Why did the influence of experts erode during the COVID-19 pandemic?

Abstract

In the face of protracted crises like climate change or pandemics, the influence of expert scientific projections on public policy is crucial yet evolves over time. This study offers an empirical demonstration of a previously fragmented theory: the diminishing influence of scientific projections on policy over time. Using a comprehensive mixed-method analysis, the article studies the relationship between expert projections, policy stringency and public support in Quebec during the COVID-19 pandemic. Scientific projections that put forward worst-case scenarios have a considerable impact on policies made in the early stages of a crisis. However, as these catastrophic projections instil a sense of fatalism as the crisis lasts, they inadvertently lead to diminished public support for both the policies and the scientific projections themselves. The implications of these findings for scientists and experts are discussed, highlighting the importance of adapting projections and knowledge communication strategies as the crisis unfolds.

Full text

1
Policy & Politics • vol XX • no XX • 120 • © Authors 2024
Print ISSN 03055736 • Online ISSN 14708442 • https://doi.org/10.1332/03055736Y2024D000000064
Accepted for publication 10 December 2024 • First published online 06 January 2025
This article is distributed under the terms of the Creative Commons Attribution 4.0
International license (http://creativecommons.org/licenses/by/4.0/).
research article
Why did the influence of experts erode during the
COVID-19 pandemic?
Antoine Claude Lemor, antoine.lemor@umontreal.ca
María Alejandra Costa, maria.alejandra.costa@umontreal.ca
Université de Montréal, Canada
Louis-Robert Beaulieu-Guay, louis-robert.beaulieu-guay@umontreal.ca
University of Saskatchewan, Canada
Éric Montpetit, e.montpetit@umontreal.ca
Université de Montréal, Canada
In the face of protracted crises like climate change or pandemics, the influence of expert scientific
projections on public policy is crucial yet evolves over time. This study offers an empirical
demonstration of a previously fragmented theory: the diminishing influence of scientific projections
on policy over time. Using a comprehensive mixed-method analysis, the article studies the
relationship between expert projections, policy stringency and public support in Quebec during
the COVID19 pandemic. Scientific projections that put forward worst-case scenarios have a
considerable impact on policies made in the early stages of a crisis. However, as these catastrophic
projections instil a sense of fatalism as the crisis lasts, they inadvertently lead to diminished public
support for both the policies and the scientific projections themselves. The implications of these
findings for scientists and experts are discussed, highlighting the importance of adapting projections
and knowledge communication strategies as the crisis unfolds.
Keywords epidemiological projections • experts • crisis policy • COVID19 • pandemic • science
and policy • time and political decision
To cite this article: Lemor, A.C., Costa, M.A., Beaulieu-Guay, L.R. and Montpetit, É. (2024) Why
did the influence of experts erode during the COVID19 pandemic?, Policy & Politics, Early View,
DOI: 10.1332/03055736Y2024D000000064
Introduction
The power of scientic experts in shaping public policy is well documented. Experts
inuence the formation of governmental agendas, problem denition, policy design,
instrument choice and regulatory decisions (Jasano, 1990; Baumgartner and Jones,
1993; Rochefort and Cobb, 1993; Schneider and Ingram, 1997). What is less well
Unauthenticated | Downloaded 02/06/25 10:04 PM UTC

Antoine Claude Lemor et al
2
known, however, is the duration of this inuence and its relationship to other policy
inputs, notably public opinion. Evidence suggests that the impact of scientic expertise
on governmental decision-making is not consistent over time. For instance, in the
context of the climate crisis, studies have highlighted the growing resistance to
scientic advice from inuential interest groups (Oreskes and Conway, 2010; Dryzek
et al, 2011; Mildenberger and Leiserowitz, 2017). During the COVID-19 crisis, some
research showed that the growth of scientic controversies eroded the inuence of
experts (Cairney, 2021; Armingeon and Sager, 2022; Easton et al, 2022; Eichenberger
et al, 2023). Nevertheless, there is still a great deal of work to be done to theorise
uctuations in expert inuence over time, a central aim of this article.
The COVID-19 crisis provides a valuable opportunity to build on this research. Not
only did experts play a central role during the pandemic, but the decisions made were
also grounded in a form of scientic knowledge particularly suited to support the policies
implemented. However, integrating scientic information into public policy can be a
challenge (Sarkki et al, 2014). The COVID-19 crisis allows us to focus on scientic
projections, a type of scientic information that lends itself to immediate use by policy
makers in the context of an emergency. Despite their utility, scientic projections may
not maintain their inuence throughout the entire duration of a crisis, partly due to
their relationship with public opinion (Norgaard, 2011).To what extent can experts’
scientic projections, even when catastrophic, sustain their inuence over time? And what
mechanisms underpin this inuence? These are the questions this article aims to address.
Using a unique and diverse data set, we empirically show that catastrophic expert
projections do inuence public policies, but their impact diminishes over time as
public and interest-group support for government policies wanes. Our study focuses
on Quebec’s COVID-19 response and utilises original data on expert projections
of hospitalisations, public opinion, interest-group activity and public adherence to
health measures (Rocheleau, 2017; Lemor et al, 2023; Lemor, 2024). We additionally
analyse the full transcripts of all press conferences given by Quebec policy makers
during the COVID-19 crisis.
Theory
Scientic modelling and projections occupy a central position in disciplines like
climatology, biology and public health. Providing future estimations based on current
conditions, they can also inform policy decisions. Projections and models, in climate
change, for example, alert policy makers to potential consequences of inaction,
such as failing to reduce CO2 emissions or prepare for extreme weather events
(Pachauri et al, 2015; Martel-Morin and Lachapelle, 2022). Projections oer experts
a platform to inuence public policy through clear communication of complex
actionable knowledge. In the realm of public health, epidemiological projections are
instrumental during epidemics. Using mathematical models, they predict future health
outcomes based on alternative health interventions (Djidjou-Demasse et al, 2020).
Epidemiological projections provide policy makers with information on the impact
of the various options available to them before they make a decision.
In the context of global warming, expert projections often depict catastrophic
scenarios, such as the occurrence of extreme events (Koubi, 2019). Likewise, during
the COVID-19 pandemic, experts provided a range of projections, with early
scenarios suggesting catastrophic outcomes like hundreds of thousands of deaths and
Unauthenticated | Downloaded 02/06/25 10:04 PM UTC

Why did the influence of experts erode during the COVID-19 pandemic?
3
overwhelmed healthcare systems (Pueyo, 2020; James et al, 2021; Ioannidis et al, 2022).
These catastrophic projections called for policy responses, which they successfully
triggered. While it seems obvious that the catastrophic nature of scientic projections
often induces an immediate response from decision makers, we might ask whether
it encourages sustained political commitment over time.
Temporal dynamics of experts’ influence
At a crisis’s onset, the urgency to act against the problem can endow experts’ scientic
projections with considerable inuence (Auld et al, 2021: 710). The uncertainty
inherent in the early stages of a crisis induces policy makers to welcome the advice
of individual experts or epistemic communities (Haas, 1992; van Asselt and Vos,
2006; Lemor and Montpetit, 2024). During the COVID-19 pandemic, it was at the
beginning of the crisis that expert recommendations indeed appeared to have been
followed most by policy makers (Eichenberger et al, 2023), an observation also made
in other health crises such as H1N1 (Salajan et al, 2020).
However, once the initial stage of the crisis has passed, once expert advice has
successfully reduced uncertainty, the demand for expert information diminishes.
The attention of policy makers then shifts towards public opinion, the competition
of interests and other such consideration (Löblová, 2018: 181). Policy learning may
be at the origin of the shift (Dunlop, 2014). Research has shown that in the early
stages of problem solving, instrumental learning gives signicant inuence to experts.
However, in subsequent stages, policy makers learn that solving problems is sometimes
insucient and need to give heed to psychosocial factors, including public discontent
(Zaki etal, 2023).
Furthermore, the consensus that may exist at the onset of a crisis can dissipate
as time passes (Montpetit, 2011). When a problem requires instrumental learning,
policy makers usually turn to trusted experts who enjoy some form of government
certication (Dunlop, 2014). Some researchers have shown how certied experts
have exerted signicant inuence at the start of the pandemic, at the expense of
other experts (Cairney, 2021; Cairney and Toth, 2023). In some places, however,
these experts were quickly challenged. In Sweden, for example, experts who felt their
perspective was being ignored by the government publicly disagreed with the approach
proposed by the public health agency, the source of certied expertise in that country
(Vetenskapsforum COVID-19, 2023; Lemor et al, 2024). Such disagreements between
experts can reduce their inuence and benet other inputs, such as public opinion.
The power of worst-case scenarios
Given that scientic projections are uncertain, they often lead to multiple scenarios,
ranging from the most optimistic to the most pessimistic (Pearce, 2020; Rhodes
and Lancaster, 2020). Experts from the Intergovernmental Panel on Climate
Change (IPCC), for example, estimate temperature increases based on various CO2
emission levels, because of the uncertainty surrounding future emissions. Dierent
assumptions on emission levels thus lead to various projections, some optimistic
and others more pessimistic (Pachauri et al, 2015). Considering these looming
Unauthenticated | Downloaded 02/06/25 10:04 PM UTC

Antoine Claude Lemor et al
4
uncertainties, cautious climate modellers might opt to give weight to the worst-
case scenario (Brister etal, 2021).
Uncertainties, coupled with the fear of losses, tend to elicit profound human
reactions (Kahneman and Tversky, 1979; Kahneman, 2011), making worst-case
scenarios more powerful than optimistic ones. Research has also shown that framing
issues negatively bolsters public policy support (Avdagic and Savage, 2021). Hausfather
and Peters (2020) explain that modellers may thus lean towards the worst-case scenario,
even when they acknowledge its low likelihood, because it maximises impact. During
the COVID-19 pandemic, a grim scientic projection from Imperial College is
believed to have expedited initial lockdowns in many countries, although it was later
deemed to have been poorly estimated (James et al, 2021). Worst-case scenarios were
indeed inuential early in the pandemic of COVID-19, despite knowledge of the
limitations of scientic projections (Pearce, 2020).
Human cognition plays a pivotal role here, interpreting threats in ways that trigger
feelings of fear, which in turn focus attention and elicit protective reactions (for
example, Caplin and Eliaz, 2003; Kahneman, 2011; Maor, 2020; Martel-Morin and
Lachapelle, 2022). The admission of the Finnish prime minister of having closed
schools ‘out of fear’ provides a good illustration (Maor and Howlett, 2020). The fear
caused by worst-case scenarios among the public can, in turn, foster momentum
for reassuring public policies (Maor et al, 2020). This dynamic frequently underpins
sudden and bold policy changes (Jones and Baumgartner, 2005; Sha and Mallinson,
2023), a phenomenon observed in many health crises (Wilkinson et al, 2010). In other
words, the literature shows that both scientic modellers and policy makers, in many
circumstances, are led to prioritise worst-case scenarios over more optimistic scenarios.
The double-edged sword of catastrophic projections
Worst-case scenarios and their inherent negativity can have deleterious eects in the
long run (Martel-Morin and Lachapelle, 2022). Also having to deal with uncertainty,
the public and interest groups might be receptive to the direst experts’ projections
early in a crisis, even if these projections mandate sacrices on their part (Kreps and
Kriner, 2020). They might in fact be willing to bear a substantial cost to prevent
foreseen disasters, if convinced the endeavour is worthwhile (Zografakis et al, 2010).
This willingness, however, might not endure.
After experiencing a crisis and shouldering the necessary sacrices, individuals could
acclimatise to the new reality to the extent they become less responsive to repeated
catastrophic projections (Martel-Morin and Lachapelle, 2022), especially when these
forecasts become controversial in a polarised political climate (Kreps and Kriner, 2020).
As the public gradually becomes desensitised to a crisis, there may emerge a sentiment
that scientists are overstating risks (Borick and Rabe, 2017). Persistent catastrophic
projections, in light of prolonged collective sacrices, might have a demoralising eect,
fostering fatalism. The public may grow sceptical towards scientic projections, even
if they originate from credible sources (Hausfather and Peters, 2020). They could
become increasingly reluctant to bear the cost of policies – whose impacts are slow
to alleviate the crisis – and become more critical of policy makers who continue
to take experts seriously (Mayer and Smith, 2019). As seen during the COVID-19
pandemic, trust in scientists can wane over time (Algan et al, 2021).
Unauthenticated | Downloaded 02/06/25 10:04 PM UTC

Why did the influence of experts erode during the COVID-19 pandemic?
5
The longer the crisis drags on, the more policy makers will grapple with deciding
whether experts or the public should inform their policy decisions. During the
pandemic, the relation between policy makers and experts indeed became, over time,
increasingly contentious and experts’ inuence diminished (Armingeon and Sager,
2022; Easton et al, 2022; Eichenberger et al, 2023; Zaki et al, 2023). In a democratic
setting, elected ocials cannot wholly distance themselves from public opinion and
other interests. And as the public becomes distrustful of scientic advice, the inuence
of experts on policies to overcome a crisis may diminish over time.
In the eld of climate change, projections have prompted the public to accept
the sacrices involved in policies aimed at keeping temperatures below a threshold
beyond which, according to experts, catastrophes will occur. These grim climate
change forecasts can also instil a sense of fatalism and powerlessness within the public,
stemming from a sentiment that the worst is, in any case, inevitable (Lertzman, 2015).
Thus, even when confronted with compelling evidence and catastrophic projections,
many individuals might resort to denial, seeking solace in the belief that individual
sacrices are incapable of addressing the problem (Norgaard, 2011). Some researchers
indeed highlight that the current communication strategy by experts on climate
change, which often emphasises impending disasters, might not be the most eective
in motivating action (Lertzman, 2015; Martel-Morin and Lachapelle, 2022).
Hypotheses about the influence of scientific projections over time
This study investigates two related hypotheses regarding the inuence of experts
and their projections during the COVID-19 pandemic. Our rst hypothesis suggests
that expert projections were initially inuential, echoing ndings from other health
crises (Salajan et al, 2020), with this inuence being contingent on public support.
Specically, we hypothesise that expert projections during the pandemic informed the
stringency of non-pharmaceutical interventions (NPIs) to the extent that the public
supported them. The second hypothesis explores the idea that the impact of experts’
catastrophic projections on public policy declines over time. We thus hypothesise that
the longer the COVID-19 crisis persisted, the less support there was for NPIs, and
the less expert projections informed decisions.
In a democratic setting, policy makers must maintain public support and if the
public loses faith in expertise, the inuence of experts declines. The challenge
comes from catastrophism, which appeals to scientic modellers and provokes rapid
political reactions. In the short term, reactions take the form of policies, often bold
ones, that impose major sacrices on the public and groups, which accept them for
a period. In the longer term, however, reactions to catastrophism take the form of
fatalism, which feeds distrust and denial of expert advice. The inuence of scientic
experts then diminishes. We formulated these hypotheses in response to recent calls
for a more comprehensive understanding of how experts inuence public policy
(Christensen, 2021).
The existing literature on crisis management, pandemics in particular, has not
fully examined the relative inuence of experts in time. Studies examine the role of
epidemiological data, of hospital capacities, of trust, of inter-country diusion, and of
policy makers’ attitudes towards the rapidity or stringency of policies, but none have
explored the inuence of experts, in relation to other forces, throughout the course
Unauthenticated | Downloaded 02/06/25 10:04 PM UTC

Antoine Claude Lemor et al
6
of a public health crisis (Sebhatu et al, 2020; Forster and Heinzel, 2021; Toshkov etal,
2021; Bourdin et al, 2022; Jalloh et al, 2022; Wang et al, 2024). Eichenberger etal
(2023) comes closest to doing so but fail to examine the relationship between expert
inuence and public support.
Our study employs a dual database approach, utilising mixed methods to analyse
both quantitative and qualitative data. On one hand, it includes expert hospitalisation
projections, data on interest groups, public adherence and public opinion (Lemor etal,
2023). On the other, it encompasses all transcriptions of press conferences held by decision
makers in Quebec, Canada, throughout the pandemic (Lemor, 2024). This comprehensive
methodology allows for an in-depth analysis that highlights the diminishing inuence of
experts over time in relation to other factors that matter in a democratic setting.
Methodology and data
The COVID-19 crisis provides an excellent opportunity to test our hypotheses. First,
infectious diseases like COVID-19, akin to climate change, call for scientic modelling,
as seen through widespread epidemiological projections reported in the media
during the pandemic’s early stages (Ford, 2020). Second, given the contagiousness
of SARS-CoV-2 and the virulence of its induced disease for a signicant portion of
the population, epidemiologists and mathematicians can produce projections that can
be construed as catastrophic without jeopardising their credibility, just as modelling
scientists can do it for climate change. Third, the shorter duration of the COVID-19
crisis compared to climate change oers an opportunity for continuous observation
from beginning to end. Finally, the rapid and successive case increases and decreases
during the COVID-19 crisis provide an opportunity to examine causal relationships
between scientic expertise, public opinion and policy, a more challenging task in
climate studies due to slower variations over time, notably in projections.
Quebec as a case study
Quebec, a Canadian province, represents an exemplary case for examining our
hypothesis due to its strategic use of NPIs throughout the pandemic, such as lockdowns
and venue closures (Han and Breton, 2022). Provincial authorities in Canada had
the responsibility for NPIs, except for managing travel restrictions. Acknowledging
its reliance on epidemiological projections, the Quebec government began publicly
releasing these projections from the second wave in September 2020. Our research
incorporates public opinion and adherence data from March 2020 to February 2022
and extensive qualitative analysis from press conference transcripts, providing insight
into policy making during the initial and later pandemic waves (Lemor et al, 2023;
Lemor, 2024). Despite the unavailability of hospitalisation projections in the rst
wave, our data set remains relatively comprehensive, covering this crisis extensively.
Epidemiological projections
The epidemiological projections employed in our study were produced by the
National Institute for Excellence in Health and Social Services (INESSS) at the
Unauthenticated | Downloaded 02/06/25 10:04 PM UTC

Why did the influence of experts erode during the COVID-19 pandemic?
7
request of the Quebec government. These projections, based on a statistical model,
forecasted weekly hospitalisations in both regular and intensive care units. While
the Quebec government initially used informal epidemiological insights, INESSS
was formally commissioned to produce weekly projections from the pandemic’s
second wave, with data being public from September 2020. These projections,
which incorporated cumulative infection numbers and vaccination counts, led us
to use death-count data in our model to avoid multicollinearity. Notably, unlike
climate change forecasts, these epidemiological projections exhibited signicant
short-term uctuations, reecting the various stages and severity of the pandemic.
This variability was critical for assessing the inuence of epidemiology on policy
decisions, particularly how projections of increased hospitalisations might lead to
more stringent NPIs.
Policy
Our second hypothesis suggests that catastrophic expert projections initially inuenced
public policies, but gradually lost their inuence over time. During the pandemic, NPIs
had been closely monitored and quantied. The Oxford Coronavirus Government
Response Tracker’s method for assessing policy stringency, adapted by the Institute
for Research on Public Policy (IRPP) for the Canadian context, serves as a basis for
our analysis (Hale et al, 2021; IRPP, 2021; 2022). We use the IRPP’s stringency index,
which rates policy stringency on a scale from 0 to 100, based on 12 NPIs like curfews,
mask mandates and closures of various venues. A weekly average of this index was
created to ensure consistency in our analysis.
Time and public support
According to our theory, the diminishing inuence of scientic modellers over time
is closely tied to public support. As time progresses, the public tends to become less
responsive to experts’ catastrophic projections and, consequently, less supportive of
policies based on these forecasts. To measure this, we utilise Quebec government-
mandated surveys that assess public support for actual NPIs (Gouvernement
du Québec, 2020). However, public opinion on policies is known to uctuate
thermostatically, dissatisfaction leading to policy adjustments, which then alter
public satisfaction levels (Soroka and Wlezien, 2009). This dynamic is particularly
evident during a pandemic, where the strictness of NPIs may gradually erode
public support, nudging the government towards more permissive measures. In
turn, more permissiveness can either restore or boost public support levels, making
it dicult for the government to determine, based on opinion data, whether the
public would be likely to accept greater NPI stringency in the face of a situation
that would justify it.
Therefore, we consider public adherence as an additional indicator. We use an
adherence index from the Quebec National Institute of Public Health (INSPQ),
which is based on surveys evaluating public attitudes towards general public
health measures like hand hygiene, physical distancing and participating in large
gatherings (INSPQ, 2022). Negative attitudes towards these measures, indicative
Unauthenticated | Downloaded 02/06/25 10:04 PM UTC

Antoine Claude Lemor et al
8
of low adherence, suggest to governments that intensifying NPIs might negatively
impact public support, irrespective of what expert projections might indicate about
potential hospitalisations.
Interest groups
Interest-group activity might compete with expert projections as an input into
decisions on NPIs’ stringency, as shown in studies on the inuence of interest groups
on regulatory decisions (Beaulieu-Guay et al, 2021). Early in a crisis, uncertainty
may lead groups to cautiously consider expert projections, even if the associated
policies adversely aect their interests. Over time, however, as the burden of stringent
policies intensies, these groups might increasingly pressure governments to lessen
NPI stringency, regardless of whether projections suggest maintaining it to avoid
hospital overloads.
For our analysis, we assess interest-group activity through media content, using
a comprehensive Canadian media database (Rocheleau, 2017). We focus on articles
from Quebec newspapers related to COVID-19, specically targeting four sectors
most aected by NPIs: restaurants, cultural venues, retail businesses and non-essential
services. Out of the 12,687 articles gathered about these groups, we randomly selected
1,630 for manual coding to identify mentions of interest-group activity. We further
distinguished articles indicating dissatisfaction with Quebec’s COVID-19 NPIs. We
subsequently computed a monthly average of articles conveying the dissatisfaction
of interest groups in the four sectors.
Empirical test and methodology overview
Our study utilises a comprehensive database to empirically examine the trade-o that
emerges over time between expert modellers’ inuence on policy and that of the public
and groups. This database includes qualitative data from press conferences since the
pandemic’s beginning (Lemor, 2024) and quantitative data on NPIs’ stringency, public
support, hospitalisation projections, public adherence to health measures, and interest-
group activity, all indexed or scaled out of 100 and updated weekly (Lemor et al, 2023).
We rst conduct a descriptive analysis to observe data trends supported by
qualitative analysis. We then perform an OLS regression to examine the impact of
expert projections on policy stringency while accounting for mortality and public
adherence. We nally incorporate into the model an interaction between expert
projections and public adherence to gauge the extent to which expert projections
are contingent on public support. This approach enables estimations of the extent to
which expert projections had inuence, as well as the extent to which public support
mediates their impact on policy over time.
Results
Hypothesis 1: Expert projections during the pandemic infor med the stringency
of NPIs, to the extent that the public supported them.
Unauthenticated | Downloaded 02/06/25 10:04 PM UTC

Why did the influence of experts erode during the COVID-19 pandemic?
9
Descriptive and qualitative analyses
The Quebec government implemented strict NPIs as early as 12 March 2020, to
‘atten the curve’ of COVID-19 contagion, as stated by the National Director of
Public Health. Beyond the suite of measures announced on 12 March to reduce
transmission, a health emergency decree was issued on 13 March. Our qualitative
analysis indicates that these initial policies were shaped by experts predicting an
imminent disaster if the government opted for a tepid response.
For instance, on 12 March, the National Director of Public Health referenced
projections (see Appendix, Table 1, citation 1) including alarming contagion scenarios
(see Appendix, Table 1, citation 3), while acknowledging their uncertainty (see
Appendix, Table 1, citations 2 and 3). He also mentioned public fears (see Appendix,
Table 1, citation 4) and suggested that this justied serious consideration of the bleakest
Figure 1: Distribution of descriptive data over time
Unauthenticated | Downloaded 02/06/25 10:04 PM UTC

Antoine Claude Lemor et al
10
projections and the adoption of particularly strict policies (see Appendix, Table 1,
citation 5). The analysis of press conferences from the rst wave thus conrms that
catastrophic projections by experts exerted signicant inuence during the crisis’s
early stages, both on the public and governmental authorities.
Our quantitative data, starting from the second wave, are presented in Figure1
and further bolster our hypothesis. While the top graph shows raw data, the bottom
graph presents smoothed data. The descriptive data suggest that stringency varied
along expert projections during the early waves of the crisis. Following a slowdown
in the summer of 2020, INESSS experts anticipated a new disaster in the autumn,
predicting a peak of nearly 900 hospitalisations in December. These projections have
given rise to serious concerns over Quebec’s limited hospital capacity compared
to other jurisdictions (Corniou, 2020). Throughout the pandemic, policy makers
endeavoured to bolster this capacity by instructing hospitals to defer less urgent
surgeries. In December 2020, hospitals were advised to discharge 50 per cent of
surgical patients, postponing numerous cancer treatments in anticipation of projected
COVID-19 hospitalisations (Lacoursière and Jean, 2020; Radio-Canada, 2022).
Unsurprisingly, the successive expert projections of autumn 2020 were followed by
incremental escalations in the stringency of NPIs, lasting through to January 2021.
The index peaked in early 2021 following the government’s decision to implement
a curfew, a unique move in North America. Intriguingly, this heightened stringency
did not undermine public support for the government’s crisis approach.
While our data indicates that interest groups were relatively prominent in the
press at the onset of the second wave, Figure 1 illustrates their seldom criticism of
the government, contrary to subsequent waves. Consequently, there was a prevailing
consensus both from the public and interest groups concerning health measures and
the projections informing them. This early crisis consensus mirrors ndings in various
opinion surveys conducted in Quebec during this period (Leger, 2023).
Hospitalisation forecasts began waning in January 2021 and persisted at
comparatively lower levels during Waves 3 and 4, relative to their previous levels. In
alignment with expert projections, the stringency of NPIs had markedly subsided by
June 2021 and remained at relatively low levels during Wave 4, a wave wherein the
projections were least worrisome. Unsurprisingly, public support remained robust
throughout Waves 3 and 4.
Quantitative analysis
To further investigate our hypothesis, we employed a regression model (Table 1) and a
projection plot to determine whether expert projections, moderated by public support,
inuenced public policies. In the OLS analysis, projections concerning future COVID-19
hospitalisations (variable ‘Projections’) were statistically signicant, indicating a negative
association with the stringency of NPIs (coecient = −0.302, p < .01). Although this
relationship might seem counterintuitive, it aligns with our hypothesis. The interaction
term (Projections * Adherence, coecient = 0.011, p < .01) reveals that expert projections
do increase the stringency of NPIs, but only to the extent that the public adheres to
health measures. The negative coecient for the variable Projections alone is explained
by low adherence during certain stages of the crisis. In summary, the relationship between
projections and stringency is mediated by public adherence to health measures.
Unauthenticated | Downloaded 02/06/25 10:04 PM UTC

Why did the influence of experts erode during the COVID-19 pandemic?
11
The population’s adherence level to the NPIs (variable ‘Adherence’) has a signicant
positive coecient of 0.632 (p < .001), highlighting its substantial role in determining
stringency. The variable measuring deaths from COVID-19 is not statistically
signicant in this model.1 With an adjusted R^2 of 0.967, the model explains a large
part of the variation in the stringency of NPIs.
In addition to the regression table, we produce an interaction plot to visualise the
relationship between Projections and Adherence, as estimated by our model (Figure2) .
The lines of the graph represent three levels of public adherence to NPIs. The
interaction graph facilitates the visualisation of the relationship between hospitalisation
projections and NPIs’ stringency and highlights dierences according to adherence
levels. For instance, if a line displays a steeper gradient at a higher adherence level,
it suggests that when the public strongly adheres to the measures, hospitalisation
projections have a more pronounced eect on stringency.
Visualising these predictions in the interaction plot, several patterns emerge.
For the minimal adherence level, the relationship between hospitalisation
projections and the severity of health measures is negative, indicating that low
adherence is associated with a decrease in policy stringency, even as the projected
hospitalisations increase. At the median adherence level, the relationship becomes
positive, suggesting that moderate adherence is related to an increase in policy
stringency when projected hospitalisations increase. At the maximum adherence
level, the slope is markedly positive, indicating that a very high level of adherence
produces a substantial increase in policy stringency, even with modest increases
in the projected hospitalisations.
In summary, the ndings indicate that expert projections were indeed inuential
during the pandemic in Quebec, but their impact was strongly mediated by
public adherence. Decisions to increase the stringency of NPIs in Quebec were
Table 1: Effects of experts’ projections and adherence (including interaction effects) on
policy stringency
OLS 1
(Intercept) −7.075 (3.377)*
Stringency −1 0.617 (0.060)***
Death −0.015 (0.080)
Projections −0.302 (0.106)**
Adherence 0.632 (0.125)***
Projections * adherence 0.011 (0.004)**
Num. Obs. 80
R2 0.967
R2 Adj. 0.964
AIC 431.2
BIC 447.9
Log.Lik. 429.802
F 3.28
Notes: + p < .1, * p < .05, ** p < .01, *** p < .001.
Unauthenticated | Downloaded 02/06/25 10:04 PM UTC

Antoine Claude Lemor et al
12
informed by expert projections, but only when the public showed willingness
to accept NPIs.
Hypothesis 2: The longer the COVID-19 crisis persisted, the less support
there were for NPIs, and the less expert projections
informed decisions.
Overall analysis
The emergence of the Omicron variant in late November 2021 altered the course of
the pandemic. Towards the end of the fourth wave, before Omicron’s detection, the
Quebec government adopted a reassuring stance. Bolstered by favourable projections
in October and an exceptional vaccination coverage, the government committed to
relaxing restrictions on gatherings for the holiday season, enabling family reunions
(see Appendix, Table 1, references 8 and 9). The National Director of Public Health
acknowledged then the eort of the population and cautioned against a potential
descent into public fatalism if the stringency of NPIs remained high (see Appendix,
Table 1, references 10 and 11). However, this optimism was excessive in the face of
an ever-evolving virus. Indeed, the arrival of the of Omicron variant brought back
catastrophic projections. Omicron being considerably more contagious than previous
variants, INESSS forecasts predicted deadly hospital saturation were NPIs’ stringency
30
40
50
60
70
0255075100
Projections
Stringency
Adherence leve
l
min
med
max
Interaction between projections and adherence
Figure 2: Effects of experts’ projections on stringency by adherence levels (minimum,
median, maximum)
Unauthenticated | Downloaded 02/06/25 10:04 PM UTC

Why did the influence of experts erode during the COVID-19 pandemic?
13
to remain at the level of the fourth wave. As illustrated in Figure 1, INESSS projected
hospitalisations to peak at 2,045 by early January 2022, the highest level since the
pandemic’s onset, despite high vaccination coverage. Two years into the crisis, these
projections were particularly alarming, not only due to the unprecedented number
of hospitalisations, but also because healthcare capacity continued to be a challenge
(Lacoursière and Chouinard, 2021).
The January 2022 projections prompted the government to retract its earlier
announcement regarding holiday-season gatherings, in addition to boosting the overall
stringency of NPIs. The string of announcements culminated with the reinstatement
of the curfew. The grim forecasts from INESSS once again wielded signicant
inuence on governmental policy at the start of the fth wave. This time, however,
public support began to decline, while interest groups responded unfavourably to
the increase in stringency (Figure 1).
Public support reached a low point in December 2021. Interestingly, it rebounded
in February 2022, only after the government’s announcements of the end of the
curfew and the gradual relaxations in NPIs’ stringency. The adherence index, a less
thermostatic gauge of public support, also began a decline in December 2021 and
consistently decreased until the end of the observation period (Figure 1). Interest
groups vocally aired their grievances and discontent. While 49 per cent of the articles
citing interest-group activity indicated policy dissatisfaction during the pandemic’s
second wave, dissatisfaction surged to 61 per cent during the fth wave. Policy makers
did not conceal their sensitivity to public discontent. Premier François Legault,
expressing his concerns about hospitals, mentioned the public’s growing exasperation
(see Appendix, Table 1, reference 12). He further suggested that he had to start caring
for ‘social cohesion’ and ‘peace’ (see Appendix, Table 1, reference 13). On 8 February
2022, responding to a journalist’s query, the premier hinted that reverting to strict
social distancing policy was now o the table, regardless of projections.
Nonetheless, according to the INESSS projections shown in Figure 1, the crisis
was far from over. A sixth wave was announced in April 2022, with hospitalisation
forecasts approaching those of the fth wave. Yet, this time, the government faced
a challenging decision as public adherence gures signalled a reluctance to support
further escalations in NPIs stringency (Figure 1). Indeed, expert projections were
met with scepticism and fatalism by the public. Surveys, for instance, indicated
that at this stage, most of the population perceived the crisis as denitively over,
marking the public’s widespread fatalism (Leger, 2022: 15). For the rst time
since the pandemic’s onset, grim forecasts failed to drive any increase in the
stringency of NPIs.
The lower graph in Figure 1 shows that at the crisis’s onset (see references 6 and
7, Appendix, Table 1), the public was deferential to experts’ catastrophic predictions.
The population accepted the sacrices deemed necessary to prevent a disaster in
the hospitals. However, after enduring signicant sacrices over a long period and
becoming accustomed to living amid the threat of hospital overcrowding, the public
became increasingly deant of experts’ scientic projections that called for increases
in NPIs stringency (Figure 1). The same was true for interest groups. Less active
during the early stages of the crisis despite the costs imposed on them by severe
public health measures, they became more reactive to increases in policy stringency
as the crisis prolonged. Individuals and groups alike were gradually acclimatising to
the crisis, becoming more fatalistic towards alarming projections. The willingness to
Unauthenticated | Downloaded 02/06/25 10:04 PM UTC

Antoine Claude Lemor et al
14
accept policies based on these forecasts hence waned, as did the expert modellers’
inuence on policy.
Conclusion
This article examines two hypotheses about the inuence of scientic projections
during a crisis and within a democratic context. While we have shown that catastrophic
projections can signicantly impact policy decisions at the onset of a crisis, we have
also shown that the inuence of such projections wanes as time passes. As a crisis starts
to feel interminable, public and interest-group support for costly preventive measures
wanes, along with the inuence of experts advocating bold policy decisions. Over
time, groups become increasingly contentious and the public more critical of policy
makers. Accountable to the public, a democratic government in such circumstances
may gradually cease to rely on alarming scientic projections to inform its policies.
This article contributes to the theorisation of the inuence of scientic expertise
in policy making, particularly in crisis situations (Christensen, 2021). It highlights a
specic category of expertise, scientic projection, which is useful when scientic
information is required quickly, given the urgency of a situation (Sarkki et al, 2014).
Right from the start of the pandemic, policy makers were able to rely on modellers
who were ready to provide useful information for decision making. Projections are
also a type of scientic information that lends itself to catastrophism, a position that
generally does not inspire indierence when its source is credible. Indeed, we have
shown that scientists capable of producing catastrophic projections are likely to have
a signicant inuence on decisions at the start of a crisis. Finally, scientic expertise –
even that of modellers, tailored to the needs of decision makers – does not have a
constant eect over time. Indeed, scientic experts who succeed in convincing policy
makers to take bold decisions, at great cost to the population, may end up losing all
inuence (Hausfather and Peters, 2020). When faced with the choice of which of
the experts or the public should inform their decisions, governments in democratic
countries generally prefer the latter (Soroka and Wlezien, 2009).
It is also worth noting that expert projections are rarely monolithic. In most crises,
including the COVID-19 pandemic, multiple groups of experts may produce diering
or even conicting projections. For example, during the pandemic, policy makers had
access to various projections – some more optimistic, others more catastrophic – which
probably inuenced their decision making dierently at dierent stages of the crisis.
While this study focuses on projections requested by the government from one of its
agencies, the availability of multiple expert projections is likely to add complexity to
how scientic advice is received by policy makers (Montpetit, 2011). This plurality
of projections could, in turn, aect the dynamics of inuence over time and warrants
further examination in future studies.
This study has some limitations worth underlining. First, our single-case study
approach advises caution in generalising. While the description of the Quebec
COVID-19 crisis we provide here shares many common elements with those of other
democratic countries, it also has a few peculiarities. In particular, the strong support
of the Quebec public for health measures lasted longer than elsewhere, which may
suggest singular attitudes in the province towards expertise. Second, we lack time-
series public attitude data explicitly related fatalism during the COVID-19 crisis.
Although we argue that catastrophist projections feed public fatalism over time, we
Unauthenticated | Downloaded 02/06/25 10:04 PM UTC

Why did the influence of experts erode during the COVID-19 pandemic?
15
have no direct observation of this dynamic. This is important, because fatalism plays
a key role in the decline in public support for NPIs, which is the causal mechanism
for the erosion of modellers’ inuence. Similarly, we do not compare dierent
levels of catastrophism, preventing us from examining its precise role in the causal
relationship. Third, our media analysis does not consider the media’s inuence on
public opinion, although a recent study suggests that public opinion precedes media
coverage (Wlezien, 2023). Lastly, our empirical study does not dwell on some notions
that appear in our theoretical reasoning. The eventual abandonment of expert advice
during a crisis could be the result of a learning process, as indicated in our theoretical
section (Dunlop, 2014). For the purposes of this study, however, we did not consider
it necessary to use learning indicators. We leave these tasks for future research on the
inuence of scientic experts.
This article lays the groundwork for research in public policy, as well as having
normative implications in elds as diverse as climate change, political psychology,
communication and crisis management. While much needs to be done to fully
demonstrate the eect of catastrophism over time, our ndings nonetheless advise
caution. Policy makers, modellers, crisis managers and communicators should be aware
that catastrophism, if initially powerful, can also become counterproductive over time.
Note
1 Other controls such as the number of cases, the number of vaccinated individuals or the
number of hospitalisations could have been used. However, several reasons prevent the
inclusion of these controls. First, the models used by INESSS to produce hospitalisation
projections utilised the number of cases, which precludes the use of the number of cases
as a control due to multicollinearity. Second, these models also consider the number of
vaccinated individuals, which also prevents their inclusion as a control for the same reason.
Finally, hospitalisations present signicant challenges because it remains impossible to
determine how many hospitalisations on which date inuence the severity of measures.
Indeed, hospitalisations cannot be reliably used as a predictor of severity in the following
weeks, as an increase in hospitalisations indicates that contamination has already occurred.
Funding
This work was supported by the Social Sciences and Humanities Research Council
(SSHRC) of Canada [grant number 435 2021 0332] and the Fonds de recherche du
Québec (FRQ) [grant number 203845].
Acknowledgements
Authors would like to thank the Centre Interuniversitaire de Recherche sur la Science
et la Technologie (CIRST) for their valuable assistance and collaboration in writing this
article. Olivier Santerre deserves specic thanks for his precious help and insights.
AI declaration
This research article was developed independently by the author, with AI assistance
limited to facilitating the writing of code and translating the text from French to English
using ChatGPT.
Conict of interests
The authors declare that there is no conict of interest.
Unauthenticated | Downloaded 02/06/25 10:04 PM UTC

Antoine Claude Lemor et al
16
References
Algan, Y., Cohen, D., Davoine, E., Foucault, M. and Stantcheva, S. (2021) Trust in
scientists in times of pandemic: panel evidence from 12 countries, PNAS, 118(40):
art e2108576118. doi: 10.1073/pnas.2108576118
Armingeon, K. and Sager, F. (2022) Muting science: input overload versus scientic
advice in Swiss policy making during the COVID-19 pandemic, Political Quarterly,
93(3): 424–32. doi: 10.1111/1467-923X.13151
Auld, G., Bernstein, S., Cashore, B. and Levin, K. (2021) Managing pandemics as super
wicked problems: lessons from, and for, COVID-19 and the climate crisis, Policy
Sciences, 54(4): 707–28. doi: 10.1007/s11077-021-09442-2
Avdagic, S. and Savage, L. (2021) Negativity bias: the impact of framing of immigration
on welfare state support in Germany, Sweden and the UK, British Journal of Political
Science, 51(2): 624–45. doi: 10.1017/s0007123419000395
Baumgartner, F.R. and Jones, B.D. (1993) Agendas and Instability in American Politics,
Chicago: University of Chicago Press.
Beaulieu-Guay, L.R., Tremblay-Faulkner, M. and Montpetit, É. (2021) Does
business inuence government regulations? New evidence from Canadian impact
assessments, Regulation & Governance, 15(4): 1419–35. doi: 10.1111/rego.12313
Borick, C.P. and Rabe, B.G. (2017) Personal experience, extreme weather events, and
perceptions of climate change, in Oxford Research Encyclopedia of Climate Science.
doi: 10.1093/acrefore/9780190228620.013.311
Bourdin, S., Ben Miled, S. and Salhi, J. (2022) The drivers of policies to limit the
spread of COVID-19 in Europe, Journal of Risk and Financial Management, 15(2):
art 67. doi: 10.3390/jrfm15020067
Brister, E., Holbrook, J.B. and Palmer, M.J. (2021) Conservation science and the ethos
of restraint, Conservation Science and Practice, 3(4): art e381. doi: 10.1111/csp2.381
Cairney, P. (2021) The UK Government’s COVID-19 policy: what does ‘guided
by the science’ mean in practice?, Frontiers in Political Science, 3. doi: 10.3389/
fpos.2021.624068
Cairney, P. and Toth, F. (2023) The politics of COVID-19 experts: comparing winners
and losers in Italy and the UK, Policy and Society, 42(3): 392–405. doi: 10.1093/
polsoc/puad011
Caplin, A. and Eliaz, K. (2003) AIDS policy and psychology: a mechanism-design
approach, RAND Journal of Economics, 34(4): 631–46. doi: 10.2307/1593781
Christensen, J. (2021) Expert knowledge and policymaking: a multi-disciplinary
research agenda, Policy & Politics, 49(3): 455–71. doi: 10.1332/030557320x15898
190680037
Corniou, M. (2020) Nombre de lits en soins intensifs: comment le Québec se
positionne-t-il face à la COVID-19?, Québec Science, 25 March, https://www.
quebecscience.qc.ca/sante/nombre-lits-soins-intensifs-quebec-covid19/.
Djidjou-Demasse, R., Selinger, C. and Sofonea, M.T. (2020) Épidémiologie
mathématique et modélisation de la pandémie de Covid-19: enjeux et diversité, Revue
Francophone des Laboratoires, 2020(526): 63–9. doi: 10.1016/s1773-035x(20)30315-4
Downs, A. (2016) Up and down with ecology: the ‘issue-attention cycle’, in D. Protess
and M.E. McCombs (eds) Agenda Setting: Readings on Media, Public Opinion, and
Policymaking, Abingdon: Routledge, pp 27–36.
Dryzek, J.S., Norgaard, R.B. and Schlosberg, D. (eds) (2011) The Oxford Handbook of
Climate Change and Society, Oxford: Oxford University Press.
Unauthenticated | Downloaded 02/06/25 10:04 PM UTC

Why did the influence of experts erode during the COVID-19 pandemic?
17
Dunlop, C.A. (2014) The possible experts: how epistemic communities negotiate
barriers to knowledge use in ecosystems services policy, Environment and Planning
C: Government and Policy, 32(2): 208–28. doi: 10.1068/c13192j
Easton, M., De Paepe, J., Evans, P., Head, B.W. and Yarnold, J. (2022) Embedding
expertise for policy responses to COVID-19: comparing decision-making structures
in two federal democracies, Public Organization Review, 22(2): 309–26. doi: 10.1007/
s11115-022-00629-6
Eichenberger, S., Varone, F., Sciarini, P., Stähli, R. and Proulx, J. (2023) When do
decision makers listen (less) to experts? The Swiss government’s implementation of
scientic advice during the COVID-19 crisis, Policy Studies Journal, 51(3): 587–605.
doi: 10.1111/psj.12494
Ferguson, N.M., Laydon, D., Nedjati-Gilani, G., Imai, N., Ainslie, K., Baguelin,
M., et al (2020) Report 9: Impact of Non-Pharmaceutical Interventions (NPIs)
to Reduce COVID-19 Mortality and Healthcare Demand, 16 March, London:
Imperial College COVID-19 Response Team. doi: 10.25561/77482
Ford, J. (2020) The battle at the heart of British science over coronavirus, Financial Times,
15 April, https://www.ft.com/content/1e390ac6-7e2c-11ea-8fdb-7ec06edeef84.
Forster, T. and Heinzel, M. (2021) Reacting, fast and slow: how world leaders shaped
government responses to the COVID-19 pandemic, Journal of European Public Policy,
28(8): 1299–320. doi: 10.1080/13501763.2021.1942157
Gouvernement du Québec (2020) Rapports de sondages sur la COVID-19,
Gouvernement du Québec, https://www.quebec.ca/sante/problemes-de-sante/a-z/
coronavirus-2019/rapports-sondages-covid19.
Haas, P.M. (1992) Introduction: epistemic communities and international
policy coordination, International Organization, 46(1): 1–35. doi: 10.1017/
S0020818300001442
Hale, T., Angrist, N., Goldszmidt, R., Kira, B., Petherick, A., Phillips, T., et al (2021)
A global panel database of pandemic policies (Oxford COVID-19 Government
Response Tracker), Nature Human Behaviour, 5(4): 529–38. doi: 10.1038/
s41562-021-01079-8
Han, J.Y. and Breton, C. (2022) Obligation vaccinale: le curieux cas du
Québec, Policy Options Politiques, 15 February, https://policyoptions.irpp.
org/fr/magazines/february-2022/provincial-paths-diverge-on-vaccine-
mandates-and-passports/.
Hausfather, Z. and Peters, G.P. (2020) Emissions – the ‘business as usual’ story is
misleading, Nature, 577(7792): 618–20. doi: 10.1038/d41586-020-00177-3
INSPQ (Institut national de santé publique du Québec) (2022) Sondages sur les
attitudes et comportements des adultes québécois, INSPQ, https://www.inspq.
qc.ca/covid-19/sondages-attitudes-comportements-quebecois/17-mai-2022.
Ioannidis, J.P.A., Cripps, S. and Tanner, M.A. (2022) Forecasting for COVID-
19 has failed, International Journal of Forecasting, 38(2): 423–38. doi: 10.1016/j.
ijforecast.2020.08.004
IRPP (Institut de recherches en politiques publiques) (2021) Nouveau guide
de codage des mesures pour contrer la COVID-19, IRPP, https://centre.irpp.
org/fr/2021/07/nouveau-guide-de-codage-des-mesures-pour-contrer-la-
covid-19/.
IRPP (Institut de recherches en politiques publiques) (2022) COVID-19 provincial
policies, IRPP, https://centre.irpp.org/data/covid-19-provincial-policies/.
Unauthenticated | Downloaded 02/06/25 10:04 PM UTC

Antoine Claude Lemor et al
18
Jalloh, M.F., Zeebari, Z., Nur, S.A., Prybylski, D., Nur, A.A., Hakim, A.J., et al (2022)
Drivers of COVID-19 policy stringency in 175 countries and territories: COVID-
19 cases and deaths, gross domestic products per capita, and health expenditures,
Journal of Global Health, 12: art 05049. doi: 10.7189/jogh.12.05049
James, L.P., Salomon, J.A., Buckee, C.O. and Menzies, N.A. (2021) The use and
misuse of mathematical modeling for infectious disease policymaking: lessons
for the COVID-19 pandemic, Medical Decision Making, 41(4): 379–85. doi:
10.1177/0272989x21990391
Jasano, S. (1990) The Fifth Branch: Science Advisors as Policymakers, Cambridge, MA:
Harvard University Press.
Jones, B.D. and Baumgartner, F.R. (2005) The Politics of Attention: How Government
Prioritizes Problems, Chicago: University of Chicago Press.
Kahneman, D. (2011) Thinking, Fast and Slow, New York: Farrar, Straus and Giroux.
Kahneman, D. and Tversky, A. (1979) Prospect theory: an analysis of decision under
risk, Econometrica, 47(2): 263–92. doi: 10.2307/1914185
Koubi, V. (2019) Climate change and conict, Annual Review of Political Science, 22:
343–60. doi: 10.1146/annurev-polisci-050317-070830
Kreps, S.E. and Kriner, D.L. (2020) Model uncertainty, political contestation, and
public trust in science: evidence from the COVID-19 pandemic, Science Advances,
6(43): art eabd4563. doi: 10.1126/sciadv.abd4563
Lacoursière, A. and Chouinard, T. (2021) Situation ‘très critique’ dans les hôpitaux:
Des patients ‘risquent d’en subir les conséquences’, La Presse, 11 January, https://
www.lapresse.ca/actualites/covid-19/2021-01-11/situation-tres-critique-dans-les-
hopitaux/des-patients-risquent-d-en-subir-les-consequences.php.
Lacoursière, A. and Jean, O. (2020) Hôpitaux sur la corde raide, La Presse, 8 December,
https://www.lapresse.ca/actualites/covid-19/2020-12-08/hopitaux-sur-la-corde-
raide.php.
Leger (2022) North American tracker, formerly available at: https://legermarketing.
wpenginepowered.com/wp-content/uploads/2022/04/Legers-North-American-
Tracker-April-14th-2022.pdf.
Leger (2023) Leger’s North American tracker, Leger, 10 May, https://leger360.com/
surveys/legers-north-american-tracker/.
Lemor, A. (2024) QC.Uncertainty_COVID. Quebec COVID-19 press conferences:
uncertainty analysis, codes, and textual data repository, Zenodo, 20 January. doi:
10.5281/zenodo.10542039
Lemor, A. and Montpetit, É. (2024) Exploring the role of uncertainty, emotions,
and scientic discourse during the COVID-19 pandemic, Policy and Society, 43(2):
289–303. doi: 10.1093/polsoc/puae010
Lemor, A., Costa, M.A., Beaulieu-Guay, L.R. and Montpetit, E. (2023) QC.COVID_
Data: v1.0.0, Zenodo, 31 October. doi: 10.5281/zenodo.10059777
Lemor, A., Montpetit, É., Téhinian, S., Van Belleghem, C., Eichenberger, S., Öberg, P.,
et al (2024) Network dynamics in public health advisory systems: a comparative
analysis of scientic advice for COVID-19 in Belgium, Quebec, Sweden, and
Switzerland, Governance. doi: 10.1111/gove.12885
Lertzman, R. (2015) Environmental Melancholia: Psychoanalytic Dimensions of Engagement,
Hove: Routledge.
Löblová, O. (2018) When epistemic communities fail: exploring the mechanism of
policy inuence, Policy Studies Journal, 46(1): 160–89. doi: 10.1111/psj.12213
Unauthenticated | Downloaded 02/06/25 10:04 PM UTC

Why did the influence of experts erode during the COVID-19 pandemic?
19
Maor, M. (2020) Policy over- and under-design: an information quality perspective,
Policy Sciences, 53(3): 395–411. doi: 10.1007/s11077-020-09388-x
Maor, M. and Howlett, M. (2020) Explaining variations in state COVID-19 responses:
psychological, institutional, and strategic factors in governance and public policy-
making, Policy Design and Practice, 3(3): 228–41. doi: 10.1080/25741292.2020.1824379
Maor, M., Sulitzeanu-Kenan, R. and Chinitz, D. (2020) When COVID-19,
constitutional crisis, and political deadlock meet: the Israeli case from a
disproportionate policy perspective, Policy and Society, 39(3): 442–57. doi:
10.1080/14494035.2020.1783792
Martel-Morin, M. and Lachapelle, E. (2022) Code red for humanity or time for broad
collective action? Exploring the role of positive and negative messaging in (de)
motivating climate action, Frontiers in Communication, 7, https://www.frontiersin.
org/articles/10.3389/fcomm.2022.968335.
Mayer, A. and Smith, E.K. (2019) Unstoppable climate change? The inuence of
fatalistic beliefs about climate change on behavioural change and willingness to pay
cross-nationally, Climate Policy, 19(4): 511–23. doi: 10.1080/14693062.2018.1532872
Mildenberger, M. and Leiserowitz, A. (2017) Public opinion on climate change: is
there an economy–environment tradeo?, Environmental Politics, 26(5): 801–24. doi:
10.1080/09644016.2017.1322275
Montpetit, É. (2011) Scientic credibility, disagreement, and error costs in 17
biotechnology policy subsystems, Policy Studies Journal, 39(3): 513–33. doi:
10.1111/j.1541-0072.2011.00419.x
Norgaard, K.M. (2011) Living in Denial: Climate Change, Emotions, and Everyday Life,
Cambridge, MA: MIT Press.
Oreskes, N. and Conway, E.M. (2010) Merchants of Doubt: How a Handful of Scientists
Obscured the Truth on Issues from Tobacco Smoke to Global Warming, New York:
Bloomsbury.
Pachauri, R.K., Mayer, L. and IPCC (Intergovernmental Panel on Climate Change)
(eds) (2015) Climate Change 2014: Synthesis Report – Contribution of Working
Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel
on Climate Change, Geneva: IPCC.
Pearce, W. (2020) Trouble in the trough: how uncertainties were downplayed in the
UK’s science advice on COVID-19, Humanities and Social Sciences Communications,
7: art 122. doi: 10.1057/s41599-020-00612-w
Pueyo, T. (2020) Coronavirus: why you must act now, Medium, 10
March, updated 19 March, https://tomaspueyo.medium.com/
coronavirus-act-today-or-people-will-die-f4d3d9cd99ca.
Radio-Canada (2022) Le Québec face à une imminente ‘avalanche de nouvelles
hospitalisations’, Radio-Canada, 6 January, https://ici.radio-canada.ca/
nouvelle/1852218/hospitalisations-hausse-covid-hopitaux-delestage-quebec.
Rhodes, T. and Lancaster, K. (2020) Mathematical models as public troubles in
COVID-19 infection control: following the numbers, Health Sociology Review, 29(2):
177–94. doi: 10.1080/14461242.2020.1764376
Rochefort, D.A. and Cobb, R.W. (1993) Problem denition, agenda access, and
policy choice, Policy Studies Journal, 21(1): 56–71. doi: 10.1111/j.1541-0072.1993
.tb01453.x
Rocheleau, S. (2017) The information ow observatory [home page], Observatoire
de la circulation de l’information, https://oci-ifo.org/en/home/.
Unauthenticated | Downloaded 02/06/25 10:04 PM UTC

Antoine Claude Lemor et al
20
Salajan, A., Tsolova, S., Ciotti, M. and Suk, J.E. (2020) To what extent does evidence
support decision making during infectious disease outbreaks? A scoping literature
review, Evidence & Policy, 16(3): 453–75. doi: 10.1332/174426420x15808913064302
Sarkki, S., Niemelä, J., Tinch, R., van den Hove, S., Watt, A. and Young, J. (2014)
Balancing credibility, relevance and legitimacy: a critical assessment of trade-os
in science–policy interfaces, Science and Public Policy, 41(2): 194–206. doi: 10.1093/
scipol/sct046
Schneider, A.L. and Ingram, H.M. (1997) Policy Design for Democracy, University Press
of Kansas.
Sebhatu, A., Wennberg, K., Arora-Jonsson, S. and Lindberg, S.I. (2020) Explaining the
homogeneous diusion of COVID-19 nonpharmaceutical interventions across
heterogeneous countries, PNAS, 117(35): 21201–8. doi: 10.1073/pnas.2010625117
Sha, S. and Mallinson, D.J. (2023) Evaluating punctuated equilibrium dynamics
within a crisis context, Policy & Politics, 51(4): 647–72. doi: 10.1332/030557321x
16891538754098
Soroka, S.N. and Wlezien, C. (2009) Degrees of Democracy: Politics, Public Opinion, and
Policy, New York: Cambridge University Press.
Toshkov, D., Carroll, B. and Yesilkagit, K. (2021) Government capacity, societal trust
or party preferences: what accounts for the variety of national policy responses
to the COVID-19 pandemic in Europe?, Journal of European Public Policy, 29(7):
1009–28. doi: 10.1080/13501763.2021.1928270
van Asselt, M.B.A. and Vos, E. (2006) The precautionary principle and the uncertainty
paradox, Journal of Risk Research, 9(4): 313–36. doi: 10.1080/13669870500175063
Vetenskapsforum COVID-19 (2023) Science forum COVID-19, Vetenskapsforum
COVID 19, https://vetcov19.se/en/.
Wang, C., Kim, Y. and Mossberger, K. (2024) Governor’s political aliation and
stringent COVID-19 policy, Public Administration Review, 84(1): 40–55. doi: 10.1111/
puar.13649
Wilkinson, K., Lowe, P. and Donaldson, A. (2010) Beyond policy networks: policy
framing and the politics of expertise in the 2001 foot and mouth disease crisis, Public
Administration, 88(2): 331–45. doi: 10.1111/j.1467-9299.2010.01831.x
Wlezien, C. (2023) News and public opinion: which comes rst?, Journal of Politics,
86(1): 1–17. doi: 10.1086/726940
Zaki, B.L., Pattyn, V. and Wayenberg, E. (2023) Policy learning type shifts during
creeping crises: a storyboard of COVID-19 driven learning in Belgium, European
Policy Analysis, 9(2): 142–66. doi: 10.1002/epa2.1165
Zografakis, N., Sifaki, E., Pagalou, M., Nikitaki, G., Psarakis, V. and Tsagarakis, K.P.
(2010) Assessment of public acceptance and willingness to pay for renewable energy
sources in Crete, Renewable and Sustainable Energy Reviews, 14(3): 1088–95. doi:
10.1016/j.rser.2009.11.009
Unauthenticated | Downloaded 02/06/25 10:04 PM UTC