Vaccination talks on Twitter. Semantic social networks and public
views from Greece
DIMITRIOS KYDROS
Economic Sciences
International Hellenic University
Terma Magnisias, 62124, Serres
GREECE
VASILIKI VRANA
Business Administration
International Hellenic University
Terma Magnisias, 62124, Serres
GREECE
Abstract: - Social media are increasingly used as a source of health information. Opinions expressed on social
media, including Twitter, may contribute to opinion formation and impact positively or negatively the
vaccination decision-making process. The paper creates networks of Greek users that talk about vaccination on
Twitter, during the last quarter of 2021 and analyzes their structure and grouping. Furthermore, some content
analysis is also produced by creating networks of words found within tweets. The main purpose is to locate and
present the Greek public views on COVID-19 vaccination. Results show that the network of Greek users may
be considered as fragmented but by all means not polarized between two different opinions. Anti-vaccination
ideas were clearly present during the first period of our study but were rapidly diminished in the following
months, maybe due to a large number of deaths and the advent of the Omicron strain. The persisting large
percentage of the population refusing to vaccinate may be expressed in other social media platforms.
Key-Words: - COVID 19, Twitter, public health, vaccination, anti-vaccination, semantic social networks
Received: March 20, 2021. Revised: January 23, 2022. Accepted: March 15, 2022. Published: April 13, 2022.
1 Introduction
The outbreak of coronavirus (COVID-19) has
caused more than 5 million deaths and posed
significant threat to people existence [1]. History
has shown that vaccines have played critical roles in
reducing mortality rates in cases of major infectious
diseases [2]. As of today, vaccines are essential to
accelerate herd immunity, reduce the number of
active cases, limit the fatality rate, enable social
measures of restricting and disease’s spread to relax,
and socioeconomic activity to resume [3, 4].
In today’s international health crisis, more than
130 vaccines are in clinical development [3] and
194 in pre-clinical development [4], while 18
vaccines against COVID-19 like Pfizer-BioNTech,
Oxford-AstraZeneca, Moderna, Johnson &
Johnson’s Janssen have been approved by at least
one country [6] and industrialized for use in a
relatively short period of time compared to other
vaccines developed in the past [7]. The rapid
development of vaccines has raised concerns about
vaccines’ safety and the probability of side effects,
and is one of the primary reasons for vaccine
hesitancy [6, 8, 9]. Troiano & Nardi[10] in a review
study identified the reasons why people refuse
vaccination against COVID-19. Reasons include
general attitude being against vaccines or
considering the vaccine useless, general lack of
trust, mistrust of health authorities, concerns about
safety as the vaccines were developed in a short
time and thus are supposed to be too dangerous,
doubts about the efficiency and the provenience of
the vaccines etc. Due to these reasons a sizable
proportion of people around the globe exhibit
reluctance to getting vaccinated [10. 11, 12, 13].
Vaccination uptake relies on a person’s weigh of the
risks vs benefits perceptions and may be
significantly influenced by misinformation [8, 14].
A growing number of people use the web and
social media to obtain health information, including
information about vaccines [15]. Accurate, reliable,
and up-to-date information are disseminated by
official websites of public health organizations that
are also increasingly invest to wisdom of the crowd
WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS
DOI: 10.37394/23209.2022.19.5
Dimitrios Kydros, Vasiliki Vrana
E-ISSN: 2224-3402
44
Volume 19, 2022
[17]. Stahl et al. [17] identified the major role of
social media in disseminating information about
vaccination. They claim that social media modify
the doctor/patient relationship and impact
vaccination decision-making process and vaccines
acceptance. The network structure, who delivers the
message on social media and how the message is
framed, are affecting the vaccine decision-making
process [18]. Issues against vaccination are being
discussed, but also social media offer an avenue to
fight against vaccine hesitancy. Love et al. [19]
performed a content analysis of posts about
vaccinations on Twitter and found that 33% of the
tweets were positive regarding vaccines, 54% were
neutral, and 13% were negative. Negative attitudes
claim alleged dangers, neutrals share immunization
experiences and the positive ones comment on
effectiveness and promote vaccines. Substantial
misinformation is also widely available through
online organized anti-vaccination groups [20, 21].
Government, medical, pharmaceutical conspiracy
theories and morality, civil liberties, themes of
effectiveness and safety of the vaccines, illnesses
that the vaccines cause, alternative medicine and
corruption of the mainstream medicine are the most
common arguments around anti-vaccination [20].
Previous studies have found that information
provided on anti-vaccine websites exert influence on
peoples’ decisions to vaccinate themselves or their
children [21, 22, 23].
The mass uptake of social media has
significantly contributed the ‘anti-vax’ COVID 19
movement [21]. The Center for Countering Digital
Hate [25] recorded 400 anti-vax social media
accounts with 58 million of followers. Since 2019,
the accounts have increased their followers by 7.8
million people. The anti-vaxers use social media to
publish false information and discourage people
from up-taking vaccines[26]. Social media
platforms may develop acts against the anti-vaccine
movement. Twitter and YouTube announced that
they would label anti-vaccination content [25] and
YouTube removed advertisements from anti-vaccine
videos and Twitter ensured that first results for
anyone searching for vaccine-related topics would
be the content created of the National Health
Service in the UK or the Department of Health and
Human Services in the USA [27]. Facebook
announced that it would down-rank or hide anti-
vaccination content [25] and offered free advertising
space to WHO and national health authorities [27].
Opinions regarding what should be done are
contradictory. De-platforming individuals or
shutting down social spaces is claimed to be the
only effective tool [25]. However, “this is an issue
of freedom of speech” according to Professor
Viswanath [27] and the scientific community agrees
that the individual’s right to determine to uptake a
vaccine and should be preserved [28].
“The world shares a collective responsibility in
fighting this pandemic; therefore, continued
research on COVID-19 vaccine acceptance and
hesitancy should be a priority” claimed
Machingaidze & Wiysonge [6]. Learning about
vaccination and anti-vaccination content on social
media is of great importance for health organization
and advocates in order to establish communication
and education strategies to resolve main doubts and
effectively react, respond and develop anti-vaccine
arguments. Research on vaccination and anti-
vaccination COVID 19 content on social media is in
its infancy and to our best knowledge research on
semantic social network analysis regarding
discussion on Twitter doesn’t exist. “Twitter can
provide a great opportunity to understand public
opinion regarding COVID-19 vaccines” mentioned
Karami et al. [29].
2 Vaccination and Anti-Vaccination
Against COVID 19 on Social Media
One of the first studies investigating perception of
social media users regarding COVID-19 vaccine
was that of Adebisi et al. [30]. They conducted a
cross-sectional survey among social media users in
Nigeria asking whether users will take the vaccine
when it will be available. According to their
findings three out of four of the responders intended
to take the COVID-19 vaccine. The major reason
for non-acceptance was unreliability of the clinical
trials, followed by the belief that their immune
system is sufficient to combat the virus. Eguia et
al.[11] also recruited Twitter users to answer an
online questionnaire about users’ intention to be
vaccinated and the main reason for their answers. A
percentage of 22.43% stated that they would not be
vaccinated. Lack of effectiveness, of the
vaccination, and possibly dangerous adverse effects
were the main reasons provided.
In favor of vaccination against COVID-19, the
ministry of Health of the Government of Spain
started a campaign in Twitter using the hashtag
#yomevacuno (igetvaccinated). Herrera-Peco et al.
[31] analysed the role of healthcare professionals
during the start of the campaign. Dissemination of
information within the #yomevacuno was found
scarce among healthcare professionals. They were
not sharing information about vaccines or
vaccination. However, the majority of them had a
WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS
DOI: 10.37394/23209.2022.19.5
Dimitrios Kydros, Vasiliki Vrana
E-ISSN: 2224-3402
45
Volume 19, 2022
favourable storytelling on the vaccine. Piltch-Loeb
et al. [32] tried to investigate the use of different
media channels for COVID-19 vaccine information
and their relationship to vaccine acceptance.
Increased vaccine acceptance was found to be
related with acquisition of information from
traditional channels like TV and newspapers while
those who are using social media or both traditional
and social media as their source of information are
less likely to get the vaccine. The findings suggest
the significant role of social media platforms in
educating users to accept the vaccine.
Puri et al. [33] tried to identify the role of social
media in spreading anti-vaccination content and the
impact of the content on public health and vaccine
hesitancy. The global pandemic situation, domestic
vaccination policies, priority groups and challenges
from COVID-19 variants are the main topics
discussed on Twitter and Weibo, in U.S.A and
China. Twitter users’ use the platform to share their
individual vaccination experiences and express anti-
vaccine attitudes while Weibo users express more
positive feelings toward the COVID-19 vaccines
and manifest evident deference to authorities. Biden
administration's evolving control plans, and
vaccination efforts were also discussed on Twitter
[34]. Vaccine hesitancy on Twitter was investigated
by Thelwall, Kousha & Thelwall [35]. Vaccine
development speed, vaccine safety and conspiracies
were the main themes discussed. Political topics are
also discussed by the majority of vaccine refusers
who express right-wing opinions, fear of a deep
state and conspiracy theories. Vaccine refusers who
do not discuss in political contexts seem to reach a
larger audience outside right-wing areas of Twitter.
Vaccine opposition was found to be demonstrated
through vaccine hesitancy, direct opposition and
adverse reactions in the study of Criss et al. [36]
who described themes of tweets related to vaccines,
ethnicity and race. Political misinformation,
scientific misinformation, and race extermination
conspiracies were also recorded. In contrast, vaccine
support was demonstrated through vaccine
affirmation, a need for a vaccine, advocacy through
reproach, vaccine development and efficacy,
COVID-19 and racism, racist vaccine humor, and
news updates. Political motivations of the vaccine
opposition movement were also found by Bonnevie
et al. [37] who examined shifts in vaccine
opposition on Twitter. Conversation about federal
health authorities, research and clinical trials and
vaccine ingredients were the main themes of the
discussions. The study also revealed that vaccine
opposition on Twitter increased by 80% across time
periods.
Sentiments of tweets containing terms related to
the COVID-19 vaccine in the U.S.A were
investigated by Karami et al. [29]. Overtime they
found that non-negative and negative sentiments are
increasing and decreasing respectively, showing that
public sentiment become less negative during the
two months after starting the vaccination. Regarding
discussion topics, negative tweets include topics
about vaccine effectiveness and stories of getting
vaccines while non-negative tweets discuss vaccine
immunity, vaccination hesitancy, mask, and social
distancing. Liu & Liu [38] characterized behavioral
intentions toward vaccines on the Twitter. On the
one hand, positive intentions were affected by the
positive values of vaccination such as return to
normal life, socioeconomic recovery and reduced
risk of infection. On the other hand, negative
intentions were associated with lack of knowledge,
underestimation of disease severity and low vaccine
effectiveness, distrust of vaccines or government
and greater confidence and trust in the natural
immune system. Sentiments and attitudes of
Australian Twitter users were studied by Kwok et
al. [39]. According to their findings two-thirds of
the tweets expressed positive opinions about the
vaccine and one-third negative. Trust and
anticipation were the top positive sentiments, while
fear was the top negative recorded in the tweets.
They concluded that Twitter users in Australia
refuted misinformation and supported infection
control measures however the level of positive
sentiment was not sufficient to increase vaccination
coverage to accelerate herd immunity. A different
perceptive was adopted by Scannell et al. [40] who
examined the persuasion frameworks of tweets.
Celebrity figures are used as persuasion techniques
on both Anti-Vaccination and Pro-vaccinations
tweets. Anti-Vaccination tweets use
Humor/Sarcasm and Anecdotal stories, while Pro-
Vaccine tweets use also Information and
Participation.
Finally, Kydros, Argyropoulou & Vrana [41]
have investigated the Twitter discussion on COVID-
19 in the case of Greece, including data collected in
the second half of 2020. In this paper a full analysis,
including semantics and sentiment analysis was
held, however, in early 2021, vaccination became
possible, so these discussions were altogether
changed. The present paper can be seen as a follow-
up of the above mentioned study.
Having all the above in mind, this paper aims at
investigating the actual situation in the Greek
Twitter sphere, after vaccines were introduced. We
have collected a rather large data set of tweets over
a period of four months at the end of 2021 and
WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS
DOI: 10.37394/23209.2022.19.5
Dimitrios Kydros, Vasiliki Vrana
E-ISSN: 2224-3402
46
Volume 19, 2022
venture mainly to find out what are the actual
discussions about vaccination on Twitter in the
Greek environment. Social Network Analysis is
used to identify main characteristics of users’
networks, continued by semantic analysis on the
actual content of tweets by using networks of word-
pairs within tweets.
Overall we venture to answer the following
Research Questions (RQs):
RQ1: Is there any structural evidence that Tweeters
in Greece are fragmented or polarized regarding
COVID-19 vaccination?
RQ2: Are there different categories of users in these
conversations? What is their position?
RQ3: Is there a clear, prevailing public view? Are
there any objections? What are they about?
The paper is structured as follows: Background
on vaccination and anti-vaccination on social media
was presented in order to situate the contribution of
the paper. The methodology section presents the
collections of tweets’, network’s formation and
semantic analysis. Results are presented in the next
section. The paper concludes with our final remarks
and propositions for further research.
3 Methodology
In all subsequent procedures (data mining, cleaning
and filtering, metric computations and
visualizations), we used NodeXL Pro [42], an Excel
template quite suitable for such investigations. We
performed the Twitter importing procedures during
the last four months of 2021 (from September to
December), in order to capture possible shifts in
Greek users’ opinions. At this point it should be
noted that Twitter’s API (the protocol that facilitates
searching and importing) usually returns data over a
range 7 to 10 days prior to the date of importing.
Thus, we chose to repeat our importing procedures
in order to collect tweets that span the whole time
window. In all our searches we used a number of
keywords in Greek, such as εμβόλιο – vaccine,
αντιεμβολιαστές – antivacciners, εμβολιασμός –
vaccination etc. Due to a certain difficulty in
searching for Greek characters in Twitter, we first
transformed all our keywords in percent notation.
However, this exact property (Greek characters)
assures us that our data sets were produced to a
large extend only by Greek people.
A rather large data set was produced from the
above mentioned procedure, forcing some further
preprocessing and filtering, such as the complete
removal of retweets. Actually, according to
Kantriwitz [43], even the engineer that created the
retweet button and procedure considers this feature
not only as useless but even as dangerous, since it
only adds on to information noise rather than
producing useful content. Thus, in the following
discussions, only tweets with actual content are
identified and processed. Furthermore, in order to
proceed with readable data, we merged out results in
three different data sets, having in mind three
different time-periods within our four-month study.
These time periods, along with the volumes of our
data sets are shown in Table 1.
Table 1. Data sets and time-periods
Time
period
Number of
tweeters
(Nodes)
Number of
tweets
(Links)
September
to end of
October,
2021
4364
10971
November
to mid-
December,
2021
3915
8878
Mid to end
of
December
2021 (last
day of our
study was
the 28th of
December).
5877
13676
WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS
DOI: 10.37394/23209.2022.19.5
Dimitrios Kydros, Vasiliki Vrana
E-ISSN: 2224-3402
47
Volume 19, 2022
A network is comprised of nodes and links. In
our case, each tweet is produced by a tweeter, so
tweeters are the nodes. If a tweet is left “floating
alone” (without producing any reply or mentioning)
then a link is drawn connecting the tweeter to
himself (a self-loop) and the respecting tweeter is an
isolated node. More importantly, each reply or
mentioning actually connects two tweeters. Hence a
directed link is created connecting these two
tweeters, who are now joined through a
“conversation”. Links can be duplicated, if two
users “discuss” over different tweets. By creating all
those links, full networks that can be examined
through Social Network Analytic techniques are
formed.
4 Results and Discussion
4.1 Visualizations and Structural Results
The three networks discussed in the previous section
are pictorially shown in Figure1, while some
structural characteristics are shown in Table 2.
(a) Septem
ber
(b) Novem
ber
(c) December
Fig. 1: Visualizations of the produced networks.
Isolates are shown in the top left, while the largest
group is shown in the bottom left.
Table 2. Some structural characteristics of the three
networks
Networ
k
Uniq
ue
Link
s
Links
with
Duplica
tes
Isolat
es
Total
Grou
ps
Grou
ps
over
5
perso
ns
Nod
es in
large
st
Gro
up
Septem
ber
5027
5944
1325
303
64
306
Novem
ber
4634
4244
1146
254
49
314
Decem
ber
7313
6363
1604
331
59
423
In Figure 1 and Table 2, we use the notion of
community [44] to bring closely interacting nodes in
the same group. A community is a group where
more links are created between nodes of this group
than with other nodes. One rather unexpected result
is the rather large proportion (about 20% in all three
cases) of isolated nodes (nodes that correspond to
tweeters who did not produce any reaction by any
other tweeter). Actually, isolates are the largest
group in our networks. A number of isolates is
certainly expected, since in the social media world
not everyone has his/hers followers and triggers
discussions. However, it can be deduced that in the
Greek social media sphere, at least within our search
framework, a rather large proportion of users are left
“shouting and unanswered”. This result however
does not mean that all those tweets were left unread;
some of them must have been read and influenced
others, but did not create a discussion.
Another proportion of users are actively engaged
in discussions over different tweets, as seen by the
almost equal numbers of plain and duplicate links.
Of course, a “discussion” has a rather loose
definition with respect to the actual number of
persons involved. It is interesting to note that over
all groups in all cases, about 20% of groups are
formed with five or more users, while the largest
groups are comprised of about 7% of the total
number of users, despite the fact that intergroup
links are definitely present (Figure 1), denoting
intergroup discussions.
Hence it can be deduced that the overall
discussion on COVID-19 vaccination in the Greek
Twitter case seems to be rather fragmented but not
clearly polarized between pros and cons on
vaccination. The above discussion answers our
RQ1, since no actual polarization was found in
structural terms.
4.2 Opinions Within Groups
We now turn our attention on our RQ2, which deals
with actual users and their status. All calculations
were again performed through NodeXL Pro. In
Table 3 we present the status (user, media,
politician, political party, etc.) of the top-ten users,
together with their position on vaccination (pro/con,
if clearly implied) by visiting their relevant personal
pages in Twitter. We do not include actual names
but preferred to show aggregate results for a clearer
view.
Table 3. The top ten Tweeters and their positions
Top
Tweeters
Top
Replied-to
Top Mentioned
September
Users: 4 /
pro: 4
Users: 7 /
pro: 3 / con:
2 /
undefined: 2
Users: 4 / pro: 1 /
con: 3
WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS
DOI: 10.37394/23209.2022.19.5
Dimitrios Kydros, Vasiliki Vrana
E-ISSN: 2224-3402
48
Volume 19, 2022
Media: 6 /
pro:6
Politicians: 2
/ pro: 2
Media: 2 / pro: 1 /
con: 1
State: 1 / pro:
1
Ministers: 3 / pro:
3
Political Parties: 1
/ con: 1
November
Users: 9 /
pro: 5 / con:
2 /
undefined: 2
Users: 6 /
pro: 4 / con:
2
Users: 4 / pro:2 /
con: 2
Media: 1 /
pro: 1
Ministers: 1 /
pro: 1
Ministers: 3 / pro:
3
Politicians: 3
/ pro: 3
Politicians: 1 / pro:
1
State: 1 / pro: 1
Media: 1 / pro: 1
December
Users: 5 /
pro: 4 / con:
1
Users: 7 /
pro: 5 / con:
1 /
undefined: 1
Users: 1 / con: 1
Media: 5 /
pro: 5
Politicians: 2
/ pro: 2
Politicians: 4 / pro:
4
Media: 1 /
con: 1
Media: 5 / pro: 5
Table 3 includes some important results that
clearly answer our RQ2. More particularly, we
identified 5 different categories of tweeters: plain
users, politicians (including ministers, the prime
minister, the head of opposition etc.) only one
political party and a number of media such as news
agencies, social media or blogs. It is interesting to
note that to tweeters are mainly users and media in
all three networks. All media and the majority of
users support vaccination; however, a clear minor
group of users does have their objections against.
When it comes to top replied-to and top mentioned,
the situation changes with the presence of members
of the political life, such as the prime minister, the
leader of opposition, ministers, one member of the
European parliament and one former member of the
European parliament. Actually, one of the top
mentioned users is the only active political party in
Greece that has a clear position against vaccination
(“Popular Orthodox Alert”), together with the only
media page that supports it.
Apart from the far-right above mentioned party,
all other political parties in Greece have stated their
support in the vaccination campaign. However,
Table 3 tells a somehow different story: it should be
expected that many more active politicians should
belong to the first column of Table 3, if they were
really active in order to persuade more people to
vaccinate. The same thing happens with political
parties, which are absent whatsoever. No such
presence was found in the data. A possible
explanation might consider the fact that the anti-
vaccination movement in Greece bares no political
barriers and is spilled throughout the political
spectrum.
In order to further clarify the actual reasons for
this poor behavior and also in order to investigate
the actual content of the discussions, in the next
subsection we create and analyze networks of words
found in the actual tweets.
4.3 Content Though Word-Pairs
Networks of word are created having words within
tweets as nodes. A link is created between two
words when they are adjacent (word-pairs) in the
same tweet. This procedure, after been applied to all
tweets and all words of our data sets, created the
networks shown in Figures 2, 3 and 4, which can be
magnified. Certainly, some words were omitted
(such as articles, particles etc.) and all left words
were translated from Greek to English, in order to
maintain readability. In the above mentioned
Figures, we again created groups (communities) of
words and also used a metric, betweenness
centrality, to size words/nodes. Not all word-pairs
are taken into account for reasons of clarity.
Actually we filtered out all word-pairs that appear in
all tweets less than 10 times. The thickness of links
is proportional to the count of appearances of a
specific word-pair. As a final remark, it should be
noted that Figures 2 to 4 do not show simple word
clouds. The presence of links is of outmost
importance, since by magnifying specific areas of
the visualized word network one can identify
sentences or small phrases.
As an example, in Figure 2 that follows, at the
top left (light blue) group, the discussion considers
the situation in northern Greece (around the city of
Thessaloniki), where the counter-vaccination
movement was indeed stronger and was partly
supported by far right, sectarian religious groups.
Actually, such words do penetrate in all larger
groups in the September word network, in some
cases with truly heavy accusations: in the middle
left word groups one can see that the government
acts as in a junta, or a discussion on myocarditis (a
quite rare symptom after vaccination) is
exaggerated. In a similar manner, some discussion
also deals with immigrants, supposed “carriers” and
not vaccinated. Critique to the government
measures is present, but to a mild extend. It is
however true that the majority of words, small
sentences or phrases do show a strong confidence on
vaccination.
WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS
DOI: 10.37394/23209.2022.19.5
Dimitrios Kydros, Vasiliki Vrana
E-ISSN: 2224-3402
49
Volume 19, 2022
Fig. 2: September’s content
In Figure 3, discussions are slightly shifted,
although similar views are circulating. One
important difference here is the questions on the
vaccination schema, since many citizens preferred
to take the one-dose vaccine and are now
questioning on the second or third dosage. In this
network, the mandatory vaccinations in some
professional groups is present, however there is no
obvious agreement about its correctness. Also, some
tweets deal with “freedom” and the supposed rights
of citizens to be unvaccinated if they choose so.
However, the general “feeling” in this network is
that anti-vaccinators have somehow retreated, or at
least they are not so prone to stand for their
opinions.
Fig. 3: November’s content
Finally, in December’s content (Figure 3), it
seems that the anti-vaccination movement is
faded out, at least in the Greek twitter.
Discussions continue on Greek, European and
global news (mainly US and Israeli) and new
measures. The Omicron variation, together with
discussions on the effect of vaccines is clearly
here. Some critique on the non-obligatory
vaccination of police force is also present,
together with some discussion on newly elected
party headers and their views on the subject.
Fig. 4: Decembrer’s content
The above analyzed results clearly answer our
RQ3. In the Greek Twitter sphere, it seems that the
discussions are prevailed by the idea that
vaccination for COVID-19 is quite important and
should be continued by all means. Different
opinions and critique on some of the taken measures
is also present, but not in favor of the anti-
vaccination movement.
However, it is interesting to note that even after
(about) one year of vaccinations and the scientific
knowledge retrieved on its safety and importance, in
Greece there exists a rather large anti-vaccination
movement, since a significant percentage of the
population still refuses to vaccinate. On January 5
2022, Greece Coronavirus Full Vaccination Rate
was 68.10% [45]. This is a large percentage but its
representatives were not traced (to this extend) in
our study. A possible explanation on this could be
the fact that Twitter is not so popular in the Greek
environment, at least to the people involved in such
conversations, since they probably prefer to express
their views in other social media platforms which do
not use limits in character count. Twitter’s platform
cannot be easily used for propaganda due to the
barrier in message length, unless a truly devoted
WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS
DOI: 10.37394/23209.2022.19.5
Dimitrios Kydros, Vasiliki Vrana
E-ISSN: 2224-3402
50
Volume 19, 2022
user or group are continuously twitting on a specific
subject.
4 Conclusions
In this paper we investigated the Greek Twitter
discussions on COVID-19 vaccination during the
last quarter of 2021. More specifically, we searched
for tweets containing Greek words such as
“vaccination”, “anti-vaccination”, “vaccine” etc.
through the Twitter API. We formed three different
data sets, according to generally accepted “periods”
of the above mentioned time period. We then
created networks of users and networks of words
and analyzed them.
Our results show that the network of Greek users
may be considered as fragmented but by all means
not polarized between two different opinions.
Structurally, users can be algorithmically amortized
in groups; however, a clear polarization would not
justify more than 3 to 4 different groups, which was
not found in our case.
Moreover, an analysis of the content of tweets,
made through the creations of networks of adjacent
words again showed that anti-vaccination ideas were
clearly present during the first period of our study
(September to mid-October 2021) but were rapidly
diminished in the following months. The December
period which coincides with a heavily congested
National Health System, a strangely large number of
deaths (when compared to other similar in
population countries) and the advent of Omicron
strain, definitely bears no anti-vaccination
discussions on Greek Twitter.
The persisting large percentage of population
refusing to vaccinate may be expressed in other
social media platforms, although it is known that
most platforms make efforts to stop or prohibit such
talks. However, it seems that other means of
communication is still used for this purpose
including word-of-mouth, a quite old but still
extremely important means to communicate ideas,
especially when time allows for it. Policymakers
and public health officials must prioritize effective
COVID-19 vaccine-acceptance messaging for
Greeks, emphasizing trust in vaccine safety and
dispelling potential myths and spread them across
all media.
It seems that Twitter users do not extensively
share anti-vaccination information. This finding is
encouraging and in accordance with the findings of
Love at al. [19], suggesting that Twitter users
critically think the situation and evaluate the shared
medical content. Governments should take into
consideration public opinion expressed in social
media toward COVID-19 vaccination, understand
the public psychology and evolution of thoughts
through time and implement strategies to promote
COVID-19 vaccination. Governments have to plan
the communication of vaccination health messages
based on evidence [46]. Negative emotions like
anger and fear should be acknowledged so to be
manipulated while governments have to find ways
to activate positive emotions like hope and altruism
[47].
Findings of the study have several limitations.
The study gives evidence of a specific period of
time including the Omicron variant emergence.
However, it provides insights into comprehending
public opinions about vaccination and in in addition
to other research like those of Luo et al. [34] who
focused on users’ opinions toward the vaccine at its
initial stage could give a longitudinal perspective.
Future research should also focus on other social
media platforms to extend the validity of the
findings.
References:
[1] WHO, WHO Coronavirus (COVID-19)
Dashboard. 2021, https://covid19.who.int/ (10-
11-2021)
[2] Harrison, E. A., & Wu, J. W. Vaccine
confidence in the time of COVID-19. European
journal of epidemiology, Vol. 35, No. 4, 2020,
pp. 325-330.
[3] Mehan, A., Venkatesh, A., & Girish, M.
COVID-19: should oral vaccination strategies
be given more consideration?. Therapeutic
advances in vaccines and immunotherapy, Vol.
8, 2020, 2515135520946503,.
[4] Porter, J. R. Vaccines, social measures and
Covid19-A European evidence-based analysis,
2021.
[5] WHO, COVID-19 vaccine tracker and
landscape, 2021, https://www.who.int/
publications/ m/item/draft-landscape-of-covid-
19-candidate-vaccines
[6] Machingaidze, S., & Wiysonge, C. S.
Understanding COVID-19 vaccine hesitancy.
Nature Medicine, Vol. 27, No. 8, 2021, pp.
1338-1339.
[7] Mohamad, O., Zamlout, A., AlKhoury, N.,
Alsalkini, M., & Shaaban, R. Factors
associated with the intention of Syrian adult
population to accept COVID19 vaccination: A
cross-sectional study. BMC Public Health, Vol.
21, No. 1, 2021, pp. 1-10.
[8] Dror, A. A., Eisenbach, N., Taiber, S.,
Morozov, N. G., Mizrachi, M., Zigron, A., ... &
WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS
DOI: 10.37394/23209.2022.19.5
Dimitrios Kydros, Vasiliki Vrana
E-ISSN: 2224-3402
51
Volume 19, 2022
Sela, E. Vaccine hesitancy: the next challenge
in the fight against COVID-19. European
journal of epidemiology, Vol. 35, No. 8, 2020,
pp. 775-779.
[9] Solis Arce, J. S. S., Warren, S. S., Meriggi, N.
F., Scacco, A., McMurry, N., Voors, M., ... &
Omer, S. COVID-19 vaccine acceptance and
hesitancy in low and middle income countries.
Nature Medicine, Vol. 27, No. 8, 2021, pp.
1385-1394
[10] Troiano, G., & Nardi, A. Vaccine hesitancy in
the era of COVID-19. Public Health, No. 195,
2021, pp. 245-251.
[11] Eguia, H., Vinciarelli, F., Bosque-Prous, M.,
Kristensen, T., & Saigí-Rubió, F. Spain’s
Hesitation at the Gates of a COVID-19
Vaccine. Vaccines, Vol. 9, No.2, 2021, 170.
[12] Freeman, D., Loe, B. S., Chadwick, A.,
Vaccari, C., Waite, F., Rosebrock, L., ... &
Lambe, S. COVID-19 vaccine hesitancy in the
UK: the Oxford coronavirus explanations,
attitudes, and narratives survey (Oceans) II.
Psychological medicine, 2020, pp. 1-15.
[13] Peretti-Watel, P., Seror, V., Cortaredona, S.,
Launay, O., Raude, J., Verger, P., ... & Ward, J.
K. A future vaccination campaign against
COVID-19 at risk of vaccine hesitancy and
politicisation. The Lancet Infectious Diseases,
Vol. 20, No. 7, 2020, pp. 769-770.
[14] Dhawan, D., Bekalu, M., Pinnamaneni, R.,
McCloud, R., & Viswanath, K. COVID-19
News and Misinformation: Do They Matter for
Public Health Prevention?. Journal of health
communication, 2021, pp.1-10.
[15] Betsch, C., Brewer, N. T., Brocard, P., Davies,
P., Gaissmaier, W., Haase, N., & Stryk, M.
Opportunities and challenges of Web 2.0 for
vaccination decisions. Vaccine, Vol. 30, No.25,
2012, pp. 3727-3733.
[16] Bean, S. J. Emerging and continuing trends in
vaccine opposition website content. Vaccine,
Vol. 29, No. 10, 2011, pp. 1874-1880.
[17] Stahl, J. P., Cohen, R., Denis, F., Gaudelus, J.,
Martinot, A., Lery, T., & Lepetit, H. The
impact of the web and social networks on
vaccination. New challenges and opportunities
offered to fight against vaccine hesitancy.
Medecine et maladies infectieuses, Vol. 46. No.
3, 2016, pp. 117-122.
[18] Limaye, R. J., Holroyd, T. A., Blunt, M.,
Jamison, A. F., Sauer, M., Weeks, R., ... &
Gellin, B. Social media strategies to affect
vaccine acceptance: a systematic literature
review. Expert review of vaccines, Vol. 20, No.
8, 2021, pp. 959-973.
[19] Love, B., Himelboim, I., Holton, A., &
Stewart, K. Twitter as a source of vaccination
information: content drivers and what they are
saying. American journal of infection control,
Vol. 41, No. 6, 2013, pp. 568-570.
[20] Kata, A. A postmodern Pandora's box: anti-
vaccination misinformation on the Internet.
Vaccine, Vol. 28, No. 7, 2010, pp. 1709-1716.
[21] Kata, A. Anti-vaccine activists, Web 2.0, and
the postmodern paradigm–An overview of
tactics and tropes used online by the anti-
vaccination movement. Vaccine, Vol. 30, No.
25, 2012, pp. 3778-3789.
[22] Dannetun, E., Tegnell, A., Hermansson, G., &
Giesecke, J. Parents’ reported reasons for
avoiding MMR vaccination: a telephone
survey. Scandinavian journal of primary health
care, Vol. 23, No.3, 2005, pp.149-153.
[23] Mason, B. W., & Donnelly, P. D. Impact of a
local newspaper campaign on the uptake of the
measles mumps and rubella vaccine. Journal of
Epidemiology & Community Health, Vol 54,
No.6, 2000, pp. 473-474.
[24] Armitage, R. Online ‘anti-vax’campaigns and
COVID-19: censorship is not the solution.
Public Health, No.190, 2021, e29
[25] Center for Countering Digital Hate, The Anti-
Vaxx Industry. How Big Tech powers and
profits from vaccine misinformation, 2020,
https://www.counterhate.com/anti-vaxx-
industry
[26] Wawrzuta, D., Jaworski, M., Gotlib, J., &
Panczyk, M. Characteristics of Antivaccine
Messages on Social Media: Systematic Review.
Journal of Medical Internet Research, Vol. 23,
No. 6, 2021, e24564.
[27] Burki, T. The online anti-vaccine movement in
the age of COVID-19. The Lancet Digital
Health, Vol. 2, No.10, 2020, e504-e505.
[28] D’Errico, S., Turillazzi, E., Zanon, M., Viola,
R. V., Frati, P., & Fineschi, V. (2021). The
Model of “Informed Refusal” for Vaccination:
How to Fight against Anti-Vaccinationist
Misinformation without Disregarding the
Principle of Self-Determination. Vaccines, Vol.
9, No.2, 2021, 110.
[29] Karami, A., Zhu, M., Goldschmidt, B.,
Boyajieff, H. R., & Najafabadi, M. M. COVID-
19 Vaccine and Social Media in the US:
Exploring Emotions and Discussions on
Twitter. Vaccines, Vol. 9, No.10, 2021, 1059.
[30] Adebisi, Y. A., Alaran, A. J., Bolarinwa, O. A.,
Akande-Sholabi, W., & Lucero-Prisno, D. E.
When it is available, will we take it? Social
media users’ perception of hypothetical
COVID-19 vaccine in Nigeria. The Pan African
Medical Journal, 2021, 38.
WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS
DOI: 10.37394/23209.2022.19.5
Dimitrios Kydros, Vasiliki Vrana
E-ISSN: 2224-3402
52
Volume 19, 2022
[31] Herrera-Peco, I., Jiménez-Gómez, B., Peña
Deudero, J. J., Benitez De Gracia, E., & Ruiz-
Núñez, C. Healthcare Professionals’ Role in
Social Media Public Health Campaigns:
Analysis of Spanish Pro Vaccination Campaign
on Twitter. Healthcare, Vol. 9, No. 6, 2021,
662.
[32] Piltch-Loeb, R., Savoia, E., Goldberg, B.,
Hughes, B., Verhey, T., Kayyem, J., ... &
Testa, M. Examining the effect of information
channel on COVID-19 vaccine acceptance.
Plos one, Vol.16, No.5, 2021, e0251095.
[33] Puri, N., Coomes, E. A., Haghbayan, H., &
Gunaratne, K. Social media and vaccine
hesitancy: new updates for the era of COVID-
19 and globalized infectious diseases. Human
vaccines & immunotherapeutics, Vol.16,
No.11, 2020, pp. 2586-2593.
[34] Luo, C., Chen, A., Cui, B., & Liao, W.
Exploring public perceptions of the COVID-19
vaccine online from a cultural perspective:
Semantic network analysis of two social media
platforms in the United States and China.
Telematics and Informatics, Vol. 65, 2021,
101712.
[35] Thelwall, M., Kousha, K., & Thelwall, S.
Covid-19 vaccine hesitancy on English-
language Twitter. Profesional de la
información (EPI), Vol.30, No.2, 2021,
e300212.
[36] Criss, S., Nguyen, T. T., Norton, S., Virani, I.,
Titherington, E., Tillmanns, E. L., ... & Gee, G.
C. Advocacy, Hesitancy, and Equity: Exploring
US Race-Related Discussions of the COVID-
19 Vaccine on Twitter. International Journal of
Environmental Research and Public Health,
Vol. 18, No.11, 2021, 5693.
[37] Bonnevie, E., Gallegos-Jeffrey, A., Goldbarg,
J., Byrd, B., & Smyser, J. Quantifying the rise
of vaccine opposition on Twitter during the
COVID-19 pandemic. Journal of
communication in healthcare, Vol.14, No. 1,
2021, pp. 12-19.
[38] Liu, S., & Liu, J. Understanding Behavioral
Intentions Toward COVID-19 Vaccines:
Theory-Based Content Analysis of Tweets.
Journal of Medical Internet Research, Vol. 23,
No. 5, 2021, e28118.
[39] Kwok, S. W. H., Vadde, S. K., & Wang, G.
Tweet topics and sentiments relating to
COVID-19 vaccination among Australian
Twitter users: Machine learning
analysis. Journal of medical Internet
research, Vol.23, No.5, 2021, e26953.
[40] Scannell, D., Desens, L., Guadagno, M., Tra,
Y., Acker, E., Sheridan, K., ... & Fulk, M.
COVID-19 vaccine discourse on Twitter: A
content analysis of persuasion techniques,
sentiment and mis/disinformation. Journal of
health communication, Vol. 26, No.7, 2021,
443-459.
[41] Kydros, D., Argyropoulou, M. Vrana, V. A
Content and Sentiment Analysis of Greek
Tweets during the Pandemic. Sustainability, 13,
2021, 6150.
[42] Smith, M., Ceni A., Milic-Frayling, N.,
Shneiderman, B., Mendes Rodrigues, E.,
Leskovec, J., Dunne, C. NodeXL: a free and
open network overview, discovery and
exploration add-in for Excel
2007/2010/2013/2016 from the Social Media
Research Foundation, 2010,
https://www.smrfoundation.org
[43] Kantriwitz, A. The Man Who Built the Retweet:
We Handed A Loaded Weapon To 4 Year-Olds.
2019, https://rb.gy/hrpget
[44] Clauset, A., Newman, M., Moore, C. Finding
community structure in very large networks.
Physical review. E, Statistical, nonlinear, and
soft matter physics. Vol.70, 2005, 066111.
[45] Ycharts. Greece Coronavirus Full Vaccination
Rate, 2022. https://ycharts.com/indicators/
greece_coronavirus_full_vaccination_rate
[46] Borah, P., Hwang, J., & Hsu, Y. C. COVID-19
Vaccination Attitudes and Intention: Message
Framing and the Moderating Role of Perceived
Vaccine Benefits. Journal of health
communication, Vol. 26, No. 8, 2021, pp. 523-
533.
[47] Chou, W. Y. S., & Budenz, A. Considering
emotion in COVID-19 vaccine communication:
addressing vaccine hesitancy and fostering
vaccine confidence. Health communication,
Vol.35, No.14, 2020, pp. 1718-1722.
Contribution of Individual Authors to the
Creation of a Scientific Article (Ghostwriting
Policy)
Dimitrios Kydros carried out Data curation and
Formal analysis, Visualization, Writing
Vrana Vasiliki has implemented Conceptualization,
Methodology, and Writing
Creative Commons Attribution License 4.0
(Attribution 4.0 International, CC BY 4.0)
This article is published under the terms of the
Creative Commons Attribution License 4.0
https://creativecommons.org/licenses/by/4.0/deed.en
_US
WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS
DOI: 10.37394/23209.2022.19.5
Dimitrios Kydros, Vasiliki Vrana
E-ISSN: 2224-3402
53
Volume 19, 2022