Academic Edi o s: Concha Pé ez
Cu iel and Dolo s Palau-Sampio
Recei ed: 14 Ma ch 2025
Re ised: 11 May 2025
Accep ed: 25 June 2025
Published: 26 June 2025
Ci a ion: Peña-Fe nández, Simón,
Aina a La ondo-U e a, and Jo di
Mo ales-i-G as. 2025. S addling Two
Pla o ms: F om Twi e o Mas odon,
an Analysis o he E olu ion o an
Un inished Social Media Mig a ion.
Social Sciences 14: 402. h ps://
doi.o g/10.3390/socsci14070402
Copy igh : © 2025 by he au ho s.
Licensee MDPI, Basel, Swi ze land.
This a icle is an open access a icle
dis ibu ed unde he e ms and
condi ions o he C ea i e Commons
A ibu ion (CC BY) license
(h ps://c ea i ecommons.o g/
licenses/by/4.0/).
A icle
S addling Two Pla o ms: F om Twi e o Mas odon, an
Analysis o he E olu ion o an Un inished Social
Media Mig a ion
Simón Peña-Fe nández * , Aina a La ondo-U e a and Jo di Mo ales-i-G as
Depa men o Jou nalism, School o Social and Communica ion Sciences, Uni e si y o he Basque
Coun y (UPV/EHU), 48940 Leioa, Spain; [email p o ec ed] (A.L.-U.); jo [email p o ec ed] (J.M.-i.-G.)
*Co espondence: [email p o ec ed]
Abs ac
Social media ha e been undamen al in he daily li es o millions o people, bu hey ha e
aised conce ns abou con en mode a ion policies, he managemen o pe sonal da a, and
hei comme cial exploi a ion. The acquisi ion o Twi e (now X) by Elon Musk in 2022
gene a ed conce ns among Twi e use s ega ding changes in he pla o m’s di ec ion,
p omp ing a mig a ion campaign by some use g oups o he ede a ed ne wo k Mas odon.
This s udy e iews he onboa ding o use s o his decen alised pla o m be ween 2016 and
2022 and analyses he mig a ion o 19,000 use s who iden i ied hemsel es as suppo e s
o he pla o m swi ch. The esul s show ha he mig a ion campaign was a eac i e
esponse o Elon Musk’s acquisi ion o Twi e and was led by a g oup o highly ac i e
academics, scien is s, and jou nalis s. Howe e , a comple e ansi ion was no ealised, as
use s p e e ed o s addle hei p esence on bo h pla o ms. Mas odon’s decen alisa ion
made i di icul o exac ly eplica e Twi e ’s communi ies, esul ing in a pa ial loss o
hese use s’ social capi al and g ea e agmen a ion o hese use communi ies, which
highligh s he in insic di e ences be ween bo h pla o ms.
Keywo ds: Mas odon; Twi e ; mig a ion; social media; social ne wo ks; use be-
ha iou ; pla o m
1. In oduc ion
Since hei eme gence in he mid-2000s, social media ha e become an in eg al pa o
daily li e o millions o people (Pew Resea ch Cen e 2024). Thei mul idimensional na u e
has allowed hem o quickly adap o a ious en i onmen s (pe sonal and social, public
and p i a e, wo k and leisu e, e c.), leading o hei apid sp ead.
Social media a e mee ing poin s wi h la ge audiences ha enable he quick publica ion
o con en and i s dis ibu ion h ough ne wo ks c ea ed by he use s hemsel es, who
can, in u n, in e ac wi h con en c ea ed by o he s by ecommending, edis ibu ing, o
commen ing on i (Boyd and Ellison 2007;Kup e schmid 2022).
In hei de elopmen , he mos success ul pla o ms ha e been cen alised ne wo ks,
meaning hose ha a e owned by a single company (Al-Kha eeb 2022), whe e all asks—
iden i y, au hen ica ion, da a s o age, add essing, go e nance, con en mode a ion, e c.—
a e managed by a single p o ide (Abbing e al. 2023).
Howe e , a he same ime ha he gene al public has emb aced he use o mains eam
social media, hei da a managemen policies ha e also aised conce ns among use s.
Soc. Sci. 2025,14, 402 h ps://doi.o g/10.3390/socsci14070402
Soc. Sci. 2025,14, 402 2 o 15
Scandals like he Camb idge Analy ica case led many use s o exp ess conce n abou he
p i acy o hei da a and how i is used (Guidi e al. 2020).
This exploi a ion o da a o comme cial pu poses, passi ely accep ed by use s, has
maximised a ge ed ads and con en pe sonalisa ion, also con ibu ing o he c ea ion o
in o ma ion diso de s such as il e bubbles o echo chambe s (La Ca a e al. 2022a). Many
use s also began o pe cei e ha social media had mo ed beyond hei o iginal pu pose as
in e connec ion pla o ms and we e dedica ed o comme cially exploi ing hei connec ions.
All o his was su ounded by a hea ed deba e abou oxici y in social media (Boyle e al.
2021), wi h highly pola ised posi ions be ween hose who eel ha no enough is being
done o p o ec use s om ex eme o undesi able con en and, on he o he hand, hose
who belie e hese policies limi eedom o exp ession (Ande son 2017).
Fo use s seeking al e na i es o Twi e , Mas odon ep esen s an appealing op ion.
Launched in Oc obe 2016, his mic oblogging pla o m is a ede a ed social ne wo k ha
os e s di e si y in use s, small bu in e connec ed communi ies, and use owne ship o
hei da a (Zignani e al. 2018;Shaw 2020). This ype o se e ede a ion, commonly e-
e ed o as he Fedi e se, is based in he in e connec ion o independen se e s (ins ances)
ha sha e common p o ocols, which allows o he exchange o in o ma ion ac oss di e en
ins ances ega dless o he so wa e being used on each (Raman e al. 2019;La Ca a e al.
2021;Láza o-Rod íguez 2024).
Mas odon allows use s o w i e messages ( oo s) up o 500 cha ac e s long and in e ac
wi h hem in a e y simila way o Twi e , by epos ing, sha ing, and mo e. As o he
decen alised pla o ms, Mas odon is buil on web s anda ds and open-sou ce ools (La
Ca a e al. 2022b). In a way, hey aim o eclaim he o iginal spi i ha d o e he in e ne as
a decen alised ne wo k (Boyle e al. 2021;B embs e al. 2023;O ihuela 2023).
F om he pe spec i e o p i acy p o ec ion, Mas odon o e s i s use s much mo e
con ol (Zulli e al. 2020;Wang 2021). In con as o o he pla o ms, da a a e only s o ed on
he se e whe e he accoun esides. Addi ionally, due o i s ede a ed na u e, swi ching
ins ances is s aigh o wa d o hose who disag ee wi h he policies o he se e hey
a e subsc ibed o (Shaw 2020). This ensu es g ea e anspa ency and elimina es he use
o ecommenda ion o mone isa ion sys ems ha would o he wise in luence use s in a
condi ioned way. Like o he decen alised ini ia i es, Mas odon is non-comme cial and
ad- ee, meaning i is no d i en by p o i bu a he suppo ed h ough olun a y wo k,
dona ions, and sponso ships (Lase e al. 2023).
Howe e , as a social media, Mas odon also has some limi a ions (B aun 2023). On
one hand, i is di icul o ha e a global iew o he ne wo k’s con en , as use s can only
ead public messages sha ed by o he use s wi hin hei own ins ance and he con en
pos ed by use s hey ollow om o he ins ances, bu hey canno access he en i e se o
messages published (Shaw 2020). On he o he hand, his e y ede a ed na u e can make
he e adica ion o ques ionable con en a di icul ask, as i depends on he willingness o
he se e adminis a o s hos ing i (Abbing e al. 2023).
In he same ein, al hough he di e si y be ween se e s is high, he ede a ed na u e
also seems o p oduce, in some cases, a ce ain insula i y, whe e ela ionships occu be ween
use s o he same ins ance o om ins ances in he same coun y (Zignani e al. 2019;E z
2022). This can lead o a ac u ed pe cep ion o eali y, whe e use s in e ac in silos wi h
people who sha e hei as es and opinions (Shaw 2020). Despi e i s ede a ed na u e, use s
coming om comme cial social media end o ins inc i ely sign up o he la ges se e
(mas odon.social). This e ec esul s in 52% o use s being egis e ed on only 10% o he
se e s (Raman e al. 2019).
Wi h a s uc u e and objec i es so di e en om hose o comme cial pla o ms,
Mas odon had exis ed since i s c ea ion in 2016 as a pe iphe al and esidual al e na i e o
Soc. Sci. 2025,14, 402 3 o 15
Twi e . Howe e , i s oppo uni y as a eal al e na i e o Twi e came wi h he umo s,
and he subsequen con i ma ion, o Elon Musk’s acquisi ion o Twi e . I was hen ha
some use s launched a campaign ad oca ing o mig a ion, using hash ags like #Good-
byeTwi e and #Twi e Mig a ion, o including hei Mas odon accoun de ails in hei
p o ile desc ip ions.
Among hei mo i a ions was he accumula ed de e io a ion o hei ela ionship
wi h Twi e , which led long-es ablished communi ies o become disillusioned and seek
al e na i es (Swogge 2023). A mo e di ec in luence came om Musk’s own s a emen s
ega ding his new mode a ion policies and he po en ial impac o cu s and layo s on he
global con e sa ion as well as he in oduc ion o paid accoun e i ica ion (Kup e schmid
2022), ques ions o which Mas odon seemed a na u al al e na i e (Lee and Wang 2023;
Sabo and Ges huizen 2024).
In his con ex , he main objec i e o his s udy is o analyse he mig a ion o use s
om Twi e o Mas odon and, in pa icula , he impac o hei ne wo ks.
Based on his gene al objec i e, he ollowing speci ic objec i es a e p oposed:
RQ1. Wha has been he empo al e olu ion in he onboa ding o use s o Mas odon,
and wha ha e been i s main miles ones?
RQ2. Is he e any co ela ion be ween he change in owne ship o Twi e and he
mig a ion o Mas odon?
RQ3. Wha a e he cha ac e is ics o he communi ies o use s wi h accoun s on bo h
social media?
RQ4. Wha is he impac o mig a ion on use s’ ne wo ks? Can we speak o a mig a ion
o i ual communi ies?
2. Theo e ical F amewo k
The s udy o mig a ion on social media has a b oad academic ajec o y. F om he
ea lies in es iga ions in o use s’ mo emen s be ween pla o ms, he loss o econ igu a ion
o digi al social capi al, o he e ec s o echnological inno a ions and mode a ion policies,
he li e a u e has add essed hese phenomena as deeply social p ocesses, going beyond
me e shi s in echnical in as uc u e.
In hei adap a ion o he push–pull–moo ing (PPM) heo y, Bansal e al. (2005) ans-
e ed his mig a o y amewo k, o iginally de eloped in human geog aphy, o he con ex
o consume beha iou , speci ically o analysing se ice p o ide swi ching in digi al
en i onmen s. The push–pull–moo ing model explains swi ching beha iou h ough h ee
ypes o ac o s: push ac o s, which d i e use s away om hei cu en p o ide ; pull
ac o s, which a ac hem o al e na i es; and moo ings, which ei he acili a e o inhibi
change. Acco ding o his model, e en when use s a e dissa is ied o pe cei e a ac i e
al e na i es, s ong moo ings—such as swi ching cos s o es ablished habi s—can play a
decisi e ole in main aining hei loyal y.
The PPM model has been widely adop ed o explain mig a ions wi hin he digi al
en i onmen (Handa kho and Ha josepu o 2020;Lin e al. 2021;Xia e al. 2023), and in
all o hese cases, moo ing e ec s play a c ucial ole. Fo example, ine ia can be one o
he main ac o s de e mining whe he use s ul ima ely decide o s ay, e en when push
and pull o ces poin owa ds change. I se es as a s abilising mechanism, ampli ying he
in luence o emo ional a achmen , pe cei ed swi ching cos s, and habi ual beha iou s,
he eby educing he impac o dissa is ac ion o a ac i e al e na i es (Sun e al. 2017).
The PPM model has also been applied speci ically o he case o Twi e . Jeong e al.
(2024a) highligh ha ac o s such as dissa is ac ion wi h go e nance (push), a ac ion o
decen alised s uc u es (pull), and use s’ p io social capi al and habi s (moo ing) in e ac
o shape digi al mig a ion ajec o ies.
Soc. Sci. 2025,14, 402 4 o 15
In line wi h he s ong in luence o moo ing ac o s in he PPM heo y, Solomonik
and Heue (2025) poin ou ha mig a ion ac oss social pla o ms is d i en mo e by
social, emo ional, and con ex ual ac o s han by echnical unc ions. Pee popula i y, local
pe cep ions o sa e y, and he alignmen o pla o ms wi h li e s ages all in luence decisions.
Acco ding o hese au ho s, such mig a ions e lec pe sonal ajec o ies a he han pu ely
a ional choices.
S anding ou in hese mig a ion p ocesses a e use s who hold a p ominen leade ship
posi ion. As Kuma e al. (2011) no e, use s mos likely o mig a e be ween social ne wo ks
a e hose who display high le els o ac i i y, main ain ex ensi e social ne wo ks (i.e., many
iends o ollowe s), and enjoy high ex e nal isibili y. These use s, e e ed o as high
ne -wo h indi iduals (HNWI), a e pa icula ly in luen ial wi hin he pla o m ecosys em
and o en se mig a ion ends. Mo eo e , when hese use s become inac i e, hey end
o p ecede b oade wa es o abandonmen , making hei beha iou a use ul indica o o
p edic ing shi s in a pla o m’s popula i y.
The di usion o inno a ions heo y (Roge s 2003), o i s pa , explains digi al mig a-
ion no only h ough he in luence o ea ly adop e s, bu also h ough pe cei ed ela i e
ad an age, compa ibili y wi h p io usage, echnical complexi y, ialabili y, obse abili y
o bene i s, and social p essu e—all o which shape he adop ion o new pla o ms. These
ac o s mean ha inno a ion does no occu uni o mly bu a he g adually and une enly,
shaped by communica i e, social, and cul u al in luences. The p ocess depends no only on
he cha ac e is ics o he inno a ion i sel bu also on he social con ex in which i ci cula es.
3. Ma e ials and Me hods
To conduc his analysis, a p elimina y lis o o e 60,000 Twi e use s was compiled,
all o whom s a ed on hei Twi e p o iles ha hey also had a Mas odon accoun , ei he in
hei use name o p o ile desc ip ion. This iden i ica ion was ca ied ou by sea ching he
Twi e academic API a he end o 2022 o use s on Twi e who sha ed hei Mas odon
p o ile. The iden i ica ion p ocess was conduc ed using a sea ch on he Twi e Academic
API 1.1 in No embe 2022. A e analysing he selec ed da a, a o al o 19,919 use s wi h
accoun s on bo h pla o ms we e iden i ied.
Fi s , he da a ega ding he mig a ion p ocess om one pla o m o ano he (RQ1)
we e analysed, speci ically, he egis a ion da e o he selec ed use s on Mas odon du ing
he pe iod be ween 1 Oc obe 2016 and 2 No embe 2022.
Secondly, o y o es ablish he co ela ion be ween he change in Twi e ’s owne ship
and mig a ion o Mas odon (RQ2), he e olu ion o egis a ion da a on his pla o m was
compa ed wi h he e olu ion o wee s con aining he wo ds “Musk” and “Twi e ” ha
had achie ed 1000 e wee s du ing he same analysis pe iod.
Fo he analysis o he communi ies exis ing on bo h social media (RQ3), he ollowing
ela ionships o he use s analysed we e ob ained om Twi e using API 1.1. In con as ,
on Mas odon, al hough i is echnically possible o e ie e he da a h ough he o icial
API (Schoch and Chan 2023;Bin Zia e al. 2024), he ela ional da a we e collec ed using
web sc aping echniques, pa sing public HTML om use p o iles wi h s anda d Py hon
lib a ies like Reques s ( . 2.29.0) and Beau i ulSoup ( . 4.12.2). This app oach was necessa y
due o di icul ies in consis en ly ob aining comp ehensi e ollowe g aphs ac oss he
ede a ed ne wo k o ou speci ic use se ia he API a he ime o collec ion. Conse-
quen ly, his da a collec ion ocuses on publicly a ailable ela ional in o ma ion, excluding
connec ions in ol ing p i a e ins ances o p o ec ed accoun s. These collec ed ela ional
da a we e used o cons uc ollowe g aphs o bo h pla o ms. We hen employed Social
Ne wo k Analysis (SNA) using he Py hon lib a y Ne wo kX ( . 3.1) and Gephi ( . 0.10.1)
o calcula e key ne wo k me ics (including densi y, clus e ing coe icien , modula i y,
Soc. Sci. 2025,14, 402 5 o 15
cen alisa ion, a e age deg ee, and numbe o communi ies, as de ailed in Table 1) and o
de ec communi ies using he Lou ain algo i hm ( esolu ion = 1). Ca ego ical asso a i i y
was also calcula ed o assess communi y p ese a ion.
Table 1. Cha ac e is ics o he communi ies on Twi e and Mas odon.
Twi e Followe s Ne wo k Mas odon Followe s Ne wo k
Be o e Oc obe 2022
4.583 nodes
Densi y 0.002 Densi y 0.001
Clus e ing coe icien 0.14 Clus e ing coe icien 0.087
Modula i y 0.524 Modula i y 0.509
No. o communi ies 283 No. o communi ies 732
Indeg ee cen alisa ion 0.135 Indeg ee cen alisa ion 0.187
Ou deg ee cen alisa ion 0.022 Ou deg ee cen alisa ion 0.046
A e age deg ee 8.537 A e age deg ee 5.436
Twi e Communi ies’ asso a i i y 0.667
A e Oc obe 2022
15.336 nodes
Densi y 0.001 Densi y 0.001
Clus e ing coe icien 0.13 Clus e ing coe icien 0.091
Modula i y 0.542 Modula i y 0.528
No. o communi ies 278 No. o communi ies 1141
Indeg ee cen alisa ion 0.137 Indeg ee cen alisa ion 0.149
Ou deg ee cen alisa ion 0.013 Ou deg ee cen alisa ion 0.021
A e age deg ee 15.182 A e age deg ee 9.193
Twi e Communi ies’ asso a i i y 0.584
Sou ce: Au ho ’s own wo k.
Finally, we measu ed whe he he mo e o Mas odon esul ed in abandoning Twi e
in o de o de e mine whe he we can speak o mig a ion o he simul aneous use o bo h
pla o ms (RQ4). To do his, wee s om a sample o oughly 2000 Twi e use s we e
cap u ed—gua an eeing be ween 200 and 300 use s pe clus e o in e es —o e a pe iod
o
±
1 mon h om hei egis a ion da e on Mas odon (60 days in o al) o assess whe he a
change in hei ac i i y le el could be obse ed.
4. Resul s
4.1. E olu ion o Mas odon Sign-Ups and he “Elon Musk E ec ”
The analysis o he Mas odon egis a ions, i s and o emos , shows ha his is a shi
ha occu s a speci ic momen s, d i en by bu s s and in ba ches, a he han g adually and
s eadily o e ime. The analysis o he e olu ion o sign-ups on he pla o m e eals i e
clea ly di e en ia ed peaks (Figu e 1). Al hough Mas odon was launched in Oc obe 2016,
he i s signi ican peak in use egis a ions did no occu un il Ap il 2017. The second
su ge in sign-ups ook place in No embe o ha same yea , when some p ominen igu es
in social media announced hei mo e o his pla o m. Among hem we e a is s, w i e s,
and ech pe sonali ies and en ep eneu s such as Chuck Wendig, John Scalzi, Melanie
Gillman, o John O’Nolan. A signi ican hi d peak occu ed in Augus 2018.
In any case, du ing i s ea ly yea s o exis ence, access o Mas odon could be conside ed
ma ginal compa ed o he numbe o ac i e use s on comme cial social media. Al hough
ini ial su ges o use s we e al eady obse ed as a esul o he i s campaigns o lea e
Twi e , he in lux o use s o Mas odon emained modes o e he nex ou yea s.
Soc. Sci. 2025,14, 402 6 o 15
Figu e 1. E olu ion o Mas odon egis a ions. Sou ce: Au ho ’s own wo k.
Howe e , coinciding wi h he announcemen on 25 Ap il 2022 o Elon Musk’s pu chase
o Twi e , he low o new egis a ions on Mas odon su ged wi h he la ges in lux o use s
o he pla o m since i s c ea ion, and six mon hs la e , a e a hea ed legal con o e sy o e
he e ms o he pu chase, Elon Musk o icially acqui ed Twi e on 28 Oc obe 2022 o
USD 44 billion. A e ha , sign-ups o Mas odon sky ocke ed.
In pa icula , in he las wo cases, a clea co ela ion can be obse ed be ween he
change in Twi e ’s owne ship and he beginning o he mig a ion o Mas odon. Rega dless
o he possible easons specula ed o explain his shi —whe he i is conce ns o e changes
in he pla o m’s policies, ea o losing i s spi i , o ejec ion o he new owne — he
connec ion be ween hese wo e en s is clea . Du ing he pe iod om 25 Ap il o 1 May
2022, 7.27% o he use s we acked we e egis e ed, and he igu e ises o 75.64% in he
pe iod om 27 Oc obe o 22 No embe 2022.
To ein o ce his analysis and isualise he empo al ela ionship, Figu e 2plo s bo h
he daily Mas odon egis a ions om he sampled Twi e use s (le Y-axis) and he daily
coun o high-impac wee s ela ed o ‘Musk’ and ‘Twi e ’ ( igh Y-axis) o e he analysis
pe iod. The igu e illus a es he sha p inc eases in bo h se ies, pa icula ly du ing he key
pe iods in 2022. The s a is ical analysis o bo h a iables shows a mode a e–high posi i e
co ela ion (R = 0.737), meaning hey inc ease oge he . Fu he mo e, he p- alue < 0.0001
indica es ha he esul is highly signi ican , meaning i is e y unlikely ha his co ela ion
occu ed by chance.
Figu e 2. Co ela ion be ween Mas odon egis a ions and ac i i y on Twi e . Sou ce: Au ho ’s
own wo k.
While his s ong posi i e co ela ion indica es a clea empo al associa ion be ween
heigh ened discussion abou Musk’s acquisi ion on Twi e and he su ge in Mas odon
Soc. Sci. 2025,14, 402 7 o 15
sign-ups among hese use s, es ablishing di ec causali y would equi e con olling o
ex e nal ac o s, such as b oade media co e age o p e-exis ing use dissa is ac ion, which
is beyond he scope o his speci ic analysis. None heless, he da a s ongly sugges ha
he mig a ion wa e had a signi ican eac i e componen linked o he change in Twi e ’s
owne ship. When he deba e o e Elon Musk’s pu chase o Twi e in ensi ied wi hin he
social media i sel , he e was an explosion o use egis a ions on Mas odon.
4.2. In e ac ion and Cohesion o he Communi ies and Rela ional Capi al
Rega dless o he easons behind i , he case o Twi e and Mas odon is o pa icula
in e es , as i allows us o cha ac e ise how use and i ual communi y mig a ions occu
on social media.
The shi om one social media o ano he o he adop ion o a new pla o m does
no simply in ol e a pe sonal use egis a ion on he new pla o m. Social media a e he
ools h ough which indi iduals c ea e and pa icipa e in i ual communi ies, a ela ional
capi al hey also wish o ans e o he new pla o m, pa icula ly he mos ac i e ones.
The e o e, Twi e use s who exp essed hei in en ion o mig a e ied o b ing hei
ela ional capi al o Mas odon as well. In o he wo ds, hey sough o ebuild on Mas odon
he i ual communi y hey had p e iously es ablished on Twi e .
We c ea ed wo e sions o he Twi e and Mas odon ollowe ne wo ks, dis inguish-
ing be ween use s who egis e ed on Mas odon be o e o a e Oc obe 2022, esul ing
in ou ne wo ks in o al (Table 1). This allowed us o compa e he s uc u e o he Twi -
e ollowe ne wo k o each g oup wi h i s co esponding Mas odon ne wo k o assess
how he mig a ion a ec ed use s’ connec i i y and communi y s uc u e. Ou analysis
indica es ha he mig a ion om Twi e o Mas odon has been associa ed, bo h be o e
and a e Oc obe 2022, wi h se e al s uc u al changes in he ne wo k, a ec ing cohesion,
agmen a ion, and cen alisa ion.
The densi y o a ne wo k measu es how in e connec ed i s nodes a e by compa ing
he numbe o exis ing connec ions o he maximum possible numbe . A highe densi y in-
dica es a mo e igh ly connec ed ne wo k, while a lowe densi y sugges s a mo e dispe sed
s uc u e. Among use s who egis e ed on Mas odon be o e Oc obe 2022, he Twi e
ne wo k had a densi y o 0.002, while he Mas odon ne wo k showed a lowe alue o
0.001. Fo hose who egis e ed a e Oc obe 2022, he densi y in Twi e was also 0.001,
emaining he same in Mas odon. These alues indica e ha despi e he mig a ion, he
o e all le el o connec i i y among use s did no g ea ly dec ease in Mas odon, sugges -
ing ha while use s e-es ablished hei connec ions, he o e all s uc u e emained e y
simila o Twi e ’s.
Despi e his ini ial insigh , se e al me ics sugges ha Mas odon e-elabo a ed ne -
wo ks a e less cohesi e and mo e agmen ed han o iginally in Twi e . Fo ins ance, he
clus e ing coe icien measu es he likelihood ha a use ’s connec ions a e also connec ed
o each o he , o ming closed iangles. A highe clus e ing coe icien sugges s ha use s
belong o igh ly kni communi ies. Among use s who joined Mas odon be o e Oc obe
2022, he Twi e ne wo k had a clus e ing coe icien o 0.14, whe eas he co esponding
Mas odon ne wo k showed a lowe alue o 0.087. Fo hose who mig a ed a e Oc obe
2022, he Twi e ne wo k’s clus e ing coe icien was sligh ly lowe a 0.13, while he
Mas odon ne wo k had a alue o 0.091. The consis en ly lowe clus e ing coe icien in
Mas odon sugges s ha ega dless o when use s mig a ed, hei ne wo ks became less
cohesi e compa ed o Twi e , wi h ewe s ongly in e connec ed g oups.
In he same ein, he a e age deg ee, which ep esen s he mean numbe o connec-
ions pe use , illus a es di e ences in connec i i y be ween he pla o ms. Among ea ly
adop e s (i.e., be o e Oc obe , 2022), he a e age deg ee in Twi e was 8.537, while in
Soc. Sci. 2025,14, 402 8 o 15
Mas odon, i was lowe a 5.436. Among la e adop e s, he a e age deg ee also dec eased
om 15.182 in Twi e o 9.193 in Mas odon. These alues indica e ha use s who mi-
g a ed la e ended o belong o mo e cohesi e communi ies, bu hei Mas odon ne wo ks
emained less connec ed compa ed o hei co esponding Twi e ne wo ks in any case.
The numbe o communi ies de ec ed wi h he Lou ain algo i hm wi h esolu ion =
1—which ha e yielded e y simila alues o Modula i y in he ou ne wo ks, wi h igu es
anging om 0.509 o 0.542—p o ides addi ional insigh s in o ne wo k agmen a ion.
Among ea ly adop e s, Twi e had 283 communi ies, while Mas odon had mo e han
wice ha numbe , wi h 732. Among la e adop e s, Twi e showed a simila numbe
o communi ies, wi h 278, whe eas in Mas odon, he numbe o communi ies inc eased
sha ply o 1141. This signi ican ise sugges s ha la e adop e s on Mas odon o med
e en mo e agmen ed communi ies han ea ly adop e s, bu he mig a ion om one
medium o he o he implied in bo h cases a dec ease in ne wo k cohesion and an inc ease
in agmen a ion.
On he o he hand, indeg ee and ou deg ee cen alisa ion me ics measu e he ex en
o which ce ain nodes accumula e mo e connec ions han o he s, indica ing hie a chical
s uc u es wi hin he ne wo k. Highe cen alisa ion alues sugges ha a ew use s hold
disp opo iona e in luence o e he ne wo k. Among ea ly adop e s, indeg ee cen alisa-
ion in Twi e was 0.135, while in Mas odon, i was sligh ly highe a 0.187. Among la e
adop e s, his alue inc eased ma ginally, eaching 0.137 in Twi e and 0.149 in Mas odon.
Simila ly, ou deg ee cen alisa ion among ea ly adop e s was 0.022 in Twi e and 0.046 in
Mas odon, while among la e adop e s, i was 0.013 in Twi e and 0.021 in Mas odon. The
consis en ly highe cen alisa ion in Mas odon sugges s ha ega dless o when use s mi-
g a ed, he ne wo k became sligh ly mo e hie a chical, wi h ce ain accoun s accumula ing
a la ge sha e o he connec ions.
An addi ional measu e, Twi e communi ies’ asso a i i y in Mas odon, e alua es
whe he use s who belonged o he same communi y on Twi e end o emain connec ed
in Mas odon. Among ea ly adop e s, his alue was 0.667, while o la e adop e s, i
dec eased sligh ly o 0.584. These alues sugges ha Twi e communi ies we e la gely
p ese ed in Mas odon, al hough he e was a sligh decline in communi y s uc u e among
la e adop e s.
All hese indings sugges ha al hough he mig a ion allowed use s o e ain an
impo an pa o hei ela ional capi al, as sugges ed by me ics such as densi y o ca ego -
ical asso a i i y, he mig a ion om Twi e o Mas odon also led o impo an s uc u al
changes in he ne wo k. The dec ease in densi y and clus e ing coe icien s, combined wi h
he inc ease in he numbe o de ec ed communi ies, indica es ha he Mas odon ne wo k
became less cohesi e and mo e agmen ed compa ed o Twi e . A he same ime, he
inc ease in indeg ee cen alisa ion sugges s ha Mas odon’s ne wo k is no only mo e
decen alised o e all bu also exhibi s some deg ee o hie a chy, whe e ce ain accoun s
hold mo e in luence.
In summa y, he ans e (o ec ea ion) o Twi e ’s i ual communi ies o Mas odon
has esul ed in a signi ican loss o ela ional capi al. I should be no ed ha his s udy
was conduc ed among he mos ac i e and mo i a ed use s and ha he p ocess ook place
in ba ches—meaning i occu ed on a la ge scale wi hin a sho pe iod o ime. Twi e
use s who announce hei p esence on Mas odon ans e pa o hei ela ional capi al o
he new social media, al hough he mig a ion educes he size o he communi y and i s
cohesion while inc easing he o al numbe o i ual communi ies on he new pla o m.
Soc. Sci. 2025,14, 402 9 o 15
4.3. Typology o he Main Mig a ed Communi ies
F om he analysis o he use clus e s common o bo h pla o ms (Figu e 3), i is also
possible o de e mine how he p e-exis ing communi ies om Twi e ha e been ( e)c ea ed
upon mig a ing o Mas odon. This p ocess is de ailed in Table 2and isualised in he Sankey
diag am p esen ed in Figu e 4. Figu e 4illus a es hese communi y ansi ions using a
Sankey diag am. The le side ep esen s he main use communi ies iden i ied on Twi e
(labelled C.X—Twi e ), and he igh side shows he communi ies iden i ied on Mas odon
(labelled C.Y—Mas odon). The wid h o he lows connec ing communi ies indica es he
p opo ion o use s mo ing om a speci ic Twi e communi y o a speci ic Mas odon
communi y. This isualisa ion highligh s bo h he endency o communi ies o eplica e
(e.g., ‘Digi al ac i is s and jou nalis s USA’) and ins ances o eg ouping o agmen a ion
in o new o mul iple Mas odon communi ies (e.g., use s om ‘Scien is s and p o esso s USA’
dis ibu ing ac oss se e al Mas odon g oups). As he da a show, he as majo i y o he
communi ies de ec ed in Mas odon using he Lou ain algo i hm eplica e in a pe cen age
o o e 70% some communi y ha al eady exis ed on Twi e and ha we ha e p oceeded
o label wi h he same name; his is he case o he communi ies Independen jou nalis s
USA, Digi al ac i is s and jou nalis s USA, Uni e si y p o esso s GER, Scien is s and
p o esso s UK, Scien is s and p o esso s USA, Scien is s and p o esso s FRA, Scien is s and
p o esso s BEL and NL, and Scien is s and p o esso s ESP. O he Mas odon communi ies,
such as 6 (Scien is s and p o esso s USA and EUR) o 9 (As onome s and physicis s USA),
eg oup use s who we e in se e al communi ies, wi h Twi e communi y 8 (Scien is s
and p o esso s USA) being he main one in bo h cases. On he o he hand, Mas odon
communi y 13 (Sci. and p o . EUR, CAN, and AUS) comes mainly om Twi e communi y
3 (En i onmen al sci. and p o . USA) bu also eg oups use s om o he communi ies.
Use Clus e s on Twi e Use Clus e s on Mas odon
Figu e 3. Compa ison o clus e s on Twi e and Mas odon. Sou ce: Au ho ’s own wo k. No e:
Colo s ep esen di e en use clus e s in he ne wo k.
Beyond he ac ha he ans e s o mig a ions om Twi e o Mas odon seem o
happen in blocks, meaning ha en i e communi ies mo e and eplica e hei ela ionships
as much as hey can—no wi hou losing a ce ain amoun o ela ional capi al along
he way—i is s iking ha all he communi ies iden i ied a e academics, jou nalis s, and
ac i is s om he US and Eu ope. The e o e, he use s who led he campaign o lea e came
mos ly om he in ellec ual and jou nalis ic ields a he han om o he a eas, like poli ics
o business.
