Citizen social science in practice: the case of the Empty Houses Project

The significant development of citizen science over the last five years is notable, seen in increasing professionalisation and institutionalisation, recognition at policy levels, and the launch of a specialised journal. Debates resist narrow definitions because of the field’s heterogeneity and encourage collaboration among actors to tailor designs to context. CSS extends these debates into the social sciences, drawing on participatory research traditions that engage citizens across the research process. It raises questions about who can collect and analyse data and how it can be used. CSS is still emerging, often framed as crowdsourced data collection enabled by technological advancements, and presents both opportunities for collecting otherwise unavailable data and challenges to established research practices. This article addresses the paucity of practical examples by empirically examining the Empty Houses Project—set up to explore crowdsourcing social data on empty houses to contribute to a policy issue. The project comprised a pilot, an awareness campaign, and a data collection window, followed by stakeholder walking interviews and policy/practitioner interviews to unpack reporting challenges, opportunities, and potential data use. The article contributes by problematising crowdsourcing as mere data gap-filling, arguing that CSS entails epistemological and political implications. It reviews relevant literature, describes the case and methods (including walking interviews), and analyses opportunities (reflection, new data, shared responsibilities) and challenges (data quality, meaningful participation and ethics, data use). It discusses how CSS reconfigures roles and responsibilities and raises questions about knowledge construction.

CSS emerges from citizen science and participatory social science traditions, notably participatory action research (PAR) and co-production. PAR emphasises community capacity, reflexivity, and challenges to expert power, positioning it as an orientation to inquiry rather than a single method. Co-production similarly interrogates the boundaries of knowledge production and has roots in planning and public services. Foundational citizen science work (Irwin) links science to citizens’ concerns and contextual knowledges, aligning with public sociology and the interrelation of facts and values. In the literature, CSS is described as enabling public assistance in research and large-scale recording of beliefs and observations, often pragmatically aimed at scaling human effort for data analysis (e.g., social media datasets). Some propose CSS as new methodological and theoretical territory that makes citizens co-learners able to transform research and policy systems. CSS should be considered within renewed attention to the politics and social life of methods, where methodological choices raise fundamental questions about knowledge limits and relationships between cultural, social, and material. CSS thus opens methods to public involvement and more socially engaged practice, connecting private troubles with public issues.

The Empty Houses Project was designed in stages to prototype a crowdsourced CSS approach addressing empty homes. Initial stages: (a) a two-week pilot with five randomly selected participants to refine instructions and categories; (b) an awareness campaign; and (c) a data collection window. Instructions and submission interface were hosted on a WordPress blog. The Google Form collected: property address and postcode; property type; indicators suggesting vacancy; estimated duration of vacancy; whether it had been reported to the local authority; and any other relevant information. Promotion used Twitter, Facebook, university mailing lists and contacts, printed flyers in public spaces, blog posts (e.g., Policy@Manchester, Big Issue North), and local TV appearances (That’s Manchester TV). The project was open for three months, then extended by one month; 20 responses were received. Data were collated via Google Forms and stored on a password-protected hard drive. Due to low submissions, two further stages were added: (d) eight walking interviews with prior submitters, housing project participants, and housing activists to explore reporting barriers and ethical considerations; and (e) nine semi-structured interviews with housing practitioners and policy officials (local authority teams, campaign groups, networks, student union adviser, etc.) to understand potential uses of data and broader contextual issues. Walking interviews: participants provided informed written consent; ethics approval was obtained from the University of Manchester. Interviews were semi-structured, participant-led routes in their neighbourhoods, typically one hour, audio-recorded (Zoom H4), transcribed, and used to discuss CSS, ethics, and reporting challenges. Practitioner/policy interviews were mostly phone/Skype (extensive notes taken; face-to-face interviews recorded and transcribed). Analysis followed thematic analysis (Braun & Clarke, 2006), with transcripts thoroughly read, memos written, and open/selective coding applied. Themes were iteratively refined, triangulating with field notes, observations, and contextual data, keeping data and emergent concepts in dialogic relationship.

- Participation as reflection and meaning making: Walking interviewees reported that trying to spot empty houses made them re-engage with and see their local environment anew, fostering reflection and generating situated, meaningful knowledge about place. Participation could engender ‘active research subjects’ and distinctive agency, especially when moving from solitary noticing to structured, collective activity. - Opportunities to produce new data: Local authority and practitioner interviewees indicated citizen-generated data could supplement official statistics, add nuance, and help target efforts to bring properties back into use—particularly valuable amid resource constraints. Local authority teams recalled past capacity for ward surveys but highlighted current budget cuts. Citizen data was viewed as a potential resource to offset limited survey capacity. - New responsibilities and power dynamics: CSS redistributes roles across data collection and analysis. Participants argued that citizens should also analyse data, not only collect it, critiquing top-down models where professionals retain analytical control. CSS can scaffold bottom-up social science that values everyday enquiry, but tensions arise between applying frameworks (which can ‘professionalise’ participation) and allowing participant control. - Data quality and robustness challenges: Open-ended tasking led to ambiguity over criteria for identifying ‘empty houses’ and highlighted tensions between positivist and constructivist approaches. Participants emphasised subjectivity, the need for clear standards or verification, and careful design that builds capacity rather than tests it. The organisation and positioning of citizens in CSS critically affect data quality perceptions and practices. - Meaningfulness of participation and ethics: Participants questioned practical benefits, shared end goals, and whether participation risks unpaid labour for experts. Some saw participation as civic duty; others worried about exploitation or lack of meaningful engagement beyond ‘donkeywork’. Embedding data collection into daily routines was debated; some preferred purposeful walking over online microtasks, while others stressed the need for committed practice rather than passive habit. - Concerns over how data will be used: Motivation hinged on clarity about data purpose, users, and potential consequences. Interviewees raised risks of misuse, preferred uses such as trend analysis over time, and suggested community-led data collection to hold authorities to account versus simply handing over data. - Project scale and outputs: The project received 20 public submissions during the collection window; it conducted eight walking interviews and nine policy/practitioner interviews. Contextual statistics underline the policy salience: in England on 1 Oct 2018 there were 216,186 long-term vacant dwellings (0.9% of stock) and 634,453 total vacant dwellings (2.6% of stock).

The study set out to explore how CSS operates in practice when mobilising citizens to identify empty houses. Findings show that CSS is not merely a mechanism for building data repositories; rather, it is an epistemic and political practice that prompts reflection, co-learning, and renegotiation of roles between researchers, practitioners, and citizens. The project revealed that participation can cultivate new ways of seeing and generate situated knowledges, addressing the research question by demonstrating both the affordances (reflection, supplemental data, civic engagement) and the challenges (data quality, ethics of labour, meaningfulness, and data governance). The significance lies in showing that design choices—task framing, analytical involvement, transparency about data use—shape participant motivation and data robustness. CSS can support policy by enriching official datasets in resource-constrained settings, but only if ethical concerns are addressed, verification strategies are in place, and participants have clarity and, where possible, influence in analysis and application. Ultimately, CSS in this context functions as a dialogic, ‘working it out together’ process that reconfigures knowledge production rather than a simple crowdsourcing pipeline.

The study demonstrates that citizen-generated data is complex, personal, and not easily used to plug data gaps, yet it is unique and otherwise uncollected. CSS creates opportunities for public reflection on social life and can challenge top-down data practices, scaffolding towards bottom-up social science that values everyday enquiry and situated knowledges. In policy contexts, CSS is attractive as a strategic tool for data generation amid constrained resources, but this raises questions about who benefits and the emancipatory claims of participatory methods. CSS adds to the social science methodological repertoire at a time dominated by ‘bigger, faster’ data, aligning with ‘technologies of humility’ that acknowledge ambiguity and partiality. CSS prompts critical questions about the nature of data, who can collect and analyse it, and how it should be used. These questions are valuable outcomes in themselves and should not be ignored. Future work should develop designs that build participant capacity, include citizens in analysis, implement clear data verification and governance, and ensure transparency about data use to support both policy relevance and ethical practice.

- Limited public submissions (20) during the collection window constrained the scale of the dataset and prompted methodological adaptations (walking and policy/practitioner interviews). - Ambiguity in task interpretation affected consistency and perceived data quality; verification procedures were not fully established within the project’s scope. - Practitioner/policy interviews conducted via phone/Skype were not recorded (extensive notes taken), potentially limiting verbatim analytical depth compared to transcribed audio. - Context-specific focus (Greater Manchester, UK) may limit generalisability; findings are situated and contingent on local policy, resources, and housing dynamics. - Participant concerns about data use and ethics indicate potential participation bias, where only those comfortable with the project’s aims and governance engaged.