Citizen Science Design in Practice: A Taxonomic Study of Configurations, Clusters, and Disciplinary Influences in German Initiatives

Introduction

Over the past decade, research on citizen science has increasingly demonstrated that the field is shaped not only by a growing number of projects, but by the diverse ways in which citizen involvement is conceptualized and implemented. Rather than following a single logic of public participation in research, citizen science encompasses a wide range of approaches that differ in collaboration structures, roles, and institutional settings (Kullenberg and Kasperowski 2016). Several scholars have emphasized that these differences reflect both disciplinary traditions and varying epistemic, historical, and organizational lineages within the field, and therefore require an analytical perspective that moves beyond simple typological distinctions (Strasser et al. 2019; Wiggins and Wilbanks 2019; Weingart and Meyer 2021). This diversity poses a central challenge for the research community, as existing participation frameworks, such as those proposed by Haklay (2013) and Shirk et al. (2012), primarily distinguish citizen science projects by the level of participant involvement. These frameworks capture an important dimension, namely the degree of decision-making power, but the practical design of citizen science projects extends far beyond this single axis.

A range of additional project features, including communication, tasks, collaboration, tools, incentives, and dissemination of outcomes, also shape how citizen science is experienced and enacted. These design decisions, which are usually taken by project initiators, coordinators, or institutions, largely reflect a top-down perspective on project organization. While some design features are shaped by methodological constraints, many reflect explicit choices made by project initiators. Making these choices visible is essential to understanding variation across projects.

A growing body of empirical work has begun to map parts of this heterogeneity. Prior studies highlight that citizen science varies not only in participation levels but also in its epistemic and organizational foundations (Strasser et al. 2019). Further research shows that cultural, institutional, and disciplinary contexts shape project design (Hecker et al. 2018; Golumbic 2024; Weingart and Meyer 2021). However, systematic knowledge about the specific design decisions made during project implementation remains limited. This gap is particularly relevant given ongoing discussions on ethics, inclusivity, and quality criteria in citizen science (Heigl et al. 2020; ECSA 2015), which emphasize transparency, communication, and feedback but whose implementation in practice remains unclear.

Communication and feedback are also central to broader debates on the democratization of knowledge production (Irwin 1995; Weinhardt et al. 2024). Understanding how projects structure interaction and engagement is therefore crucial, particularly given the disciplinary diversity of citizen science. Different research domains may require more standardized or data-intensive approaches, while others allow for more collaborative or deliberative designs.

Against this background, a taxonomy-based perspective offers a promising approach for systematically examining how citizen science projects are designed in practice. The taxonomy for participatory processes developed by Stein et al. (2023a, c) provides a structured framework for capturing key design dimensions, including communication, collaboration, incentives, tasks, moderation, and outcomes. Although originally developed for participatory processes in other domains, the taxonomy contains elements directly applicable to citizen science and allows for a broader analytical perspective beyond participation levels alone.

We apply this taxonomy to a sample of 60 citizen science projects listed on Germany’s national citizen science platform “mit:forschen!” (formerly Bürger schaffen Wissen). The sample is based on the full population of registered initiatives (n = 240), resulting in a response rate of approximately 25%. While not fully representative, this response rate is substantial given the voluntary and decentralized nature of many citizen science initiatives. Germany constitutes one of the most active and institutionally developed citizen science ecosystems in Western Europe, with long-standing funding structures, a national support infrastructure, and substantial disciplinary diversity. The platform represents the largest curated and transdisciplinary registry of citizen science projects in the German-speaking region. Using this platform allows us to capture a broad and heterogeneous landscape of projects across disciplines and organizational contexts. This makes the platform a suitable empirical basis for an exploratory examination of design configurations in citizen science. While the empirical analysis focuses on Germany, the research questions are conceptual and relevant beyond this context. Replication in other countries remains necessary. Building on this foundation, the study pursues the following research questions:

Theoretical Background: A Taxonomy for Digital Involvement Projects

In science, politics, and economics alike, the inclusion of citizens in formerly exclusive processes and structures is gaining increasing relevance. In the digital realm, multiple participation formats coexist. While the academic field discusses this involvement as citizen science (Haklay et al. 2021), research on political participation examines the phenomenon of e-participation (Macintosh 2004; Sanford and Rose 2007), and in the private sector context, crowd-X processes are a present practice (Howe 2006; Estellés-Arolas, Navarro-Giner, and Guevara 2015). Despite disciplinary differences, these projects share similar design and implementation challenges. To systematically classify such projects across domains, Stein et al. (2023a, b, c) introduce the concept of “Digital Involvement Projects” (DIP) and develop a taxonomy of their key characteristics. DIP refers to projects that utilize digital platforms to involve external participants in structured processes (Stein et al. 2023a). The taxonomy comprises 19 design dimensions derived through an iterative process of theory and practice. The dimensions encompass three classifications determining the level of participation within a project, five dimensions delineating the project’s implementation process, two dimensions concerning incentives for both participants and organizers, four dimensions depicting communication structures within the projects, three dimensions addressing project stakeholders, and two dimensions focusing on the projects’ outcomes (see also Table 1) (Stein et al. 2023a).

During the taxonomy’s development, several established citizen science frameworks have been used to construct the dimensions (Stein et al. 2023a, b, c). The typologies and principles of Shirk et al. (2012), Haklay (2013), ECSA (2015), and Levy and Germonprez (2017) provoked the inclusion of design characteristics for the dimensions participatory degree (see Table 1, D1), participatory structure (D6), reasons for the participatory design and incentives (D9, 10), communication direction (D11), project stakeholders (15, 16), and project outcomes and publicity (D18, 19), enabling a structured link into the citizen science literature. Other dimensions and characteristics have been based either on the crowdsourcing and e-participation literature (e.g., Table 1, D3), empirical project examples stemming from all three fields (e.g., D2), or feedback from participation practitioners (e.g., D4), potentially going beyond classical citizen science frameworks.

While the taxonomy overlaps with existing citizen science frameworks, its openness enables a broader characterization of project design. Early frameworks (e.g., Shirk et al. 2012; Haklay 2013) differentiate projects primarily by participation level. However, scholarly attention in the field of citizen science has recently shifted away from rigid definitions toward a more open and inclusive approach. The ECSA has outlined ten principles of good citizen science practices, offering general guidance to practitioners regardless of specific project types (ECSA 2015). Haklay et al. (2021) similarly emphasize the plurality of citizen science and caution against rigid definitions. By adopting the general perspective of DIP that extends to approaches such as e-participation or crowdsourcing, this open and inclusive approach to citizen science practices can be ensured.

Furthermore, by focusing on the configuration of individual design choices, the taxonomy shifts the focus from abstract definitions to their tangible, empirical implementations. While prior work has focused on definitions and recommendations, less attention has been paid to making design decisions transparent. Spasiano et al. (2021) endeavor to develop a transdisciplinary framework encompassing citizen science practices, methods, and associated issues. However, their analysis primarily draws upon reviewing articles from academic journals. Similarly, Stein et al. (2023b) conducted a design review of 16 multi-project citizen science platforms, yet their emphasis remained on platform features rather than on their use in initiatives. In addition, Golumbic (2024) and Hecker et al. (2018) offer valuable empirical insights into project goals and participation models. Golumbic highlights how tensions between scientific and societal goals manifest in project design, while Hecker et al. identify dominant patterns of contributory and collaborative engagement across European initiatives. Yet, both stop short of providing a systematic, taxonomy-based comparison of design characteristics across multiple dimensions—a gap this study seeks to address.

Conversely, scholars such as Monzón Alavarado et al. (2020) and Moczek et al. (2021) have undertaken research analyzing real-life projects and their design choices. These studies highlight participation patterns but cover only a limited set of design dimensions. However, several other design choices remain unexplored due to limitations imposed by the availability of public information on the projects. Similarly, Moczek et al. (2021) analyzed 79 German citizen science initiatives concerning their funding, project roles and staffing, levels of participation, task organization, diversity, and inclusiveness. By engaging directly with projects on the “mit:forschen!” platform, Moczek et al. (2021) leveraged insights from project initiators regarding their projects’ design and volunteers’ involvement. However, key aspects of implementation and their relation to participation levels remain underexplored.

In summary, existing work only partially captures the diversity of design decisions in citizen science. The DIP taxonomy offers a structured approach to deepen insights into project design decisions. It focuses on making individual design choices tangible, including both traditional citizen science focus areas (e.g., the degree of participation), as well as less-explored design criteria, some drawn from adjacent participatory fields. Furthermore, the taxonomy supports the classification of citizen science projects within the broader field of participatory practices.

Methodological Approach

To address our research questions, we use the DIP taxonomy (Stein et al. 2023a) to classify citizen science projects in Germany. Taxonomies present a suitable approach to structurally describe a target domain by focusing on similarities and differences (Nickerson, Varshney, and Muntermann 2013). Existing taxonomies focus on epistemic and organizational aspects rather than design configurations (Strasser et al. 2019). As no design-oriented taxonomy exists, we adopt the broader participatory-process perspective of the DIP taxonomy. Although digital citizen science formats have become increasingly common, many projects continue to operate in predominantly analog settings. Within the DIP taxonomy, the dimensions “Format” and “Implementation” (see Table 1) are specifically oriented toward the digital domain. We extend the taxonomy by adding the characteristics “analog” and “none” to capture non-digital projects. “Design dimensions” refer to the 19 categories defined by the DIP taxonomy. Each dimension contains a set of “design characteristics”, that is, the selectable attributes that describe a project’s design choices. “Design decisions” refer to the concrete selections made by each project within these characteristics. We apply this terminology consistently to avoid ambiguity.

We collaborated with the citizen science platform “mit:forschen!” to obtain a project sample, which has served as an anchor for citizen research before (e.g., Moczek et al. (2021)). “mit:forschen!” (online since 2014) is the central citizen science platform in Germany. To include a project on the platform, registration is required. As part of the registration process, the project is reviewed by two independent experts and assessed to ensure it meets the platform’s criteria. The following criteria guide this evaluation:

• Adherence to citizen science principles: Projects must pursue a scientific research question or aim to establish scientific infrastructure.

• Active citizen involvement: Citizens should be actively engaged in significant phases of the research process, such as developing research questions, collecting data, analyzing results, or disseminating findings.

• Transparency: The roles and contributions of all participants must be clearly and openly communicated.

• Accessibility and support: Projects must have a dedicated German-language website and appoint a contact person to address inquiries and assist participants.

• Regional and linguistic focus: Projects should be conducted in German and have a clear regional focus within Germany.

• Non-commercial orientation: Projects primarily serving commercial purposes are excluded.

The platform started with 14 projects; at the time of the survey, 240 projects were registered on the platform.

Drawing on the experiences of Monzón Alvarado et al. (2020) and Moczek et al. (2021), we opt for a direct engagement with project initiators to accurately characterize projects in line with the taxonomy’s dimensions. To circumvent subjective and ambiguous classifications that may arise with this approach (Stein et al. 2023a), we developed a web application that provides detailed descriptions of the taxonomy’s characteristics and facilitates the application process (Stein et al. 2025). The web application guides the user through the taxonomy’s dimensions and allows selection of one or more characteristics for each dimension to adequately describe their project (see also Figure 1). Finally, it saves the final configuration for research if the user initiates it. Between June and August 2023, we contacted all projects (n = 240) and received 66 responses. After removing duplicates, 60 unique projects remained. We also assigned each project to a primary disciplinary domain. Overall, we distinguished between the five domains “Humanities, social and cultural sciences,” “Engineering and planning sciences,” “Mathematics and computer science,” “Medicine and health sciences,” “Natural sciences,” and a remaining category of “Others.”

Figure 1

For each project classification, a binary data set was considered (i.e., 0 or 1 for each of the taxonomy’s characteristics). Due to the limited sample size, we restricted the analysis to descriptive statistics and suitable clustering methods to identify preliminary design patterns between the projects. Given the exploratory nature of the study and the relatively small sample size, we report descriptive proportions rather than confidence intervals around the estimates. We applied hierarchical, agglomerative clustering following guidance by Kassambara (2017). Agglomerative clustering supports similarity evaluation through a tree-based representation of objects (Romesburg 2004; Kassambara 2017). Due to the binary nature of the data, object distance was measured based on the Yule similarity (Choi, Cha, and Tappert 2010), which supports clustering based on overall profiles rather than magnitudes (Kassambara 2017). To assess the resulting clusters, we first used the silhouette coefficient (Range: –1 (poor) to 1 (optimal) (Rousseeuw 1987; Kassambara 2017), then qualitatively interpreted the clusters with two domain experts. With this approach, we followed previous applications of the DIP taxonomy (Stein et al. 2025). We tested domain differences using chi-squared tests.

The study used only project-level data provided voluntarily by project representatives. According to the ethical guidelines of our institution, the study therefore does not fall under the category of human-subjects research requiring formal ethics committee approval. Participation was voluntary and based on informed consent.

Results

In the following, we first report descriptive results by dimension before providing insights into the clustering results and the associated differences across research domains. An overview of the descriptive results is presented in Table 2.

Cluster analysis

The initial cluster analysis yielded a two-cluster solution with an average silhouette coefficient of 0.36, whereas all alternative solutions yielded substantially lower silhouette values. This indicated that a two-cluster structure provided the best balance between internal cohesion and separation. The resulting configuration consisted of a larger cluster of 42 projects (Cluster 1) and a smaller cluster of 18 projects (Cluster 2). We conducted a sensitivity analysis excluding low-variance dimensions (<0.10). In the sensitivity analysis, we excluded design characteristics with extremely low variance across projects, as such variables contribute little to distance calculations and may distort cluster solutions. The exclusion applied only to individual design characteristics (variables), not to projects themselves. All 60 projects remained in the dataset throughout the analysis. This step follows recommendations to remove low-variance variables in cluster analysis to avoid disproportionate influence on cluster structure (Dalmaijer et al. 2022). Across all applied clustering methods, the two-cluster solution again performed best, and the silhouette coefficient increased compared with the initial analysis, suggesting improved discriminatory power.

To assess potential collinearity among design characteristics, we examined pairwise associations among binary variables using phi coefficients, which are equivalent to Pearson correlations for binary variables (Dormann et al. 2013). No strong correlations were observed that would indicate problematic redundancy between variables. Consequently, all remaining design characteristics were retained for the cluster analysis.

Taken together, both analyses consistently point to a moderate two-cluster structure. In the following, we describe the characteristics of these clusters based on their average choices across the dimensions of the taxonomy, with a visualization provided in Figure 2.

Figure 2

Cluster 1

Projects in Cluster 1 (C1) tend to exhibit a lower degree of participation, as reflected in high shares of informative and consultative projects as well as single or low-effort tasks (see Figure 2). Implementation is predominantly digital and individually structured, with all low-time-requirement projects falling into C1. A representative example is ZOWIAC, in which citizens report sightings of invasive species such as raccoons and American mink via app or website; verified data feed into distribution maps supporting conservation and health monitoring—a clearly bounded, impact-incentivized contribution with no community interaction required. More broadly, volunteer incentivization in C1 centers on impact, while participatory design choices are more frequently driven by access/resource and acceptance/legitimation considerations. Communication structures are varied but generally less intensive and community-oriented than in C2: nearly all projects with single/two-sided communication and external or no moderation are found in C1, and expert feedback prevails over crowd feedback. Stakeholder configurations reflect an open yet expert-centric character—most projects are initiated and led by experts, with broadly open target groups—and outcomes are predominantly made publicly accessible.

Cluster 2

Projects in C2 demonstrate a high degree of participation, with the majority exhibiting a consultative or democratic character and offering multiple, effort-based tasks to participants (see Figure 2). Implementation is primarily collaborative, often combining digital and analog formats with team-based, synchronous work structures; time requirements range from high to self-selected. The „Digital Active Women“ project illustrates this profile well: Recently migrated women act as co-researchers, evaluating digital services and co-creating recommendations through workshops and surveys, with findings fed back to local authorities and counseling centers to improve outreach—a design that is collaborative, mixed-method, and aimed at social participation rather than data collection alone. Compared with C1, incentive structures are more self-related, and funding- and value-based rationales for participatory design are more frequently cited. Communication is pronounced and community-oriented: nearly all projects facilitate multi-sided communication with crowd and expert feedback and moderation, and the majority actively support community building. Stakeholder configurations are more varied than in C1—experts typically act as initiators and drivers, but equal partnerships also occur—and target groups range from open to restricted or closed. Outcomes similarly show variability; relative to C1, decisions as outputs are more common and outcomes are more frequently access-restricted.

Project domains

Figure 2 visualizes the sample’s distribution across thematic domains (left) and relates it to cluster affiliation (right). It shows all taxonomy characteristics for both clusters (C1 in blue, C2 in orange), with bars indicating the share of projects exhibiting each characteristic. The six panels represent the dimensions Degree of Participation, Implementation, Incentives, Communication, Stakeholders, and Outcomes, enabling direct comparison of characteristic frequencies between clusters.

Figure 3 shows the distribution of project domains across the two clusters, indicating distinct disciplinary patterns. Overall, a dominance of the domains “Humanities, social and cultural science” and “Natural sciences” is apparent, enriched by a few instances of medicine- and engineering-related projects in the sample. Investigating relations between project domain and cluster affiliation, we find that C1 primarily consists of projects in natural sciences (0.48) and humanities, social, and cultural sciences (0.36). At the same time, C2 is primarily composed of projects in humanities, social, and cultural sciences (0.50) and from medicine and health sciences (0.28). A Fisher’s Exact Test indicates a significant association between domain and cluster affiliation (p < 0.01). Given the exploratory nature of the analysis and the limited sample size, p-values were interpreted cautiously. These domain–cluster patterns persisted when the analysis was repeated with the reduced variable set, suggesting that disciplinary differences are not driven by over- or underrepresented design features.

Figure 3

Discussion and Conclusion

This study characterizes the design of 60 German citizen science projects using a taxonomy-based approach. Applying the DIP taxonomy reveals broad design diversity across the sample, yet several characteristics are absent or rare (<10%): low-effort participation types, profit-maximization incentives, one-sided communication, external moderation, equal citizen–researcher partnerships, crowd-initiated and crowd-driven structures, closed target groups, decision outcomes, and non-public outcomes. Conversely, a small set of characteristics appears in nearly all projects (≥ 90%): organization/expert-initiated and -driven structures, knowledge outcomes, and public outcomes.

Comparing our results with Stein et al. (2025), which covers the broader field of participatory practices, reveals both shared and divergent patterns. Several characteristics underrepresented in our sample are similarly rare across the broader field, including profit-maximization incentives, external moderation, equal partnerships, crowd initiation, and closed target groups, suggesting these reflect general participatory constraints rather than citizen science–specific tendencies. Other patterns are more distinctive. The near-universal prevalence of knowledge and public outcomes in citizen science reflects the ECSA principles’ institutional embedding of open knowledge production, whereas the broader participatory field shows a more diversified outcome profile (e.g., 30% product, 52% knowledge, 24% decision, 24% sharing in Stein et al. 2025). One-sided communication is similarly more common in broader participatory practice (41% in Stein et al. 2025) than in citizen science, consistent with the ECSA principles’ emphasis on active involvement and feedback.

That said, a non-trivial share of projects in our sample report no explicit feedback mechanisms, raising questions about the fulfillment of ECSA principle 5. Several factors may explain this: large-scale projects may rely on aggregated or delayed formats rather than individualized responses; respondents may interpret “feedback” narrowly; and sustained feedback mechanisms require resources that may be deprioritized under funding constraints. Given the self-reported nature of the data, variation in how taxonomy categories were understood may further influence reporting. Notably, expert feedback predominates over crowd feedback in citizen science, whereas the reverse holds in the broader participatory field (22% expert versus 43% crowd feedback in Stein et al. 2025). The scarcity of profit- and funding-based incentives likely reflects the dominance of nonprofit academic institutions, though such incentives are rare across participatory approaches more generally (13% funding, 7% profit in Stein et al. 2025). The absence of closed target groups may partly reflect our sampling via “mit:forschen!”, as projects with closed target groups likely benefit less from listing on a public meta-platform.

Several characteristics explicitly emphasized in citizen science typologies—equal partnerships and crowd initiation or leadership—remain rare in practice. Compared with the broader participatory sample (78% institutional drivers, 89% institutional owners in Stein et al. 2025), institutional leadership is similarly prevalent in citizen science, pointing to a tension between normative ideals and structural realities. The curated nature of “mit:forschen!” may further amplify this by overrepresenting institutionally anchored projects.

Citizen science projects nonetheless show considerably more heterogeneity in participation formats, implementation, and moderation than the broader participatory field, where projects tend to be predominantly digital (96%) and largely unmoderated (91%) (Stein et al. 2025). This internal heterogeneity—and its structured co-variation with other design dimensions—is what our clustering analysis helps to illuminate.

Our clustering analysis identifies two loose design clusters distinguishing projects with lower from those with higher participation. The silhouette coefficient (0.36) indicates moderate separation, suggesting indicative rather than sharply distinct patterns. While the centrality of participation degree is well established in citizen science typologies (Haklay 2013; Shirk et al. 2012), the clustering solution extends these typologies by empirically demonstrating systematic co-variation between participation degree and further design dimensions—including communication structures, implementation formats, and task scope—that prior frameworks had anticipated theoretically but not jointly examined. We label the clusters “crowdsourcing research design” (C1) and “participatory research design” (C2). C1 aligns with “Crowdsourcing” or “Volunteer Sensing” (Haklay 2013) and “contributory projects” (Shirk et al. 2012); C2 resembles “Participatory Science” or “Extreme Citizen Science” (Haklay 2013) and “collaborative” or “co-created projects” (Shirk et al. 2012).

The two configurations reflect practical trade-offs rooted in differing research requirements. C1 projects typically address questions that rely on large-scale, standardized observations—distribution mapping, environmental monitoring, species identification—for which scalable infrastructure, clearly specified tasks, and asymmetric feedback arrangements are functional necessities rather than mere design preferences. C2 projects, by contrast, address exploratory or socially embedded questions requiring collaborative problem framing and joint interpretation, making deliberative and co-creative formats methodologically coherent, albeit more resource- and coordination-intensive. The observed cluster differences can thus be read as adaptations to distinct epistemic requirements within citizen science, alongside organizational trade-offs between scale and depth of engagement.

These design configurations also reflect institutional constraints (e.g., funding structures, coordination capacity, organizational resources) rather than purely deliberate choices. The significant association we find between cluster affiliation and disciplinary domain is best understood not as a direct disciplinary effect, but as reflecting how disciplinary traditions shape typical research questions and validation practices: Disciplines vary in their reliance on large-scale standardized data collection versus interpretive and qualitative co-analysis, and these tendencies increase the likelihood of adopting C1- or C2-type designs within particular domains. Future research with larger samples and qualitative follow-up is needed to disentangle how disciplinary norms, problem types, and structural constraints interact in shaping citizen science design.

Finally, the two clusters may reflect broader epistemic orientations in contemporary science. C2 resonates with traditions emphasizing collaboration, reflexivity, and participatory knowledge practices, including Mode 2 science (Nowotny et al. 2001), post-normal science (Funtowicz and Ravetz 1993), and co-production (Jasanoff 2004), while C1 aligns more closely with standardized and centralized research formats. Since our data do not directly measure epistemic orientations, future qualitative research would be well positioned to examine these commitments more explicitly and to trace how they become operationalized in citizen science project design.

Conclusion

With this work, we aim to stimulate discussion among practitioners and researchers on the design of citizen science projects, encouraging openness to diverse approaches within and beyond the field. While defining desirable standards and principles remains important, it is equally crucial to quantify the status quo to identify challenges in practice. Our work provides a foundation for systematically assessing design practices in Germany, and supports critical reflection on the current distribution of approaches. It raises key questions: Is this the design landscape we aspire to? Which practices should be adopted more widely—or avoided? What changes in infrastructure, guidance, or capacity-building are needed to support such decisions?

Beyond offering a descriptive overview, our study contributes an empirically grounded extension of traditional citizen science typologies by mapping additional design dimensions and identifying two recurring configurations, while emphasizing that these configurations organize heterogeneity around two dominant patterns without eliminating meaningful variation across design dimensions. However, rather than reinforcing binary classifications, we aim to foster an empirically driven dialogue around key questions: Are certain design configurations better suited to achieving specific citizen science project goals? To what extent do factors such as disciplinary origin shape these choices, and should they?

Concluding, we hope this work encourages reflection and debate, not only about how citizen science is currently designed, but how it might evolve in response to both internal ambitions and external inspirations.

Data Accessibility Statement

The dataset contains project-level information that may allow the identification of individual initiatives listed on the “mit:forschen!” platform. To protect the privacy of organizations and project teams, the full dataset cannot be made publicly available. An anonymized and documentation-based version of the dataset is available from the authors upon reasonable request.

Author Contributions

Carolin Stein – Conceptualization, Methodology, Data Collection, Data Analysis, Writing – Original Draft, Writing – Review & Editing, Moritz Müller – Conceptualization, Methodology, Data Collection, Writing – Original Draft, Writing – Review & Editing, Jonas Fegert – Conceptualization, Methodology, Data Analysis, Writing – Review & Editing, Project Administration, Submission and Revision Management.

The authors used AI-assisted tools (Claude and Grammarly) for language refinement and copyediting of the manuscript. All content was reviewed and approved by the authors.