With time comes trust? The development of misinformation perceptions related to COVID-19 over a six-month period: Evidence from a five-wave panel survey study in the Netherlands

1 Introduction

The outbreak of COVID-19 resulted in a global health crisis from March 2020 onwards. Even though media dependency was high, information on the pandemic was not always accurate – as misleading information abounded especially online (see e. g., Brennen et al., 2021; Pennycook et al., 2020). Importantly, people also perceived that information surrounding them was false or misleading, which could have had severe ramifications for their information-seeking behaviors or even compliance (e. g., Bridgman et al., 2020). Not surprisingly, survey research across 40 nations found that more than half of all participants were concerned about misinformation spread through digital means (Newman et al., 2020). As concerns about misinformation can have real consequences for behaviors in crisis settings (e. g., Bridgman et al., 2020; Hameleers et al., 2021), this paper explores how beliefs related to the prevalence of misinformation developed over a six-month period, and whether these changes were associated with people’s information diets, trust, and selective exposure patterns.

In the context of concerns about the trustworthiness of information, the high-choice and fragmented digital information setting may motivate the active approach of alternative information sources that challenges expert knowledge and evidence disseminated by established information sources (Brennen et al., 2021). At the same time, the increasingly more relative status of facts and empirical evidence may relate to selective avoidance of established information (e. g., Van Aelst et al., 2017). Especially considering that online and social media channels may contain relatively high levels of misinformation, conspiracies, and counter-factual narratives (Bridgman et al., 2020; Gabarron et al., 2021), the avoidance of established media and the approach of social media content may cultivate misinformation perceptions over time. Yet, we know little about how different media diets correspond to over-time changes in perceived misinformation.

Even though believing that (some) information is misleading or inaccurate is justified at times (e. g., Yang et al., 2021), too distrusting or even cynical beliefs related to the omnipresence of false information may be disproportionate and not in line with the actual share of misinformation (Allen et al., 2020). Hence, although false information makes up less than five percent of people’s online media diets (e. g., Acerbi et al., 2022), mis- and disinformation are often discursively weaponized (e. g., Egelhofer and Lecheler, 2019) and salient as beliefs in public opinion (Hameleers et al., 2021). As exemplified by Jones-Jang et al.’s (2021) findings in the context of the 2018 U.S. midterm elections, perceptions of false information can strengthen political cynicism. As perceiving false information or constructing “fake news” labels have real consequences (e. g., Jones-Jang et al., 2021; Van Duyn and Collier, 2021), it is important to explore how mis- and disinformation perceptions may be related to the approach and avoidance of different information sources, as well as people’s trust in news.

To explore how misinformation perceptions changed across different landmark moments of the outbreak, and how these potential over-time changes corresponded with (digital) media preferences, we rely on a five-wave panel survey study conducted in the Netherlands. As a main contribution, we offer insights into the dynamics between media consumption and beliefs about misinformation. As perceptions of misinformation and exposure to misinformation can reinforce each other over time (Xu et al., 2022), believing that information is false and deceptive may lead to news avoidance, cynicism and a higher likelihood of encountering deceptive information in alternative channels. Although it is beyond the scope of our data collection to make strong causal inferences, we aim to offer first insights into the dynamics between media use and avoidance and misinformation perceptions in times of a highly salient infodemic.

2 Misinformation perceptions over time

Misinformation can be understood as factually inaccurate and misleading information in general (Wardle, 2017), or information found to be misleading or erroneous based on relevant empirical evidence and/or expert knowledge (Vraga and Bode, 2020). Here, we also regard it as an umbrella term that may refer to both unintentional untruthfulness (also referred to as misinformation) and the goal-directed creation and dissemination of dishonest information (also referred to as disinformation).

We do not focus on the dissemination of actual misinformation as an informational genre, but on beliefs related to the prevalence of false information about the pandemic. As such beliefs related to accuracy and honesty of information have real consequences for (health) behaviors (Hameleers et al., 2021), we consider it important to focus on the attitudinal component of misinformation. In addition, as believing that information is false may cultivate cynicism over time (Jones-Jang et al., 2021), it is important to consider how levels of misinformation beliefs evolve over the course of a salient health crisis, and how they may correspond to trust and information seeking.

Misinformation may especially be prominent in times of high uncertainty or a lack of expert consensus and empirical evidence (Pennycook et al., 2020). Hence, in times of the outbreak of a new disease, much information is unclear at the start. At the same time, experts experience pressure to inform the media and the public swiftly on new developments – even if their information is characterized by a high level of uncertainty. Health-related misinformation is thus likely to occur in the context of a pandemic (Drezde et al., 2016; Oyeyemi et al., 2014), although such information is not necessarily disseminated to intentionally mislead citizens.

Our conceptualization of misinformation perceptions moves beyond generic measures of media trust tapping the perception that the media are able to fulfill their role in society, for example, by informing them in an accurate and honest way. Based on the credibility literature that distinguishes between competence and trustworthiness (Hovland et al., 1953; McCroskey and Young, 1981), we more specifically assess credibility evaluations in terms of the extent to which information on COVID-19 is seen as accurate, based on expert knowledge, and honest in terms of the intentions of disseminators to inform citizens in a truthful manner. We argue that, in an information context where concerns about “Fake News” and false information are mounting, whereas the status of facts and evidence has become more relative (Van Aelst et al., 2017) or even politicized (Waisbord, 2018), we have to respond with measures of credibility that are sensitive to the socio-political reality of factual relativism.

In line with this approach to credibility beliefs during the pandemic, extant research shows that people are likely to associate COVID-19 information with misinformation in particular. Nielsen et al. (2020), for example, show that people believe that there is a high level of misinformation on COVID-19 coming from different sources, especially online. Although we do not directly ask people to report on levels of misinformation in this paper, across the different time periods included in our multi-wave study, news users were expected to associate novel information on the pandemic with misinformation, as it is surrounded by high levels of uncertainty and a lack of (expert) consensus. But how do these misinformation perceptions change as the pandemic progresses from a more to a less uncertain reality?

First, it can be expected that the more information on COVID-19 has been disseminated, the more likely people associate it with misinformation. In that sense, we can speak of a potential accumulation effect over time. Over time, people have encountered more stories that refute expert knowledge and empirical evidence, or introduce conspiracies and incorrect explanations (see e. g., Brennen et al., 2021; Pennycook et al., 2020) – which may cultivate misperceptions and beliefs in conspiracies among news users. As these stories became more prominent in people’s newsfeed over time, and as many fact-checking platforms and legacy media flagged these claims as misinformation, the public’s perception of untruthfulness may also increase over time. In other words, the more alternative factual claims and disputed narratives are available to news users and accumulate in their news diets, the stronger their perception of misinformation may become. Importantly, we do not argue that this accumulation is completely unjustified or not based on the actual accumulation of incorrect or misleading information – irrespective of the accuracy of people’s verdicts, we aim to explore how beliefs related to the prevalence of misinformation changed over time.

We can contrast this to a potential explanation of over-time certainty and consensus on the ongoing pandemic. Hence, while novel media events are surrounded by high levels of uncertainty and conflicting evidence (e. g., Boukes et al., 2021), (official) information on COVID-19, for example disseminated by the WHO, national governments, and independent scientists, became less conflicting and more certain as more knowledge was available after the initial phase of the outbreak. Over time, uncertainty decreased, whereas expert consensus and evidence expanded. This corresponds to the higher levels of trust in authorities and official sources as the pandemic progressed. In the Netherlands, for example, official sources and authorities were trusted more during the pandemic than before or at the start of the outbreak of COVID-19 (Newman et al., 2022). Although the first months of the outbreak may be associated with higher levels of false information due to a lack of expert agreement and verified evidence, the higher level of consensus and availability of empirical evidence at later stages might have lowered the perception that misinformation flourishes.

Based on these competing explanations of the development of misinformation beliefs over time, and the lack of extant research on misinformation perceptions related to erupting crises, we forward the following research question:

Again, we do not make any assumptions on the extent to which such beliefs are justified or disproportionate – but rather aim to explore how beliefs related to the prevalence of misinformation are dynamic and contingent upon changing circumstances.

3 The role of information exposure and avoidance

Misinformation beliefs may emphasize distrust in mainstream media channels as well as online and social media. Survey research has indicated that 56 % of citizens across 40 nations were concerned about misinformation spread via online channels (Newman et al., 2020). Other comparative findings indicate that people associate news media frequently with the dissemination of false or deceptive content (Hameleers et al., 2021). This indicates that – irrespective of whether these concerns are always justified – people are likely to associate certain media types with the dissemination of falsehoods. The consumption of information from traditional or social media may correspond with increased perceptions of misinformation. As information consumed in the media may both contain actual misinformation (misinformation as a genre) and misinformation as an accusation or salient issue (misinformation as a label, see Egelhofer and Lecheler, 2019), the consumption of information may cultivate the perception that misinformation abounds.

Here, we can again arrive at competing expectations for the direction of the influence of information consumption on perceived misinformation as the crisis developed: media exposure may either (1) lower misinformation perceptions because more accurate information and expert consensus was consumed over time, or may (2) amplify misinformation perceptions due to the increasing amount of uncertainty and conflicting evidence in the information people are exposed to. As it reaches beyond the scope of this endeavor to assess the level of certainty and veracity of the information people were exposed to, we have to use self-reported indicators of media use and avoidance.

Regarding the first explanation that hints at a reassuring role of increased media consumption, the approach of more information may imply that news users are capable of finding trustworthy sources of information that they can use to update their beliefs, and herewith lower perceptions of misinformation. Misinformation perceptions may, then, be reconciled by approaching more trustworthy and credible information that reassures distrust. We refer to this as a first-person media literacy mechanism in response to misinformation (e. g., Jang and Kim, 2018). More specifically, people who approach more information may feel confident and capable of solving the lack of expert knowledge and evidence on COVID-19. They may thus compensate for the omnipresence of misinformation by actively seeking out more content to verify sources and statements on the pandemic. When people consume more information about the pandemic, they may come across more expert knowledge and empirical evidence that can offer higher levels of certainty and verified knowledge. Moreover, news users may be exposed to fact-checks and other refutations of false information (Yang et al., 2021). Such exposure to information may lower the perception that misinformation is a salient issue as people can restore uncertainty and the lack of expert knowledge by seeking out more information in the media.

Alternatively, the second contrasting explanation – which we call the falsehood accumulation mechanism – implies that the consumption of more information results in the confrontation with more inaccurate content over time, as a high share of people’s information diets contains information that turned out to be untrue (Nielsen et al., 2020). As people select more news on COVID-19, they also come across more misinformation (Brennen et al., 2021; Pennycook et al., 2020), as well as factual inconsistencies and retractions of earlier falsehoods. In addition, consuming more information on the pandemic may also imply that people are confronted with accusations of false information or even Fake News accusations (e. g., Tagliabue et al., 2020).

Here, we distinguish between mainstream (offline) versus digital sources of information (social media). As misinformation has mainly been associated with online and social media sources (e. g., Nielsen et al., 2020), it could be argued that people who consume more information from social media are also exposed to more false information over time. Bridgman et al. (2020), for example, find that COVID-19 misinformation is more prevalent on social media compared to news media that mostly emphasize public health regulations. As a consequence, misperceptions regarding facts about COVID-19 are associated more with social media compared to regular news media exposure.

As conflicting information and false narratives on COVID-19 are more prevalent online than offline (e. g., Bridgman et al., 2020; Gabarron et al., 2021), people who mainly consumed online sources of information may have encountered more accusations of false information, more conflicting information from alternative sources, and more information that turned out to be untrue. As a consequence of this, the level of distrust in the honesty and veracity of information may be stronger in people who mainly consumed social media

Applied to beliefs related to misinformation, it can be expected that people consuming more news in contexts that contain high levels of false information as well as discussions about false information (i. e., Fake News accusations online, see e. g., Egelhofer and Lecheler, 2019) may also perceive misinformation to be a salient issue. Against the backdrop of these explanations on the relationship between media use and misinformation perceptions, we ask:

4 Selective exposure and avoidance related to misinformation perceptions

Beyond mapping the relationship between self-reported media exposure and misinformation perceptions, we look at the relationship between the intentional approach or avoidance of news and changes in misinformation perceptions. Overall, we expect that over-time increases in news avoidance correspond to increases of misinformation perceptions (also see Hameleers et al., 2021). Considering that the intentional avoidance of information is related to distrust and news media skepticism (Strömbäck et al., 2020), we expect that higher levels of news avoidance are also related to higher misinformation beliefs.

Arguably, most information that people encounter in the news is based on factual information and expert knowledge. As indicated by Acerbi et al. (2022), who build on various studies that have looked at the prevalence of false information in people’s information diets, the prevalence of misinformation in media diets is expected to be lower than five percent. However, distrusting news users that avoid established media whilst selectively exposing themselves to alternative outlets are found to have a stronger perception that the news media spreads Fake News and misleading information (e. g., Müller and Schulz, 2021). In addition, alternative media approached by people who avoid the news are likely to contain delegitimizing references to mainstream information and news (e. g., Albright, 2017). Considering that people who avoid news and information on COVID-19 may seek out alternative information and knowledge outside of the news media realm, the avoidance of news may relate to higher levels of perceived misinformation.

On the other hand, when people approach more information over time, they may do so because they trust the accuracy and veracity of such information. Established news are most likely to contain factually accurate information, and people who actively approach the news may not perceive that misinformation is likely to occur. Indeed, previous literature hints at a positive association between media use and trust (Tsfati and Ariely, 2014). Hence, the more people avoid the news, the more likely they are to perceive an increase in misinformation in the channels they avoid. Increased news approach patterns, on the other hand, may relate to more trustworthy perceptions over time, and lower beliefs about misinformation as people are more likely to be confronted with trustworthy information. We therefore hypothesize:

For active news approach, we expect the opposite pattern:

5 Trust and misinformation perceptions

As indicated by Merkley and Loewen (2021), anti-intellectualism, conceptualized as distrust of experts and intellectuals, corresponds with misperceptions about COVID-19. As people distrusting expert sources of knowledge may be more likely to believe alternative counter-factual narratives and conspiracies, they may be more vulnerable to misperceptions. In a similar vein, using a two-wave survey design, Jones-Jang et al. (2021) found that perceived exposure to mis- and disinformation enhanced political cynicism. Extending these findings, we aim to assess whether people with lower levels of trust in the authorities and the media are also more likely to hold beliefs about misinformation related to COVID-19.

Trust can be conceptualized as the evaluation that an actor, group or institution fulfills the expectations that people have about them (Baier, 1986). Political actors are, for example, expected to be honest and transparent about their decisions. In a similar vein, the media are supposed to inform the public in an accurate and honest way. In that sense, trust in media can be related to misinformation perceptions, and the competence and trustworthiness dimensions of credibility in particular (Hovland et al., 1953).

We forward the argument that when people negatively evaluate the competence and trustworthiness of institutions – such as the political elites and the mainstream media – they may also hold the belief that false information about the pandemic is omnipresent. In other words, the authorities were expected to report in a transparent and honest way about the coronavirus and its implications – and the news media played in important role in disseminating information from the authorities to the public (Merkley and Loewen, 2021). Hence, the authorities, the national government, and the national institution for disease control in particular, were important actors of information supply during the high demand for accurate information during the pandemic (Dunwoody, 2020). When people distrust these authorities, their corresponding evaluations of the veracity of the overall information supply may also be negative (i. e., misinformation beliefs).

This corresponds to the linkage between populist beliefs and anti-intellectualism and misinformation perceptions identified in previous research (e. g., Hameleers et al., 2021; Merkley and Loewen, 2021). The more people distrust elite sources of information communicating about COVID-19 (i. e., the mass media, and authorities such as the government or the national center for disease control), the more they may consider that these actors and other sources are spreading falsehoods. We therefore hypothesize:

6 Context

These hypotheses are tested in the context of the Netherlands. It should be emphasized that, overall, trust in the media is relatively high in this country. According to the 2020 Reuters Institute Digital News Report, 52 % of the Dutch population trusted the news most of the time during the pandemic (Newman et al., 2020). Trust levels were substantially higher in the Netherlands as compared to the US (29 %), the UK (28 %), and France (23 %), for example. The same report also indicates that Dutch citizens were the least concerned about their ability to distinguish between real and false news online (only 32 % of participants was concerned about misinformation and disinformation). In the US (67 %) and the UK (63 %), the level of concern was twice as high as compared to the Dutch setting. Based on these numbers, we should note that the Netherlands offers, relative to many other countries, a context in which citizens may be relatively resilient to misinformation due to their high trust in established news and high confidence in their ability to separate true from false news (Humprecht et al., 2020).

7 Method

8 Results

The development of COVID-19 misinformation perceptions over time

To answer RQ1, the COVID-19-specific misinformation perceptions across the five waves spanning a six-month period are presented in Figure 1. The findings clearly show an increase in misinformation beliefs as the pandemic progressed. Time positively affected misinformation beliefs, and people became more likely to perceive false and/or misleading information as prevalent across the five panel waves. In addition, a panel model with fixed effects explaining misinformation perceptions was run. This panel structure controls for unobserved time-invariant factors. Model 1 in Table 1 summarizes the effects of wave dummies on misinformation perceptions. The positive association of all dummies signals again the significant rise of misinformation perceptions over time. To answer RQ1, then, misinformation perceptions became significantly more pronounced as the pandemic evolved.

Table 1:

Results of panel analysis explaining variation in misinformation perceptions regarding COVID-19.

	(1)	(2)	(3)	(4)	(5)	(6)
Social media exposure		.04* (.02)	.04* (.02)	.04* (.02)	.05** (.02)	.05** (.02)
Traditional media exposure		–.04† (.02)	–.04† (.02)	–.02 (.02)	.00 (.02)	.02 (.02)
Fear perceptions			.05* (.02)	.05* (.02)	.05* (.02)	.03 (.01)
News avoidance				.05*** (.01)	.03** (.01)	.03* (.01)
News seeking				–.03* (.01)	–.02 (.01)	–.02* (.01)
Trust in official actors					–.13*** (.02)	–.15*** (.02)
Trust in news officials					–.07*** (.02)	–.07*** (.02)
Wave 2	.19*** (.04)	.19*** (.04)	.19*** (.04)	.18*** (.04)	.15*** (.04)
Wave 3	.23*** (.05)	.22*** (.05)	.22*** (.05)	.20*** (.05)	.17*** (.05)
Wave 4	.26*** (.05)	.25*** (.05)	.25*** (.05)	.23*** (.05)	.19*** (.05)
Wave 5	.37*** (.05)	.37*** (.05)	.37*** (.05)	.35*** (.05)	.26*** (.06)
Constant	4.66*** (.03)	4.54*** (.11)	4.54*** (.11)	4.49*** (.12)	5.42*** (.16)	5.82*** (.14)

Note: Cells contain unstandardized regression coefficients, standard errors in parentheses.

† p < 0.10, * p < 0.05, ** p < 0.01, *** p < 0.001

Figure 1:

Means of misinformation perception across five waves, separated for those respondents who primarily use traditional media or social media for their news consumption.

9 Discussion

At times of crisis and uncertainty, such as the outbreak of COVID-19, misinformation may thrive in people’s newsfeeds. The context of uncertainty may also correspond to distrust in information sources: As different (online) sources communicated different narratives and as expert interpretations were subject to change as new information came to light, people became uncertain about whom to trust. Thus, the outbreak of COVID-19 was arguably not only surrounded by high levels of misinformation (e. g., Pennycook et al., 2020), people also perceived a lack of expert knowledge and empirical evidence as crucial issues during the pandemic (Nielsen et al., 2020). In this challenging setting, this paper asked how misinformation perceptions developed over time during different stages of the outbreak, and how these potential changes corresponded to (social) media diets and preferences. We answered this main question through a five-wave panel survey study among a representative sample of Dutch participants conducted at different stages of the pandemic, spanning the different waves of the actual pandemic and the different lockdowns and restrictions on public life that people went through.

The first main finding of this paper is that misinformation perceptions increased over time. We thus find support for a “cumulation effect” of misinformation perceptions. This can be explained by the increasing salience of actual (refuted) misinformation in people’s newsfeeds as the pandemic progressed. In line with this finding, content analyses have shown that digital information channels in particular contained high levels of false and/or misleading information (see e. g., Brennen et al., 2021; Pennycook et al., 2020), which arguably trickled down to the credibility evaluations of news users (also see Newman et al., 2020). The current hybrid media environment may have played a role in the amplification and mainstreaming of misinformation (e. g., Kim et al., 2018; Lukito, 2020). Specifically, mainstream media channels have reported extensively on conspiracies, refuted false statements (i. e., misperceptions about face masks) and counter-factual narratives originating online (i. e., spread via Twitter but reported by legacy media). This over-time attention to false storylines and refuted statements that were first disseminated as truthful may have cultivated a stronger perception of information pollution among news users. It is relevant to note that increases in misinformation perceptions are at odds with increasing levels of trust in authorities and health officials, especially during the earlier phases of the pandemic (e. g., Skirbekk et al., 2023). One explanation could be that people’s concerns about misinformation mostly related to online information and unverified sources. Such concerns about online falsehoods may motivate the acceptance of information from authorities and official sources, explaining increasing levels of trust in official sources during the pandemic.

Our second main finding is that misinformation beliefs corresponded to online and social media use, although the relationship between media use and misinformation beliefs remained stable across the different waves. This means that people who predominately used online news via social media were more likely to perceive information on COVID-19 as false than mainstream news consumers. This relationship between mainstream versus social media consumption and misinformation perceptions is in line with Nielsen et al.’s (2020) findings indicating that news consumers are more likely to associate online than offline sources with misinformation. This finding can be understood in light of the higher likelihood to encounter misinformation, conspiracies, and counter-factual narratives in online compared to offline settings (Van Aelst et al., 2017). As more counter-factual narratives, conspiracies and alternative truth epistemologies are found on social media than in mainstream media (Waisbord, 2018), people may be more likely to increasingly associate this online setting with misinformation. In other words, as more contradicting and delegitimizing narratives are present, this may increase the belief of information pollution online. We found that the relationship between social media use and misinformation perceptions was mostly driven by Facebook and Instagram use. WhatsApp and Twitter played a less central role, which can be explained as they are, at least in the Netherlands, typically regarded as channels for interpersonal communication (WhatsApp) or elite communication (Twitter). The more community-oriented structure of the other platforms, as well as the different target audience, may correspond with higher misinformation levels – both in terms of perceptions and actual falsehoods spread via these social media channels.

Another important finding is that the relationship between media use or media trust and misinformation perceptions did not change over time. Although we see an increase in misinformation beliefs across waves, the role that media exposure, avoidance and trust played remained stable over time. We interpret this as evidence that – in the context of continuing relevance and salience of COVID-19 across all the different waves – media diets and perceptions did not take on a different role at different stages of the pandemic. We can potentially explain this as media behaviors may have changed between a routine versus pandemic period, but not within the actual crisis context where they have taken on more stable routines. Future longitudinal research on media perceptions in times of crisis should ideally rely on data collected across routine periods and different stages of the crisis to arrive at more refined insights on the development of media use and trust (misinformation beliefs) over time.

Our main findings can be translated to practical implications for journalism and policymakers. First of all, in order to increase the likelihood that people comply with risk-averse behaviors demanded by governments throughout different stages of the pandemic, it is important to enhance trust in established sources of information and keep misinformation perceptions as low as possible – as they may correspond to the selective avoidance of established information sources. Over-time cultivated levels of uncertainty, disagreement and cumulative misinformation exposure may be important barriers for policy makers that ask citizens to adjust their behaviors multiple times over an extensive time-period. In this challenging context, it is recommended to restore trust in established sources of information and motivate citizens’ approach to novel information on the pandemic.

Second, it is important to make high-quality information available, comprehensive, and accessible through conventional information channels. As people with higher misinformation perceptions may approach social media platforms that actually have a higher likelihood to contain falsehoods (e. g., Yang et al., 2023), it is important to think about ways to present conventional information in a more attractive and trustworthy format. Arguably, enhancing transparency related to uncertainty (of expert claims or new evidence) and acknowledging the distrust of citizens in times of crisis may be ways to engage with people with higher misinformation perceptions through conventional information channels.

In the context of these recommendations, it should be noted that we did not analyze the perceived and actual causes of misinformation beliefs. As there was a lot of false and even deliberately deceptive information present in people’s media diets, a certain prevalence of misinformation beliefs is justified, and may actually correspond to the supply of information disseminated during the pandemic. However, we found that misinformation perceptions, on average, exceeded the midpoint of the scale substantially and significantly, which may point to severe distrust in the accuracy of information. As some studies indicate that less than one percent of people’s media diets may consist of false or misleading information (e. g., Allen et al., 2020), it could be argued that the misinformation beliefs identified in this study are disproportionate and may indicate a level of concern that exceeds the actual threat. Perceiving and discussing misinformation may also have real consequences – which also shows that it is important to focus on the effects of discussions on the threat of misinformation (see also Van Duyn and Collier, 2019). We leave it up to future research to more directly assess the extent to which misinformation beliefs are justified, and whether the perceived prevalence of false information corresponds to the actual level of misinformation that citziens are exposed to. Against this backdrop, it is important to assess to what extent concerns about the prevalence of misinformation are proportionate to the actual threat posed by false information in people’s media diets.

Although our findings offer important new insights into the development of misinformation beliefs related to COVID-19 over an extensive time period, our study is not without limitations. First of all, we collected data in one single country (the Netherlands). Although many nations faced similar consequences and revelations of the virus over the sampled period of six months, there are important regional variations in the severity of the pandemic and the restrictions on public and economic life that have not been taken into account here. We suggest that future research relies on a comparative (most-different-systems) design and factor in regional differences in the severity of the crisis and national responses to the threat. Second, we rely on self-reported measures of misinformation perceptions as well as media use. Such self-reported measures are known to come with biases. We suggest that future research triangulates self-reported measures with observational data and media tracking data that may together offer more comprehensive insights into people’s behaviors during the different stages of the pandemic.

Here, we should also note that we only had data available for four social media platforms. Although they reflected the most widely used platforms by different segments of the population at the time of data collection, the omission of YouTube as a platform may have excluded exposure to more alternative counter-factual narratives that became more central on this platform when the pandemic progressed. Finally, although we rely on multiple waves, our data may not offer insights into the causal mechanisms that explain the relationships postulated here. We do not exclude the option of reversed causal pathways or mutually reinforcing mechanisms in which (alternative) social media diets reinforce misinformation perceptions and the other way around. We suggest that future research relies on more sophisticated methods and causal models that also include information about the veracity of the content people were exposed to.

Despite these limitations, we offer novel insights into how people’s misinformation beliefs may change as a global health crisis presents itself as a new (temporal) reality in the digital information ecology. To make sure that people are willing to continuously update their responses in line with the information disseminated by the establishment, it is crucial to make sure that misinformation perceptions are not amplified to the point where segments of society systematically reject the truth claims of the authorities, and seek shelter in alternative communities that reconcile their distrust.