Our publications

Academic resilience—students’ capacity to succeed despite socioeconomic disadvantage—is widely studied, but little consensus exists on how it should be operationalized. Inconsistent definitions make it difficult to compare findings across studies and to draw conclusions about what promotes resilience. The present study asks whether different operationalizations and thresholds lead to different estimates of resilience and different conclusions about its predictors. Using PISA 2018 data from 75 education systems and based on an existing measurement framework, we compared three ways of operationalizing academic resilience: identifying high-achieving disadvantaged students, identifying students who outperform expectations given their socioeconomic status, and predicting achievement within a disadvantaged sample. We applied each at three relative threshold levels of disadvantage and/or high achievement, and assessed the consequences for the estimated prevalence of resilient students and for the associations of four factors (gender, self-perceived reading competence, reading enjoyment, attitude toward learning) with resilience. Prevalence estimates varied nearly fivefold across operationalizations, and rankings of education systems differed substantially. When associations were pooled meta-analytically, they were broadly consistent across operationalizations and threshold levels. When education systems were analyzed separately, however—as is standard in PISA-based research—findings were consistent in only 20–72% of systems, depending on the factor, and threshold level alone altered conclusions in 3–37%. A complementary interaction analysis of risk and protective factors, which requires no threshold, further showed that some motivational factors were associated with a flatter socioeconomic gradient in achievement and others with a steeper one—a pattern that threshold-based definitions cannot detect. Operationalization and threshold levels can meaningfully affect the conclusions studies reach about academic resilience and its predictors. While part of this dependence is mechanical, given how the operationalizations are constructed, the magnitude of this dependence across education systems suggests that findings should not be interpreted independently of the operationalization used to produce them. We provide recommendations for matching operationalization choices to research questions.

This study investigates longitudinal associations between peer relations, affective well-being, and academic achievement during early adolescence. Data were drawn from a two-wave study of a large nationally stratified sample of Czech students transitioning from Grade 6 to Grade 7 (N = 2,609, ages 12–13). Cross-lagged panel models were used to examine reciprocal links between friendship, social exclusion, positive and negative affect, and achievement in mathematics and Czech language, assessed using item-response-theory-scaled tests. Friendship predicted increases in subsequent positive affect, whereas social exclusion predicted declines in positive affect. Exclusion was also associated with lower later achievement in Czech language across students and with reduced mathematics achievement specifically among girls, indicating domain- and gender-differentiated academic vulnerability. Affective well-being did generally predict achievement. These findings highlight the developmental importance of peer relations for emotional adjustment and suggest that experiences of exclusion may carry direct academic consequences, particularly for girls in mathematics.

International student mobility is often framed as an integrative policy instrument, yet mobility systems frequently display durable inequalities and core–periphery structures. This study examines whether a regionally bounded program outside Western Europe’s historical academic core functions primarily as an infrastructure of integration or as a mechanism that reproduces stratification. Using openly accessible administrative records from the CEEPUS, the analysis reconstructs a 20-year longitudinal bipartite network of higher education institutions and thematic network agreements (2005/2006–2024/2025). To identify the micro-mechanisms governing network change, the study estimates a bipartite stochastic actor-oriented model in which institutions adjust their memberships over time, while accounting for the creation and discontinuation of thematic networks. The model tests closure mechanism, cumulative advantage, and selection processes based on geopolitical position, research prestige, and institutional type. Results indicate that cumulative advantage operates strongly at the institutional level: institutions already embedded in multiple thematic networks are substantially more likely to expand further. Stratification is further reinforced by prestige-based activity and prestige homophily, with higher-prestige institutions both participating more and clustering with similarly positioned partners. Relative importance analysis confirms that prestige-based effects constitute the dominant attribute-based force governing tie formation decisions throughout the observation period, consistently outweighing both geopolitical and organizational mechanisms. Geopolitical and organizational boundaries also shape collaboration: institutions in European Union member states expand less on average, yet institutions cluster by shared European Union status, and public and private institutions show both differential activity and strong within-type clustering. Overall, the program expands and stabilizes over time but evolves toward patterned integration rather than uniform cohesion, with stratified collaboration dynamics consistent with persistent academic hierarchy.

This study examines how Generative Artificial Intelligence (GenAI) use is associated with academic achievement among Czech seventh-grade students. Grounded in the deep versus surface approaches to learning framework, we distinguish between the frequency of GenAI use and students’ endorsement of reliance on GenAI over understanding, defined as the belief that knowing how to prompt AI effectively is more important than developing one’s own conceptual understanding. Using cross-sectional survey data from a nationally stratified sample of 2,307 students, we estimated structural equation models predicting IRT-scaled achievement in mathematics and Czech language. Frequency of GenAI use was not positively associated with achievement in either subject. In contrast, stronger endorsement of reliance on GenAI over understanding was consistently and negatively associated with achievement, with a more pronounced relationship in Czech language. Neither GenAI use frequency nor reliance significantly moderated the associations between learning-related dispositions and achievement. These findings suggest that the educational implications of GenAI depend less on how often students use AI tools and more on whether AI is approached as a support for understanding or as a substitute for cognitive engagement. The results contribute to debates on epistemic offloading and the role of AI in shaping students’ orientations toward learning.

This paper examines how the taboos and problems perceived by doctoral students are associated with their well-being and their intentions to persist in or leave academia. We distinguish between doctoral problems—stressors associated with doctoral study—and doctoral taboos, understood as issues that remain difficult to voice in relationships with supervisors or peers because their disclosure may evoke shame or threaten these relationships. Drawing on a mixed-methods design conducted in the Czech Republic, we first developed core perceived doctoral problems and taboos based on a qualitative analysis of 21 in-depth interviews. Doctoral problems and taboos were subsequently examined in a national survey of 190 doctoral students using machine learning models. Financial difficulties constituted the most prevalent doctoral problem, reported by roughly 75% of students. Partial dependence analyses indicated that several doctoral taboos (e.g., finances, parenthood, supervisor quality) were among the strongest predictors of burnout, depression, anxiety, stress and intentions to drop out or leave academia. The study contributes to debates on well-being in doctoral education by illuminating which issues within doctoral education become taboo and how such tabooization relates to doctoral well-being and intentions to remain in academia.

This study examines emotional-state homophily—the tendency to form relationships with peers sharing similar emotions—among K-12 students in the Czech Republic. Using anonymized data from 9,176 students in 357 classrooms, we constructed communication networks for both present-state and desired communication frequencies. Employing network modeling, we tested hypotheses about emotional-state homophily, as well as the moderating role of age, while controlling for network structures and gender effects. Our analysis revealed significant homophily tendencies for happiness, loneliness, and stress in both present-state and desired communication networks. Additionally, happiness-based homophily increased with age, while loneliness-based homophily diminished. Unexpectedly, stress and hopelessness were associated with higher tie frequencies, challenging assumptions about their isolating effects.

This study applies psychological network analysis (Gaussian graphical model) to examine cross-sectional links between bullying and victimization (physical, verbal, rumor-spreading, exclusion, and cyber), internalizing symptoms (depression, anxiety, social phobia, low self-esteem), and externalizing problems. Adolescents (N = 1,519; Mage = 12.5) from 18 schools completed self-reports at two time points (3 months apart). Network models were estimated separately for each wave. The analysis showed that bullying and victimization overlapped most in verbal and physical forms, less so in rumor-spreading, peer exclusion, and cyber forms. Victimization types were linked to distinct psychological outcomes: physical victimization to self-related worries, verbal victimization to feelings of worthlessness, peer exclusion to sadness, rumor-spreading to social anxiety and appetite problems, and cyber victimization to social phobia and low self-esteem.

Research shows victims have lower peer status, but whether bullying participants differ in (dis)liking them and if classroom context matters remains unclear. Early adolescents (N?=?751) were assessed twice over six months for (dis)liking and reputational nominations on physical bullying, defending, and victimization. SAOMs showed that victims did not experience a decline in being liked or an increase in being disliked over six months and almost no differences in how bullies, victims, and defenders (dis)liked (other) victims were found, except that victims developed disliking toward other victims over time. Despite assumptions of the healthy context paradox, classroom victimization did not moderate (dis)liking toward victims. Overall, the (dis)liking was stable and largely consistent across bullying participants and classroom contexts, though concerning dislike emerged among victims.

This multi-authored article explores the emerging concept of critical GenAI literacy and its postdigital configurations. Fourteen authors contributed with different sections, followed by five author-reviewers who examined the article as a whole. The objective was to invite diverse perspectives, constructive critique, and promote collaborative efforts to develop a clearer and more inclusive understanding of critical GenAI literacy. The introduction focuses on establishing some initial assumptions about what we mean by ‘critical’ in general, and for GenAI literacy in particular, and why a conjunction of postdigital configurations of the construct emerges as an adequate methodological solution. In the main part of the article, some authors focus on the concept of ‘literacies’ as a starting point for their reflection, while others directly examine critical GenAI literacy through a postdigital perspective. We conclude that critical GenAI literacy requires moving beyond technical skills to engage with AI’s epistemological, ethical, and relational dimensions, ensuring learners critically interrogate its role in knowledge production.

This social network study investigated whether dyad- and group-level (dis)liking relationships affect perceptions of classmates' involvement in physical aggression and explored the moderating roles of classroom moral disengagement and defending norms. The study employed a longitudinal design with two time points 6 months apart, encompassing 27 classrooms and 632 early adolescents. Using multiplex stochastic actor-oriented modeling, we found that liking, but not disliking, significantly influenced perceptions. Liking a classmate increased the likelihood of perceiving them as a defender. Moreover, students' own perceptions (aggressor, victim, and defender nominations) were shaped by the perceptions of classmates they liked, while classroom moral disengagement reduced this influence for defender nominations. Results on classroom defending norms were mixed.

This study explores the social integration of Ukrainian refugee students in Czech schools, focusing on peer network dynamics in classrooms with a significant proportion of ethnic minority students. We employ longitudinal social network analysis combined with qualitative insights to investigate the evolution of friendship and exclusion ties among 266 students (ages 11–15, 45.11% girls, 21.05% Ukrainian) in 12 classrooms across two time points during the 2022/2023 school year. Our study introduced socioeconomic status (SES) and sociocultural adaptation as key explanatory variables, hypothesizing their roles in fostering cross-ethnic interactions. Overall, we found that ethnic homophily decreased over time, indicating greater cross-ethnic mixing, even as overall number of friendships in classrooms declined. Ukrainian students were initially less integrated into peer networks, exhibiting fewer friendships and strong preferences for intra-ethnic ties. However, this disparity became less pronounced as the school year progressed. Sociocultural adaptation emerged as a significant predictor, with adapted Ukrainian students more likely to receive friendship ties and less likely to receive exclusion ties, highlighting the critical role of pro-social behavior in social integration. Conversely, SES did not significantly affect the formation of cross-ethnic ties.

This study examines the long-term effects of pre-primary education on peer relationships in sixth grade using a stratified random sample of 2,397 sixth-grade students from the Czech Republic. The findings reveal that students who attended pre-primary education for more than one year, are less rejected compared to those who did not attend. Ethnic minority students and those with parents without higher education are identified as more prone to peer rejection, underscoring the challenges these groups face in school. Importantly, this study shows that pre-primary education provides some protective effect against peer rejection for these vulnerable groups.

A multi-group stochastic actor-oriented model was employed to examine the effects of social selection and peer influence through friendship ties on Czech language grades and reading literacy among lower-secondary students, controlling for students' SES, gender, and network structure. The sample comprised 276 sixth-grade students across twelve classrooms, measured at two time points with an interval of seven months. No evidence of selection or influence in terms of grades and literacy was found. However, effects of gender and SES assortativity were observed.

With the opportunities and challenges stemming from the artificial intelligence developments and its integration into society, AI literacy becomes a key concern. Utilizing quality AI literacy instruments is crucial for understanding and promoting AI literacy development. This systematic review assessed the quality of AI literacy scales using the COSMIN tool aiming to aid researchers in choosing instruments for AI literacy assessment. This review identified 22 studies validating 16 scales targeting various populations including general population, higher education students, secondary education students, and teachers. Overall, the scales demonstrated good structural validity and internal consistency. On the other hand, only a few have been tested for content validity, reliability, construct validity, and responsiveness. None of the scales have been tested for cross-cultural validity and measurement error. Most studies did not report any interpretability indicators and almost none had raw data available. There are 3 performance-based scale available, compared to 13 self-report scales.

This study provides insights into the demographic predictors of public speaking anxiety, using a sample of 1745 students from a large public university in the Czech Republic. We employed the Personal Report of Confidence as a Speaker in a short form to assess public speaking anxiety levels. Our multivariate regression analysis identified gender, type of high school, and study level as significant predictors of public speaking anxiety, whereas age, nationality, and field of study were not. The study found that women, non-binary individuals, graduates of academic high schools, and bachelor’s students are more prone to public speaking anxiety.

The aim of this meta-analysis was to evaluate the effects of performance-based financial incentives on higher education students using causal evidence. A total of 18 randomized controlled trials involving 20,286 students were included. Performance-based financial incentives increased the number of college credits earned, marginally improved student grade point averages, and improved exam scores when targeted at a single subject. No evidence was found for differential effects on student populations by gender or first-generation college status. Neither the incentive amount nor a focus on low-income students influenced the incentive effectiveness in improving student grade point averages or credits earned. Tentative evidence suggests that incentive designs in which the number of reward recipients is limited are more beneficial to students from the upper median level of high school grade point average. Hence, the study implies that it is more beneficial to provide lower-amount incentives attainable by a larger number of students than to provide higher-amount incentives attainable only by a small number of students.

We employed dynamic actor-oriented models on a sample of 145 Czech lower-secondary students in 62 small groups and pooled the results from the groups with a meta-analytical procedure. We found bursty behavior resulting from endogenous structural mechanisms of reciprocity, transitivity, cyclicity, and preferential attachment. Students gave preference to initiating interactions with those they initiated interactions with before and off-task interaction contributed to the development of on-task interaction. Students strongly preferred interactions with friends. Those students who talked a lot during regular whole-classroom lessons and students with high levels of literacy tended to both initiate and receive more interactions in group work, and students similar in these attributes preferred to interact with each other. Group size did not affect preferential attachment tendencies in interaction, but smaller groups made the effect of friendship ties on interactions stronger, and communication group norms shifted with changing group composition.

This study investigates Slovak political discourse on Facebook in the weeks before and after the invasion began. We examine the discourse through the lens of Discourse Network Analysis, combining qualitative content analysis of MPs’ Facebook posts with quantitative bipartite social network analysis. During the two weeks, we retrieved all posts from all MPs who had an active public Facebook page. We manually coded all posts and created a bipartite discourse network consisting of MPs connected to each other via shared discourse codes in two timepoints. We performed a series of exploratory analyses that identified the content of the political discourse, the structure of the political discourse network, and the mechanisms driving the change of the political discourse network. Our results show that the invasion dramatically changed political discourse in Slovakia, with the domestic coalition-opposition struggles losing prominence among the invasion-related topics. The structure of the political discourse network showed a strong coalition-opposition split. While coalition MPs had largely pro-Ukrainian sentiments, opposition MPs largely communicated pro-Russian propaganda. A cluster of opposition MPs consistently spread conspiracy theories both before and after the invasion began, supporting a “conspiracy singularity” theory—the tendency of actors to spread multiple different conspiracy theories and interconnect various conspiracy theories into one overarching narrative. The change of the discourse network at the beginning of the invasion was largely driven by the agenda setting of several parties, agenda reinforcement, and increasing political polarization.

We provide a brief insight into the integration of Ukrainian refugees in school social networks in the Czech Republic following the mass migration caused by the Russian invasion of Ukraine. Our sample contains twelve classrooms with a total of 266 students in grades 5 to 9; 21.05% of the students were of Ukrainian origin. We employed multiplex exponential random graph modelling to assess the level to which Ukrainian refugees were integrated within peer networks, capturing both friendship and exclusion ties. We then employed a meta-analytical procedure to aggregate the results from the individual classrooms and a meta-regression to study the relationship between classroom ethnic composition and the level of integration of Ukrainian refugees. We found social networks to be formed heavily along ethnic lines with strong ethnic homophily in friendship ties and a propensity of the Ukrainian students to both send and receive fewer friendship ties than their Czech classmates. We found no evidence that the Ukrainian students sent or received more exclusion ties than their Czech classmates, suggesting that the Ukrainian students did not face explicit rejection from classmates; rather, we saw a tendency of the Ukrainian students to be neglected. Our findings stand in contrast to reports from school headmasters who asserted that the social integration of Ukrainian students was seamless. We further found a higher proportion of Ukrainian students in classrooms to be related to stronger homophilic behaviour and a lower tendency of Ukrainian students to make friends. Our results therefore imply that increased classroom diversity may negatively influence the integration of refugees in social networks.