Developing the Citizen Summit Method: Understanding citizens’ views on digital surveillance technologies

Introduction

Citizen summits are growing in popularity in policymaking. In their original form policymakers used this participatory forum to poll opinions and engage voters in deliberating matters of political and social importance. Summits have proven effective in engaging publics on issues such as climate change (Jun & Bryer, 2017; Bedsted et al., 2011), immigration policy, health care, environmental issues and the economy (Badger & Quealy, 2019), and have contributed to democratic reform (Fung, 2015). In urban planning, they have addressed the democratic deficit, ensured diverse public input into policy matters and supported the development of citizens’ civic skills (Durrant & Cohen, 2023; Grönlund et al., 2022; Beauvais & Warren, 2019). Yet, in their original form, the quality of the scientific evidence citizen summits produce has been criticised (Macnaghten et al., 2005; Sturgis, 2014; Sturgis & Allum, 2004).

This article sets out several methodological enhancements to the original citizen summit format that retain its deliberative benefits while enabling higher-quality scientific evidence to be gathered. These enhancements rely on careful theoretical integration of interdisciplinary perspectives on the phenomena being investigated. This new Citizen Summit Method (CSM) adopts a robust mixed methods approach which can enable interdisciplinary and inter-sectoral research teams to improve their understanding of public responses to policy interventions. The new CSM was developed during a large project in which citizens across nine countries in the European Union (EU) deliberated about the acceptability of using digital surveillance technologies (DSTs) to protect national security. Each event enjoyed a high level of local political support which enhanced their legitimacy.

The new CSM has huge potential for application in internet policy, where citizens are affected by many complex, high-risk policy and regulatory challenges. While the research reported in this article addresses surveillance, a topic within the remit of internet policy (Hintz & Dencik, 2016; Gstrein, 2020), there are many other areas to consider. Examples include online safety (Neudert, 2023), digital inclusion (Carmi &Yates, 2020; Quyoum et al., 2025), and ensuring free and fair democratic elections (Bennett & Lyon, 2019; Bakir, 2021). The privacy and human rights risks for citizens in these policy areas demand high-quality evidence concerning their impact on democratic rights and processes. Given the diverse multi-stakeholder nature and distributed decision-making structure of internet governance (Radu et al., 2024), deliberative research that garners citizens’ views can create a robust representative evidence base to drive effective policymaking.

The article begins by describing the study context for developing the method, then it describes the original citizen summit approach before explaining the enhancements made by the new CSM. Finally, the CSM’s ability to support informed debate and generate policy recommendations, as well as its limitations and the need for further development, are considered.

1. The study context

The study that developed the CSM gathered citizens’ views on how national security and human rights can be protected in a digitised world. Terror attacks across Europe have led to the legal sanctioning and procurement of DSTs, such as smart Closed Circuit Television (CCTV), geolocation tracking, and enhanced communications surveillance as solutions to homeland and international security challenges. These measures polarise opinion among the public, who are expected to trade off their privacy in exchange for promised increased security (European Agenda on Security, 2015).

All DSTs present security benefits and privacy risks (Siegrist & Cvetkovich 2000), yet the idea of a security–privacy trade-off rests on an incentivistic logic which assumes citizens will accept a DST if they believe the security benefits outweigh the privacy risks (Acquisti et al., 2015). This acontextual argument has been criticised for presenting privacy and security as abstract categories rather than as enacted social practices emerging from the interaction between people and their social and institutional contexts (Pavone & Esposti, 2012; Dourish & Anderson, 2006). Prior empirical work has also failed to challenge this trade-off assumption (Strickland & Hunt, 2005) and revealed the need for the complex issues underlying privacy concerns and public scepticism towards DSTs to be better understood (Ball et al., 2019).

In practice, citizens have little control over whether their data are gathered for national security purposes and the changing institutional context is likely to influence their attitudes toward DSTs (Acquisti et al., 2016; Nissenbaum, 2009). This study captured these underlying dynamics to reveal whether European publics really were ‘trading off’ privacy for security or had other priorities. The CSM was developed to examine citizens’ views on these issues, internationally at scale, by studying the acceptability of DSTs and attitudes towards the security agencies that deploy them (Solove, 2011).

2. Citizen summits: origins of the Citizen Summit Method

Citizen summits originated in the United States’ local government (Moynihan, 2003). In 1999, the new mayor of Washington D.C. worked with the Office of Neighborhood Action and AmericaSpeaks to create a new model of public participation that could contribute to district strategic planning. The 3,000 citizens who attended reviewed a draft strategic plan prepared by the Office of Neighborhood Action. Grouped on tables of ten, they spent over seven hours discussing city-wide priorities, using a networked laptop computer and wireless polling keypads to vote on questions posed at different points in the summit. Trained facilitators helped to promote dialogue and keep the conversations focused. The voting results were immediately displayed on large screens at the front of the room. Following the discussion, each group agreed a message to share through the networked laptop, to which the mayor was able to respond during the summit.

This original citizen summit approach is able to mobilise democratic participation, raise awareness about issues of public importance, engage citizens in open debate to make recommendations directly to policymakers, and facilitate immediate feedback from both policymakers and fellow citizens. These are the hallmarks of deliberative research, in which there is capacity to share and weigh up different arguments, before reaching a decision and making policy recommendations (Leshner, 2003; Wilsdon & Willis, 2004).

Despite their popular use in policy settings, several criticisms of the original summit method require it to be redesigned for use in academic or inter-sectoral settings, where policymakers and academics collaborate in deliberative research. These concerns include the extent to which the original method conforms to social scientific methodological principles and questions about the quality of the gathered evidence (Macnaghten et al., 2005; Sturgis, 2014; Sturgis & Allum, 2004). Citizen summits primarily gather quantitative polling data with simple ‘yes/no’ responses rather than validated scales. The qualitative reasons behind these views are also typically not captured or fed into policy processes (Macnaghten et al., 2005). Participatory theorists have also expressed concern about the extent to which summits are truly participatory. The predesigned formats, procedures, and questions used appear more consultative in nature (Vaughn & Jacquez, 2020). Wider criticisms of deliberative research also apply, including questions about summits’ level of reflexivity (European Commission, 2008); whether theoretical and empirical questions are fully addressed (Jasanoff, 2005; Wynne, 2006); if social empowerment is an inevitable consequence (Harmon et al., 2013); who and how the involved publics are constituted (Felt & Fochler, 2010), and who decides what will be discussed (Stilgoe et al., 2014).

3. Using the CSM to study digital surveillance technologies

Using citizen summits to study public opinion robustly on the polarising issue of DSTs required several elements to be added to the original format. First, research questions, a theoretical model and key variables were identified by theoretically integrating the results of an interdisciplinary literature review. These formal insights underpinned and integrated all aspects of the research design and materials. Second, the CSM’s methodological components were redesigned for use in a deliberative plenary setting, while remaining scientifically robust. Established scales rather than the polling-style ‘yes/no’ questions used in the original summits were used. Third, unlike in the original summit approach, qualitative data were formally gathered to understand citizens’ reasoning on themes linked to the key variables identified through theoretical integration. These elements are set out in more detail below.

3.1 Integrating the methodological components

Citizen summits have four integrated methodological components (see Table 1): (i) information materials shared in advance to familiarise citizens with the issues being discussed; (ii) video vignettes used as stimulus material to engage citizens in the debate: (iii) the use of citizen table groups, supported by a table facilitator, to debate and make policy recommendations; and (iv) an attitude survey administered via an audience response system, using polling keypads to instantly capture and relay the data back to citizens, promoting vitality and participation to the proceedings.

In the new CSM these traditional summit components were deconstructed and re-integrated using theoretical integration logic (Poth & Shannon-Baker, 2022). A detailed theoretical model was developed to guide the overall research design and provide crucial connections linking each of the methods (see Figure 1). This theoretical integration took place during the study design and piloting stages to ensure each methodological component supported the others. Common constructs which connected the different methodological components were first identified and then embedded within each component (Poth & Shannon-Baker, 2022; Alavi et al., 2018; Tunarosa & Glyn, 2017).

The interdisciplinary research team whose expertise spanned Science and Technology Studies (STS), Surveillance Studies, Law and Management Studies, first conducted a wide-ranging literature review to identify relevant phenomena of interest which might represent, influence or intervene in the security-privacy trade off. This review integrated constructs and dimensions from across the team’s disciplinary interests, as shown in table 2. A further literature review then identified the consequences of the phenomena under investigation and the relationships between them were mapped (Fisher et al., 2021a; Fisher et al., 2021b). For example, those with higher general concerns about security threats and positive general views on DSTs could be more likely to believe DSTs were effective and hence more likely to accept them. The resulting theoretical model, shown in Figure 1, featured the public acceptability of DSTs as the dependent variable (see Table 2), which refers to the extent to which citizens approve of the DST and believe it should be routinely implemented (Sanquist et al., 2008).

The theoretical model explored how the public acceptability of DSTs was influenced by a range of constructs, including: general familiarity with, and attitudes towards, technology; familiarity/attitudes towards different DSTs; perceived intrusiveness/effectiveness of these technologies; perceived temporal, spatial and social proximity of risks associated with their use; perceived trustworthiness of public bodies operating DSTs; and citizens’ substantive privacy concerns. Variables concerning the technologies’ perceived effectiveness and intrusiveness represented the different sides of the security–privacy trade-off. As these were attitudinal variables, they were not suitable for objectively measuring ‘amounts’ of security and privacy to be ‘traded-off’; rather, they represented the balance of citizens’ views on these issues. The subsequent analysis examined whether there was a direct relationship between perceived effectiveness and intrusiveness with the public acceptability of the different DSTs, or whether other variables were influential.

Understanding and observing these different constructs required their integration into the different data collection methods used in the summits, with relevant theoretical constructs becoming part of the attitude survey component of the method. Then, it was ensured that these core constructs featured in the table group debates and information materials (see Table 4). The video vignettes and information material were designed to describe these core constructs and set up the matters for debate in the table groups. All material was translated into each participating country’s national language¹.

During this re-integration of the summit components, care was taken to foreground citizens’ voices in the data gathering, while producing internationally comparable, mutually interdependent and individually robust quantitative and qualitative datasets. The questionnaire and table group discussion designs were adapted for this purpose, as were the information materials and video vignettes. These adaptations are described next.

3.2 Developing the questionnaire for a deliberative plenary setting

Applying the methods in plenary rather than a one-to-one setting had implications for the interaction between researchers and participants, the questions posed and how the data were gathered. Both quantitative and qualitative elements of the data collection were modified to ensure citizens’ ease of participation and that the methods were theoretically mutually interdependent.

In the original citizen summits, the survey questions are displayed on a central screen, responses are recorded on keypads and immediate feedback is given on the same screen. This approach shows how participants’ opinions are formed at various points in the event and is crucial to the participatory experience. However, these questions mainly seek simple ‘yes’, ‘no’ or ‘don’t know’ answers, partly due to display limitations, but also because traditional summits poll opinions rather than test theoretical models using scaled questions.

For a more theoretically informed approach, the length and style of the survey questions were modified. To avoid participant fatigue, the attitude survey was limited to around one hundred individual questions, split into shorter sections. Each question and its accompanying Likert scale needed to be clearly visible on the central display screen and easy to read aloud. This means those involving lengthy sentences, using complex or reverse scales, or those requiring trade-offs can be incomprehensible.

The questionnaire development followed an iterative process, involving multiple testing and piloting stages (Gomez et al., 2014). As the original scales were shown to be too complex to use at the summits, the research team rigorously refined these measures. Where possible, to ensure content and construct validity, they consulted with the original authors of the scales. Beginning with the existing scales three researchers iteratively reduced the number of scale items and refined the wording. Further testing was undertaken of the content and wording with other academic and security experts from the research team. The questionnaire, booklet and films were also piloted with members of the public. This rigorous testing of the questions, survey instrument and other materials across nine EU countries, underpins the method’s reliability. Table 3 illustrates the process for the trustworthiness questions and shows the original literature scales and the final shortened versions.

3.3 Developing qualitative data collection in the table groups

Turning to the qualitative component, the table group discussions foregrounded citizen voices, open debate, and recommendations as independent qualitative data sources. They also added depth and context to the attitude survey findings. In the original summit method, the table group discussions enabled participants to form opinions and then vote in the plenary poll. For the new CSM, these discussions were modified according to focus group methodology, which can also be readily combined with other methods (Morgan, 1997). The ability of focus groups to explore “processes of attitude formation and the mechanisms involved in interrogating and modifying views” (Barbour, 2008, p. 32) was salient in revealing the reasoning behind the views in the questionnaire and the recommendations.

The trained facilitators ensured different views were exchanged by the diverse table groups, moderating the discussion to ensure consensus in the shared recommendations that were reached (Pinto da Costa, 2021). Although the facilitators did not ask prescribed questions, if the conversation faltered, they used a series of prompts reflecting the constructs in the video vignettes, information material and questionnaire (see Table 4). At the start of each discussion round and before the facilitator opened the debate, each participant completed a template about their views of the benefits and limitations of each DST.

Qualitative data sources included notes taken by table facilitators and dedicated note-takers who captured the argument flow and wrote down memorable quotes. The facilitators also shared written reflections after the event with the research team. These qualitative data were joined with other qualitative data sources generated in the discussions. Participants were encouraged to note their individual thoughts on postcards and ‘post’ them into a post-box located in the room. At the end of the discussion round, each table group produced a reasoned policy recommendation on a pre-prepared template, which also featured in the project report to the European Commission.

Each qualitative component was transcribed for thematic analysis in NVivo. These data were then coded and analysed to identify recurrent themes and explore the main arguments used (Marshall & Rossman, 2011). Although the amount of background noise precluded verbatim recording of the discussions, this challenge could be addressed by future technological advancement.

3.4 Information materials

The information materials informed participants about the technical features of the DSTs and considered current and potential privacy and security controversies associated with their use. The material was initially written in plain English and was subsequently translated as required by the different study countries. It included an overview of surveillance, privacy, and security informed by academic, legal, and journalistic resources and was illustrated with real-world examples. The purpose, function, benefits, and limitations of each DST in relation to these security and privacy implications were also considered as a way of stimulating debate.

The document progressed through four rounds of internal review to ensure an engaging format, easily digestible information, and well-balanced arguments. One round of external review involved the project’s advisory board, with three rounds of piloting with citizens. The content was proof-read and edited by a professional magazine editor, before being converted into a magazine by graphic designers. The booklet³, which was well-illustrated and prepared to the highest production standards, was mailed to participants before the summits with a letter encouraging them to peruse the content.

3.5 Short films (video vignettes)

The research team collaborated on producing the video vignettes, which integrated the same constructs as the survey and the information materials. Video vignettes are appropriate to use in settings which involve challenging situations, extreme emotions, or questions of ethics or human rights (Caro et al., 2012) and using films favourably compares to similar phenomena when experienced first-hand (Eifler, 2007). An iterative approach driven by theory was taken to the production process. Following Johnson (2000), constructs from the research model underpinning the questionnaire were used as the basis for the film-making.

Experts from industry, regulatory bodies, security bodies, campaign organisations, and academia were interviewed on camera about the uses, benefits, and limitations of each DST to create twelve one-hour interviews. Three seven-minute films were produced, each focusing on one of the DSTs. Content analysis of the interviews was used to distill source material for each film. The films were documentary in style, used high production values and were created by a small company with extensive experience of producing television for the BBC. Five rounds of editing of the raw interviews took place, reducing one hour and 45 minutes of material to seven minutes for each film. Images were then added and copyright clearances obtained. Piloting the films⁴revealed the need for different national conventions for subtitling or voice overs. Some countries (e.g. Germany, Denmark) accepted subtitles in their respective national languages but in others (e.g. Italy), voice overs had to be recorded. The organisation of summits using the new CSM approach is discussed next, focusing on the specific case of the UK summit.

4. Running the summits

Twelve summits were organised in nine EU countries in North-West Europe (Denmark, Norway, and the UK), Central Europe (Austria, Germany, and Switzerland), and Southern Europe (Hungary, Italy, and Spain). In each country, participants’ views were sought on two DSTs from a group of three: Smart Closed-Circuit Television (CCTV), Smartphone Location Tracking (SLT), and Deep Packet Inspection (DPI). To ensure equal coverage of all DSTs across the project, the nine countries were grouped into three clusters using Hofstede’s cross-cultural framework (Hofstede, 2001) as a guide (see also Grimmelikhuijsen et al., 2013). As Table 5 shows, the DSTs were then considered in at least two of the three countries in each cluster⁵.

3.1 Organisation of the UK summits

The UK summits focused on Smart CCTV and DPI. Since the early 1990s, the UK has used video-based CCTV systems extensively in public spaces, with a further wave of expansion from 2010 of digital or smart CCTV, where a camera collects digital images and matches them to known images in a database. DPI is used by companies, intelligence services, and governments to read the content of communications sent via the internet (Wehner, 2013). At the time of the study, controversies had emerged over these technologies’ use, including in the wake of the Snowden revelations, when internet or surveillance facilitated by DPI led to concerns about privacy.

The summits took place in the UK’s second largest city of Birmingham in central England, which is the administrative headquarters of the West Midland’s metropolitan county and a major manufacturing, engineering, commercial, and service centre. The city is served by an international airport and has a strong cultural and educational tradition, with four universities located there. The multicultural diversity of its citizens, which is highly representative of ethnically rich Britain, made Birmingham a suitable location. The city has also been at the heart of debate about DSTs, with a public outcry following the dense installation of Automatic Number Plate Recognition (ANPR) cameras in a predominantly Muslim area of the city in 2010 (Isakjee & Allen, 2013).

3.2 Event sampling and recruitment

The two UK citizen summits took place in a city-centre hotel, with citizens recruited by a market research company. Some of those sampled were approached in a busy shopping mall in central Birmingham, others were drawn from the company’s contact database or recruited by phone. The sample was strictly representative of the national demographic profiles in terms of age, gender, ethnicity, educational attainment, and occupation (Cresswell, 2013). This composition is designed to reflect the diverse views of the UK population, rather than of any specific group. Although not a representative random population sample, this approach supports a balanced and inclusive brainstorming exercise. The information magazine outlining the issues for discussion was sent to participants, along with the timetable, location, travel, and expenses details. Each participant was offered a £100 shopping voucher as an incentive for participation.

214 participants took part in the UK summits: 105 at the first event, and 109 at the second. 205 (97%) were British citizens. 25% of participants claimed to belong to a minority ethnic group, although some did not reveal their ethnicity. British society is highly multicultural and ethnic groups of South-East Asian or Black African origin account for 14% of the total population of England and Wales. 53% of participants were men and 47% were women; 39% had vocational qualifications, with 17% being university graduates. 47% were employed as managers or professionals, with 37% working in clerical, services, sales or technician roles. The remainder in paid work were employed in craft or trade roles, such as agricultural, fisheries or forestry workers, or were plant or machine operators or assemblers. 42% reported earnings well below the average annual British salary, although this included retirees. Of the 13% who claimed to earn well over the average, 37% were university postgraduates, 35% were managers or senior officials, and 35% were professionals.

3.3 Registration, timetable, and responsibilities

The scale, length, and complexity of the summits required careful management, with formal training sessions prior to the event with briefing packs outlining each role and responsibility shared within the teams. Each summit required a team of circa 40 staff members, comprising three technical support roles, one head facilitator, one head prompter, one introductory speaker, two floor managers, circa 14 table facilitators and 14 note takers, and three handling registration.

The day-long event was organised into several segments, during which citizens viewed the short films, discussed the content in their table group, answered questions using the voting system, and then listened to feedback in plenary (see timetable in Table 6). Participants were registered, given a unique registration code and issued with a remote-control keypad for use with the audience response system. They were allocated to table groups of eight or nine participants, mixed in terms of age, gender, education, and occupational level. Participants were encouraged to spend a few minutes introducing themselves and chatting informally with others at their table, promoting an informal atmosphere and helping to kick-start subsequent discussion.

During the introduction session, the head facilitator described the summit purpose and reviewed the practical arrangements. A short speech by a popular local media broadcaster helped to motivate those attending. Attendance by local or national policymakers or politicians lends legitimacy to the summit. The first voting session followed, in which participants answered questions about their views on security, privacy, and the use of security technologies. These responses benchmarked participants’ initial views against which later views could be gauged.

Two 90-minute sessions relating to each of the two DSTs followed. Each began by screening the short film, followed by a table group discussion. The discussion was deliberately free-flowing, and table moderators intervened only to move the conversation forward when needed, or to encourage those who were more reluctant to join in. Participants then used their clickers to answer questions about the DST featured in the film. After each question, the Head Facilitator shared the voting results via a central screen. To foster shared engagement and reinforce the value of citizens’ views, some table groups shared their discussions with the rest of the room. The research team created a short film about the summits, which shows how they were administered, the experiences of the participants and the high level of political support they received.⁶

Discussion

The effectiveness of the CSM re-design is demonstrated in three ways: first, by the explanatory power of the mutually interdependent qualitative and quantitative datasets; second by the extent to which the redesigned summits engaged participants with DSTs and the issues raised; and finally, through the generation of constructive recommendations for policy.

The new CSM generated nuanced quantitative data which could be subject to more sophisticated multivariate analysis. Accordingly, the relationships between key variables and citizens’ reasoning behind it were connected and made visible, recorded, and analysed scientifically. In robustly exploring these relationships and gathering rich qualitative insights, the heterogeneity of different citizen voices and views were more clearly revealed (Ball et al., 2018; Esposti et al., 2021).

Although detailed discussion is beyond the scope of this article, headline findings from the UK summits and snapshots of the quantitative and qualitative data are reported below. These data reveal different heterogeneous responses in how surveillance is experienced. Current research, for example, suggests that ethnic minorities often feel heavily scrutinised and at greater risk of digital privacy harms, while others view DSTs as protective (Quyoum et. al., 2025). The mixed-method design captures these contrasting perspectives, while quantile regression of the survey data moves beyond ‘average views’. Overall it was found that institutional trustworthiness – rather than the privacy or security aspects of DSTs – crucially shapes whether citizens oppose or support digital surveillance. These relationships hold internationally: European citizens were not directly trading off privacy for security when deliberating about the acceptability of DSTs (Esposti et al., 2021).

The mutual interdependence between the qualitative and quantitative datasets underpinned the effectiveness of the theoretical integration of these mixed methods early in the research design process. The findings associated with Institutional Trustworthiness illustrate this point. The quantitative data gathered via the audience response system showed that 30% agreed that security agencies that use SCCTV are trustworthy, 31% considered them competent, 45% thought they were concerned about citizen welfare and national security, but only 16% considered these agencies did not abuse their power. The figures were similar for DPI, with 31% believing the agencies using this technology are trustworthy, 29% considered them competent, 41% thought they were concerned about citizen welfare and national security, and 12% believed they did not abuse their power. Table 7 summarises the UK summit’s Institutional Trustworthiness results.

The qualitative data gave richer insights into the factors which underpinned these citizen views, with one individual who commented that, “Someone needs to tell us that it [the data] is really being used for security” (Male, older); while another argued it “needs to be made clearer to users who and what is looking at their activities etc. and for what reason” (Postcard 10). In the same vein, one table group called for: “the government and security forces to be more open with statistics showing how DPI has benefited us. How many interceptions have taken place?” Another group complained about the “… overall lack of communication by all agencies involved [and] lack of consistency between agencies and their procedures and poor response to incidents”. In the words of one participant: “…who are they [security agencies] and who do they answer to?” (Male, older). The vast majority considered that access to this information should only be granted to security agencies or certain government departments, because: “We need guarantees on who has our information, and we need accountability” (Female, younger). Most were opposed to the involvement of private sector organisations in national security matters and did not want such firms to operate DSTs or manage the data, as this recommendation stated: “Personal security has been compromised due to privatization of security services”. Many wanted to avoid the information being exploited for other purposes, as this citizen emphasised: “Private companies should not be involved in operating DSTs or have access to the information/data that is produced”.

Turning now to participant engagement, while the UK participants expressed broad support for DSTs at the end of the summits, they had also scrutinised privacy issues in a new light. 90% believed DSTs help enhance national security. 79% said they had gained new insight into the topics discussed, with 62% believing that the summit generated valuable insights for politicians. Just under half said their attitudes towards DSTs had changed during the summit; with 30% becoming more negative and 15% more positive. These changing views were reflected in citizens’ comments about privacy and how these technologies could affect it. While fears about public privacy rose from 46% to 65%, concerns about personal privacy increased from 35% to 69%. Worries about the use of DPI were greater than about smart CCTV, perhaps because of the ubiquitous and familiar use of CCTV on the UK’s streets. 77% of citizens considered Smart CCTV to be an effective security tool, 80% believed it improved national security, with 68% considering the level of intrusiveness associated with the technology were justified. DPI attracted greater criticism, with concerns that it might spy on everyone, irrespective of whether they were doing anything wrong. In the words of one citizen: “DPI should only be used if you have been charged with a crime not just to look at what your habits are” (Postcard 14). Table 8 summarises the UK summit’s citizen engagement results.

The summit recommendations also reflected these concerns. Despite having high expectations of national security, British citizens are unwilling to forfeit their personal privacy. This nervousness about DSTs’ privacy implications, which included concerns about who might access the data collected, how it might be used, and with what implications, reinforces the paradoxes of the security–privacy trade off. Participants were keen to see more governmental transparency about how information was used and called for greater openness about the benefits and risks of DSTs. They also referenced the need for better regulation and oversight of how DSTs are used. One table group recommended:

“The use of DSTs should be governed by transparent and easy to understand legislation. In order to ensure accountability an independent regulatory body should be established that has responsibility for overseeing the use of DSTs, and which sets rules about handling the gathering of information/data.”

Another recommendation was to: “form a publicly elected independent either national or worldwide body who monitors and controls the security agencies. Report findings to the public so that we can see who is accessing our data and what they are using it for”. Speaking specifically about Smart CCTV, one citizen made the following request: “Do not let CCTV get too advanced so that we end up in 1984, big brother is watching you!” (Postcard 71). Taken together, these findings show the CSM’s capacity to capture quantitative data on citizens’ views, in combination with qualitative insights into the thinking behind them.

Conclusion

This article has reported the methodological developments undertaken to create a robust CSM approach to study citizens’ views on DSTs. The adaptations respond to concerns about the limitations of the original summit method (Macnaghten et al., 2005) and the dilemmas associated with deliberative methods and participation experiments (Roberts, 2004). Interdisciplinarity was central to the redesign, which involved the theoretical integration of interdisciplinary constructs concerning the acceptability of DSTs. Augmenting the original method with a theoretically derived model ensured the CSM’s four methodological elements cross-referenced each other and could be modified, tested, and applied in deliberative plenary settings. By uncovering the factors shaping citizens’ views the data clarified citizens’ concerns beyond the security–privacy trade-off (Pavone et al., 2018). Recommendations were delivered to the European Commission in a series of country reports and an aggregated report which featured statistics and qualitative data analysis (SurPRISE 2015).

The article provides a template for those seeking to pursue high quality deliberative policy research involving the public. We recommend wide ranging cross-disciplinary literature reviews to identify variables which may explain differences in public attitudes towards policy matters and generate deliberation about them. Inclusive dialogue about what those variables might be, involving both researchers and policymakers, should focus on the purpose of the research rather than on disciplinary boundaries. The selected variables and the relationship between them then need to be modelled and incorporated into the CSM’s main elements: attitude survey, table group discussions, information materials, and short films. Exhaustive testing for the plenary deployment of data collection instruments is crucial. Subsequent analyses will then be able to explore heterogeneity in responses across citizen groups and the reasons for it.

While this article focused on the acceptability of DSTs and featured some technologies which have since been superseded, the CSM can be deployed where public deliberation of internet policy matters is warranted. The variables explored have relevance when considering internet policy areas including online safety, free and fair elections, and the digital divide as well as digital surveillance. In all of these cases, institutional trustworthiness, the impacts on security or implications for privacy or other human rights matters are relevant. Other variables can be added. Drawing on insights from urban planning and addressing the democratic deficit, the CSM method could more avowedly incorporate policymakers not only into the events themselves but also in research design and implementation as co-creative partners within intersectoral research teams. High level support from policymakers contributed to the success of the research described in this paper.

The CSM still has room to develop into an inclusive and deliberative tool which may generate robust evidence about some of the most complex issues of our times. The overt use of a theoretical model to integrate and mix the methods still risks marginalising the inductive element in data collection and interpretation, even though the table group element is more free form. In this way the CSM still tends towards the more consultative form of participatory research in that its conceptual boundaries are preset and not defined by the participants (Vaughn & Jacquez, 2020). To overcome this limitation, a further recommendation includes using the CSM as a component in longitudinal participatory research in which participants determine an agenda to be scaled up or implement their results in a wider participative policy process.