https://arxiv.org/pdf/2304.09914) Abstract Populist rhetoric employed in online media is characterized as deeply impassioned and most often imbued with strong emotions. This paper investigates the differences in affective non-verbal communication of political leaders. We use a deep-learning approach to process a sample of 220 YouTube videos depicting political leaders from 15 different countries, analyze their facial expressions of emotion, and then examine differences in average emotion scores representing the relative presence of six emotional states (anger, disgust, fear, happiness, sadness, and surprise) and a neutral expression for each frame of the processed YouTube video. Based on a sample of manually coded images, we find that this machine learning approach has 53−60% agreement with human annotation. We observe statistically significant differences in the average score of expressed negative emotions between groups of leaders with varying degrees of populist rhetoric. Overall, our contribution provides insight into the characteristics of non-verbal emotional expression among political leaders, as well as an open-source workflow for further computational studies of their affective communication. ## 2024 ### [Dynamic exploratory graph analysis of emotions in politics](https://advances.in/psychology/10.56296/aip00021/) **Tomašević, A.**, & Major, S. *advances.in/psychology*, 2, e312144 · [Cited by 2](https://openalex.org/works?filter=cites:W4401643362) · [Open access](https://advances.in/psychology/wp-content/uploads/10.56296_aip00021.pdf) · [GitHub](https://github.com/atomashevic/dynega-emotions) Abstract This study explores the dynamics of emotions in political leaders' communication using network psychometric methods applied to facial expression recognition (FER) data extracted from YouTube videos. The analysis covers 220 videos of global political leaders and employs zero-shot machine learning via the transforEmotion R package. It focuses on six emotions (happiness, excitement, hope, anger, fear, and sadness) and a neutral expression. Dynamic Exploratory Graph Analysis reveals a two-dimensional network structure for FER scores and their rate of change, showing distinct patterns between positive and negative emotions. The first derivative model indicates a negative correlation between anger and most other emotions, suggesting a more autonomous expression of anger. Significant differences in network structure emerge between leaders with varying degrees of populist rhetoric. More populist leaders exhibit less connected and more autonomous expression of anger, while happiness becomes more contingent on other emotions. In the discussion, we consider the universality of the network structure, the autonomy of anger expression, and the implications of emotional connectivity within the estimated models. The results offer valuable insights for future computational studies of affective political communication, particularly in the context of rising global populism. ### Emotional Dynamics in Online Videos: Integrating Machine Learning and Content Analysis Major, S., & **Tomašević, A.** In *Navigating the Digital Age. An In-Depth Exploration into the Intersection of Modern Technologies and Societal Transformation* (pp. 128–152). Institute for Philosophy and Social Theory ### [The role of basic psychological needs in the relationships between identity orientations and adolescent mental health: A protocol for a longitudinal study](https://doi.org/10.1371/journal.pone.0296507) Jovanović, V., **Tomašević, A.**, Šakan, D., Lazić, M., Gavrilov-Jerković, V., Zotović-Kostić, M., & Obradović, V. *PLOS ONE*, 19(1), e0296507 · [Cited by 10](https://openalex.org/works?filter=cites:W4390495843) · [Open access](https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0296507&type=printable) Abstract Research into the role of identity orientations (the relative importance an individual places on different personal and social attributes and characteristics when defining her or his identity) in adolescent mental health is extremely limited. Furthermore, the potential mechanisms that might explain the associations between identity orientations and adolescent mental health are poorly understood. This study protocol describes a one-year longitudinal study across three time points to be initiated with the purpose of investigating the mediating role of basic psychological needs satisfaction and frustration in the relationship between identity orientations and various mental health indicators in adolescence. We aim to recruit a large sample of Serbian adolescents (N = 2,000 at Time 1), using a two-stage stratified random sampling. The data will be analyzed using the random intercept cross-lagged panel model (RI-CLPM), and the results will be contrasted with the traditional CLPM. The goal of this study is to make a theoretical contribution to research in the fields of identity, self-determination theory, and adolescent mental health, as well as to provide insights towards the development of evidence-based recommendations for creating prevention and promotion programs aimed at improving the well-being of adolescents. ## 2023 ### [Sustainability of Stack Exchange Q&A communities: The role of trust](https://doi.org/10.1140/epjds/s13688-023-00381-x) Vranić, A., **Tomašević, A.**, Alorić, A., Mitrović Dankulov, M. *EPJ Data Science*, 12(1), Article 4 · [Overview](https://atomasevic.xyz/papers/stack-exchange-trust.html) · [Cited by 4](https://openalex.org/works?filter=cites:W4321784003) · [Open access](https://epjdatascience.springeropen.com/counter/pdf/10.1140/epjds/s13688-023-00381-x) · [GitHub](https://github.com/ana-vranic/Stack-Exchange-communities) Abstract Knowledge-sharing communities are fundamental elements of a knowledge-based society. Understanding how different factors influence their sustainability is of crucial importance. We explore the role of the social network structure and social trust in their sustainability. We analyze the early evolution of social networks in four pairs of active and closed Stack Exchange communities on topics of physics, astronomy, economics, and literature and use a dynamical reputation model to quantify the evolution of social trust in them. In addition, we study the evolution of two active communities on mathematics topics and two closed communities about startups and compare them with our main results. Active communities have higher local cohesiveness and develop stable, better-connected, trustworthy cores. The early emergence of a stable and trustworthy core may be crucial for sustainable knowledge-sharing communities. ### [The good, the bad, and the ugly: Revisiting the Dark Core](https://doi.org/10.1007/s12144-021-01829-x) Dinić, B. M., Wertag, A., Sokolovska, V., & **Tomašević, A.** *Current Psychology* · [Cited by 50](https://openalex.org/works?filter=cites:W3160046272) · [OSF](https://osf.io/mty95/) Abstract The aim of this research (N = 268) was to explore several compelling candidates for the Dark Core of the broader set of dark and light traits by using network analysis and analyses of shared variance. Several previously proposed cores were tested: primary psychopathy, narcissistic rivalry, Honesty-Humility, Aggressiveness, Selfishness, Antagonism and its callousness facet as well as emotional competency, including its emotionality facet as negative aspect of callousness. The results showed that central elements in this network are primary psychopathy, callousness, pathological selfishness, Machiavellianism, and narcissistic rivalry. Furthermore, among basic traits, Antagonism shared the highest percentage with the dark traits, especially its facet callousness, which is the best candidate for the Dark Core. The combination of callousness, deceitfulness, and grandiosity shared 92% of the total common variance with dark traits and it could be seen as the Antagonistic Triad or alternative Dark Triad. ### Political Participation of Young People in Serbia: Activities, Values, and Capability Stanojević, D., Vukelić, J., & **Tomašević, A.** In *Young People Shaping Democratic Politics: Interrogating Inclusion, Mobilising Education* (pp. 31–53). Springer International Publishing Abstract Focusing on the case of Serbia, in this chapter we examine the levels of young people’s political activism, conventional and unconventional, by analyzing the factors that influence their formation. At the same time, we provide an overview of the current debates on the transformation of political activity in the twenty-first century, with special emphasis on the practices of young people in the post-socialist context. Our main research questions of the chapters are: How are the values of young people related to the types of political engagement? How does trust in institutions and political actors affect involvement in the political field? Does a sense of capability to participate in political life also lead to real involvement? The specific context of the prolonged social crisis limited economic opportunities, low institutional and social trust, as well as conflicts of different value orientations, finds a specific place in youth participation shaped by instrumental goals, clientelism, and authoritarianism. On the one hand, conventional politics, with its informal rules of the game, developed clientelism, the still traditional structure of political parties advocating mass numbers mobilizes a relatively large number of young people compared to other European countries. But this activity is only partly inspired by political and ideological goals, and the dominant strategy of young people is getting a job and social promotion through these channels. On the other hand, the semi-authoritarian regime in which young people live often leads to self-organizing and various alternative and protest politics which are grounded in the traditions of the 1990s. ### [The role of trust in sustainability of knowledge-sharing social groups: the case of Stack Exchange Q&A communities](https://pos.sissa.it/427/217/) **Tomašević, A.**, Vranić, A., Alorić, A., & Mitrović Dankulov, M. In *11th International Conference of the Balkan Physical Union* (p. 217) Abstract Knowledge-sharing communities are fundamental for the development and evolution of any knowledge-based society. Their emergence, function, and disappearance determine the course of evolution of a knowledge-based society. The sustainability of these groups is crucial for the success of the knowledge-transfer process in modern societies and the efficiency and success of this process. This work explores the role of the structure of social interactions and social trust in the emergence of sustainable knowledge-sharing communities. We combine tools and methods from complex networks theory, statistical physics, computer science, and sociology to explore roles mentioned in the sustainability of StackExchange communities on four different topics: astronomy, physics, economics, and literature. StackExchange is one of the most successful online knowledge-sharing networks that hosts more than 150 communities on various topics. To control the influence of the subject, we select a pair of active and one closed community for each topic and analyze and compare their early evolution. We adapt the dynamical reputation model to quantify the change in social trust in these communities. We analyze the evolution of the social interaction network and social trust between members during the first 180 days of their existence. Our results show that sustainable communities have higher local cohesiveness and develop stable, more strongly connected cores. The social trust between members is more heightened in sustainable communities. In these communities, the trust between core members develops early and remains high over time. This work shows that the emergence of a stable, trustworthy core may be determining factor in building a sustainable knowledge-sharing community. ## 2022 ### [The narcissism network and centrality of narcissism features](https://doi.org/10.1007/s12144-020-01250-w) Dinić, B. M., Sokolovska, V., & **Tomašević, A.** *Current Psychology* · [Cited by 29](https://openalex.org/works?filter=cites:W3119317537) · [OSF](https://osf.io/hfdsx/) Abstract The aim of this research was to explore the underlying structural organization and centrality of narcissism features by using network analysis. The research was conducted on a sample from the general population using subscales of four narcissism instruments (Narcissism Personality Inventory, Grandiose Narcissism Scale, Pathological Narcissism Inventory, Five-Factor Narcissism Inventory Short Form). The results revealed detection of four communities, one of which was interpreted as vulnerable narcissism, while the remaining three were interpreted as aspects of grandiose narcissism (grandiose exhibitionism, narcissistic antagonism, and authority). The results suggested that although features of grandiose narcissism (grandiose exhibitionism and leadership/authority) show consistent higher strength centrality across networks, entitlement/exploitativeness, followed by grandiose fantasy, has a central role in bringing together maladaptive grandiose and vulnerable aspects of narcissism. In addition, leadership/authority brings together various aspects of grandiose narcissism. The results support the role of entitlement features as the core of narcissism. ### Catalytic capital investment as an enabler of affordable rental and cooperative housing in Central and South-Eastern Europe: Full Research Report Pósfai, Z., Jelinek, C., Dević, S., **Tomašević, A.**, & Pavlović, I. Periféria Policy and Research Center Abstract This Research Report looks at the potential of catalytic capital investments in the housing sector of Central and South-Eastern Europe (CSEE), and more specifically in eight countries of this region - Bosnia and Herzegovina, Bulgaria, Czechia, Croatia, Hungary, Northern Macedonia, Serbia, Slovenia - as explored by a research consortium from five countries during 2022. We define catalytic capital investments as “investments that are more patient, risk-tolerant, concessionary, and flexible than conventional capital”. Countries of the CSEE region have experienced quickly increasing housing inequalities in the past decades. The public housing stock is marginal and severely residualized, and private rental housing markets are small and severely under-regulated. Thus, they do not provide adequate stability and are too expensive for the majority of the population. 20-30% of the population, including middle-class households, struggles with housing affordability issues, and housing deprivation rates are high. ## 2021 ### [Housing and family trajectories of young adults in five countries: Sweden, Germany, United Kingdom, Italy and Serbia: Sequence analysis of European social survey data](https://doi.org/10.2298/SOC2102262S) Stanojević, D., & **Tomašević, A.** *Sociologija*, 63(2), 262–288 · [Cited by 3](https://openalex.org/works?filter=cites:W3185880380) · [Open access](http://www.doiserbia.nb.rs/ft.aspx?id=0038-03182102262S) Abstract The aim of this paper is to analyse housing and family transitions among the young and young adults in five countries: Sweden, Germany, United Kingdom, Italy and Serbia, representing the Social-democratic, Conservative, Liberal, Mediterranean and (SEE) Post-socialist models of welfare regimes. For the purposes of our analysis, we used round 9 of European Social Survey data. The focus of our analysis was on the rotating module ?Timing of life? which aims to capture the views of European citizens about their life courses and their strategies to plan their own lives, as well as measures the timing of key life events. Variables from this module were used to construct life trajectories of respondents which are statistically modelled as sequences. Interpretation of the obtained results leads to two important conclusions. First, the differences in the types of family transitions of young people between countries are significant. Second, these differences can be explained both by individual characteristics and by the social and cultural context that determines the horizon of opportunities for young people. Even after controlling the effects of individual characteristics such as gender, age, education, parental education, religious affiliation, statistical differences between countries persist, indicating that a significant part of variability cannot be explained on an individual-level but exclusively by social and institutional context. ### [Measuring the influence of beliefs in belief networks](https://arxiv.org/abs/2110.09154) **Tomašević, A.** *arXiv preprint arXiv:2110.09154* · [Cited by 1](https://openalex.org/works?filter=cites:W3207125685) · [Open access](https://arxiv.org/pdf/2110.09154) Abstract Influential beliefs are crucial for our understanding of how people reason about political issues and make political decisions. This research proposes a new method for measuring the influence of political beliefs within larger context of belief system networks, based on the advances in psychometric network methods and network influence research. Using the latest round of the European Social Survey data, we demonstrate this approach on a belief network expressing support for the regime in 29 European countries and capturing beliefs related to support for regime performance, principles, institutions, and political actors. Our results show that the average influence of beliefs can be related to the consistency and connectivity of the belief network and that the influence of specific beliefs (e.g. Satisfaction with Democracy) on a country level has a significant negative correlation with external indicators from the same domain (e.g. Liberal Democracy index), which suggests that highly influential beliefs are related to pressing political issues. These findings suggest that network-based belief influence metrics estimated from large-scale survey data can be used a new type of indicator in comparative political research, which opens new avenues for integrating psychometric network analysis methods into political science methodology. ## 2020 ### [Centrality and redundancy of the Dark Tetrad traits](https://doi.org/10.1016/j.paid.2019.109621) Dinić, B. M., Wertag, A., **Tomašević, A.**, & Sokolovska, V. *Personality and Individual Differences*, 155, 109621 · [Cited by 52](https://openalex.org/works?filter=cites:W2987182636) · [OSF](https://osf.io/gmnky/) Abstract The aim of this research was to examine centrality and redundancy of the Dark Tetrad traits (psychopathy, Machiavellianism, narcissism, and sadism) using network analysis. The first sample (N = 546) was assessed using a short instrument, the second (N = 404) and the third (N = 410) samples were assessed with full instruments for the first three dark traits, while for sadism the same instrument was used in all three studies. The results showed that psychopathy is the central feature across all networks, especially its facets which correspond to primary psychopathy or interpersonal manipulation and callousness. Narcissism seemed redundant when total scores of the dark traits were analyzed, but these results should be interpreted with caution given the small number of variables in the network. However, on the facet level, some facets of psychopathy were redundant (secondary psychopathy or lack of cognitive responsiveness). These results reiterate the importance of psychopathy in the core of the dark traits and provide a deeper insight into the relations between the Dark Tetrad traits.