The media morphosis of science communication during crises
Abstract
- Abstract
- en Effective science communication is crucial in informing citizens in times of crisis. The ICT-driven metamorphosis of the media industry (media-morphosis) has facilitated the shift from a rather one-directional information deficit model in science communication to a more bidirectional communication corresponding to a dialogue model. Little is known about the preferences for social media communication from the supply side of institutional and scholarly communication. Analyzing survey data from nearly 8,700 scientists, we compared their preferences for press media and social media, proxies of traditional centralized communication channels and modern decentralized channels, respectively. Our empirical results show that trust in governments and prioritizing access to all perspectives on the crisis are important factors that influence scientists’ preferences regarding forms of communication. Trust favors centralized systems, while social media serves as an alternative in contexts lacking trust in politics. Prioritizing the diffusion of different perspectives drives a shift to stated preferences for decentralized communication systems. We also show that scientists’ stated preferences for communication systems depend on the socio-political context.
Description
- Full text
-
The media morphosis of science
communication during crises
Rocco Caferra1, Giuseppe Di Liddo2, Andrea Morone2 & David Stadelmann3
Effective science communication is crucial in informing citizens in times of crisis. The ICT-driven
metamorphosis of the media industry (media-morphosis) has facilitated the shift from a rather
one-directional information deficit model in science communication to a more bidirectional
communication corresponding to a dialogue model. Little is known about the preferences for social
media communication from the supply side of institutional and scholarly communication. Analyzing
survey data from nearly 8,700 scientists, we compared their preferences for press media and social
media, proxies of traditional centralized communication channels and modern decentralized channels,
respectively. Our empirical results show that trust in governments and prioritizing access to all
perspectives on the crisis are important factors that influence scientists’ preferences regarding forms
of communication. Trust favors centralized systems, while social media serves as an alternative in
contexts lacking trust in politics. Prioritizing the diffusion of different perspectives drives a shift to
stated preferences for decentralized communication systems. We also show that scientists’ stated
preferences for communication systems depend on the socio-political context.
Keywords Trust, Crises, Media, Knowledge
The advent of the digital age has paved the way for innovations in the interaction between scientific institutions,
scientists and citizens1,2. Social media platforms are increasingly taking center stage at the expense of traditional
print media, leading to a transformation of science journalism3. These platforms, supported by the development
of e-government and mobile technologies hold the promise to change science communication, stimulate
innovation and rethink the flow of information in the “knowledge society”4,5. The media industry is undergoing
a significant transformation, commonly referred to as the “media-morphosis”6. In this transformation process, it
is essential to understand the innovations of scientific information dissemination5. The accessibility and type of
(science) information published is highly dependent on the media platform used7. Given the disruptive innovation
experienced by the media industry8, it is relevant investigating emerging trends in science communication9 and,
in particular, understanding scientists’ preferences for traditional and modern communication channels.
The media-morphosis can be analyzed by observing the preferences for two types of systems: traditional
centralized media, such as print media, and modern social media. These models have different characteristics
including the potential quality, accessibility, and effectiveness of the disseminated information. In fact, as analyzed
in this paper, traditional centralized models tend to view science as a one-way knowledge transfer, but this has
been criticized for being top-down and paternalistic. Modern decentralized models, on the other hand, favor
accessibility and the speed of information flow, but they might make filtering and quality checking processes more
complicated due to the multitude and heterogeneity of agents present within them. This relative decentralization
marks a shift from “information deficit model” which sees public skepticism because of insufficient knowledge,
to a potentially more inclusive “dialogue model” which emphasizes mutual understanding and participatory
engagement between scientists and the public10.
There is a lack of research on the impact of this media-morphosis on the supply of insights by scientists. We
aim to gain a deeper understanding of scientists’ preferences regarding science communication during crises, a
showcase where the dissemination of the correct information is vital to foster human coordination.
Understanding the factors that influence scientists’ communication preferences is particularly important in
times of crisis, whether natural disasters, public health emergencies or other unforeseen events, the importance
of rapid and accurate communication is paramount. The information driven coordination and engagement of
human efforts can be critical to effectively managing and mitigating the adverse effects of a crisis11,12. Ideally,
well-structured crisis communications ensure that all stakeholders including politicians, experts, and citizens
have access to the necessary information. In this context, the role of the scientific community is crucial, as
scientists have the knowledge and insights needed to inform policymakers and the public at large. Indeed, science
1University of Rome Unitelma Sapienza, Rome, Italy. 2University of Bari “Aldo Moro”, Bari, Italy. 3University
of Bayreuth, Bayreuth, Germany. email: rocco.caferra@unitelmasapienza.it; giuseppe.diliddo@uniba.it;
a.morone@uniba.it; david.stadelmann@uni-bayreuth.de
OPEN
Scientific Reports | (2025) 15:5506
1
| https://doi.org/10.1038/s41598-025-88973-7
www.nature.com/scientificreports
communication and policy mutually influence each other, and emerging social media science communication
could influence the use of scientific evidence in the policy-making process9.
To investigate factors affecting scientists’ communication preferences, we analyzed a survey of scientists
with nearly 8,700 responses on items related to more centralized or decentralized means of communication. By
analyzing scientists’ responses, we explore their propensity for different channels of information dissemination
as a function of scientists’ trust in government, their assessment of knowledge in the scientific community, and
the importance they place on the diversity of perspectives, including diverging perspectives such as potential
conspiracy theories. Specifically, we compare scientists’ preferences for more centralized, top-down systems,
such as traditional print media, which are typically curated and disseminated by professional journalists and
editors, with more decentralized systems, such as social media, where content is generated and shared by a
diverse range of users.
We investigate two hypotheses related to scientists’ preferences for means of communication during
emergencies and provide evidence to support them. Firstly, we hypothesize that scientists who trust the
government are more likely to favor centralized communication forms, such as print media, while being less
inclined towards decentralized communication forms, like social media. As a second hypothesis we propose
that when scientists believe that all perspectives on crises should be considered, their preferences will be
reversed, favoring decentralized communication over centralized platforms. Through the analysis of survey data
collected during the early period of the COVID-19 pandemic (from May 4 to June 3, 2020), we substantiate both
hypotheses.
Our empirical findings reveal that trust in government is positively associated with scientists’ preferences
for centralized information dissemination. On the other hand, when scientists emphasize the importance of
diverse perspectives, they show a tendency to favor decentralized communication channels like social media.
Furthermore, our results uncover an additional influencing factor on scientists’ communication preferences.
Scientists who perceive that the scientific community has a clear understanding of the costs and benefits of
different courses of action during the crises are more inclined to support centralized communication forms
while no effect is found for decentralized communication forms. This indicates that the perception of the
scientific community’s knowledge plays a significant role in shaping communication preferences of scientists
during crises.
The remainder of this paper is organized as follows: Sect. "Literature review and theoretical framework"
overviews the literature motivating our article, while we discuss the data and empirical strategy in Sect. "Data
and empirical strategy". Our main results, robustness tests and refinements are presented in Sect. "Results".
Section "Conclusions" offers concluding remarks.
Literature review and theoretical framework
Science communication plays a vital role in enhancing societal culture by fostering scientific awareness and
increasing access to information7. However, scientific knowledge, often seen as the epitome of ‘true knowledge,’
also relies on mass media to bridge any potential gap between scientists and society. Although different types of
outreach events and educational venues (such as the European Researchers’ Night) play a role in communicating
scientific knowledge, mass media offers the advantage of large-scale, continuous dissemination.
Media services have experienced significant innovations in recent times13,14. Despite its importance, the
relationship between innovation theory and media studies remains underexplored. Media innovation includes
both process and product changes, such as the emergence of new platforms and novel types of information.
These developments are crucial for the transformation of the media industry8. Within this framework, digital
social media platforms have emerged as disruptive innovations in the internet era, representing a media-
morphosis responsible for significant sociopolitical changes6,15–17. The shift in the use of media can be analyzed
by comparing centralized and decentralized systems of information dissemination. In the following section, we
will explain the reasons behind this shift and shed light on why such analyses become particularly relevant in
times of crisis.
Traditional print media, operating within centralized, top-down systems, relied on professional journalists
and editors to curate content. By contrast, contemporary internet-based models adopt a decentralized, bottom-
up approach that fundamentally alters the dissemination dynamics. While traditional media ensures a certain
editorial rigor, social media platforms, driven by a diverse range of independent users, can amplify the risk of
misinformation and fake news18,19.
Social media platforms are often seen as enhancing democracy through open and equal deliberation among
citizens, representatives, and policymakers20. On the one hand, they facilitate social networking and user-centered
innovation21,22, enabling the open-source diffusion of information. On the other hand, they risk degrading the
quality of information compared to traditional, authoritative mass media. The democratization related to social
media may introduce challenges such as misinformation, power imbalances, and cultural impacts23,24. This
duality highlights the need for improved communication strategies that address these evolving challenges.
Over recent decades and with the advent of social media in particular, science communication has also
transitioned from a “information deficit model” to the more participatory approach of a “dialogue model”.
This transition has coincided with changes in the creation and dissemination of informational content. In fact,
the information deficit model assumes a one-way transfer of knowledge from informed experts to a passive,
uninformed audience. The audience functions as a receiver of information. Centralized systems tend to correspond
to deficit model because they limit possible interactions between the expert and the audience. This approach
has been critiqued for oversimplifying the complex relationship between scientific knowledge and society10,25.
The dialogue model, characterized by reciprocal engagement and mutual understanding between scientists
and the public26has been argued to promote active engagement with non-expert communities, emphasizing
trust-building rather than mere dissemination27. This may foster inclusive and socially robust solutions, further
Scientific Reports | (2025) 15:5506
2
| https://doi.org/10.1038/s41598-025-88973-7
www.nature.com/scientificreports/
enhancing the societal relevance and impact of research28, potentially at the expense of reliability and accuracy.
These models find their potential empirical correspondence in in decentralized systems, such as social media,
where users receive information and have the opportunity to interact both among themselves and with the
information provider. While our empirical analyses relate the dialogue model to social media, it is important to
note that it is not limited to social media; rather, it includes all dialogic approaches aimed at fostering a two-way
conversation, enabling the exchange of information, perspectives, perceptions, and concerns between scientists
and the public27.
During crises, the intersection of science and media becomes central to coordinating individual responses
and making informed decisions29–32. Preferences for communication models, whether centralized (in line with
the information deficit model) or decentralized (in line with the dialog model), might depend on trust33and
the perceived value of diverse perspectives. The limitations of centralized gatekeeping, as highlighted during
the early COVID-19 pandemic, underscore the importance of open systems for addressing public confusion
and misinformation34–36. The expertise of scientists and their communication strategies play a pivotal role in
mitigating these issues.
The spread of misinformation during crises37,38 underscores the need for robust communication systems.
As discussed above, traditional media’s centralized filtering often provides higher-quality information, while
decentralized networks promote crowdsourced knowledge influenced by informal networks39. Social media
platforms facilitate rapid access to and dissemination of information, enabling greater diversity of perspectives40,41.
However, as Melki et al.42 argue, information verification practices are critical, given the questionable credibility
of user-generated content. Furthermore, a broader participation in the information diffusion, as in the dialogue
model, may improve the effectiveness of communication. These contrasting systems shape communication
preferences, particularly during crises when rapid, accurate information is essential.
Data and empirical strategy
Data
To comprehensively assess the range of perspectives of scientists regarding centralized and decentralized forms
of communication, we explore a survey designed during the period of late March to April 2020, i.e., at the start
of the COVID-19 pandemic. At that time, the global community was grappling with the uncertainties associated
with the disease and the intricate nuances of knowledge present in the scientific community. The data collection
phase spanned from May 4 to June 3, 2020, a period that coincided with the initial wave of COVID-19 in many
countries, during which the world’s attention was focused on contrasting the pandemic. The survey allows us
to gather unique insights that may inform future discussions during crises when expert knowledge is required.
Ethical clearance and survey information
Ethical clearance and protocol approval for data collection was granted by the Ethics Commission of the
Frankfurt School of Finance and Management. We follow the legal duty of General Data Protection Regulation
(GDPR) (EU) 2016/679. No personal or sensitive information that can be used to identify the respondents
were collected. Besides, the consent of the respondents to partake in the online survey were seek before the
survey was executed by including an electronic informed consent in the online survey form. All procedures were
performed in accordance with relevant guidelines. The participants were self-selected and informed that the
survey pertained to the crisis related to the pandemic. As an incentive, participants were offered a $500 lottery
prize for themselves and $500 for a charity of their choice if they provided a contact email address at the very
end of the survey.
The survey was conducted through SurveyMonkey in 2020 (surveymonkey.com). SurveyMonkey is one of
the world’s most popular survey platforms. It targeted scientists who had been corresponding authors of articles
published in top-ranked journals across 55 scientific fields from 2015 to early 2020. The selection of journals was
based on their 2019 SCImago Journal Ranking (SJR) in 13 scientific areas defined by Scopus: Arts and Humanities;
Business, Management and Accounting; Economics, Econometrics and Finance; Energy; Health Professions;
Immunology and Microbiology; Medicine; Multidisciplinary; Neuroscience; Nursing; Pharmacology, Toxicology
and Pharmaceutics; Psychology; Social Sciences. To obtain responses from corresponding authors, 849 journals
were sampled, and 318,251 corresponding author email addresses were extracted from all over the world43. To
ensure representation of social scientists, we also included 68,470 scholars registered in the RePEc bibliographic
database. Two-thirds of the sample pool, amounting to 220,923 invitations, were randomly sent out from May
4 to June 3, 2020, excluding Sundays. The response rate based on opened emails was 13.93%, 5.76% based on
accepted invitations, with 98% completing the survey within 24 h of opening the link. The survey covered
various topics related to COVID-19, including a special section on the media. Participation was voluntary, and
subjects could skip questions or quit the survey at any time. Consequently, there is a proportion of missing values
for different variables of about 33%.
Main variables of interest
Our primary objective is to investigate preferences for centralized and decentralized information channels,
particularly focusing on print media vs. social media. To capture scientists’ communication preferences, we
utilized a question where participants were asked to indicate the extent to which they obtained information from
print media and social media on a 7-point Likert scale, ranging from 1 (not at all) to 7 (very much). The specific
question was “To what extent do you get your information about the COVID-19 epidemic from the following
sources? [item 1 – print media], [item 2 – social media]”. It is worth noting that the two items are negatively
correlated, with a correlation coefficient of −0.205 (p < .01). On the one hand, there might be some key variables
determining the trade-off between the two means of information; on the other hand, the absence of a close
Scientific Reports | (2025) 15:5506
3
| https://doi.org/10.1038/s41598-025-88973-7
www.nature.com/scientificreports/
inverse correlation can be partially attributed to other determinants explaining the preferences for both the
means of communications. We aim to explore these determinants.
A central aspect of our research is investigating how trust in government influences preferences of means
of communication. To measure trust in government’s truthfulness and care for citizens, we closely follow the
literature and adopt the approach used by Fetzer et al. (2020) and explicitly asked participating scientists whether
they trust their country’s government to take care of its citizens (“How much do you trust the government of the
country where you live in to take care of its citizens?”). Thereby we ensure the quality and comparability of the
results with our trust measure43 (Forrester and Nowrasteh, 2023). As in the literature, our measure ranges from 1
to 5 Likert scale with higher values reflecting higher levels of trust in governments and, in turn, greater perceived
truthfulness of government.
Furthermore, we delve into the direct connection between the media and information dissemination. To
explore the hypothesis that scientists favor decentralized communication when they believe in considering all
perspectives on a crisis, i.e., including those perspectives deemed as a potential conspiracy, we incorporated a
specific question in the survey. Participants were asked to rate their agreement level on a 7-point Likert scale,
ranging from 1 (strongly disagree) to 7 (strongly agree) regarding whether the news media should report all
perspectives on COVID-19 even in relation to conspiracy (“The news media should report all perspectives on
COVID-19, even those considered to be conspiracy theories.”).
In addition, we aim to understand how scientists’ perceptions of the scientific community’s knowledge
influence their communication preferences for centralized vs. decentralized means of communication. To
measure this, we asked participants to rate their perceptions of whether the scientific community has a clear
understanding of the costs and benefits of different courses of action regarding the coronavirus (“The scientific
community has a clear idea about the costs and benefits of different courses of action in relation to the coronavirus.”).
The responses were recorded on a 7-point Likert scale, ranging from 1 (strongly disagree) to 7 (strongly agree).
In Table 1, we present descriptive statistics for the main variables that provide an overview of the responses
received from the participating scientists. It is evident that preferences for decentralized information
systems are, on average, higher than those for Centralized ones (4.96 vs. 3.9), i.e. during crisis the overall
support for a communication approach in line with the dialogue model is somewhat higher than support for
an approach following the deficit model. In addition to these variables, we also control for a wide range of
personal characteristics, views on the goals of government policy during the crisis, fixed effects for countries
where scientists live, fixed effects for scientific fields, and other relevant covariates, including COVID19 case
numbers, among others. These additional variables can help us to better isolate the relationships of interest and
the potential influences of trust on scientists’ communication preferences during a critical point of the crisis.
Descriptive statistics of all other control variables are relegated to the online supporting information, Table A1.
Empirical strategy
Our empirical strategy is a regression control framework. Our first main estimation Eq. (1) focusses on
preferences for centralized information provision as a dependent variable:
CentralizedInfoi = α + β 1Trusti + β 2AllPerspectivesi + β 3KnowledgeCommunityi + Xiβ + Ciϕ + ϵ
(1)
Our second main estimation Eq. (2) focusses on preferences for decentralized information provision.
DecentralizedInfoi = γ + δ 1Trusti + δ 2AllPerspectivesi + δ 3KnowledgeCommunityi + Xiδ + Ciγ + ϵ
(2)
The unit of observation in both equations is the individual scientist. We always control for all personal
characteristics, scientific domains, and other covariates summarized by Xi. We also account for country-fixed
effects, Ci. The inclusion of country-fixed effects is able account for potential for country-level heterogeneity,
such as cultural differences or the relevance of country-specific institutions. Media usage may differ significantly
depending on national or cultural backgrounds. Similarly trust in government may depend on institutions.
By accounting for country-fixed effects, we hold the influence of country-specific factors constant. ϵ is an
idiosyncratic error term. Results are based on OLS estimators, and we calculate robust standard error estimates.
Variable
Obs.
Mean
Std.
Dev.
Min
Max
CentralizedInfo: Preference for Centralized Information - To what extent do you get your information about the COVID-19 epidemic
from the following sources? (Print media)
8493
3.901
2.314
1
7
DecentralizedInfo: Preference for Decentralized Information - To what extent do you get your information about the COVID-19 epidemic
from the following sources? (Social Media)
8701
4.96
1.965
1
7
Trust: How much do you trust the government of the country where you live in to take care of its citizens?
8509
2.939
1.395
1
5
AllPerspectives: The news media should report all perspectives on COVID-19, even those considered to be conspiracy theories.
8438
2.823
1.8
1
7
KnowledgeCommunity: The scientific community has a clear idea about the costs and benefits of different courses of action in relation to
the coronavirus.
8497
3.968
1.573
1
7
Table 1. Descriptive statistics (Likert Scales). Notes: Descriptive statistics for main variables of interest based
on responses of individual scientists. Descriptive statistics for all variables are presented in the supporting
information.
Scientific Reports | (2025) 15:5506
4
| https://doi.org/10.1038/s41598-025-88973-7
www.nature.com/scientificreports/
Both estimation equations differ only regarding the dependent variable. Our hypothesis is that scientists who
exhibit higher trust levels (Trust) tend to favor centralized information provision (CentralizedInfo) more than
decentralized information provision (DecentralizedInfo), i.e. β 1 > δ 1 (Hypothesis 1). It might even be the case
that higher trust levels are positively associated to centralized information provision but negatively associated to
decentralized information provision, i.e. β 1 > 0 and δ 1 < 0.
Moreover, we hypothesize that scientists having higher preferences for reporting all perspectives on the crisis
(AllPerspectives), have a lower tendency to support centralized information provision but a higher tendency to
support decentralized information provision, such that β 2 < δ 2 (Hypothesis 2). It might even be the case that
scientists who care for all perspectives do not favor centralized information provision, i.e., β 2 < 0, but favor
instead decentralized information provision, i.e., δ 2 > 0.
Finally, if scientists think that the scientific community has a clear idea about the costs and benefits related
to the crisis (KnowledgeCommunity), then we hypothesize that they rather support a centralized information
provision in comparison to a decentralized information provision, i.e., β 3 > δ 3 (Hypothesis 3). We suppose
that if scientists think that their community has a comparatively low knowledge of the costs and benefits related
to the crisis, then they do not express any clear preference for either centralized or decentralized information
provision.
In our empirical analysis, we account for numerous other covariates ( Xi) to reduce potential omitted variable
bias in our coefficient estimates. For example, if older scientists tend to prefer more centralized information
provision and if they also have more trust in government, controlling for the age of the scientists responding to
our survey is relevant. The large array of covariates and country effects in our analysis gives us some confidence
that the relationships found between trust, preference for all perspectives and the belief regarding knowledge
in the scientific community may shape preferences for different communication channels. Importantly, we
hypothesized that trust in government, preferences for all perspectives to be reported, and the belief regarding
knowledge in the scientific community systematically and differentially correlate with preferences for centralized
and decentralized information.
Thus, our setting gains the reliability because for precisely the same individuals, the same control variables, the
same estimation strategy, we predict that β 1 > δ 1, β 2 < δ 2, and β 3 > δ 3, i.e., we predict the differential
influences for our main variables regarding centralized versus decentralized information provision. Our results
will provide empirical support for all three hypotheses.
Results
Main results
Table 2 presents the key findings of our study. In specifications (1) to (4), we examine scientists’ preferences for
centralized information as a dependent variable, while in specifications (5) to (8), we investigate preferences for
decentralized information provision as a dependent variable. To estimate specifications (4) and (8), we make use
of ordered probit models. To ensure the robustness of our results and account for potential confounding factors,
we always incorporate a comprehensive set of control variables across all specifications, along with country
effects. Coefficients for all control variables are presented in the supporting information.
If scientists express trust in their country’s government to take care of citizens, they generally show a preference
for centralized information provision, such as print media. The estimated point coefficient associated with the
variable Trust exhibits a positive and statistically significant association, β 1 > 0, in all specifications where
the dependent variable is preferences for centralized information. On the other hand, higher levels of trust in
government are negatively associated ( δ 1 < 0) with preferences for decentralized information provision, such
as social media. This association is, however, not statistically significant. Consequently, our findings support
Hypothesis 1 that the coefficient for preferences of centralized information is greater than the coefficient for
preferences of decentralized information, β 1 > δ 1. Scientists who exhibit higher trust levels tend to favor
centralized information provision.
Next, we delve into the association between the support expressed by scientists for news media reporting
all perspectives, including those considered as conspiracy theories. The variable representing this support,
AllPerspectives is found to be negatively and statistically significantly associated with preferences for centralized
information provision, β 2 < 0. In contrast, when considering decentralized information provision as the
dependent variable, the coefficient for AllPerspectives is positively and statistically significantly associated with
this variable, δ 2 > 0. This implies that scientists who express a stronger preference for news media reporting
all perspectives on the crisis are less likely to support centralized information provision but are more inclined
to support decentralized information provision. In essence, their willingness to consider diverse perspectives
appears to influence their communication preferences, favoring decentralized channels over centralized ones.
Thereby providing support for Hypothesis 2, i.e., β 2 < δ 2.
Finally, we explore the significance of scientists’ beliefs regarding the knowledge of the scientific community
concerning the costs and benefits of various courses of action related to the crisis. Our results indicate that
as the perceived knowledge of the scientific community increases, scientists’ preferences for centralized
information also increase significantly. However, we find no statistically significant relationship between the
variable KnowledgeCommunity and preferences for decentralized information provision. Providing evidence
for Hypothesis 3, our findings suggest that the higher the level of perceived knowledge within the scientific
community, the more likely an expert is to support centralized information provision over decentralized
information provision, β 3 > δ 3. This underscores the relevance of expert perceptions of the scientific
community’s expertise in shaping their communication preferences during the crisis.
These results highlight that individuals who have higher trust in government are more likely to prefer top-
down systems, supporting information providing in line with the deficit model during times of crisis, where
one-directional information diffusion between experts and non-experts is the standard. Still, it is important to
Scientific Reports | (2025) 15:5506
5
| https://doi.org/10.1038/s41598-025-88973-7
www.nature.com/scientificreports/
note that preferences for dialogue model, based on the use of decentralized social media networks, are generally
higher, supporting the earlier argument about the social transition toward this type of system (as highlighted in
Table 1). Moreover, these preferences are neutral with respect to institutional trust (the impact is not statistically
significant), do not depend on the perception of the quality of information held by experts, but rather follow
a bottom-up model where scientists support the democratization and heterogeneity of the information being
disseminated.
Robustness and refinements: the relevance of social trust and quality of government
We further explore our main hypotheses concerning the relevance of scientists’ trust and preferences for reporting
all perspectives in the media. Considering the model outlined in the empirical strategy section, we made several
extensions using the sample splitting approach. To provide a deeper analysis, we considered two additional
dimensions: social trust in society and the quality of government, since formal and informal institutions can play
an important role in defining social organization44,45. To achieve this, we divided the sample of scientists based
on the countries in which they are active. Specifically, we formed groups of scientists from countries with low
levels of social trust and high levels of social trust, as well as groups from countries with low government quality
and high government quality. This setting allows us to investigate the relevance of institutions (including social
trust in a country as well as government quality) next to individual responses of scientists. We then estimate
our main models including all control variables and country effects. The results of this analysis are presented in
Table 3.
We have observed that the individual trust of scientists in the government’s ability to take care of its citizens
during the crisis is consistently positively associated with preferences for centralized information provision. This
association holds true regardless of whether the country has a low trust or high trust society and irrespective of
the quality of government as show in Table 3. Thus, the coefficient β 1 for Trust remains positive regardless of
the specific institutional context, indicating a consistent pattern of trust positively influencing preferences for
centralized information provision. Conversely, we find that trust is never positively associated with preferences
for decentralized information provision. In certain settings where there is high social trust and high government
quality, there is even a slightly negative association ( δ 1 < 0). Thus, our main hypothesis regarding the relevance
of trust in government among scientists remains robust. Across different contexts, we systematically observe
that β 1 > δ 1 (Hypothesis 1) indicating that scientists with higher levels of trust in government tend to prefer
centralized information provision over decentralized alternatives, favoring the one-way communication of the
information deficit model.
When investigating the variable AllPerspectives, a more nuanced picture emerges. As before, we consistently
find a positive relationship in all institutional settings with a preference for decentralized information ( δ 2 > 0
). However, in societies with high social trust and high government quality, scientists who express a higher
preference for having all perspectives reported tend to show a lower preference for centralized information
provision ( β 2 < 0), supporting participatory in line with the dialogue model. On the other hand, there is
no correlation between preferences for all perspectives and centralized information provision in societies with
low social trust and low government quality. This observation can be viewed as consistent, as in such societies,
there might be a perception of a relevant risk when reporting all perspectives, including conspiracy theories.
Our empirical result also highlights the complex interplay between societal factors, trust, and preferences for
information dissemination during crises. The role of social trust and government quality adds further depth
to our understanding of how scientists’ communication preferences are influenced by diverse perspectives,
underscoring the importance of context in shaping their preferences. Overall, our empirical results offer further
support for Hypothesis 2. Scientists who favor that all perspectives on a crisis should be reported in the media
display a lower tendency to support centralized information provision but exhibit a higher tendency to support
decentralized information provision ( β 2 < δ 2).
Finally, our analysis considers the influence of scientists’ perceptions regarding the knowledge of the scientific
community on costs and benefits related to the crisis depending on social trust and government quality. We find
Centralized Information
Decentralized Information
(1)
(2)
(3)
(4 ordered probit)
(5)
(6)
(7)
(8 ordered probit)
Trust
0.079***
0.077***
0.036***
− 0.031
− 0.026
− 0.014
(0.023)
(0.024)
(0.012)
(0.02)
(0.02)
(0.011)
AllPerspectives
− 0.033**
− 0.032**
− 0.014*
0.086***
0.084***
0.046***
(0.016)
(0.016)
(0.008)
(0.014)
(0.014)
(0.008)
KnowledgeCommunity
0.088***
0.081***
0.078***
0.039***
0.019
0.021
0.022
0.014
(0.018)
(0.018)
(0.018)
(0.009)
(0.015)
(0.015)
(0.015)
(0.009)
Control Variables
yes
yes
yes
yes
yes
yes
yes
yes
Country effects
yes
yes
yes
yes
yes
yes
yes
yes
Observations
7535
7394
7379
7379
7707
7556
7540
7540
(Pseudo) R2
0.12
0.121
0.122
0.036
0.079
0.083
0.083
0.027
Table 2. Differential influence of trust and preferences for reporting all perspectives in the media on .
Centralized vs. decentralized information provision. Notes: Robust standard errors are in parentheses; ***
p < .01, ** p < .05, * p < .1.
Scientific Reports | (2025) 15:5506
6
| https://doi.org/10.1038/s41598-025-88973-7
www.nature.com/scientificreports/
that if scientists believe the scientific community possesses a clear understanding of such aspects, they consistently
express support for centralized information provision. This association remains irrespective of whether they are
in a low social trust society or a high social trust society, as well as whether they reside in a country with low
government quality or high government quality. In essence, the coefficient β 3 for KnowledgeCommunity remains
positive and aligned with our hypotheses. On the other hand, we do not observe any significant association
between the variable KnowledgeCommunity and preferences for decentralized information provision. As such,
the perceived knowledge within the scientific community does not seem to impact scientists’ preferences
for decentralized information dissemination ( δ 3). In conclusion, we observe β 3 > δ 3 (consistent with
Hypothesis 3) such that the results demonstrate that scientists who have confidence in the scientific community’s
understanding of the costs and benefits related to the crisis tend to support centralized information provision
consistently across diverse societal contexts. However, their perceptions of knowledge do not seem to play a
significant role in shaping their preferences for decentralized information provision.
Discussion and limitations
We analyze how different models of science communication are perceived and accepted by scientists, particularly
during crises when the dissemination of accurate information is vital for social coordination. Our results
show that scientists with higher levels of trust in government are more likely to favor centralized information
dissemination, consistent with the information deficit model. Conversely, scientists who prioritize the inclusion
of all perspectives in media reporting, even those considered conspiratorial, demonstrate a stronger preference
for decentralized communication channels, in line with the dialogue model. Finally, when scientists believe that
the scientific community has a clear understanding of the costs and benefits related to a crisis, they are more
likely to support centralized information provision. This suggests that in situations where scientific consensus
exist, the need for feedback and dialogue through decentralized models may be perceived as less critical. Our
findings overall suggest a growing preference for decentralized, participatory communication models aligned
with the dialogue model. Interestingly, the preference for decentralized models persists even in times of crisis
but it is dependent on trust, perceptions regarding knowledge of the scientific community and the relevance
assigned to the inclusion of all perspectives in science communication.
While the study encompasses a large sample of scientists globally, making the findings less susceptible to
biases stemming from national, cultural, or institutional differences as when the analysis was based on scientists
from a single country, it is important to recognize the contextual limitations. Our data were collected during
a specific and unique crisis: the peak of the first wave of the COVID-19 pandemic. This context may limit the
generalizability of our results to other crises and to how scientists view the role of science communication in
general. Nevertheless, studying a crisis scenario offers valuable insights, as such situations often intensify conflicts
between centralized and decentralized information provision, providing a critical testbed for understanding
scientists’ communication preferences.
Although our analysis accounts for country-fixed effects and a range of control variables, it relies on
observational data from a survey. This methodological approach inherently limits the ability to establish
causality. Unobservable factors potentially correlated with our main variables of interest and the preference for
centralized or decentralized communication forms might affect the coefficient estimates. However, the observed
variations in scientists’ preferences, as well as the influence of trust and the perceived importance of including
all perspectives, suggest systematic differences in how centralized and decentralized information systems are
viewed and which model of science communication matters during crises.
It is also important to acknowledge that we treat all forms of decentralized media, such as social media, as a
single category which is a limitation of our study. We do not differentiate between platforms like X, Facebook,
YouTube, or blogs, each of which may have distinct user characteristics and attract varying preferences from
scientists. Scientists’ attitudes toward these platforms could vary based on factors such as audience reach,
Centralized Information
Decentralized Information
Centralized Information
Decentralized Information
(1) Low Social
trust
(2) High
Social
trust
(3) Low Social trust
(4) High
Social Trust
(5) Low
Government
Quality
(6) High
Government
Quality
(7) Low
Government
Quality
(8) High
Government
Quality
Trust
0.071*
0.078***
0.01
− 0.043*
0.082*
0.075***
0.011
− 0.041*
(0.042)
(0.029)
(0.033)
(0.025)
(0.044)
(0.029)
(0.035)
(0.025)
AllPerspectives
0.016
− 0.053***
0.096***
0.078***
0.029
− 0.057***
0.072***
0.089***
(0.028)
(0.02)
(0.023)
(0.018)
(0.029)
(0.02)
(0.023)
(0.017)
KnowledgeCommunity
0.089***
0.07***
0.036
0.017
0.111***
0.063***
0.028
0.021
(0.034)
(0.021)
(0.027)
(0.019)
(0.035)
(0.021)
(0.028)
(0.018)
Control Variables
yes
yes
Yes
yes
yes
yes
yes
yes
Country effects
yes
yes
Yes
yes
yes
yes
yes
yes
Observations
1982
5397
2036
5504
1817
5562
1870
5670
(Pseudo) R2
0.16
0.113
0.157
0.059
0.184
0.103
0.156
0.057
Table 3. The relevance of social trust and government quality for the differential influence of individual trust
and preferences for reporting all perspectives in the media. Notes: Robust standard errors are in parentheses;
*** p < .01, ** p < .05, * p < .1.
Scientific Reports | (2025) 15:5506
7
| https://doi.org/10.1038/s41598-025-88973-7
www.nature.com/scientificreports/
content format, or perceived credibility. However, we believe that the stark contrast between centralized and
decentralized modes of science communication highlighted in our study remains highly relevant. Our findings
suggest that, despite platform-specific nuances, social media in general is perceived as fundamentally different
from traditional, centralized media forms. This distinction provides meaningful insights into how decentralized
systems are viewed as an alternative to conventional communication channels.
Conclusions
By analyzing a comprehensive survey of up to 8,700 scientists, we investigate their communication preferences
during crises, shedding light on the factors that influence their choices between centralized and decentralized
information channels. The findings reveal the significant impact of three key variables: trust in government,
consideration of diverse perspectives, and perceptions of scientific community knowledge.
Our first result shows that scientists who trust the government tend to favor information provision by means
of traditional print media. This preference for centralized systems emphasizes their belief in government capacity
and coordination during crises. In contrast, the second hypothesis highlights that when scientists prioritize
the reporting of all perspectives, including those considered as conspiracy theories, their preferences shift
towards decentralized information platforms, like social media. This observation indicates a desire for diversity
in information sources, allowing for open debates and the exploration of alternative solutions. Moreover, our
analysis uncovers the relevance of scientists’ perceptions regarding the knowledge of the scientific community
on costs and benefits related to the crisis. When scientists perceive consensus in this area, they show a stronger
preference for centralized information provision. This suggests their trust in the expertise of the scientific
community to guide coordinated mass action during crises.
Trust in institutions and the consensus of scientific thinking strengthen preferences for the use of traditional
media and thus for the centralization of information to coordinate human action during crises. Following the
preferences of scientists, innovations in social media become important in stimulating the debate on different
alternative theories which might potentially be regarded as a kind of “democratization of science” which itself
depend less on the quality of institutions and central information systems.
Developing effective crisis communication strategies requires careful consideration of these factors, fostering
open dialogue, and ensuring that diverse perspectives are considered.
Future studies could build on this work, expanding it in several directions. From a methodological
standpoint, it would be interesting to distinguish the specific characteristics and diversity of actors within social
media platforms. In current systems, both authoritative sources (such as expert journalists in print media) and
other types of sources within decentralized networks, like social media, coexist. Moreover, it would be relevant
to understand the dynamics within decentralized networks to assess whether dialogue model works better in
terms of information quality when expert involvement as ambassadors facilitates the promotion and spread of
virtuous behaviors, or whether they are influenced by other actors who may hinder the exchange of high-quality
information.
Our study is the first to comprehensively examine scientists’ preferences for centralized vs. decentralized forms
of communication. It explores the pivotal roles of trust, the belief that all perspectives should be considered, and
the perceived knowledge within the scientific community as key factors influencing preferences for centralized
or decentralized communication channels. These findings shed light on the broader relevance of the information
deficit model versus the dialogue model in shaping science communication strategies, particularly in contexts
requiring effective dissemination and engagement.
Data availability
The dataset, scripts and the survey questions employed in this study are available as an online resource on
https://doi.org/10.6084/m9.figshare.28263500.v1.
Received: 3 May 2024; Accepted: 3 February 2025
References
1. Weingart, P. Science and the media. Res. Policy. 27 (8), 869–879 (1998).
2. Brossard, D. & Scheufele, D. A. Science, new media, and the public. Science 339 (6115), 40–41 (2013).
3. Massarani, L., Entradas, M., Neves, L. F. F. & Bauer, M. W. Global Science Journalism Report. Available online: (2021). https://ww
w.scidev.net/wp-content/uploads/Global-Science-Journalism-Report-2021.pdf (accessed November 21, 2024).
4. Jaeger, P. T., Bertot, J. C. & Shilton, K. Information policy and social media: Framing government—citizen web 2.0 interactions. In
Web 2.0 Technologies and Democratic Governance: Political, Policy and Management Implications (pp. 11–25). (2012).
5. Hyland-Wood, B., Gardner, J., Leask, J. & Ecker, U. K. Toward effective government communication strategies in the era of
COVID-19. Humanit. Social Sci. Commun. 8(1), 30 (2021).
6. Fidler, R. F. Mediamorphosis: Understanding new Media (Pine Forge, 1997).
7. Burns, T. W., O’Connor, D. J. & Stocklmayer, S. M. Science communication: a contemporary definition. Public. Underst. Sci. 12 (2),
183–202 (2003).
8. García-Avilés, J. A., Carvajal-Prieto, M., De Lara-González, A. & Arias-Robles, F. Developing an index of media innovation in a
national market: the case of Spain. Journalism Stud. 19 (1), 25–42 (2018).
9. Szüdi, G. et al. New trends in science communication fostering evidence-informed policymaking. Open. Res. Europe. 2, 78. https:/
/doi.org/10.12688/openreseurope.14769.1 (2022).
10. Bucchi, M. Of deficits, deviations and dialogues: theories of public communication of science. In Handbook of Public
Communication of Science and Technology (71–90). Routledge. (2008).
11. Botzen, W. J. W. et al. Individual hurricane evacuation intentions during the COVID-19 pandemic: insights for risk communication
and emergency management policies. Nat. Hazards. 111, 507–522 (2022).
Scientific Reports | (2025) 15:5506
8
| https://doi.org/10.1038/s41598-025-88973-7
www.nature.com/scientificreports/
12. Khan, S. et al. Rethinking communication in risk interpretation and action. Nat. Hazards. 88, 1709–1726. https://doi.org/10.1007
/s11069-017-2942-z (2017).
13. Xu, J. Does the medium matter? A meta-analysis on using social media vs. traditional media in crisis communication. Public.
Relations Rev. 46 (4), 101947 (2020).
14. Storsul, T. & Krumsvik, A. H. Media Innovations: A Multidisciplinary Study of Change (University of Gothenburg, 2013).
15. Winston, B. Media, Technology and Society: A History: from the Telegraph to the Internet (Routledge, 2002).
16. Shirky, C. The political power of social media: Technology, the public sphere, and political change. Foreign Aff. 90, 28–41 (2011).
17. Schroeder, R. Social Theory after the Internet: Media, Technology and Globalizationp. 210 (UCl, 2018).
18. Allcott, H. & Gentzkow, M. Social media and fake news in the 2016 election. J. Economic Perspect. 31 (2), 211–236 (2017).
19. Gabarron, E., Oyeyemi, S. O. & Wynn, R. COVID-19-related misinformation on social media: a systematic review. Bull. World
Health Organ. 99 (6), 455 (2021).
20. Loader, B. D. & Mercea, D. Networking democracy? Social media innovations and participatory politics. Inform. Communication
Soc. 14 (6), 757–769 (2011).
21. Von Hippel, E. Democratizing innovation: the evolving phenomenon of user innovation. J. für Betriebswirtschaft. 55, 63–78 (2005).
22. Lüthje, C., Herstatt, C. & Von Hippel, E. User-innovators and local information: the case of mountain biking. Res. Policy. 34 (6),
951–965 (2005).
23. Gaenssle, S. Income distribution and nudity on social media: attention economics of Instagram stars. Kyklos 77 (1), 184–212
(2024).
24. Pugno, M. Social media effects on well-being: the hypothesis of addiction of a new variety. Kyklos 77 (3), 690–704 (2024).
25. Seethaler, S., Evans, J. H., Gere, C. & Rajagopalan, R. M. Science, values, and science communication: competencies for pushing
beyond the deficit model. Sci. Communication. 41 (3), 378–388 (2019).
26. Trench, B. Towards an analytical framework of science communication models. In Communicating Science in Social Contexts (eds
Cheng, D. et al.) (Springer, Dordrecht, 2008).
27. Nerghes, A., Mulder, B. & Lee, J. S. Dissemination or participation? Exploring scientists’ definitions and science communication
goals in the Netherlands. Plos one 17(12), e0277677 (2022).
28. Anderson, P. S., Odom, A. R., Gray, H. M., Jones, J. B., Christensen, W. F., Hollingshead,T., … Seeley, M. K. (2020). A case study
exploring associations between popular media attention of scientific research and scientific citations. PLoS One, 15(7), e0234912.
29. Thelwall, M. & Stuart, D. RUOK? Blogging communication technologies during crises. J. Computer-Mediated Communication. 12
(2), 523–548 (2007).
30. Farrow, K., Grolleau, G. & Mzoughi, N. What in the word! The scope for the effect of word choice on economic behavior. Kyklos 71
(4), 557–580 (2018).
31. Pollett, S. & Rivers, C. Social media and the new world of scientific communication during the COVID-19 pandemic. Clin. Infect.
Dis. 71 (16), 2184–2186 (2020).
32. Eichenberger, R. et al. Certified coronavirus immunity as a resource and strategy to cope with pandemic costs. Kyklos 73, 464–474
(2020).
33. Hanitzsch, T. & Berganza, R. Explaining journalists’ trust in public institutions across 20 countries: media freedom, corruption,
and ownership matter most. J. Communication. 62 (5), 794–814 (2012).
34. Nielsen, R., Fletcher, R., Newman, N., Brennen, J. & Howard, P. Navigating the ‘infodemic’: How People in Six Countries Access and
rate news and Information about Coronavirus (Reuters Institute for the Study of Journalism, 2020).
35. Malecki, K. M., Keating, J. A. & Safdar, N. Crisis communication and public perception of COVID-19 risk in the era of social
media. Clin. Infect. Dis. 72 (4), 697–702 (2021).
36. Lwin, M. O., Lee, S. Y., Panchapakesan, C. & Tandoc, E. Mainstream news media’s role in public health communication during
crises: Assessment of coverage and correction of COVID-19 misinformation. Health Commun. 38 (1), 160–168 (2023).
37. Rocha, Y. M. et al. L. D. The impact of fake news on social media and its influence on health during theCOVID-19 pandemic: a
systematic review. J. Public Health, 31, 1007–1016 (2023).
38. Azzimonti, M. & Fernandes, M. Social media networks, fake news, and polarization. Eur. J. Polit. Econ. 76, 102256 (2023).
39. Tsfati, Y. et al. Causes and consequences of mainstream media dissemination of fake news: literature review and synthesis. Annals
Int. Communication Association. 44 (2), 157–173 (2020).
40. 50, Yoo, E., Rand, W., Eftekhar, M. & Rabinovich, E. Evaluating information diffusion speed and its determinants in social media
networks during humanitarian crises. J. Oper. Manag. 45, 123–133 (2016).
41. Kim, J., Bae, J. & Hastak, M. Emergency information diffusion on online social media during storm Cindy in US. Int. J. Inf. Manag.
40, 153–165 (2018).
42. Melki, J. et al. Mitigating infodemics: The relationship between news exposure and trust and belief in COVID-19 fake news and
social media spreading. Plos one 16(6), e0252830 (2021).
43. Aranzales, I. et al. Scientists have favorable opinions on immunity certificates but raise concerns regarding fairness and inequality.
Sci. Rep. 11, 14016 (2021).
44. Forrester, A. C. & Nowrasteh, A. Trust plays no role in regional U.S. economic development - and five other problems with the trust
literature. Kyklos 76 (3), 461–477 (2023).
45. Ahlerup, P., Olsson, O. & Yanagizawa, D. Social capital vs institutions in the growth process. Eur. J. Polit. Econ. 25 (1), 1–14 (2009).
Author contributions
R.C., G.D.,A.M. and D.S. wrote the main manuscript text and reviewed the manuscript.
Funding
Open Access funding enabled and organized by Projekt DEAL.
Declarations
Competing interests
The authors declare no competing interests.
Funding
Funding for this work was provided by the Deutsche Forschungsgemeinschaft (DFG, German Research
Foundation) under Germany’s Excellence Strategy – EXC 2052/1–390,713,894 (D.S.). Funded by the Open
Access Publishing Fund of the University of Bayreuth (D.S.).
Scientific Reports | (2025) 15:5506
9
| https://doi.org/10.1038/s41598-025-88973-7
www.nature.com/scientificreports/
Additional information
Supplementary Information The online version contains supplementary material available at https://doi.org/1
0.1038/s41598-025-88973-7.
Correspondence and requests for materials should be addressed to R.C., G.L., A.M. or D.S.
Reprints and permissions information is available at www.nature.com/reprints.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which
permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give
appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and
indicate if changes were made. The images or other third party material in this article are included in the article’s
Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included
in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or
exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy
of this licence, visit http://creativecommons.org/licenses/by/4.0/.
© The Author(s) 2025
Scientific Reports | (2025) 15:5506
10
| https://doi.org/10.1038/s41598-025-88973-7
www.nature.com/scientificreports/
People & roles
Origins & context
- Title
- The media morphosis of science communication during crises
- Publication type
- Article
- Language
- English
- Journal
- Scientific Reports
- Year
- 2025
- volume
- 15
- pages
- 5506
- Status
- Peer reviewed
- Relation
- Open Access Publizieren
Identifiers & sources
- Source ID (eref-/epub-)
- eref-96031
- ISSN
- 2045-2322
Loading the visualisations…
Knowledge graph
Loading the knowledge graph…