Voting in online surveys on open government policies in Moldova and Ukraine

Introduction

Governments introduce e-participation formats, including internet voting (i-voting) in consultations, referenda, or elections to reduce budgetary expenses, modernise public administration, increase voter turnout at elections, or elevate public trust. Usually, they execute this within a broader electoral reform, digitalisation policy, or open government initiative. In this context, of special interest is the Open Government Partnership (OGP). The initiative unites 75 countries (including 28 national members on the European continent) and 106 local governments (including 35 local members in Europe) and thousands of civil society organisations for promoting open government as a government more accessible, responsive and accountable to citizens and for improving the relationship between people and their government (Open Government Partnership, n.d.-a; Open Government Partnership, n.d.-b; Open Government Partnership, n.d.-c). OGP encourages its member countries to create national action plans (NAPs) for new policies in a dialogue between authorities and the public (the co-creation process), especially using digital technologies. But of all e-participation and online consultation formats digitally powered voting in online surveys for open government policy priorities is studied the least.

Voting using digital means is important because it can bring not only more evident practical, technical and financial benefits but also more subtle yet profound socio-political effects. The very introduction of popular i-voting for public policy can empower voters and organised civil society, make authorities more open to input from the public and strengthen open government overall. These effects could be even more profound if there is a public-government discussion about the i-voting format. As stakeholders engage, deliberate and co-create a policy, they may build higher trust toward each other. Finally, the i-voting campaign can develop the skills and habits of direct democracy and shape the practice of more innovative, participatory and accountable governance. Adversely, i-voting can harm the participatory process by distorting results and undermining legitimacy. I-voting hardware or software systems and their cybersecurity may fail; inaccurate voter registers and misidentification may prevent people from voting or allow hacked voting; certain voter groups may not be able to vote because of the digital divide. Mass media discourse may frame a voting issue, microtargeting ads may manipulate public opinion; forced vote disclosure, group pressure, vote coercion and buying can malform voting results; corrupt vote storage and counting can contort voting outcomes. Overall, these issues may discredit the whole i-voting campaign.

As the available studies of the impact of voting in online surveys are limited spatially, temporally, dimensionally, or methodologically and lack the link to open government, more research is required. Thereby, this inquiry aims to assess the impact of voting in online surveys for open government policy priorities – on voters, civil society organisations, authorities and open government (viewed as collaborative policy-making between authorities and the public) overall. Of all 28 European countries participating in OGP (roughly evenly distributed across the continent; see the full list at Open Government Partnership, n.d.-c), only two (Moldova and Ukraine) used online surveys for choosing policy priorities for co-creating NAPs within the OGP process. Both countries have been members of OGP since 2011 and introduced online surveys for NAP co-creation in 2016. This makes Moldova and Ukraine perfect subjects for a comparative study of the socio-political impact (in relation to change in attitudes, practices, or policies of target groups) of online surveys on open government.

To examine the socio-political impact of voting in online surveys on open government, this inquiry raises four research questions about Moldova and Ukraine: (1) What is the socio-political impact of open government co-creation online surveys on voters? (2) What is the socio-political impact of open government co-creation online surveys on civil society organisations? (3) What is the socio-political impact of open government co-creation online surveys on government authorities? (4) What is the socio-political impact of open government co-creation online surveys on open government as a system? To perform international comparative research, we analysed open government online surveys and their socio-political impact in Moldova and Ukraine as exploratory case studies using a mixed methods approach to data collection and analysis. We examined applied reports assessing open government co-creation and policy implementation, open government-related documentation, social media posts by national MSFs, digital tools for online consultations and opinions of public and government stakeholders via online semi-structured expert interviews.

Core concepts

As a universal technology, voting using digital means can facilitate numerous e-participation formats (elections, referenda, participatory budgeting (PB), consultations, surveys, collaborative drafting etc.). Here, we understand e-participation (or online participation) according to Dijk (2012, p. 12) as “taking part in public affairs in a particular phase of the institutional policy process”. Some scholars denote i-voting generically as “voting using the internet and computer technologies at least for vote casting” (Khutkyy, 2020b, p. 1). Similarly, some practitioners define i-voting inclusively as a system that “allows voters to cast their ballots from any internet-connected computer anywhere in the world” (e-Estonia, n.d.). Even linguistically a collective choice among alternative candidates or policy options is “voting”. However, considering the level of technological and organisational efforts, the degrees of user identification and cybersecurity and political stakes, the community of i-voting scholars tends to use the term “i-voting” exclusively for internet-based elections and referenda (Germann & Serdült, 2017; Krimmer et al., 2019; Solvak & Vassil, 2018; Spycher-Krivonosova, 2022). A broader concept of electronic voting (e-voting) includes, besides remote i-voting, also non-remote e-voting using electronic voting machines, usually within polling stations. With respect to such conventional usage of concepts, this study also reserves i-voting for internet-based elections and referenda and utilises different concepts for other e-participation formats that use digitally powered voting. In this paper, for linguistic clarity, we denote the act of choosing a person or a policy as “voting” (sometimes characterising it as “digitally powered” or “using digital means”). Its core attribute is remote vote casting – performed via the internet. In this paper, we focus on “online surveys” following the Council of Europe’s (2009, p. 8) definition of “e-survey” that “allows opinions to be obtained informally, by electronic means, from random or selected persons, usually in connection with a proposal and a set of possible responses”. We also view “online consultations” as synonymous with “e-consultations” – “collecting opinions of designated persons or the public at large on a specific policy issue without necessarily mandating the decision maker” (Krimmer & Kripp, 2009, p. 6). Thereby, the core characteristic of an online survey within an online consultation is its non-binding character. Online surveys enable a random or systematic sampling representative of the whole target population, yet often anyone can participate risking to generate a sampling bias (e.g., receiving unproportionately more responses from an aware, digitally savvy and socio-politically active public). Nevertheless, this instrument is useful in obtaining instant and structured policy input from the public.

Since open government (in this context – not a state, but a system of governance) is both a policy phenomenon and a subject of scientific study, the core term of open government merges both academic and practitioner perspectives. From an academic viewpoint, open government in a broad sense “designates a system of governance where governments not only selectively inform citizens or occasionally consult them about public decisions, but where they should do so and where citizens can choose what to do with government information and how to engage with public institutions” (Schnell, 2020, p. 1598). From this perspective, governments open up not only by consistently communicating with citizens and involving them in decision-making with the use of digital technologies and online tools but also by ensuring space for independent citizens and CSO activism. This accords with the applied definition of open government as “a culture of governance based on innovative and sustainable public policies and practices inspired by the principles of transparency, accountability and participation that fosters democracy and inclusive growth” (OECD, 2016, pp. 3–4). Essentially, open government means a more inclusive and collaborative (as compared to a one-way, service deliver-focused and paper-based government) approach to public policy-making between citizens and authorities, especially using digital technologies. We will use this definition in this study. The central components of open government are public transparency (supply side), civic participation (demand side) and public accountability (feedback loop between the two sides). The study views citizens (voters), civil society organisations (CSOs) and authorities collectively as stakeholders and examines the socio-political impact of online surveys on them as well as open government overall.

Investigated socio-political effects of i-voting and online surveys

As this paper explores the socio-political impact of voting in online surveys on open government, it is useful to perform a literature review of principal identified effects of i-voting and online surveys on voters, civil society organisations, government authorities and on open government as a system. Beyond the usually and solidly covered practical (such as convenience, speed of usage and procedures), technical (e.g., reliability and cybersecurity) and financial (e.g., cost of introduction and maintenance) aspects, the available research on the impact of i-voting and online surveys on open government is limited. More “soft” socio-political outcomes of the use of i-voting and online surveys are a rather rare focus of research. The most relevant ones are discussed further.

Probably the most common and contested argument in the literature is that due to reduced use costs, i-voting has the potential to mobilise voters, include some disenfranchised groups, such as out-of-country voters and thus increase voter turnout. These effects were not detected everywhere and there is no guarantee that i-voting will increase voter turnout. For example, Musiał–Karg (2012) found little or no increase in voter turnout specifically induced by i-voting neither in Estonian i-voting at elections nor in Swiss i-voting at referenda. Yet, in some cases such voter mobilisation effects are evident. Thus, an online survey of Swiss nationals abroad revealed that most of them, including those with i-voting experience, perceived i-voting as easy to use, useful, efficient and trustworthy; moreover, the ease of use correlated with willingness to utilise e-voting systems (Pleger & Mertes, 2018). A study of referenda in Geneva, Switzerland found that i-voting technology reduced avoidable voter mistakes and thereby increased the effective turnout in terms of the number of valid votes cast (Germann, 2021b). A study of Geneva residents’ voting patterns identified that i-voting increased turnout among abstainers and occasional voters (Petitpas et al., 2021). The research on i-voting in eight Swiss cantons found that in comparison to postal-only voting, i-voting increased registered expatriate voter turnout by 4.1-6.4 percentage points (Germann, 2021a). Similarly, a study of municipal elections in Ontario, Canada found that i-voting can increase turnout by 3.5 percentage points (with larger increases when a vote by mail is not yet adopted and greater use when registration is not required) (Goodman & Stokes, 2020). A survey of the Brazilian state of Rio Grande do Sul found that voting for participatory budgeting projects using digital means attracted citizens with no prior experience of contact with the government and previously disengaged from the PB process and thereby spurred an estimated increase in turnout of around 8.2 percent (Spada et al., 2016). Overall, when the mobilisation effect is manifest, it is small-scale, yet important for enfranchising and mobilising some voter groups.

Furthermore, i-voting is applicable for choosing policies or electing leadership and thereby affects political parties. Thus, some “pirate parties” applied i-voting for electing leadership (representative democracy), voting on policies (direct democracy) and delegating votes to experts (liquid democracy) with outcomes depending on the voting design and party democraticness (Khutkyy, 2019c). The Italian Five Star Movement and the Spanish Podemos party also applied i-voting in their online primaries, although their candidate selection process was mixed and demonstrated elements indicating undemocratic tendencies (Mikola, 2017). At large, i-voting is able to reveal the democratic or undemocratic character of a political party.

There is a limited number of studies on the i-voting impact on authorities. For instance, a survey of 47 electoral management bodies (EMBs) in Ontario municipalities of Canada revealed strong satisfaction with i-voting and strong support for it, citizen-centred (as opposed to possible administration-centred) rationales for adoption (accessibility, improved participation and convenience) and benefits for adoption (convenience and accessibility), although admitted some challenges (mostly related to digital literacy, outreach campaign and internet access; to a lesser extent – to security) (Goodman & Spicer 2019). Albeit, in the Netherlands’ case the public sector became so dependent on the private sector that the government lost ownership and control over both the e-voting system and the election process (Oostveen, 2010). Similarly, a multi-methods study of the impact of i-voting on Estonian election administration found that i-voting shifts and imposes new roles and responsibilities on street-level bureaucrats of elections and in extreme cases of delegation, the election administration neither owns nor rents nor understands the i-voting system (Spycher-Krivonosova, 2022). Yet, an international survey of EMBs in 72 countries found no negative impact of the introduction of e-voting technology on the independent position of EMBs (Loeber, 2020). These reports indicate the importance of i-voting design, training and implementation for shaping the experiences of EMBs and other authorities.

Finally, in some cases, due to magnifying and empowering e-engagement, i-voting and online surveys can shift the balance of power in citizen-government relationships for the benefit of citizens, promote other (including innovative) forms of civic participation in policy-making online (e-participation) and thereby influence open government as a whole. But digitally-enabled voting does not always strengthen e-participation. Thus, one qualitative case study of Estonia’s e-participation found that the e-consultation platform with voting functions Osale.ee lacked impact and failed in its adoption and outcomes (Toots et al., 2016; Toots, 2019). Another study of Ukraine’s PB found that the outcomes for e-participation integrity due to using digitally-enabled voting for choosing PB projects are disputable. For example, in Odessa, a civic activist stated that municipality officials allegedly misused passport data for fake voting (which is possible, as the voting platform shows only the number of votes, but not voters, thus lacking requisite transparency), while in Kyiv another civic activist assessed that digital voting technology shielded against third-party manipulation and miscount (which is corroborated by the virtual absence of voting criticism in social media) (Khutkyy & Avramchenko, 2019). An expert survey of municipal administrators in charge of the Decidim platform (that employs the function of voting using digital means, in particular for choosing PB projects of local development) discovered that 50% of them agreed (in contrast to 17% who disagreed and 33% who neither agreed nor disagreed) that the use of the Decidim platform gave more decision capacity to citizens (Borge et al., 2022). The introduction of PB with voting for local development projects increases voting in regular elections (People Powered, 2022): in New York City, United States, after voting in participatory budgeting, voters’ likelihood to vote in ordinary elections increased by 8.4 percentage points (Johnson et al., 2021); in Prague, Czech Republic, in districts that introduced PB (compared to districts without PB), voter turnout in local elections increased by 3 percentage points (Kukucková & Bakos, 2019). Another case study of the Wasauksing First Nation indigenous community in Ontario, Canada discovered that i-voting in referendum facilitated innovation and modernisation of community governance, improved community connectedness, self-determination and self-governance (Budd et. al, 2019). Probably, e-governance and e-participation are rather experimental processes and require several rounds to evolve and demonstrate measurable outcomes. The examples of practiced i-voting models in other municipalities in Canada and their own experimentation in Australia contribute to policy learning at the municipal scale, such as the accumulation of new information and examples over time by competing advocacy coalitions that use technical expertise and other resources in the attempt to influence public policy (Goodman & Smith, 2016). Lessons drawn from research and practice in 31 countries reveal that online surveys as part of online consultations are suitable for creative experimentation, effectively securing citizen control, protecting against arbitrariness and contributing to performance improvements (Schwab et al., 2017).

The viewed studies are often confined to a particular community or country, a single election, a concrete target group or institution, a specific facet or voting format, or one data source. Therefore, this paper assesses the impact of voting in online surveys on the open government in previously understudied countries, voting campaigns, a wider set of stakeholders, voting dimensions, formats and data sources. This enables us to discern more (in)direct links and draw more encompassing conclusions of the impact of online surveys on open government as a totality in the cases of Moldova and Ukraine.

Research methods

To inspect multiple contexts of online survey impact, we applied the international comparative research design of the Moldova and Ukraine case studies. Since the available scholarship on the impact of voting in online surveys on open government is scarce, we carried out this inquiry as exploratory research. To gather diverse and rich data and to cross-validate findings, we performed a mixed methods approach to data collection and analysis from multiple sources.

In this inquiry, we examined applied reports assessing open government co-creation and policy implementation. Our analysis covered eight Independent Reporting Mechanism (IRM) reports: three about Moldova (Mirza-Grisco, 2018; IRM staff & Mirza-Grisco, 2021; Independent Reporting Mechanism, 2022) and five about Ukraine (Presniakov, 2015; Kotliar, 2016; Khutkyy, 2018; IRM staff & Khutkyy, 2020; Independent Reporting Mechanism, 2021) designed to evaluate the OGP co-creation process and implementation. We also analysed six government self-assessment reports: three in Moldova (The Government of Moldova, 2012; The Government of Moldova, 2018; State Chancellery of Moldova, 2021) and three in Ukraine (The Government of Ukraine 2016; The Government of Ukraine 2017; The Government of Ukraine 2020). In addition, we reviewed two relevant independent reports: one about Moldova (Transparency International Moldova, 2018) and one about Ukraine (Khutkyy, 2019c). The analysis focused on the descriptions of online survey designs, decision-making procedures and online survey impact on open government policies.

Furthermore, we inspected open government-related documentation using policy analysis. Our review processed nine government decrees adopting NAPs: four in Moldova (Republic of Moldova, 2012; Republic of Moldova, 2013; Republic of Moldova, 2016; Republic of Moldova, 2018) and five in Ukraine (Kabinet Ministriv Ukrayiny, 2012; Kabinet Ministriv Ukrayiny, 2014; Kabinet Ministriv Ukrayiny, 2016; Kabinet Ministriv Ukrayiny, 2018; Kabinet Ministriv Ukrayiny, 2021). Besides, we considered six information notes: two from Moldova (Guvernul Republicii Moldova Cancelaria de Stat, 2018a; Guvernul Republicii Moldova Cancelaria de Stat, 2018b) and four from Ukraine (Hromadianske suspilstvo i vlada 2016a; Hromadianske suspilstvo i vlada 2016b; Uriadovyi portal, n.d.; Departament komunikatsii Sekretariatu KMU, 2021). We analysed the change in government policies of public transparency, civic participation and public accountability.

Additionally, we scrutinised social media posts by national MSF and linked comments using manual qualitative content analysis. Our examination covered nine co-creation-related Facebook posts (seven on the Agenția de Guvernare Electronică webpage (Agenția de Guvernare Electronică, n.d.) in Moldova and two on the OGP Ukraine webpage (OGP Ukraine, n.d.) in Ukraine). We examined government transparency, the depth of stakeholder deliberation and indications of the impact of online surveys on authorities, the public and their interaction.

We also assessed digital tools by manual qualitative and quantitative content analysis. Specifically, we inspected four digital instruments: the online consultation e-platform Particip.gov.md (Particip.gov.md, n.d.-a) in Moldova and the Verkhovna Rada Ukrayiny draft law discussion e-platform (Verkhovna Rada Ukrayiny, n.d.) in Ukraine, the online survey SurveyMonkey e-form (Centrul de Guvernare Electronica, n.d.) in the Moldovan case and Discuto e-platform (OGP Ukraine, 2018b) in the Ukrainian case. Considering that some e-platforms (especially, Discuto) allowed a wider range of e-participation beyond online survey (most importantly, e-deliberation), we audited the digital tools for the possibilities and actual scope of deliberation (peculiarly, the number of “votes” and comments), online survey vote casting and the presentation of results.

Finally, we obtained the opinions of stakeholders from the public and the government via 10 online semi-structured expert interviews. For a balanced perspective, we approached civic activists, development specialists, independent experts and government officials potentially knowledgeable about voting in online surveys during the OGP co-creation process in their countries. We identified potential informants through publications on dedicated OGP-related government websites (checking the lists of current and former members of MSFs) and via references by interviewees themselves (using the snowball technique). We approached a total of 17 most potentially aware persons (eight from Moldova and nine from Ukraine) via publicly available email addresses (e.g., on social media profiles). With a 58.8% response rate, two researchers conducted 10 interviews (five with stakeholders from Moldova and five from Ukraine) from 17 July – 29 August 2022. Considering the narrow circle of persons informed about the specific topic of the impact of online surveys on OGP priorities in the two countries (to the best of our knowledge, up to 20 persons in both countries in total, 10 in Moldova and 10 Ukraine), this is a reasonable resulting sample of gaining expert input from every second person of the entire target population. We finalised the resulting sample after reaching a saturation point of obtaining complementing, mutually validating and exhaustive data and assuming that the remaining three uncontacted potential informants who work in the same organisations have similar experiences and would provide similar answers. Respondents granted informed consent that the interviewer audio-records, transcribes and cites their answers in publications. Two Moldovan informants requested to stay confidential. On average, interviews lasted for 30 minutes. Considering the wartime context of ongoing Russia’s full-scale invasion of Ukraine, this was a reasonable interview time. We transcribed all but one (following the informant’s request) interview audio recording. The authors structured the text by three themes: the role of respondents in the co-creation process, online survey voting design and voting impact. Upon request, we shared citations with informants who verified their accuracy.

Besides the above mentioned data on OGP co-creation, we also reviewed general literature and country-specific materials on online consultations, i-voting and open government. Such literature on the digital participation process was used as a horizon-broadening optics for the analysis of both focused case studies aiming to put them in a broader country and co-creation context.

Online consultations, online surveys, i-voting and open government in Moldova

Online consultation is the most widely announced e-participation format in Moldova, rolled out through the online participation platform Particip.gov.md. According to Moldovan law, to be legally adopted by a central authority a legal act requires a public discussion (Khutkyy, 2019a). During 2016 and 2018 (the years of the analysed OGP co-creation process), the government agencies announced 637 and 1,146 public consultations of draft documents respectively (Particip.gov.md, n.d.-a). However, the utility and use of the platform is questioned by civil society members, as the platform is not sufficiently promoted nor used by civil society (Mirza-Grisco, 2018). CSOs also criticise the government for shortening the discussion period up to one day, for inviting the same government-friendly CSOs and for sending short-notice notifications about on-site public consultations. Moreover, in 2016 and 2018, the most commented draft legislation received 216 comments, the second-commented – 39 comments, while the rest – 10 comments or less, in most cases – zero, indicating a lack of public interest in this kind of e-participation (Particip.gov.md, n.d.-b). Allegedly, even CSOs don’t use the platform very often (Mirza-Grisco, 2018). Moldova’s parliament conducts online consultations too, but rather as announcements of public hearings with the possibility to provide input by email. Between 2016-2023 the Parliament listed only 13 public consultations (The Parliament of the Republic of Moldova, n.d.). Civil society is critical of parliament’s online consultations for conducting them so rarely, adopting sensitive laws very quickly on first reading and arranging public consultations only upon advocacy by civil society (Khutkyy, 2019a). Online surveys in Moldova are used occasionally. For example, in 2016 e-Government Agency held a survey aiming to map the “Revolution of Data” in Moldova, but its results were unclear (Government Open Data Portal, n.d.). Regarding i-voting, Moldova only approved the concept of an i-voting system (an alternative option to paper elections) in 2022 (Moldpres, 2022) and has not introduced it yet. In sum, the use of online consultations and online surveys in Moldova varies depending on the government branch and agency. These processes are rather formal than substantive and lack transparency and accountability.

Since joining OGP in 2011, Moldova’s open government approach has evolved unevenly. During this time, by December 2022, Moldova had created four NAPs and implemented three of them (the development of the fifth NAP was delayed) (Open Government Partnership, n.d.-d). Reportedly, public consultations for respective NAPs were held both offline and online in 2012 (The Government of Moldova, 2012) and 2013 (Caraseni, 2014). Yet, the online channel of consultations was merely for collecting inputs. According to IRM assessment reports, in 2016 (Mirza-Grisco, 2018) and 2018 (IRM staff & Mirza-Grisco, 2021), an additional method was launched – an online survey, where the public voted or prioritised OGP policy commitments in a de facto i-voting. It is noteworthy that Moldova’s government invested substantial efforts in engaging civil society in the 2012 and 2013 co-creation processes, facilitated less public participation in 2016 and encouraged more civic involvement again in 2018. Thus, the introduction of i-voting during the last two co-creation cycles paralleled the retrieval of civil society engagement by the government.

Online surveys during the 2016 OGP co-creation in Moldova

In 2016, OGP priorities in Moldova were ranked in an online survey. On 13 April 2016, Moldova’s government agency made a Facebook post calling for public input to the NAP (Agenția de Guvernare Electronică, 2018). Facebook was a relevant communication channel, because in several years (2018) it reached 990 million Moldovan users, around 36.6% of the total population (Statista Research Department, 2022). The public had 18 days to provide inputs, which was sufficient for succinct suggestions. The survey itself has been held as a SurveyMonkey e-service online questionnaire with no prior user identification (Centrul de Guvernare Electronica, n.d.). The first multiple choice question asked to choose five of 20 (including the open “other” option) priority thematic areas for the next NAP. Subsequent two open-ended questions inquired about problems and solutions in open government widening the input. Some stakeholders clearly did not label the format as i-voting arguing that it was rather mapping ideas than shortlisting (Veronica Cretu, independent expert) and clarifying that the goal was to hear people’s opinions without saying that the results will become final priorities (Stanislav Ghiletchi, civic activist). As the survey aim was collecting and prioritising ideas, non-binding for authorities, according to our conceptual framework, we classify it as an online survey (and online consultation in a broader sense).

The 2016 public outreach campaign in Moldova was multi-channel yet narrow in actual readership. According to the IRM report, an online survey was disseminated via social media, news portals and governmental websites, but this initiative was unknown beyond the usual channels, especially outside of the capital (Mirza-Grisco, 2018). The government posted the call on Facebook six times (Agenția de Guvernare Electronică, 2018). These posts collectively received 147 “likes”, one “comment” and 27 “reposts”. This showed some public interest in the issue. As a result, only 29 people participated in the survey and shared their opinions on NAP content (Mirza-Grisco, 2018). Compared to the online consultation e-participation rates found above (between zero, 39 and 216 comments to a draft document), the survey participation rate is lower than the maximum, but average for the Moldovan policy-making context, although a very low level in absolute numbers. Probably, the most informed and motivated civic activists participated.

Online surveys during the 2018 OGP co-creation in Moldova

In 2018, the online survey for Moldova OGP priorities was similar in format to the one in 2016. On 31 July 2018, the governmental webpage on Facebook announced a call for public input to the 2019–2020 NAP (Agenția de Guvernare Electronică, 2018). A sufficient 20 days were allocated for the survey. The call contained a link to an online questionnaire administered by SurveyMonkey. Judging from the survey results (Guvernul Republicii Moldova Cancelaria de Stat, 2018b), the questionnaire contained a question requesting to rank 21 open government domains in the order of priority for the next NAP. The questionnaire also had open-ended questions about problems and solutions in interacting with authorities thus allowing a deeper input. This was an online consultation in an online survey format.

The 2018 Moldova awareness-raising campaign was minimal. The original Facebook post received only five “likes” and two “shares”. Reportedly, there were also press releases on the website of the State Chancellery and the Electronic Government Agency, as well as on the web pages particip.gov.md and data.gov.md (Guvernul Republicii Moldova Cancelaria de Stat, 2018a). As a result, 41 people participated in the survey (Guvernul Republicii Moldova Cancelaria de Stat, 2018b). This is slightly more than in 2016, but still reflects the country’s typical online consultation e-participation rate of dozens of people and rather a low absolute number indicating the involvement of a narrow circle, probably of regularly engaged civic activists. The implications of this pattern are discussed further.

The impact of 2016 and 2018 online surveys on open government in Moldova

The 2016 Moldova online consultations in the form of online surveys, although non-binding, might have had some basic objective impact on the 2016–2018 open government policies. An independent researcher discovered that three areas, voted by the majority of the e-survey respondents, were included in the final NAP (Mirza-Grisco, 2018). Indeed, the third NAP contained these policy commitments (Republic of Moldova, 2016). However, there is no certainty that the decisions of policymakers were based on these votes and not influenced by other political processes or even made beforehand. Moreover, the accountability aspect was problematic. The government has not published survey results online (the cited researcher gained individual access to the outcomes only 19 months later). Open-access government feedback on how public inputs were considered was missing. Consequently, the level of public influence during the development of the NAP was assessed as “consult” meaning that “The public could give inputs” (Mirza-Grisco, 2018, p. 20).

The impact of the 2018 non-binding online survey in Moldova was more evident than that of the 2016 one, especially in the accountability aspect. Alike in 2016, the top-five open government priorities were incorporated in the fourth NAP (Republic of Moldova, 2018). Similarly to 2016, we cannot attribute this outcome exclusively to the online survey. Nevertheless, we know for sure that the survey results were published online (Guvernul Republicii Moldova Cancelaria de Stat, 2018b) strengthening government accountability. Consequently, the level of public influence during the development of the NAP was assessed higher than in 2016 – as “involve” meaning that “the government gave feedback on how public inputs were considered” (IRM staff & Mirza-Grisco, 2021, p. 13). Specifically, the government published a consolidated table of the proposals to the NAP collected during the online consultations accompanied by comments from the authorities. It should be noted that the increase in the level of public influence does not necessarily mean higher citizen engagement but clearly reflects higher government responsiveness to public input.

When each stakeholder group is viewed separately, expert opinions reveal hardly any influence of the 2016 and 2018 online surveys for OGP themes on the attitudes and practices of voters, civil society and authorities in Moldova. Thus, a government official was sceptical about the impact of online surveys for OGP themes on any stakeholder group. Similarly, Veronica Cretu (independent expert) doubted that the survey either transformed the attitudes of its participants or there was any significant impact on public authorities. Nonetheless, considering the low number of voters (29 and 41 in 2016 and 2018 respectively) we can surmise the role of online surveys on voters (who most probably came from civil society) in Moldova. Within such a small sample, the voice of each voter and CSO should be more vocal thus strengthening the power of a small civil society group on OGP co-creation. As OGP guidelines do not set up a minimum participation threshold, this e-participation exclusiveness does not necessarily impede the quality of input (on the contrary, it may even raise it) or the legitimacy of the process, but it raises questions about the inclusiveness of online consultations in the country.

However, when the interaction between the public and the government is examined, Moldovan national experts did report some impact of the 2016 and 2018 voting in online surveys for OGP priorities on open government in the country. A government official acknowledged that the questionnaire facilitated the participation of civil society in the decision-making process and guided the course of action of the MSF. As specified by Stanislav Ghiletchi (civic activist), although the MSF did not entirely outsource the priority ranking to people, the analysis of the online questionnaire results helped the MSF validate and adjust NAP priority areas and later discuss them during respective thematic workshops between the civil society and public authorities. As an independent expert admitted, these activities made it evident that the policy priorities of the public of the government coincided.

It is noteworthy that the effect on open government was not due to online surveys alone, but because of its combination with other formats of cooperation between civil society and the authorities. A government official believed that the effect of voting in online surveys was positive especially due to open-ended questions (not voting, but rather a collection of ideas) that allowed free expression of opinions about areas not yet covered by the draft NAP and because of other face-to-face and online discussions about NAP content. In line with this, an independent expert stressed that to discuss complex issues public consultations should include both interactive forms like workshops and voting. Similarly, Stanislav Ghiletchi (civic activist) emphasised that the MSF engaged multiple methods and approaches to collect a wider, more representative and balanced pool of opinions of the people, civil society and public authorities. Also, Veronica Cretu (independent expert) mentioned that the draft NAP was shared with colleagues via Google Docs for comments and contributions and was later published on the particip.gov.md platform for public consultation. Nina Catirev (government official) admitted that the publication of the draft NAP on the government online portal secured the transparency of decision-making regarding specific policy commitments. Thereby, we can summarise that online surveys did prompt some civic participation as well as public transparency and accountability, but their impact on open government is more profound when combined with online crowdsourcing of ideas, discussions and other online consultation forms.

Online consultations, online surveys, i-voting and open government in Ukraine

In Ukraine, online consultations remain a rather underused e-engagement format (Khutkyy, 2019a). Largely, because online consultations are optional for authorities to conduct. Since 2017, the national parliament has conducted only 27 online consultations on the text of draft laws (the last one was in 2019), of which the most popular received a total of 863 “votes” and 61 comments, the second one – 84 reactions and one comment, all others – 15 or less “votes” (Verkhovna Rada Ukrayiny, n.d.). During 2018–2023, each ministry, central government agency and regional authorities announced their own consultations, mostly offline with the option of providing input via email (Uriadovyi portal, n.d.). The most numerous online surveys are in the Diia portal and mobile app of e-services. Launched in 2022 (years after the analysed OGP co-creation campaigns and are thereby incomparable to them), they have attracted over 2.1 million participants for a single survey (Ukrinform, 2023). However, civil society heavily criticised Diia online surveys for violating the law requirements for the discussion period, being unrepresentative and excluding Ukrainian citizens not using Diia (Lisovskyi, 2023) and for the equivocal choice of topics (Antonov, 2022). The sociological community also criticised Diia surveys arguing that the sample of Diia users probably skewed towards young urban males loyal to authorities, that biased question framing, excluded negative answers and that the surveys lacked transparency and accountability (Antoniuk, 2022). Other cases of nationwide i-voting include elections to civic councils at some agencies and ministries held since 2015 (Khutkyy, 2020a). The former elections mobilised over 40,000 voters in 2017 (Khutkyy, 2019a) – only around 0.1% of the total population. Despite the enthusiasm of some public officials (Morozov, 2022), i-voting at elections to public offices in Ukraine has not been introduced yet. Overall online consultations and online surveys in Ukraine are held, but their usage, transparency, and accountability greatly vary across government agencies, and in most cases lack popular interest and influence on public policy.

During its more than 11 years in OGP Ukraine has demonstrated evident progress in open government practice. The country joined OGP in 2011. Since then, by December 2022 it had developed five NAPs and implemented four of them (the fifth NAP 2020–2022 was on pause) (Open Government Partnership, n.d.-e). Over this time the country’s co-creation process demonstrated a vivid growth of civic engagement formats as well as the diversity and number of contributors (Khutkyy, 2019b). Simultaneously, the level of dialogue between the public and the government in developing NAPs evolved from limited and controversial consultations in 2011–2012 (Presniakov, 2015) to a comprehensive and iterative collaboration in 2017–2018 (IRM staff & Khutkyy, 2020). As documented in IRM assessment reports, online surveys with voting for priority OGP policy commitments in Ukraine occurred twice: in May 2016 (Khutkyy, 2018) and in July 2018 (IRM staff & Khutkyy, 2020). This indicates that the introduction of online surveys as a part of online consultations regarding open government policy priorities in Ukraine reflects a broader advance of civic engagement, citizens–authorities cooperation and public influence within the national OGP process.

Online surveys during the 2016 OGP co-creation in Ukraine

The 2016 voting for OGP priorities in Ukraine was an online survey. On 17 May 2016, the Secretariat of the Cabinet of Ministers of Ukraine (SCMU) posted an online announcement calling the public to vote online for five priority action points of the 2016–2018 NAP (Hromadianske suspilstvo i vlada 2016a). The hyperlink led to a Google Form, an online survey tool. The call also provided the link to the full 2016–2018 NAP draft, nudging the voters to be informed about the initiative and prospective reforms. This was an element of civic education. According to the announcement, the voting lasted 14 days. This is a reasonable time for voting, though not for a discussion. Google Form is not suitable for that purpose, so public discussions were conducted in person – at deliberations in the regions and the world café workshop in the capital.

The outreach and scope of 2016 Ukraine’s communication campaign promoting the online survey for OGP priorities were limited. The publication of the announcement online showed government transparency, albeit not perfect. According to the assessment in the IRM report, the public did not widely read the government website and the government disseminated the information about the online survey within the inner circle of CSOs that had collaborated with the government during previous NAPs (Khutkyy, 2018). Oksana Prykhodko (civic activist) also noted that the co-creation lacked a wide awareness-raising campaign and that she learned about the possibility to vote via a personal invitation. This indicates that the 2016 online survey for OGP priorities was in fact expert voting. This might have increased the weight of each vote, shifted the distribution of votes from a more popular demand to a more expert perspective and increased the readiness of authorities to value and incorporate the expert survey results. Yet, these are just speculations not supported by evidence. On 1 June 2016, SCMU published the online survey results. As reported in the official statement, 99 persons voted and five commitments received the most votes (Hromadianske suspilstvo i vlada, 2016b). The number of voters reflects the scale of proactive interest in the initiative and the scope of informing – rather small-scale. The very fact of publishing the online survey results demonstrates the accountability of the government on this issue. Though the government provided the number of votes for each top-priority draft commitment, the number of votes for all draft commitments would have brought more clarity.

Online surveys during the 2018 OGP co-creation in Ukraine

The 2018 Ukraine online survey voting for OGP priorities advanced in the diversity of digital media channels and the scale of communication activities. On 5 July 2018, the government used the OGP Ukraine’s Facebook webpage to call the public to vote for specific policy commitments for the 2018–2020 NAP (OGP Ukraine, 2018c). Facebook was a reasonable communication channel because in 2018 it had around 12.5 million users in Ukraine (Statista Research Department, 2023). The Facebook post received 36 “likes” and eight “comments”, and gained 58 “shares” signalling average visibility (other government’s Facebook posts usually receive dozens “likes”, “comments”, and “shares”).

The scale and depth of online consultations and voting in online surveys for Ukraine OGP priorities in 2018 progressed too. The abovementioned social media post forwarded readers to a dedicated webpage on the Discuto platform (OGP Ukraine, 2018b). The platform contains a description of OGP, deliberation activities, and i-voting guidelines. Notably, the webpage also contains a hyperlink to a YouTube video about OGP Ukraine (OGP Ukraine, 2018a). These activities reflect relevant transparency and civic education efforts by the government. Natalia Oksha (government official) admitted that the discussion and the voting facilitated the informing and engaging of the public and civil society in the OGP process. The e-platform allowed registered users to read, deliberate in the form of comments and vote for or against any of the 24 draft commitments. During 5-20 July 2018, the platform attracted 262 contributors who made 37 comments and made 2,712 votes (any user could vote for multiple commitments). The number of 262 users reflects a wider e-participation audience than in 2016, but still less than maximum and rather average e-participation (compared to the range between zero, 84 and 863 reactions in other government-held nationwide online consultations mentioned above). As Olena Ursu (development programme specialist) admitted, an intensive outreach campaign was required to encourage people to vote and later the e-platform user registration procedure limited the circle of voters to very motivated persons. The 37 comments indicate a low number for deliberation on nationwide policies, especially since they were made only by 13 users. This was a narrow discussion. Yet, presuming that the voters read the general description of the initiative and specific draft commitment texts, this should have been an informed deliberation and voting. Of all votes, 2,309 (85.1%) were in favour of proposals (positive votes), while 403 (14.9%) were against proposals (negative votes). Each commitment received overwhelmingly more positive than negative votes (commitments were ranked by the number of positive votes). This was a high level of agreement among the non-representative sample of voters. The detailed voting results are available over four years after the online survey period. This indicates a high level of government transparency in this respect. In addition, the government made a Facebook post and in comments to it provided the list of all 24 draft commitments ranked by the number of positive votes (OGP Ukraine, 2018c). These activities show some government accountability on this matter.

The impact of 2016 and 2018 online surveys on open government in Ukraine

The 2016 online survey of voting for OGP priorities in Ukraine likely had an objective impact on open government commitments for 2016–2018. The comparison of the voting results in the 2016 online survey (Hromadianske suspilstvo i vlada, 2016b) and the 2016–2018 NAP (Kabinet Ministriv Ukrayiny, 2016) revealed that all the top-five most voted draft commitments were adopted as actual policies by the government. IRM report corroborated this by assessing that during the development of the NAP, the voting procedure indicated that the level of public influence reached the level of “collaborate” meaning that “there was iterative dialogue and the public helped set the agenda” (Khutkyy, 2018, p. 20). Evidently, the actual “public” that voted was most probably active civil society, but it was still the “public” as opposed to the “government”. The OGP framework does not set up a quantitative threshold for counting an input as “expert” or “public”. Yet, we consider this in our analysis. Partially, this level of civil society impact (“public influence” in OGP terms) was not the highest because the voting for OGP priorities was not binding, but advisory for the government. In sum, this analysis demonstrates that although the 2016 online survey was an advisory online consultation, in practice its results shaped Ukraine’s policies of the 2016–2018 NAP. Since both the government and civil society formally committed to implementing them, the online survey influenced the responsibilities of both stakeholder groups.

The 2018 online survey with voting for Ukraine OGP priorities had a clear objective impact on open government commitments for 2018–2020. The comparison analysis of the 2016 online survey results (OGP Ukraine, 2018b) and the 2018–2020 NAP (Kabinet Ministriv Ukrayiny, 2018) demonstrated that 19 of the top 20 most-voted draft commitments were adopted as actual policies by the government. According to the IRM assessment, of the top 20 voted draft commitments only one was not included in the final NAP since it had been carried out by the time of NAP adoption (IRM staff & Khutkyy, 2020). The IRM report also explicitly specified that the MSF considered the ranking of the online survey by focusing on the top 20 priorities and merging some of them (IRM staff & Khutkyy, 2020). Partially due to the online consultations with the online surveys, the IRM assessment marked the level of public influence during the co-creation of the 2018–2020 NAP as remaining at the “collaborate” level (IRM staff & Khutkyy, 2020, p. 10). Thereby, similarly to the 2016 one, the 2018 online survey was advisory, but de facto its results defined Ukraine’s policies of the 2018–2020 NAP. In turn, this influenced government and civil society stakeholders responsible for implementation.

Subjective opinions of the interviewed national experts about the impact of the 2016 and 2018 online surveys indicate that the latter rather amplified the existing awareness, engagement and self-confidence of the active public, almost identical to the organised civil society. On one side there is the risk of broadening the gap between this group and the remaining citizens. On the other side, this process may reflect a specialisation of active citizens professionally working in the area of open government. As Olena Ursu (development programme specialist) reasoned, understanding NAP themes required some expertise on the topics, therefore the voting involved persons and CSOs professionally working on open government-related projects. She assumed that the voters were affected by the voting because they analysed prospective reforms deeper, voiced their opinions about country priorities and thereby felt involved in shaping the NAP and creating positive change in the life of compatriots. For example, as reported by Anna Pakhno (government official), the 2018 co-creation process included an extra 2018 online survey voting (beyond the ones analysed in earlier sections) among participants of an offline event dedicated to choosing priority Sustainable Development Goals. This public discussion engaged around 100 participants from the government and civil society, who deliberated on draft policy commitments in a world café format (IRM staff & Khutkyy, 2020). According to Natalia Oksha (government official), the voting at this event had the biggest impact on open government, because the voting audience consisted of informed persons interested in the voting results and eager to further develop those priorities into specific draft policies.

These reflections demonstrate that because of participating in online surveys strengthened by other online consultation forms (such as online discussions) civil society gained more advocacy power (but not the lobbying power of corporate interests, which was not found in this study), while the authorities became more open to public input, resulting in more concerted NAPs. Moreover, this was not only a civil society-government influence but a joint multistakeholder effort. In this aspect, Oksana Prykhodko (civic activist) admitted that the very approach of discussing policies with the public had a profound influence on facilitating collaboration among diverse stakeholders. From the perspective of Olena Ursu (development programme specialist), specialised CSOs mobilised their target audiences for voting and also believed that it is realistic to cooperate with the government in planning joint projects. Such voter mobilisation by SCO was an efficient civil society advocacy mechanism. Olena Ursu (development programme specialist) also noticed that the government, being aware of the voting role in the OGP process tended to consider policy priorities vital for the people and to adopt a more human-centred policy approach. This was most manifest in the 2018 voting, which the interviewed experts explicitly assessed as influential. In particular, as Natalia Oksha (government official) admitted, the voting empowered civil society to voice their opinions and to persuade the authorities to edit draft policy commitments. In addition, she acknowledged that comments on the online platform prompted partial re-wording of draft commitments thereby indicating some influence of voters on the final policy text. This finding highlights the importance of conducting comprehensive multi-format online consultations including online surveys and discussion forums for a productive policy co-creation.

Finally, the analysed online survey voting for OGP policy priorities had a notable impact on wider open government. It facilitated the introduction of i-voting for other purposes. As Oleksandr Ryzhenko (independent expert) acknowledged, the use of voting using digital means increased voters’ trust in the whole domain of electronic democracy. However, such trust in digitally powered voting with lower (as compared to more rigorous i-voting systems) standards of voter identification, cybersecurity, and other aspects bears the risk of expressing trust in nationwide referenda or elections of public officials with lower standards. Olena Ursu (development programme specialist) reflected that this voting demonstrated that e-voting is efficient, should be applied for government-public cooperation on reforms and thus laid the foundation for wider e-voting utilisation. Indeed, e-voting was a recurrent theme in NAPs themselves: in the framework of the e-democracy roadmap in the 2016–2018 NAP (Kabinet Ministriv Ukrayiny, 2016) that mandated e-voting as a component of the online platform for civil society-authorities interaction in the 2018–2020 NAP (Kabinet Ministriv Ukrayiny, 2018) that foresaw e-voting for elections of civic councils at executive authorities (Departament komunikatsii Sekretariatu KMU, 2021).

Conclusion and discussion

The study identified several differences between the two countries regarding the socio-political impact of voting in online surveys on open government. In Moldova, the influence of the online surveys for OGP themes (especially compared to offline discussions) was minor. Likewise, the level of visible civil society influence on the contents of NAP in Moldova was lower than in Ukraine. In contrast, in Ukraine, the effects of online surveys about OGP commitments on open government were more profound. Specifically, voting in the online surveys enhanced the awareness, engagement and self-confidence of the active public and facilitated the introduction of i-voting for other purposes in the country. Considering the reviewed in the country context overall advance of civic engagement, citizens–authorities’ cooperation and public influence within the national OGP process, the observed higher impact in Ukraine was probably due to more intensive multistakeholder efforts to implement OGP-related civic education, transparency and accountability.

Moreover, the inquiry found several divergences in allegedly resemblant practices in OGP processes in the studied countries. Co-creation processes in the two countries included deliberation but in Moldova only offline, whereas in Ukraine – both offline and online. In addition, Moldovan online survey voting was exclusively on the internet, while Ukrainian i-voting was once performed both on an online platform and as an extra exercise at an in-person event – in a hybrid format. This indicates experimentation with a wider spectrum of civic engagement formats by Ukrainian stakeholders. Notably, voting for policies in online consultations is more powerful when crowdsourcing of ideas, deliberation and crowd law policy drafting precede voting. Besides, although the voting for OGP policies was introduced in the same years, in Moldova it coincided with unstable developments, while in Ukraine – with an overall gradual advance of open government. Furthermore, although prioritising policies seemed similar in format, in Moldova the government perceived it as an advisory crowdsourcing online survey, while in Ukraine – as a prioritising online survey designed to steer formal decision-making. Partially because in Moldova voting options were broad themes, while in Ukraine – specific commitments. Also, the Moldovan online survey included open-ended questions later turned into concrete proposals, whereas in Ukraine one online survey was reinforced by interactive online deliberation about precise commitment formulations. The MSFs held online surveys at dissimilar stages of the policy-making cycle: in Moldova – at the agenda-setting stage, while in Ukraine – at the policy-drafting and decision-making stages. Also, in both countries, the voting for OGP priorities affected citizens-authorities cooperation. But in Moldova, the focus was entirely on OGP policies, while in Ukraine – both on policies and on civil society empowerment and government openness.

Finally, the research discovered common patterns in the impact of the online survey voting for OGP policies in both Moldova and Ukraine. It is highly likely that the voters belonged to the active public – organised civil society. Therefore, the impact of voting on voters was almost equivalent to the voting impact on organised civil society. Further, the expert-assessed impact of the online surveys on civil society was more manifest than the impact on the authorities. Moreover, the most far-reaching influence of online surveys is the magnified collaboration between the public and the authorities. Regarding the impact of online surveys on open government overall, the convergent objective outcome was the inclusion of top-voted draft policy areas or commitments into final NAPs. Although de jure the online surveys were advisory, de facto they probably influenced OGP priorities as argued further. Legally, only the central government was authorised with the ultimate decision-making mandate on NAP. However, the MSF, having representatives from both civil society and authorities, due to its sectoral mandate and negotiation abilities had a persuasive power sufficient to advocate priority OGP policies, voted by the active public. Thereby, due to consensus among stakeholders, formally advisory online survey voting for OGP policy priorities probably became an informally influential exercise. These case studies illustrate how an emerging semi-formalised policy-making innovation evolves.

The analysed practice of online surveys also reveals a controversy. The major disputable feature of the voting-powered online consultations in both countries is the low e-participation rate (as counted in absolute numbers). And although the rates correspond to the usual ones in other nationwide online consultations, they probably indicate insufficient outreach and low interest of the public. As a result, a narrow circle of the most informed and motivated civic activists votes in online surveys, deliberate on online fora, and shape open government policy priorities. On one side, this raises questions about the representativeness of public input and the consideration of the preferences of diverse societal groups. On the other side, the complex issues of nationwide government reforms naturally engage primarily experts in the field and result in a professional discussion and well-reasoned choice of priorities. Such expertocracy is not good or bad per se. As long as such online consultation is open for the inclusion of new participants, conforms to OGP co-creation requirements, and ensures high-quality input to open government policies, it functions as an efficient real-life implementation of an ideal vision of participatory democracy. The results of our research demonstrate that the cases of OGP co-creation in Moldova and Ukraine are largely aligned with this idea.