En i onmen al Science and Policy 124 (2021) 451–460
A ailable online 30 July 2021
1462-9011/© 2021 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).
How has he COVID-19 pandemic a ec ed he clima e change deba e
on Twi e ?
Ma ia L. Lou ei o
a
,
*, Ma ia All´
o
b
a
ECOBAS, Uni e si y o San iago de Compos ela, Spain
b
Uni e si y o A Co u˜
na, Spain
ARTICLE INFO
Keywo ds:
Clima e change
COVID-19
Public bads
Social no ms
Twi e
ABSTRACT
Clima e change and he COVID-19 pandemic sha e many simila i ies. Howe e , in he pas mon hs, conce ns
ha e inc eased abou he ac he heal h eme gency has pu on hold du ing he pandemic many clima e adap-
a ion and mi iga ion policies. We ocus ou a en ion on unde s anding he ole o he ecen heal h eme gency
on he ansmission o in o ma ion ela ed o clima e change, join ly wi h o he socio-economic a iables, social
no ms, and cul u al dimensions. In doing so, we c ea e a unique da ase con aining he numbe o wee s w i en
wi h speci ic clima e ela ed keywo ds pe coun y wo ldwide, as well as coun y speci ic socio-economic
cha ac e is ics, ele an social no ms, and cul u al a iables. We ind ha socio-economic a iables, such as
income, educa ion, and o he isk- ela ed a iables ma e in he ansmission o in o ma ion abou clima e
change and Twi e ac i i y. We also ind ha he COVID-19 pandemic has signi ican ly dec eased he o e all
numbe o messages w i en abou clima e change, pos poning he clima e deba e wo ldwide; bu pa icula ly in
some ulne able coun ies. This shows ha in spi e o he exis ing clima e eme gency, he cu en pandemic has
had a de imen al e ec o e he sho - e m planning o clima e policies in coun ies whe e clima e ac ion is
u gen .
1. In oduc ion
The clima e c isis is expec ed o be he mos dange ous p oblem e e
aced by Humani y. Howe e , some au ho s a gue ha in he pas ew
mon hs and due o he u gency du ing he ongoing COVID-19 pandemic,
conce ns abou clima e change ha e been pu on hold in some coun ies.
Fo example, in B azil, du ing 2020 a signi ican numbe o pieces o
legisla ion ha elax en i onmen al laws, om easing o es p o ec ions
o declassi ying he oxici y o pes icides we e app o ed du ing he
second wa e o COVID-19(Vale e al., 2021). In he same pe iod, he
Uni ed S a es has olled back ce ain en i onmen al egula ions and
appea ed o di ec s imulus unds owa d ein igo a ing he ossil uel
indus y (Rosenbloom and Ma ka d, 2020). In he mean ime, he
Ge man Council o Economic Expe s (2020) p oduced a epo on he
co ona i us c isis wi hou including any en i onmen al conside a ions,
o men ioning he wo ds clima e change o sus ainabili y. The Columbia
Clima e School (2020) also epo s he delays o COP-26 and in e na-
ional nego ia ions ela ed o he p o ec ion o he en i onmen ha may
allow coun ies o mo e hei ac ions away om he igh agains
clima e change. The In e na ional Ene gy Agency has no ed delays in
clima e iendly policies and ela ed in es men s (In e na ional Ene gy
Agency (IEA, 2020). In addi ion, some p i a e companies ha e inc eased
he p oduc ion o was e and pollu ion, due o he consequence o he
hea y use o plas ics, and p i a e anspo a ion. These a e jus a ew
examples ha illus a e he complex e ec s o COVID-19 in ela ion o
clima e change adap a ion and mi iga ion policies.
While he cu en heal h pandemic also o e s unp eceden ed in-
sigh s in o how he global clima e c isis may be managed, i also man-
i es s a clea need o ackle clima e change as an u gen manne (Klene
e al., 2020; Mandanedo and Manning, 2020). Fuen es e al. (2020)
examine bo h global c ises, showing ha bo h a e global public bads,
ca ying e y la ge economic cos s, while sha ing ha ex-an e mi iga ion
and p e en ion a e cheape han managing ex-pos impac s.
The e a e wo cha ac e is ics ha de ine he na u e o p o ision o
public goods, as hose ep esen ing clima e con ol policies. In pa ic-
ula , many indi iduals can ob ain gains om hem, while exclusion o
o he bene icia ies om such associa ed impac s is a di icul o almos
impossible ask (non-excludabili y and non- i al y). These in insic
cha ac e is ics may gene a e ce ain incen i es o misbeha io , by
which indi iduals may no be willing o con ibu e o push o wa d wi h
* Co esponding au ho .
E-mail add ess: [email p o ec ed] (M.L. Lou ei o).
Con en s lis s a ailable a ScienceDi ec
En i onmen al Science and Policy
jou nal homepage: www.else ie .com/loca e/en sci
h ps://doi.o g/10.1016/j.en sci.2021.07.011
Recei ed 10 Feb ua y 2021; Recei ed in e ised o m 7 July 2021; Accep ed 12 July 2021
En i onmen al Science and Policy 124 (2021) 451–460
452
he p o ision o he public good (in ou case, clima e-con ol ela ed
ac ions), gi en ha hey will ha e a gain o loss based on wha he es o
he socie y will do anyway. This ee- iding p oblem may be pa icula ly
acu e in cases whe e indi iduals eel ha i is un ai o hem o ix a
p oblem caused by hi d pa ies. This a gumen has been pu o wa d by
many coun ies ecen ly, in he con ex o he Pa is Ag eemen . In-
di iduals may no be willing o con ibu e wi h hei e o ei he o his
public good p o ision o clima e con ol, because hey may eel ha
hei con ibu ion is meaningless, wi h espec o he magni ude o he
p oblem a hand; and no ma e how much e o hey pu indi idually,
he p oblem will no be sol ed. O he s may conside ha he p oblem
will no be ixed in hei li espan, p e e ing o enjoy he p esen , and
lea ing he issue o be sol ed by u u e gene a ions. These empo al
dimensions may unde mine he unde s anding o he cu en u gency o
he p oblem. In sho , mul iple easons can be ound o explain he
unde -p o ision o public goods due o economic sel ishness.
The e o e, in o de o unde s and he dynamics o he p o ision o-
wa ds clima e change policies, i is o in e es o analyze how in o ma-
ion abou clima e change is being ansmi ed. Twi e con e sa ions
ela ed o clima e change a e being conside ed as a p oxy o socie al
conce ns and awa eness o he p oblem, as in p e ious s udies (Jacques
and Know, 2016; Mayna d and Bon che a, 2015). In pa icula , we
in es iga e wha ype o ela ionship exis s be ween clima e
change- ela ed Twee s, and o he socio-economic cha ac e is ics, cul-
u al a iables, social no ms, and ex e nal shocks. I is pa icula ly
ele an o unde s and how he cu en pandemic has impac ed he
clima e change deba e. In o de o unde ake his analysis, we ga he
in o ma ion ob ained om Twi e om 2018 onwa ds. In o al, mo e
han 48 million wee s (48,234,241 wee s) ha e been ga he ed wo ld-
wide in mo e han 210 coun ies (See Appendix A).
1.1. Objec i es and hypo heses
Ou main esea ch objec i e is o show ha al hough he ecen
pandemic and clima e change a e bo h a e public bads, he a en ion
equi ed o igh he cu en pandemic and he subsequen educ ion o
emissions due o massi e lockdowns o he popula ion, and ini ial
mobili y es ic ions, ha e diminished cu en conce ns abou clima e
change, educing he public deba e abou clima e change in many
coun ies. A second objec i e o ou esea ch is o unde s and he ole o
cul u al a iables and social no ms ela ed issues when communica ing
abou clima e change.
Speci ic esea ch hypo heses examined in his pape a e ela ed o
he mechanisms by which social media in o ma ion abou clima e
change is being ansmi ed. In pa icula , we aim o es he ollowing
esea ch hypo heses:
1 Risk a e sion inc eases he dissemina ion abou clima e change.
2 Al uism inc eases he dissemina ion o in o ma ion abou clima e
change.
3 The cu en COVID-19 pandemic has educed he communica ion
ac i i y a ound he clima e change deba e.
The es o he pape is s uc u ed as ollows: Sec ion 2 p esen s a
sho li e a u e e iew abou public goods, clima e change and Twi e
usage. Sec ion 3 con ains he da a desc ip ion employed in ou analysis,
while Sec ion 4 shows he empi ical model and esul s. Finally, he pape
concludes in Sec ion 5.
2. Li e a u e e iew
Resea ch, mainly based on beha io al economics, shows he g ea
impo ance o human in e ac ions in e ms o ecip oci y and con o mi y
(Ba dsley and Sausg ube , 2005), al uism (And eoni, 1989, 1990; Fong
and Lu me , 2009); cul u al di e ences (Casson e al., 2002); social
capi al (Ande son e al., 2004); in o ma ion aming (Cookson, 2000),
and o he beha io al anomalies (Gowdy, 2008) in he p o ision o
public goods. These con ibu ions ha e shown he cogni i e limi a ions
o adi ional economic heo y, based on sel -in e es ed and a ional
indi iduals. In summa y, s udies om beha io al economics ha e
shown ha en i onmen al jus ice and social no ms also a ec indi idual
decisions, and he e o e should be aken in o accoun in adi ional
economic models. As i u ns ou , when conside ing a global p oblem
(bad) such as clima e change, indi iduals can be in luenced by alues
and belie s sha ed in g oups o which hey eel a sense o belonging
(Ho man, 2011).
The idea o coope a ion in he p o ision o public goods has been pu
o wa d by many schola s, including Os om (2000), who sugges ed ha
a icula ed plans based on sha ing common social no ms can de e mine
a “good” p o ision and conse a ion o he public good. B ekke and
Johansson-S enman (2008), sugges ha wha i may be a ional o a
single coun y o indi idual, may be subop imal o he g oup; and
pa icula , i all ollow a ional sel -in e es ed pa hs, a global good o a
ce ain le el o comp ise o p o ision will no be achie ed.
P e ious a emp s o unde s anding clima e change in o ma ion and
communica ion dynamics on Twi e ha e been unde aken. In pa ic-
ula , and in e ms o assessing he impac o Twi e con e sa ions on he
di usion o in o ma ion ela ed o clima e change, Ki ilenko and S ep-
chenko a (2014) analyzed messages in i e di e en languages (English,
Ge man, Russian, Po uguese and Spanish), inding geog aphical di -
e ences ela ed o wee ing, di e en ime pa e ns, and he impo ance
gi en o he impac o e en s in he discussion. They ound la ge a i-
a ions ac oss me opoli an a eas and by opic. Cody e al. (2015)
analyzed he collec i e sen imen h ough di e en episodes ela ed o
clima e change, concluding ha na u al disas e s and o he phenomena
ela ed o clima e change con ibu ed o a dec ease in he le el o o e all
happiness. Ki ilenko e al. (2015) also analyzed he use o Twi e in he
USA o assess whe he people ela ed a change in he pe cei ed em-
pe a u e wi h he global d i e o clima e change. Mayna d and
Bon che a (2015) used Twi e o analyze and unde s and he social
engagemen ega ding clima e change, wi h he aim o helping o gani-
za ions o ca y ou be e campaigns o in o ma ion and imp o ing
socie al unde s anding. Holmbe g and Hells en (2015) s udied whe he
gende di e ences a e impo an when explaining he communica ion o
clima e change. They ound ha o e all, emale and male wee e s used
a simila language, deno ing di e ences ega ding he use o hash ags
and use names.
In addi ion, he social impac o speci ic na u al disas e s has also
been analyzed wi h social media. Fo example, K y asheyeu e al.
(2016) s udied he impac o he Hu icane Sandy on Twi e con e -
sa ions, inding a ela ionship be ween he p oximi y o he hu icane
pa h and social media ac i i y. Jacques and Knox (2016) also analyzed
wee s abou hu icane Sandy o unde s and he denial discou se o
clima e change. They ound h ee majo g oups o denials: hose who
ejec clima e science because clima e science is a conspi acy a o ing
g ow h o go e nmen ; hose who oppose enewable ene gy and ene gy
axa ion; and inally, a hi d g oup exp essing ea o go e nmen al
abuse o powe . Sisco e al. (2017) analyzed Twi e con e sa ions o
assess how ex eme wea he e en s gene a e a en ion o clima e
change, inding ha he inancial damage linked o hese e en s is a
good p edic o o he a en ion paid o clima e change in he USA.
Roxbu gh e al. (2019) assess he ole clima e change may o may no
ha e played in in luencing h ee high-magni ude ex eme wea he
e en s– Hu icane I ene, Hu icane Sandy and Snows o m Jonas,
inding ha clima e change con e sa ions ma e in di e en ways in he
h ee e alua ed e en s. Mo e ecen ly, Lou ei o and All´
o (2020) assess
how social media da a om Twi e acili a e in e na ional compa isons
in e ms o p e e ences and emo ions owa ds clima e and ene gy pol-
icies in wo di e en coun ies (Spain and he UK). Consequen ly, he
p e iously exis ing li e a u e highligh s he impo ance and alidi y o
social media e lec ing ha “human senso s” may an icipa e economic
impac s (Ki ilenko e al., 2015).
M.L. Lou ei o and M. All´
o
En i onmen al Science and Policy 124 (2021) 451–460
453
We ex end hese p e ious analyses by conside ing he ole o addi-
ional socioeconomic a iables no p e iously s udied in in o ma ion
ansmission, in he con ex o con e sa ions ela ed o clima e change.
We conside he impac o isk pe cep ions, ime p e e ences, and social
ac o s when ansmi ing in o ma ion abou clima e change. In sum-
ma y, we look a he ole o social no ms and cul u al a iables in
shaping he clima e change deba e du ing he pe iod o analysis. In a
second s ep, we speci ically conside he impac o COVID-19 in
Twee ing ac i i y ela ed o clima e change.
3. Da ase desc ip ion
3.1. Twi e da a
Twi e is a social ne wo k which has a ound 313 million mon hly
ac i e use s in 2020 (OBe lo, 2021). In o de o unde ake he objec i es
o his esea ch, a mon hly da abase has been c ea ed con aining in-
o ma ion abou wee ing ac i i y ela ed o clima e change in 213
coun ies, egis e ing mo e han 100 wee s pe loca ion (See Appendix
A) since 2018 and un il Sep embe 2020 ( eco ding a ound 48,234,241
wee s).
The da a collec ion p ocess has been based in he sea ch o wee s
h ough he use o hash ags and keywo ds. Thus, eques s ha e been
made h ough he lib a y Tweepy o Py hon which wo ks wi h Twi e
API (2021) speci ying a “keywo d” o speci ic “hash ags”. A he same
ime a wee was ecei ed, ano he eques was made abou he use ha
published he message. Wi h espec o he use , in o de o gain in o -
ma ion abou hei gende , he Gende API was employed (Gende API,
2019). This API can de ec gende om social media by que ying use -
names. The assignmen ype is p obabilis ic, and when he use name has
egis e ed unde a name no included in i s da abase, he unknown
ca ego y is assigned as a de aul . As desc ibed by San ama ía and
Mihalje i´
c (2018), i s da abase size is by a one o he la ges employed
o gende iden i ica ion s udies (pa icula ly when compa ed wi h
simila ools). This so o iden i ica ion me hod is qui e popula , and i s
accu acy is one o he highes achie ed. Resul s show ha abou 25 % o
he o al egis e ed wee s ha e been w i en by emale (38 % i we only
conside use s whose gende has been iden i ied).
As he e a e some es ic ions in e ms o downloads om Twi e , a
ime able has been es ablished o collec he da a. Wo ldwide da a
collec ion is done a ge ing wo main languages: English and Spanish.
Speci ically, Mondays, Wednesdays, F idays and Sundays he API
collec ed wee s in Spanish ha we e published in he las 30 days; while
Tuesdays, Thu sdays and Sa u day collec ed wee s in English. I is
impo an o speci y ha bo h wee s and e wee s ha e been eco ded.
Al hough da a collec ion is e y ex ensi e, he inal da ase con ains
in o ma ion om 20 coun ies,
1
eco ding a o al o 36,205,609 usable
wee s (75 % o he o al). Table 1 p esen s he dis ibu ion o wee s and
in o ma ion abou coun ies wi h he highes a e o numbe o wee s
(in o al, hey ep esen mo e han 92.82 % o wee s om he sample o
20 coun ies). As can be seen, h ee coun ies ep esen o e hal o he
numbe o wee s in his da ase ; he USA accoun s o a ound 33.42 % o
wee s, ollowed by Spain wi h a ound 15.37 % o wee s, and he UK
ep esen s a ound 10.31 % o he o al. In addi ion, i can also be
obse ed ha he mos p edominan hash ags a e #clima echange,
#wea he and #clima e. Mo eo e , use s also wee abou he disas e s
su e ed ela ed o clima e change, including wild i es, d ough s o
hu icanes. G aph 1 shows he numbe o o al wee s by coun y, and as
can be seen he USA, UK and Aus alia a e he mos ac i e coun ies on
hese social media; whe eas in he case o Spanish-speaking coun ies,
he mos p edominan is Spain, ollowed by A gen ina, Mexico, Chile
and Colombia.
In e ms o empo al a ia ions, he highes numbe o wee s was
egis e ed du ing he all 2019 and ea ly 2020, wi h a clea d op om
Ma ch 2020 onwa ds. G aph 2 shows he dis ibu ion o wee s o e
ime o he inal sample. G aph 3 shows he e olu ion o wee s o e
ime bu conside ing he Hemisphe e whe e wee s we e pos ed, in o de
o con ol o seasonal a ia ions. This ein o ces he impo ance o he
impac o COVID-19 on social media communica ion abou clima e
change.
3.2. Sociodemog aphic a iables da ase
The Twi e da ase has been augmen ed wi h socio-economic a i-
ables collec ed om he O ganisa ion o Economic Co-ope a ion and
De elopmen (OECD) (2020) and om he Wo ld Bank S a is ics. In
pa icula , he na ional GDP comes om he Qua e ly Na ional Ac-
coun s published by he OECD (a a iable ha has been ans o med and
exp essed as GDP/numbe o in e ne use s). Fu he mo e, and in o de
o collec a p oxy o educa ion, he public expendi u es on educa ion,
exp essed as a % o GDP om The Wo ld De elopmen Indica o s
da ase (Wo ld Bank, 2020) has been included (5.2 %). In addi ion, we
accoun o he language in he coun y, wi h Spanish-speaking coun-
ies gene a ing abou 25 % o he sample o wee s (See Table 2).
3.3. Expe iencing clima e change impac s
To analyze he main ac o s a ec ing how people communica e
abou clima e change on Twi e , addi ional a iables ha e been added
o his social media da ase . Thus, i is expec ed ha in coun ies whe e
impo an clima e- ela ed ca as ophes o ex eme e en s ha e
occu ed, an inc ease in Twi e ac i i y may be egis e ed a ound hese
phenomena. To e lec hese ypes o e en s and na u al disas e s, in-
o ma ion om he In e na ional Disas e Da abase (EM-DAT) (2020)
has been collec ed. We ind ha mainly eigh ype o na u al disas e s
ha e been expe ienced du ing he pe iod o analysis: d ough s, ea h-
quakes, ex eme empe a u e episodes, loods, landslide, s o ms, ol-
canic ac i i y and wild i es. All o hese, excep ea hquakes, can be
ela ed o clima e change (ei he as a cause o a di ec consequence). As
an example, and acco ding o S enchiko (2016), olcanic e up ions
may p oduce long- e m impac s in oceans and in he a mosphe e, ac i-
a ing complex clima e eedbacks.
3.4. O ganiza ion o UNFCC clima e change con e ences
Addi ional a iables e e ing o he o ganiza ion o he in e na ional
clima e change Con e ences o he Pa ies (COP con e ences) by Uni ed
Na ions Clima e Change (UNFCC, 2020) ha e been conside ed, in o de
o assess he impac o hese e en s on he a en ion gi en in social
media o clima e change. Fi e di e en UNFCC COP con e ences ook
place du ing 2018-Sep embe 2020 a ound he wo ld, as desc ibed in
Table 2.
Table 1
Pe cen age o numbe o wee s and hash ags by coun y.
Coun y Pe cen age (%) Mos popula hash ags
Uni ed S a es 33.42 #clima echange, #wea he , #hu icane
Spain 15.37 #clima echange, #clima e, # lood
Uni ed Kingdom 10.31 #clima echange, #wea he , #clima eac ion
A gen ina 8.64 #clima echange, #hu icane, #wild i e
Mexico 8.39 #clima echange, #clima e, #hu icane
Chile 7.04 #clima echange, #d ough , #wild i e
Aus alia 4.88 #clima echange, #clima eac ion, #d ough
Colombia 4.77 #clima echange, #clima e, #globalwa ming
To al 92.82
Sou ce: Own elabo a ion
1
The inal da ase con ains da a om 20 coun ies due o he exis ence o
missing alues when me ging in o ma ion e ie ed om Twi e wi h addi-
ional a iables. An impo an limi a ion is ha he Wo ld Values Su ey has
only been conduc ed in a educed se o coun ies.
M.L. Lou ei o and M. All´
o
En i onmen al Science and Policy 124 (2021) 451–460
454
3.5. Risk, ime and social a iables
Finally, wi h he aim o es ing he e ec o isk pe cep ions, ime
p e e ences, and o he social ac o s when ansmi ing in o ma ion
abou clima e change, a iables e e ing o social no ms we e included.
Speci ically, a iables ex ac ed om he Global P e e ence Su ey
(2020) which is a “globally ep esen a i e da ase on isk and ime p e e -
ences, posi i e and nega i e ecip oci y, al uism and us ” ha e been
conside ed (al hough in he empi ical speci ica ions, wo o hem ha e
been d opped due o high co ela ion p oblems
2
). These a iables ep-
esen ing p e e ences a e based on 12 su ey i ems (see Falk e al., 2018,
o mo e de ails). The a iables inally conside ed a e: a isk- aking in-
dica o ha exp esses he “willingness o ake isks in gene al,” al uism
(“willingness o gi e o good causes”), us (“people ha e only he bes
in en ions”), and nega i e ecip oci y (“willingness o punish un ai
beha io owa d o he s”) (See Table 2 o de ailed desc ip ion). I is
impo an o no e ha in o de o in e p e hese measu es, Falk e al.
(2018) indica e ha each p e e ence was s anda dized a he indi idual
le el; he e o e, each measu e has a mean o ze o and a s anda d de i-
a ion o one. Thus, we ha e in o ma ion abou he di e ence o he
wo ld mean exp essed in s anda d de ia ions.
3.6. Poli ical iews
In o de o unde s and how poli ical iews shape he deba e abou
clima e change, an indica o o poli ical o ien a ion was also included.
This in o ma ion comes om he Wo ld Values Su ey, Wa e 7:
2017-2020 and i is an index ha anges om 1 (le ) o igh (10).
McC igh e al. (2015) ha e analyzed whe he he poli ical ideology may
imply di e ences in e ms o pe cep ion abou clima e change; inding
G aph 1. Numbe o wee s by coun y.
Sou ce: own elabo a ion
G aph 2. Tempo al a ia ion o wee s.
Sou ce: own elabo a ion
2
Pa ience shows a co ela ion coe icien o 0.775 wi h GDP. Posi i e eci-
p oci y shows a co ela ion coe icien o 0.779 wi h al uism.
M.L. Lou ei o and M. All´
o
En i onmen al Science and Policy 124 (2021) 451–460
455
ha in Eu ope, hose close o igh - wing pa ies a e less likely o wo y
abou he clima e p oblem han hose close o le - wing pa ies.
3.7. Co id-19 pandemic
To con ol o he po en ial impac o he COVID-19 pandemic, while
conside ing he ac ha he impac has been he e ogeneous ac oss
coun ies, we ha e included a a iable deno ing mon hly dea hs
G aph 3. Tempo al a ia ion o wee s by Hemisphe e.
Sou ce: own elabo a ion
M.L. Lou ei o and M. All´
o

En i onmen al Science and Policy 124 (2021) 451–460
456
eco ded pe coun y acco ding o he Eu opean Cen e o Disease
P e en ion and Con ol (2021) as a consequence o COVID-19. In
addi ion, indi idual c oss-p oduc s iden i ying he e ec o COVID-19
dea hs pe coun y ha e been included.
4. Empi ical model and esul s
4.1. Empi ical models
Ou empi ical s a egy is based on he es ima ion o a baseline
eg ession ha models wee ing abou clima e change wi h espec o
socioeconomic a iables (income, educa ion gende ), social no ms, po-
li ical p e e ences, expe iences wi h clima e- ela ed ex eme e en s, and
he celeb a ions o he UNFCC COP con e ences.
In pa icula , he eg ession o be es ima ed has been speci ied as
ollows:
Ln(Yi ) =
α
+βXi +γZi +δDi +
η
Ti +
ε
i (1)
Whe e Yi ep esen s he numbe o wee s no malized by he numbe o
in e ne use s pe coun y (i) and mon h ( ), which is being modeled as a
unc ion o a ec o o socio-economic a iables Xi , social no ms, ime
and isk p e e ences, and poli ical iews ep esen ed by Zi , ac ual im-
pac s caused by clima e change Di , and he a iable Ti ha con ols o
he impac o he COVID-19 pandemic (numbe o dea hs). An ex ended
speci ica ion has been also conside ed, including c oss-p oduc a iables
be ween he coun y indica o s and he COVID-19 indica o . This may
e lec he e ogeneous pe cep ions abou he ela ionship be ween
clima e change and he COVID-19 pandemic.
4.2. Resul s
Baseline esul s a e epo ed in Table 3. In e ms o socio-economic
ac o s ha may a ec he ansmission o in o ma ion in Twi e , we
ind ha weal h (GDP) and educa ion play signi ican oles inc easing
he numbe o Twee s su ounding he opic o in e es . Fu he mo e,
Spanish-speaking coun ies a e mo e likely o wee abou clima e
change han he es . This may be explained by he high deg ee o
clima e ulne abili y o La in Ame ican coun ies and Spain. We also
ind ha women wee less abou clima e change han males, as in
gene al. P e ious wo k has shown a posi i e and clea impac o ac ual
expe iences wi h ca as ophic e en s on wee ing ac i i y (Roxbu gh
e al. (2019). Ou esul s ein o ce hese p e ious indings, by showing
Table 2
Summa y s a is ics.
Va iable Desc ip ion Mean S d.
De .
Ln wee s Ln[Numbe o wee s/in e ne
use s)*100,000]
3.559 1.889
Socioeconomic cha ac e is ics
Women Pe cen age o wee s pos ed by
women
24.877 6.588
GDP/In e ne Use s GDP/Numbe o in e ne use s 0.053 0.016
Spanish speaking coun ies 1, i he coun y is a Spanish-
speaking coun y; 0 o he wise
0.250 0.433
Educa ion expendi u e Gene al go e nmen
expendi u e on educa ion
exp essed as a pe cen age o
GDP.
5.168 1.005
Na u al disas e s
D ough Numbe o d ough s su e ed
acco ding o he EM-DAT o
each speci ic pe iod o ime
0.008 0.087
Ea hquake Numbe o ea hquakes
su e ed acco ding o he EM-
DAT o each speci ic pe iod o
ime
0.027 0.172
Ex eme empe a u es Numbe o ex eme
empe a u es episodes su e ed
acco ding o he EM-DAT o
each speci ic pe iod o ime
0.036 0.203
Flood Numbe o loods su e ed
acco ding o he EM-DAT o
each speci ic pe iod o ime
0.136 0.433
Landslide Numbe o landslides su e ed
acco ding o he EM-DAT o
each speci ic pe iod o ime
0.011 0.103
S o m Numbe o s o ms su e ed
acco ding o he EM-DAT o
each speci ic pe iod o ime
0.098 0.379
Volcanic ac i i y Numbe o olcanic ac i i y
episodes su e ed acco ding o
he EM-DAT o each speci ic
pe iod o ime
0.015 0.134
Wild i es Numbe o wild i es su e ed
acco ding o he EM-DAT o
each speci ic pe iod o ime
0.017 0.150
Clima e change mee ings
CCmee ing_ap il_Bonn18 1, indica ing he mee ing abou
clima e change ha ook place
in Ap il 2018 in Bonn;
0 o he wise
0.030 0.172
CCmee ing_sep _Bangkok18 1, indica ing he mee ing abou
clima e change ha ook place
in Sep embe 2018 in Bangkok;
0 o he wise
0.030 0.172
CCmee ing_dec_Ka owice18 1, indica ing he mee ing abou
clima e change ha ook place
in Decembe 2018 in Ka owice;
0 o he wise
0.030 0.172
CCmee ing_june_Bonn19 1, indica ing he mee ing abou
clima e change ha ook place
in June 2019 in Bonn;
0 o he wise
0.030 0.172
CCmee ing_dec_Mad id19 1, indica ing he mee ing abou
clima e change ha ook place
in Decembe 2019 in Mad id;
0 o he wise
0.030 0.172
Va iables om he Global P e e ence Su ey
Risk aking Global P e e ence Su ey: Two
i ems o measu e his
p e e ence. 1)Sel -assessmen :
willingness o ake isks in
gene al 2) Lo e y choice
sequence using s ai case
me hod
−0.058 0.152
Nega i e ecip oci y Global P e e ence Su ey:
Th ee i ems o measu e his
p e e ence. 1)Sel -assessmen :
willingness o ake e enge. 2)
0.034 0.226
Table 2 (con inued)
Va iable Desc ip ion Mean S d.
De .
Sel -assessmen : willingness o
punish un ai beha io owa d
sel . 3)Sel -assessmen :
willingness o punish un ai
beha io owa d o he s
Al uism Global P e e ence Su ey: Two
i ems o measu e his
p e e ence .1) Dona ion
decision. 2)Sel -assessmen :
willingness o gi e o good
causes
−0.040 0.280
T us Global P e e ence Su ey: 1)
Sel -assessmen : people ha e
only he bes in en ions
0.023 0.226
Va iables om he Wo ld Values Su ey
Le - igh index Index ha measu es he
poli ical o ien a ion. I anges
om 1 (le ) o 10( igh )
5.510 0.440
COVID-19 Va iables
Dea hs COVID-19 Numbe o dea hs eco ded as a
consequence o COVID-19/
1000
0.996 4.795
M.L. Lou ei o and M. All´
o
En i onmen al Science and Policy 124 (2021) 451–460
457
ha coun ies su e ing wild i es signi ican ly inc ease he numbe o
wee s abou clima e change (also loods o he Robus OLS eg ession),
whe eas d ough s and olcanic ac i i y ha e a nega i e impac on his
ype o communica ion. The media a en ion caused by wild i es,
oge he wi h he immedia e impac s o hese e en s on socie y, clea ly
show immedia e eac ions in Twi e ; while in he case o d augh s, he
associa ed e ec s may become ele an a e se e al weeks o mon hs o
he popula ion. In addi ion o p io expe iences wi h ca as ophic
e en s, we also ind ha social no ms and cul u al a iables help
explaining he numbe o wee s abou clima e change. In pa icula , we
ind e idence ha al uis ic socie ies and hose us ing scien i ic ad ice
a e mo e willing o con ibu e and sha e messages abou clima e change.
We also ind ha he communica ion dynamics changes o e ime,
inc easing almos always a e in e na ional UNFCC COP con e ences
(wi h he excep ion o he one celeb a ed in Bonn 2018), and dec easing
in gene al wi h he COVID-19 epidemic.
When con olling o he cul u al and social-no ms dimensions, we
ind ha ci izens om coun ies whose willingness o ake isks is
highe - no e ha he isk- aking measu e is nega i e- a e less likely o
wee abou clima e change. This may sugges ha a e ing beha io
may be a common p ecau ion aken by coun ies ha a e mo e isk
ad e se, and in o ma ion ansmission may encou age a e ing
beha io . Ci izens om coun ies wi h highe scales on al uism and
us a e also mo e willing o wee and sha e in o ma ion abou clima e
change. These indings a e in ui i e since hey may be mo e willing o
sha e hei expe iences, bu also hei neighbo s’ expe iences in e ms o
clima e impac s. On he con a y, and in line wi h p e ious li e a u e,
coun ies wi h igh wing poli ical o ien a ions a e less willing o wee
abou clima e change. In summa y, we ind ha social no ms ma e ,
and help shaping he clima e deba e.
When analyzing he e ec o COVID-19 on wee ing abou clima e
change (Table 4), esul s show a nega i e impac o he numbe o dea hs
as a consequence o COVID-19, implying ha since he pandemic s a -
ed, he numbe o con e sa ions abou clima e change has dec eased.
Looking a coun y-speci ic e ec s, we ind ha o mos coun ies, he
end o clima e ela ed wee s is nega i e o e he pandemic pe iod,
wi h in e es ing excep ions. Coun ies such as I aly, F ance, Finland, he
USA and he UK ha e shown a posi i e end in con e sa ions o clima e
change- ela ed opics a e COVID-19, whe eas o he La in Ame ican
coun ies (such as A gen ina, Chile, Colombia) and Asian (Indonesia)
bu also o he Eu opean coun ies (Sweden, Swi ze land and Aus ia)
ha e expe ienced he opposi e e ec . This may show ha al hough
economic g ow h has been comp omised qui e se e ely wo ldwide, he
e ec s o his impac on inequali y and po e y a e much mo e acu e in
low and mid-income coun ies. In hese a eas, he need o es o e
adi ional economic ac i i ies is a he u gen , and as a consequence,
economic impac s ela ed o clima e change become ela i ely less
impo an in he scope o daily con e sa ions. On he con a y, Twi e
con e sa ions om weal hie coun ies, in gene al, ha e shown a mo e
posi i e ou look abou clima e change mi iga ion and adap a ion pol-
icies, making he clima e deba e mo e p edominan du ing he COVID-
19 pandemic. This shows ha o de eloped coun ies, he e is a clea
complemen a i y in he discussion o bo h opics; and no necessa ily, a
subs i u ion be ween opics (clima e change deba e o COVID deba e) as
in mid-income and low-income coun ies.
5. Conclusions
In his pape we in es iga e he ela ionship be ween con ibu ing o
he clima e change deba e on Twi e and he occu ence o he COVID-
19 pandemic, while conside ing socio-economic ac o s, poli ical p e -
e ences and aspec s ela ed o social no ms. Ou esul s con i m ou
main esea ch hypo heses, showing ha he public deba e abou clima e
change has been slowed down in many coun ies a e he pandemic.
This is a oubling si ua ion, pa icula ly because some coun ies whe e
he clima e policies ha e been educed a e de eloping coun ies and a e
he e o e mo e ulne able o clima e impac s. We also ind e idence
abou he impac o cul u al a iables and social no ms in he commu-
nica ion o clima e change.
Ou esul s also show in e es ing in e na ional di e ences in e ms o
Table 3
Baseline esul s.
Baseline Robus OLS Baseline Tobi
Ln wee s Coe . Robus S d. E . P>| | Coe . S d. E . P>| |
Socioeconomic cha ac e is ics
Women 0.003 0.011 0.802 0.004 0.007 0.535
GDP_In e ne use s 60.031 3.089 0.000 57.893 3.153 0.000
Educa ion expendi u e 0.293 0.060 0.000 0.276 0.049 0.000
Spanish-speaking coun ies 3.716 0.112 0.000 3.647 0.118 0.000
Expe ience wi h clima e change
D ough −1.293 0.456 0.005 −1.356 0.462 0.003
Ea hquake −0.209 0.257 0.416 −0.133 0.278 0.632
Ex eme empe a u es 0.242 0.220 0.271 0.230 0.229 0.315
Flood 0.061 0.086 0.481 0.035 0.110 0.753
Landslide 0.045 0.333 0.892 −0.092 0.484 0.850
S o m 0.105 0.113 0.353 0.105 0.123 0.393
Volcanic ac i i y −0.674 0.518 0.193 −0.441 0.365 0.227
Wild i es 0.789 0.217 0.000 0.780 0.300 0.010
CCmee ing_ap il_Bonn18 −1.751 0.257 0.000 −1.707 0.260 0.000
CCmee ing_sep _Bangkok18 0.575 0.185 0.002 0.540 0.252 0.032
CCmee ing_dec_Ka owice18 0.671 0.199 0.001 0.630 0.255 0.014
CCmee ing_june_Bonn19 0.133 0.184 0.470 0.121 0.256 0.636
CCmee ing_dec_Mad id19 0.739 0.211 0.000 0.736 0.253 0.004
Cons an −2.178 0.339 0.000 −1.982 0.336 0.000
Sigma 1.390 0.074
N 759 759
Roo MSE 1.238
F 84.060
P ob >F 0.000
R-squa ed 0.584
Loglikelihood −1191.831
LRChi2 674.900
P ob >Chi2 0.000
Pseudo R2 0.221
M.L. Lou ei o and M. All´
o
En i onmen al Science and Policy 124 (2021) 451–460
458
he ela ionship be ween COVID-19 and clima e change con e sa ions
on Twi e . Con e sa ions abou clima e change inc eased a e COVID-
19 in he case o Aus alia, he UK, and F ance, while hey dec eased in
mos La in Ame ican coun ies, and o he coun ies wi h a less in e -
en ionis app oaches o COVID-19 policies (Aus ia and Sweden, o
example). Ou esul s esemble he ic ions be ween clima e change
adap a ion policies and pos COVID-19 accele a ion policies, which ha e
been clea ly demons a ed in some coun ies. Ne e heless, we also
show ha hese con adic ions be ween economic g ow h and sus ain-
able eco e y a e no uni e sal. Clima e and COVID-19 conce ns seem o
be mo e closely aligned in de eloped coun ies, while mo e ensions and
con adic ions seem o be eme ging in de eloping coun ies.
In his con ex , awa eness and in o ma ion campaigns abou he
linkages be ween bo h c ises may be qui e ele an o build esilience
agains such global public bads in he nea u u e. Fu u e esea ch may
conside he ela ionship and e olu ion o Twee s conce ning bo h
opics. In summa y, we belie e ha unde s anding he ole played by
social no ms and cul u al dimensions is c ucial o o m a global and
cohe en s a e o opinion abou in e na ional opics ha a e becoming
impo an h ea s o he u u e o Humani y. I is only hen ha we
would be able o design success ul and meaning ul con ol policies
wo ldwide.
Table 4
Ex ended esul s.
Ex ended Robus OLS Ex ended Tobi
Ln wee s Coe . Robus S d. E . P>| | Coe . S d. E . P>| |
Socioeconomic cha ac e is ics
Women −0.023 0.012 0.064 −0.018 0.008 0.020
GDP_In e ne use s 42.005 5.241 0.000 39.799 4.482 0.000
Educa ion expendi u e 0.444 0.097 0.000 0.393 0.069 0.000
Spanish-speaking coun ies 3.922 0.220 0.000 3.815 0.180 0.000
Expe ience wi h clima e change
D ough −1.651 0.666 0.013 −1.743 0.501 0.001
Ea hquake 0.324 0.213 0.129 0.347 0.261 0.184
Ex eme empe a u es 0.261 0.221 0.238 0.257 0.211 0.225
Flood 0.162 0.080 0.045 0.146 0.106 0.172
Landslide 0.226 0.319 0.480 0.015 0.443 0.973
S o m 0.174 0.114 0.127 0.182 0.115 0.114
Volcanic ac i i y −0.917 0.506 0.071 −0.646 0.352 0.067
Wild i es 0.553 0.204 0.007 0.539 0.278 0.053
CCmee ing_ap il_Bonn18 −1.737 0.238 0.000 −1.776 0.253 0.000
CCmee ing_sep _Bangkok18 0.532 0.183 0.004 0.496 0.242 0.041
CCmee ing_dec_Ka owice18 0.718 0.201 0.000 0.673 0.246 0.006
CCmee ing_june_Bonn19 0.107 0.185 0.564 0.094 0.247 0.705
CCmee ing_dec_Mad id19 0.742 0.199 0.000 0.745 0.244 0.002
Risk, ime and social ac o s (+)
Risk aking 1.324 0.464 0.004 1.315 0.387 0.001
Al uism 1.118 0.209 0.000 1.020 0.217 0.000
T us 0.369 0.278 0.185 0.481 0.265 0.070
Nega i e ecip oci y −0.091 0.253 0.718 −0.157 0.252 0.535
Le - igh index −0.321 0.110 0.004 −0.303 0.117 0.010
Dea hs COVID-19 −0.054 0.009 0.000 −0.055 0.018 0.003
Dea hs*Spain 0.028 0.018 0.124 0.027 0.057 0.634
Dea hs*F ance 0.085 0.013 0.000 0.086 0.050 0.087
Dea hs*UK 0.146 0.025 0.000 0.145 0.041 0.001
Dea hs*A gen ina −0.171 0.036 0.000 −0.167 0.119 0.163
Dea hs*I aly 0.071 0.014 0.000 0.073 0.053 0.165
Dea hs*Mexico 0.003 0.016 0.837 0.003 0.037 0.931
Dea hs*Aus alia 1.419 0.871 0.104 1.251 2.148 0.561
Dea hs*Aus ia −2.273 1.161 0.051 −2.226 2.016 0.270
Dea hs*Chile −0.325 0.113 0.004 −0.320 0.167 0.057
Dea hs*Colombia −0.170 0.046 0.000 −0.172 0.084 0.040
Dea hs*Finland 3.036 1.252 0.016 3.075 4.532 0.498
Dea hs*Indonesia −0.576 0.086 0.000 −0.929 0.331 0.005
Dea hs*Japan −0.341 0.920 0.711 −0.693 1.576 0.660
Dea hs*Ne he lands 0.062 0.060 0.303 0.060 0.242 0.804
Dea hs*Poland −1.612 0.973 0.098 −1.824 1.088 0.094
Dea hs*Sweden −0.574 0.263 0.030 −0.576 0.319 0.072
Dea hs*Ge many −0.016 0.057 0.773 −0.019 0.165 0.910
Dea hs*Swi ze land −0.486 0.231 0.036 −0.459 0.845 0.587
Dea hs*USA 0.048 0.010 0.000 0.050 0.022 0.024
Cons an 0.501 1.036 0.629 0.706 0.888 0.427
Sigma 1.111 0.063
N 660 660
Roo MSE 1.120
F 141.280
P ob >F 0.000
R-squa ed 0.671
Loglikelihood −961.836
LRChi2 742.1
P ob >Chi2 0
Pseudo R2 0.2784
(+) Dea hs*B azil is he omi ed c oss-p oduc .
M.L. Lou ei o and M. All´
o
En i onmen al Science and Policy 124 (2021) 451–460
459
CRediT au ho ship con ibu ion s a emen
Ma ia L. Lou ei o: Concep ualiza ion, Me hodology, Funding,
W i ing-Re iewing and Edi ing. Ma ia Allo: Da a cu a ion, Es ima ion
and W i ing.
Decla a ion o Compe ing In e es
The au ho s epo no decla a ions o in e es .
Acknowledgmen s
Bo h au ho s hank he edi o (James Bu le ) and he anonymous
e iewe s o hei commen s and sugges ions.
Au ho s acknowledge inancial suppo om Spanish Agency o
Resea ch (Agencia Es a al de In es igaci´
on). G an numbe “PID2019-
111255RB-100”.
Appendix A
Coun y Numbe o
wee s
Coun y Numbe o
wee s
Coun y Numbe o
wee s
Coun y Numbe o
wee s
Coun y Numbe o
wee s
Uni ed S a es 12,100,000 Singapo e 62,219 Aze baijan 11,935 Sie a Leone 4255 Anguilla 691
Spain 5,565,081 Ghana 57,239 Sudan 11,275 Macedonia 4007 Macau 626
Uni ed
Kingdom
3,731,300 Came oon 53,988 Bahamas, The 11,271 Ando a 4004 Mon se a 595
A gen ina 3,128,437 Uganda 53,308 Somalia 10,795 Madagasca 3981 E i ea 590
Mexico 3,038,144 Sou h Ko ea 52,017 C oa ia 10,581 Bu kina Faso 3913 Dominica 586
Chile 2,548,771 Czech Republic 50,747 Slo akia 10,314 Man, Isle o 3895 S . Vincen and he
G enadines
464
Canada 2,332,737 G eece 49,653 Iceland 9942 Kazakhs an 3687 S alba d 448
Aus alia 1,766,554 Hai i 49,087 Papua New
Guinea
9895 New Caledonia 3530 Be muda 445
Colombia 1,725,805 Denma k 48,567 Slo enia 9780 Togo 3516 F ench Guiana 436
Venezuela 1,370,689 Uni ed A ab
Emi a es
47,993 Oman 9295 Gib al a 3372 Sou h Geo gia and
he Sou h Sandwich
Is
402
India 1,139,674 Sou h A ica 47,587 Namibia 9053 Seychelles 3038 Como os 255
Ecuado 702,154 Tu key 46,362 Jo dan 9014 F ench Polynesia 2912 Sao Tome and
P incipe
219
F ance 617,979 Poland 45,623 Cambodia 8936 Gabon 2814 San Ma ino 197
Ge many 592,990 Bangladesh 43,641 Rwanda 8897 Ne he lands
An illes
2787 Cook Islands 195
Pe u 385,123 Tanzania, Uni ed
Republic o
39,984 Malawi 8827 Gue nsey 2729 Pi cai n Islands 147
Pue o Rico 361,334 Nepal 35,053 A ghanis an 8793 Uzbekis an 2653 Fede a ed S a es o
Mic onesia
141
I aly 295,924 Saudi A abia 34,104 Alge ia 8793 B unei 2603 Ki iba i 118
Pa aguay 294,500 Uk aine 32,566 Belize 8442 Guyana 2603
I eland 266,055 Romania 30,398 Tunisia 8430 G eenland 2517
U uguay 245,844 Is ael 28,665 Byela us 8297 Nige 2301
B azil 243,768 Zimbabwe 28,281 Zai e 7957 S . Lucia 2199
Kenya 223,761 I an 28,080 Sy ia 7950 A uba 2018
Panama 217,152 Bosnia and
He zego ina
26,471 Es onia 7833 Bu undi 1946
Belgium 200,155 Vie nam 26,106 Yemen 7743 Su iname 1920
Nige ia 178,432 Luxembou g 26,067 La ia 7629 Guadeloupe 1836
Philippines 175,570 Jamaica 25,712 Mal a 7450 Congo 1730
Dominican
Republic
172,428 S i Lanka 25,695 Mongolia 7401 Bhu an 1685
Pakis an 171,872 Mo occo 23,147 Je sey 7263 Guam 1675
Boli ia 165,326 Egyp 23,103 Ba bados 7251 Cen al A ican
Republic
1631
New Zealand 157,839 Taiwan 21,927 Li huania 6947 Fiji 1495
Japan 156,723 Hunga y 21,490 Mau i ius 6677 Reuni´
on 1414
Thailand 149,290 G enada 20,904 Bo swana 6625 S . Ki s and Ne is 1365
El Sal ado 146,910 Se bia 20,087 Kuwai 6541 An igua and
Ba buda
1357
Gua emala 142,810 I aq 18,332 No h Ko ea 5795 Tajikis an 1209
Cuba 134,951 Ky gyzs an 17,194 A menia 5620 Equa o ial Guinea 1204
Sweden 132,192 Qa a 17,003 Libya 5523 Nau u 1204
Indonesia 128,457 Laos 16,005 Tu kmenis an 5354 Falkland Islands
(Islas Mal inas)
1144
Hondu as 126,878 Angola 15,502 Senegal 5322 Guinea-Bissau 1117
Swi ze land 124,667 Geo gia 15,381 Mali 5147 Ame ican Samoa 1061
Cos a Rica 122,267 Zambia 14,908 Mozambique 5077 Mau i ania 1054
Malaysia 105,041 E hiopia 14,639 Guinea 4916 Swaziland 1052
Ne he lands 100,830 T inidad and
Tobago
14,346 Albania 4812 No he n Ma iana
Islands
958
Russia 100,251 I o y Coas 13,991 Bah ain 4768 Cayman Islands 936
Po ugal 97,832 Libe ia 13,022 Mon eneg o 4643 Gaza S ip 870
(con inued on nex page)
M.L. Lou ei o and M. All´
o