Content uploaded by Robert Fahey
Author content
All content in this area was uploaded by Robert Fahey on Oct 15, 2020
Content may be subject to copyright.
Belief in Conspiracy Theories and Socio-Political
Identity in Japan
Robert A. Fahey
Waseda University
Airo Hino
Waseda University
Stefano Camatarri
Catholic University of Louvain
Sebastian Jungkunz
University of Duisburg-Essen
& University of Bamberg
& Zeppelin University
September 2020
Draft: Not for citation or circulation without permission of the authors.
1 Introduction
While belief in conspiracy theories has been recognised as part of the socio-
political landscape for many decades (Aaronovitch 2010; Davis 1971), in recent
years there has been an uptick in interest in this phenomenon due to the pro-
posed link between belief in conspiracies (specifically, in conspiracies by elite
groups to control and oppress the “good people” of a nation) and populism
(Castanho Silva et al. 2017; Hawkins 2009; Mudde 2004), and their connec-
tion to propaganda operations or “fake news”. While a dearth of high-quality
empirical research into the incidence of conspiracy beliefs makes comparison
difficult, media attention alone suggests that the 21st century has seen a surge
in such beliefs — ranging from beliefs that the U.S. government was involved
in the terrorist attacks of September 11th 2001 (a belief held in some form by
1
around a third of U.S. adults, according to Stempel et al. (2007)) or that school
shooting incidents are staged “false flag” events by gun control proponents, to
anti-science conspiracy theories holding that climate change is not real or that
vaccines cause illnesses such as autism. U.S. President Donald Trump effectively
launched his political career as a prominent proponent of the conspiracy theory
that former President Barack Obama was not born in the United States and
had falsified his birth certificate (the so-called “birther” movement), and his
presidency has received vocal support from believers in a collection of conspir-
acy theories called “QAnon”, which hold that Trump is secretly waging a war
against a cabal of wealthy, high-profile paedophiles and child traffickers, who
include many of his political opponents and critics. Most recently, conspiracy
theories have arisen stating that the COVID-19 novel coronavirus pandemic was
engineered — either by China or by Trump’s political opponents, depending on
the specific theory — or does not exist at all, and is merely a manufactured
panic designed to allow philanthropist and former Microsoft CEO Bill Gates to
inject microchips into citizens under the pretence of vaccinating them. Even
more “classic” conspiracy theories, such as the belief that the government is
hiding evidence of contact with aliens, that the moon landings were faked or
even that the Earth is flat, also appear to have gained fresh adherents in recent
years, perhaps due to a kind of “halo effect” from the broad climate promoted
by other conspiracy theories of not merely questioning authority, but regarding
all authorities as intrinsically duplicitous and untrustworthy.
While earlier work on conspiracy theories focused on examinations of the the-
ories themselves (e.g. Clarke 2002), more recent work in this field has concerned
itself with understanding the individuals and communities who believe in and
actively spread the theories (Brotherton et al. 2013; Douglas et al. 2019; Oliver
and Wood 2014; Radnitz and Underwood 2017). A core question is whether a
2
tendency to believe in conspiracy theories is attributable to specific personality
traits, or “pathological”, as Radnitz and Underwood (2017) would have it, or
whether it is situational, with conspiracy beliefs being formed in response to
specific events for which the individual finds it impossible, for some reason, to
accept the standard, scientific or evidence-based explanation. In parallel with
this, a further strand of research has emphasised the community aspects of
modern conspiracy theories, noting the concern that such (largely online) com-
munities effectively form “epistemological counter-publics” (Klein et al. 2018;
Schatto-Eckrodt et al. 2020), and the powerful sense of camaraderie and partic-
ipation in a “great struggle” that participation in such a community can offer,
especially to individuals who feel isolated or lacking control in their ordinary
daily lives. Douglas et al. (2017) combines several of these factors into a three-
pronged explanation of the psychology of conspiracy belief, in which the motives
for developing or holding such beliefs are a combination of epistemic (meaning
that conspiracies provide consistent explanations in the face of uncertainty and
confusion), existential (in that conspiracy beliefs provide a sense of control and
security to individuals who may otherwise feel powerless), and social (in that
conspiracies generally blame powerful, amoral external actors for bad outcomes
and thus allow the individual to uphold a positive image of their in-group).
From the perspective designing research to explore the correlates of conspir-
acy belief, we hypothesised that these first two factors — epistemic and existen-
tial — could be captured effectively by the measurement of anomia. “Anomie”
describes the perception that society is in a state of decline or breakdown (Srole
1956); more recent work has identified the perceived breakdown of social fabric
(decline in morality and trust, etc.) and the perceived failure of leadership (lack-
ing legitimacy and effectiveness) as two distinct dimensions within this concept
(Teymoori et al. 2016). Anomia, then, is the individual’s psychological response
3
to those perceptions — a sense of confusion, anger and helplessness in the face
of rapid or accelerating social change. The concept encapsulates both the confu-
sion and uncertainty behind epistemic motivations for conspiracy belief, and the
sense of powerlessness behind the existential motivations. Combined with ques-
tions which aim to establish the individuals’ level of satisfaction with their social
relationships (and thus the degree to which they might embrace conspiracy the-
ories in order to make up for a lack of satisfaction with those current in-group
relationships), this should allow a wide range of the hypothesised motives for
conspiracy beliefs to be explored.
1.1 Conspiracy Theories in Japan
Although many of the examples given thus far are of American conspiracy the-
ories (though belief in these often extends far beyond the borders of the United
States, as shown by Swami et al. (2010), which examined belief in 9/11 con-
spiracy theories among British adults), the existence of and belief in conspiracy
theories is considered to be universal (Prooijen and Douglas 2018). Such theo-
ries are found throughout history — Davis (1971) documents conspiracy theories
dating back to the American Revolution, while Prooijen and Douglas (2017) de-
tails formation of conspiracy theories around the Great Fire of Rome in 64AD
— and in every part of the world. Japan is no exception in this regard. Perhaps
the most famous — and tragic — instance of a conspiracy theory in modern
Japanese history is the widespread rumour that ethnically Korean residents of
Tokyo were poisoning wells in the wake of the 1923 Great Kanto Earthquake,
resulting in violent reprisals in which an estimated 6000 Koreans and other mi-
nority residents were murdered over the course of several weeks. In line with the
universal nature of conspiracy theories, this sequence of events directly echoed
historical pogroms against Jewish communities in Europe, in which false ac-
4
cusations of poisoning wells had often been an inciting factor. More recently,
conspiracy theories asserting that the government was hiding the extent of ra-
dioactivity in Tokyo and North-Eastern Japan emerged in the wake of the 2011
Fukushima Disaster, while claims that Japan’s media (in some versions of the
theory, secretly controlled from behind the scenes by foreign forces) is seeking
to undermine the nation by promoting false or exaggerated information about
war crimes committed during the Second World War have gained traction in
online right-wing communities. Most recently of all, a popular YouTuber who
mostly posts videos promoting conspiracy theories ran as a candidate in Tokyo’s
2020 gubernatorial election on a campaign slogan stating that “COVID-19 is
just a common cold”, and later organised “maskless” demonstrations against
the government’s public health countermeasures in central Tokyo locations.
Despite the existence of such contemporary conspiracy theories in Japan,
until recently almost no research had been conducted on either the psycho-
logical underpinnings of belief in these theories1or even of their incidence
within Japan’s citizenry. Majima and Nakamura (2019) developed and tested a
Japanese version of the Generic Conspiracist Beliefs Scale introduced by Broth-
erton et al. (2013), finding that the scale met standards of survey validity and
functioned as expected in the Japanese context. Although the original scale
measured five conspiracy factors, Majima and Nakamura (2019) suggested that
a two-factor approach (dividing conspiracies involving extraterrestrial life or
contact into a factor separate from all other conspiracies) might be appropriate
in Japan. However, while they tested the correlation between conspiracy be-
liefs and some psychological factors (such as supernatural beliefs), they did not
examine any connection with political ideas such as populism.
In this study, we conduct a large-scale national survey of conspiracy beliefs
1Kagoya (1998) suggests a game theory approach to understanding the spread of conspiracy
theories, but this seems intended as a universal model and the study did not consider any
Japan-specific data.
5
and their correlates, building upon existing survey research into conspiracies
in other countries (Brotherton et al. 2013; Castanho Silva et al. 2017). In
addition to further establishing the prevalence of conspiracy beliefs within the
electorate, the study has two core aims. Firstly, we aim to provide some insight
into why Japanese citizens — who are generally not considered to hold strong
“anti-science” or “anti-expert” views, as is the case for many citizens in other
countries where conspiracy theories have been studied — believe in conspiracy
theories rather than conventionally accepted or scientific explanations. In line
with our interpretation of the three motives outlined by Douglas et al. (2017),
we hypothesise that both anomia and social isolation may be factors involved in
these beliefs. Secondly, we wish to explore the consequences of conspiracy belief
on political alignment, especially in light of the growing literature connecting
conspiracy belief to populism in other countries, and the emergence of two new
populist challenger parties (Reiwa Shinsengumi and The Party to Protect the
People from the NHK ) as national political forces during the 2019 election at
which our survey was fielded.
2 Data and Method
In order to investigate the prevalence and nature of conspiracy theory belief in
Japan, we added a battery of questions on conspiracies to a large-scale, nation-
ally representative survey conducted in the weeks after the House of Councillors
election in July 20192. This survey had 5106 completed entries; after removing
respondents who failed satisficing checks, we were left with 3965 samples.
Ten items were included in the conspiracy theories battery of the survey.
Five of these were drawn from the scale of generic conspiracist beliefs proposed
2The survey was conducted from 26 August to 2 September by the Survey Research Center
in Japan using the samples owned by Rakuten Insight.
6
by Brotherton et al. (2013) (also used in Castanho Silva et al. (2017)), with
one item chosen from each of the five sub-categories of that scale - government
malfeasance, extraterrestrial cover-up, malevolent global conspiracies, personal
well-being related conspiracies, and control of information. These five “general”
conspiracies are items CONS1 through CONS5 in Table 1. In addition, a fur-
ther five items were devised by the authors to attempt to capture domestic,
Japan-specific conspiracy theories, such as belief that ethnically Korean resi-
dents (zainichi Koreans) are secretly controlling aspects of Japanese society, or
that details of nuclear accidents have been hidden from the public. These are
items CONS6 through CONS10 in Table 1. Of these, four were strongly po-
litically polarised — CONS6 (the cover-up of nuclear accidents) and CONS10
(that the nation is being secretly controlled by the right-wing lobby group Nip-
pon Kaigi) are in essence “left-wing” conspiracy theories, while CONS7 (that the
nation is secretly controlled by powerful zainichi Korean groups) and CONS9
(that Japanese war crimes during the Second World War have been exagger-
ated) are “right-wing” conspiracy theories. CONS8, which states that Japan’s
governance is being guided by an unspecified religious group, is more open to
interpretation and general in nature — depending on which religious group they
imagine to be involved, this belief may be held by people of any (or neither)
political orientation.
Respondents were asked to rate each of these items on a five-point Likert
scale, with responses of 5 and 4 indicating that they believe the statement and
believe it “somewhat”, respectively. The average responses to these conspir-
acy items, and the percentage of respondents who believe the statement either
entirely or to some extent, are shown in Figure 13. The item with the lowest
3We note here that while a response of 3 (“cannot say either way”) may be considered
neutral in many survey batteries, since many of these conspiracy items are by their nature
untrue or in defiance of common knowledge, it might reasonably be considered troubling for
citizens to respond that they are unsure of whether they are true or not.
7
Table 1: Conspiracy Theory Items
Item Wording
CONS1 The government uses people as patsies to hide its involvement in criminal ac-
tivity.
CONS2 A small, secret group of people is responsible for making all major world deci-
sions, such as going to war.
CONS3 Evidence of alien contact is being concealed from the public.
CONS4 Experiments involving new drugs or technologies are routinely carried out on
the public without their knowledge or consent.
CONS5 A lot of important information is deliberately concealed from the public out of
self-interest.
CONS6*The government is concealing important facts about the nuclear power plant
accidents.
CONS7*Japanese society is basically run by people of Korean origin.
CONS8*A specific religious group is determining the path of this country’s politics and
economy.
CONS9*Massacres and the Comfort Women issue during the Second World War are
being exaggerated.
CONS10*Important decisions in this country are basically made by Nippon Kaigi.
*indicates conspiracy items specific to the Japanese context.
Items in italics have a strong political orientation and were excluded from analysis of general atti-
tudes to conspiracy theories.
average response and lowest percentage of respondents expressing any degree of
belief is CONS3, “Evidence of alien contact is being concealed from the public”,
with around 16% of respondents indicating belief, while the most commonly held
conspiratorial belief is CONS5, “A lot of important information is deliberately
concealed from the public out of self-interest”, with almost 63% of respondents
indicating belief in this statement.
In addition to the conspiracy theory related items, the survey also included
batteries on political views, populism and anomia, attitudes towards social
change and self-evaluation of respondents’ social positions and life satisfaction.
These batteries were included with a view to allowing us to explore both the
causes of conspiracy belief and its possible effects on citizens’ political views
and activities.
8
Figure 1: Survey Responses to Conspiracy Survey Items (n=3965)
2.1 Constructing Conspiracy Factors
Given the inclusion of politically polarised items in the 10-item battery (CONS6
and CONS10 being “left-wing”, CONS7 and CONS9 being “right-wing”), we
did not anticipate that the items would form a coherent scale4. The correlation
matrix for the battery seen in Figure 2 confirms our expectation, as it is clear at
a glance that the politically polarised items, especially CONS9, do not correlate
well with the other items in the battery. These items appear to be more closely
connected to respondents’ political positions than to their underlying belief in,
or susceptibility to, conspiracy theories in general; on this basis we chose to
exclude the politically polarised items from the current study and focus instead
on the more general conspiracy beliefs encapsulated in CONS1 through CONS5
4While Wood et al. (2012) found that individuals holding conspiratorial beliefs often believe
multiple contradictory conspiracy theories, this finding was limited to individuals believing in
multiple versions of the same essential belief — for example, individuals may believe in the
overarching conspiracy theory that the authorities engaged in a cover-up over the death of
Princess Diana, while believing in multiple conflicting versions of what was being covered up.
As such, the finding would not seem to extend to believing in conspiracy theories whose core
beliefs are essentially polarised to opposite ends of the political spectrum.
9
and CONS8.
Figure 2: Correlation Matrix for Conspiracy Survey Items (n=3965)
We tested the validity of these six items using confirmatory factor analysis
both as a single scale (a 1-factor model) and as a two-factor model which divided
the variables into those referring to “classic”, anti-globalist conspiracy theories
(CONS2, CONS3 and CONS8) and more strongly “anti-government” conspir-
acy theories (CONS1, CONS4 and CONS5). The results are shown in Table 2
and demonstrate that the validity of the 2-factor model showed a marked im-
provement over the 1-factor model. Following the logic of Wood et al. (2012)
regarding the monological nature of conspiracy belief (see also Clarke 2002;
Swami et al. 2010), this 2-factor model divides up the conspiracies according
to the actor whom respondents see as orchestrating the conspiracy — secretive,
largely non-governmental actors in the case of “classic” conspiracies, and gov-
ernmental actors in the case of “anti-government” conspiracies. The 2-factor
model is shown graphically in Figure 3 and the loadings on the two factors are
shown in Table 3. We proceeded to use the factors extracted from this model
10
in our subsequent analyses of the correlates of conspiracy beliefs.
Table 2: Confirmatory Factor Analysis Models
Model N RMSEA SRMR Chi2 CFI Avg.
Loading
Min.
Loading
Lowest
Loading
1-factor 3965 0.112 0.05 378.89 0.906 0.619 0.456 CONS5
2-factor 3975 0.084 0.04 193.32 0.953 0.666 0.513 CONS5
Figure 3: CFA Model (Two Factor)
CONS1
CONS2
CONS3
cons antigov
cons classic
CONS4
CONS5
CONS8
Table 3: Conspiracy Theory Two Factor Model / Loadings
cons classic cons antigov
CONS2 (0.844) CONS1 (0.781)
CONS3 (0.692) CONS4 (0.536)
CONS8 (0.630) CONS5 (0.513)
2.2 Other Variables and Factors
In addition to the conspiracy factors, a number of other variables from the
survey were used in our analysis. Demographic factors — age, gender and level
of education — were used in all regression analyses, with age being divided into
four cohorts (under 35, 35 to 54, 55 to 69, and 70-plus) and level of education
being represented by a dummy variable indicating whether the respondent had
graduated university. Respondents’ self-assessments of their political positions
were used to create dummy variables indicating holding left-wing (values below
4 on the 10-point scale) and right-wing (values above 6) views. A set of variables
11
asking respondents to self-report their level of satisfaction with four key areas
of their lives (relationships with friends, relationships with family, their job and
their finances) were also used in some models.
Furthermore, the survey included a three-item set of variables designed to
measure anomia. The three items used in the survey to measure this sentiment
are shown in Table 4 and combined into a single scale for use in regression
analyses.
Table 4: Anomia Items
Item Wording
ANO1 The world is changing too fast and I don’t feel like I can keep up.
ANO2 Everything is so confusing these days, I no longer know where we are headed.
ANO3 The Japan I love is slipping away and changing too quickly.
3 Factors Contributing to Conspiracy Belief
Previous studies on the formation of conspiracy beliefs have primarily focused on
the question of whether susceptibility to such beliefs is pathological, with specific
personality traits making individuals more likely to believe in conspiracies (e.g.
Oliver and Wood (2014)), or situational, with a wide range of people being likely
to form conspiracy beliefs in response to the correct set of triggers (e.g. Radnitz
and Underwood (2017)). Our study makes no specific assumption in this regard
— it seems likely that elements of both personality and situation play a role in
conspiracy belief formation — but draws on this prior literature to construct a
number of explanatory models for belief in the two conspiracy factors identified
in our survey data.
The first model simply incorporates basic demographic data — age, gen-
der and level of education. We hypothesised that conspiracy theories would be
more widely held by young people, as they are commonly spread online, and
12
that those with a higher level of education would be less likely to express be-
lief in conspiracies. The second model incorporates individuals’ self-reported
political alignment. Here, we hypothesised that people with right-wing views
would be less likely to believe in anti-government conspiracies; this hypothesis
is very specific to the Japanese context, where self-description as “right-wing”
is often indicative of support for the ruling Liberal Democratic Party rather
than of more ideological support for small government or, indeed, general suspi-
cion of government5. The third model incorporates the anomia factor described
above, which we expected to be positively correlated with conspiracy beliefs,
as individuals experiencing anomia seem likely to turn to conspiracies as per-
versely comforting explanations for rapid change and perceived social decline.
A fourth model tests the role of individuals’ life satisfaction on their conspiracy
beliefs; we hypothesised that low satisfaction with certain aspects of social or
professional life might be a factor driving people to turn to conspiracies either
as an explanation for difficult aspects of their lives, or as a welcoming alterna-
tive community of fellow believers. Finally, the fifth model incorporates both
anomia and life satisfaction variables into a single model.
The results of these models are shown for the “classic” conspiracy theories
in Table 5, and for the “anti-government” conspiracy theories in Table 66. The
effects of age and education are confirmed in these analyses across all five mod-
els for both conspiracy factors, with younger cohorts (notably those under 35)
being significantly more likely to endorse conspiracy beliefs, while those over 70
and those with university degrees are significantly less likely to do so. Politi-
5Since it is difficult to establish the direction of the causal relationship between conspiracy
belief and political alignment — either seems equally likely to be the cause of the other as to
be caused by the other — we excluded this factor in the other models, but present alternative
regression tables including the left- and right-wing alignment variables in the Appendix.
6These regression models were run without weights, which may introduce some uncertainty
in the results; while the survey sample was representative stratifying age, gender and region
categories with census data, the removal of satisficers and non-response rates on certain items
may have skewed the data in some areas. Appropriate weighting of the data is included in
the future planning for this research project.
13
cal alignment is strongly related to “anti-government” conspiracy beliefs, with
people self-identifying as right-wing being much less likely to hold these beliefs,
while those identifying as left-wing are much more likely to do so. This ef-
fect is also seen to some degree for “classic” conspiracies, where the negative
correlation of right-wing views is sustained although the positive correlation of
left-wing views is much less pronounced.
Anomia emerges as by far the strongest correlate of both forms of conspiracy
theory, with the anomia factor being significant at the P < 0.001 level in both
analyses, with a coefficient of 0.348 for classic conspiracies and 0.387 for anti-
government conspiracies — notably higher than even the effect of university
education on the conspiracy factors. Finally, the life satisfaction factors reveal
a potentially important difference between the two types of conspiracy belief
(which otherwise seem to occur in largely the same kinds of individual). While
the results suggest some relationship between dissatisfaction with one’s financial
status and classic conspiracy beliefs, this disappears once anomia is added to the
model7; however, the relationship is notably stronger for those who hold “anti-
government” conspiracy beliefs and persists even after the addition of anomia.
On the other hand, those who hold “classic” conspiracies show some degree of
dissatisfaction with both friendships and family relationships, which supports
the hypothesis that their conspiracy belief may be partially predicated on social
isolation and the sense of community offered by conspiracy groups.
7As the life satisfaction items score higher as respondents become more satisfied, the cor-
relation coeffecients here are negative when dissatisfaction is related to conspiracy belief.
14
Table 5: Regression Analysis of “Classic” Conspiracy Belief
Model 1 Model 2 Mo del 3 Model 4 Mo del 5
Female 0.066∗0.061∗0.009 0.099∗∗ 0.044
(0.033) (0.049) (0.772) (0.002) (0.152)
UniversityGrad −0.194∗∗∗
−0.198∗∗∗
−0.166∗∗∗
−0.173∗∗∗
−0.158∗∗∗
(<0.001) (<0.001) (<0.001) (<0.001) (<0.001)
Age: Under 35 0.236∗∗∗ 0.250∗∗∗ 0.282∗∗∗ 0.202∗∗∗ 0.273∗∗∗
(<0.001) (<0.001) (<0.001) (<0.001) (<0.001)
Age: 35-54 0.140∗∗∗ 0.144∗∗∗ 0.159∗∗∗ 0.093∗0.132∗∗∗
(<0.001) (<0.001) (<0.001) (0.019) (<0.001)
Age: 55-69 0.000
(base)
Age: 70 Plus −0.148∗∗
−0.157∗∗
−0.196∗∗∗
−0.093 −0.145∗∗
(0.002) (0.001) (<0.001) (0.064) (0.002)
PolRight1
−0.154∗∗∗
(<0.001)
PolLeft10.084∗
(0.043)
Anomia20.349∗∗∗ 0.340∗∗∗
(<0.001) (<0.001)
Life Satisfaction:3
Friends −0.045∗
−0.044∗
(0.031) (0.026)
Family −0.062∗∗
−0.065∗∗∗
(0.001) (<0.001)
Job −0.010 0.013
(0.597) (0.483)
Finances −0.047∗∗
−0.005
(0.005) (0.761)
Intercept −0.006 0.028 0.003 −0.568∗∗∗ 0.384∗∗∗
(0.877) (0.483) (0.939) (<0.001) (<0.001)
N 3226 3226 3226 3226 3226
r20.034 0.044 0.145 0.053 0.154
*** p < 0.001, ** p < 0.01, * p < 0.05. P-values in parentheses.
1PolLeft and PolRight are dummy variables for values below 4 and above 6 respectively in
the political position self-identification question.
2Anomia is a factor made up of three items - see Table 4.
3Life Satisfaction items are respondents’ self-reported level of satisfaction (higher values =
more satisfied).
15
Table 6: Regression Analysis of “Anti-Government” Conspiracy Belief
Model 1 Model 2 Mo del 3 Model 4 Mo del 5
Female 0.071∗0.070∗0.008 0.104∗∗ 0.045
(0.022) (0.022) (0.780) (0.001) (0.129)
UniversityGrad −0.168∗∗∗
−0.175∗∗∗
−0.136∗∗∗
−0.131∗∗∗
−0.114∗∗∗
(<0.001) (<0.001) (<0.001) (<0.001) (<0.001)
Age: Under 35 0.192∗∗∗ 0.206∗∗∗ 0.246∗∗∗ 0.151∗∗ 0.228∗∗∗
(<0.001) (<0.001) (<0.001) (0.001) (<0.001)
Age: 35-54 0.108∗∗ 0.112∗∗ 0.131∗∗∗ 0.062 0.104∗∗
(0.004) (0.003) (<0.001) (0.111) (0.004)
Age: 55-69 0.000
(base)
Age: 70 Plus −0.126∗∗
−0.138∗∗
−0.179∗∗∗
−0.070 −0.125∗∗
(0.007) (0.003) (<0.001) (0.156) (0.006)
PolRight1
−0.137∗∗∗
(<0.001)
PolLeft10.166∗∗∗
(<0.001)
Anomia20.389∗∗∗ 0.366∗∗∗
(<0.001) (<0.001)
Life Satisfaction:3
Friends −0.024 −0.022
(0.255) (0.246)
Family −0.022 −0.026
(0.216) (0.121)
Job −0.033 −0.007
(0.090) (0.681)
Finances −0.081∗∗∗
−0.035∗
(<0.001) (0.022)
Intercept −0.005 0.010 0.002 0.489∗∗∗ 0.290∗∗∗
(0.903) (0.800) (0.945) (<0.001) (<0.001)
N 3226 3226 3226 3226 3226
r20.025 0.040 0.168 0.048 0.170
*** p < 0.001, ** p < 0.01, * p < 0.05. P-values in parentheses.
1PolLeft and PolRight are dummy variables for values below 4 and above 6 respectively in
the political position self-identification question.
2Anomia is a factor made up of three items - see Table 4.
3Life Satisfaction items are respondents’ self-reported level of satisfaction (higher values =
more satisfied).
16
4 Effects of Conspiracy Beliefs on Political Pref-
erences
The final stage of our analysis model tests the effects of holding conspiracy beliefs
(again divided into two factors, classic and anti-government ) on individuals’
political preferences. For this analysis, we focused on respondents’ sentiment
towards parties (measured in the survey on a scale from 0, representing strong
dislike, to 100, representing strong favour) rather than on their voting behaviour
in the House of Councillors election. This was done partially in order to exclude
the effects of low turnout from the results (although examining the effect of
conspiracy belief on political participation, as distinct from preferences, is an
interesting avenue for future research), and also in order to more effectively
see the effects as they relate to the new populist parties, Reiwa Shinsengumi
(“Reiwa”) and The Party to Protect the People from the NHK (“N-Koku”),
whose low vote share could obscure any notable relationship between support for
them and conspiracy belief. By using sentiment rather than vote choice as our
dependent variable, we can include in the analysis individuals who voted for the
large parties (potentially for tactical reasons) but who feel warmly towards the
populist challenger parties nonetheless. Along with Reiwa and N-Koku, we also
analysed sentiment towards the ruling, centre-right Liberal Democratic Party
(LDP) and the main opposition party, the centre-left Constitutional Democratic
Party of Japan (CDPJ). As with all other analyses presented in this draft paper,
these regression analyses were conducted on unweighted data, and the results
are subject to change slightly once weighting to account for data loss through
non-response and satisficing is carried out.
The results reveal that conspiracy beliefs are particularly linked to support
for the right-wing parties — not just the populist challenger party N-Koku, but
17
also the mainstream LDP. However, while belief in “classic” conspiracy theories
(which generally posit global, extra-governmental conspiracy actors) is posi-
tively related to support for both of these parties, belief in “anti-government”
conspiracies is very strongly negatively correlated with LDP support, which
is of course unsurprising given the LDP’s long-term status as a ruling party.
Notably, support for both of these parties is also slightly negatively correlated
with anomia, even though anomia itself was positively correlated with conspir-
acy belief, suggesting a complex interaction between these factors and political
preferences8. Among the left-wing parties, meanwhile, “anti-government” con-
spiracies are slightly correlated with CDPJ support at the P < 0.05 level, while
belief in both “classic” and “anti-government” conspiracies is slightly correlated
with Reiwa support, though only at the very limits (P=ca.0.05) of the sig-
nificance tests. No correlation between anomia and support for either of the
left-wing parties was found.
5 Conclusions
Though it remains in its early stages, this study has shown that there is a signif-
icant incidence of belief in a variety of conspiracy theories within the Japanese
public. Japan’s hypothesised low levels of the anti-science sentiment which
underpins conspiracy theories such as climate change denial, the anti-vaccine
movement or various types of disbelief in the facts around COVID-19 in western
countries does not, it seems, extend to inoculating the public against conspiracy
theories more broadly, with even more outlandish theories such as the existence
of a cover-up of contact with extraterrestrial life being given some degree of
credence by as many as one in seven respondents. We found that conspiracy
8Table 10 and 11 in the Appendix show the models without conspiracy factors and political
leaning.
18
Table 7: Effect of Conspiracy Belief on Party Sentiment
LDP CDPJ Reiwa N-Koku
Female −4.030∗∗∗ 5.733∗∗∗ 3.520∗∗
−5.544∗∗∗
(<0.001) (<0.001) (0.001) (<0.001)
UniversityGrad −2.235∗2.628∗∗ 1.785 −3.946∗∗∗
(0.021) (0.007) (0.091) (<0.001)
Age: Under 35 4.686∗∗
−3.212∗2.294 11.167∗∗∗
(0.001) (0.021) (0.128) (<0.001)
Age: 35-54 2.065 −2.898∗1.049 2.445
(0.074) (0.011) (0.404) (0.060)
Age: 55-69 0.000
(base)
Age: 70 Plus 0.443 4.853∗∗ 0.815 0.022
(0.761) (0.001) (0.611) (0.989)
PolRight117.417∗∗∗
−13.129∗∗∗
−11.883∗∗∗
−2.816∗
(<0.001) (<0.001) (<0.001) (0.017)
PolLeft1
−15.748∗∗∗ 8.152∗∗∗ 9.919∗∗∗
−0.003
(<0.001) (<0.001) (<0.001) (0.999)
Anomia2
−2.971∗∗∗ 0.619 0.096 −1.837∗∗
(<0.001) (0.298) (0.882) (0.006)
Conspiracy Beliefs:
Classic 6.596∗∗∗
−1.183 2.594∗4.806∗∗∗
(<0.001) (0.317) (0.044) (<0.001)
Anti-Government −10.310∗∗∗ 2.963∗2.620 0.995
(<0.001) (0.016) (0.050) (0.471)
Intercept 49.210∗∗∗ 41.965∗∗∗ 35.854∗∗∗ 33.879∗∗∗
(<0.001) (<0.001) (<0.001) (<0.001)
N 2766 2766 2766 2766
r20.252 0.137 0.124 0.069
*** p < 0.001, ** p < 0.01, * p < 0.05. P-values in parentheses.
1PolLeft and PolRight are dummy variables for values below 4 and above 6 respectively in
the political position self-identification question.
2Anomia is a factor made up of three items - see Table 4.
19
theories are broadly divided into two factors, “classic” and “anti-government”
conspiracies, with the dividing line seeming to be the type of actor — extra-
governmental or governmental — that the individual views as being behind the
conspiracy, rather than the specific content of the conspiracy itself.
In analysing the correlates and possible reasons for individuals holding con-
spiracy beliefs, we found that both types of belief are more commonly held by
younger and less educated people, with feelings of anomia being very strongly
correlated with both factors. In general, respondents self-reporting their polit-
ical views as right-wing were less likely to hold conspiracy beliefs, though we
note that the direction of causality in this case is unclear, as it is possible the
conspiracy belief pushes people away from right-wing attitudes (in the specific
Japanese sense of “right-wing”, which is generally seen as a pro-government
stance) rather than vice versa. Perhaps the most significant difference we found
between the two varieties of conspiracy belief is that while both forms of be-
lief are correlated with dissatisfaction with one’s financial status, only belief in
“anti-government” conspiracies is associated with dissatisfaction with one’s ca-
reer, while only belief in “classic” conspiracies is associated with dissatisfaction
with relationships with friends and family. We speculate that this last point
could indicate that belief in “classic” conspiracies fits with the social aspect of
conspiracy belief described by Prooijen and Douglas (2018), in that social isola-
tion may push individuals to embrace the online communities that have sprung
up around conspiracy beliefs.
Finally, our analysis of conspiracy factors as an independent variable influ-
encing individuals’ political preferences suggested that conspiracy believers tend
to support both the mainstream and niche right-wing parties (the LDP and N-
Koku) respectively, although anti-government conspiracy belief specifically was
strongly negatively correlated with support for the ruling LDP. Perhaps sur-
20
prisingly, conspiracy belief was only very weakly correlated with support for
the mainstream and niche left-wing parties (CDPJ and Reiwa), and only at the
lowest level of significance (P < 0.05).
Within these findings regarding political support, it is worth focusing briefly
on the specific nature of support for the N-Koku populist challenger party that
is revealed in this analysis. Respondents with positive views of N-Koku stand
out in a number of ways; they are markedly younger, more likely to be male and
less likely to be a university graduate than those with positive views of other
parties. More uniquely, they do not identify strongly with either the left or right
of the political spectrum (despite N-Koku’s platform including a number of far-
right elements), but do tend to believe in classic conspiracy theories. Notably,
this set of factors resembles the make-up of online communities that went on to
form part of the nativist “alt-right” in the U.S. and elsewhere — young, male,
less educated, often viewing themselves as politically agnostic, and tending to
distrust established authorities and experts. While the alt-right that ultimately
emerged as a political force incorporated several other groups — perhaps most
notably the long-standing white nationalist and white supremacist groups whose
influence eventually supplanted the more anarchic online forums on which the
alt-right’s younger supporters had started out (Hodge and Hallgrimsdottir 2020)
— the groups who engaged in the vitriolic online culture wars documented by
Nagle (2017) remain significant within the demographics of the modern alt-right
(Forscher and Kteily 2019). Such individuals and groups have almost certainly
existed for just as long if not longer in Japan as they have in the U.S. and
elsewhere, so there is little support for the conclusion that Japan is following
the same path a few steps behind in this regard, but the correlates of favourable
political views of N-Koku nonetheless suggest that conspiracy theories are indeed
playing a role in the political identity formation of this younger demographic.
21
6 Future Research Plans
Aside from a number of elements noted elsewhere in the text (such as the cur-
rent lack of sample weighting in our analyses), the current study has a number
of limitations and drawbacks which can be addressed with future work. Perhaps
the most obvious is that at present there is no cohesive model for the role of
conspiracy belief factors; while the regression analyses suggest a mechanism that
flows from individuals’ personal circumstances and sentiments through conspir-
acy belief and eventually on to political preferences, we intend to eventually
integrate these steps into a structural equation model rather than considering
them independently.
The research has uncovered a number of aspects which we consider to merit
further inquiry and future survey research. The politically polarised items
(CONS6, CONS7, CONS9 and CONS10) which we were forced to exclude from
the factors, for example, could be expanded upon in an improved survey bat-
tery to give us a more comprehensive basis on which to separate out political
beliefs from underlying propensity towards belief in conspiracies. The role of
strong political polarisation in the formation or expression of conspiracy beliefs
has not been extensively studied either in Japan or abroad, and is deserving of
more in-depth consideration, especially given the current importance of totally
politically polarised conspiracy movements such as QAnon, or the degree of
political polarisation which appears to exist in the conspiracy theories around
COVID-19.
Another future avenue for this research lies in building on the work by Cas-
tanho Silva et al. (2017) by exploring the connection between conspiracy beliefs
and populist sentiment, measured independently from party support. This is
especially relevant in Japan given the “broad tent” nature of the LDP, as we
would expect to find significant variation among supporters of that party (and
22
others) which is better captured by populism survey batteries that focus on the
constituent factors of populism rather than on party politics.
23
References
Aaronovitch, David (2010). Voodoo histories: the role of the conspiracy theory
in shaping modern history. 1st American ed. New York: Riverhead Books.
Brotherton, Robert, Christopher C. French, and Alan D. Pickering (2013).
“Measuring Belief in Conspiracy Theories: The Generic Conspiracist Beliefs
Scale”. English. In: Frontiers in Psychology 4.
Castanho Silva, Bruno, Federico Vegetti, and Levente Littvay (2017). “The Elite
Is Up to Something: Exploring the Relation Between Populism and Belief in
Conspiracy Theories”. en. In: Swiss Political Science Review 23.4, pp. 423–
443.
Clarke, Steve (June 2002). “Conspiracy Theories and Conspiracy Theorizing”.
en. In: Philosophy of the Social Sciences 32.2. Publisher: SAGE Publications
Inc, pp. 131–150.
Davis, David Brion (1971). The fear of conspiracy: images of un-American sub-
version from the Revolution to the present. Ithaca [N.Y.]: Cornell University
Press.
Douglas, Karen M., Robbie M. Sutton, and Aleksandra Cichocka (Dec. 2017).
“The Psychology of Conspiracy Theories:” en. In: Current Directions in Psy-
chological Science. Publisher: SAGE PublicationsSage CA: Los Angeles, CA.
Douglas, Karen M., Joseph E. Uscinski, Robbie M. Sutton, Aleksandra Ci-
chocka, Turkay Nefes, Chee Siang Ang, and Farzin Deravi (2019). “Under-
standing Conspiracy Theories”. en. In: Political Psychology 40.S1. eprint:
https://onlinelibrary.wiley.com/doi/pdf/10.1111/pops.12568, pp. 3–35.
Forscher, Patrick S. and Nour S. Kteily (Nov. 2019). “A Psychological Profile
of the Alt-Right:” en. In: Perspectives on Psychological Science. Publisher:
SAGE PublicationsSage CA: Los Angeles, CA.
24
Hawkins, Kirk A. (Aug. 2009). “Is Ch´avez Populist?: Measuring Populist Dis-
course in Comparative Perspective”. en. In: Comparative Political Studies
42.8, pp. 1040–1067.
Hodge, Edwin and Helga Hallgrimsdottir (Aug. 2020). “Networks of Hate: The
Alt-right, “Troll Culture”, and the Cultural Geography of Social Movement
Spaces Online”. In: Journal of Borderlands Studies 35.4. Publisher: Rout-
ledge eprint: https://doi.org/10.1080/08865655.2019.1571935, pp. 563–580.
Kagoya, Kazuhiro (1998). “The Unintended Consequences of the Refutation
against Rumor”. In: Japanese Sociological Review 49.4, pp. 584–599.
Klein, Colin, Peter Clutton, and Vince Polito (2018). “Topic Modeling Reveals
Distinct Interests within an Online Conspiracy Forum”. English. In: Fron-
tiers in Psychology 9. Publisher: Frontiers.
Majima, Yoshimasa and Hiroko Nakamura (Dec. 2019). “Development of the
Japanese Version of the Generic Conspiracist Beliefs Scale (GCBS-J)”. In:
Japanese Psychological Research.
Mudde, Cas (2004). “The Populist Zeitgeist”. en. In: Government and Opposi-
tion 39.4, pp. 541–563.
Nagle, Angela (June 2017). Kill All Normies: Online Culture Wars From 4Chan
And Tumblr To Trump And The Alt-Right. en. John Hunt Publishing.
Oliver, J. Eric and Thomas J. Wood (2014). “Conspiracy Theories and the
Paranoid Style(s) of Mass Opinion”. en. In: American Journal of Political
Science 58.4, pp. 952–966.
Prooijen, Jan-Willem van and Karen M. Douglas (June 29, 2017). “Conspiracy
theories as part of history: The role of societal crisis situations:” in: Memory
Studies.
25
Prooijen, Jan-Willem van and Karen M. Douglas (2018). “Belief in conspiracy
theories: Basic principles of an emerging research domain”. In: European
Journal of Social Psychology 48.7, pp. 897–908.
Radnitz, Scott and Patrick Underwood (2017). “Is belief in conspiracy theo-
ries pathological? A survey experiment on the cognitive roots of extreme
suspicion”. In: British Journal of Political Science 47.1, p. 113.
Schatto-Eckrodt, Tim, Svenja Boberg, Florian Wintterlin, and Thorsten Quandt
(2020). “Use and Assessment of Sources in Conspiracy Theorists’ Communi-
ties”. In: Disinformation in Open Online Media. Ed. by Christian Grimme,
Mike Preuss, Frank W. Takes, and Annie Waldherr. Lecture Notes in Com-
puter Science. Cham: Springer International Publishing, pp. 25–32.
Srole, Leo (1956). “Social Integration and Certain Corollaries: An Exploratory
Study”. In: American Sociological Review 21.6, pp. 709–716.
Stempel, Carl, Thomas Hargrove, and Guido H. Stempel (June 2007). “Media
Use, Social Structure, and Belief in 9/11 Conspiracy Theories”. In: Journal-
ism & Mass Communication Quarterly 84.2. Publisher: SAGE Publications
Inc, pp. 353–372.
Swami, Viren, Tomas Chamorro-Premuzic, and Adrian Furnham (2010). “Unan-
swered questions: A preliminary investigation of personality and individual
difference predictors of 9/11 conspiracist beliefs”. en. In: Applied Cognitive
Psychology 24.6, pp. 749–761.
Teymoori, Ali, Jolanda Jetten, Brock Bastian, Amarina Ariyanto, Fr´ed´erique
Autin, Nadia Ayub, Constantina Badea, Tomasz Besta, Fabrizio Butera,
Rui Costa-Lopes, Lijuan Cui, Carole Fantini, Gillian Finchilescu, Lowell
Gaertner, Mario Gollwitzer, ´
Angel G´omez, Roberto Gonz´alez, Ying Yi Hong,
Dorthe Høj Jensen, Minoru Karasawa, Thomas Kessler, Olivier Klein, Mar-
cus Lima, Tuuli Anna M¨ah¨onen, Laura Megevand, Thomas Morton, Paola
26
Paladino, Tibor Polya, Aleksejs Ruza, Wan Shahrazad, Sushama Sharma,
Ana Raquel Torres, Anne Marthe van der Bles, and Michael Wohl (July
2016). “Revisiting the Measurement of Anomie”. en. In: PLOS ONE 11.7,
e0158370.
Wood, Michael J., Karen M. Douglas, and Robbie M. Sutton (Nov. 2012). “Dead
and Alive: Beliefs in Contradictory Conspiracy Theories”. en. In: Social Psy-
chological and Personality Science 3.6. Publisher: SAGE Publications Inc,
pp. 767–773.
27
Appendices
Appendix I: Alternative Regression Models
28
Table 8: Alternative Regression Analysis of “Classic” Conspiracy Belief
Model 1 Mo del 2 Model 3 Model 4
Female 0.066∗0.061∗0.005 0.094∗∗
(0.033) (0.049) (0.876) (0.004)
UniversityGrad −0.194∗∗∗
−0.198∗∗∗
−0.169∗∗∗
−0.177∗∗∗
(<0.001) (<0.001) (<0.001) (<0.001)
Age: Under 35 0.236∗∗∗ 0.250∗∗∗ 0.295∗∗∗ 0.216∗∗∗
(<0.001) (<0.001) (<0.001) (<0.001)
Age: 35-54 0.140∗∗∗ 0.144∗∗∗ 0.164∗∗∗ 0.097∗
(<0.001) (<0.001) (<0.001) (0.014)
Age: 55-69 0.000
(base)
Age: 70 Plus −0.148∗∗
−0.157∗∗
−0.203∗∗∗
−0.105∗
(0.002) (0.001) (<0.001) (0.036)
PolRight1
−0.154∗∗∗
−0.142∗∗∗
−0.130∗∗∗
(<0.001) (<0.001) (<0.001)
PolLeft10.084∗0.082∗0.103∗
(0.043) (0.037) (0.016)
Anomia20.348∗∗∗
(<0.001)
Life Satisfaction:3
Friends −0.047∗
(0.025)
Family −0.059∗∗
(0.001)
Job −0.015
(0.440)
Finances −0.043∗∗
(0.009)
Intercept −0.006 0.028 0.027 −0.590∗∗∗
(0.877) (0.483) (0.471) (<0.001)
N 3226 3226 3226 3226
r20.034 0.044 0.154 0.062
*** p < 0.001, ** p < 0.01, * p < 0.05. P-values in parentheses.
1PolLeft and PolRight are dummy variables for values below 4 and above 6 respectively in
the political position self-identification question.
2Anomia is a factor made up of three items - see Table 4.
3Life Satisfaction items are respondents’ self-reported level of satisfaction (higher values =
more satisfied).
29
Table 9: Alternative Regression Analysis of “Anti-Government” Conspiracy
Belief
Model 1 Mo del 2 Model 3 Model 4
Female 0.071∗0.070∗0.009 0.103∗∗
(0.022) (0.022) (0.299) (0.001)
UniversityGrad −0.168∗∗∗
−0.175∗∗∗
−0.141∗∗∗
−0.139∗∗∗
(<0.001) (<0.001) (<0.001) (<0.001)
Age: Under 35 0.192∗∗∗ 0.206∗∗∗ 0.259∗∗∗ 0.167∗∗∗
(<0.001) (<0.001) (<0.001) (<0.001)
Age: 35-54 0.108∗∗ 0.112∗∗ 0.135∗∗∗ 0.066
(0.004) (0.003) (<0.001) (0.086)
Age: 55-69 0.000
(base)
Age: 70 Plus −0.126∗∗
−0.138∗∗
−0.191∗∗∗
−0.085
(0.007) (0.003) (<0.001) (0.081)
PolRight1
−0.137∗∗∗
−0.124∗∗∗
−0.123∗∗∗
(<0.001) (<0.001) (<0.001)
PolLeft10.166∗∗∗ 0.163∗∗∗ 0.174∗∗∗
(<0.001) (<0.001) (<0.001)
Anomia20.387∗∗∗
(<0.001)
Life Satisfaction:3
Friends −0.027
(0.193)
Family −0.020
(0.271)
Job −0.038∗
(0.047)
Finances −0.076∗∗∗
(<0.001)
Intercept −0.005 0.010 0.008 −0.503∗∗∗
(0.903) (0.800) (0.820) (<0.001)
N 3226 3226 3226 3226
r20.025 0.040 0.181 0.061
*** p < 0.001, ** p < 0.01, * p < 0.05. P-values in parentheses.
1PolLeft and PolRight are dummy variables for values below 4 and above 6 respectively in
the political position self-identification question.
2Anomia is a factor made up of three items - see Table 4.
3Life Satisfaction items are respondents’ self-reported level of satisfaction (higher values =
more satisfied).
30
Table 10: Party Sentiment Regression (without conspiracy factors)
LDP CDPJ Reiwa N-Koku
Female −4.142∗∗∗ 5.766∗∗∗ 3.638∗∗
−5.487∗∗∗
(<0.001) (<0.001) (0.001) (<0.001)
UniversityGrad −1.921 2.405∗0.967 −4.941∗∗∗
(0.050) (0.013) (0.363) (<0.001)
Age: Under 35 3.922∗∗
−2.789∗3.687∗12.919∗∗∗
(0.005) (0.044) (0.015) (<0.001)
Age: 35-54 1.792 −2.735∗1.758 3.301∗
(0.125) (0.017) (0.166) (0.012)
Age: 55-69 0.000
(base)
Age: 70 Plus 0.918 4.578∗∗
−0.082 −1.000
(0.533) (0.002) (0.960) (0.553)
PolRight117.865∗∗∗
−13.348∗∗∗
−12.485∗∗∗
−3.494∗∗
(<0.001) (<0.001) (<0.001) (0.004)
PolLeft1
−16.887∗∗∗ 8.556∗∗∗ 10.725∗∗∗
−0.759
(<0.001) (<0.001) (<0.001) (0.596)
Anomia2
−4.606∗∗∗ 1.342∗1.984∗∗ 0.196
(<0.001) (0.015) (0.001) (0.754)
Intercept 49.319∗∗∗ 41.960∗∗∗ 35.874∗∗∗ 33.942∗∗∗
(<0.001) (<0.001) (<0.001) (<0.001)
N 2766 2766 2766 2766
r20.229 0.134 0.102 0.042
*** p < 0.001, ** p < 0.01, * p < 0.05. P-values in parentheses.
1PolLeft and PolRight are dummy variables for values below 4 and above 6 respectively in
the political position self-identification question.
2Anomia is a factor made up of three items - see Table 4.
31
Table 11: Party Sentiment Regressions (without political leaning)
LDP CDPJ Reiwa N-Koku
Female −4.727∗∗∗ 6.376∗∗∗ 3.952∗∗∗
−5.349∗∗∗
(<0.001) (<0.001) (<0.001) (<0.001)
UniversityGrad −3.361∗∗ 3.268∗∗ 2.352∗
−3.942∗∗∗
(0.002) (0.001) (0.032) (<0.001)
Age: Under 35 6.863∗∗∗
−4.768∗∗ 0.959 10.949∗∗∗
(<0.001) (0.001) (0.541) (<0.001)
Age: 35-54 2.774∗
−3.299∗∗ 0.656 2.370
(0.030) (0.006) (0.617) (0.069)
Age: 55-69 0.000
(base)
Age: 70 Plus −1.564 6.181∗∗∗ 2.225 0.203
(0.331) (<0.001) (0.182) (0.903)
Anomia2
−2.407∗∗∗ 0.311 −0.198 −1.872∗∗
(<0.001) (0.619) (0.770) (0.005)
Conspiracy Beliefs:
Classic 7.599∗∗∗
−1.665 1.862 4.795∗∗∗
(<0.001) (0.180) (0.165) (<0.001)
Anti-Government −13.100∗∗∗ 4.590∗∗∗ 2.620 1.169
(<0.001) (<0.001) (0.050) (0.395)
Intercept 52.353∗∗∗ 38.916∗∗∗ 33.580∗∗∗ 32.945∗∗∗
(<0.001) (<0.001) (<0.001) (<0.001)
N 2766 2766 2766 2766
r20.080 0.046 0.045 0.067
*** p < 0.001, ** p < 0.01, * p < 0.05. P-values in parentheses.
1PolLeft and PolRight are dummy variables for values below 4 and above 6 respectively in
the political position self-identification question.
2Anomia is a factor made up of three items - see Table 4.
32