Trustworthiness: Public reactions to COVID-19 crisis communication

1 Introduction

In the wake of the global COVID-19 pandemic, questions about how citizens assessed the trustworthiness of their governments’ information and regulations have become essential (Offerdal et al., 2021). In the Scandinavian countries, known as “high-trust societies” long before the pandemic (see e.g., Andreasson, 2017), a majority declared that they were satisfied with the response of their governments, and expressed high levels of trust in the health authorities during the first stages of the pandemic (Nielsen and Lindvall, 2021). This was the case in all three Scandinavian countries—Denmark, Norway, and Sweden—even though the respective governments in the early stage of the COVID 19-crisis, Spring 2020, chose different strategies to curb the pandemic. Whereas the governments in Norway and Denmark implemented lockdowns and strict regulations on citizens’ movements, Sweden relied on recommendations and remained open (Nord, 2022).

To contextualize the study, the Scandinavian countries—Denmark, Norway, and Sweden—are small, stable, welfare states that were led by minority coalition governments in 2020: Social Democratic Prime Minister (PM) Mette Frederiksen in Denmark, Conservative PM Erna Solberg in Norway, and Social Democrat PM Stefan Löfven in Sweden. The Danish Cabinet and PM Mette Frederiksen fronted the crisis management through TV speeches, press conferences, and decision-making processes. A similar role was taken by the Norwegian government where the PM and the Minister of Health led the crisis management, advised and aided by expert institutions that did not always agree (Christensen and Laegreid, 2020; NOU 2021, p. 6). While Denmark and Norway implemented strict regulatory measures such as lockdowns and closing of schools and kindergartens from mid-March to April 2020, Sweden used an advisory strategy relying on voluntary compliance (Anderberg, 2021; Claeson and Hanson, 2021). In Sweden, the pandemic crisis management was—by law—delegated from the Cabinet to the Public Health Agency in Sweden (PHAS), and accordingly, the expert institution had a more visible role than the PM (Andersson and Aylott, 2020; Pierre, 2020). The different strategies led to radically different numbers of deaths and infections in Sweden compared to its Nordic neighbours in the first two years of the pandemic. As of February 21, 2022, the death tolls related to COVID-19 were approximately 1,656 per million in Sweden (16,897 deaths, population 10.2 million), compared to 669 per million in Denmark (3,884 deaths, population of 5.8 million), and 281 per million in Norway (1,549 deaths, population of 5.5 million)^[1]. Despite national differences, trust in the national political and health authorities remained remarkably high and stable in all three countries (Organisation for Economic Co-operation and Development, 2022).

Recent research has pointed to the “rally around the flag” hypothesis as an explanation of the tendency of citizens to support their governments in times of crisis (Esaiasson et al., 2021; Schraff, 2021). Such effects were far from being global phenomena during the COVID-19 pandemic (Kritzinger et al., 2021; Van Aelst and Blumler, 2021); however, in Scandinavia, there is substantial quantitative evidence that governments could rely on their citizens to comply with measures to curb the virus (Johansson et al., 2023). Likewise, survey studies show that not only health authorities, but also nationwide news media gained from rallying effects around the national media institutions (Knudsen et al., 2023). We take these quantitative findings as a starting point for this qualitative study that investigates how members of the public assess and negotiate the trustworthiness of health authorities. We seek to unpack why members of the public decide to trust and comply with the crisis measures imposed by authorities. By doing this, our study contributes to understanding the reasons why citizens “rally around the flag.” Building on insights of theories and studies of risk communication, rhetoric, and authority performances, we have formulated the research question as follows:

RQ1: How did members of the public assess the trustworthiness of health authorities?

RQ2: How did members of the public discuss and negotiate reasons for complying (or not) with policies and measures?

Below, we discuss the theoretical perspectives and previous studies we draw on, followed by a methods section that presents our data material and the methods we have used to analyze the focus group data.

2 Trustworthiness and trust in public health authorities

In times of public health crises, such as the COVID-19 pandemic which forms the context of this study, securing the public’s trust in health authorities is considered essential for securing compliance with strict measures to curb the infection (Liu and Mehta, 2021; Majid et al., 2021; Johansson et al., 2023). Yet, in the literature, there are many definitions of trust. Sztompka (1999, p. 25) defines trust as “a bet about the future contingent actions of others.” Trust has also been seen as based on shared moral values with others resulting in the belief that the others will not take advantage of us (Uslaner, 2002). Rational choice theory sees trust as a form of encapsulated interest —“the potentially trusted person has an interest in maintaining a relationship with the trustor, an interest that gives the potentially trusted person an incentive to be trustworthy” (Hardin, 2006, p. 17). Here, we rely on a definition of trust from organizational research: “the willingness of a party to be vulnerable to the actions of another party based on the expectation that the other will perform a particular action important to the trustor, irrespective of the ability to monitor or control that other party” (Mayer et al., 1995, pp. 712). Organizational research commonly points to two antecedents of trust: trusting dispositions, and trustworthiness (Baer and Colquitt, 2018; Mayer et al., 1995; Rousseau et al., 1998). The former is a stable tendency to ascribe good intentions to other people, and, thus, a reliance of the trustor on what is said and done by others. It also implies how such dispositions might help explain trust in many situations, but also that trust might differ related to different topics and situations (Baer and Colquitt, 2018). Organizational scholars frequently discuss trustworthiness as a construct with the following three elements (Haynes et al., 2012; Mayer et al., 1995): ability (the trustor believes that the trustee has the needed knowledge, skills, expertise, and competence to perform x’ in a specific domain or to tackle some specific tasks) (van Dijck and Alinejad, 2020); integrity (the trustor believes that the trustee adheres to a set of values shared or accepted by the trustor, for instance, that there will be consistency between word and deed); and benevolence (the trustor believes that the trustee cares for the wellbeing of the trustor for their own sake, rather than the self-interest of the trustee). Baer and Colquitt (2018, p. 170) concluded that not only are these facets strongly correlated to trust, but trust is also “primarily and essentially a function of perceived trustworthiness.” The three elements found to increase trustworthiness are similar to the Aristotelian rhetorical tradition of seeing competence (ability), character (integrity), and goodwill (benevolence) as ways to strengthening ethos (Aristotle, ca. 340 B. C. E./2007; Fiskvik et al., 2023).

Previous studies of health communication during COVID-19 have shown that all three elements have been described as important. Van Dijck and Alinejad (2020, p. 2) singled out ability as crucial for the perceived trustworthiness of experts, whereas other studies have pointed to integrity, drawing attention to the perceived difference in trustworthiness between experts and politicians (Hendriks et al., 2023; Janssen et al., 2021). Public health authorities are often perceived as more trustworthy, presumably lacking a political agenda (Quinn et al., 2013). Another recent study pointed to how factors such as personal contact with experts, as well as the perceived independence from political elites, might increase perceptions of trustworthiness (Mihelj et al., 2022).

Walls et al. (2004) employed focus group data to develop the concept of critical trust, portraying trust as “provisionally conceptualized as multifaceted, potentially dynamic, and dependent upon a range of contextual variables” (p. 135). Their findings suggested that “people employ a range of rationales when assessing government” (p. 146), negotiating scepticism and reliance when they assessed the trustworthiness of public institutions for health and risk management. Many contextual factors played into the evaluations, and only rarely did the authors note expressions of naïve or blind trust, or the opposite, outright distrust among the participants (Walls et al., 2004).

Most citizens will gain information about political issues from different news and social media, and during the pandemic, when access to most social settings outside of the family were non-existent or limited, news media became overly important both as sources of information and manufacturers of trust (Knudsen et al., 2023). Koivunen and Vuorelma (2022, p. 394), offering a nuanced view of how trustworthiness is constructed in mediatized and networked contexts, argue that “instead of approaching trust through indexes that examine the level of citizens’ trust towards institutions, we treat it as a key element in authority performances.” The point in their reasoning is that the manufacturing of trust requires authorities to act and respond, but also to pay attention to the mediatized manufacturing of trust: “The media does not only report on and analyze authority performances of political actors but also themselves enact authority performances, which include manufacturing trust” (Koivunen and Vuorelma, 2022, p. 395). Authority performances were abundant in all phases of the COVID-19 pandemic in the Scandinavian countries, regularly enacted through governments’ press conferences and news reporting about crisis management. In a related but not similar vein, Bjørkdal and collegues (2021) argue that the pandemic could be seen as a set of rhetorical situations—a question inviting a response—and analyze the responses from the heads of state, the Danish, Norwegian and Swedish Prime Ministers (PMs) in the first stages of the crisis when they communicated to the people in daily press conferences broadcast in radio, TV, and online. The perspectives outlined above will be applied in the analysis of the data in the sections below.

3 Method and data

We used focus group interviews to research how group participants negotiated these dimensions and highlight the contingent character of trustworthiness, contributing to a nuanced in-depth understanding of how trustworthiness is situational (Delia, 1976; O’Keefe, 2016). Unlike survey data, our data does not yield statistically generalizable data, yet we aim for conceptual generalizability by employing an inductive, reflexive analytical approach (Polit and Beck, 2010; Tjora, 2019).

To collect data, we procured the services of the market analysis company Opinion to organize 12 focus groups across Scandinavia, four in each country, in May 2020. Opinion operates across the Scandinavian countries, and the company recruited participants from their pre-existing database according to the specifications of the project. Participants were rewarded with a gift card (DKK/NOK/SEK 500, approximately €50). Originally planned as physical focus groups, owing to the scale of the pandemic, they were turned into virtual groups and the conversations were online live chats. The chats were synchronous in text format with no visual communication.

Each group consisted of six to nine participants from four different age groups: Young without children (18–30); Families with small children (30–45); “Empty nesters,” i. e., persons without small children in the household (46–65); and “Seniors” aged 65 or older. The number of participants per group was set to secure that each group would allow for a certain diversity of opinions and beliefs, and, at the same time, allow enough time and space for all participants to actively take part in the conversation. Within the groups, there were variations between participants concerning jobs, state of employment, education, class, and cultural background. The selection of participants also considered that people had experienced the lockdowns and other measures differently, varying with life and job situations. For example, students were locked out from studies and part-time jobs and reported experiencing economic insecurity, whereas health and care workers had to work in high risk-environments.

Each live chat lasted approximately two hours and was led by professional moderators, one in each country, speaking Danish, Norwegian, and Swedish respectively. They opened the sessions by inviting the participants to present themselves and talk about how they initially heard about a global pandemic and how they had reacted to this information, explicitly noting that participants should not share confidential information leaving them vulnerable to harm. In the transcribed chats that were submitted to the research team, the participants were anonymized, except for first names, to fulfil research ethics requirements concerning privacy and anonymity.

The interviews were guided by theory-based and detailed interview guides, translated, and adapted to the different national contexts. The common parts related to questions guiding the participants’ experiences, whereas the interview guides used specific Danish, Norwegian, and Swedish examples when asking about political leaders and governments, public health institutions, media reporting and measures for curbing the pandemic. Addressing aspects of trust and trustworthiness, the questions probed topics such as experiences of the pandemic, views on the public health information and crisis communication of political leaders and public health agencies, and experiences and views on the measures to combat the pandemic. They were asked about their personal experiences with the restrictions and recommendations set out by the health authorities to curb the pandemic, whether they found the measures reasonable, and whether and why they did or did not trust the representatives of the health authorities. The interview guides were designed to entice views, discussions, and negotiations on the credibility and efficiency of measures and the participants’ compliance—or lack thereof. In particular, the participants were invited to discuss and reason about how and why they assessed these actors as competent, displaying moral character, or showing goodwill. The moderators encouraged and asked participants to engage and discuss with each other and give reasons for their views.

The participants were also presented with one or two news stories and opinion pieces from major news media in each country: Dagbladet (N), Jyllands-Posten (DK), TV2 (DK), Svenska Dagbladet (S), and Aftonbladet (S). The pieces were selected as examples of the current public debates on the measures taken in each of the countries. The stories for Norway discussed criticism generated by the re-opening of schools and kindergartens; the pieces for Denmark illustrated disagreements on the strictness of the measures and sanctions for breaking them; and the items for Sweden showed the intense debate on the Swedish strategy of recommending rather than regulating and sanctioning public behavior. The media items were included to stimulate debate and discussions within the groups. The news stories’ veracity and the relevance of the opinion pieces were checked and discussed within the cross-national research team prior to being included as examples in the interview guides. The interview guide and stimuli are available on request.

The transcripts of the online chats were coded qualitatively in NVivo using a stepwise inductive-deductive approach (SDA; Tjora, 2019) involving many rounds of reading and analysis. Using an inductive-deductive approach means that the codes were initially inductively drawn from the transcripts of the chats in the focus groups. SDA has many similarities with a grounded theory approach (e.g., Corbin and Strauss, 2008), but is more deductive when drawing on previous theoretical insights in the analysis.

The discussions in the focus groups were not identical, and revealed different views and concerns related to the trustworthiness of political leaders and public health institutions in general as well as their spokespersons. The different group dynamics sometimes led to some aspects taking precedence over others, reflecting both different demographics, experiences of the participants, and national contexts. In the analysis of the transcripts, we delved into the expressions and negotiations between the focus group participants on the trustworthiness of health authorities concerning how they had experienced the first two months of the pandemic.

4 Results and analysis

We set out with two research questions asking how members of the public assessed the trustworthiness of health authorities (RQ1 and how they discussed and negotiated reasons for trusting and/or complying—or not—with the national COVID-19 measures (RQ2). The interviews and discussions show that the participants assessed and had various reasons for trusting and not trusting public health authorities. Most participants formed their impressions watching and listening to televised press conferences and media reporting; only a few mentioned that they also had access to relevant information through their jobs or family.

When asked a set of questions seeking to understand how participants assessed the trustworthiness and validity of the information and measures issued by health authorities, they replied by referring to several lines of reasoning based on their perception of political leaders and the spokespersons of public health agencies.

5 Discussion

In line with other studies, we find that focus group participants in Denmark, Norway, and Sweden in May 2020 assessed the trustworthiness of public health authorities according to the well-established concepts of ability, goodwill and integrity (Baer and Colquitt, 2018; Haynes et al., 2012). As expected, the ancient insights of how trustworthiness—ethos—is built through exposition of ability, integrity and benevolence go a long way to understand and interpret the replies and negotiations found in the focus group data. All three rhetorical components were pointed to by participants who found health authorities trustworthy. The political leaders, the governments, were commended for their ability to act under extreme and risky circumstances and for listening to and applying the knowledge of the experts, in line with what Hendriks et al. (2023) and Janssen et al. (2021) also found. The public health experts were seen as trustworthy for providing unbiased and independent expertise, as Milhelj et al. (2022) also pointed out, even in situations of uncertainty and scarcity of both knowledge and resources.

As shown above, there was no consensus either within the groups or between them that the measures chosen by their individual governments were the “best” or most “correct” measures for curbing the virus. Participants frequently stated that they did not trust authorities’ abilities, as shown when they expressed doubts about the national strategies, and which one would be the best in the long run. They often doubted that authorities, be they experts or not, had adequate competence to assess the effects of the measures they implemented (Denmark and Norway) or recommended (Sweden). Still, they chose to trust their goodwill and comply with regulations because of the high risks for themselves, their families, communities, and society connected to not doing so, or because of loyalty to the political system. The conversations and negotiations in the focus groups showed that the trustworthiness of authorities was contested, divided, and negotiated, particularly between the expert agencies and the political leaders. The participants distinguished between different types of competencies, the scientific expertise of the public health agencies, and the political competence of governments and ministers. Accordingly, our findings relate closely to the previous study of Walls et al. (2004), who coined the concept of critical trust reflecting that members of the public have a variety of reasons for trusting authorities, and that they weigh and balance different concerns against each other. Focus group participants expressed critical and, we may add, “conditional trust” towards political leaders and health experts. In their own words, they chose to trust authorities despite not believing that they necessarily implemented the best measures. They assessed some individual experts as more trustworthy than others on the basis of how their were perceived in the news and press conferences, echoing Koivunen and Vuorelma’s (2022) point that in the networked society, “authority performances” are integral to the production and maintenance of trust. Studies have analyzed how the press conferences became the main arenas both for informing the public and exercising authority (Young, 2022). The form and content were not similar in the three countries (Bjørkdal et al., 2021), yet this study adds to the understanding of how the public perceived and interpreted how public health authorities exerted leadership and expertise through the press conferences. Authority performances such as the press conferences and media reporting on authorities managing the crisis were situations in which trustworthiness and authority seemingly were produced and reproduced. As found in many previous studies (Cvetkovich and Lofstedt, 1999; Frederiksen, 2014; Gillman et al., 2023), displaying uncertainty and risk about the situation did not seem to weaken the trustworthiness of either experts or governments. Rather, openness and transparency play into building trustworthiness (Ihlen et al., 2022).

Returning to our starting point, the unpacking of the “rally around the flag” hypothesis, that citizens will support their governments in states of crisis, our findings shed light on the reasons why citizens choose to be loyal. As pointed to above, citizens’ trust is both critical (Walls et al., 2004) and conditional. The COVID-19-crisis displayed many national contexts of low trust in public health authorities, e.g. the U. K., where Baker and Lilleker (2022) pointed to scandals and rule-breaking among the elites as reasons for mistrust. Our findings show that the public perceived public authorities as conditionally trustworthy, that is, those who chose to trust authorities did so despite uncertainty. Equally important, system loyalty weighed in as a separate and strong reason for compliance with regulations and recommendations. Even those who did not trust that the authorities had implemented the correct measures, and those not supporting the sitting government, argued in favor of complying with the COVID-19 measures.

Obviously, our findings do not, as they are qualitative, contribute to support or weaken the hypothesis. Rather, they add insights into what reasons may explain support or lack of support and contribute to refining the hypothesis. The rallying hypothesis brings to mind an all too simplified image of citizens blindly supporting their leaders in times of crisis, as Van Aelst and Blumler (2021) already questioned. Our study suggests that rallying effects should also be seen as conditional and linked to system adherence. As stated initially, the Scandinavian countries were high-trust democracies before the pandemic, and this context is likely to be an important explanation for the willingness of most respondents to comply with recommendations and regulations. Citizens expressed few reasons for distrust, but they were neither uncritical nor expressing blind loyalty. They simply seemed, individually and collectively, to agree that in the risky and insecure situation they experienced, they chose to trust the experts for expertise and the political authorities for balancing the choices of protecting and saving lives against the harms of shattered economies and other social problems.

6 Conclusion

The focus groups data were collected two months after the three Scandinavian governments declared the COVID-19 pandemic national crises, and the analyses generated two distinct insights. First, although we acknowledge that “rally around the flag” effects will often be present in crisis situations, this study suggests that the explanations for why they arise are not that citizens look for strong leaders or simple answers. Rather, citizens choose to believe in weakly founded and insecure knowledge because the consequences of not doing so imply major risks for both individual citizens and the social order. Second, the trustworthiness of experts and leaders was continually negotiated, manufactured, and renegotiated, producing critical (Walls et al., 2004) and conditional trust. Expressions of trust in spokespersons and leaders were not distinguished by national borders but by the credibility of messages and the authority performances of the leaders. The data from three countries tell a story of similarities between citizens of states that selected—at this stage—quite different strategies. Willingness to and reasons for complying with measures to curb the disease were expressed and justified by participants from Sweden, Norway, and Denmark, notwithstanding the national policies consisting of harsh regulations or mild pressure.