3. Evidence-based and data-driven public communication

Abstract

Key data on audiences, their perceptions, behaviours, and on the performance of different content and messages can make communication more strategic. This chapter provides an overview of governments’ gathering and use of insights about their publics and the application of behavioural science in this field. It also explores the potential of these practices to foster a better understanding between government and citizens and contribute to more data-driven and inclusive communication. The chapter concludes with forward-looking approaches to further mainstream the use of evidence in the design and delivery of public communication.

Data and evidence as prerequisites for effective public communication

Public communication is no different from any other policy area in that evidence and data are essential to its effectiveness against stated objectives. Indeed, communication can only be deemed strategic when it is grounded in analysis of the audience it aims to engage. This foundation allows activities and content to be designed based on a sound understanding of how specific societal groups perceive a given issue, consume information and trust government messages.

Evidence-driven communication applies data and insights at all stages of development and delivery – from the initial objective-setting and planning to the final stages of evaluation and learning. Such an approach seeks to build a strong understanding of the trends in public discourse around core policy issues; as well as audience perceptions, attitudes and habits; and the short-, medium- and long-term effects of communication activities (OECD, 2020[1]).

As mentioned in Chapter 1, the information ecosystem is increasingly crowded with a large volume of content. Multiple stakeholders across media, the private sector and civil society are competing for the public’s attention. More than ever, the current environment creates imperatives for making communication more targeted and compelling to specific publics. Doing so requires an understanding of behavioural, cognitive and psychological characteristics of specific groups of citizens, and knowledge of the communication channels they use.

Without a solid evidence base, institutions are casting their messages to an unspecified public in a crowded communication space, with little knowledge of whether and how information is received. Similarly, they have no means of monitoring and measuring the impact of their activity and whether it is serving any pre-defined objectives. For this reason, evidence – in the form of audience insights research, social listening 1 or behavioural insights (BI) – is a prerequisite for a strategic approach to communication.

Thanks to the digital transformation of communication and the vast quantities of data generated by online activities, insights into audiences have not only become more diverse and precise but also cheaper, faster, and more easily accessible. Big data and software as a service (SaaS) platforms for online analytics have made acquiring and processing this information simpler, even for teams that lack specialised research or data science competencies. More qualitative and nuanced insights still require applied research methods such as surveys, polling, focus groups or behavioural experiments. Indeed, responses to the OECD survey on “Understanding Public Communication” (2020) suggest these are important tools in many institutions.

The role of data in particular in shaping public communication is visible among OECD survey responses, where 30 out of 38 centres of government (CoGs) and 17 out of 24 ministries of health (MHs) reported utilising different sources of data to inform the design of digital initiatives (Figure 3.1). OECD survey results suggest however that governments have yet to exploit its strategic value for the delivery of more responsive and effective communication. As Figure 3.2 illustrates, most CoGs and MHs primarily collect data directly from audiences, and only a small share of them use data on public service uptake for example.

Mainstreaming the use of data in the design, delivery and evaluation of public communication is also at the core of using this function to enable social listening capabilities. Analysis of public discourse as well as attitudes and sentiment towards a given issue are essential not only to craft attuned messages and content, but also to improve policy and ensure it is in line with the needs of its intended beneficiaries.

Figure 3.1. Institutions that rely on audience insights or data about public services to inform digital communication efforts

Figure 3.2. Types of data used to inform digital communication efforts in CoGs and MHs

This chapter will explore the use of evidence, particularly with regards to audience and behavioural insights, in the design and delivery of public communication activities across CoGs and MHs in OECD countries and beyond. It will conclude with a series of reflections regarding the broader challenges to effectively leveraging this type of data, including ethical considerations and data privacy concerns that are crucial to keep in mind in this field. The subsequent chapter will complement the discussion of evidence-based communication with a focus on evaluation.

Research and insights to better understand and speak to citizens

Understanding the public is fundamental to communicating effectively and inclusively with all groups in society, including minorities and underrepresented groups. Gathering and using audience insights can help make public communication more relevant and impactful by tailoring the choice of different channels and messages to specific target audiences. Although research on audiences is a well-established practice in many countries, specific cases were markedly scarce across OECD survey responses. For this reason, this section draws more heavily on aggregated data, and less on practical illustrations, which is an area of focus for future data collection.

As defined in the OECD survey, audience insight refers to research activity that helps gain a deeper understanding of the public’s motivations, impeding factors, fears or barriers, as well as their understanding of the subject to be communicated and their attitude towards it, as well as their media consumption habits. Efforts to understand stakeholders have become more prominent as digital platforms that generate vast data on users’ demographic features and attitudes have increased in number. The same platforms also allow for a more precise targeting of diverse publics, making this knowledge highly useful.

Responses to the OECD survey suggest that there is still significant scope to adopt more sophisticated methods for insight gathering and make this a more regular feature of designing communications. Indeed, 41% of CoGs and 21% of MHs report using audience insights to inform communication planning only on an ad-hoc basis, while 21% of MHs do not do so at all (Figure 3.3). Slightly over a quarter of CoGs reported gathering audience insights at least every three months. These frequencies can be suitable for communications around issue areas or policies that are slower-moving. However, the COVID-19 pandemic and the spread of false or misleading content around it have highlighted the importance of understanding rapid shifts in public perceptions and demands for information that call for real-time capacity to conduct accurate research and adjust approaches accordingly.

Figure 3.3. Frequency in which CoGs and MHs used audience insight to inform communication planning in the past 3 years

The prevalence of less-than-frequent insight gathering can perhaps be attributed to the demanding nature of this area of communication in terms of time, resources and specialised skills required. Indeed, among the multiple different sources and methods of collecting this information, each has advantages, disadvantages, and contexts to which it is best suited. For instance, desktop research, such as media monitoring or social media analytics, is associated with low costs and rapid results, but tends to return top-line insights that may not be complete or accurate, or cut to the core of public opinion. Demographic data gathered by digital platforms for example is still largely self-reported data, so some biases can continue to affect the analysis in this field. More advanced data science skills can enable more sophisticated analysis but may be less common in communication teams.

By contrast, methods such as surveys, focus groups or interviews, which require extensive set-up, preparatory research and expertise, can return deep qualitative results but are highly resource-intensive. Despite this, data shows that between 70% and 50% of CoGs and MHs rely on surveys and focus groups for their analysis, with desktop research also being prevalent in over half of respondents (see Figure 3.4).

The availability of skilled professionals dedicated to this competency is an important enabler of a more frequent and advanced use of audience insights. Out of the 65% of CoGs reporting to have a dedicated person or team working on insight gathering, three-quarters have this staff embedded in the communication unit (although they sometimes cover multiple roles) and a small minority have separate teams with whom communicators co-ordinate. Similarly, 8% of MHs surveyed rely on external offices for their audience insights. Building these capacities and upskilling teams to benefit from new tools will be instrumental to strengthen the use of insights in communication.

Figure 3.4. Most important sources used by CoGs and MHs to gather audience insights

Leveraging insights for more targeted and inclusive communication

Understanding the public facilitates better segmentation of diverse audience groups, allowing a greater differentiation of what content is best suited to each group and the communication channels on which different segments are more likely to see and engage with. In this respect, audience insights create opportunities for government messages to be elaborated into formats that make them resonant and relevant for diverse audiences and also help this content reach the intended groups through informed use of platforms. For example, the United States Centers for Disease Control and Prevention (CDC) used responses to a Tweet where individuals stated they were preparing for a zombie apocalypse, alongside responses related to terrorism or natural disaster, to develop a campaign that resonated with the public. Building on these responses, the CDC disseminated information on how to react in the context of a zombie apocalypse (and other emergencies) as a way to spread important messages about pandemic preparedness. The campaign ‘went viral’ and increased traffic to the CDC’s Emergency Preparedness webpage by 1.143% compared to the same date the year before (CDC, 2018[2]). Such a use of insights can in turn enable more impactful communication but also support efforts to include harder-to-reach or vulnerable groups, such as youth and minorities, and ensure they are not left out of information-sharing and engagement opportunities.

Data from the OECD survey highlights that MHs engage in segmented audience targeting more than their CoG counterparts. Youth and people with disabilities are prominent target groups across CoG and MH respondents (Figure 3.5). Furthermore, close to a quarter of CoGs and 14% of MHs reported that they do not target specific groups. Indeed, only 23% of CoGs and 22% of MHs reported that audience segmentation and tailoring of communications was among the top three objectives of their audience insight gathering. Overall, these responses could indicate that diversified approaches and inclusiveness of communication are areas for development to implement more strategic communications.

Figure 3.5. Specific groups targeted in the communication of CoGs and MHs

Using the appropriate channels to reach all audiences is essential to the segmentation and targeting of communication. According to the OECD survey, the reach of a given channel, both in general and for specific target groups, is the primary criterion for its selection – above cost and timing (Figure 3.6). Insight-gathering is similarly instrumental in this respect, helping to understand which platforms each segment of the public consumes and how they engage with content. Survey responses show that using insight to select channels is somewhat established as a practice, with 50% of CoGs and 29% of MHs reporting that they do so.

Figure 3.6. Criteria used to determine communication channels by CoGs and MHs

From audience insights to social listening

Beyond its use in developing more tailored and effective communication, insight gathering can evolve into a mechanism for social listening – the practice of following online conversations and “listening” to citizens. An emerging theory on organisational listening, proposed most prominently by Jim Macnamara (2015[3]), elaborates the potential for communicators to listen to the public and understand the demand for information or engagement in order to respond to it. Organisational listening does not refer to snooping or tracking any individual or group’s speech or actions, which is contrary to data privacy and the democratic principles that guide this report (for more on ensuring privacy see the sections below and Chapter 5). Instead, listening refers to the legitimate practice of extrapolating trends from aggregated open data that can shape the communication agenda based on citizens’ needs, rather than the communication agenda driving the gathering of insights.

This is in contrast to the more common practice of communicating through “speaking” on a schedule determined by the government or institution’s own priorities and timings. Indeed, Macnamara’s research indicates that 80% to 90% of the communication by the organisations studied consists of “speaking”, meaning it is focused on one-way communication of information and content (Macnamara, 2017[4]). It suggests that public communication may be underutilised to identify and address citizens’ needs and that it is not sufficiently conceived as a two-way avenue for dialogue outside dedicated feedback initiatives.

As discussed in the OECD Principles of Good Practice for Public Communication Responses to Help Counter Mis- and Disinformation (OECD, forthcoming[5]), tracking real-time trends from aggregated data on discussions and interest in given topics is a useful practice. It goes hand-in-hand with interventions to fill so-called “information voids”, or gaps in reliable sources on a given subject, especially in relation to sensitive topics that are vulnerable to rumours and falsehoods (see Box 3.1). This approach is gaining ground in counter-disinformation efforts, and governments can leverage similar practices to make communication more responsive to citizens.

To this end, evolving the practices for gathering audience insights and building capacity to “listen” could help fully realise the potential of public communication to reinforce better governance and rebuild public trust. In turn, insights from listening activities could feed into policy making, and service design and delivery to ensure citizens’ voices translate into government action.

Box 3.1. The use of social listening to identify information voids to inform vaccination efforts in the United States during COVID-19

Using a mix of quantitative and qualitative data, including the use of social listening, the CDC publishes publicly available reports on citizens’ perception of vaccination efforts in the United States at least once a month. The reports have been instrumental in identifying the extent to which false information has spread and impacted the vaccination efforts. Moreover, these reports have enabled the government to pinpoint interventions that respond to prevalent mis- and disinformation and fill information voids with accurate and clear messaging. For instance, the report published June 7th, 2021 found that vaccinated individuals had concerns about the effectiveness of the doses against the Delta variant. This insight enabled the CDC to recommend, among other measures, that messages target the effectiveness of the vaccines on the most common variants in the United States, and the need to continue partnering with trusted figures to amplify these messages.

Source: https://www.cdc.gov/vaccines/covid-19/vaccinate-with-confidence.html ; https://www.cdc.gov/vaccines/covid-19/downloads/SoVC-report-9-508.pdf.

The use of behavioural insights for communication

BI is defined as the “lessons derived from behavioural and social sciences, including decision making, psychology, cognitive science, neuroscience, organisational and group behaviour” (OECD, 2017[6]). It acknowledges that human behaviour is shaped by systematic biases that can hinder the ability of individuals to act in their best interest (OECD, 2019[7]). While drawing on BI can enable governments to improve communications to better prompt behavioural change, BI does not refer to changing individuals’ behaviours against their own will. Rather, when citizens struggle to make choices in their own interests, such as quitting smoking for example, communications can leverage BI to craft messages that help individuals overcome their own biases while preserving their freedom. In this regard, its use can make public communication more efficient in improving the welfare of citizens.

Disseminating information and conveying clear messages to citizens is often only half of the goal of a public communication campaign. If the messages transmitted are intended to lead to a change in behaviour, it is important to integrate behavioural insights early on. To ensure concrete outcomes, communicators can build on evidence about what actually influences behaviours, instead of relying on beliefs of “perfect rationality”, broad assumptions or declared intentions.

Sharing information on its own is not always enough to lead to behavioural change. Even when the target audience has a good understanding and adequate awareness of a specific policy issue, an “intention–action” gap can undermine the effectiveness of communication efforts. Indeed, research points to the relatively weak relationship between awareness, intention and action. For example, pro-environmental behaviours may not be adopted by those professing pro-environmental attitudes and beliefs (Dryzek, Norgaard and Schlosberg, 2011[8]; Eom, Kim and Sherman, 2018[9]). Survey results reveal that CoGs and MHs cite raising awareness and informing citizens about their rights and responsibilities more often than behaviour-specific objectives (such as engaging stakeholders or promoting the uptake and improving the delivery of public services).

This section will explore how public communication can use BI to better understand and encourage shifts in citizen’s behaviours by incorporating insights from different segments of society in its various stages. It will first examine how CoGs and MHs are interacting with BI experts to design communication activities. It will then outline potential opportunities to leverage BI at each step of the communication planning process. Examples from OECD countries and beyond are used to illustrate good practices in this field and reflect on recent lessons from the COVID-19 crisis.

How are governments leveraging BI in the context of public communication?

Over the past decade, BI units have started to emerge within government institutions across the world. While many are already collaborating with communicators, OECD survey results point to an important number of countries that have yet to engage with BI experts to inform the design and delivery of communication efforts (see Figure 3.7). As a matter of fact, close to 63% (24 out of 38) of CoGs and 57% (13 out of 23) of MHs claimed to engage with behavioural experts within government or in academia and civil society. This is consistent with the fact that only 10 out of 24 CoGs and 6 out of 9 MHs report having a communication strategy or plan aiming at addressing behaviour change.

Figure 3.7. Share of CoGs and MHs that interact with BI experts

Diverse institutional structures supporting BI in public communication

Despite their growing popularity, OECD survey results reveal that dedicated BI units are yet to be institutionalised within many CoGs and MHs (see Table 3.1). In fact, half of the CoGs whose public communicators engage with BI experts have no such expertise within government and engage solely with external actors (including civil society and academia). In addition, 6 out of 13 MHs that reported to engage with BI experts also stated that they have such expertise within government.

Several countries engage with both internal and external BI experts to enrich their communication planning. For example, the United Kingdom has both internal BI capabilities within Public Health England (PHE), the National Health Service (NHS) and other parts of government, including the communication team within the Cabinet Office (OECD, 2020[10]). In the case of France, the French Interministerial Directorate for Public Transformation (Direction interministérielle de la transformation publique or DITP) functions as an inter-ministerial body providing tools and recruiting external experts for specific ministry projects (OECD, 2020[10]).

Table 3.1. Types of interactions with BI experts in CoGs and MHs
	Has a BI expert(s) on staff	A BI expert(s) working in the communication team of other government entities	A BI expert(s) inside government (not in comms team)	Interacts with external BI experts	Other	Does not interact with BI experts
Australia	■	●	●■	●■
Belgium			■	●■
Canada			●■
Chile	●■	●■		●
Colombia				●		■
Costa Rica						●
Czech Republic				●
Estonia	●			●
Finland				■
France		●
Germany		●
Greece	■			■
Hungary				●		■
Iceland						■
Ireland	■	■	■	■		●
Israel					●
Italy				●	●
Japan					■
Korea						●
Latvia				●
Lithuania				●		■
Luxembourg						■
Mexico						●
Netherlands	●	●	●	●
Norway						●
Poland						●
Portugal				■
Slovak Republic				●
Slovenia						●
Spain	■			■
Sweden						●■
Switzerland				●■
Turkey				●
United Kingdom	●	●	●	●
Total OECD CoG ●	4	6	4	15	2	8
Total OECD MH ■	5	2	4	8	1	6
Armenia	■			●
Brazil						●
Ecuador						●■
Jordan				●		■
Lebanon	●					■
Morocco						●
Paraguay						●
Philippines				●
Romania						●■
Thailand	■	■		●■
Tunisia						●
Total all CoG ●	5	6	4	19	2	14
Total all MH ■	7	3	4	9	1	10
Notes: Austria CoG and Turkey MH did not provide a response to this question. Black circles (CoGs) and black squares (MHs) represent those respondents that engage in certain type(s) of interaction(s) with BI experts.
Source: OECD 2020 Survey “Understanding public communication”.

Evidence from the OECD survey also indicates diverse institutional set-ups among OECD and partner countries (see Table 3.1) in terms of BI experts within or outside government. For example, public communicators in Hungary state that they regularly consult with experts in social sciences, while the government of Mexico follows a multi-stakeholder model, referring to experts in different policy domains prior to the deployment of whole-of-government campaigns. Structurally, 4 out of 24 CoGs (Australia, Canada, France and Germany) and 2 out of 13 MHs (Belgium and Canada) stated that they did not have a BI expert within their communication team in 2019 but interacted with BI experts located in other areas of government. In some cases, there is a centralised BI function that offers expertise across government. This is the case in Germany, where BI experts at the Federal Chancellery are available to support other parts of government on a range of projects, including communication.

Additionally, survey results reveal that BI is not necessarily applied evenly throughout government. Indeed, survey data indicates that MHs are more likely to have internal BI resources than CoGs. For example, in Armenia, Australia, Ireland and Thailand MHs indicated to have BI experts within the Ministry’s staff while the CoGs did not. In addition, countries such as the United Kingdom and Ireland have created specialised BI units or have units with expertise in BI within MHs. In fact, 7 of the 13 (54%) MHs that reported engaging with BI expertise have such a person in their team, whereas the same is only true for 5 out of 24 (21%) CoGs. The COVID-19 pandemic has highlighted the need for such arrangements to be as flexible as possible. For example, Ireland created a specialised subgroup, temporarily combining internal expertise (from the Ministry’s communication unit and research unit) with the expertise of external partners from other state organisations and academia to inform the deliberations of the National Public Health Emergency Team (NPHET). The role was to provide insights and to carry out targeted behavioural research studies to provide insight for communication about the virus and public health behaviours (see Box 3.2).

Box 3.2. Ireland’s use of behavioural insights to support public communication

The National Public Health Emergency Team (NPHET) in Ireland leveraged BI as a means to support public communication activities. A Subgroup on Behavioural Change was set up in March 2020 including social and behavioural scientists as well as communicators.

The aim was to better understand and mitigate issues surrounding messaging fatigue, misunderstanding or non-compliance with public health guidance. This was done through active listening of citizens as well as by identifying public perceptions and norms that may affect the country’s National Action Plan to fight the pandemic. The Subgroup supported communicators through the sharing of evidence and provided advice on tone, content as well as outputs of communication messages.

Members of the Subgroup also worked on specific projects and collaborated on COVID-19 response actions across the Ministry. The Ministry of Health published two brief cases studies on the application of behavioural insights to two such areas: hand washing and the COVID Tracker App for Ireland. First, a behaviourally informed hand washing poster was issued to all households. The poster was informed by international public health guidance but also insights from a literature review of behavioural science and hand hygiene. The poster focused on capability (how to) but also motivation, emphasising emotions of threat and disgust (“Kill the virus”) and affiliation (“To protect you and others” and “Save lives”). Second, the design and launch of the COVID Tracker App for Ireland was informed by insights from a user experiences study, expert review, and pre-testing of content options. The wider evidence-informed communication strategy enabled the creation and dissemination of accurate, impactful information to citizens and empower the public to make informed choices as well as increased compliance.

Having met its terms of reference the Behavioural Change Subgroup was stood down following its last meeting on 24 July 2020. A COVID-19 Communications and Behavioural Advisory Group (CBAG) was formed to advise the Department of Health on these matters. The CBAG consisted of members of the former Subgroup on Behavioural Change but with wider representation from health sector. CBAG provided insights on the Health Service Executive (HSE) and the Department’s broader communication strategies which plays a significant role in overcoming the challenges that existed around vaccine uptake. The CBAG regularly reviewed public opinion data on COVID-19 vaccination intentions. Members of the CBAG undertook a systematic view of evidence from 20+ surveys on COVID-19 vaccine uptake intentions and built on this to undertake predictive analysis using data collected from Ireland. The CBAG also provided insights for actions to support the uptake of COVID-19 vaccinations such as the content of SMS messages after registering to get a vaccine and a vaccine leaflet for households. The latter was also informed by a meta-analysis of international randomised control trials testing the effect of individual correspondence on uptake of influenza vaccines. In addition, members of the CBAG produced a short paper identifying 15 key points to support uptake of COVID-19 vaccines. These research reports and insights from the members of CBAG helped to inform the work of the Department of Health and the HSE in public communications.

Furthermore, the Government Information Service (the unit responsible for co-ordinating communications centrally, located in the Department of the Taoiseach) commissioned the Economic and Social Research Institute (ESRI) to undertake a behavioural study to inform policy and communication in respect of their response to the pandemic. From the outset of the COVID-19 pandemic, and in particular in advance of the development of vaccines, it was clear to the government that understanding human behaviour was a key way to protect citizens.

The Social Activity Measure (SAM) is a behavioural study that records the public response to the risk of COVID-19 infection and COVID-19 guidelines. Designed by the ESRI’s Behavioural Research Unit (BRU), SAM is an anonymous, interactive, online study that surveys people about their recent activity. It examines where and how risks of COVID-19 transmission arise. SAM aims to inform policy regarding the opening of the economy and society, while keeping the virus under control. The research is funded by the Department of the Taoiseach.

Method

SAM is a “prompted recall” study that uses methods from behavioural science to help people to recall their activities. It asks about times when people left their homes via factual neutral questions. Questions cover locations, people visited and visitors to their home during the previous week. Follow-up questions gather greater detail about the previous two days: how many people participants met, for how long, ease of keeping a 2m distance, use of hand sanitiser and face masks, and so on. The study concludes with questions about the pandemic more generally.

SAM has produced many specific findings, which can be grouped into different sorts:

trends in the locations people visit and how these change as restrictions change and the public mood shifts
measures of riskier behaviour, including the proportion of people having close contacts, the number of individuals people meet up with from outside their household, the prevalence of social visits to other homes
indicators of day-to-day risk mitigation, such as mask wearing and maintaining social distance
breakdowns of these behaviours by location and social group, identifying where the largest risks are occurring and who is taking the most risk
associations between perceptions, attitudes and behaviours, such as the strong relationship between perceiving restrictions as coherent, increased compliance and reduced risk taking
non-associations, such as the weak impact on behaviour of the perceived threat of fines
changes in expectations as the pandemic evolves and policy changes
figures for vaccine uptake and reasons for hesitancy among a minority
estimates of how vaccination is changing behaviour
measures of compliance with public health guidance and the perceived compliance of others
trends in travel across county boundaries, to Northern Ireland and overseas
perceptions of different aspects of the policy response.

Source: https://www.gov.ie/en/collection/3008f6-the-national-public-health-emergency-team-nphet-covid-19-subgroup-be/; “National Public Health Emergency Team Behavioural Change Subgroup Draft Paper: Overview of the Work of the Subgroup to date” (2020). Department of Health, “National Public Health Emergency Team Behavioural Change Subgroup Draft Paper: Overview of the Work of the Subgroup to date” (2020). https://www.gov.ie/en/collection/3008f6-the-national-public-health-emergency-team-nphet-covid-19-subgroup-be/; Murphy, R. and R. Mooney (2020), A brief case study on using a behaviourally informed poster to improve hand washing in homes, Research Services and Policy Unit, Department of Health, https://www.gov.ie/en/collection/3c5bc8-health-research-and-statistics/; Murphy, R., S. Gibney and G. Mac Criosta (2020), Refining Tracker App content with user experience, expert review and an experimental study, Research Services and Policy Unit, Department of Health, https://www.gov.ie/en/collection/3c5bc8-health-research-and-statistics/; Murphy, R. (2020), Using Behavioural Science to Improve Hand Hygiene in Workplaces and Public Places, Research Services and Policy Unit, Department of Health, https://assets.gov.ie/73447/7989b01eb9844f1aaa636d0ba7c254f7.pdf; Gibney, S., L. Bruton and P. Doherty (2020), COVID Contact Tracing App: User Perspectives and Experience Research, Research Services and Policy Unit, Research and Development and Health Analytics Division, Department of Health, https://igees.gov.ie/wpcontent/uploads/2020/07/Research-Report-App-user-experience-and-perspectives-May-2020.pdf ; Murphy, R. (2020), Summary of Comments by the Sub-group on Behavioural Change on the content and design of the COVID-19 Contact Tracing App, Research Services and Policy Unit, Research and Development and Health Analytics Division, Department of Health, https://igees.gov.ie/wp-content/uploads/2020/07/Behavioural-Change-Subgroup-Report-April-2020.pdf; Julienne, Hannah and Lavin, Ciarán and Belton, Cameron and Barjaková, Martina and Timmons, Shane and Lunn, Pete (2020) Behavioural pre-testing of COVID Tracker, Ireland’s contact-tracing app. ESRI Working Paper 687 December 2020; Muldoon, O., Bradshaw, D., Jay, S., Kinsella, E., Maher, P., Taaffe, C., Murphy, R. (2021). A research paper produced for the COVID-19 Communications and Behavioural Advisory Group, 2021 https://www.gov.ie/en/collection/3c5bc8-health-research-and-statistics/#behavioural-insights-and-patient-public-engagement; Murphy, R., Taaffe, C., Ahern, E. (2021). A meta-analysis of the impact of individual correspondence on flu vaccination rates: considerations for COVID-19 vaccination. Department of Health and Dublin City University. https://www.gov.ie/en/collection/3c5bc8-health-research-and-statistics/#behavioural-insights-and-patient-public-engagement; 15 Key Points to Support Uptake of COVID-19 Vaccines, Purcell, K., and Murphy, R. members of the COVID-19 Communications and Behavioural Advisory Group.

Organisational and institutional challenges for BI integration in public communication

Based on the above findings, communication practitioners must overcome several challenges to ensure BI is used optimally. First, it is crucial that communicators have access to BI expertise within or beyond government. However, OECD survey data revealed that a significant portion of CoGs do not interact with BI experts or do so only on an occasional basis. Integrating BI experts into communication teams – whether through internal hiring processes or institutionalised partnerships with external parties – can lower barriers to collecting, using and evaluating BI and raise awareness around their benefits.

However, institutionalisation on its own is not enough to systematise the solicitation of BI expertise. Even where access to experts is provided, if the inclusion of those experts occurs on an ad-hoc or case-by-case basis, there is a risk of missing opportunities to enhance the impact of a campaign.

A key means to overcome this is to integrate an assessment of the potential use of BI in the design process for campaigns, including the identification of opportunities to include BI and an analysis of which fields of expertise would be most relevant.

BI practitioners within and outside government also require support from senior officials. Decision makers can advocate for an early-on integration of BI experts and call for ambitious experiments and the use of results for future communication campaigns.

The role of BI in public communication delivery

BI can be solicited at every stage of the communication process, from the objective setting and the preliminary collection of insights to the evaluation phase (see Table 3.2). Including these types of insights throughout the process allows governments to have a rigorous and coherent approach to promoting behaviour change through communication. Indeed, BI can help detect and better understand biases, propose levers of behavioural change, and evaluate the impact attained.

Table 3.2. Four-step model of BI implementation in communication campaigns
Stage	What can BI bring?	How?
Identify communication objectives	Identify the target behaviours of a given campaign Identify what success looks like from the behaviour change angle Identify enablers and barriers to the target behaviour(s) Identify whether the target behaviour is indeed likely to be influenced by communication material or other tools Check whether the target audience is relevant, Help set a realistic goal based on previous experiences and research.	Identify target behaviours before engaging in strategic communication thinking Identify a metric for success for behaviour change (e.g. 70% of the target audience is washing their hands) Identify what determines behaviours through observation, interviews and surveys. Model the factors influencing behaviours through the lense of BI models framing motivational and capability factors, for example such as the COM-B framework. Conduct literature reviews on the topic, to know what research has identified as important factors for this behaviour in other settings. Look for evaluations of similar campaigns in order to acknowledge the effects that were obtained and their extent to know what can be expected.
Identify, design and inform a communication strategy or plan	Offer several levers to trigger behavioural change through the general message and anticipate potential spillover effects. Identify the most effective timing, channels, and formats for the communication activity. Adapt material to different audience segments. Reflect on the framing of the message, the visual and objective of the organisation.	Use a model to map identified obstacles and potential levers and assess the chances for each to succeed. Integrate possible feedback in the model to anticipate unintended consequences of the campaign. Use audience insights and literature on specific groups to assess the channels and contexts most susceptible to influence. Test messages and framing to assess their respective efficiency, through A/B testing or other experiments.
Reflect BI on the delivery of campaigns	Monitor the timing of the intervention and that the real-time measures show no unintended effects. Collect relevant data in prevision for evaluation	Run regular surveys to assess the efficiency of the strategy, sentiments, perceptions and trust in the campaign. Adjust the message or format of the campaign if necessary
Evaluate effects on behaviour change	Assess whether an actual behavioural change occurred and its effects on intended audiences. Analyse possible reasons for failures or shortcomings and propose improvements Synthetise, archive and publish the results to allow accumulation of experience and learning	Run rigorous evaluations, using relevant measures of behaviours and counterfactuals whenever available. Adapt the methodology of testing to the data available, the resources and the time scale.
Note: COM-B stands for Capability, Opportunity, Motivation, Behaviour. It is a model of behaviour and behaviour change that helps identify where adjustments need to be made to enable and encourage specific actions.
Source: Author’s own work, based on UK GCS (2021).

Mobilising BI to segment and understand audiences

A key task for any communicator is to identify an audience and build an understanding of its existing beliefs, expectations and behaviours. Applying BI can aid in the segmentation of groups and in the definition of objectives to tailor communications to different needs. Its potential to build an understanding of the drivers, fears and media consumption patterns of different target groups was acknowledged by a significant share of CoGs and MHs as a priority objective for their use of BI (see Figure 3.8). This can help identify which audiences are most likely to change their behaviour and which messages and messengers will be most effective.

Figure 3.8. Objectives of CoGs and MHs for using audience insights

Indeed, standard research methods such as surveys and focus groups can only go so far in achieving these aims as self-reported behaviours and motives do not always overlap with actual behaviours (Kollmuss and Agyeman, 2002[11]) (Kormos and Gifford, 2014[12]). They reflect attitudes rather than real actions, which can prevent the identification of the most adequate communication channels, and of the actual drivers of the target audience. BI research has a long history of narrowing the gap between declared and actual behaviours.

Evidence drawn from audience insights was indicated to influence campaign objectives for just under one-third of CoGs and MHs in OECD and partner countries (see Figure 3.5 above). BI can enhance the use of insights with research-based and tested models adapted to each type of audience, and can be a useful means to identify the extent to which behaviour is likely to be influenced by a communication campaign.

To capture actual behaviours, it is also important to reflect upon the ways in which the insights are collected. OECD survey results indicate that the primary methods for the collection of insights include the use of surveys (24 of 35 CoGs; 12 of 18 MHs), focus groups (18 of 35 CoGs; 9 of 18 MHs) and desktop research (18 of 35 CoGs; 10 of 18 MHs).2 These are common methods utilised in evidence-based communication to collect “self-reported” behaviours and intentions. However, it is important to complement these insights with information about actual behaviours, beyond “self-reported” behaviours and intentions. Using BI to design and inform communication strategies and activities

Despite the growing adoption of BI, behaviourally-informed insights are rarely used at the planning stages of communications. According to OECD survey results, the evaluation of behaviours (13 out of 24 CoGs and 5 out of 9 MHs), biases (7 CoGs and 4 MHs) and identification of initiatives to address behaviour change (10 CoGs and 6 MHs) are prioritised in less than half of communication strategies and plans across surveyed countries. These results reveal that even though individuals’ perceptions and actions are sometimes measured, BI is not systematically used in the planning processes, even though it can yield valuable results. In the case of the Government of Canada, for example, the administrationconducted monthly surveys that were behaviourally-informed to feed into the design of ongoing crisis communication efforts (see Box 3.3).

Box 3.3. Behaviourally informed surveys informing communication in Canada during COVID-19

Throughout the COVID-19 crisis, a team of behavioural scientists surveyed a panel of more than 2 000 Canadian citizens on a close to monthly basis to assess the evolution of their perception of risk, their trust in government and media, and their fatigue. The latter are all crucial factors influencing compliance with the recommended health practices communicated by the government. These insights also aimed to identify the least adhered-to health practices, the segments most at risk of infringing them and the spread of misinformation about the virus. The survey highlighted the value of personal storytelling and trusted messengers for promoting vaccine acceptance, which is now informing public communication strategies.

Source: https://impact.canada.ca/en/challenges/cosmo-canada/methods.

Furthermore, BI is well-positioned to provide ex ante advice on the most effective timings, channels and formats for communicating to different audiences. Whether it is because people are already thinking about the topic, because they are in a phase of behaviour change or simply because they have time to perform the behaviour, BI can greatly inform communication activities in this regard.

BI can also help public communicators fine-tune the details of ongoing campaigns, such as message framing and visual identity to maximise efficiency, readability and cognitive salience. In this regard, a good practice observed in the field is to test the framing of messages and visual content prior to the launch of a campaign to assess its efficiency, for example through A/B testing.3

Interestingly, testing campaigns through focus groups, comparison of options and user-testing is a common practice in 29 out of 37 CoGs and 17 out of 22 MHs and (see Figure 3.9). However, these methods focus on gathering declared intentions, which might not reflect actual behaviours. As such, bigger and more representative samples might be needed, complemented by both qualitative and quantitative sources of data.

Figure 3.9. Methods employed by CoGs and MHs to test campaigns before their implementation

Reflecting on behavioural factors for the delivery of key communication campaigns

The field of BI has rapidly evolved over the last decade to advance the application of cutting-edge theory into practice. To this end, several frameworks have emerged to summarise complex behavioural science literature and map the main factors affecting human behaviour. Some frameworks are particularly relevant to analyse behavioural challenges in terms of barriers and enablers to behavioural change. For example, the COM-B model is widely used to analyse behaviours using three main factors that contribute to behavioural change: capability, opportunity and motivation (Michie, van Stralen and West, 2011[13]). Other frameworks summarise how to apply behavioural science in practice to improve communications. For example, the EAST framework developed by the UK Behavioural Insights Team (BIT), outlines four main ways to facilitate behavioural change: communications campaigns can encourage behavioural change by making the desired behaviour Easy, Attractive, Social and Timely (EAST) (The Behavioural Insights Team, 2014[14]).

OECD survey results reveal practical examples of how BI factors can be integrated from theory to practice in key government campaigns (see Box 3.4). In Australia, behavioural insights were employed to identify at-risk groups in a campaign aimed at reducing the number of road accidents by personalising messages to the identified groups. Furthermore, BI has been used by the Government of Canada for redesigning communication material to increase the uptake of the online governmental service MyBenefits and resulting in its increased utilisation. Finally, the Netherlands used behavioural evidence to facilitate the work of check-out staff in groceries to lower the number of minors buying alcohol or cigarettes. By showcasing a weakness in the current system of ID checks, the government pushed citizens to systematically provide their IDs to checkout staff through communication materials which in turn, deterred youth from attempting to buy these products.

Box 3.4. Examples of BI-driven communication campaigns in OECD countries

Australia

The Western Australian Road Safety Commission partnered with external consultants in order to improve road safety. BI were used to support change in this area through a five-year communication campaign strategy aiming at reducing the occurrences of accidents linked to speeding, drunk driving, mobile phone distractions and seat belts.

The “Zero Heroes” campaign targeted a segment of the population for positive reinforcement communications. It reached high awareness level amongst the population (71%) and had behavioural impacts: 34% took some sort of action, as a result of the campaign; 21% checked their demerit points on the website of the Department of Transports. Campaigns such as “Time with Mum” and “Priorities” were developed in order to target higher risk and resistant segments of the population. The motivation and deterrents to behaviour change were identified in these groups and subsequently used to provide personalised campaigns as well as increase uptake of the desired behaviour. It reached relatively high levels of awareness in the specific target populations (39%) an audience that is renowned for not listening and not relating to Government messages.

Canada

In 2015, Ontario’s Behavioural Insights Unit (BIU) became the first government unit dedicated to applying behavioural insights to enhance public services in Canada. The BIU conducts rigorous experimentation (e.g., through randomized controlled trials (RCTs)) to assess the impact of interventions such as enhanced communications on compliance, uptake, and efficiency of government programs and services. For example, the BIU designed letters to encourage eligible student uptake of the Human Papillomavirus (HPV) vaccine, emphasising the student’s time-limited eligibility for a free vaccine or the cancer risk associated with HPV infection. This trial also made enhancements to the appointment booking system, enabling multiple email reminders with planning prompts to help overcome students’ prospective memory failure and to reduce appointment no-shows. The RCT revealed that students who received the redesigned letters were twice as likely to get vaccinated than students in the control condition, with more than a thousand students getting the vaccine during the trial period alone. Another notable achievement of the BIU was increasing fine collections under the Ontario Provincial Offences Act, leading to an anticipated $9.3 million increase in collected fines each year. As part of the trial, the BIU adapted three versions of a fine notice to include new salient formatting with concrete steps, along with statements that highlighted the social norm that most recipients make on-time payments, or used a loss aversion framing that emphasized the penalties for inaction. The BIU continues to collaborate with ministries across the Ontario government to facilitate the improvement of public programs and a shift to online services.

The Netherlands

The NIX18 campaign by the Ministries of Health, Welfare & Sport and General Affairs in the Netherlands aimed at making not drinking and smoking under the age of 18 the norm. Behavioural insights found that employees in supermarkets struggled to conduct ID checks, while customers reported annoyance at being considered too young. A communication campaign followed these findings in order to push customers to systematically provide their ID when buying cigarettes or alcohol. Materials, including posters and checkout dividers were provided to supermarkets. By urging vast swaths of the population to show their identification systematically, this also convinced youth that they would not be able to buy such products easily.

Source: https://www.kantarpublic.com/our-work/improving-road-safety-in-australia https://www.health.gov.au/health-topics/medicare-compliance/how-we-ensure-compliance/behavioural-insights-and-interventions https://www.ontario.ca/page/behavioural-insights-ontario-update-report-2020#section-2 https://www.nix18.nl/ ; a wealth of behavioural insights : 2017 edition and Author’s own work, based on Budge, M., Deahl, C., Dewhurst, M., Donajgrodzki, S., & Wood, F. (2009). Communications and behaviour change.

During the COVID-19 pandemic, BI practitioners have demonstrated the value of applying BI by adapting the content of communication campaigns to the perceptions of risk and awareness levels of citizens (see Box 3.5). In France, the behavioural analysis unit within the Direction Interministérielle de la Transformation Publique (DITP) evaluated key prevention campaigns and provided communicators with related advice. Furthermore, Finland, the Republic of Korea, Switzerland and the United Kingdom, partnered with influencers and other trustworthy public figures to amplify governmental messaging during the crisis. Finally, emotional responses were leveraged in communication campaigns to increase civic duty and moral responsibility among the population. Positive emotions such as pride, joy or hope have been identified as more efficient to trigger voluntary action, as opposed to negative ones which may lead to inaction or self-protection (Brennan and Binney, 2010[15]). The integration of BI in campaigns allowed countries to fine-tune previous actions and collect relevant data that can be made use of in the evaluation stage.

Box 3.5. Examples of BI campaigns in the context of COVID-19

Canada

Emotions are a strong driver of behaviour. During the COVID-19 pandemic, the Government of Canada ran a campaign to promote physical distancing by appealing to individuals’ values, moral responsibility as well as civic duty to protect the most vulnerable groups in society. This was coupled with actionable solutions, in this case: staying home. Campaigns which promote clear solutions to issues can empower individuals by providing them with a sense of control and deterring feelings of hopelessness.

Switzerland

Citizens tend to care about what others do and the sources from which information comes from can impact its trustworthiness. Governments around the world have taken advantage of this, asking influencers to participate in the efforts to contain the spread of the virus. In Switzerland, social media campaigns such as #soschützenwiruns (“let’s protect ourselves”) were leveraged. Trusted influencers shared images and information through these channels to convince citizens to stay home and protect vulnerable groups.

Source: Author’ s own work based on (Durantini et al., 2006[16]) and (Bavel et al., 2020[17]); additional sources include https://impact.canada.ca/en/challenges/covid-communications/campaigns https://www.oecd.org/coronavirus/policy-responses/regulatory-policy-and-covid-19-behavioural-insights-for-fast-paced-decision-making-7a521805/

Communication campaigns aiming to promote vaccination confidence also benefitted from the use of behavioural insights. For instance, the Government of Canada deployed methodologies and techniques to monitor knowledge, perception, fears and behaviours of citizens related to COVID-19. These insights found that people are more likely to respond to personal narratives regarding vaccination experiences than to information campaigns (OECD, 2021[18]). In Ireland, the Department of Health’s communication strategies were informed by the COVID-19 Communications and Behavioural Advisory Group (CBAG). CBAG played a strategic role by providing advice on communication tools that could be leveraged to increase the uptake of vaccinations such as the use of SMS messages after registering to get a vaccine confirming registration as well as reminding individuals of their appointments. Finally, in Colombia, a Randomized Control Trial was carried out to compare the intent of individuals to get vaccinated before and after exposure to a message which read “Healthcare workers will be the first to receive the vaccine. To help them fight COVID-19, when it’s your turn, they need you to get the vaccine too.” The research which was conducted in partnership with the British Embassy found that those exposed to the message were more likely to be vaccinated (The Behavioural Insights Team, 2021[19]).

Evaluating the impact of public communication in changing behaviours

Evaluating observed behaviour change at the end of a campaign can be a key means to demonstrate success and may, in turn, serve to encourage the application of BI. As Figure 3.10 illustrates, the most commonly used evaluation metrics among CoGs and MHs include measures such as number of people reached and awareness levels.

Figure 3.10. Metrics used to evaluate communication activities by CoGs and MHs

The evaluation of behavioural change has been particularly prevalent in the health sector and is increasingly acknowledged since the onset of the COVID-19 pandemic. According to OECD survey results, 11 out of 24 MHs in OECD and partner countries evaluated behaviour change in populations to measure the success or failure of a given campaign. For instance, the Public Health England campaign “Change4Life: sugar smart” evaluated changes in household consumption patterns resulting from the introduction of the Sugar Smart app and advertising across 750 supermarkets. In the context of COVID19, the Irish Department of Health has run weekly national surveys, focus groups and behavioural studies to better understand changes in attitudes, perception and media consumption patterns.

While the practice of BI entails rigorous experimental methods, OECD survey results indicate that only 6 CoGs and 2 MHs use experimental methods to evaluate the results of campaigns. The most popular methods employed by 22 CoGs and 10 MHs are surveys, whose limitations for evaluating behaviour change have been discussed in the previous sections. The evaluation of BI requires highly technical expertise, including the ability to plan and run rigorous evaluations, use relevant measures of behaviours, apply counterfactuals, and adapt the test methodology according to available data, resources and timeframe. Evaluating changes in behaviour can therefore be considered a discipline, one in which the synthesis and publication of results allows for the accumulation of experiences to build the necessary capabilities for public communicators to draw on post hoc insights in the early steps of future campaigns.

Towards a systemic use and dissemination of evidence in public communication

This chapter illustrated the strategic value of evidence-driven communication. It explored the utility of audience insights for improving the targeting of communication campaigns as well as its potential to shift to a two-way model of communication. It also provided an analysis of the critical role of behavioural insights in informing the design, delivery and evaluation of citizen-centric communication.

However, public communication and the integration of evidence do not happen in a vacuum. In this regard, the opportunities and challenges for employing insights more efficiently and effectively in the broader data ecosystem in which public communicators must operate merit further reflection. While not exhaustive, the following section reflects on issues for future research in this field, including:

ensuring sound data governance models to foster value creation within and beyond government
leveraging emerging technologies to build social listening capabilities and facilitate the collection and analysis of insights
addressing legal and privacy concerns through the ethical management and use of data.

First, the data-intensive character of the public communication profession raises important questions about data governance. Notably, growing volumes of information in an environment where institutional “legacy challenges”4 remain unsolved is inhibiting the ability of public sector institutions to access, share and extract value from data (OECD, 2019[20]). Barriers include lack of incentives, standards, and interoperable systems for storing and processing data (OECD, 2019[21]). Efforts to tap into the strategic value of data for public communication could be accompanied by a reflection on the role of data quality principles, the sharing of protocols and the establishment of relevant training programmes. In this regard, the OECD Recommendation of the Council on Enhancing Access to and Sharing of Data sets out general principles that could further guide related conversations (see Box 3.6).

Box 3.6. OECD Recommendation of the Council on Enhancing Access to and Sharing of Data

The OECD Recommendation provides a series of principles on how governments can maximise the benefits of enhancing data access and sharing arrangements while protecting individuals’ and organisations’ rights and accounting for other legitimate interests and objectives. The principles include:

Establish and implement policy measures to enhance data access and sharing alongside broader activities such as engaging with and empowering relevant stakeholders to increase the trustworthiness of the data ecosystem. This could be done by promoting transparency, inclusivity and through the encouragement of data sharing across and between public and private sectors.
Adopt a whole-of-government approach to data access and sharing to ensure that specific legal, societal and policy engagements are met. This could be done through the adoption of arrangements that achieve these aims, the implementation of flexible and scalable governance frameworks through fostering regulatory environments supporting access and sharing.
Promote and enable a culture of responsibility for data governance by taking into account the rights and interests of different stakeholders, including those in the private and civil spheres.
Provide incentive mechanisms as well as develop sustainable business models and markets for data access and sharing. The recommendations outline, for instance, the promotion of self- and co- regulation mechanisms to achieve these aims.
Improve conditions for cross-border data access and sharing through reducing restrictions and promoting international co-operation. The Recommendations also state that conditions upon cross-border access and sharing should be non-discriminatory, transparent, necessary and proportionate to the level of risk.
Ensure that data can be found as well as re-used across and within organisations by providing associated documentation in a timely, transparent manner. The Recommendations also promote the adoption of standards for data models and formats.
Educate stakeholders to enable all parties to use data responsibly. Namely, the OECD recommends enhancing data-related skills and competencies needed, including by public servants, to harness the benefits of data access, sharing, and use.

Source: OECD (forthcoming[5]), Principles of Good Practice for Public Communication Responses to Help Counter Mis- and Disinformation, OECD Publishing, Paris..

The role of public communicators in supporting the effective dissemination of government datasets to promote their use and re-use remains an area to be further explored. In fact, while there are established communication channels with the private sector, the OECD Survey on Open Government Data (OGD) found that only 18 out of 33 countries consider civil society and journalists as priority communities to engage with on OGD initiatives and/or policies (OECD, 2018[22]). Similarly, only 14 out of 33 countries were found to have concrete communication strategies to raise awareness on OGD, its benefits and existing datasets (OECD, 2018[22]). As discussed in Chapter 5, these findings highlight opportunities for public communication and digital government units to co-ordinate in promoting the effective dissemination of data across and beyond the public sector.

Second, the rise of emerging technologies offers multiple opportunities to ground public communication in evidence. The use of chat bots is a common example. In Brazil, the Secretariat of Social Communication (SECOM) is employing intelligent machine learning processes to conduct sentiment analysis, monitor the effects of messages, and identify information gaps that may require refocusing content (see Chapter 4 for more details). In the United States, the Centre for Disease Control has similarly developed advanced social listening tools that triangulate diverse sources of data on public discourse and media in relations to COVID-19 and vaccination (see Box 3.1).

Third, public communicators increasingly face ethical dilemmas related to the use of insights in light of the growing reliance on personal information and artificially intelligent technologies. On the one hand, data privacy concerns emerge over how population data is gathered, used and reported. On the other, the programming of machine learning and natural language processors inherently reflect biases that may skew the collection and interpretation of data from different population groups, in particular among underrepresented segments of society. Some countries such as the United Kingdom have begun to disseminate general guidelines for data ethics to support the work of communicators (see Box 3.7). The OECD has also developed a set of principles, which could be of use for public communicators to reflect on the value and practical implications of data ethics (see Box 3.8).

Box 3.7. Data ethics framework in the United Kingdom

In recognising the important role of public communicators, the UK Government established a set of guidelines defining the scope of ethics in governmental communication, together with a dedicated Data Ethics Framework as well as a guide to using artificial intelligence.

The UK Government’s Data Ethics Framework builds upon the overarching principles of transparency, accountability and fairness. This particular document guides the responsible use of data in government and the wider public sector by enabling civil servants to understand ethical issues as well as handle them in the context of their work. The framework is designed to be revisited frequently throughout the different stages of a project from data collection, to the sharing of information. It includes a self-assessment table for review at each stage of the process to empower communicators and other government employees to evaluate progress.

As Artificial Intelligence is increasingly used in communication, the Government set a framework guiding its use and creation within the public sector. The joint guidance developed by the Government Digital Service (GDS) and the Office for Artificial Intelligence (OAI) covers three broad issues, including: to assess whether AI can help meet user needs, the public sector can leverage AI and to use AI ethically, fairly as well as safely. It encourages civil servants to build upon the ethical values of

1. Respecting the dignity of individuals.
2. Connect with each other sincerely, openly and inclusively.
3. Care for the well-being of all.
4. Protect the priorities of social values, justice, and public interest.

Source : https://www.gov.uk/government/publications/data-ethics-framework

Box 3.8. OECD Good Practice Principles for Data Ethics in the Public Sector

The OECD Good Practice Principles for Data Ethics in the Public Sector outline the value and practical implications of data ethics in government. These principles call on governments to:

1. manage data with integrity
2. be aware of and observe relevant government-wide arrangements for trustworthy data access, sharing and use
3. incorporate data ethical considerations into governmental, organisational and public sector decision-making processes
4. monitor and retain control over data inputs, in particular those used to inform the development and training of AI systems, and adopt a risk-based approach to the automation of decisions
5. be specific about the purpose of data use, especially in the case of personal data
6. define boundaries for data access, sharing and use
7. be clear, inclusive and open
8. publish open data and source code
9. broaden individuals’ and collectives’ control over their data
10. be accountable and proactive in managing risks.

Source: OECD (2021[23]), Good Practice Principles for Data Ethics in the Public Sector, OECD Publishing, Paris.

Key findings and way forward

For public communication to be deemed strategic, it needs to be informed by evidence, for example in the form of audience insights research, social listening or behavioural insights (BI). While a majority of countries make use of evidence to inform the design and delivery of public communication, there remains scope to collect, employ and disseminate insights on audiences, behaviour change and uptake of services in more systematic and strategic ways – from the planning to the ex post evaluation phases.
Audience insights provide communicators with a real-time understanding of public concerns and sentiments. Beyond simple demographic traits, understanding the habits, attitudes and information consumption patterns from different segments of society is key to making communication more inclusive, especially for underrepresented or disengaged groups.
Survey data revealed room for governments to more systematically embed audience insights into the planning, design and delivery of communication activities, given that a majority of CoGs and MHs state that they conduct this type of research on an ad hoc basis. Similarly, greater audience segmentation and diversification of content across channels and target groups that is based on audience insights can contribute to more impactful, communication.
Tapping into more sophisticated social listening capabilities by evolving the gathering of insights is the next frontier for promoting a two-way dialogue with citizens and making use of public feedback to improve policy making. Further research into the different types of insights and collection methods could help build a state of the art understanding and model for this field.
Emerging technologies have opened new possibilities for public communicators to gather and analyse evidence to inform communication activities. For example, big data, cloud computing, smart algorithms and analytical softwares have unlocked a vast trove of insights and diminished the cost of acquiring and processing information about audiences. Further research into existing tools could help build an understanding of the benefits and potential limitations in building stronger social listening capabilities, in particular those which may raise ethical, privacy and security concerns.
Behavioural insights provide key evidence on cognitive factors and biases that can enable communication to be more responsive and effective in reaching citizens amid competition for their attention in a crowded media ecosystem. Tapping into behavioural factors can help deploy communications that encourage desired actions in line with key policy goals.
Efforts toward strengthening institutional capacities and ensuring expertise are available to collect and embed BI in a scientific way could aid countries in reaching more effective communications.
Moving beyond a siloed collection and management of different types of behavioural data and audience insights could help ensure they are more widely used across public institutions and for relevant campaigns. Governments should reflect on data ethics and data governance arrangements to promote a whole-of government culture of evidence, avoid duplications and reduce costs.

References

[17] Bavel, J. et al. (2020), “Using social and behavioural science to support COVID-19 pandemic response”, Nature Human Behaviour, Vol. 4/5, http://dx.doi.org/10.1038/s41562-020-0884-z.

[15] Brennan, L. and W. Binney (2010), “Fear, guilt, and shame appeals in social marketing”, Journal of Business Research, Vol. 63/2, pp. 140-146, http://dx.doi.org/10.1016/j.jbusres.2009.02.006.

[2] CDC (2018), “Zombie Talk and Preparedness”, Center for Disease Control and Prevention, https://emergency.cdc.gov/cerc/cerccorner/article_050918.asp (accessed on 14 September 2021).

[8] Dryzek, J., R. Norgaard and D. Schlosberg (eds.) (2011), The Oxford Handbook of Climate Change and Society, Oxford University Press, http://dx.doi.org/10.1093/oxfordhb/9780199566600.001.0001.

[16] Durantini, M. et al. (2006), “Conceptualizing the influence of social agents of behavior change: A meta-analysis of the effectiveness of HIV-prevention interventionists for different groups.”, Psychological Bulletin, Vol. 132/2, pp. 212-248, http://dx.doi.org/10.1037/0033-2909.132.2.212.

[9] Eom, K., H. Kim and D. Sherman (2018), “Social class, control, and action: Socioeconomic status differences in antecedents of support for pro-environmental action”, Journal of Experimental Social Psychology, Vol. 77, pp. 60-75, http://dx.doi.org/10.1016/j.jesp.2018.03.009.

[11] Kollmuss, A. and J. Agyeman (2002), “Mind the Gap: Why do people act environmentally and what are the barriers to pro-environmental behavior?”, Environmental Education Research, Vol. 8/3, pp. 239-260, http://dx.doi.org/10.1080/13504620220145401.

[12] Kormos, C. and R. Gifford (2014), “The validity of self-report measures of proenvironmental behavior: A meta-analytic review”, Journal of Environmental Psychology, Vol. 40, pp. 359-371, http://dx.doi.org/10.1016/j.jenvp.2014.09.003.

[4] Macnamara, J. (2017), Creating a “Democracy for Everyone”: Strategies for increasing listening and engagement by government, The London School of Economics and Political Science and University of Technology Sydney, London and Sydney, https://www.uts.edu.au/sites/default/files/FASS-Research-Report-Creating-a-New-Democracy-for-Everyone.pdf?no-cache (accessed on 12 April 2021).

[3] Macnamara, J. (2015), Creating an ‘architecture of listening’ in organizations: The basis of engagement, trust,healthy democracy, social equity, and business sustainability., https://www.uts.edu.au/sites/default/files/fass-organizational-listening-report.pdf.

[13] Michie, S., M. van Stralen and R. West (2011), “The behaviour change wheel: A new method for characterising and designing behaviour change interventions”, Implementation Science, Vol. 6/1, http://dx.doi.org/10.1186/1748-5908-6-42.

[18] OECD (2021), Enhancing public trust in COVID-19 vaccination: The role of governments, OECD Publishing, Paris, https://dx.doi.org/10.1787/eae0ec5a-en.

[23] OECD (2021), Good Practice Principles for Data Ethics in the Public Sector, OECD, Paris, https://www.oecd.org/gov/digital-government/good-practice-principles-for-data-ethics-in-the-public-sector.pdf (accessed on 8 April 2021).

[10] OECD (2020), Regulatory policy and COVID-19: Behavioural insights for fast-paced decision making, OECD Publishing, Paris, https://dx.doi.org/10.1787/7a521805-en.

[1] OECD (2020), “Transparency, communication and trust: The role of public communication in responding to the wave of disinformation about the new Coronavirus”, OECD Policy Responses to Coronavirus (COVID-19), OECD Publishing, Paris, https://doi.org/10.1787/bef7ad6e-en.

[21] OECD (2019), Enhancing Access to and Sharing of Data: Reconciling Risks and Benefits for Data Re-use across Societies, OECD Publishing, Paris, https://dx.doi.org/10.1787/276aaca8-en.

[20] OECD (2019), The Path to Becoming a Data-Driven Public Sector, OECD Digital Government Studies, OECD Publishing, Paris, https://dx.doi.org/10.1787/059814a7-en.

[7] OECD (2019), Tools and Ethics for Applied Behavioural Insights: The BASIC Toolkit, OECD Publishing, Paris, https://dx.doi.org/10.1787/9ea76a8f-en.

[22] OECD (2018), Open Government Data Report: Enhancing Policy Maturity for Sustainable Impact, OECD Digital Government Studies, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264305847-en.

[6] OECD (2017), Behavioural Insights and Public Policy: Lessons from Around the World, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264270480-en.

[5] OECD (forthcoming), Principles of Good Practice for Public Communication Responses to Help Counter Mis- and Disinformation, OECD Publishing, Paris.

[19] The Behavioural Insights Team (2021), How can we use behavioural science to promote vaccination against COVID-19 in Latin America?, https://www.bi.team/blogs/how-can-we-use-behavioural-science-to-promote-vaccination-against-covid-19-in-latin-america/.

[14] The Behavioural Insights Team (2014), EAST: Four Simple Ways to Apply Behavioural Insights, https://www.behaviouralinsights.co.uk/publications/east-four-simple-ways-to-apply-behavioural-insights/ (accessed on 13 May 2021).

Notes

← 1. The OECD survey on which this report is based was administered in 2020 to cover the year 2019. Although the responses refer to the pre-COVID-19 era, several respondents have reflected the experience of the pandemic in some of their answers.

← 2. N= 35 CoGs and 19 MHs that claimed to use audience insights to inform its communication planning (question 9).

← 3. A/B testing refers to randomised experiments to compare two versions of a single piece of communication (i.e. message, visual, slogan) to determine which one is more effective.

← 4. Legacy challenges include: “outdated data infrastructures and data silos to skill gaps, regulatory barriers, the lack of leadership and accountability, and an organisational culture which is not prone to digital innovation and change” (OECD, 2019[20]).

╳

Metadata, Legal and Rights

This document, as well as any data and map included herein, are without prejudice to the status of or sovereignty over any territory, to the delimitation of international frontiers and boundaries and to the name of any territory, city or area. Extracts from publications may be subject to additional disclaimers, which are set out in the complete version of the publication, available at the link provided.

https://doi.org/10.1787/22f8031c-en

The use of this work, whether digital or print, is governed by the Terms and Conditions to be found at http://www.oecd.org/termsandconditions.