192 2024, olume 27, issue 1, pp. 192–210, DOI: 10.15240/ ul/001/2024-5-001
In o ma ion Managemen
A compa a i e analysis o mul i a ia e
app oaches o da a analysis in managemen
sciences
Rizwan Raheem Ahmed1, Dalia S eimikiene2, Jus as S eimikis3,
Ind e Siksnely e-Bu kiene4
1 Indus Uni e si y, Facul y o Managemen Sciences, Depa men o Business Adminis a ion, Pakis an, ORCID:
0000-0001-5844-5502, [email p o ec ed];
2 Li huanian Spo s Uni e si y, Li huania, ORCID: 0000-0002-3247-9912, [email p o ec ed] (co esponding
au ho );
3 Li huanian Cen e o Social Sciences, Ins i u e o Economics and Ru al De elopmen , Li huania; Uni e si y
o Economics and Human Science in Wa saw, Facul y o Managemen and Finances, Poland, ORCID: 0000-0003-
2619-3229, [email p o ec ed];
4 Kauno Kolegija Highe Educa ion Ins i u ion, Li huania, ORCID: 0000-0002-0927-4847, [email p o ec ed].
Abs ac : The esea che s use he SEM-based mul i a ia e app oach o analyze he da a
in di e en ields, including managemen sciences and economics. Pa ial leas squa e s uc u al
equa ion modeling (PLS-SEM) and co a iance-based s uc u al equa ion modeling (CB-SEM) a e
powe ul da a analysis echniques. This pape aims o compa e bo h models, hei e iciencies
and de iciencies, me hodologies, p ocedu es, and how o employ he models. The ou comes
o his pape exhibi ed ha he PLS-SEM is a echnique ha combines he s eng hs o s uc u al
equa ion modeling and pa ial leas squa es. I is impe a i e o know ha he PLS-SEM is a powe ul
echnique ha can handle measu emen e o a he highes le els, im and unbalanced da ase s,
and la en a iables. I is bene icial o analyzing ela ionships among la en cons uc s ha may no be
candidly wi nessed and migh no be applied in si ua ions whe e adi ional SEM would be in easible.
Howe e , he CB-SEM app oach is a p ocedu e ha pools he s eng hs o bo h s uc u al equa ion
modeling and con i ma o y ac o analysis. The CB-SEM is a dominan mul i a ia e echnique ha
can g ip mul iple g oups and indica o s; i is bene icial o analyzing ela ionships among la en
a iables and mul iple mani es a iables, which can be di ec ly obse ed. The pape concluded
ha he PLS-SEM is a mo e sui able echnique o analyzing ela ions among la en cons uc s,
gene ally o a small da ase , and he measu emen e o is high. Howe e , he CB-SEM is sui able
o analyzing compound la en and mani es cons uc s, mainly when he goal is o gene alize
esul s o speci ic popula ion subg oups. The PLS-SEM and CB-SEM ha e speci ic e iciencies and
de iciencies ha de e mine which echnique o use depending on esou ce a ailabili y, he esea ch
ques ion, he da ase , and he a ailable ime.
Keywo ds: Pa ial leas squa e-SEM (PLS-SEM), co a iance-based-SEM (CB-SEM), SEM-based
mul i a ia e app oach, mul iple mani es a iables, PLS-SEM s. CB-SEM modeling.
JEL Classi ica ion: C8, C42, C52.
APA S yle Ci a ion: Ahmed, R. R., S eimikiene, D., S eimikis, J., & Siksnely e-Bu kiene, I.
(2024). A compa a i e analysis o mul i a ia e app oaches o da a analysis in managemen
sciences. E&M Economics and Managemen , 27(1), 192–210. h ps://doi.o g/10.15240/
ul/001/2024-5-001.
Ea ly Access Publica ion Da e: Janua y 23, 2024.
193
2024, olume 27, issue 1, pp. 192–210, DOI: 10.15240/ ul/001/2024-5-001
In o ma ion Managemen
In oduc ion
The esea che s use co a iance-based s uc u -
al equa ion modeling (CB-SEM) and pa ial leas
squa e s uc u al equa ion modeling (CB-SEM)
o analyze he da a o complica ed connec-
ions among he la en and mani es cons uc s
(Ahmed e al., 2021; Hai e al., 2022). S ill, he e
a e some i al di e ences be ween he wo mul-
i a ia e echniques; o example, PLS-SEM and
CB-SEM modeling handle collinea i y di e en ly
(Ahmed e al., 2022; Hai e al., 2019; Sa s ed
e al., 2019). Howe e , he PLS-SEM is e y
bene icial o managing da a wi h a high deg ee
o collinea i y because i di ides he da a in o la-
en a iables unco ela ed using he PLS-SEM
me hod (Sa s ed e al., 2022). On he o he
hand, he CB-SEM modeling is ounded on mul-
i a ia e no mali y and necessi a es he da a
o be unco ela ed, as highligh ed by Lu e al.
(2020) and Hai J . e al. (2017). When he da a
is highly co ela ed, CB-SEM may gene a e
un eliable o inconsis en esul s (Becke e al.,
2022; Lega e e al., 2022). Ano he di e ence
is how he models a e es ima ed, as Sa s ed
e al. (2019) demons a e. The PLS-SEM mod-
eling uses a echnique called boo s apping
o es ima e he model pa ame e s. This me hod
can be compu a ionally in ensi e bu allows
o a us wo hy app oxima ion o ac o s
in he p esence o ou lie s and non-no mali y.
The CB-SEM uses maximum likelihood ap-
p oxima ion, which is compu a ionally e icien
bu may no wo k well wi h non-no mal da a o
ou lie s, acco ding o Hai J . e al. (2017) and
Muelle and Hancock (2018). The CB-SEM
bases i s assump ions on he mul i a ia e no -
mali y hypo hesis and demands ha he da a
be unco ela ed (Ahmed e al., 2021; Hai e al.,
2019; Hayes e al., 2017). Fo handling da a
wi h non-no mali y, ou lie s, and missing alues,
he PLS-SEM is no based on dis ibu ional
assump ions and is, he e o e, mo e lexible
(Ringle e al., 2022; Sa s ed & Cheah, 2019).
In ligh o his, i has been p o en by Hwang
e al. (2020), Ringle e al. (2015), and Hai e al.
(2018) ha PLS-SEM is a mo e eliable and
adap able me hod o assessing complex and
co ela ed da a han CB-SEM, which is based
on mul i a ia e no mali y assump ions. Howe -
e , he echnique chosen depends on he goals,
esea ch ques ions, and da ase cha ac e is ics
(Hai e al., 2022). The CB-SEM and PLS-SEM
a e mul i a ia e me hodologies, bu each has
s eng hs and weaknesses.
The PLS-SEM is a s a is ical echnique ha
combines he bene i s o s uc u al equa ion
modeling pa ial and leas squa es o e alua e
complex associa ions be ween la en a i-
ables and obse able da ase s, as highligh ed
by Sa s ed e al. (2019), and Ahmed e al.
(2022). The PLS-SEM is pa icula ly help ul
in handling da a wi h a high deg ee o collinea -
i y since i uses he PLS-SEM echnique o b eak
he da ase down in o unco ela ed la en a i-
ables. The PLS-SEM employs a mo e sui able
pa ame e es ima ion echnique o examining
he model’s pa ame e s in he p esence o ou -
lie s and non-no mali y (Memon e al., 2019).
PLS-SEM has excellen lexibili y because
i does no ely on dis ibu ional assump ions
and can handle da a wi h non-no mali y, ou li-
e s, and missing alues, acco ding o Hai e al.
(2010) and Sa s ed e al. (2021). The PLS-SEM
echnique can es ima e la en a iables ha
symbolize unobse ed o unde lying cons uc s
in he da a (Hai & Sa s ed , 2021; Lega e
e al., 2022). The PLS-SEM echnique pe mi s
he s udy o nume ous g oups/subpopula ions
in he da a, which can help compa e g oups
o measu emen in a iance es s. Acco ding
o Sa s ed e al. (2019) claim, he PLS-SEM can
also handle co ela ions be ween cons uc s ha
a e no linea . The PLS-SEM enables an unde -
s anding o he in e ac ions be ween cons uc s
by p o iding de ails on he in ensi y and di ec-
ion o he associa ions and he compa a i e
signi icance o each cons uc in he conside ed
model (Lega e e al., 2022). The PLS-SEM anal-
ysis can be pe o med using a ious p og ams,
including he Sma -PLS, Wa p-PLS, XLSTAT,
and R packages o PLS-SEM (Memon e al.,
2019; Pa ma e al., 2022). Hence, i can be
supposed ha he PLS-SEM is an e ec i e ech-
nique o e alua ing complex, highly connec ed
da a since i enables he modeling and handling
o non-linea ela ionships in a obus , lexible,
and unde s andable manne (Hai e al., 2014;
Hai e al., 2019).
The CB-SEM is used by Hayes e al. (2017)
and Lu e al. (2020) o examine a complica ed
ela ionship be ween la en cons uc s and
obse able da a. The CB-SEM uses maximum
likelihood app oxima ion o es ima e he model
pa ame e s, which is compu a ionally e -
icien as one o i s essen ial cha ac e is ics
(Ahmed e al., 2022; Hoope e al., 2008).
Gi en ha he CB-SEM echnique is ounded
on he assump ions o mul i a ia e no mali y,
194 2024, olume 27, issue 1, pp. 192–210, DOI: 10.15240/ ul/001/2024-5-001
In o ma ion Managemen
an unco ela ed da ase is needed (Hai e al.,
2018). La en a iables, o unseen o unde ly-
ing cons uc s in he da a, can be es ima ed
using CB-SEM (Hai e al., 2011; Hoope e al.,
2008). The CB-SEM o e s many i indices ha
could be employed o examine he model i and
spo any po en ial issues unde conside a ion
(Ben le , 1990). Se e al g oups o subpopula-
ions can be analyzed using CB-SEM, allowing
o compa ing g oups o es ing in a iance
measu emen s (Sa s ed e al., 2021). Ac-
co ding o Sa s ed e al. (2019) and Rigdon
(2016), he CB-SEM enables model adjus men
by in oducing o elimina ing la en s uc u es
o ou es. Va ious so wa es a e a ailable o
CB-SEM analysis, including AMOS, LISREL,
and M-Plus (Becke e al., 2022; Hai e al.,
2018). By p o iding de ails on he s eng h and
di ec ion o he link and he ela i e p ominence
o each cons uc in he model unde conside -
a ion, CB-SEM enables he analysis o ela ion-
ships be ween a iables (Pa ma e al., 2022).
Thus, i is deba ed ha CB-SEM is a s a is ical
me hodology using he compu a ionally e ec-
i e maximum likelihood me hod o examine
he model’s pa ame e s. I is p edica ed on mul-
i a ia e no malcy and necessi a es he ab-
sence o co ela ion in he da a. Addi ionally,
i o e s many goodness-o - i s a is ics, allows
o model adjus men and es ima ion o la-
en a iables, and makes so wa e a ailable
(Hai e al., 2014).
This pape ’s goal is o assess and con as
PLS-SEM s. CB-SEM modeling. This s udy
may be help ul o u u e esea che s, who may
use i o decide which app oach o use unde
pa icula ci cums ances. The CB-SEM and
PLS-SEM mul i a ia e app oaches a e also co -
e ed comp ehensi ely in his s udy. The CB-SEM
and PLS-SEM mul i a ia e app oaches ha e
also been desc ibed in ea lie esea ch, bu ha
li e a u e does no discuss e e y aspec o bo h
mul i a ia e echniques (Becke e al., 2022; Hai
e al., 2019; Lega e e al., 2022; Ringle e al.,
2022), and se e al o he s udies, which had
se e al d awbacks. P e ious li e a u e, o in-
s ance, does no add ess se e al ea u es, such
as sample size, mul icollinea i y issues, compo-
nen s, ypes o CB-SEM and PLS-SEM, model
i indices, measu emen , and s uc u al models.
The cu en s udy p o ides an in-dep h analy-
sis o he CB-SEM and PLS-SEM mul i a ia e
echniques’ ea u es, bene i s, sho comings,
and me hodology.
The emaining sec ions o he pape a e
di ided in o nume ous sec ions, such as sec-
ion 2 (Theo e ical Backg ound), sec ion 3
(Resea ch me hodology), sec ion 4 (Resul s
and discussion), sec ion 5 (Conclusions), and
Limi a ions and u u e esea ch o ien a ions.
1. Theo e ical backg ound
P e ious li e a u e has explo ed he di e ence
be ween PLS-SEM and CB-SEM modeling.
The li e a u e di e en ia ed hei ca ego ies
and demons a ed he e iciencies and de-
iciencies o bo h models (Hai e al., 2006;
Hai & Sa s ed , 2021). Se e al s udies ha e
demons a ed ha PLS-SEM is an SEM o ex-
plo e complex ela ionships be ween nume ous
pa ame e s (Hai e al., 2022; Hensele e al.,
2015). Simila ly, p e ious li e a u e exhibi ed
ha CB-SEM models could be used depend-
ing on he esea ch goals, esea ch ques ions,
and da a a angemen s. PLS-PM (PLS pa h
modeling) is a PLS-SEM a ian used o e alu-
a e associa ions be ween obse ed and unob-
se ed elemen s in he model and o es ima e
he pa h coe icien s connec ing hese a i-
ables. PLS-SEM and CB-SEM modeling come
in a mul iplici y o di e en o ms (Memon e al.,
2019; Ringle e al., 2015). Acco ding o Hai
e al. (2022) and Sa s ed and Cheah (2019),
PLS-CFA (PLS-con i ma o y ac o analysis)
is used o gauge heo ies abou he s uc u e
o he measu emen model, including heo ies
abou he numbe o componen s, ac o load-
ings, and measu emen e o s. The PLS-SEM
is a me hod o es ima e he pa h coe icien s
be ween cons uc s and in es iga e associa-
ions be ween unobse ed elemen s in a model
(Ringle e al., 2022). PLS- eg ession is accus-
omed o e alua ing he associa ion among
p ede e mined p edic o s o a cons uc o in-
e es , such as a dependen a iable (Hai
& Sa s ed , 2021; Lega e e al., 2022). By iden-
i ying he linea blend o componen s ha
maximizes he co a iance among cons uc s,
PLS-canonical analysis is used o ca ego ize
he unde lying s uc u e o a da ase , as dem-
ons a ed by Rich e e al. (2020) and Hai e al.
(2017). A se o da a is di ided in o g oups using
PLS-DA (PLS-disc iminan analysis), a kind
o PLS-SEM g ounded on he alues o p edic-
o s (Hai e al., 2022). PLS-SEM wi h small
da a is he ype o PLS-SEM ha is e y help ul
when se e al cons uc s a e mo e inc edibly
associa ed wi h a small sample size; missing
195
2024, olume 27, issue 1, pp. 192–210, DOI: 10.15240/ ul/001/2024-5-001
In o ma ion Managemen
da a, non-no mali y, and mul icollinea i y can
all be accommoda ed by i (Ma hews, 2017;
Ringle e al., 2015).
Acco ding o p e ious s udies, he e a e
o he a ie ies o CB-SEM modeling, including
con i ma o y ac o analysis, a so o CB-SEM
accus omed o es ing a conside ed measu e-
men model based on ixed la en a iables and
p ese mani es ac o s (Ahmed e al., 2022;
Hai e al., 2022). I could suppo o disp o e
a scale’s o measu e’s ac o s uc u e (Hai
e al., 2010; Hussain & Ahmed, 2020). Pa h
analysis assesses he di ec ion, s eng h,
and co ela ion be ween a ious pa ame e s.
I can e alua e heo ies o causal ela ionships
be ween many componen s (Hayes e al.,
2017). This kind o CB-SEM, known as la en
g ow h cu e modeling (LGCM), looks a how
a iables change o e ime. Examining a a i-
able’s a e o change and i s consis ency ac oss
ime is a common p ac ice (Hai e al., 2011).
G ounded on he p o isions o answe s o a se
o obse ed ac o s, la en class analysis (LCA)
is equen ly used o asce ain subdi isions
o classes wi hin a popula ion (Nunkoo e al.,
2020). The CB-SEM me hod, known as mul i-
g oup SEM, compa es an associa ion among
cons uc s ac oss a ious g oups o popula ions.
I can be used o look o a ia ions o pa e ns
in he ela ionships be ween a iables be ween
a ious g oups (Cohen, 1992; Hayes e al.,
2017). This kind o CB-SEM, SEM wi h miss-
ing da a, deals wi h missing da a in he analy-
sis. I is cus oma y o look a he pa ame e s
o he model and missing da a simul aneously
(Hai e al., 2006). SEM wi hou no mali y da a
is he ype o CB-SEM ha wo ks wi h non-
no mal da a o he analysis. Using eliable
es ima ing app oaches, he model’s pa ame e s
could be e alua ed (Hensele e al., 2015).
Acco ding o Ringle e al. (2015) and Hai
e al. (2010), he sample size o PLS-SEM
should be sizable o ensu e adequa e powe
o he s a is ical analysis and o ob ain a sui -
able le el o gene alizabili y (Hai e al., 2010;
Ringle e al., 2015). Howe e , PLS-SEM
sample size ecommenda ions a e less ac-
cu a e han hose o adi ional SEM (Sha ma
e al., 2021). PLS-SEM is conside ed a mo e
eliable me hod han adi ional SEM ega ding
sample size and measu emen e o because
i can ole a e highe le els o measu emen
e o (Ahmed e al., 2019; Hai e al., 2019). As
a esul , PLS-SEM equen ly has mo e lexible
sample size equi emen s han ypical SEM.
I is i al o keep in mind ha sample size
is always de e mined by he s udy pu pose,
he esou ces a ailable, and he amoun o ime
a ailable, e en i some s udies ha e shown ha
PLS-SEM may be employed wi h da ase s as
low as 50–100 ins ances (Hayes e al., 2017).
P e ious li e a u e also discussed he e-
qui ed sample o PLS-SEM modeling; acco d-
ing o Hussain and Ahmed (2020), Hussain
e al. (2021), and Zaidi e al. (2022), he sample
size equi ed o achie e a speci ic powe le el,
o ins ance, 80% o 90%, can be de e mined
in a ious ways, including simula ion s udies
and powe analysis echniques. The sample
size is calcula ed conside ing he esea ch
opic, he esou ces a ailable, and he amoun
o ime a ailable. I is c ucial o emembe ha
sample size es ima ions a e equen ly ap-
p oxima e. I is also c i ical o emembe ha
he PLS-SEM sample size equi emen s a y
depending on he numbe o p edic o s and
model complexi y (Sa s ed & Cheah, 2019).
As models become mo e complica ed, sample
sizes become mo e c i ical. The sample size
mus be p o icien a ensu ing he accu acy and
objec i i y o he es ima ed alues (Shmueli
e al., 2019). Simila ly, p e ious li e a u e also
discussed ha he sample size is a i al conce n
in he CB-SEM echnique since i can a ec
he alua ion o he conside ed model’s pa am-
e e s and he model’s capaci y o ix he da ase .
A la ge sample size will p oduce mo e p ecise
pa ame e es ima es and a be e model-da a
i (Hai e al., 2014; Sha ma e al., 2021).
Acco ding o Hai e al. (2011), he op imal
sample size will depend on he complexi y
o he model, he numbe o indica o s, he la-
en ac o s quan i y, and he le el o measu e-
men e o . The e a e nume ous me hods o
compu ing he sample size o he CB-SEM.
The ecommenda ions o he sample size
o he CB-SEM depend on a ious aspec s,
among hem he numbe o ac o s, he num-
be o es ima ed pa ame e s, and he amoun
o measu emen e o (Hai e al., 2011). One
o se e al ecommended sample size c i e ia
is he “10:1 ule”, which desc ibes ha he sam-
ple size mus be en imes he pa ame e s,
which has o be e alua ed. This ule, hough,
only unc ions unde ce ain ci cums ances
(Hussain & Ahmed, 2020; S eukens & Le oi-
We elds, 2016). Powe analysis echniques
o simula ion s udies can be used o calcula e
196 2024, olume 27, issue 1, pp. 192–210, DOI: 10.15240/ ul/001/2024-5-001
In o ma ion Managemen
he sample size equi ed o a ain a gi en powe
le el, o ins ance, 80% o 90%, o es ablish
he sample size ha p o ides a high likelihood
o de ec ing a pa icula e ec . In gene al,
SEM equi es a sample size o 200 o mo e.
Howe e , his guideline migh only apply
o pa icula models; hus, employing mo e so-
phis ica ed sample size es ima ion echniques
is always a good idea.
Acco ding o Kang (2021) and Hoenig and
Heisey (2001), powe analysis can decide
he desi ed sample size o iden i y a pa icula
e ec size a a speci ic powe le el. Powe anal-
ysis can conside he magni ude o measu e-
men e o , he complexi y o he model, and
he numbe o indica o s. Simula o s de Mon e
Ca lo – his me hod simula es da a and igu es
ou he sample size necessa y o accu a ely
es ima e model pa ame e s (Hayes e al., 2017;
K oese e al., 2014). I can be accomplished
by ela ing he Akaike in o ma ion c i e ion (AIC)
o he Bayesian in o ma ion c i e ion (BIC) o
a ious sample sizes. I is c ucial o emembe
ha sample size is only one conside a ion when
e alua ing he i o a model. Se e al addi ional
elemen s, o ins ance, he numbe o indica o s,
he conside ed model’s complexi y, he da a
dis ibu ion, and he es ima ion me hod, impac
he model i (Hoenig & Heisey, 2001).
P e ious li e a u e demons a ed ha mul-
icollinea i y is a common p oblem in PLS-SEM
and CB-SEM, which occu s when wo o mo e
p edic o a iables a e closely associa ed
(G ewal e al., 2004). I occu s when wo o
mo e independen a iables exhibi s ong co -
ela ions, and es ima ing models and explaining
hei esul s can be challenging (Wondola e al.,
2020). Fo example, a co ela ion ma ix can
de e mine how e e y independen cons uc
connec s wi h o he s o ind mul icollinea -
i y in PLS-SEM and CB-SEM. Mul icollinea -
i y may be p esen i he e is a signi ican
co ela ion be ween wo o mo e independen
cons uc s (Wondola e al., 2020). The deg ee
o mul icollinea i y in a mul iple eg ession
model is measu ed by he a iance in la ion
ac o (VIF). The VIF o 1 shows he absence
o mul icollinea i y, while a VIF bigge han 1
speci ies he occu ence o mul icollinea i y.
High mul icollinea i y is equen ly indica ed
by a VIF mo e signi ican han i e (Chan e al.,
2022; Hussain & Ahmed, 2020). The a iance
amoun in a p edic o , which o he p edic o s
canno desc ibe, is ep esen ed by ole ance,
which is he ecip ocal o VIF. The e is high
mul icollinea i y when he ole ance alue is be-
low 0.2. The condi ion index gauges he le el
o mul icollinea i y in a mul iple eg ession mod-
el. Mul icollinea i y is indica ed by a numbe
highe han 30 (A minge & Schoenbe g, 1989;
Chan e al., 2022). The p e ious li e a u e has
discussed and iden i ied se e al posi i e and
nega i e aspec s o PLS-SEM and CB-SEM
echniques, howe e , nume ous ac o s a e s ill
missing o es ablish he di e en ia ion be ween
bo h modeling echniques, hus he cu en
s udy answe s hose ques ions.
2. Resea ch me hodology
2.1 Resea ch design and es ima ion
echniques
The unde aking is a compa a i e s udy, which
has di e en ia ed PLS-SEM and CB-SEM mod-
eling; he s udy also conside s he e iciencies
and de iciencies o bo h models in he manage-
men sciences ield. The compa a i e s udies
could be pe o med quali a i ely o quan i a i e-
ly. Howe e , he esea ch design o his s udy
is quali a i e, and esea che s ha e s a ed
he p os and cons o PLS-SEM and CB-SEM
echniques; hey also compa e di e en pa am-
e e s o bo h echniques. This s udy has used
p e ious li e a u e and ho oughly e iewed
p e ious s udies, books, and o he ele an
publica ions o analyze bo h models. This s udy
also used g aphical analysis o dis inguish be-
ween PLS-SEM and CB-SEM modeling.
The s udy examined he c i e ia o alida e
measu emen models, such as con e gen
and disc iminan alidi ies, using ac o loading
o i ems, C onbach’s alpha, composi e eliabili y,
and a e age a iance ex ac ed o cons uc s
o alida e he con e gen alidi y and eliabil-
i y in bo h PLS-SEM and CB-SEM echniques.
Mo eo e , his s udy analyzed HTMT, Fo nell-
La cke c i e ion, and c oss-loading o alida e
disc iminan alidi y o bo h SEM echniques.
Simila ly, his s udy also examined he pa ame-
e s o alida ing a s uc u al model o PLS-SEM
modeling. Fo his pu pose, he esea che s
used he coe icien o de e mina ion (R2), e ec
size ( 2), pa h coe icien analysis (di ec , indi ec
ela ionship o cons uc s), goodness o i mea-
su es, and p edic i e ele ance (Q2).
This esea ch used con i ma o y ac o
analysis, s uc u al equa ion modeling, pa h
coe icien analysis (di ec , indi ec ela ionship
o cons uc s), and goodness o i measu es
197
2024, olume 27, issue 1, pp. 192–210, DOI: 10.15240/ ul/001/2024-5-001
In o ma ion Managemen
o alida e s uc u al models in CB-SEM ech-
niques. This s udy also used he g aphical
analysis o examine he obse ed, unobse ed,
con e gen , and disc iminan alidi y o endo se
he measu emen model o bo h PLS-SEM
and CB-SEM echniques. The g aphical analy-
sis also de ined he pa h coe icien ela ionship
(di ec and indi ec ela ionship o cons uc s)
o alida e he s uc u al model o bo h
PLS-SEM and CB-SEM echniques.
2.2 Ac onyms and ull names
Tab. 1 exhibi ed he ac onyms and ull names
o di e en abb e ia ions used in his pape .
3. Resul s and discussion
The esul s o his s udy demons a ed he pa am-
e e s o he measu emen and s uc u al models
o bo h PLS-SEM and CB-SEM echniques.
3.1 The measu emen model in CB-SEM
and PLS-SEM modeling
In PLS-SEM & CB-SEM modeling, alida ing
he measu emen model en ails e alua ing
he i ness o he da ase and he eliabili y o in-
dica o s chosen o ep esen he la en a iables
(Hai e al., 2019). This p ocedu e includes
he ollowing s eps as he ac o loadings indi-
ca e how in ensely indica o s and unobse ed
ac o s a e linked. Signi ican ac o loadings
show ha he indica o s and la en a iables a e
closely connec ed (Hai e al., 2014; Rigdon,
2016). Fac o loadings ha e a con en ional
cu -o o 0.7, which can change depending
on he esea ch en i onmen (Ringle e al.,
2015). The measu ing model mus be alida ed
by e alua ing he indica o s’ eliabili y and a-
lidi y. While alidi y na a es how well he indi-
ca o s measu e he la en a iable, eliabili y
Ac onyms Full names Ac onyms Full names
PLS-SEM Pa ial leas squa e s uc u al
equa ion modeling PLS-CFA Pa ial leas squa e con i ma o y
ac o analysis
CB-SEM Co a iance-based s uc u al
equa ion modeling PLS-DA Pa ial leas squa e disc iminan
analysis
SEM S uc u al equa ion modeling LGCM La en g ow h cu e modeling
Sma -PLS Sma pa ial leas squa e so wa e LCA La en class analysis
Wa p-PLS
Va iance-based and ac o -based
s uc u al equa ion modeling
so wa e
SRMR S anda dized oo mean squa e
esidual
XLSTAT Excel s a is ical so wa e HTMT He e o ai mono ai a io
o co ela ion
AMOS Analysis o momen s uc u es AVE A e age a iance ex ac ed
LISREL Linea s uc u e ela ions D_ULS The squa ed euclidean dis ance
M-Plus Mic odia plus so wa e AGFI Adjus ed goodness o i index
RMSEA Roo mean squa e e o
o app oxima ion RNI Rela i e non-cen ali y index
CFI Compa a i e i index PCFI Pa simonious-adjus ed i index
GFI The goodness o he i index PNFI Pa simony-adjus ed no med i index
TLI Tucke Lewis index G_D Geodesic dis ance
NFI No mal i index VIF Va iance in la ion ac o
Sou ce: own
Tab. 1: Ac onyms and ull names
198 2024, olume 27, issue 1, pp. 192–210, DOI: 10.15240/ ul/001/2024-5-001
In o ma ion Managemen
e e s o he indica o s’ consis ency ac oss ime
(By ne, 2013; Ringle e al., 2015).
The model mus o e a good ma ch o
he da ase ; in PLS-SEM modeling, se -
e al i indices, including R2 and Q2, can be
applied o measu e how well he model ixes
he da ase . These indices show he pe cen -
age o he a iance in ou come cons uc s
he model jus i ica ions (Hensele e al., 2015;
Pa ma e al., 2022). Simila ly, many i indices,
including he RMSEA, chi-squa e s a is ic, and
compa a i e i index (CFI), can be applied
o measu e he model’s i ness in CB-SEM
modeling (Hoope e al., 2008). The impo ance
o pa h coe icien s and he o e all model should
be es ed by examining he s uc u al model
(Ben le & Bone , 1980; Hai e al., 2022).
Suppose he ac o loadings o he model
do no ma ch he da a well. In ha case, i may be
essen ial o e-speci y he model by modi ying
he pa h coe icien s, adding o emo ing a i-
ables, o making o he modi ica ions (Sa s ed
e al., 2022). I is essen ial o emembe ha
measu emen model alida ion is an i e a i e
p ocess, and he model should be e ined and
e-e alua ed as needed un il an accep able le -
el. I is c ucial o emembe ha when wo king
wi h CB-SEM, using mul iple da a sou ces, such
as sel - epo su eys, beha io al obse a ions,
and physiological measu es, can inc ease
he a ionali y o he measu emen model. Ad-
di ionally, he alida ion p ocess should be done
wi h he sample used in he s udy and no jus
in he popula ion in gene al (Hai e al., 2014).
The anno a ed g aphical o m o he measu e-
men model o PLS-SEM is p o ided in Fig. 1
(Ahmed e al., 2021). Fig. 1 demons a ed ha
he ac o loadings o each i em a e highe
han 0.70, and he a e age a iance ex ac ed
is mo e signi ican han 0.50, which ul illed
he con e gen alidi y equi emen . Mo eo e ,
he pa h analysis be ween he cons uc alida -
ed he disc iminan alidi y; hus, his endo sed
he measu emen model.
Fig. 1: Measu emen model in PLS-SEM modeling
Sou ce: Ahmed e al. (2021)
199
2024, olume 27, issue 1, pp. 192–210, DOI: 10.15240/ ul/001/2024-5-001
In o ma ion Managemen
The anno a ed g aphical o m o he mea-
su emen model o CB-SEM is p o ided in Fig. 2
(Ash a e al., 2018). Fig. 2 also demons a ed
ha each obse ed a iable has a ac o load-
ing o mo e han 0.70, alues o pa h coe -
icien be ween he unobse ed a iables,
and alues o goodness o i measu es ha e
ollowed he cu -o s. Thus, Fig. 2 demons a-
es ha he measu emen model is alida ed
in CB-SEM modeling.
3.2 The s uc u al model in CB-SEM and
PLS-SEM modeling
In CB-SEM and PLS-SEM modeling, alida ing
he s uc u al model en ails analyzing he mod-
el’s i ness o he da ase , de e mining he im-
po ance o he pa h coe icien , and e iewing
he o e all model (Kline, 2015). This p ocedu e
includes se e al s eps; o example, he pa h
coe icien s show how s ong and in wha di-
ec ion he la en a iables a e ela ed. High
posi i e pa h coe icien s indica e a s ong posi-
i e ela ionship be ween he la en a iables,
while high nega i e pa h coe icien s indica e
a s ong nega i e ela ionship (Hai e al., 2019;
Raza e al., 2021). T- es s o boo s apping ech-
niques can be used o conclude he signi icance
o he pa h coe icien s. I he pa h coe icien
is signi ican , he la en a iables mus be s a-
is ically ela ed (Hai e al., 2014; Hayes e al.,
2017; Hensele e al., 2015).
In PLS-SEM modeling, i indices like R2
and Q2 could be applied o measu e he o e all
i ness o he model. These indices indica e
he a iance p opo ion in dependen ac o s ha
he model explains (Ben le , 1990). The o e all
i ness o he CB-SEM model can be e alu-
a ed using i indices such as he RMSEA, chi-
squa e s a is ic, and compa a i e i index (CFI;
Hoope e al., 2008). Disc iminan alidi y ex-
amines how li le la en a iables connec wi h
measu emen s o un ela ed cons uc s. I can
Fig. 2: Measu emen model in CB-SEM modeling
Sou ce: Ash a e al. (2018)
200 2024, olume 27, issue 1, pp. 192–210, DOI: 10.15240/ ul/001/2024-5-001
In o ma ion Managemen
be assessed by con as ing he la en a iables’
a e age a iance ex ac ed wi h hei squa ed
co ela ion o un ela ed ac o s (Ahmed e al.,
2021; Fo nell & La cke , 1981; Hai J . e al.,
2017; Malho a e al., 2006).
The model may need o be e-speci ied
by adding o emo ing a iables, changing
he pa h coe icien s, o modi ying he model
in o he ways i i does no i he da a well
o i he pa h coe icien s a e no signi ican
(Kau mann & Gaeckle , 2015; Sa s ed e al.,
2022). I is c ucial o emembe ha s uc-
u al model alida ion is an i e a i e p ocess.
The model mus be polished and eexamined
un il an accep able i le el and signi icance
a e achie ed (Sa s ed e al., 2019). I is es-
sen ial o emembe ha when wo king wi h
CB-SEM, using mul iple da a sou ces, such
as sel - epo su eys, beha io al obse a-
ions, and physiological measu es, can
inc ease he alidi y o he s uc u al model.
Addi ionally, he alida ion p ocess should be
done wi h he sample used in he s udy and no
jus in he popula ion in gene al (Malho a e al.,
2006). The anno a ed g aphical o m depic ed
he s uc u al model o PLS-SEM in Fig. 3
(Ahmed e al., 2021). Fig. 3 demons a ed
he pa h coe icien be ween he cons uc s
(di ec and indi ec ela ionship), which shows
signi ican alues; mo eo e , R-squa e alues
showed he impac o exogenous a iables
on endogenous a iables. Thus, Fig. 3 alida -
ed he s uc u al model in PLS-SEM modeling.
The anno a ed g aphical shape o he
s uc u al model o CB-SEM is p o ided
in Fig. 4 (Ash a e al., 2018). Fig. 4 demon-
s a es he pa h coe icien be ween he con-
s uc s (di ec and indi ec ela ionship), which
shows signi ican alues. Mo eo e , i in-
dices alues mee he equi ed h eshold.
Thus, Fig. 4 alida ed he s uc u al model o
CB-SEM modeling.
Fig. 3: S uc u al model in PLS-SEM
Sou ce: Ahmed e al. (2021)
207
2024, olume 27, issue 1, pp. 192–210, DOI: 10.15240/ ul/001/2024-5-001
In o ma ion Managemen
Managemen , 23(1), 82–104. h ps://doi.o g/
10.3846/jbem.2021.15664
A minge , G., & Schoenbe g, R. J. (1989).
Pseudo maximum likelihood es ima ion and
a es o misspeci ica ion in mean and co a i-
ance s uc u e models. Psychome ika, 54(3),
409–425. h ps://doi.o g/10.1007/b 02294626
Ash a , M., V einha d , J., Ahmed, R. R.,
Š eimikienė, D., & Mangi, R. A. (2018). Explo -
ing in e ening in luence o in e ac ional jus ice
be ween p ocedu al jus ice and job pe o -
mance: E idence om Sou h Asian coun ies.
Am i ea u Economic, 20(47), 169–184. h ps://
doi.o g/10.24818/ea/2018/47/169
As achan, C. B., Pa el, V. K., & Wanzen-
ied, G. (2014). A compa a i e s udy o CB-SEM
and PLS-SEM o heo y de elopmen in amily
i m esea ch. Jou nal o Family Business S a e-
gy, 5(1), 116–128. h ps://doi.o g/10.1016/j.j bs.
2013.12.002
Ba bić, D., Lučić, A., & Chen, J. M. (2019).
Measu ing esponsible inancial consump-
ion beha iou . In e na ional Jou nal o Con-
sume S udies, 43(1), 102–112. h ps://doi.o g/
10.1111/ijcs.12489
Becke , J.-M., P oksch, D., & Ringle, C. M.
(2022). Re isi ing Gaussian copulas o handle
endogenous eg esso s. Jou nal o he Acade-
my o Ma ke ing Science, 50(1), 46–66. h ps://
doi.o g/10.1007/s11747-021-00805-y
Ben le , P. M. (1990). Compa a i e i in-
dexes in s uc u al models. Psychological Bul-
le in, 107(2), 238–246. h ps://doi.o g/10.1037/
0033-2909.107.2.238
Ben le , P. M., & Bone , D. G. (1980).
Signi icance es s and goodness o i in he
analysis o co a iance s uc u es. Psycho-
logical Bulle in, 88(3), 588–606. h ps://doi.o g/
10.1037/0033-2909.88.3.588
Blackwell, R. D., Minia d, P. W., & Engel, J. F.
(2001). Consume beha io . Ha cou College
Publishe s. h ps://doi.o g/10.2307/1250212
Bollen, K. A. (1989). S uc u al equa ions
wi h la en a iables. John Wiley. h ps://doi.o g/
10.1002/9781118619179
Bollen, K. A., & Da is, W. R. (2009). Causal
indica o models: Iden i ica ion, es ima ion, and
es ing. S uc u al Equa ion Modeling: A Mul idis-
ciplina y Jou nal, 16(3), 498–522. h ps://doi.o g/
10.1080/10705510903008253
By ne, B. M. (2013). S uc u al equa ion
modeling wi h AMOS: Basic concep s, applica-
ions, and p og amming (2nd ed.). Rou ledge.
h ps://doi.o g/10.4324/9780203805534
Chan, J. Y.-L., Leow, S. M. H., Bea, K. T.,
Cheng, W. K., Phoong, S. W., Hong, Z.-W.,
& Chen, Y.-L. (2022). Mi iga ing he mul icol-
linea i y p oblem and i s machine lea ning ap-
p oach: A e iew. Ma hema ics, 10(8), 1283.
h ps://doi.o g/10.3390/ma h10081283
Cheah, J. H., Ting, H., Ramayah, T.,
Memon, M. A., Cham, T. H., & Cia olino, E.
(2019). A compa ison o i e e lec i e- o ma-
i e es ima ion app oaches: Reconside a ion
and ecommenda ions o ou ism esea ch.
Quali y & Quan i y, 53(3), 1421–1458. h ps://
doi.o g/10.1007/s11135-018-0821-7
Chin, W. W. (2010). Boo s ap c oss-
alida ion indices o PLS pa h model assess-
men . In V. W. W. Esposi o Vinzi, J. Hensele ,
& H. Wang (Eds.), Handbook o pa ial leas
squa es: Concep s, me hods and applica ions
(Sp inge Handbooks o Compu a ional S a is-
ics Se ies) (Vol. 2, pp. 83–97). Sp inge . h ps://
doi.o g/10.1007/978-3-540-32827-8_4
Cohen, J. (1992). A powe p ime . Psycho-
logical Bulle in, 112(1), 155–159. h ps://doi.o g/
10.1037/0033-2909.112.1.155
Cohen, J. (1994). The ea h is ound (p < .05).
Ame ican Psychologis , 49(12), 997–1003.
h ps://doi.o g/10.1037/0003-066x.49.12.997
Fo nell, C., & La cke , D. F. (1981). S uc u -
al equa ion models wi h unobse able a iables
and measu emen e o : Algeb a and s a is ics.
Jou nal o Ma ke ing Resea ch, 18(3), 382–388.
h ps://doi.o g/10.1177/002224378101800313
F anke, G., & Sa s ed , M. (2019). Heu-
is ics e sus s a is ics in disc iminan alidi y
es ing: A compa ison o ou p ocedu es. In e -
ne Resea ch, 29(3), 430–447. h ps://doi.o g/
10.1108/in -12-2017-0515
G ewal, R., Co e, J. A., & Baumga ne , H.
(2004). Mul icollinea i y and measu emen e o
in s uc u al equa ion models: Implica ions o
heo y es ing. Ma ke ing Science, 23(4), 519–529.
h ps://doi.o g/10.1287/mksc.1040.0070
Hai , J. F., Ande son, R. E., Babin, B. J.,
& Black, W. C. (2010). Mul i a ia e da a analy-
sis: A global pe spec i e. Pea son Educa ion.
Hai , J. F., Babin, B. J., & K ey, N. (2017). Co a-
iance-based s uc u al equa ion modeling in he
Jou nal o Ad e ising: Re iew and ecommen-
da ions. Jou nal o Ad e ising, 46(1), 163–177.
h ps://doi.o g/10.1080/00913367.2017.
1281777
Hai , J. F., Black, W. C., Babin, B. J., & An-
de son, R. E. (2018). Mul i a ia e da a analysis.
Cengage.
208 2024, olume 27, issue 1, pp. 192–210, DOI: 10.15240/ ul/001/2024-5-001
In o ma ion Managemen
Hai , J. F., Black, W. C., Babin, B. J., Ande -
son, R. E., & Ta ham, R. L. (2006). Mul i a ia e
da a analysis. Pea son P en ice Hall.
Hai , J. F., Hul , T., Ringle, C., & Sa s ed , M.
(2014). A p ime on pa ial leas squa es s uc-
u al equa ion modeling (PLS-SEM). Sage
Publica ions.
Hai , J. F., Hul , G. T., Ringle, C. M.,
& Sa s ed , M. (2022). A p ime on pa ial
leas squa es s uc u al equa ion modeling
(PLS-SEM). Sage Publica ions.
Hai , J. F., Ringle, C. M., & Sa s ed , M. (2011).
PLS-SEM: Indeed, a sil e bulle . Jou nal o
Ma ke ing Theo y and P ac ice, 19(2), 139–151.
h ps://doi.o g/10.2753/m p1069-6679190202
Hai , J. F., Rishe , J. J., Sa s ed , M., & Ring-
le, C. M. (2019). When o use and how o epo
he esul s o PLS-SEM. Eu opean Business
Re iew, 31(1), 2–24. h ps://doi.o g/10.1108/
eb -11-2018-0203
Hai , J. F., & Sa s ed , M. (2021). Explana-
ion plus p edic ion: The logical ocus o p ojec
managemen esea ch. P ojec Managemen
Jou nal, 52(4), 319–322. h ps://doi.o g/
10.1177/8756972821999945
Hayes, A. F., Mon oya, A. K., & Rock-
wood, N. J. (2017). The analysis o mecha-
nisms and hei con ingencies: PROCESS
e sus s uc u al equa ion modeling. Aus al-
asian Ma ke ing Jou nal, 25(1), 76–81. h ps://
doi.o g/10.1016/j.ausmj.2017.02.001
Hensele , J. (2021). Composi e-based
s uc u al equa ion modeling: Analyzing la en
and eme gen a iables. Guil o d P ess.
Hensele , J., Ringle, C. M., & Sa s ed , M.
(2015). A new c i e ion o assessing dis-
c iminan alidi y in a iance-based s uc u al
equa ion modeling. Jou nal o he Academy o
Ma ke ing Science, 43(1), 115–135. h ps://doi.
o g/10.1007/s11747-014-0403-8
Hoenig, J. M., & Heisey, D. M. (2001).
The abuse o powe . The Ame ican S a is i-
cian, 55(1), 19–24. h ps://doi.o g/10.1198/
000313001300339897
Hoope , D., Coughlan, J., & Mullen, M.
(2008). S uc u al equa ion modeling: Guide-
lines o de e mining model i . Elec onic
Jou nal o Business Resea ch Me hods, 6(1),
53–60. h ps://doi.o g/10.21427/D7CF7R
Hu, L. T., & Ben le , P. M. (1999). Cu -o c i e-
ia o i indexes in co a iance s uc u e analysis:
Con en ional c i e ia e sus new al e na i es.
S uc u al Equa ion Modeling, 6(1), 1–55.
h ps://doi.o g/10.1080/10705519909540118
Hul , G. T. M., Hai , J. F., P oksch, D.,
Sa s ed , M., Pinkwa , A., & Ringle, C. M.
(2018). Add essing endogenei y in in e na ional
ma ke ing applica ions o pa ial leas squa es
s uc u al equa ion modeling. Jou nal o In e -
na ional Ma ke ing, 26(3), 1–21. h ps://doi.
o g/10.1509/jim.17.0151
Hussain, S., & Ahmed, R. R. (2020). Sma -
phone buying beha io s in a amewo k o b and
expe ience and b and equi y. T ans o ma ions
in Business & Economics, 19(2), 220–242.
Hussain, S., Ahmed, R. R., & Shamsi, A. F.
(2021). Technology con i ma ion is associa ed
o imp o ed psychological well-being: E idence
om an expe imen al design. T ans o ma ions
in Business & Economics, 20(53), 177–196.
Hwang, H., Sa s ed , M., Cheah, J. H.,
& Ringle, C. M. (2020). A concep analysis o
me hodological esea ch on composi e-based
s uc u al equa ion modeling: B idging PLSPM
and GSCA. Beha io me ika, 47(1), 219–241.
h ps://doi.o g/10.1007/s41237-019-00085-5
Kang, H. (2021). Sample size de e mina-
ion and powe analysis using he G*Powe
so wa e. Jou nal o Educa ional E alua ion
o Heal h P o essions, 18, 17. h ps://doi.
o g/10.3352/jeehp.2021.18.17
Kau mann, L., & Gaeckle , J. (2015).
A s uc u ed e iew o pa ial leas squa es in
supply chain managemen esea ch. Jou nal
o Pu chasing and Supply Managemen , 21(4),
259–272. h ps://doi.o g/10.1016/j.pu sup.2015.
04.005
Kline, R. B. (2015). P inciples and p ac-
ice o s uc u al equa ion modeling (4 h ed.).
The Guil o d P ess.
K oese, D. P., B e e on, T., Taim e, T.,
& Bo e , Z. I. (2014). Why he Mon e Ca lo
me hod is so impo an oday. WIREs Compu-
a ional S a is ics, 6(6), 386–392. h ps://doi.
o g/10.1002/wics.1314
Lega e, A. E., Hai , J. F., Ch e ien, J. L.,
& Rishe , J. J. (2022). PLS-SEM: P edic ion-
o ien ed solu ions o HRD esea che s. Hu-
man Resou ce De elopmen Qua e ly, 34(1),
91–109. h ps://doi.o g/10.1002/h dq.21466
Liengaa d, B. D., Sha ma, P. N.,
Hul , G. T. M., Jensen, M. B., Sa s ed , M.,
Hai , J. F., & Ringle, C. M. (2021). P edic ion:
Co e ed, ye o saken? In oducing a c oss-
alida ed p edic i e abili y es in pa ial
leas squa es pa h modeling. Decision Sci-
ences, 52(2), 362–392. h ps://doi.o g/10.1111/
deci.12445
209
2024, olume 27, issue 1, pp. 192–210, DOI: 10.15240/ ul/001/2024-5-001
In o ma ion Managemen
Lu, J., Ren, L., Zhang, C., Rong, D.,
Ahmed, R. R., & S eimikis, J. (2020). Modi ied
Ca oll’s py amid o co po a e social esponsi-
bili y o enhance o ganiza ional pe o mance
o SME indus y. Jou nal o Cleane P oduc-
ion, 271, 122456. h ps://doi.o g/10.1016/
j.jclep o.2020.122456
Malho a, N. K., Kim, S. S., & Pa il, A.
(2006). Common me hod a iance in IS e-
sea ch: A compa ison o al e na i e app oaches
and a eanalysis o pas esea ch. Managemen
Science, 52(12), 1865–1883. h ps://doi.o g/
10.1287/mnsc.1060.0597
Ma hews, L. (2017). Applying mul i-g oup
analysis in PLS-SEM: A s ep-by-s ep p ocess.
In H. La an & R. Noonan (Eds.), Pa ial leas
squa es pa h modeling (pp. 219–243). Sp inge
In e na ional Publishing. h ps://doi.o g/10.1007/
978-3-319-64069-3_10
Maydeu-Oli a es, A., Shi, D., & Rosseel, Y.
(2018). Assessing i in s uc u al equa ion mod-
els: A Mon e-Ca lo e alua ion o RMSEA e sus
SRMR con idence in e als and es s o close
i . S uc u al Equa ion Modeling: A Mul idisci-
plina y Jou nal, 25(3), 389–402. h ps://doi.o g/
10.1080/ 10705511.2017.1389611
McDonald, R. P., & Ho, M.-H. R. (2002).
P inciples and p ac ice in epo ing s a is i-
cal equa ion analyses. Psychological Me h-
ods, 7(1), 64–82. h ps://doi.o g/10.1037/
1082-989x.7.1.64
Memon, M. A., Cheah, J.-H., Ramayah, T.,
Ting, H., Chuah, F., & Cham, T. H. (2019). Mod-
e a ion analysis: Issues and guidelines. Jou nal
o Applied S uc u al Equa ion Modeling, 3(1),
i–xi. h ps://doi.o g/10.47263/jasem.3(1)01
Miles, J., & She lin, M. (2007). A ime and
a place o inc emen al i indices. Pe sonal-
i y and Indi idual Di e ences, 42(5), 869–874.
h ps://doi.o g/10.1016/j.paid.2006.09.022
Mou i, N. (2005). A consume -based assess-
men o alliance pe o mance: An examina ion o
consume alue, sa is ac ion and pos -pu chase
beha io . Uni e si y o Cen al Flo ida.
Muelle , R. O., & Hancock, G. R. (2018).
S uc u al equa ion modeling. The e iewe ’s
guide o quan i a i e me hods in he social sci-
ences. Taylo & F ancis G oup, Rou ledge.
Nunkoo, R., Tee oo engadum, V., Ring-
le, C. M., & Sunnassee, V. (2020). Se ice qual-
i y and cus ome sa is ac ion: The mode a ing
e ec s o ho el s a a ing. In e na ional Jou nal
o Hospi ali y Managemen , 91, 102414. h ps://
doi.o g/10.1016/j.ijhm.2019.102414
Oli e , R. L. (2014). Sa is ac ion: A beha -
io al pe spec i e on he consume . Rou ledge.
h ps://doi.o g/10.4324/9781315700892
Pa ma , V., Channa , Z. A., Ahmed, R. R.,
Š eimikienė, D., Pahi, M. H., & S eimikis, J.
(2022). Assessing he o ganiza ional commi -
men , subjec i e i ali y and bu nou e ec s on
u no e in en ion in p i a e uni e si ies. Oeco-
nomia Cope nicana, 13(1), 251–286. h ps://doi.
o g/10.24136/oc.2022.008
Raza, S. A., Qazi, W., Khan, K. A., & Salam, J.
(2021). Social isola ion and accep ance o he
lea ning managemen sys em (LMS) in he ime
o COVID-19 pandemic: An expansion o he
UTAUT model. Jou nal o Educa ional Compu -
ing Resea ch, 59(2), 183–208. h ps://doi.o g/
10.1177/0735633120960421
Rich e , N. F., Schub ing, S., Hau , S.,
Ringle, C. M., & Sa s ed , M. (2020). When p e-
dic o s o ou comes a e necessa y: Guidelines
o he combined use o PLS-SEM and NCA.
Indus ial Managemen and Da a Sys ems,
120(12), 2243–2267. h ps://doi.o g/10.1108/
imds-11-2019-0638
Rigdon, E. E. (2016). Choosing PLS pa h
modeling as analy ical me hod in Eu opean man-
agemen esea ch: A ealis pe spec i e. Eu o-
pean Managemen Jou nal, 34(6), 598–605.
h ps://doi.o g/10.1016/j.emj.2016.05.006
Ringle, C. M., Wende, S., & Becke , J. M.
(2015). Sma PLS 3. Sma PLS GmbH
Boennings ed .
Ringle C. M., Wende, S., & Becke J. M.
(2022). Sma PLS 4. Sma PLS Os s einbek.
Sa s ed , M., & Cheah, J.-H. (2019). Pa -
ial leas squa es s uc u al equa ion modeling
using Sma -PLS: A so wa e e iew. Jou nal o
Ma ke ing Analy ics, 7(3), 196–202. h ps://doi.
o g/10.1057/s41270-019-00058-3
Sa s ed , M., Hai , J. F., Cheah, J.-H., Beck-
e , J.-M., & Ringle, C. M. (2019). How o speci y,
es ima e, and alida e highe -o de cons uc s
in PLS-SEM. Aus alasian Ma ke ing Jou -
nal, 27(3), 197–211. h ps://doi.o g/10.1016/
j.ausmj.2019.05.003
Sa s ed , M., Radomi , L., Moisescu, O. I.,
& Ringle, C. M. (2022). La en class analysis
in PLS-SEM: A e iew and ecommenda ions
o u u e applica ions. Jou nal o Business Re-
sea ch, 138, 398–407. h ps://doi.o g/10.1016/
j.jbus es.2021.08.051
Sa s ed , M., Ringle, C. M., & Hai , J. F.
(2021). Pa ial leas squa es s uc u al equa-
ion modeling. In C. Hombu g, M. Kla mann,
210 2024, olume 27, issue 1, pp. 192–210, DOI: 10.15240/ ul/001/2024-5-001
In o ma ion Managemen
& A. E. Vombe g (Eds.), Handbook o ma ke e-
sea ch (pp. 1–47). Sp inge In e na ional Publishing.
h ps://doi.o g/10.1007/978-3-319-05542-8_15-2
Sha ma, P. N., Shmueli, G., Sa s ed , M.,
Danks, N., & Ray, S. (2021). P edic ion-o ien -
ed model selec ion in pa ial leas squa es pa h
modeling. Decision Sciences, 52(3), 567–607.
h ps://doi.o g/10.1111/deci.12329
Shi, D., & Maydeu-Oli a es, A. (2020).
The e ec o es ima ion me hods on SEM i in-
dices. Educa ional and Psychological Measu e-
men , 80(3), 421–445. h ps://doi.o g/10.1177/
0013164419885164
Shmueli, G., Sa s ed , M., Hai , J. F.,
Cheah, J.-H., Ting, H., Vai hilingam, S.,
& Ringle, C. M. (2019). P edic i e model as-
sessmen in PLS-SEM: Guidelines o using
PLSp edic . Eu opean Jou nal o Ma ke ing,
53(11), 2322–2347. h ps://doi.o g/10.1108/
ejm-02-2019-0189
Tanaka, J. S. (1993). Mul i ace ed con-
cep ions o i in s uc u al equa ion models.
In K. A. Bollen & J. S. Long (Eds.), Tes ing
s uc u al equa ion models. Sage Publica ions.
Toda, H. Y., & Yamamo o, T. (1995). S a-
is ical in e ence in ec o au o eg essions
wi h possibly in eg a ed p ocesses. Jou nal o
Econome ics, 66(1–2), 225–250. h ps://doi.
o g/10.1016/0304-4076(94)01616-8
Tucke , L. R., & Lewis, C. (1973). A eli-
abili y coe icien o maximum likelihood ac o
analysis. Psychome ika, 38(1), 1–10. h ps://
doi.o g/10.1007/b 02291170
Wondola, D. W., Aulele, S. N., & Lem-
bang, F. K. (2020). Pa ial leas squa e (PLS)
me hod o add essing mul icollinea i y p ob-
lems in mul iple linea eg essions (Case
s udies: Cos o elec ici y bills and ac o s a -
ec ing i ). Jou nal o Physics: Con e ence Se-
ies, 1463, The 5 h In e na ional Con e ence on
Basic Sciences, 1463(1), 012006. h ps://doi.
o g/10.1088/1742-6596/1463/1/012006
Zaidi, E. Z., Ahmed, R. R., & Raza, S.
(2022). Role o social media ma ke ing in
SME sec o pe o mance. T ans o ma ions in
Business & Economics, 21(56), 275–304.
Zhang, M. F., Dawson, J. F., & Kline, R. B.
(2020). E alua ing he use o co a iance-
based s uc u al equa ion modelling wi h
e lec i e measu emen in o ganiza ional and
managemen esea ch: A e iew and ecom-
menda ions o bes p ac ice. B i ish Jou nal
o Managemen , 32(2), 257–272. h ps://doi.
o g/10.1111/1467-8551.12415
