This article utilizes exit poll data from IPSOS to examine the sociodemographic determinants of voting behavior in the 2023 Polish parliamentary elections. Applying multinomial logistic regression, we examine the independent effects of age, gender, education, occupation, and place of residence on party choice, with PiS as the reference category. The results reveal clear structural divides: younger voters, better-educated voters, and urban voters disproportionately support KO [Koalicja Obywatelska, Civic Coalition], NL [Nowa Lewica, The Left], and, in the case of young men, Konfederacja [Confederation], while older voters, less-educated voters, and rural voters remain PiS’s electoral base. Gender effects are significant, with men more inclined toward Konfederacja and women toward NL. Occupational status further differentiates preferences, with business owners leaning toward Konfederacja and pensioners toward PiS. Predicted probability plots illustrate these dynamics and confirm the persistence of class, generational, and spatial cleavages in Polish politics. Methodologically, the study demonstrates the analytical potential of exit poll data, offering a replicable framework for future electoral research in Poland and beyond.
Easy Language guidelines discourage the use of full overt pronouns and subordinate clauses. This study examines how manipulating these linguistic elements influences perceived reading difficulty, reader preference, and the cognitive effort involved in text processing. We also assessed the effects of null and overt pronouns on content comprehension. 26 adults with intellectual disabilities and 26 neurotypical adults participated in a mixed-design experiment involving two reading tasks. For both tasks, we recorded participants’ eye movements while they rated the perceived difficulty and reported their preference. Additionally, Task 1 included open-ended comprehension questions. The linguistic elements only had significant effects on fixation patterns and focal attention. Crucially, they did not significantly alter perceived text difficulty. These findings suggest that including subordinate clauses does not significantly hinder comprehension. However, further research is essential to guide the future development
of Easy Language, as evidence is still limited regarding its additional linguistic and graphic features and the needs of different target groups.
This study examines the acculturation strategies employed by international graduates of English-language management programs in Poland and their changes over time, focusing on their experiences during their studies and as graduate professionals. Following the Qualitative Longitudinal Study approach, the research conducts an in-depth exploration of participants' experiences across two time points: initially during their student years and again two years after graduation. Eleven participants (aged 20–27 at the time of the second interview) engaged in semi-structured interviews and created cultural identity maps. Through Interpretative Phenomenological Analysis, we identified four distinct acculturation strategies: separation, individualism, bicultural integration, and multicultural integration, with most participants actively engaging with multiple cultures during the second wave of interviews. Furthermore, the longitudinal analysis highlighted the participants’ perspectives on the unfolding of the acculturation process over time, revealing shifts in their identities, values, and behaviors. Participants conceptualized acculturation as an ongoing, lifelong process rather than a finite transition, particularly as their experiences intersected with developmental transitions into adulthood. This study contributes to understanding how specific contextual factors, such as international academic and professional environments, influence acculturation processes. Methodologically, this study underscores the value of employing cultural identity mapping from a longitudinal perspective to examine acculturation as a dynamic process.
Evaluating policy interventions for displaced populations requires authentic stakeholder perspectives, yet traditional methods prove inadequate during humanitarian crises. This paper presents a meta-evaluation of an AI-powered tool utilizing Retrieval-Augmented Generation technology to process 521,089 Telegram messages from Ukrainian and Russian-speaking populations. Our central finding challenges initial design assumptions: while intended to simulate stakeholder perspectives, the tool proved most valuable as a document exploration platform for qualitative data analysis. This shift from simulation to exploration represents a significant methodological insight for evaluation practice. The technical architecture successfully implemented multi-stage filtering, hierarchical clustering into 249 thematic groups, and transparent retrieval mechanisms. We argue that AI technologies offer greatest promise not in replacing stakeholder engagement, but in enhancing evaluators’ capacity to systematically process qualitative data. This research contributes to debates on responsible AI integration in evaluation methodology.
Experiments that manipulate the presence of mimicry generally find that mimicry benefits the mimicker. These results led to the "mimicry-as-a-social-glue" hypothesis, which considers mimicry as a mechanism responsible for starting and maintaining social relations. There are two novel aspects in the present pre-registered experiment. First, the experiment examines temporal aspects by including four conditions: no mimicry, mimicry during the first five minutes, the last five minutes, or mimicry present throughout the interaction. By doing so, we contribute to ongoing efforts to standardise mimicry methodology. Second, this explores the underexplored issue of potential costs associated with mimicry and challenges the "mimicry-as-a-social-glue" hypothesis. The results demonstrate a relationship between temporal factors and the effects of mimicry. Participants who were mimicked during the final five minutes of the interaction reported significantly lower self-esteem compared to those mimicked during the initial five minutes. However, no significant differences in self-esteem were found between the no mimicry condition and mimicry during the first five minutes, nor between the final five minutes condition and mimicry sustained throughout the entire interaction. Similarly, no significant differences were observed between the no mimicry condition and the condition in which mimicry was sustained for the entire interaction. These findings suggest that the costs and benefits associated with mimicry depend on its temporal dynamics.
