Content uploaded by Cary Wu
Author content
All content in this area was uploaded by Cary Wu on Jul 17, 2024
Content may be subject to copyright.
The Trump Effect? Right-Wing Populism and Distrust
in Voting by Mail in Canada
Cary Wu
1,
�, Andrew Dawson
2
1
York Research Chair and Associate Professor, Department of Sociology, York University,
Toronto, ON, Canada
2
Associate Professor, Department of Sociology, York University, Glendon Campus, Toronto,
ON, Canada
Abstract Do Donald Trump’s attacks on voting by mail influence
how some Canadians view mail-in ballots? The Trump effect on views
and behaviors surrounding voting by mail has been well documented in
the United States. North of the border, more Canadians than ever voted
by mail in the last general election. In this study, we consider how
right-wing populism is associated with trust in voting by mail among
Canadians. Specifically, we seek to test two main hypotheses. First, we
consider whether Canadians holding populist views—and, in particular,
those holding right-wing populist views (would-be Trump supporters)—
are less trusting of voting by mail. Second, we consider whether politi-
cal media exposure amplifies this association. We analyze data from
both the 2021 Canadian Election Study and Democracy Checkup
Survey. We find that those who hold populist views clearly have less
trust in voting by mail. This is especially true among right-leaning indi-
viduals. Furthermore, as in the United States, this effect is moderated by
one’s level of political media exposure, with higher levels of political
media exposure amplifying the effect of populist views on trust in vot-
ing by mail. Our findings, therefore, suggest that the politicization of
mail-in voting by President Trump has important implications for the le-
gitimacy of the electoral system not only in the United States, but also
in Canada and potentially in other parts of the world.
Introduction
Voting by mail has long been a vital component of the democratic process
in Canada. In 1993, the option of submitting a ballot by mail was extended
�Corresponding author: Cary Wu, Department of Sociology, York University, 2060 Vari Hall,
4700 Keele St, Toronto, ON M3J 1P3, Canada; email: carywu@yorku.ca.
# The Author(s) 2024. Published by Oxford University Press on behalf of American Association for Public Opinion Research.
This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs licence
(https://creativecommons.org/licenses/by-nc-nd/4.0/), which permits noncommercial reproduction and distribution of the work, in
any medium, provided the original work is not altered or transformed in any way, and that the work is properly cited. For commer-
cial reuse, please contact journals.permissions@oup.com
https://doi.org/10.1093/poq/nfae020
Public Opinion Quarterly (2024) Vol 00 No 0, 1–33
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
to all Canadian voters (Maas 2015). Despite a long history of being a secure
and reliable method of voting in Canada, mail-in voting was not commonly
used in general elections prior to the COVID-19 pandemic (Harell and
Stephenson 2022). During the 2019 federal election, only about 55,000
Canadians used the mail-in voting option to cast their ballot, representing
only about 0.3 percent of the total ballots cast. This changed during the
COVID-19 pandemic due to safety concerns about voting in person. In the
latest federal election (2021), Elections Canada received more than 700,000
special ballots by mail, a record number representing an increase of more
than tenfold from 2019 (Elections Canada 2022; Harell and
Stephenson 2022).
South of the border, nearly half of US voters cast ballots by mail during
the 2020 presidential election (United States Census Bureau 2021).
However, in the lead-up to the election, voting by mail became highly politi-
cized due to the unrelenting attacks on mail-in voting from some
Republicans and President Trump in particular. Trump began to cast doubt
on the legitimacy of mail-in voting as early as April 2020, when he tweeted
that mail-in voting was “substantially fraudulent” and that it would lead to a
“rigged election.” While evidence suggests that there is almost no meaning-
ful fraud associated with mail ballots (Thompson et al. 2020; West 2020),
claims do not have to be true to generate effects. Exposure to false claims of
election fraud can lower people’s faith in elections and beliefs in democratic
government (Justwan and Williamson 2022; Berlinski et al. 2023).
A sitting US president’s public pronouncements regarding the illegitimacy
of mail-in voting predictably received a great deal of international attention
and news coverage, including in Canada. Given the shared border and close
trade relationship between the two countries, it is unsurprising that a Pew
Research Center (2018) study that surveyed 37 countries found that interest
in news about the United States is highest in Canada, where 78 percent of
Canadians indicate that they track US news closely. As such, it is possible
that Trump’s attempts to delegitimize voting by mail had some influence on
public opinion in Canada.
On one hand, Canada has a long history of conducting fair and transparent
elections. Notably, the Canadian electoral system is not as fraught with parti-
sanship as its US counterpart, primarily because in Canada federal elections
are run by a single, nonpartisan electoral management body (Elections
Canada). In addition, compared to many jurisdictions in the United States, it
is for the most part comparatively easier to vote in person in Canada, as
measured by wait times or the absence of legal restrictions surrounding
queuing—“no one will take away your water bottle” (BBC News 2021).
Furthermore, unlike in the United States, the process of voting by mail is
standardized in Canada—all Canadians are eligible to vote by mail and must
apply for a special ballot from Elections Canada in advance and provide
2 C. Wu and A. Dawson
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
proof of eligibility to vote (i.e., proof of identity and other personal informa-
tion). This track record of successful elections and strong democratic institu-
tions is likely to inspire confidence among Canadian citizens in the integrity
of their voting system, regardless of any political rhetoric or attacks on vot-
ing methods in other countries.
On the other hand, Canadian voters may have been influenced by news
coverage of Trump’s claims of widespread voter fraud during the 2020 US
election. Although these claims refer to the US electoral system, they may
nonetheless create an atmosphere of suspicion and doubt around mail-in vot-
ing in Canadian elections. Since his election in 2016, some Canadians have
expressed their desire to support a far-right populist leader like Donald
Trump (Perry and Scrivens 2019), indicating a receptiveness to his ideas and
appeals. Indeed, recent research has documented manifestations of a “Trump
effect” in Canada. For example, Perry, Mirrlees, and Scrivens (2017) have
explored how the US politics of hate unleashed by Trump’s right-wing pop-
ulist posturing galvanized Canadian white supremacist ideologies, identities,
movements, and practices. They express concern about the rise of a vocal
far-right populist movement and that right-wing extremism is becoming in-
creasingly normalized (i.e., mainstream) in Canada—developments aided by
Canada’s far-right media (Elmer and Burton 2022).
Others remain confident that these various far-right movements have little
impact on mainstream Canadian political culture (Gillies 2023). The
People’s Party of Canada, a “populist radical right” party considered to have
imported Trumpism and other far-right ideologies, attracted less than 5 per-
cent of the vote in the 2021 federal election (Budd 2021). Furthermore,
Budd (2021) argues that Canada’s political opportunity structures and its
single-member-plurality electoral system lead to moderation in right-wing
populist ideology. Consequently, some suggest that “Canada remains some-
what of a bulwark against populism” (Gillies 2023, p. 12).
In this study, we consider how right-wing populism is associated with
trust in voting by mail among Canadians. More specifically, despite having
a less partisan electoral system, we examine whether Trump’s attacks on
voting by mail have a diffusion effect across the border and divide how
Canadians view mail-in ballots. To do so, we investigate whether Canadians
holding right-wing populist views are less trusting of voting by mail.
Second, following the literature, we consider whether news consumption
intensifies this association. Analyzing data from the 2021 Canadian Election
Study, we find strong evidence of a “Trump effect.” That is, we find a clear
negative association among Canadians between holding right-wing populist
views and trust in voting by mail. Furthermore, as in the United States, this
effect is moderated by one’s level of news consumption, with higher levels
of news consumption exacerbating the effect.
The Trump Effect? 3
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
The “Trump Effect” on Voting by Mail
In the United States, postal voting has long been a widely accepted method
of casting a ballot. The practice dates back to the Civil War, when soldiers
were able to send their votes from the battlefield to make their voices heard
(Fortier and Ornstein 2002). The use of mail-in ballots has increased in re-
cent years, particularly during the COVID-19 pandemic as a means for vot-
ers to reduce their risk of exposure to the virus, with a large share of the US
electorate (43 percent) casting ballots by mail in 2020 (United States Census
Bureau 2021). While the mail-in ballot is meant to increase the convenience
and accessibility of voting, evidence suggests that voting by mail has only a
modest effect, if any, on increasing electoral participation in recent elections
(Berinsky, Burns, and Traugott 2001; Southwell 2009; Monroe and
Sylvester 2011; Gronke and Miller 2012; Barber and Holbein 2020; Amlani
and Collitt 2022). Several studies suggest that voting by mail mainly
changes “how people vote” (or voter retention) rather than “who votes” (or
voter recruitment), including during the 2020 election amid the COVID-19
pandemic (Monroe and Sylvester 2011; Thompson et al. 2020; Yoder et al.
2021). Notably, numerous studies have failed to find evidence that voting by
mail confers any specific partisan advantage (Barber and Holbein 2020;
Amlani and Collitt 2022).
Before the Trump era, studies show that voters were not polarized in
terms of how they cast their ballots, with Republicans and Democrats shar-
ing an equal likelihood of voting by mail (Berinsky, Burns, and Traugott
2001; Thompson et al. 2020; see also Lockhart et al. 2020). However, this
changed in the 2020 election (McDonald 2022). Trump repeatedly claimed
that mail-in voting is prone to fraud and abuse, and that it would result in a
“rigged” election. Consequently, how US citizens vote—specifically, voting
by mail—became highly politicized (Lockhart et al. 2020).
The Trump effect on views and behaviors surrounding vote by mail has
been well documented in the United States (Justwan and Williamson 2022;
Berlinski et al. 2023). There are two major effects. First, Trump actively dis-
couraged his supporters from voting by mail, urging them to vote in person
instead. Consequently, many of his supporters voted in person on election
day, while many Democrats voted by mail. In 2020, roughly 60 percent of
Democrats nationally voted by mail, compared with 30 percent of
Republicans (Stewart III 2020). Moreover, using large online surveys before
the 2020 November election, Lockhart et al. (2020) find a variation of
roughly 10 percent between the proportion of Democrats and Republicans
who preferred voting by mail. Shino et al. (2022) also show that cue taking
from Trump helps explain the willingness of politically aware Republicans
who voted by mail during the 2020 election to disavow their own bal-
lot choice.
4 C. Wu and A. Dawson
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
Second, Trump’s wrongful claims had a significant impact on public per-
ception, leading to increased skepticism and mistrust of voting by mail and
the electoral process. Using a nationwide survey experiment conducted after
the 2018 midterm elections—a time when many prominent Republicans
made unsubstantiated fraud claims—Berlinski et al. (2023) show that expo-
sure to claims of voter fraud reduces confidence in electoral integrity, espe-
cially among Republicans and Trump approvers. This finding was
reproduced by Justwan and Williamson (2022) in their survey experiment of
the effect of Trump’s false claims of electoral fraud, particularly surrounding
voting by mail, following the 2020 election. Consequently, Republicans (43
percent) are much more likely than Democrats (11 percent) to identify fraud
as a major problem with voting by mail (Mitchell et al. 2020).
Perceptions of electoral integrity matter—sowing the seeds of electoral
mistrust via false claims has real effects. Pennycook and Rand (2021) show
that a majority of Trump voters—particularly those with more political
knowledge and who more closely follow election news—falsely believed
that election fraud was widespread, and that Trump had won the 2020 elec-
tion. Trump’s continued efforts to undermine confidence in the election ulti-
mately culminated in the January 6 insurrection at the Capitol, which aimed
to overturn the results of the election.
Trust in Voting by Mail: Elite Cues and Media
Amplification
There is a growing body of literature on factors that affect trust in
elections. This literature can be divided into five major camps, each with a dif-
ferent focus: (1) electoral management bodies/election administration; (2)
individual-level factors; (3) the winner effect; (4) elite cues; and (5) media ex-
posure. These five camps are, of course, not necessarily mutually exclusive.
First, the electoral management body (EMB) and its performance in ad-
ministrating elections are considered crucial determinants of trust in elec-
tions. Various facets of EMBs are considered consequential in influencing
trust, including their design (Otaola 2018), their degree of independence
from the government (Birch 2008; Kerr and L€
uhrmann 2017), their overall
administrative performance during elections (Kerr 2013; Bowler et al.
2015), and their localized performance as measured by the level of profes-
sionalism of poll workers (Hall et al. 2009) and individual experiences
(whether negative or positive) at the polling station (King 2017, 2020; King
and Barnes 2019).
Second, it is well established in the literature that individual-level traits in-
fluence one’s level of regime support more generally and trust in elections
more specifically (Birch 2008; Booth and Richard 2014). These traits are
The Trump Effect? 5
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
primarily demographic, including age, gender, education, socioeconomic sta-
tus, and race/ethnicity.
Third, an additional individual-level factor—voting for the winning candi-
date—is a strong predictor of trust in elections and is commonly known as
the “winner effect” (Anderson and LoTempio 2002; Maldonado and
Seligson 2013; Sances and Stewart 2015; Reller, Anderson, and Kousser
2022). Notably, evidence suggests that the “winner effect” is independent
from the effect of elite claims of voter fraud (Reller, Anderson, and Kousser
2022). In analyzing Trump’s victory in the 2016 election, Sinclair, Smith,
and Tucker (2018) find that the “winner effect” mitigates the effects from
strong pre-election cues of a “rigged system” from Trump, although Levy
(2021) finds this “winner effect” to be larger for voter confidence than for
beliefs about illicit voting.
This leads us to the fourth camp—elite cues. The claims of elites and,
more specifically, election candidates surrounding the legitimacy of elec-
tions have a significant effect on voter confidence (Vonnahme and Miller
2013). As mentioned, these claims need not be factual to have an impact,
such as the effects of Trump’s unsubstantiated claims of voter fraud on con-
fidence in elections (Justwan and Williamson 2022; Berlinski et al. 2023;
Shino et al. 2022). While the “winner effect” mitigated elite cues from
Trump about a rigged system during the 2016 election in the United States
(see above), in Mexico, Monsiv
ais-Carrillo (2023) finds that the effect of
elite claims eclipses the “winner effect,” paradoxically leading to a situation
where supporters of a victorious populist leader who undermined the legiti-
macy of electoral institutions are less likely to trust the integrity of elections
than the partisan losers.
Fifth, the evidence suggests that media exposure is an important predictor
of trust in elections. Referred to as the “Achilles heel of democracy,” coun-
tries with independent media tend to have less confidence in elections and
other political institutions, as the media have the freedom to criticize these
institutions (Dawson 2021; Kerr and L€
uhrmann 2017). Therefore, within de-
mocracies, media consumption, particularly exposure to negative news (i.e.,
stories of electoral maladministration), is negatively associated with trust in
elections (Karp, Nai, and Norris 2018). As the media amplifies elite cues, in-
cluding false claims regarding the legitimacy of elections, media consump-
tion also increases exposure to such claims. While claims of electoral
malpractices are often exaggerated by opposition media linked with the
defeated parties (Carreras and _
Irepo
glu 2013), it is important to note that the
negative effects of false claims are not limited to news consumption from
these media outlets. Rather, exposure to false claims from other (main-
stream) news sources, even when a corrective is provided, can nonetheless
result in an erosion of trust (Justwan and Williamson 2022; Berlinski
et al. 2023).
6 C. Wu and A. Dawson
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
As detailed above, elite cues (i.e., Trump’s unsubstantiated claims of elec-
toral fraud) and media exposure were paramount in creating a divide (be-
tween staunch Trump/Republican supporters and others) in how US citizens
view the legitimacy of voting by mail. We therefore focus on variables re-
lated to these factors in exploring the potential diffusion of the “Trump
effect” in Canada. In addition, we also assess whether the impact of these
variables holds when controlling for the other determinants of trust in elec-
tions outlined above, namely various individual-level demographic factors
and trust in Canada’s EMB.
1
Notably, we are unable to assess (or control
for) a “winner effect” given that the data are limited to a campaign period,
rather than a postelection, survey (see below).
Hypotheses
Research suggests that Trump supporters in the United States are more likely
to express lower levels of trust in postal voting as compared to other non-
Trump supporters. We test whether this “Trump effect” is also present in
Canada. That is, whether there is a trust divide in voting by mail in Canada
between would-be Trump supporters—that is, Canadians holding right-wing
populist views—and other Canadians. Consequently, we hypothesize that:
H1: Canadians with populist orientations are less trusting of voting by mail.
H1a: The negative association between populist orientations and trust in voting by
mail is stronger among Canadians identifying as right leaning.
None of the major political parties in Canada, including the populist radi-
cal right-wing People’s Party of Canada, have mounted a public campaign
attempting to delegitimize voting by mail. As such, any findings indicating
that Canadians holding right-wing populist views have less trust in the mail-
in ballot can arguably be attributable to cues from Trump, particularly given
the closeness with which Canadians follow US news.
As the “Trump effect” rests on exposure to unsubstantiated claims of elec-
toral fraud, our second hypothesis pertains to media exposure. That is, in the
case of Trump’s false claims of fraud surrounding voting by mail, we posit
that the media—all media—serve to amplify (i.e., propagate) these claims.
This follows recent research suggesting that exposure to unfounded fraud
claims—even when a corrective is provided by the news source—can
1. Given that trust in Elections Canada can be considered posttreatment of populist orientations
and a proxy for overall trust in voting, the variable is excluded from the models presented below
to avoid potential posttreatment bias of the effect of our populist orientation measures. However,
as a robustness check, we did run additional analyses (see Additional Analysis in Appendix) con-
trolling for trust in Elections Canada. This allowed us to independently statistically assess
respondents’ level of trust in voting by mail, as trust in voting by mail is worded in a relative
sense to voting in person. The results are substantively similar to those reported below.
The Trump Effect? 7
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
nonetheless lower trust in elections, particularly among those who share an
ideological orientation with the person making the claim (Justwan and
Williamson 2022; Berlinski et al. 2023). Consequently, we hypothesize that
overall political media exposure—not only consumption of partisan or
“ideological” news—will have a negative effect on trust in voting by mail
for Canadians who hold populist attitudes.
H2: Political media exposure amplifies the effect of populist views on trust in
voting by mail.
Comparing the province of Quebec to the rest of Canada provides a natu-
ral experiment to verify the mediating effect of media exposure on trust in
voting by mail. Quebec is the only province in Canada where English is a
minority language, as French is the first official language of the vast major-
ity (82 percent) of Quebecers (Statistics Canada 2022). According to the
2021 census, only about half the population (51.7 percent) has a working
knowledge of English, as compared to over 90 percent of the population in
all other provinces and territories (Statistics Canada 2023). While Trump’s
claims were translated by French-language media, recent research suggests
that direct exposure to US-based (i.e., English-language) media outlets, par-
ticularly content from US-based accounts on social media, played an impor-
tant role in the propagation of COVID-19 misperceptions among Canadians
(Bridgman et al. 2021). We test whether this mechanism is also applicable
to trust in voting by mail by comparing Quebec to the rest of Canada to as-
sess the influence of US-based media exposure. Consequently, we expect
the amplification effect of media exposure to be significantly lower in
Quebec as compared to the rest of Canada, where English is the major-
ity language.
H3: In predominantly French-speaking Quebec, the media exposure amplification
effect will be significantly weaker than in the rest of Canada.
Data and Methods
Data
We analyze two datasets: the 2021 Democracy Checkup Survey (DCS) and
the 2021 Canadian Election Study (CES). These data provide unique and un-
paralleled insight into Canadians’ social and political lives (Stephenson et al.
2022; Harell et al. 2022). Specifically, the CES is a large-scale survey of
citizens that has been conducted each election year since 1965. Continuing
this long tradition, the 2021 CES was conducted to document the attitudes of
Canadians during and after the federal election held on September 20, 2021.
The CES consists of two rounds, a campaign period survey and a
postelection survey, both of which include questions on trust in elections,
8 C. Wu and A. Dawson
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
issue positions, vote intentions, and many other political attitudes and behav-
iors. In this study, we use data from the campaign period survey conducted
from August 17 to September 19, 2021. Data from the postelection survey
are excluded from our analysis because trust in voting by mail—the focal
outcome variable of this study—was not asked.
We also analyze data from the 2021 Democracy Checkup Survey con-
ducted annually by the Consortium on Electoral Democracy (C-Dem) during
the inter-election period. The Democracy Checkup Surveys often include
core CES questions as well as a variety of timely measures of political and
social attitudes. The 2021 DCS was conducted from May 20 to June 7,
2021. Analyzing the DCS data allows us to examine how elite cues and me-
dia amplification work to shape Canadians’ trust in voting by mail during
normal times—that is, outside of an election campaign. During the campaign
period, Canadians’ political interests and/or partisan attitudes are more acti-
vated, which could influence their responses to questions pertaining to politi-
cal attitudes and opinions.
Both the CES and DCS were conducted online due to the COVID-19 pan-
demic. The Qualtrics online platform was used to administer the surveys.
Sample respondents were selected through the Leger Opinion panel (https://
leger360.com), which is the largest proprietary panel in Canada with over
400,000 double opted-in members across Canada and the United States. The
majority of the Leger Opinion panelists (over 70 percent) are recruited ran-
domly through call center outreach, while the remaining are recruited
through targeted methods such as invitations, affiliate programs, social me-
dia engagement, partner campaigns, word of mouth, and offline recruitment.
This hybrid recruitment approach is designed to enhance the quality of
the panel.
To create accurate samples reflecting the general population, the Leger
company uses a proprietary software in accordance with Canadian census
data. Furthermore, a response rate that considers age, gender, and region is
incorporated to maximize optimal representativeness. This quota sampling
methodology was used to construct both the CES sample and the DCS sam-
ple. For example, the CES respondents were selected with targets stratified
by region and balanced on gender and age based on the 2016 Canadian
Census. Respondents are Canadian citizens or permanent residents aged
18 years or older. The first wave was launched on August 17 and ran until
September 8, producing a sample of 5,820. The second wave ran from
September 9 to September 19, resulting in a sample of 13,120.
Simultaneously, an “oversample” wave, without sampling quotas, ran from
August 18 to September 19, yielding a sample of 2,028. The “oversample”
wave collected overflow responses from full quotas in the first two waves. A
daily sample of 300 respondents was set for the campaign period, but in the
last 10days of the campaign, the target was set to 1,500 respondents per
The Trump Effect? 9
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
day. These targets were achieved on most days during the total 18 days of
the collection period, resulting in a total of 20,968 valid respondents
(Stephenson et al. 2022). The DCS sample of 7,949 participants aged
18 years or older was pulled from the Leger Opinion panel following a simi-
lar process (Harell et al. 2022). More details about these surveys, including
the quota proportions, can be found on the Canadian Election Study website
(http://www.ces-eec.ca/), where data files are also publicly available. All
analyses are conducted using the weighted data according to the 2016
Canadian census. Listwise deletion was used when analyzing both datasets.
Measures
Dependent variable: we measure our key outcome variable, trust in voting
by mail, using the following question in the CES: “Voting by mail is equally
as trustworthy as voting in person.” Response categories include 1 ¼
“strongly disagree,” 2 ¼“somewhat disagree,” 3 ¼“neither agree nor dis-
agree,” 4 ¼“somewhat agree,” and 5 ¼“strongly agree.” Hence, higher
scores represent higher levels of trust in voting by mail. We use the same
question in the DCS. The only difference is that responses to this question in
the DCS include only four categories: 1 ¼“strongly disagree,” 2 ¼
“somewhat disagree,” 4 ¼“somewhat agree,” and 5 ¼“strongly agree.” We
also create a binary measure by grouping the last two response categories
(“somewhat agree” and “strongly agree”) as “trust” (1) and other categories
as “distrust” (0). This provides a means to compare the trust levels across
the two different surveys. In particular, it allows us to track the change in
this binary measure over time based on the recorded date of each response
across both surveys from May to September 2021 (for more details, please
see figure A1).
Populism: our key predictor is holding populist attitudes. Populism is gen-
erally understood as a “thin-centered” political ideology that claims to repre-
sent the interests and the general will of the ordinary or “pure people” and
denounces the privileged, out-of-touch, and “corrupt elite” (Mudde and
Kaltwasser 2017). Following the literature, we measure populist orientations
as a latent variable using multiple items from each survey (Roccato et al.
2019; Wettstein et al. 2020). The CES includes the following five items that
help capture respondents’ alignment with populist views:
1. What people call compromise in politics is really just selling out on
one’s principles.
2. Most politicians do not care about the people.
3. Most politicians are untrustworthy.
4. The people, and not politicians, should make our most important pol-
icy decisions.
10 C. Wu and A. Dawson
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
5. Most politicians care only about the interests of the rich and powerful.
Responses were coded on a 1–5 Likert scale from strongly disagree (1) to
strongly agree (5).
2
We use principal components analysis (PCA) to combine
these five measures. The PCA is a data reduction technique that combines
different indicators based on the common variance among the measures (see
also Wu et al. 2022). By using the PCA, we can generate an index that
reflects the extent of populist beliefs held by each respondent. This index is
derived from a single component that explained over half (55 percent) of the
variance, with factor loadings exceeding 0.64 for all items, indicating a reli-
able measure of populism. We calculated the predicted component score for
each respondent and used it to assess their level of alignment with popu-
list views.
The DCS includes two questions that capture two essential components of
populism: anti-elite ideology and the view that policies should follow
the general will of the people (Mudde and Kaltwasser 2017; Schulz et al.
2018 ). The first component, anti-elite ideology, is measured using the ques-
tion of whether respondents agree with the statement that “I’d rather put my
trust in the wisdom of ordinary people than the opinions of experts and
intellectuals.” The second component, similar to that of the CES, captures
the view that policies should follow the general will of the people through
the question of whether respondents would agree with the statement that
“Ordinary people, not career politicians, should make our most important
policy decisions.” Responses to both questions were coded in four catego-
ries: 1 ¼“strongly disagree,” 2 ¼“somewhat disagree,” 3 ¼“somewhat
agree,” and 4 ¼“strongly agree.” Higher scores, therefore, indicate stronger
populist orientations. As with the CES data, we applied PCA to merge these
two items and streamline the analysis. The outcome of the PCA confirmed a
reduction into a single overall measure, with one component explaining
more than half (73 percent) of the variance, and both items displaying factor
loadings exceeding 0.85, indicating high consistency. We also calculate the
predicted component score for each respondent and use it to assess their
level of alignment with populist views. Further analysis separately assessing
the individual items reveals the same pattern (see table A3).
Right-wing political ideology: to determine where respondents fall along
the political spectrum, we use the question in the CES that asks, “In politics,
people sometimes talk of left and right. Where would you place yourself on
a scale from 0 to 10 where 0 means the left and 10 means the right?”
2. We note that the use of agree-disagree questions in these populism measures may inadver-
tently capture not only genuine populist sentiments but also a tendency to agree in general.
Readers should be aware of this potential acquiescence bias when interpreting the results of
such measures.
The Trump Effect? 11
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
In DCS, this question was not asked. We use partisanship to indicate
Canadians’ political ideology. The item asks, “In federal politics, do you
usually think of yourself as a: Liberal, Conservative, NDP, Bloc Qu
eb
ecois,
Green, Another party (please specify), or No party.” We recategorized
the response into four groups in descriptive analysis: 1 ¼Liberal,
2¼Conservative, 3 ¼NDP and other, 4¼Independent. In our regression
analyses, we create a binary variable separating between those who identify
as on the political right/Conservative (1) and others (0).
Political media exposure: as political media exposure is difficult to validly
measure via self-reports (Dilliplane, Goldman, and Mutz 2013; Guess et al.
2019; Goldman and Warren 2020), we use several indicators that examine
its different aspects—news consumption, social media use, and political
news engagement. The CES includes the following items:
1. Whether respondents circulated, (re)posted, or commented on political
information, news, or content online or offline over the past 12 months
(1 ¼never, 2 ¼just once, 3 ¼a few times, and 4 ¼more than
five times)
2. Used social media to discuss politics or political issues over the past
12 months (1 ¼never, 2 ¼just once, 3 ¼a few times, and 4 ¼more
than five times)
3. How closely respondents follow politics on TV, radio, newspapers, or
the internet (1 ¼not at all, 2 ¼not very closely, 3 ¼fairly closely,
and 4 ¼very closely)
4. On average, how much time do you usually spend watching, reading,
and listening to news each day? (1 ¼none, 2 ¼1–10 minutes, 3 ¼
11–30 minutes, 4 ¼31–60 minutes, 5 ¼between 1 and 2 hours, and 6
¼more than two hours)
The DCS does not include all four items, but it does include the second
item that asks about whether respondents used social media to discuss poli-
tics or political issues, and the last item that asks about respondents’ time
spent on news in general. Our main analysis is based on the first and the last
item; however, additional analyses using the other measures yield similar
patterns (see figures A2 and A3).
Political knowledge: in the CES, we use the following item to measure
Canadians’ political knowledge/sophistication: “What is the name of the
Governor-General of Canada?” We record those who answered correctly as
1¼politically knowledgeable, others as 0. In the DCS, political knowledge
is measured using the item “Which level of government is primarily respon-
sible for the following policy areas?” Policy areas include Employment
Insurance (Federal); Healthcare (Provincial); Primary and Secondary
12 C. Wu and A. Dawson
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
Education (Provincial); Defense Policy (Federal); Public Transit
(Municipal); and Sewage and Water (Municipal). The resulting political
knowledge score ranges from 0 (all incorrect answers) to 6 (all cor-
rect answers).
Demographic controls included in the analysis are gender, age, education,
household income, marital status, and visible minority status. We also add a
binary control for generalized social trust as a rough proxy of certain person-
ality traits (e.g., trust in experts, conspiratorial ideation, and science literacy)
that may be correlated to both holding populist attitudes and trust in postal
voting (Berning and Ziller 2017). In addition, we create two binary variables
of Alberta and Quebec based on the respondents’ province of residence
given the unique political culture and dynamics in these two provinces.
Table 1 provides the coding scheme and descriptive statistics for all varia-
bles in the analysis.
Plan of Analysis
Our analysis takes three general steps. First, we explore the descriptive as-
sociation between populism and trust in voting by mail and how this asso-
ciation differs across political ideologies, political contexts, as well as
levels of political media exposure. Second, since the dependent variable is
an ordinal response variable in both CES and DCS, we estimate odds ra-
tios (ORs) of trust in voting by mail using ordinal logistic regressions
with controls, with additional analyses using simple OLS models yielding
consistent results. Specifically, we analyze three general models. In
Model (1), the base model, we include populism (PCA index), political
ideology, and political media exposure together as individual predictors.
In Model (2), we add the interaction term between populism and right
ideology to test whether the relationship between populism and trust in
mail-in voting is conditional on political ideology. In Model (3), we add
the interaction term between populism and media exposure to assess
whether the latter amplifies the effect of populism on trust in voting by
mail. In our final step, we replicate Model (3) separating Quebec and the
rest of Canada to examine whether the media exposure amplification ef-
fect is stronger outside Quebec. These three analytical steps are conducted
separately using CES and DCS data. More recent research has suggested
that ordinal variables can be treated as continuous variables (Robitzsch
2020), and OLS can be as effective as, if not more than, other models for
categorical outcomes (Biswas, Das, and Das 2019; Gomila 2021). In
Supplementary Material section A, we have provided additional analysis
using OLS models. The results are largely consistent.
The Trump Effect? 13
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
Table 1. Coding scheme and descriptive statistics.
CES DCS
Trust in voting by mail Trust in voting by mail
Trust in voting by mail (1 ¼strongly disagree,
5¼strongly agree) 3.43
Trust in voting by mail (1 ¼strongly disagree,
4¼strongly agree) 2.91
Right-wing populism (1 ¼strongly disagree, 5 ¼strongly agree) Right-wing populism
What people call compromise in politics is really just selling out
on one’s principles 3.24
Anti-elite ideology (1 ¼strongly disagree,
4¼strongly agree) 2.19
Most politicians do not care about the people 3.39 Speaking for ordinary people (1 ¼strongly
disagree, 4 ¼strongly agree) 2.61Most politicians are untrustworthy 3.40
The people, and not politicians, should make our most important
policy decisions 3.52
Most politicians care only about the interests of the rich
and powerful 3.62
Media exposure Media exposure
News consumption (1 ¼none, 6 ¼more than two hours) 3.62 News consumption (1 ¼none, 6¼more than
two hours) 3.58Circulated, (re)posted, or commented on political
information, news 1.81
Used social media to discuss politics or political issues 1.67
Respondents closely follow politics on TV, radio, newspapers, or
the internet 2.74
Political knowledge (knowing the name of the Governor-General
of Canada) 56.04%
Political knowledge (scale, 0-6) 4.65
(continued)
14 C. Wu and A. Dawson
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
Table 1. Continued.
CES DCS
Social trust: can most people be trusted? 38.65%Social trust: can most people be trusted? 44.02%
Controls Controls
Condence in Elections Canada (1 ¼none at all,
4¼a great deal) 2.81
Condence in Elections Canada (1 ¼none at
all, 4 ¼a great deal) 3.06
Female (0 ¼Male, 1¼Female) 51% Female (0 ¼male, 1¼female) 51%
Age in years 53.07 Age in years 48.06
Education (1 ¼no schooling, 10¼professional degree
or doctorate) 6.82
Education (1 ¼no schooling,
10 ¼professional degree or doctorate) 7.63
Household income before taxes (in dollars) 86,161.00 Household income before taxes (1 ¼no
income, 8¼$200,000þ) 4.17
White (0 ¼Non-white, 1¼White) 76.8% White (0 ¼Non-white, 1¼White) 80.7%
Married/living together 62.2% Married/living together 61.8%
Single, never married 23.5% Single, never married 25.1%
Divorced 7.8% Divorced 7.2%
Separated 2.8% Separated 2.8%
Widowed 3.7% Widowed 3.1%
Alberta 6.99% Alberta 11.21%
Quebec 16.22% Quebec 23.47%
The Trump Effect? 15
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
Results
To begin, we highlight three descriptive patterns. First, Figure 1 shows that
populist orientations are associated with lower trust in voting by mail.
Moreover, the negative association is particularly strong among those on the
right (CES) or Conservative supporters (DCS). Second, political context also
matters. Alberta has a long history of successful populist political move-
ments and is widely considered Canada’s most conservative province (Budd
2021; Laycock 1990). Figure 2 shows that the negative association between
populism and trust in voting by mail is more apparent in Alberta than in the
Figure 1. Association between populism and trust voting by mail by politi-
cal ideology.
Figure 2. Association between populism and trust voting by mail in Alberta
and the rest of the country (ROC).
16 C. Wu and A. Dawson
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
rest of Canada. Third, media consumption seems to amplify the negative ef-
fect of populism on trust in voting by mail. Figure 3 suggests that higher lev-
els of political news engagement (CES) or news consumption (DCS)
intensify the negative association between populism and trust in voting
by mail.
Table 2 reports results from three ordinal logistic regression models cap-
turing the “Trump effect” on trust in voting by mail using data from the
CES. Model (1), the base model, includes the populism (PCA) index, politi-
cal ideology, and political news engagement together. It shows that populist
beliefs are significantly associated with lower odds of believing that voting
by mail is equally as trustworthy as voting in person. Specifically, a one-
standard-deviation increase in the populism index score is associated with
nearly 30 percent lower odds of trusting in voting by mail (p<0.001).
Identifying as ideologically right is associated with lower trust in voting by
mail. Moving one point on the 0–10 scale from left (0) to right (10) on the
political ideology spectrum is associated with a nearly 20 percent decrease
in the odds of trusting in voting by mail (p<0.001). Further, political news
engagement as measured by whether respondents circulated, (re)posted, or
commented on political information, news, or content online or offline is
also associated with lower levels of trust in voting by mail (p<0.001).
Model (2) adds the interaction term between populism and political ideol-
ogy to the base model. The significant interaction effect (p<0.005) suggests
that the relationship between populism and trust in mail-in voting is condi-
tional on ideology—strongest among those on the right as expressed on the
0–10 self-placement scale. Model (3) adds the interaction term between pop-
ulism and political news engagement to the base model. The significant in-
teraction effect (p<0.001) suggests that political news engagement
Figure 3. Association between populism and trust voting by mail by levels of
political news engagement (left) and news consumption (right).
The Trump Effect? 17
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
Table 2. Ordinal logistic regressions capturing the Trump effect on trust in
voting by mail, CES.
Model (1) Model (2) Model (3)
The Trump effect
Populism index 0.697 0.848 0.820
(0.000) (0.010) (0.000)
Right ideology 0.804 0.809 0.806
(0.000) (0.000) (0.000)
Political news engagement 0.893 0.896 0.903
(0.000) (0.000) (0.000)
Right-wing populism (Populism#right ideology) 0.965
(0.002)
Media amplication (Populism#news engagement) 0.919
(0.000)
Covariates
Political knowledge 1.128 1.137 1.128
(0.029) (0.020) (0.029)
Social trust 1.352 1.358 1.356
(0.000) (0.000) (0.000)
Female 0.945 0.946 0.950
(0.247) (0.254) (0.292)
Age 0.990 0.990 0.991
(0.000) (0.000) (0.000)
Education 1.099 1.101 1.099
(0.000) (0.000) (0.000)
Household income 1.000 1.000 1.000
(0.019) (0.020) (0.020)
White 0.950 0.958 0.954
(0.373) (0.458) (0.416)
MS_Single (ref. Married/living together) 1.049 1.044 1.045
(0.424) (0.471) (0.458)
MS_Divorced 0.928 0.926 0.933
(0.367) (0.353) (0.404)
MS_Separated 0.776 0.773 0.770
(0.069) (0.063) (0.059)
MS_Widowed 1.071 1.075 1.068
(0.598) (0.576) (0.619)
Alberta 0.813 0.816 0.806
(0.025) (0.028) (0.020)
Quebec 0.674 0.673 0.678
(0.000) (0.000) (0.000)
Estimated cut points
cut1 0.0368 0.0386 0.0381
(0.000) (0.000) (0.000)
(continued)
18 C. Wu and A. Dawson
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
amplifies the effect of populism on trust in voting by mail. To facilitate the
interpretation of the interaction effects in Models (2) and (3), Figure 4 pro-
vides a visualization of the average marginal effects of populism on trust in
voting by mail by political ideology (left) and levels of political news en-
gagement (right).
We replicate the analysis using the data from the DCS. Table 3 reports the
results. Note that, this time, the dependent variable trust in voting by mail
has only four categories, including 1 ¼strongly disagree to 4 ¼strongly
agree. Model (1) shows that both populist beliefs and a right-wing political
ideology are significantly associated with lower odds of believing that voting
by mail is equally as trustworthy as voting in person. News consumption
seems to be associated with higher odds of trusting in voting by mail
(p¼0.000). Model (2) suggests that respondents who are Conservative sup-
porters with populist orientations are more likely to have lower trust in
Table 2. Continued.
Model (1) Model (2) Model (3)
cut2 0.114 0.121 0.119
(0.000) (0.000) (0.000)
cut3 0.215 0.228 0.224
(0.000) (0.000) (0.000)
cut4 0.738 0.783 0.771
(0.082) (0.158) (0.133)
N 10,171 10,171 10,171
Note: Exponentiated coefcients; p-values in parentheses.
Figure 4. Average marginal effects of populism with 95 percent CIs by politi-
cal ideology (left) and levels of political news engagement (right), CES.
The Trump Effect? 19
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
Table 3. Ordinal logistic regressions capturing the Trump effect on trust in
voting by mail, DCS.
Model (1) Model (2) Model (3)
The Trump effect
Populism index 0.779 0.809 1.071
(0.000) (0.000) (0.544)
Conservative 0.404 0.413 0.407
(0.000) (0.000) (0.000)
News consumption 1.208 1.210 1.210
(0.000) (0.000) (0.000)
Right-wing populism (Populism#conservative) 0.852
(0.078)
Media amplication (Populism#news consumption) 0.915
(0.002)
Covariates
Political knowledge 0.952 0.955 0.954
(0.057) (0.073) (0.066)
Social trust 1.539 1.539 1.542
(0.000) (0.000) (0.000)
Female 0.875 0.874 0.880
(0.037) (0.035) (0.047)
Age 1.002 1.002 1.001
(0.498) (0.480) (0.593)
Education 1.082 1.083 1.082
(0.000) (0.000) (0.000)
Household income 1.069 1.070 1.070
(0.002) (0.002) (0.002)
White 1.028 1.032 1.035
(0.749) (0.720) (0.690)
MS_Single (ref. Married/living together) 1.057 1.061 1.062
(0.495) (0.464) (0.455)
MS_Divorced 1.029 1.026 1.039
(0.821) (0.836) (0.762)
MS_Separated 0.880 0.882 0.860
(0.490) (0.497) (0.421)
MS_Widowed 0.800 0.795 0.818
(0.202) (0.186) (0.254)
Alberta 0.739 0.741 0.745
(0.005) (0.005) (0.006)
Quebec 0.631 0.631 0.627
(0.000) (0.000) (0.000)
Estimated cut points
cut1 0.364 0.374 0.373
(0.000) (0.000) (0.000)
(continued)
20 C. Wu and A. Dawson
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
voting by mail (p¼0.078). Model (3) suggests that social media use intensi-
fies the effect of populism on trust in voting by mail. Figure 5 visualizes the
interaction effects in Models (2) and (3) in table 3.
Finally, we compare Quebec to the rest of Canada and replicate Model
(3), which includes the interaction term between political media exposure
and populism (please see table A3 for full results). A significant media am-
plification effect is found only in primarily English-speaking regions outside
Quebec. In Quebec, media use and political news engagement do not seem
to intensify the negative effect of populist orientations on trust in voting by
mail. Figure 6 visualizes the main results. Using data from the CES, Panel A
shows that there is a significant amplifying effect of political news engage-
ment outside Quebec, while Panel B shows that in French-speaking Quebec
this effect is not significant. The DCS data provide parallel findings in indi-
cating that news consumption amplifies the negative effect of populism out-
side (Panel C), but not within (Panel D), Quebec.
Table 3. Continued.
Model (1) Model (2) Model (3)
cut2 1.383 1.424 1.418
(0.145) (0.113) (0.115)
cut3 7.263 7.485 7.476
(0.000) (0.000) (0.000)
N 3,712 3,712 3,712
Note: Exponentiated coefcients; p-values in parentheses.
Figure 5. Average marginal effects of populism with 95 percent CIs by politi-
cal ideology (left) and levels of news consumption (right), DCS.
The Trump Effect? 21
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
Conclusion
Voting by mail is considered to be among the most widespread and signifi-
cant reforms adopted in modern elections (Menger and Stein 2020). As elec-
tions are a crucial mechanism linking citizens to the political system
(Hooghe and Stiers 2016), the level of confidence of the electorate in postal
voting, and the electoral system more broadly, has important consequences.
When the trustworthiness of electoral systems is called into question, this
can harm both political participation and the legitimacy of the political sys-
tem (Birch 2008; Carreras and _
Irepo
glu 2013; Harell and Stephenson 2022)
and/or lead to the voicing of discontent and the support of populist anti-
system parties (B
elanger and Nadeau 2005; Hooghe 2018).
In the United States, Trump’s unproven accusations of fraud surrounding
the mail-in ballot during the 2020 election had important consequences. Not
only did they have a significant influence on public opinion and infamously
culminated in the Capitol Hill insurrection, but they also led to attempts to
restrict access to voting by mail and the passing of other antidemocratic leg-
islation (New York Times 2021). In this context, we examine whether the in-
fluence of Trump’s false claims spread across the border; that is, whether his
unsubstantiated claims of electoral fraud also created a division in the way
Canadians view voting by mail. The data from both the CES and DCS pro-
vide clear indirect evidence of a “Trump effect” in Canada pertaining to trust
in postal voting. First, in accordance with Hypothesis 1, we find that
Figure 6. Average marginal effects of populism with 95 percent CIs by levels
of political news engagement/news consumption, CES & DCS.
22 C. Wu and A. Dawson
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
Canadians with populist orientations are less trusting of voting by mail.
Second, individuals in Canada most likely to be sympathetic to Trump’s po-
litical pronouncements—those who hold right-wing populist views—have
significantly less trust in voting by mail, lending support to Hypothesis 1a.
Furthermore, as in the United States and consistent with Hypothesis 2, the
effect of populist attitudes on trust in postal voting is moderated by one’s
level of political media exposure, with higher levels of exposure exacerbat-
ing the effect among those who hold populist views. The results of the
Quebec experiment suggest that at least part of this political media exposure
effect in Canada is driven by exposure to US-based media. That is, the me-
dia exposure amplification effect is significantly stronger in English-
speaking Canada than in predominantly French-speaking Quebec, following
Hypothesis 3. As a whole, these findings have important implications for
Canada, where the percentage of votes cast by mail increased more than ten-
fold between the two most recent elections.
This study is unique in that it examines a case of a cross-border diffusion ef-
fect on public opinion as a result of false claims of electoral fraud from a for-
eign political leader referring to an electoral system in a different country. As
mentioned, Canada can be considered an ideal natural experiment to examine
this type of diffusion effect for two reasons. First, the absence of any of the
major political parties in Canada, including the People’s Party of Canada,
attempting to delegitimize voting by mail provides more confidence in attribut-
ing the effects of cues from Trump. Second, while many countries follow US
news, Canadians do so comparatively closely, providing heightened opportuni-
ties for exposure to cues from US politicians. Our findings therefore show that
the politicization of mail-in voting by President Trump has important implica-
tions for the legitimacy of the electoral system not only in the United States,
but also in Canada and potentially in other parts of the world.
This study is not without its limitations, however. First, the question per-
taining to trust in voting by mail was not included in earlier waves of either
the DCS or CES. Consequently, we are unable to further validate the “Trump
effect” by confirming a weaker association between right-wing populist orien-
tations and trust in postal voting prior to Trump’s pronouncements in 2020.
While this would have provided additional confidence in the nonspuriousness
of our findings, our Canada-wide analyses combined with the natural experi-
ment of Quebec provide some evidentiary support for a “Trump effect.” That
said, the evidence presented here is suggestive and indirect, as we do not have
concrete measures directly linked to Trump. Due to the cross-sectional nature
of both CES and DCS data, we are also constrained in making causal infer-
ences concerning several associations highlighted in our analyses. For more
robust evidence regarding the impact of right-wing ideology and media expo-
sure on people's mistrust in voting by mail, future research could explore the
collection of panel data or employ experimental methods.
The Trump Effect? 23
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
Second, the validity of various survey-based measures of media expo-
sure—including some used in this study—has been called into question
(Dilliplane, Goldman, and Mutz 2013; Guess et al. 2019; Goldman and
Warren 2020). Each measure of media exposure that we use is therefore not
without its specific limitations. However, given that these indicators, which
examine different aspects of political media exposure, lead to similar results,
taken together they establish the robustness of the interaction effect with
populist attitudes on trust in voting by mail. Third, the generalizability of
our findings may be limited outside the US–Canada context. While the diffu-
sion effect on public opinion that we describe may be potentially occurring
in other countries, our findings of a “Trump effect” may be due to the
unique relationship between Canada and the United States. Consequently, in-
vestigating whether this effect is replicable in other contexts is an interesting
avenue for future research.
Appendix
Figure A1. Change in trust in voting by mail over time, DES and CES. The
gure depicts the change in the probability of indicating “somewhat agree” or
“strongly agree” to the statement “Voting by mail is equally as trustworthy as
voting in person” using the recorded date for each survey response in the DCS
(May 20–June 7, 2021) and CES (August 7–September 19). It shows that
Canadians’ trust in voting by mail remained stable within the DCS and CES
periods. The lower trust from the CES is likely due to the difference in re-
sponse coding. The CES includes ve ordinal categories from 1 ¼strongly
disagree to 5 ¼strongly agree, while the DCS includes only four ordinal cate-
gories without the middle category of neither disagree nor agree. The anoma-
lous value on June 7 is likely the result of interviewing only eight respondents
that day, the nal day of the survey.
24 C. Wu and A. Dawson
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
Figure A2. Media exposure amplies the negative effect of populism on trust in
voting by mail, CES. The gure reports the amplifying effect of media use using
alternative measures of media exposure using data from the CES. Panel A is based
on the measure of “how closely follow political news on TV, radio, newspapers,
or the internet over the last 12 months.” Panel B is based on the measure of “Used
social media to discuss politics or political issues on social media.” Results show
that there is a signicant amplifying effect regardless of the measure.
Figure A3. Media exposure amplies the negative effect of populism on trust
in voting by mail, DCS. The gure reports the amplifying effect of media use
using alternative measures of media exposure using data from the DCS.
Political media exposure is measured using the item “Used social media to dis-
cuss politics or political issues over the last 12 months.”
The Trump Effect? 25
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
Table A1. Comparing codings of trust in voting by mail in CES and DES.
DES (May 20–June 7, 2021) CES (August 17–September 19, 2021)
Freq. % Freq. %
Original Original
Strongly disagree 916 11.53 Strongly disagree 1,659 12.46
Somewhat disagree 1,527 19.21 Somewhat disagree 2,187 16.42
Somewhat agree 2,823 35.53 Neither agree nor disagree 1,998 15.00
Strongly agree 2,680 33.72 Somewhat agree 3,532 26.53
Strongly agree 3,939 29.58
Binary Binary
Trust (0 ¼no, 1¼yes) 69.25 Trust (0 ¼no, 1¼yes) 56.11
Figure A4. Comparing the amplifying effect of media exposure on the nega-
tive effect of populism on trust in voting by mail among the political left and
the political right. The gure visualizes the three-way interactions between
populism, media use, and political ideology. It shows that the amplifying ef-
fect of media exposure on the negative effect of populism on trust in voting
by mail is found to be stronger among the political right than the political left.
26 C. Wu and A. Dawson
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
Table A2. Populism factor loadings (pattern matrix) and unique variances.
Factor 1 Uniqueness
CES populism items
What people call compromise in politics is really
just selling out on one’s principles 0.672 0.548
Most politicians do not care about the people 0.849 0.280
Most politicians are untrustworthy 0.725 0.475
The people, and not politicians, should make our
most important policy decisions 0.635 0.597
Most politicians care only about the interests of
the rich and powerful 0.803 0.355
DCS populism items
Rather put my trust in the wisdom of ordinary
people than the opinions of experts and
intellectuals 0.853 0.273
Ordinary people, not career politicians, should
make our most important policy decisions 0.853 0.273
The Trump Effect? 27
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
Table A3. Ordinal logistic regressions comparing the amplifying effect of
political media exposure in ROC and Quebec.
CES DCS
Model
(ROC)
Model
(Quebec)
Model
(ROC)
Model
(Quebec)
The Trump effect
Populism index 0.851 0.700 1.115 0.879
(0.004) (0.002) (0.415) (0.554)
Right ideology
a
0.794 0.866 0.386 0.575
(0.000) (0.000) (0.000) (0.010)
Political news engagement
b
0.915 0.854 1.197 1.324
(0.000) (0.001) (0.000) (0.000)
Media amplication
(populism news engagement)
c
0.907 0.962 0.896 1.009
(0.000) (0.419) (0.001) (0.882)
Covariates
Political knowledge 1.126 1.128 0.945 0.956
(0.051) (0.363) (0.054) (0.397)
Social trust 1.402 1.167 1.505 1.723
(0.000) (0.138) (0.000) (0.000)
Female 0.963 0.854 0.940 0.720
(0.487) (0.125) (0.402) (0.011)
Age 0.991 0.991 1.001 1.004
(0.000) (0.010) (0.732) (0.424)
Education 1.097 1.123 1.068 1.148
(0.000) (0.000) (0.002) (0.001)
Household income 1.000 1.000 1.076 1.035
(0.000) (0.000) (0.003) (0.468)
White 0.917 1.275 1.080 0.677
(0.166) (0.062) (0.393) (0.189)
MS_Single (ref. Married/living together) 1.013 1.178 1.040 1.038
(0.847) (0.169) (0.675) (0.822)
MS_Divorced 0.940 0.879 0.964 1.180
(0.507) (0.442) (0.797) (0.534)
MS_Separated 0.779 0.703 0.768 1.156
(0.097) (0.320) (0.205) (0.725)
MS_Widowed 1.107 0.785 0.894 0.389
(0.472) (0.497) (0.562) (0.028)
Estimated cut points
d
cut1 0.0380 0.0805 0.379 0.507
(0.000) (0.000) (0.000) (0.225)
cut2 0.117 0.282 1.272 2.979
(0.000) (0.000) (0.336) (0.051)
cut3 0.212 0.624 6.601 17.140
(0.000) (0.186) (0.000) (0.000)
(continued)
28 C. Wu and A. Dawson
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
Supplementary Material
Supplementary Material may be found in the online version of this article:
https://doi.org/10.1093/poq/nfae020.
Funding
This work was supported by the York University Research Chairs Program
(to C.W.), and the SSHRC Insight Grants Program (to C.W.).
Data Availability
Replication data can be downloaded here: 2021 Canadian Election Study
(CES)—Consortium on Electoral Democracy (C-Dem) (harvard.edu) and
here: 2021 Democracy Checkup—Consortium on Electoral Democracy (C-
Dem) (harvard.edu); full analysis code and additional analysis are contained
in section B of the Supplementary Material that accompanies this article.
References
Amlani, Sharif, and Samuel Collitt. 2022. “The Impact of Vote-By-Mail Policy on Turnout and
Vote Share in the 2020 Election.” Election Law Journal: Rules, Politics, and Policy
21:135–49.
Anderson, Christopher J., and Andrew J. LoTempio. 2002. “Winning, losing and political trust
in America.” British Journal of Political Science 32:335–51.
Barber, Michael, and John B. Holbein. 2020. “The Participatory and Partisan Impacts of
Mandatory Vote-By-Mail.” Science Advances 6:eabc7685.
Table A3. Continued.
CES DCS
Model
(ROC)
Model
(Quebec)
Model
(ROC)
Model
(Quebec)
cut4 0.712 2.504
(0.082) (0.010)
N 8,043 2,128 2,800 912
Note: Exponentiated coefcients; p-values in parentheses.
a
In DCS data, this is indicated by Conservative.
b
In DCS data, this is indicated by news consumption.
c
In DCS data, this is indicated by the interaction between populism and news consumption.
d
In DCS data, the outcome has four categories, and therefore three cuts.
The Trump Effect? 29
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
BBC News. 2021. “Canada Election: Why It’s Easier to Vote in Canada than the US.”
September 18, 2021, https://www.bbc.com/news/world-us-canada-58589809.
B
elanger,
Eric, and Richard Nadeau. 2005. “Political Trust and The vote in Multiparty
Elections: The Canadian Case.” European Journal of Political Research 44:121–46.
Berinsky, Adam J., Nancy Burns, and Michael W. Traugott. 2001. “Who Votes by Mail?: A
Dynamic Model of the Individual-Level Consequences of Voting-By-Mail Systems.” Public
Opinion Quarterly 65:178–97.
Berning, Carl C., and Conrad Ziller. 2017. “Social Trust and Radical Right-Wing Populist Party
Preferences.” Acta Politica 52:198–217.
Berlinski, Nicolas, Margaret Doyle, Andrew M. Guess, Gabrielle Levy, Benjamin Lyons, Jacob
M. Montgomery, Brendan Nyhan, and Jason Reier. 2023. “The Effects of Unsubstantiated
Claims of Voter Fraud on Condence in Elections.” Journal of Experimental Political
Science 10:34–49.
Birch, Sarah. 2008. “Electoral Institutions and Popular Condence in Electoral Processes: A
Cross-National Analysis.” Electoral Studies 27:305–20.
Biswas, Atanu, Samarjit Das, and Soumyadeep Das. 2019. “OLS: Is That So Useless for
Regression with Categorical Data?” Advances in Analytics and Applications 227–42.
Booth, John A., and Patricia Bayer Richard. 2014. Latin American Political Culture: Public
Opinion and Democracy. Washington, DC: CQ Press.
Bowler, Shaun, Thomas Brunell, Todd Donovan, and Paul Gronke. 2015. “Election
Administration and Perceptions of Fair Elections.” Electoral Studies 38:1–9.
Bridgman, Aengus, Eric Merkley, Oleg Zhilin, Peter John Loewen, Taylor Owen, and Derek
Ruths. 2021. “Infodemic Pathways: Evaluating the Role that Traditional and Social Media
Play in Cross-National Information Transfer.” Frontiers in Political Science 3:20.
Budd, Brian. 2021. “Maple-Glazed Populism: Political Opportunity Structures and Right-Wing
Populist Ideology in Canada.” Journal of Canadian Studies 55:152–76.
Carreras, Miguel, and Yasemin _
Irepo
glu. 2013. “Trust in Elections, Vote Buying, and Turnout
in Latin America.” Electoral Studies 32:609–19.
Dawson, Andrew. 2021. “The Achilles Heel of Democracy? A Macro Cross-National
Assessment of the Correlates of State Legitimacy.” Social Science Research 97:102574.
Dilliplane, Susanna, Seth K. Goldman, and Diana C. Mutz. 2013. “Televised Exposure to
Politics: New Measures for a Fragmented Media Environment.” American Journal of
Political Science 57:236–48.
Elections Canada. 2022. “Report on the 44
th
General Election of September 20, 2021.” https://
www.elections.ca/content.aspx?section=res&dir=rep/off/sta_ge44&document=p2&lang=e.
Elmer, Greg, and Anthony Burton. 2022. “Rebel Personalities: Canada’s Far-Right Media.”
First Monday. https://doi.org/10.5210/fm.v27i5.12546.
Fortier, John C., and Norman J. Ornstein. 2002. “The Absentee Ballot and the Secret Ballot:
Challenges for Election Reform.” University of Michigan Journal of Law Reform 36:483.
Gillies, Jamie. 2023. “The Authoritarian Elephant Next Door?: A Canadian and Comparative
Perspective Amidst American Democratic Backsliding & Uncertainty.” American Behavioral
Scientist 67:598–611.
Gomila, Robin. 2021. “Logistic or Linear? Estimating Causal Effects of Experimental
Treatments on Binary Outcomes using Regression Analysis.” Journal of Experimental
Psychology. General 150:700–709.
Goldman, Seth K., and Stephen Warren. 2020. “Debating How to Measure Media Exposure in
Surveys”. In The Oxford Handbook of Electoral Persuasion, edited by E. Suhay, B. Grofman,
and A. H. Trechsel, 104. USA: Oxford University Press. https://doi.org/10.1093/oxfordhb/
9780190860806.013.28.
Gronke, Paul, and Peter Miller. 2012. “Voting by Mail and Turnout in Oregon: Revisiting
Southwell and Burchett.” American Politics Research 40:976–97.
30 C. Wu and A. Dawson
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
Guess, Andrew, Kevin Munger, Jonathan Nagler, and Joshua Tucker. 2019. “How Accurate are
Survey Responses on Social Media and Politics?” Political Communication 36:241–58.
Hall, T. E., J. QuinMonson, and K. D. Patterson. 2009. “The Human Dimension of Elections:
How Poll Workers Shape Public Condence in Elections.” Political Research Quarterly
62:507–22.
Harell, Allison, and Laura B. Stephenson. 2022. “Elections During a Health Crisis: Voter
Involvement and Satisfaction across Pandemic Elections in Canada.” IRPP. May 2022.
https://centre.irpp.org/wp-content/uploads/sites/3/2022/05/Elections-During-a-Health-Crisis-
Voter-Involvement-and-Satisfaction-across-Pandemic-Elections-in-Canada.pdf.
Harell, Allison, Laura B. Stephenson, Daniel Rubenson, and Peter John Loewen. 2022. “2021
Democracy Checkup.” https://doi.org/10.7910/DVN/KCIK9D., Harvard Dataverse, V1; 2021
Democracy Checkup Survey Codebook_v1.pdf.
Hooghe, Marc. 2018. “Trust and Elections” In The Oxford Handbook of Social and Political
Trust, edited by E. Uslaner, 617–31. New York, NY: Oxford University Press.
Hooghe, Marc, and Dieter Stiers. 2016. “Elections as a Democratic Linkage Mechanism: How
Elections Boost Political Trust in a Proportional System.” Electoral Studies 44:46–55.
Justwan, Florian, and Ryan D. Williamson. 2022. “Trump and Trust: Examining the
Relationship between Claims of Fraud and Citizen Attitudes.” PS: Political Science &
Politics 55:462–69.
Karp, Jeffrey A., Alessandro Nai, and Pippa Norris. 2018. “Dial ‘F’for Fraud: Explaining
Citizens Suspicions About Elections.” Electoral Studies 53:11–19.
Kerr, Nicholas. 2013. “Popular Evaluations of Election Quality in Africa: Evidence from
Nigeria.” Electoral Studies 32:819–37.
Kerr, Nicholas, and Anna L€
uhrmann. 2017. “Public Trust in Manipulated Elections: The Role of
Election Administration and Media Freedom.” Electoral Studies 50:50–67.
King, B. A. 2020. “Waiting to Vote: The Effect of Administrative Irregularities at Polling
Locations and Voter Condence.” Policy Studies 41:230–48.
King, B. A., and A. Barnes. 2019. “Descriptive Representation in Election Administration: Poll
Workers and Voter Condence.” Election Law Journal: Rules, Politics, and Policy 18:16–30.
King, Bridgett A. 2017. “Policy and Precinct: Citizen Evaluations and Electoral Condence.”
Social Science Quarterly 98:672–89.
Laycock, David H. 1990. Populism and Democratic thought in the Canadian Prairies,
1910–1945. Toronto: University of Toronto Press.
Levy, M. 2021. “Winning cures Everything? Beliefs about Voter Fraud, Voter Condence, and
the 2016 Election.” Electoral Studies 74:102156.
Lockhart, Mackenzie, Seth J. Hill, Jennifer Merolla, Mindy Romero, and Thad Kousser. 2020.
“America’s Electorate is Increasingly Polarized along Partisan Lines about Voting by Mail
during the COVID-19 Crisis.” Proceedings of the National Academy of Sciences of the United
States of America 117:24640–42.
Maas, Willem. 2015. “Report on Access to Electoral Rights: Canada” [GLOBALCIT], EUDO
Citizenship Observatory, Electoral Rights Reports. https://hdl.handle.net/1814/36495.
Maldonado, Arturo, and Mitchell A. Seligson. 2013. “Winners, Losers and Trust in Elections in
24 Nations in Latin America and the Caribbean: An Analysis using the Americas Barometer
2004-2012.” In Prepared for delivery at the 2013 Annual Workshop, “Concepts and Indices
of Electoral Integrity”, 1–30. The Weatherhead Center for International Affairs,
Harvard University.
McDonald, Michael P. 2022. From Pandemic to Insurrection: Voting in the 2020 US
Presidential Election. Boston: De Gruyter.
Menger, A., and R. M. Stein. 2020. “Choosing the Less Convenient Way to Vote: An Anomaly
in Vote by Mail Elections.” Political Research Quarterly 73:196–207.
The Trump Effect? 31
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
Mitchell, Amy, Mark Jurkowitz, J. Baxter Oliphant, and Elisa Shearer. 2020. “Political Divides,
Conspiracy Theories and Divergent News Sources Heading into 2020 Election.” https://policy
commons.net/artifacts/1426222/political-divides-conspiracy-theories-and-divergent-news-sour
ces-heading-into-2020-election/2040620/ on 01 September 2023. CID: 20.500.12592/jb1xs1.
Monsiv
ais-Carrillo, Alejandro. 2023. “Happy Winners, Sore Partisans? Political Trust,
Partisanship, and the Populist Assault on Electoral Integrity in Mexico.” Journal of Politics in
Latin America 15:72–95.
Monroe, Nathan W., and Dari E. Sylvester. 2011. “Who Converts to Vote-by-Mail? Evidence
from a Field Experiment.” Election Law Journal 10:15–35.
Mudde, Cas, and Crist
obal Rovira Kaltwasser. 2017. Populism: A Very Short Introduction.
London: Oxford University Press.
New York Times. 2021. “In State Houses, Stolen Election Myth Fuels GOP Drive to Rewrite
Rules.” March 2, 2021, https://www.nytimes.com/2021/02/27/us/republican-voter-suppres
sion.html.
Otaola, Miguel Lara. 2018. “To Include or Not to Include? Party Representation in Electoral
Institutions and Condence in Elections: A Comparative Study of Latin America.” Party
Politics 24:598–608.
Pennycook, Gordon, and David G. Rand. 2021. “Examining False Beliefs about Voter Fraud in
the Wake of the 2020 Presidential Election.” Harvard Kennedy School Misinformation
Review https://nrs.harvard.edu/URN-3:HUL.INSTREPOS:37367184.
Perry, Barbara, and Ryan Scrivens. 2019. Right-Wing Extremism in Canada. Cham,
Switzerland: Springer Nature.
Perry, Barbara, Tanner Mirrlees, and Ryan Scrivens. 2017. “The Dangers of Porous Borders:
The” Trump Effect” in Canada.” Journal of Hate Studies 14:53–75.
Pew Research Center. 2018. “Many People in Other Countries Closely Follow News about the
U.S.” https://www.pewresearch.org/fact-tank/2018/01/16/many-people-in-other-countries-
closely-follow-news-about-the-u-s/.
Reller, Cassidy, Anthony Anderson, and Thad Kousser. 2022. “Are Elite Cues Necessary to
Drive the “Winner Effect” on Trust in Elections?” Electoral Studies 80:102541.
Robitzsch, Alexander. 2020. “Why Ordinal Variables Can (Almost) always be treated as
Continuous Variables: Clarifying Assumptions of Robust Continuous and Ordinal Factor
Analysis Estimation Methods.” Frontiers in Education 5:589965.
Roccato, Michele, Piergiorgio Corbetta, Nicoletta Cavazza, and Pasquale Colloca. 2019.
“Assessment of Citizens’ Populist Orientations: Development and Validation of the POPulist
ORientation (POPOR) Scale.” Social Science Quarterly 100:2148–67.
Schulz, A., P. M€uller, C. Schemer, D. S. Wirz, M. Wettstein, and W. Wirth. 2018. “Measuring
Populist Attitudes on Three Dimensions.” International Journal of Public Opinion Research
30:316–26.
Sances, Michael W., and Charles Stewart III. 2015. “Partisanship and Condence in the Vote
Count: Evidence from US National Elections Since 2000.”Electoral Studies 40:176–88.
Shino, Enrijeta, Daniel A. Smith, and Laura Uribe. 2022. “Lying for Trump? Elite Cue-Taking
and Expressive Responding on Vote Method.” Public Opinion Quarterly 86:837–61.
Sinclair, Betsy, Steven S. Smith, and Patrick D. Tucker. 2018. ““It’s Largely a Rigged System”:
Voter Condence and the Winner Effect in 2016.” Political Research Quarterly 71:854–68.
Southwell, Priscilla L. 2009. “Analysis of the Turnout Effects of Vote by Mail Elections,
1980–2007.” The Social Science Journal 46:211–16.
Statistics Canada. 2022. “Table 98-10-0169-01, Mother Tongue by First Ofcial Language
Spoken, Language Spoken Most Often at Home and Other Language(S) Spoken Regularly at
Home: Canada, Provinces and Territories, Census Metropolitan Areas and Census
Agglomerations with Parts.” https://doi.org/10.25318/9810016901-eng.
32 C. Wu and A. Dawson
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
Statistics Canada. 2023. “Table 15-10-0004-01, Population by Knowledge of Ofcial Languages
and Geography, 1951 to 2021.” https://doi.org/10.25318/1510000401-eng.
Stephenson, Laura B., Allison Harell, Daniel Rubenson, and Peter John Loewen. 2022. “2021
Canadian Election Study (CES).” https://doi.org/10.7910/DVN/XBZHKC., Harvard
Dataverse, V3; 2021 Canadian Election Study Technical Report and Codebook v2.pdf.
Stewart III Charles. 2020. “How We Voted in 2020: A First Look at the Survey of the
Performance of American Elections.” MIT Election Data and Science Lab 15:2020–12.
Thompson, Daniel M., Jennifer A. Wu, Jesse Yoder, and Andrew B. Hall. 2020. “Universal
Vote-By-Mail has no Impact on Partisan Turnout or Vote Share.” Proceedings of the National
Academy of Sciences of the United States of America 117:14052–56.
United States Census Bureau. 2021. “Majority of Voters Used Nontraditional Methods to Cast
Ballots in 2020.” https://www.census.gov/library/stories/2021/04/what-methods-did-people-
use-to-vote-in-2020-election.html.
Vonnahme, Greg, and Beth Miller. 2013. “Candidate Cues and Voter Condence in American
Elections.” Journal of Elections, Public Opinion & Parties 23:223–39.
West, Darrell M. 2020. “How Does Vote-by-Mail Work and Does It Increase Election Fraud?”
Brookings, June 22. https://www.brookings.edu/articles/how-does-vote-by-mail-work-and-
does-it-increase-election-fraud/.
Wettstein, Martin, Anne Schulz, Marco Steenbergen, Christian Schemer, Philipp M€
uller,
Dominique S. Wirz, and Werner Wirth. 2020. “Measuring Populism across Nations: Testing
for Measurement Invariance of an Inventory of Populist Attitudes.” International Journal of
Public Opinion Research 32:284–305.
Wu, Cary, Rima Wilkes, and David C. Wilson. 2022. “Race & Political Trust: Justice as a
Unifying Inuence on Political Trust.” Daedalus 151:177–99.
Yoder, Jesse, Cassandra Handan-Nader, Andrew Myers, Tobias Nowacki, Daniel M. Thompson,
Jennifer A. Wu, Chenoa Yorgason, and Andrew B. Hall. 2021. “How Did Absentee Voting
Affect the 2020 US Election?” Science Advances 7:eabk1755.
# The Author(s) 2024. Published by Oxford University Press on behalf of American Association for Public
Opinion Research.
This is an Open Access article distributed under the terms of the Creative Commons Attribution-
NonCommercial-NoDerivs licence (https://creativecommons.org/licenses/by-nc-nd/4.0/), which permits
noncommercial reproduction and distribution of the work, in any medium, provided the original work is not
altered or transformed in any way, and that the work is properly cited. For commercial reuse, please contact
journals.permissions@oup.com
Public Opinion Quarterly, 2024, 00, 1–33
https://doi.org/10.1093/poq/nfae020
Original article
The Trump Effect? 33
Downloaded from https://academic.oup.com/poq/advance-article/doi/10.1093/poq/nfae020/7715006 by guest on 16 July 2024
