Scholarly interest in individuals’ beliefs regarding their political competence and their capacity to influence political developments has a long tradition in the political and social sciences (Balch, 1974; Campbell et al., 1954). In this context, PE is commonly defined as “the feeling that individual political action does have, or can have, an impact upon the political process” (Campbell et al., 1954, p. 187). Within this conceptualization, two distinct but related dimensions of PE can be differentiated: external and internal PE (Balch, 1974; Beierlein et al., 2012; McDonnell, 2020; Morrell, 2003; Oser et al., 2022; Pollock, 1983). External PE has been theorized as the belief that authorities, such as governments or politicians, and the political system as a whole, are responsive to citizens’ concerns and that individuals’ political actions can influence governmental decisions (Balch, 1974; Beierlein et al., 2012; Vetter, 1997). Internal PE, by contrast, refers to the individual’s self-perception of being capable of engaging with political issues (Morrell, 2003). This includes, for instance, the perceived ability to comprehend political issues and to participate effectively in political life (McDonnell, 2020; Niemi et al., 1991; Oser et al., 2022). Accordingly, external PE pertains to beliefs about the responsiveness of the political system itself, whereas internal PE reflects self-attributed beliefs and abilities. Empirical findings support the assumption that these two constructs are independent of one another (e.g., Beierlein et al., 2012; Craig et al., 1990; Vetter, 1997). This distinction is corroborated by their divergent correlations with external criteria: external PE is positively associated with trust in political institutions and confidence in the functioning of the political system (Boulianne, 2019; Geurkink et al., 2020; Niemi et al., 1991). Internal PE, on the other hand, is positively associated with political participation, political interest, political knowledge, and educational attainment (Niemi et al., 1991; de Gil Zúñiga et al., 2017). Moreover, empirical studies demonstrate that PE tends to be experienced more strongly with regard to more proximal, local than with regard to more distant, national politics (Denters, 2002; McDonnell, 2020). From a theoretical perspective, this asymmetry has important implications for perceptions of political participation and democratic legitimacy: citizens tend to perceive institutions as more accessible, responsive, and open to influence, which strengthens their belief in the value of political engagement. Conversely, the complexity, scale, and institutional distance of national politics may attenuate perceptions of individual influence, thereby potentially weakening both institutional trust and participatory orientations at the national as compared to the local level.

THMs are generally viewed as forums where community members engage directly with policymakers (Field, 2019). They serve as platforms for two-way communication: leaders present policy updates, while participants ask questions, express concerns, and offer feedback. Their open and dialogic nature ideally supports both external and internal dimensions of PE. THMs are thus valuable for studying perceptions of institutional responsiveness and political participation although few studies have examined immediate consequences of participating, evidence suggests THMs can enhance trust in politicians (Minozzi et al., 2015; Spitzer and Weber, 2023). When citizens feel heard and see their feedback reflected, participation should therefore enhance confidence in policymakers, an expression of external PE.

Conceptually, we treat THMs not as a single intervention but as a multi-stage participatory format which may affect PE through distinct mechanisms. Early stages include invitations and public communication, which may raise awareness of THMs, as well as self-selection—randomly invited or not—into participation, whereby attendees may differ systematically from non-attendees. Subsequent stages involve agenda formation and question submission, potentially fostering internal PE through reflection and articulation. Our study focuses on a later stage of this process, namely participants’ experience of responsiveness when they see citizen’s questions acknowledged and answered, which should constitute a key pathway to external PE. This process perspective highlights that THMs combine informational, relational, and deliberative elements, and that observed effects depend critically on the stage of participation under study.

Utilizing data collected at a series of five THMs, the purpose of this study is to investigate whether THM participation has measurable effects on citizens’ perception of institutional responsiveness as well as their self-attributed ability to understand political processes. We employ a quasi-experimental pre–post design using two-wave survey data from five THM sessions. We compare changes in PE between randomly recruited active participants and passive listeners who attended independently and served as a comparison group. This approach is methodologically appropriate for examining short-term attitudinal dynamics in a field setting.

We hypothesize that participation in a THM should foster a stronger perception of responsiveness by institutions and politicians. Accordingly, we expected external PE to be higher after the event compared to before it (Hypothesis 1a). This effect should be particularly pronounced for participants who engaged actively rather than those who only listened passively (Hypothesis 1b). Moreover, because taking part in a THM may facilitate a better understanding of political processes and dynamics, we expected internal political efficacy to significantly increase following participation (Hypothesis 2a). Once again, this effect should be more pronounced for active participants than for passive listeners (Hypothesis 2b).

Active participants were drawn as a random sample from the Bielefeld population register and invited to attend. For reasons of data security, invitations were administered by our collaboration partners within the city administration. After receiving the invitation, individuals registered for a specific THM session. Passive listeners were able to join voluntarily. The THMs were publicly announced, and passive participants were able to register via an online portal. Five THMs were examined for this study, each with 90 participants: 40 active participants and 50 passive listeners. Each session addressed a different policy issue—communal life, city development, public finances, education, and public security. The Mayor was present at each THM. Moreover, since the THMs were thematically organized by municipal “departments” (Dezernate), the respective department heads participated as well, with the exception of THM 5 (Public Security), where the Chief of Police attended instead. Active participants could submit discussion points or questions anonymously, under their full name, or directly via microphone. However, not all participants submitted at least one question. Any questions that were not addressed during the THM were subsequently answered and made available on a webpage afterwards. Sessions began with brief policymaker statements, followed by an extended discussion based on participant contributions. Each meeting lasted 120–180 min and took place in a neutral venue in the city center. At t1, participants completed pen-and-paper questionnaires placed on their seats, using an anonymous personal code to link pre- and post-measures. One day later, they received an email invitation to complete an online questionnaire (t2), available for seven days. All paper questionnaires were later digitized for analysis.

To measure the two dimensions of PE, an adapted version of a validated German short scale (Beierlein et al., 2012) was used. Internal PE (“I am good at understanding and assessing important political issues” and “I take active part in discussions about political issues”; with an inter-item correlation of r = 0.50 at t1 and t2; Spearman-Brown coefficients r_SB = 0.67 at t1 and t1, respectively) and external PE (“Politicians care about what people think”, “Politicians strive to keep in close touch with the people”; with an inter-item correlation of r = 0.56, r_SB = 0.72, at t1 and r = 0.71, r_SB = 0.83 at t2) were measured with two items each. Demographic information were measured as follows: age in cohorts (16–24, 25–40, 41–65, 66 and older), gender (female, male, diverse), educational attainment (lower secondary school, intermediate secondary school, entrance qualification for higher education, not yet completed school qualification), and vocational degree (apprenticeship/vocational training, technical college qualification, Bachelor’s, Master’s, Diploma, Doctorate, and no degree obtained yet).

At t1, a total of 250 participants completed the questionnaire; 106 also took part in the post-event survey (t2). Of these, 52 were men and 52 women (49.1% each), one identified as diverse, and one gave no specification (0.9% each). Two participants (1.9%) were aged 16–25, 35 (33.0%) 25–40, 49 (46.2%) 40–65, and 20 (18.9%) were aged 66 and older. In terms of educational attainment, five (4.8%) reported not having obtained a formal qualification, 24 (22.6%) a job entrance qualification, eight (7.5%) a technical college degree, 62 (58.5%) an academic degree, and five (4.8%) a PhD; two (1.9%) did not specify. Overall, there were no substantial differences between participants who were retained at t2 and those who dropped out after t1. Two exceptions emerged: respondents who participated at t2 were marginally less likely to be aged 16–24, and among passive listeners, those retained at t2 showed significantly higher levels of internal PE than those who dropped out after t1. A detailed overview of panel attrition by sociodemographic characteristics and PE indicators is provided in Appendix A. At t1, 80.2% of participants (203) reported that they had never attended a THM before. A further 16.2% (41) indicated occasional participation (1–4 times), while only 3.6% (9) reported having attended five or more times. An overview of participation by the different topics can be seen in Table 1.

Bielefeld’s demographic profile is broadly typical of a German large-city setting, with a mean age of 43 years, a slight female majority (51.2%), and high electoral participation in the 2025 federal election (81.90%). Consistent with this positioning, Bielefeld is only marginally younger than the German benchmark for cities (cf. Statistische Ämter des Bundes und der Länder, 2026).

Data preparation and analysis was conducted using RStudio (Posit Team, 2025). To evaluate bias in dropout between t1 and t2, Little’s MCAR test (Little, 1988) using the naniar package (Tierney and Cook, 2023) was applied to the four political efficacy items at both time points. Owing to sample size, we did not conduct structural equation modeling to confirm factorial structure; however, the conceptual separation of internal and external PE was supported by high within-facet and non-significant between-facet correlations (Appendix B). To assess whether there was considerable variance between the different THM events, intraclass correlation coefficients (ICCs; Lüdecke et al., 2021) were calculated from null models. Changes from t1 to t2 were tested with paired t-tests. To examine moderation by participant role, we estimated linear multilevel models with random intercepts (per person), maximum likelihood estimation, and Satterthwaite-adjusted degrees of freedom, including time, role, and the interaction of time and role as predictors. Adding gender, age, and education as predictors into the models did not change interpretation of our findings; models without covariates are reported in Appendix C.

Participants reported higher internal (M = 4.07, SD = 0.76) than external PE (M = 2.92, SD = 0.65), Δ = 1.15, 95% CI [0.96, 1.33], t(101) = 12.36, p < 0.001. The two facets of PE were not significantly correlated, r = 0.12, t(100) = 1.22, p = 0.224. Male participants reported higher internal PE (M = 4.25, SD = 0.73) than female participants (M = 3.87, SD = 0.76), t(99.8) = 2.59, p = 0.011. No significant differences between the two facets of PE emerged by education or age group.

The MCAR test indicated data were not missing completely at random, Χ²(12) = 12.70, p = 0.394. ICC values regarding waves were well below the 0.10 threshold, which is why multilevel model effects were nested only in individuals, not in THM events. Model assumption checks are reported in Appendix D. In view of the brevity of the measurement instrument and potentially divergent effects of individual items, we also looked at the differences between individual items by participant role after t1 and t2. The effects described here did not depend on individual items (Appendix E).

As expected, external PE was significantly higher at t2 than at t1, Δ = 0.17, 95% CI [0.05, 0.29], t(103) = 2.71, p = 0.008, supporting Hypothesis 1a. The linear mixed model on external PE showed that the main effect of time, b = 0.22, SE = 0.05, t(105.32) = 4.01, p < 0.001, was qualified by a time × role interaction, b = −0.25, SE = 0.09, t(104.73) = −2.91, p = 0.004 with active participation as reference category (see Table 2 for the fixed effects of the multilevel models). As predicted, external PE increased significantly from t1 to t2 among active participants, while no such increase was observed among passive listeners, confirming Hypothesis 1b (Table 3 contains an overview of means and standard errors for internal and external PE by group).

Internal PE did not differ significantly between t1 and t2, Δ = 0.09, 95% CI [−0.01, 0.20], t(103) = 1.74, p = 0.084. The multilevel model showed no significant time × role interaction on internal PE (b = 0.05, SE = 0.08, t(104.73) = 0.72, p = 0.472). Passive listeners reported higher internal PE than active participants overall but change over time did not interact with role. Therefore, Hypotheses 2a and 2b were rejected (see Figure 1).

This study examined whether participation in THMs fosters PE, distinguishing between internal and external PE. Our findings provide ambivalent results. First, external PE increased significantly following participation, but only among active participants. Second, internal PE remained stable across time, regardless of participation role, though passive participants reported higher levels of internal PE. While these findings contribute to our understanding of the democratic consequences of participatory events such as THMs on an empirical level, they also raise theoretical considerations. The increase in external PE aligns with expectations derived from deliberative democracy theory: THMs are designed to create accessible arenas where citizens can directly address policymakers, reducing institutional distance and reinforcing perceptions of responsiveness (Habermas, 2015; Lukensmeyer and Brigham, 2002). Our results suggest that THMs can indeed strengthen citizens’ beliefs that political actors and, by extension, institutions listen to and value citizens’ inputs. This was exclusively true for active participants, whose opportunities to submit questions and engage in dialogue with policy makers may have heightened their sense of external PE. The asymmetry between the two comparison groups underscores the importance of direct interaction: deliberation appears to enhance external PE more effectively when citizens do not observe but are able to directly contribute. By contrast, internal PE proved resistant to short-term shifts. This result resonates with prior research conceptualising internal efficacy as a stable self-perception of competence, closely tied to long-term factors such as education or income (Craig et al., 1990; Oser et al., 2022; Oser et al., 2023). Being aware of or attending a single event—even in an active role—may not suffice to affect individuals’ broader confidence in their ability to understand and discuss political issues. Notably, passive listeners reported higher internal efficacy than active participants, possibly reflecting a (self-)selection process: as those who attended as passive listeners were not randomly drawn from the population register and have attended THMs for intrinsic interest, they may have already felt highly competent to follow political discussions. This assumption is also supported by the observation that passive listeners who were retained at t2 were higher in internal PE at t1 compared to those who dropped out after t1.

The findings presented carry implications for research and practice. Theoretically, they suggest that external and internal PE may be differentially responsive to participatory interventions. Whereas external efficacy may be more sensitive to positive experiences with policy makers and institutions, internal PE appears more stably anchored individually. This resonates with arguments treating the two dimensions as analytically separate rather than assuming parallel dynamics in their formation and possible changes. From a practical perspective, the results highlight potential and limits of THMs. By fostering external PE, THMs may enhance perceptions of legitimacy and trust in institutions, outcomes critical for democratic participation (Nabatchi and Leighninger, 2015). As democracies face increasing distrust in institutions and decreasing external PE (Bienstman et al., 2024; Foa et al., 2020), political actors may turn to THM-style formats. Yet THMs’ limited impact on internal efficacy suggests that they cannot substitute for broader civic education and capacity-building efforts. Some work, therefore, highlights the role of meeting style when it comes to participatory events (Collins, 2021; Nabatchi and Leighninger, 2015). If the aim is to help citizens feel both heard and capable, THMs should be paired with efforts that strengthen political literacy, deliberative skills, and ongoing opportunities for engagement.

Several limitations of the study merit attention. Although random sampling of the active group mitigated many selection concerns, questions persist about who opts to take part in THM events and studies—especially among passive listeners. Kim (2015) similarly found that in the United States, external PE was more strongly tied to face-to-face civic and political activity, whereas internal PE was more closely linked to online participation. This pattern suggests that different forms of PE align with distinct dispositions and engagement styles. The increase in external efficacy among active THM participants may therefore reflect not only the event’s influence but also characteristics of individuals already inclined toward active involvement, while the stable levels among passive listeners may relate to their stronger internal efficacy. In this view, the rise in external PE among active participants likely stems from an interaction between the THM’s effects and participants’ underlying predispositions. To circumvent these caveats, in line with a true experimental logic, a more robust approach would involve drawing participants from the population register and subsequently randomly assigning them to active and passive roles, something that for practical constraints was not feasible in the present study. Moreover, our study captured relatively immediate changes only; future research should examine whether effects of THM participation on PE persist over time. Finally, the findings derive from a specific local context and focus on locally relevant issues which might have influenced PE of citizens on a general level (McDonnell, 2020). However, such effects should be limited to the local actors and issues involved in the discussion. Participants in this study may have primarily had the respective THM actors in mind when responding to the external PE scales, rather than national representatives or the federal government. Future cross-national or cross-level (e.g., local, regional, national) comparisons would thus be valuable to assess the broader relevance of THM formats across different political contexts and levels.

THMs may offer meaningful opportunities to strengthen citizens’ perceptions of institutional responsiveness, particularly for those directly engaged in dialogue and on a local level. A comprehensive strategy for civic participation should therefore combine events involving active participation with educational and capacity-building measures to empower citizens to believe in their voice and at the same time trust that it will be heard.