Dedunu, Ha shani; Wee asinghe, Salinda; Wickc amasinghe, Ananda
A icle
Reali y is di e en om wha we see: Knowledge
managemen and i m inno a ion
Jou nal o Inno a ion & Knowledge (JIK)
P o ided in Coope a ion wi h:
Else ie
Sugges ed Ci a ion: Dedunu, Ha shani; Wee asinghe, Salinda; Wickc amasinghe, Ananda (2025) :
Reali y is di e en om wha we see: Knowledge managemen and i m inno a ion, Jou nal o
Inno a ion & Knowledge (JIK), ISSN 2444-569X, Else ie , Ams e dam, Vol. 10, Iss. 3, pp. 1-13,
h ps://doi.o g/10.1016/j.jik.2025.100693
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/327594
S anda d-Nu zungsbedingungen:
Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen
Zwecken und zum P i a geb auch gespeiche und kopie we den.
Sie dü en die Dokumen e nich ü ö en liche ode komme zielle
Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich
machen, e eiben ode ande wei ig nu zen.
So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen
(insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en,
gel en abweichend on diesen Nu zungsbedingungen die in de do
genann en Lizenz gewäh en Nu zungs ech e.
Te ms o use:
Documen s in EconS o may be sa ed and copied o you pe sonal
and schola ly pu poses.
You a e no o copy documen s o public o comme cial pu poses, o
exhibi he documen s publicly, o make hem publicly a ailable on he
in e ne , o o dis ibu e o o he wise use he documen s in public.
I he documen s ha e been made a ailable unde an Open Con en
Licence (especially C ea i e Commons Licences), you may exe cise
u he usage igh s as speci ied in he indica ed licence.
h ps://c ea i ecommons.o g/licenses/by/4.0/
Reali y is di e en om wha we see: Knowledge managemen and
i m inno a ion
Ha shani Dedunu
a
, Salinda Wee asinghe
b
, Ananda Wickc amasinghe
c,*
a
Facul y o Managemen S udies, Raja a a Uni e si y o S i Lanka, College o Business Law and Go e nance, James Cook Uni e si y, Aus alia
b
School o Managemen , Facul y o Business and law, Queensland Uni e si y o Technology, Aus alia
c
School o Business, Facul y o Business and Law, Uni e si y o Wollongong, NSW, Aus alia
ARTICLE INFO
JEL code A ea o Speci ica ion:
M10
O30
O31
O36
Keywo ds:
Knowledge managemen
Inno a ion
Banking indus y
Non-linea i y
ABSTRACT
In oday’s knowledge-d i en business en i onmen , i m inno a ion hinges on e ec i e knowledge managemen .
O ganiza ions a e hus mo i a ed o con inuously c ea e and apply knowledge o sus ain compe i i e ad an age
h ough inno a ion. This s udy in es iga es how a ious knowledge managemen dimensions uniquely impac
i m inno a ion wi hin an in eg a ed amewo k, examining bo h linea and non-linea ela ionships—an
app oach no p e iously explo ed. Using a deduc i e, quan i a i e app oach, da a we e collec ed ia an online
su ey o 437 banking employees, wi h S uc u al Equa ion Modelling (SEM) employed o analyze quad a ic
ela ionships. Findings e eal ha knowledge c ea ion has an in e ed U-shaped ela ionship wi h i m inno-
a ion, while knowledge applica ion shows a U-shaped ela ionship. In con as , knowledge sha ing, applica ion,
and p o ec ion exhibi linea ela ionships, wi h knowledge sha ing being mos impac ul in d i ing inno a ion
wi hin he banking sec o . These esul s unde sco e ha o e es ima ing he impac o knowledge managemen
can be coun e p oduc i e, as i s dimensions do no consis en ly ollow linea pa hs. The s udy o e s c i ical
insigh s o managemen , pa icula ly in knowledge-in ensi e indus ies, o moni o and calib a e each knowl-
edge managemen dimension’s in luence on i m inno a ion o op imal pe o mance.
In oduc ion
Resea ch in es iga ing Knowledge Managemen (KM) is gaining
momen um in managemen li e a u e p ima ily because o he long-
s anding iew in academia ha managing knowledge is he nex on ie
o compe i i e ad an age in a knowledge-based en i onmen (Sang,
2024). The olume o li e a u e ela ed o KM inc eased a e he pio-
nee ing wo k o Nonaka (1994) and i s empi ical applica ion by leading
o ganisa ions, such as Apple, Ta a, and Google, o business inno a-
i eness and compe i i e ad an ages (Ve ma & Dixi , 2016). Conse-
quen ly, du ing he las ew decades, KM has been en iched e ically,
di used ho izon ally ac oss a ious managemen disciplines, and inally
de eloped as a sepa a e discipline in he ield o managemen .
This in ense a en ion o KM was ecei ed no only because o i s
impo ance as a discipline o managemen , bu also because o i s
con ibu ion o o ganisa ions as a ehicle o change and inno a ion.
The li e a u e a gues ha wi hou p ope knowledge managemen , i m
inno a ion does no occu o is signi ican ly delayed (Wee asinghe &
Dedunu, 2020). Fo example, in Hau e Cuisine and culina y se ices,
symbolic knowledge d i es he inno a ion p ocess, inspi ing che s wi h
c ea i e ideas, whe eas syn he ic knowledge connec s he che ’s idea
wi h scien is s, while analy ical knowledge p o ides suppo o subse-
quen science-based de elopmen (Albo s-Ga ig´
os e al., 2017). This
e idence implies ha he p ope managemen o knowledge leads o -
ganisa ions o achie e success ul inno a ion (Liu & Zeinaly, 2021; Yang
& Rui, 2009). Thus, he challenge o i ms is o ecognise co ec
knowledge and manage i o success ul inno a ion (du Plessis, 2007).
KM and inno a ion ha e been an in e es ing a ea o in es iga ion
o e he las ew decades. P e ious s udies ha e examined KM and
inno a ion (Basadu & Gelade, 2006; du Plessis, 2007; Salehi e al.,
2021; Wang e al., 2022); p oduc inno a ion pe o mance (Yus e al.,
2021); business model inno a ion (Bashi & Fa ooq, 2019); and, c ea-
i i y and inno a ion (As u i e al., 2022; Qandah e al., 2020); and
inno a ion managemen (B iones-Pe˜
nal e e al., 2020). Toge he , hese
s udies indica e ha he ela ionship be ween KM and inno a ion is
complex and con ex -d i en, and ha KM dimensions (knowledge
* Co esponding au ho .
E-mail add esses: [email p o ec ed], [email p o ec ed] (H. Dedunu), [email p o ec ed], [email p o ec ed]
(S. Wee asinghe), [email p o ec ed] (A. Wickc amasinghe).
Con en s lis s a ailable a ScienceDi ec
Jou nal o Inno a ion & Knowledge
jou nal homepage: www.else ie .com/loca e/jik
h ps://doi.o g/10.1016/j.jik.2025.100693
Recei ed 24 July 2024; Accep ed 11 Ma ch 2025
Jou nal o Inno a ion & Knowledge 10 (2025) 100693
A ailable online 20 Ma ch 2025
2444-569X/© 2025 The Au ho s. Published by Else ie España, S.L.U. on behal o Jou nal o Inno a ion & Knowledge. This is an open access a icle unde he CC
BY license (
h p://c ea i ecommons.o g/licenses/by/4.0/ ).
c ea ion, sha ing, applica ion, and p o ec ion) a e a ec ed by di e en
ac o s, esul ing in di e en and unp edic able in luences on inno a-
ion. Fo example, A sawan e al. (2022) emphasise ha knowledge
sha ing is in luenced by inno a ion cul u e. The ex en o which his
cul u e os e s inno a ion depends on s uc u ed space, au ho ised
space, he willingness o inno a e, and he in e play be ween leade ship
and social condi ions (Aue nhamme & Hall, 2013). Ri ala e al. (2022)
ind ha knowledge p o ec ion mode a es inc emen al inno a ion,
al hough i does no signi ican ly impac adical inno a ion. Dos a e al.
(2014) ind ha , especially in he banking sec o , knowledge om
cus ome s has a posi i e impac on a i m’s inno a ion capabili y and
business ope a ions, whe eas knowledge abou cus ome s and knowl-
edge o cus ome s ha e di e en e ec s. The in es iga ion by Campa-
nella e al. (2019) in o he ans o ma ion o aci knowledge in o
explici knowledge ound ha he dimensions o socializa ion, ex e -
naliza ion, combina ion, and in e naliza ion ha e di e en (posi i e and
nega i e) in luences on a bank’s economic alue c ea ion. The dynamics
o KM pose signi ican challenges when compa ing i s o e all e ec s on
inno a ion pe o mance.
Howe e , an unsol ed ques ion in he li e a u e is how hese KM
dimensions collec i ely con ibu e o a i m’s inno a ion while main-
aining hei ela ionships. Ahuja (2002); B a ianu e al. (2020); Yang
and Rui (2009) p o ide a mo e dynamic iew o explain how hese e-
la ionships become non-linea . Yang and Rui (2009) ound a U-shaped
ela ionship be ween knowledge acquisi ion and new p oduc c ea i i y
and an in e ed U-shaped ela ionship be ween knowledge dissemina-
ion and new p oduc c ea i i y. An examina ion o a ional, emo ional,
and spi i ual knowledge in he decision-making p ocess has also iden-
i ied his non-linea e ec (B a ianu e al., 2020). In pa icula , his
s udy no es ha emo ional and spi i ual knowledge, especially in
s uc u ed ields such as inance, plays a pe iphe al ole; acco dingly,
hei in luence on decision-making is indi ec . Ahuja (2002) emphasises
ha he ela i e alue o knowledge o inno a ion diminishes o e ime
when simila knowledge is con inuously c ea ed and dissemina ed.
Chesb ough and Rosenbloom (2002) emphasise ha when knowledge is
epea edly applied wi hin simila con ex s, i s no el y dec eases o e
ime. These indings call o a mo e dynamic pe spec i e o cap u e he
beha iou o knowledge dynamism, which is c ucial o managing
knowledge and educing he a bi a iness o knowledge in e en ions in
i m inno a ion (Schilpe oo d & Ah weile , 2014). I a i m ails o
unde s and his, i may no ully u ilise he alue o KM o inno a ion.
Despi e he inc easing numbe o s udies, he non-linea i y o KM di-
mensions has no ye been su icien ly add essed, lea ing a lack o
cla i y abou how hese dimensions, bo h indi idually and collec i ely,
con ibu e o inno a ion wi hin an in eg a ed amewo k. This s udy
aims o add ess his gap by ad ancing he li e a u e on he non-linea
beha iou o KM.
This s udy di e s om p e ious s udies in se e al ways. Fi s , i
challenges he con en ional belie ha KM and i m inno a ion ha e a
linea ela ionship and asse s ha KM’s impac o KM is no always
s aigh o wa d. This no el pe spec i e sheds ligh on he non-linea
dynamics o KM in inno a ion. Acco dingly, he s udy shows ha he
KM dimensions os e inno a ion in a non-linea ashion, e en hough
hei di ec impac on business inno a ion is waning, which is e olu-
iona y. Thus, we unde sco e he signi icance o comp ehending KM
wi hin an o ganiza ional amewo k. Second, in con as o p io ob-
se a ions (A sawan e al., 2022; Aue nhamme & Hall, 2013; Li e al.,
2018; McLeod e al., 2022; Ri ala e al., 2022) ha conside ed he in-
di idual dimensions o KM in inno a ion, his s udy deployed he en i e
KM cons uc wi h i s dimensions o comp ehend which dimensions
should be p omo ed and which should no . This ue na u e is no isible
when a dimension is isola ed o obse ed.
The emainde o he pape is s uc u ed as ollows. A e iew o he
li e a u e is o ganised a ound he heo e ical basis o he esea ch, ol-
lowed by he s udy hypo hesis. The esea ch app oaches and s eps
included in he s udy a e o ganised in he Me hodology sec ion. The
ou h chap e illus a es he da a analysis in de ail, and he las chap e
p o ides he implica ions o he s udy, ollowed by u u e esea ch a eas.
Li e a u e e iew
Inno a ion
Inno a ion, de ined as he applica ion o new solu ions o he ede-
sign o exis ing solu ions o mee no el equi emen s (Bai e al., 2014),
in ol es a speci ic skill o capabili y ha dis inguishes a i m om i s
compe i o s. Inno a i e i ms, such as Apple, Sony, and Google, a e
mo e compe i i e wi hin hei indus ies (Vega e al., 2012; Xu e al.,
2022) and ac i ely engage in inno a ion, os e o wa d hinking, and
con inuously edesign alue p oposi ions (Bai e al., 2014; Un & Asa-
kawa, 2015). Va ious heo e ical lenses, including open and closed
inno a ion, p o ide and demand-side inno a ion, and p oduc and
se ice inno a ion, as well as analy ical le els such as use s, indi iduals,
g oups, and i ms, ha e been used o explo e he concep o inno a ion
(McLeod e al., 2022; Ri ala e al., 2018; Vega e al., 2012). These s udies
indica e ha inno a ion is in e wined wi h knowledge. Nonaka (1994),
a pionee in KM, asse ed ha a i m’s inno a ion s ems om expanding
o enewing i s knowledge base by blending exis ing knowledge wi h
new insigh s, aligning wi h he e olu iona y economic pe spec i e o
inno a ion, whe e new knowledge is buil upon exis ing knowledge
(Coombs & Hull, 1998). These pe spec i es emphasise ha i m inno-
a ion emains wi hin he i m’s pu iew h ough e ec i e combina-
ions o new and exis ing knowledge; he e o e, knowledge managemen
becomes he d i ing o ce behind i m inno a ion.
Fi m inno a ion becomes inc emen al when new knowledge in e-
g a ion is con ingen on exis ing knowledge, wi h inno a ion in ol ing
mino changes in echnology, unc ionali y, appea ance, and pe o -
mance (Jugend e al., 2018). Wi hin his amewo k, inno a ion is
nei he a andom e en no spon aneous; ins ead, inno a ion un olds
g adually h ough inc emen al s eps in KM. This con inual imp o emen
allows he i m o s eng hen i sel and adap p omp ly o changing
ci cums ances (Coombs & Hull, 1998; Kodama, 2017). The opposing
iew sugges s ha he in eg a ion o no el knowledge, which signi i-
can ly di e ges om exis ing knowledge, esul s in b eak h ough ideas
ha lead a i m owa ds adical inno a ion (Ri ala e al., 2018).
Pa icula ly in he banking indus y, bo h o ms o inno a ion o en
occu ac oss p oduc , p ocess, se ice, and echnology dimensions,
bene i ing indi idual i ms and he wide communi y, including s ake-
holde s (Bai e al., 2014; Un & Asakawa, 2015). D awing on his dis-
cussion, he p esen s udy ocuses on p oduc , p ocess, se ice, and
echnology inno a ion in banks, gi en ha he compe i i e na u e o he
indus y necessi a es simul aneous engagemen in hese ypes o inno-
a ion compa ed wi h less compe i i e indus ies.
Knowledge managemen
Knowledge, a widesp ead concep in socie y, e lec s an indi idual’s
belie abou eali y (Nonaka, 1994), is de i ed om pe sonal expe i-
ences and educa ion, and esides in a pe son’s mind (Ha ing on e al.,
2019; L´
opez-To es e al., 2019). F om he knowledge managemen
pe spec i e, knowledge appea s in wo o ms: aci and explici
(Nonaka, 1994). “Explici knowledge” e e s o knowledge ha is
codi ied in many o ma s, such as books, magazines, and a icles ha
can be ansmi ed in o mal and sys ema ic ways, whe eas “ aci
knowledge” e e s o knowledge ha is deeply oo ed in ac ion and
beha iou in a speci ic con ex , which is blended wi h pe sonal quali ies;
he e o e, i is ha d o o malise and communica e (Nonaka, 1994).
Nonaka s a es ha aci and explici knowledge can be con e ed in o
use ul o ganisa ional knowledge h ough socialisa ion, ex e nalisa ion,
combina ion, and in e nalisa ion (Liu e al., 2019).
Fi m knowledge, oo ed in he o ganiza ional s uc u e and
possessed by indi idual employees, is ecognised as a p ima y sou ce o
H. Dedunu e al.
Jou nal o Inno a ion & Knowledge 10 (2025) 100693
2
i m compe i i eness and inno a ion (Sang, 2024). Fi m knowledge
con ibu es o and acili a es syne gy, p oac i e lea ning, and c ea i e
p oblem sol ing h ough he unique in eg a ion o aci and explici
knowledge. De eloping a dis inc i e knowledge base and applying such
knowledge o o ganiza ional pe o mance is c i ical, especially in a
knowledge-based economy whe e knowledge dispa i y shapes a i m’s
compe i i e posi ion (Tassabehji e al., 2019).
F om a schola ly pe spec i e, KM is de ined as a mechanism by
which o ganisa ions access, sha e, apply, and s o e knowledge, he eby
c ea ing new knowledge and capabili ies ha sus ain inno a ion (du
Plessis, 2007; Wee asinghe & Sede a, 2023). The e o e, KM is a complex
p ocess wi h dis inc dimensions. Despi e nume ous s udies on his
concep , schola s ha e no ye eached an ag eemen on wha cons i u es
KM, p ima ily because o he a ying in e p e a ions o i s co e unc-
ions. Fo example, some schola s a gue ha knowledge c ea ion in-
cludes bo h c ea ion and cap u e o knowledge, whe eas many s udies
ea hese as wo dis inc dimensions o KM (Reza Rasekh e al., 2014;
Shu e al., 2012). While en iching he dep h and b ead h o KM, his
di e si y in in e p e a ion subs an ially diminishes he cons uc ’s uni-
o mi y in he li e a u e, adding ex aneous meaning o KM.
D awing on he co e meaning o KM (Nonaka, 1994), he na u e o
he cons uc i sel (Coombs & Hull, 1998), and ecen applica ions like,
Ode and Aya oo (2020); Shehzad e al. (2022); Ting e al. (2021);
Wee asinghe and Sede a (2023), his s udy iden i ies ou main di-
mensions o KM: knowledge c ea ion, knowledge sha ing, knowledge
applica ion, and knowledge p o ec ion, using a p ocess lens. These di-
mensions a e widely ecognised as he p ima y unc ions o he KM
p ocess. Technically, he p ocess begins wi h knowledge c ea ion and
concludes wi h knowledge p o ec ion; howe e , he cyclical na u e o
KM illus a es ha knowledge p o ec ion, as he inal s ep, gi es ise o
new knowledge c ea ion (Coombs & Hull, 1998). The spi al na u e o
KM p omo es exponen ial g ow h h ough he c ea i e applica ion o
knowledge o inno a ion. Howe e , he e icacy o KM in d i ing i m
inno a ion depends on he ela ionships be ween hese cons uc s and
hei espec i e dimensions.
Knowledge managemen and inno a ion: ela ionship na u e
The long-s anding discussion be ween inno a ion and KM has
demons a ed he i id na u e o hese ela ionships in he li e a u e.
Mos o hese ela ionships a e p edominan ly linea , di ec , o indi ec .
Fo example, Ting e al. (2021) s a e ha knowledge-managemen
in as uc u e ( echnology, cul u e, and s uc u e) and
knowledge-managemen p ocesses (knowledge c ea ion, sha ing, and
u ilisa ion) bo h ha e s a is ically signi ican and posi i e e ec s on i m
inno a ion pe o mance. Simila KM beha iou has been ound o g een
inno a ion (Wang e al., 2022). Thus, g een KM di ec ly a ec s a i m’s
sus ainable compe i i e ad an age h ough g een inno a ion capabil-
i ies. Howe e , Shehzad e al. (2022) s a e ha knowledge c ea ion has
an insigni ican e ec on a i m’s g een p oduc and p ocess inno a ion
compa ed o he e ec s o knowledge acquisi ion, sha ing, and appli-
ca ion. This s udy highligh s he ac ha no all dimensions o KM a e
equally impo an o i m inno a ion, emphasising he need o ca e ul
manage ial a en ion in managing knowledge o inno a ion. Howe e ,
mos cu en schola ship is s ill based on he assump ion ha KM has an
inc easing e u n ela ionship wi h i m inno a ion, and his ideology
has s ee ed hem o explo e one side o he ela ionship (linea
ela ionships).
This assump ion is challenged by Bloom e al. (2020) who explain
ha al hough esea ch e o s a e inc easing subs an ially ac oss ields,
esea ch p oduc i i y sha ply declines o e ime. This sugges s ha , in
b oade e ms, knowledge managemen (wi h esea ch as he p ima y
o m) may yield diminishing e u ns in p ac ical applica ions. Rega ding
new p oduc c ea i i y, Yang and Rui (2009) ound a non-linea ela-
ionship, speci ically a U-shaped link be ween knowledge acquisi ion
and new p oduc c ea i i y, and an in e ed U-shaped ela ionship
be ween knowledge dissemina ion and new p oduc c ea i i y. Thei
s udy highligh s ha he impac o KM is no always p edic able h ough
a linea unc ion; a he , i may in ol e a combina ion o posi i e and
nega i e e ec s, wi h bo h inc easing and diminishing e u ns.
Fo ins ance, conside inno a ions ela ed o elec ic ehicles (EVs)
in he au omobile indus y. In he ea ly s ages, he indus y aced limi ed
knowledge, pa icula ly ega ding ba e y echnology, which con-
s ained o e all EV inno a ion o elec ic ehicles. Howe e , as i ms
such as Tesla and Toyo a del ed deeply in o new knowledge a eas (e.g.
li hium-ion ba e ies), he indus y’s inno a i e capaci y in elec ic e-
hicles su ged, demons a ing he inc easing e u ns o KM on inno a-
ion. Howe e , his ex ensi e applica ion o KM has now begun o show
diminishing e u ns as he sa u a ion o cu en knowledge esul s in
only mino inc emen al impac s on i m inno a ion ela i e o
in es men .
This diminishing e u n o knowledge was explained by Ahuja
(2002) in he con ex o knowledge sha ing. Acco dingly, hey s a e ha
knowledge sha ing in i s ea ly s ages leads i ms o achie e signi ican
b eak h oughs ( adical inno a ion); howe e , hey no e ha his e ec
diminishes o e ime as i ms accumula e simila knowledge. Simila ly,
Chesb ough and Rosenbloom (2002) emphasise ha when knowledge is
epea edly applied wi hin simila con ex s, no el y dec eases, leading o
educed inno a ion ou pu . Nambisan and Zah a (2016) explain ha
knowledge managemen in he con ex o oppo uni y o ma ion is
non-linea in complex indus ies such as he au omo i e indus y
because o he in ica e, i e a i e, and dynamic na u e o demand-side
na a i es, which, in u n, in luence he non-linea pa e n o inno a-
ion g ow h. The li e a u e emphasises ha he e ec o knowledge
managemen on inno a ion is dynamic, leading o a ying ou comes.
Despi e he indi idual e ec s o knowledge managemen dimensions
demons a ing non-linea i y o e ime, hei beha iou wi hin an in e-
g a ed amewo k has been signi ican ly o e looked in he con ex o
i m inno a ion. Acco dingly, conside ing he p esen ela ionship be-
ween knowledge managemen dimensions and i m inno a ion, his
s udy de elops a concep ual amewo k, as illus a ed in Fig. 1.
Dimensions o knowledge managemen
Knowledge c ea ion (KC)
Knowledge c ea ion en ails cap u ing and de eloping he necessa y
knowledge and pu ing i in a o m ha may be use ul o applica ions
(Ala i & Leidne , 2001). Acco dingly, KC has wo unc ions: knowledge
cap u e and knowledge de elopmen . Knowledge cap u e o en e e s o
assimila ing o acqui ing knowledge om ex e nal sou ces (Zhou & Li,
2012), whe eas knowledge de elopmen in ol es he p oduc ion o
in-house knowledge (Ala i & Leidne , 2001). This s udy inco po a ed
bo h ex e nally cap u ed and de eloped knowledge in o knowledge
c ea ion. In a dynamic en i onmen , new knowledge c ea ion is
pa icula ly impo an , as in many ins ances, such as
echnology-o ien ed consume s, and he exis ing knowledge o a i m
may all sho o mee ing he e ol ing demand in he ma ke , as i ms
canno simply imi a e o eplace knowledge o in oduce inno a i e
p oduc s (Chang e al., 2014; Ola a ie a & F iedmann, 2008).
Assimila ing ex e nal knowledge enables a i m o unde s and
exis ing and p ospec i e ma ke signals ha e en ually change i s na -
u al e olu ion and pa h dependencies (Vogel & Gü el, 2012). Howe e ,
a i m’s in-house knowledge gene a ion upda es i s in e nal capabili ies
and compe encies, econ igu ing i s esou ce base o ex e nal inno a-
i e changes (Basadu & Gelade, 2006; Zhou & Li, 2012). Inno a ion
o en add esses unme needs and exis ing challenges (Wang e al., 2022).
KC assis s i ms in his con ex in apping eme ging p oblems in he
indus y, unde s anding hei oo causes, and a icula ing c ea i e so-
lu ions by in eg a ing new and exis ing knowledge. The e o e, i is
possible o an icipa e a ela ionship be ween KC and inno a ion
(Alshan y & Emeagwali, 2019).
The na u e o he ela ionship be ween KC and inno a ion is
H. Dedunu e al.
Jou nal o Inno a ion & Knowledge 10 (2025) 100693
3
impo an o comp ehending he dynamism o in luence. Fo example,
Papa e al. (2018) iden i y a posi i e in luence o knowledge acquisi ion
on i m inno a ion; howe e , his ela ionship is mode a ed by human
esou ce p ac ices. The s udy no ices ha a i m inno a i e clima e
esul ed om HR lexibili y and eels employee ee o sha e inno a i e
ideas and isions (Papa e al., 2018). Shu e al. (2012) ound ha
knowledge acquisi ion is mo e ex e nally o ien ed and makes a dis inc
con ibu ion o i m and p oduc inno a ion han o p ocess inno a ion.
Yang & Rui, (2009) ound ha his con ibu ion is no always consis en
wi h i m inno a ion, wi h a U-shaped ela ionship be ween knowledge
acquisi ion and new p oduc c ea i i y. Conside ing he ela ionship
be ween c ea i i y and inno a ion, i is easonable o assume ha
knowledge c ea ion has a linea o non-linea ela ionship wi h i m
inno a ion. Acco dingly, his s udy de elops he ollowing hypo hesis:
H1: The knowledge c ea ion has a signi ican linea o non-linea ela-
ionship wi h i m inno a ion.
Knowledge sha ing (KS)
KS in ol es he exchange o ideas, expe iences, and knowledge
among membe s o an o ganisa ion o ensu e ha he igh knowledge is
gi en o he igh pe son a he igh ime o inno a i ely pe o m he
igh ask (Saenz e al., 2012). This is a social p ocess ha akes place
e ically (among di e en le els) and ho izon ally (a he same le el)
among employees in an o ganisa ion, allowing hem o c i ically e al-
ua e exis ing pa e ns o wo k and make necessa y al e a ions o
inno a i e imp o emen s (Raelin, 1998). This comp ehensi e p ocess
minimises wo k duplica ion, epe i ion, and e en he cos o he o ga-
nisa ion by co-c ea ing no el solu ions (Liu & Zeinaly, 2021), and
ans e ing he s a egic knowledge equi ed o sha pen i m inno a-
ion (Zhao e al., 2020). Howe e , he e ec o KS on inno a ion is
con o e sial because o he quali y o sha ed knowledge and em-
ployees’ willingness o engage (Dye & Nobeoka, 2000; Vacca o e al.,
2010).
The exis ing li e a u e es ablishes a solid ounda ion o unde -
s anding he ela ionship be ween KS and inno a ion. Saenz e al. (2012)
ound ha pe sonal-in e ac ion-based KS and knowledge-embedded
managemen p ocesses signi ican ly in luenced new idea gene a ion
and inno a ion p ojec managemen . Zhao e al. (2020) no e ha bo h
inbound and ou bound KS con ibu e o inno a ion. While ou bound KS
os e s inno a ion di ec ly, inbound KS os e s inno a ion indi ec ly.
A sawan e al. (2022) s a e ha KS especially suppo s small in
achie ing compe i i e ad an ages by c ea ing an inno a i e cul u e. Xia
e al. (2021) s a e ha when cul u e p omo es ask o ien a ion, ICT
applica ion, and eam disposi ion in an o ganisa ion, collabo a i e KS
becomes mo e e icien in inno a ion. The downside o KS is also e iden
in he li e a u e, which emphasises ha inance, insu ance, and eal
es a e a e majo indus ies in which pseudo-knowledge sha ing exis s
compa ed o o he s (Came on Cock ell & S one, 2010). The s udy
u he emphasises ha his nega i e e ec can be uled ou by
es ablishing mo i a ion o knowledge-sha ing and p o iding inancial
incen i es. Ma ín C uz e al. (2009) poin ou ha when employees a e
in insically mo i a ed, hey end o engage in KS in he inno a ion
p ocess mo e han when hey a e ex insically mo i a ed. Howe e ,
Ahuja (2002); Tasi (2001) emphasise ha he signi ican e ec o KS on
inno a ion can be seen only a i s ini ial s age because i ms become
sa u a ed wi h knowledge when simila in o ma ion is sha ed, and hen
inno a ion gains diminish. Conside ing he na u e o KS in inno a ion,
his s udy es ablishes he second hypo hesis.
H2: Sha ing o knowledge has a signi ican linea o non-linea ela-
ionship wi h i m inno a ion.
Knowledge applica ion (KA)
KA, he u ilisa ion o knowledge gained h ough KC and KS (Ahuja,
2002), is he co e o KM, as knowledge pe se does no b ing any alue o
anyone wi hou i s p ope applica ion. Thus, KA is he ue use o
knowledge o he be e men o an o ganisa ion. Knowledge is pu in o
ac ion by an agency (e.g. an employee o , in a e ins ances, echnology
such as a cha box, Si i, o an au oma ed sys em). I in eg a es i id
knowledge om a ious sou ces in a dis inc i e manne . (Shin e al.,
2001). Thus, o ganisa ions mus ha e sys ems in place o acqui e he
co ec in o ma ion and mechanisms o deploy in o ma ion aligned wi h
he o ganisa ion’s goals and objec i es (Almuayad e al., 2024).
KA is an inno a i e p oblem-sol ing mechanism which dis inguishes
a i m om i s i al h ough inno a i e knowledge applica ions. The
di e si y o KA b ings abou di e en ou comes h ough a simila se o
knowledge inpu s depending on he con ex , en i onmen al dynamics,
and e icacy o applica ion (Almuayad e al., 2024) allowing he i m o
gain a sus ainable compe i i e ad an age. Consequen ly, knowledge is
in aluable wi hou p ope applica ion (Almuayad e al., 2024). In
pa icula , Eisenha d and Ma in (2000) s a e ha esou ces a e ine
and managemen needs o ac upon hem o ha e an e ec . This
pe spec i e emphasises ha inno a ion eme ges h ough he p ope
applica ion o knowledge by managemen (Li e al., 2009; Ode & Aya-
oo, 2020). Inno a ion is inhe en ly linked o isk, and he applica ion o
new knowledge o inno a ion in ol es unce ain y, o en esul ing in
unp edic able ou comes (Allen, 2013). Concening KA, Ahuja (2002)
emphasise ha i ms ha apply exis ing knowledge o inno a e o en
ini ially achie e high e u ns. Howe e , as knowledge is epea edly
applied wi hin simila con ex s, no el y diminishes, leading o a dimi-
nu ion in he e u n on KA. Simila ly, Chesb ough and Rosenbloom
(2002) ound a diminishing e u n on KA in p oduc de elopmen . Based
on he exis ing discussion, he hi d hypo hesis was de eloped as
ollows:
H3: Knowledge applica ion has a signi ican linea o non-linea ela-
ionship wi h i m inno a ion.
Knowledge p o ec ion (KP)
KP e e s o he ex en o which i ms employ speci ic p ocesses o
Fig. 1. Concep ual amewo k.
H. Dedunu e al.
Jou nal o Inno a ion & Knowledge 10 (2025) 100693
4
go e n and sa egua d p op ie a y knowledge (Nielsen & Nielsen, 2009;
No man, 2002). P e ious esea ch has con ended ha knowledge should
be s o ed and managed in u u e applica ions. Mos KM li e a u e ad-
oca es KP unde in ellec ual p ope y igh s (IPRs) (Bhuk a, 2020;
B ans e e e al., 2006; Olande e al., 2014; Oli e & Sapi , 2017;
Samaniego, 2013), including con ac s, pa en s, adema ks, and copy-
igh s. Howe e , ob aining IPRs o all ypes o knowledge, pa icula ly
ope a ional, is imp ac ical. The ocus o his s udy is no o p o ec
knowledge using in ellec ual p ope y; a he , i ocuses on sa egua ding
(e.g. documen ing, in e nalising knowledge o cul u e) cu en knowl-
edge ( aci and explici ) o u u e applica ions. Knowledge p ese a ion
main ains a i m’s dis inc i eness and acili a es knowledge ans e
ac oss gene a ions (Olande e al., 2014).
KP enhances i ms’ abili y o o ge no el solu ions o eme ging
challenges (Lee & Choi, 2003). I anscends pas knowledge and lays a
s ong ounda ion o i m inno a ion (Ode & Aya oo, 2020). Despi e
b eak h ough inno a ions, mos inc emen al inno a i e solu ions a e
g ounded in pas o p esen knowledge bases. The e o e, e ec i e KP is
c ucial o con inuous inno a ion. Ode & Aya oo (2020) ound ha KA
has a signi ican e ec on i m inno a ion. Despi e i ms employing
di e en KP measu es o become compe i i e in in e - i m inno a ion,
ex ending cul u al alues and i m policies p esen challenges (Dona e &
Guadamillas, 2010). Ri ala e al. (2022) also s a e ha KP nega i ely
mode a es enewal capi al and i m inc emen al inno a ion. Howe e ,
in he long un, excessi e eliance on pa en p o ec ion can limi i ms’
ex e nal collabo a ion and knowledge low (A o a & Ceccagnoli, 2006),
he eby slowing i m inno a ion and diminishing he e u ns o KP.
Building on his ounda ion, we de elop ou ou h hypo hesis:
H4:Knowledge p o ec ion has a signi ican linea o non-linea ela ion-
ship wi h i m inno a ion.
Me hodology
S udy con ex popula ion and he sample
As a apidly de eloping indus y in he se ice economy, he
implemen a ion o e icien KM has become essen ial o inancial in-
s i u ions o ensu e compe i i e alue c ea ion, while p o ec ing hei
inno a i e po en ial (Campanella e al., 2019). The dynamic na u e o
p oduc s uc u es, indus y compe i ion, and inc eased consume
knowledge has challenged banks’ adi ional oles, pushing hem o
inno a e in hei knowledge applica ions (Dos a e al., 2014; Sang,
2024). Banks ha e become ho spo s o inno a ion o new p oduc s,
se ices, and echnological applica ions. KM in banks has been
inc easingly emphasised owing o inancial ins i u ions’ suscep ibili y o
a ious isks, such as c edi de aul , ma ke ola ili y, and ope a ional
b eakdowns, all o which necessi a e e ec i e knowledge managemen
(Sang, 2024). Agains his backd op, his s udy ocuses on he banking
indus y.
This empi ical s udy is based on he S i Lankan banking indus y, a
well-es ablished knowledge-in ensi e indus y. A i m becomes
knowledge-in ensi e when i in es s signi ican ly in R&D o skilled la-
bou (Yang & Rui, 2009). Mos S i Lankan banks main ain in-house R&D
and ongoing collabo a ion wi h uni e si ies/ex e nal esea ch ins i u es
(Wee asinghe & Sede a, 2023), as well as adop ing a special ec ui men
scheme o di ec ly abso b skilled g adua es om uni e si ies. The in-
dus y consis s o 24 licenced banks, and we a ge ed he op en banks
in e ms o se ice capaci y. We excluded banks ha ope a ed only in
egional a eas and hose ha ope a ed only in he capi al ci y. Acco d-
ingly, ou online su ey a ge ed employees om 10 selec ed banks
(annexu e: 01). The sample size was de e mined using he PLS-SEM
sampling ma ix and Mo gan able. The PLS Ma ix equi es 70 em-
ployees (Joseph F. Hai e al., 2016), whe eas he Mo gan able equi es
382 employees (Seka an & Bougie, 2016). Wi h a 50 % esponse
expec a ion, we conduc ed an online su ey using o icial employees’
Wha sApp g oups. We app oached he egional bank managemen and
asked hem o dis ibu e ou su ey o selec ed b anches. The su ey
began in he i s week o Sep embe 2022. The i s eminde was gi en
a week la e , ollowed by a second eminde a e wo days h ough he
same channel. To p e en edundancy, a message s a ing, “Igno e his
message i you ha e al eady con ibu ed o he su ey o m” was p e-
sen ed. To con i m he ep esen a i eness o he sample ela i e o he
popula ion, an analysis o non- esponse bias was conduc ed. Howe e ,
no signi ican di e ences we e iden i ied be ween hose who esponded
a he beginning compa ed o hose who esponded a he end.
Measu emen o a iables
The s udy deploys al eady alida ed scales o measu e ou i ems
each o KC (Ting e al., 2021; Yang & Rui, 2009), KS (Liu & Zeinaly,
2021; Zhao e al., 2020), KA (Kim & Lee, 2010), and KP (Manha &
Thalmann, 2015; Olande e al., 2014), and i e i ems o inno a ion
(Liu & Zeinaly, 2021; Zhao e al., 2020) wi h sligh modi ica ions
conside ing he banking indus y, a ge audience, and expe sugges-
ions. Expe sugges ions we e ecei ed by sending a ques ionnai e o
senio esea che s in he ield. The scale anged om one o se en. The
32 inalised i ems a e lis ed below he esponden s’ a igue s a is ics
(Dassanayaka e al., 2022).
Con ol a iables
This s udy con olled o employee c ea i i y, age, and gende .
C ea i i y has a signi ican in luence on i m inno a ion, whe eas
employee age and gende a ec inno a ion h ough accumula ed
expe ience and willingness o ake isks (Gius iniano e al., 2016; Lee
e al., 2019). Thus, con olling o hese e ec s helps de e mine he eal
e ec o KM on i m inno a ion. The s udy uses he c ea i i y scale o
Kim and Lee (2010) and measu es i using ou i ems on a se en-poin
scale, whe e employee gende (0 =male, 1 = emale) and age a e
measu ed by disc e e alues.
T ea men o common me hod bias
The s udy used Podsako e al.’s (2003) ecommenda ions o p e en
a possible common me hod bias when collec ing da a om one in o -
man . Hence, we di e en ia ed measu emen scales o p edic o and
c i e ion a iables, sa egua ded esponden anonymi y, mi iga ed e al-
ua ion app ehension h ough di e se ques ions, and enhanced scales by
(a) main aining simplici y, speci ici y, and conciseness in he ques ions;
(b) excluding ague concep s; (c) elimina ing double-ba elled ques-
ions; and (d) cla i ying ambiguous o un amilia e ms. A common
me hod bias a ises when one ac o accoun s o he majo i y o he
co a iance in he da ase (Reio, 2010). Ou con i ma o y ac o analysis
con i ms ha he e a e six dis inc ca ego ies o a iables in he da ase
(annexu e 02), a common me hod bias p oblem in he da ase is mos
unlikely.
Analysis and discussion o he esul s
The collec ed da a we e igo ously cleaned be o e analysis (Joe F.
hai e al., 2020). This in ol ed add essing pa e ned esponses,
incomple e submissions and missing da a. Pa e ned esponses we e
emo ed o p e en po en ial dis up ion o he genuine e ec s o he
da ase . Missing alues we e handled by subs i u ing hem wi h he
means o he esponses (Joe F. hai e al., 2020). In o al, 437 comple ed
esponses we e included in he inal analysis. Pos -cleaning, he box plo
examina ion e ealed no ou lie s. Skewness and excess ku osis alues
ell wi hin he -1 o +1 ange, wi h sligh excep ions o KP_2 and F_2.
Thus, we in e ed ha he da ase was no mally dis ibu ed and sui able
o in e en ial analysis. In e ms o he sample cha ac e is ics, esponses
we e p edominan ly om he Bank o Ceylon (21.7 %), ollowed by he
People’s Bank (19.5 %), and he leas om HSBC Bank (1.6 %),
H. Dedunu e al.
Jou nal o Inno a ion & Knowledge 10 (2025) 100693
5
indica ing a highe esponse a e om go e nmen -sec o banks
compa ed o p i a e-sec o banks ope a ing in he coun y. The majo i y
(66.1 %) o esponden s we e aged be ween 26 and 35 yea s and held
bachelo ’s deg ees (38.5 %). This age g oup e lec s he ac i ely engaged
wo k o ce segmen , which is likely o ha e conside able expe ience and
g ow h po en ial in hei oles. Non-manage ial le el esponden s
cons i u ed 59 % o he sample, wi h a p edominance o males (57.9 %).
This may in luence he s udy ou comes by emphasising he pe spec i es
o male-dominan ope a ional employees o e s a egic-le el
employees.
Measu emen model e alua ion
This s udy employed a S uc u al Equa ion Modeling (SEM)
app oach u ilising Sma PLS so wa e. Ou selec ion o Sma PLS was
based on se e al easons. Fi s , Sma -PLS is a powe ul app oach o
examining complex models ha in ol e he ela ionships be ween la en
a iables, mode a o s, and media o s (Noonan, 2017). Second,
Sma -PLS is ecommended o la ge sample sizes (La an, 2018), and
he sample size should be la ge enough o exceed he minimum
PLS- ecommended sample size. Thi d, i p o ides ad anced ea u es (e.
g. quad a ic ela ionships and impo ance-pe o mance maps) and is
highly applicable o p edic ion- ocused esea ch (Noonan, 2017). As his
s udy includes a complex model, a la ge sample, and a quad a ic
analysis, we con end ha Sma -PLS is an app op ia e choice o he
objec i es o his s udy.
Ou analysis comp ises wo s ages: (1) he de elopmen o he
measu emen model and (2) he de elopmen o he s uc u al model.
The measu emen model dic a es he la en cons uc s o he s uc u al
model (Hana iah, 2020). A e lec i e measu emen app oach was
applied o he measu emen model, emphasising ha i ems in he
ques ionnai e a e in luenced by hei espec i e la en cons uc s, and
any al e a ion in a la en cons uc equi es a co esponding change in
he espec i e i ems (Hai e al., 2020). To assess he e lec i e mea-
su emen model, his s udy ollowed he se en s eps ou lined by Hai
e al. (2020). These s eps encompass: 1. es ima ion o loadings and
signi icance; 2. indica o eliabili y; 3. composi e eliabili y; 4. a e age
a iance ex ac ed; 5. disc iminan alidi y; 6. nomological alidi y; and
7. p edic i e alidi y.
As illus a ed in Table 1, po en ially oublesome measu es wi h low
i em loadings below he 0.7 h eshold a e iden i ied by he loading es-
ima e (Hai e al., 2020), such as KC_1 (0.649), KP_4 (0.669), and I_1
(0.544). Because he i em loadings o (KC_1) and (KP_4) a e close o 0.7,
hey we e conside ed o he analysis (Saunde s e al., 2009). Howe e ,
i em (I_1) was emo ed om u he analysis. The alues o C onbach’s
alpha and composi e eliabili y a e ound be ween 0.7 and 0.95, indi-
ca ing he in e nal consis ency o he i ems being used. The AVE also
alls be ween he anges o 0.5 and 1, ensu ing he con e gen alidi y o
he model.
The s udy ensu ed disc iminan alidi y h ough he Fo nell-La cke
es , as illus a ed in Table 2, and i em c oss-loadings in Annexu e 02.
These a iables ensu e disc iminan alidi y when he sha ed a iance
wi h he cons uc exceeds he sha ed a iance be ween cons uc s (Hai
e al., 2020). Acco ding o he Fo ne -La cke es , he squa e oo o he
cons uc ’s AVE (sha ed a iance wi h cons uc ) was g ea e han i s
highes co ela ion wi h any o he cons uc . This ensu ed disc iminan
alidi y o he da ase . Addi ionally, as indica ed in Annexu e 02, he
ou e loadings o he indica o a e highe han i s c oss-loadings wi h he
o he cons uc s, excep C_2. We u he examined he Va iance In la ion
Fac o (VIF) o e alua e collinea i y issues. A model encoun e s a
Table 1
I em loading, c oss loading, con e gen alidi y.
Va iable Indica o s Desc ip i e Con e gen alidi y In e nal consis ency
Mean O e all
mean
Ou e loadings AVE Composi e eliabili y C onbach
alpha
Range >0.7 >0.5 >0.7 0.7- 0.95
Knowledge C ea ion KC_1 5.078 5.015 0.656 0.544 0.728 0.719
KC_2 5.199 0.765
KC_3 4.725 0.733
KC_4 5.034 0.789
Knowledge Sha ing KS_1 5.201 5.385 0.759 0.638 0.826 0.813
KS_2 5.325 0.797
KS_3 5.291 0.821
KS_4 5.497 0.817
Knowledge Applica ion KA_1 5.330 5.542 0.845 0.738 0.884 0.882
KA_2 5.533 0.852
KA_3 5.556 0.867
KA_4 5.751 0.873
Knowledge P o ec ion KP_1 5.842 5.551 0.749 0.574 0.765 0.753
KP_2 5.693 0.792
KP_3 5.080 0.812
KP_4 5.590 0.671
C ea i i y C_1 5.384 5.378 0.745 0.579 0.778 0.759
C_2 5.451 0.824
C_3 5.270 0.737
C_4 5.410 0.733
Inno a ion I_1 5.722 0.544*0.582 0.760 0.758
I_2 5.577 0.712
I_3 5.588 0.704
I_4 5.773 0.781
I_5 5.952 0.783
*
Remo ed om he analysis.
Table 2
Fo nell la cke c i e ion.
Va iables IN CR KS KC KP KA
Inno a ion 0.710
C ea i i y 0.756 0.760
Knowledge Sha ing 0.696 0.587 0.799
Knowledge C ea ion 0.710 0.688 0.611 0.738
Knowledge P o ec ion 0.459 0.413 0.381 0.485 0.756
Knowledge Applica ion 0.637 0.513 0.471 0.531 0.363 0.859
IN: Inno a ion, CR: C ea i i y, KS: Knowledge Sha ing, KC: Knowledge C ea ion,
KP: Knowledge P o ec ion, KA: Knowledge Applica ion
H. Dedunu e al.
Jou nal o Inno a ion & Knowledge 10 (2025) 100693
6
collinea i y issue when he VIF alue exceeds 5 (Hai e al., 2020). Ac-
co ding o he da a, he VIF alue anges om 1.00 - 3.18, which illus-
a es ha he e a e no mul icollinea i y issues in he da a se .
When a en ion is paid o he desc ip i e s a is ics p esen ed in
Table 1, he o e all mean alues o all he conside ed a iables ep esen
alues abo e he ‘ag ee’ le el o he se en-poin scale. This implies ha
KC, KS, KA, KP, inno a ion, and employee c ea i i y a e abo e a e age
in he S i Lankan banking indus y.
S uc u al model: s age wo
The s uc u al model uses a boo s ap esampling app oach wi h
5000 subsamples (Joseph F. Hai e al., 2016), and he ou pu o he SEM
analysis is illus a ed in Fig. 2 and Table 3. The model i s a is ics
indica ed ha SRMR =0.087, d_ULS =2.646, d_G =1.049, chi-squa e =
2404.793, and NFI =0.618. As illus a ed by Hai e al. (2020), his
s udy i s con i ms he absence o mul icollinea i y among highe -o de
cons uc s h ough he VIF alue, which anges below i e. Second, he
p edic i e capabili y o he s uc u al model was assessed using he
coe icien o de e mina ion (R
2
), e ec size (
2
), and blind olding (Q
2
).
As he able indica es, he model’s R
2
was 72.7 %, indica ing ha 72.7 %
o he a ia ion in i m inno a ion was explained by he a iables
conside ed in he model. E ec size illus a es he p edic i e powe o
each independen a iable. When
2
>0.35, he e ec is high; 0.35 >
2
>0.15, he e ec is medium, and he e ec becomes low i
2
<0.02
(Hai e al., 2020). As shown in he able, KS (
2
=0.231), inno a ion (
2
=0.190), and KA (
2
=0.132) showed he highes e ec sizes on i m
inno a ion p ocesses. The e ec o hese a iables compa ed o he es
is g aphically isible h ough he Impo ance Pe o mance Map (IPM)
Fig. 2. S uc u al equa ion model.
Table 3
Ou pu o s uc u al equa ion modelling.
Pa h Pa h
coe icien
T S a is ics
(IO/STDEVI)
2
CI (95 %) Q2
KC ->
Inno a ion
0.048 1.079 0.002 (-0.035,
0.135)
0.694
KS ->
Inno a ion
0.359*9.704 0.231 (0.287,
0.431)
KA ->
Inno a ion
0.263*7.209 0.132 (0.190,
0.334)
KP ->
Inno a ion
0.098*2.775 0.022 (0.031,
0.166)
QE (KC) ->
Inno a ion
(-0.049)*2.179 0.011 (-0.093,
-0.004)
QE (KS) ->
Inno a ion
0.022 1.145 0.002 (-0.015,
0.058)
QE (KA) ->
Inno a ion
0.129*5.482 0.061 (0.084,
0.176)
QE (KP) ->
Inno a ion
0.034 1.399 0.007 (-0.015,
0.081)
Age ->
Inno a ion
0.085*3.060 0.024 (0.029,
0.139)
C ea i i y ->
Inno a ion
0.345*10.117 0.190 (0.277,
0.412)
Gende ->
Inno a ion
0.006 0.114 0 (-0.087,
0.105)
SRMR =0.087, d_ULS =2.646, d_G =1.049, Chi-squa e =2404.793, NFI =
0.618.
*
Signi ican a 95 % con idence in e al.
H. Dedunu e al.
Jou nal o Inno a ion & Knowledge 10 (2025) 100693
7
shown in Fig. 3. The p edic i e ele ance o he model is (Q2 =0.694).
When he alue is abo e ze o, he model es ablishes p edic i e
ele ance.
The ou pu o he s uc u al model is used o e alua e he de eloped
hypo heses. As pe he able, he e ec o QE KC [β =-0.049, =2.179,
(-0.093 -0.004), p <0.05] and QE KA [β =0.129, =5.482, (0.084
-0.176), p <0.05] on i m inno a ion is non-linea and s a is ically
signi ican . As a esul , his s udy con i ms H1 and H3. The ela ionship
be ween KC and i m inno a ion is an in e ed U-shape, whe eas ha
be ween KA and i m inno a ion is a U-shape. Acco ding o he
quad a ic unc ion (Fig. 4), i m inno a ion inc eases in line wi h KC up
o a ce ain le el, a e which i begins o decline as KC con inues.
Howe e , as shown in Fig. 5, he le el o inno a ion dec eases when a
i m s a s o apply knowledge; howe e , a e a ce ain poin , KA en-
cou ages he inno a ion p ocess.
The ela ionships KS [β =0.359, =9,704, (0.287, 0.431), p <0.05],
and KP [β =0.098, =2,775, (0.031, 0.166), p <0.05] a e s a is ically
signi ican , and linea wi h i m inno a ion, con o ming o H2 and H4.
The model con ols o he e ec o employee c ea i i y, age, and
gende . The e ec s o c ea i i y [β =0.345, =10.117, (0.277, 0.412), p
<0.05] and age [β =0.085, =3.060, (0.029, 0.139), p <0.05] a e
s a is ically signi ican , whe eas gende [β =0.006, =0.114, (-0.087,
0.105), p >0.05] is insigni ican .
Discussion o he esul
This s udy explo es he na u e o he ela ionship be ween KM and
i m inno a ion h ough ou hypo heses, inding ha KM has bo h a
linea and a non-linea ela ionship wi h i m inno a ion.
The hypo hesis (H1) assumes ha KC has a signi ican linea o non-
linea ela ionship wi h i m inno a ion. Desc ip i e s a is ics p esen ed
in Table 1 indica e ha KC, in he S i Lankan banking sec o , is a an
‘ag eed’ le el (5.015) on a se en-poin scale anging om s ongly
disag ee o s ongly ag ee. Ex ensi e employee encou agemen o ollow
p o essional and academic p og ams (5.199), new esea ch knowledge
(5.034), a ious aining p og ams (5.078), and ne wo king wi h o he
en i ies (4.725) we e he main easons o a highe le el o KC. Thus, KC
c ea ion is an ac i e and ongoing endea ou in he S i Lankan banking
indus y. As pe he s a is ics p esen ed in Table 3, he linea e ec o KC
on i m inno a ion is no s a is ically signi ican (p =0.281). Howe e ,
Fig. 3. Impo ance-pe o mance map.
Fig. 4. KC and inno a ion.
Fig. 5. KA and inno a ion.
H. Dedunu e al.
Jou nal o Inno a ion & Knowledge 10 (2025) 100693
8
