Spangenbe g, H.; The on, C. C.
A icle
P omo ing e hical ollowe beha iou h ough leade ship
o e hics: The de elopmen o he e hical leade ship
in en o y (ELI)
Sou h A ican Jou nal o Business Managemen
P o ided in Coope a ion wi h:
Uni e si y o S ellenbosch Business School (USB), Bell ille, Sou h A ica
Sugges ed Ci a ion: Spangenbe g, H.; The on, C. C. (2005) : P omo ing e hical ollowe beha iou
h ough leade ship o e hics: The de elopmen o he e hical leade ship in en o y (ELI), Sou h
A ican Jou nal o Business Managemen , ISSN 2078-5976, A ican Online Scien i ic In o ma ion
Sys ems (AOSIS), Cape Town, Vol. 36, Iss. 2, pp. 1-18,
h ps://doi.o g/10.4102/sajbm. 36i2.623
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/218324
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/
S.A .J.Bus.Manage.2005,36(2) 1
P omo ing e hical ollowe beha iou h ough leade ship o e hics: The
de elopmen o he e hical leade ship in en o y (ELI)
H. Spangenbe g*
Cen e o Leade ship S udies, G adua e School o Business,
Uni e si y o S ellenbosch, S ellenbosch 7600, Republic o Sou h A ica
[email p o ec ed]
C.C. The on
Depa men o Indus ial Psychology, Uni e si y o S ellenbosch,
S ellenbosch 7600, Republic o Sou h A ica
Recei ed Decembe 2004
The Cen e o Leade ship S udies de eloped a model ha de ines and desc ibes leade ship beha iou equi ed o
c ea ing an e hical and high pe o ming o ganisa ion. Pu pose ully building an e hical o ganisa ional en i onmen is a
p e equisi e o e hical o ganisa ional beha iou . To be an e hical high pe o mance leade , a pe son mus be bo h an
e ec i e leade and a leade o e hics. The pu pose o his s udy is o de elop a 360° ins umen ha can be used o assess
he quali y o e hical leade ship o middle, senio and execu i e manage s in public, p i a e and no - o -p o i
o ganisa ions. The esul s epo ed he e p o ide easonable suppo o he use o he ELI. The possibili y o causal
in luences exis ing amongs speci ic i s -o de leade ship dimensions, howe e , needs o be in es iga ed.
*To whom all co espondence should be add essed.
In oduc ion
The o ganisa ional misbeha iou as epo ed in daily
newspape s and academic jou nals in he USA owa ds he
end o he 1980’s, lead o he obse a ion ha people ha e
los ai h in he alues o hei economic socie y and ha
hey needed a spi i ual ebi h in business leade ship. The
ques ion a ose whe he indus y needed a new eligion, o a
leas a be e one han i has had (Ohmann, 1989). Re lec ing
on his p emise, Kanungo and Medonca (1996) a gue ha a
business co po a ion in i sel is incapable o doing good o
e il in socie y. Such ou comes a e due en i ely o he
une hical beha iou o indi iduals, whe he i is employees
o manage s. Al hough all employees a e esponsible o
achie ing o ganisa ional objec i es in an e hical manne , he
accoun abili y o p o iding p ope di ec ion and high
pe o mance s anda ds lies wi h o ganisa ional leade s. In
he Sou h A ican con ex , whe e epo s on co up ion and
co up ion- ela ed inciden s in public, business, spo s, and
o he o ganisa ions egula ly o m pa o he majo e ening
news bulle ins o SABC, he o icial b oadcas e , he same
easoning applies. The du y and esponsibili y o p o iding
e hical high pe o mance leade ship es s wi h execu i e and
senio manage s o Sou h A ican business, public se ice,
non-go e nmen al, spo s and o he leisu e o ganisa ions.
The King Repo on Co po a e Go e nance in Sou h A ica
(King Commi ee on Co po a e Go e nance, 2002:101)
de ines o ganisa ional e hics as ‘ he p inciples, no ms and
s anda ds ha i p omo es o he guidance and conduc o
i s ac i i ies, in acco dance wi h es ablished alues’. The
King Repo u he con ends ha adhe ence o es ablished
p inciples o e hical conduc p o ides a s ong measu e o
o ganisa ional in eg i y and s a es ha ‘a company’s e hical
p inciples ep esen a majo mo i a o o s akeholde
in ol emen wi h i and, as such, should pe mea e i s
cul u e, mo i a ing i s s a egy, business goals, policies and
ac i i ies’ (p.102).
Wi h ega d o he implemen a ion o co po a e e hics,
And ews (1989:99) desc ibes h ee challenges ha
o ganisa ions ace, namely ‘ he de elopmen o he
execu i e as a mo al pe son; he in luence o he co po a ion
as a mo al en i onmen ; and he ac ions needed o map a
high oad o economic and e hical pe o mance – and o
moun gua d ails o keep co po a e way a e s on ack’.
Bass and S eidlmeie (1999) p opose h ee pilla s o e hical
leade ship ha suppo he impo ance o he e hical
de elopmen o he o ganisa ional leade and u he build
ou And ews’s challenges. The pilla s a e he mo al
cha ac e o he leade and his/he conce ns o sel and
o he s, e hical alues being embedded in he leade ’s
isioning, and he mo ali y o he choices and ac ions o
leade s and hei ollowe s.
Jose and Thibodeaux (1999), quo ing esea ch, desc ibe a
numbe o speci ic oles o he e hical leade :
1. Wi h ega d o isioning, hey sugges ha leade s se
he mo al s anda ds o he o ganisa ion and ocus on
he in eg i y o he common pu pose (Hosme , 1987).
2 S.A .J.Bus.Manage.2005,36(2)
2. Leade s explici ly cla i y he e hical dimensions o
managemen decisions, and o mula e and jus i y
e hical p inciples ha go e n decision-making
(Ende le, 1987).
3. Leade s signal hei commi men o e hical p inciples
h ough hei in luence on co po a e cul u e.
• They do his by de ining o ganisa ional e hics and
alues, and by ac ing as ole models in es ablishing a
posi i e e hical clima e. Employees no mally ecei e
hei e hical guidance om obse ing he beha iou s o
hei supe io s (Jose & Thibodeaux, 1999; O`Boyle &
Dawson, 1992).
• Leade s build cul u e by wha hey moni o and
con ol, how hey eac o c i ical e en s and wha
c i e ia hey use o ec ui ing, selec ing, ewa ding and
dismissing o ganisa ional membe s (Schein, 1992).
• In addi ion o a s ong impac on he e hical beha iou
o employees, e hical leade s acili a e us (B ien,
1998).
4. The CEO aces he challenge o c ea ing a high deg ee
o cong uence be ween he o ganisa ions’s guiding
belie s and employees` e e y-day belie s (Jose &
Thibodeaux, 1999).
In o de o desc ibe leade ship equi ed o c ea ing an
e hical and high pe o ming o ganisa ion and o pu i in
p ope o ganisa ional con ex , a heo e ical model was
de eloped. The model comp ises ou phases. I is p oposed
ha he e hical o ien a ion o leade s (Phase 1); exempli ied
by building an e hical o ganisa ion, i.e. applying
ans o ma ional in luencing p ocesses, building o e hical
s uc u es, c ea ing an e hical cul u e and clima e, and
de eloping p oduc i e s akeholde ela ions (Phase 2);
media ed by ex e nal ac o s (Phase 3); will c ea e an e hical
o ganisa ional en i onmen and posi i e ou comes o he
o ganisa ion, i s employees and o he s akeholde s (Phase
4).
Phase 2 o he model, namely ‘Building an e hical
o ganisa ion’ se ed as heo e ical basis o he de elopmen
o he leade ship o e hics assessmen ins umen . The i s
componen o building an e hical o ganisa ion is applying
ans o ma ional in luencing s a egies. This b oad cons uc
can be equa ed wi h he co e leade ship p ocesses o
analysing he o ganisa ion’s ex e nal and in e nal
en i onmen s, de eloping and communica ing he ision,
p epa ing he people and he o ganisa ion o implemen ing
he ision and, inally, implemen ing he ision and s a egy
(Conge & Kanungo, 1998; House, 1995; Kanungo &
Medonca, 1996; Spangenbe g & The on, 2002). The o he
h ee elemen s o Phase 2, namely c ea ing e hical s uc u es
(B ickley, Smi h & Zimme man, 1994; James, 2000;
Lindsay, Lindsay & I ine, 1996; T e ino & Nelson, 1995);
c ea ing an e hical cul u e and clima e (Cullen, Vic o &
B onson, 1993; Deshpande, 1996; T e ino & Bu e ield,
1998; Upchu ch & Ruhland, 1996; Vaicys, Ba ne &
B own, 1996; Va di, 2001; Vic o & Cullen, 1987; 1990);
and building p oduc i e s akeholde ela ionships Hummels,
1998; Hus ed, 1998; Maignan & Fe ell, 2000; S ong,
Ringe & Taylo , 2001) a e also key esponsibili ies o he
leade o e hics.
Du ing he de elopmen o he model i was ealised ha
e hical leade ship comp ises an impo an bu sepa a e
dimension o leade ship. While he aims o a adi ional
o ganisa ional ision is di ec ed mainly a o e all
o ganisa ional and people pe o mance and e ec i eness, an
e hical ision is aiming a achie ing e hical o ganiza ional
beha iou . An e hical ision may include impe a i es such
as, o example, gi ing people a sense o mo al pu pose,
econciling business e iciency wi h social esponsibili y,
os e ing he uniqueness and in eg i y o he indi idual and
g oup, and ea ing all in e nal and ex e nal s akeholde s
wi h espec , hones y and in eg i y.
In o de o c ea e an e hical high pe o mance o ganisa ion,
i is sugges ed ha he c ea ion and applica ion o an e hical
ision could ei he o m pa o he o e all o ganisa ional
isioning p ocess, o un pa allel bu in conce wi h he
la ge isioning p ocess. E hical leade ship can hus be
ope a ionalised and measu ed in e ms o he same p ocess
model on which he Leade ship Beha iou In en o y (LBI)
is based (Spangenbe g & The on, 2002). In e ms o his
model, leade ship o e hics comp ises he c ea ion and
sha ing o an e hical ision (based on a ca e ul diagnosis o
he ex e nal and in e nal en i onmen s in which all ele an
pa ies pa icipa e); p epa ing he leade , ollowe s, and he
o ganisa ion – pa icula ly i s s uc u es and cul u e - o
implemen ing he ision; and he ac ual implemen a ion
p ocess i sel . This obse able, beha iou ally-based
leade ship p ocess lends i sel eadily o beha iou
assessmen by means o 360° me hodology.
An issue ha needs o be cla i ied is he ques ion whe he
he pe son who does well in a leade ship assessmen
ques ionnai e [ ha no mally would include some e hical
leade ship dimensions ( o example he LBI1)] can be
conside ed a leade o e hics? One could p obably answe
such a ques ion by a guing ha such a pe son could be
conside ed an hones leade wi h in eg i y who c ea es us
in he wo k uni and in him/he sel . Howe e , o be an
e hical high pe o mance leade , a pe son mus be bo h an
e ec i e leade (as pe LBI o MLQ2, o example) and a
leade o e hics. In essence, he LBI p edic s ha an
indi idual will display leade ship e ec i eness (measu ed
by a la ge a ie y o dimensions) and e hical beha iou
(measu ed by a ew impo an e hical dimensions).
Howe e , o enac leade ship o e hics, a pe son mus
pu pose ully and consciously demons a e and cul i a e
e hical beha iou in his/he wo k uni . The leade mus ,
he e o e, pu pose ully acili a e he de elopmen o an
e hical o ganisa ional cul u e by ins illing and cul i a ing
1The Leade ship Beha iou In en o y (LBI), measu es a ou -phase
p ocess o leade ship (Spangenbe g & The on, 2002), including
ans o ma ional leade ship dimensions and a numbe o dimensions
ha ha e a dis inc e hical con en , e.g. enabling he leade ( wo
dimensions) enabling ollowe s, in e pe sonal ela ions, conce n o
o he s, and cha ac e - ela ed dimensions such as hones y and in eg i y,
and decisi eness and ha diness.
2Mul i- ac o Leade ship Ques ionnai e (Bass & A olio, 1997).
S.A .J.Bus.Manage.2005,36(2) 3
e hical belie s and alues (T e ino, Ha man & B own,
2000). To conclude his a gumen , he King Repo
(2002:106) s a es ha ‘ he mo al objec i es o he company,
exempli ied by he boa d collec i ely and he di ec o s
indi idually, needs o be made isible o bo h in e nal and
ex e nal s akeholde s and seen o be in eg al wi h o he
co po a e objec i es’.
Since leade ship o e hics is a c ucial o ce in c ea ing an
e hical o ganisa ion, i was decided o i s de elop an
e hical leade ship in en o y, o be ollowed by he
de elopmen o an e hics audi o o ganisa ions a a la e
s age. By assessing he quali y o he leade ship o e hics o
membe s o a managemen eam on a 360° basis, insigh will
be inc eased conside ably in o hei own e hical leade ship
beha iou as well as unde s anding o he en i e domain o
leade ship o e hics and he impac he eo on hei uni s.
G ea e sel -awa eness and inc eased unde s anding o
e hics in o ganisa ional dynamics will p o ide a posi i e
clima e o majo indi idual and eam lea ning ini ia i es.
The pu pose o his s udy is, he e o e, o de elop a 360°
ins umen ha can be used o assess he quali y o
leade ship o e hics o middle, senio and execu i e
manage s in public, p i a e and no - o -p o i o ganisa ions.
In e ms o he o egoing a gumen i p obably would ha e
been app op ia e o e e o such an ins umen as he
Leade ship o E hics In en o y. In o mal consul a ion wi h
leade s pa icipa ing in he s udy, howe e , indica ed ha
ma ke accep ance o he ins umen would mos likely be
be e i he mo e con en ional i le E hical Leade ship
In en o y (ELI) would be adop ed.
De elopmen o he ques ionnai e
S a egy o de eloping he ques ionnai e
1. Gene a ion o i ems. The main sou ce o i ems was he
li e a u e on he de elopmen o e hics in
o ganisa ions. A second sou ce o i ems was he
Leade ship Beha iou In en o y (LBI). A signi ican
numbe o i ems om he LBI esea ch copy,
consis ing o 237 i ems, had e hical con en and wi h
mino adap a ions could be included in he ELI
esea ch copy. Resea ch indica es ha LBI i ems
measu e obse able leade ship ac ions and beha iou
wi h a high deg ee o eliabili y and alidi y
(Spangenbe g & The on, 2002).
2. Using he s uc u e o he Leade ship Beha iou
In en o y (LBI) as a basis o classi ying newly
de eloped i ems. The a ionale was ha he LBI
comp ises ou ac ionable phases ha a e equi ed o
implemen ing a ision, i espec i e o whe he i is a
pe o mance o e hics- ela ed ision.
3. Using a Delphi echnique in ield esea ch. This means
ha a c i ical s ages in he de elopmen p ocess, a
panel o expe s e i y he p ocess and p opose
imp o emen s as equi ed. Posi i e expe ience on h ee
occasions in he pas o using he Delphi echnique as
an aid in ques ionnai e de elopmen has sugges ed ha
his i e a i e, consensus seeking echnique should be
employed again (Spangenbe g & The on, 1997; 2002;
The on & Spangenbe g, 2004).
Field esea ch o de eloping he ques ionnai e
Two ounds o ield esea ch by means o he Delphi
echnique we e conduc ed. Du ing he i s ound he d a
ques ionnai e was sen o a panel o 13 indus ial
psychologis s, known as expe s in he ield o leade ship
assessmen and/o de elopmen . They we e asked o wo k
h ough he ques ionnai e as ollows: e i y ele ance o
dimensions and p opose new dimensions i equi ed; a e he
app op ia eness o each i em on a h ee-poin scale, and
p opose new i ems as equi ed.
The i s d a ques ionnai e consis ed o 24 dimensions,
measu ed by means o 276 i ems. In o de o p o ide a
be e unde s anding o a e s, dimension de ini ions we e
included and i ems we e ca ego ised acco ding o
dimensions. On e u n o ques ionnai es a wo-s eam
app oach was ollowed o imp o e o cull i ems. Fi s ly, in
line wi h he Delphi echnique, commen s on speci ic
dimensions and i ems we e s udied. Commen s om a e s
we e indeed ex emely use ul in imp o ing i ems. Secondly,
desc ip i e s a is ics we e conduc ed on i em a ings and he
culling o i ems was based mainly on a i hme ical decision
ules. The imp o ed and sho ened d a ques ionnai e s ill
consis ed o 24 dimensions, bu he numbe o i ems was
educed o 189.
I was decided o conduc a second ound o ield esea ch
by means o he Delphi echnique in which o ganisa ional
leade s would ac as a e s. I was ealised ha in o de o
acili a e sound co po a e go e nance ha simul aneously
inc eases e hical, high o ganisa ional pe o mance and
dec eases o ganisa ional malp ac ices, he suppo and
con inued in ol emen o a signi ican body o high-le el
o ganisa ional leade s would be equi ed. F anklin Sonn,
o me Sou h A ican ambassado o he USA and
p ominen businessman, in a pe sonal discussion (May
2003), pu i blun ly ha an e hical leade ship ins umen
de eloped by an ins i u ion o lea ning, wi hou consen and
suppo o he leade ship communi y, will no ha e he
desi ed e ec .
The second ound Delphi was conduc ed along he same
lines as he i s one, wi h 16 execu i e manage s ac ing as
a e s. The g oup included CEO’s om some o Sou h
A ica’s la ges and mos espec ed companies. Ra ings and
a e commen s we e e y use ul in educing dimensions by
collapsing hem in o o he dimensions o by culling hem.
Ra ings and a e commen s we e again use ul in culling
i ems ha ecei ed unsa is ac o y a ings, and o u he
imp o e i ems ha we e e ained. The d a ques ionnai e
was educed o 19 dimensions, measu ed by 103 i ems. The
dimensions measu ed by he ELI, wi h hei de ini ions, a e
p esen ed in Table 1.
4 S.A .J.Bus.Manage.2005,36(2)
Table 1: Fi s -o de la en leade ship dimensions measu ed by he ELI
1. C ea ing and sha ing an e hical ision
Unde s anding he e hical dynamics in he ex e nal and in e nal en i onmen s (En i on)
Diagnoses e hical dynamics in he ex e nal and in e nal en i onmen s in o de o de elop an e hical ision.
De eloping a challenging ision (E ision)
De elops a collec i e e hical ision ha inspi es people and gi es hem a sense o pu pose, is cus ome - ocused and ad ances
di e si y o people.
Building us in he leade and he uni (Bu us )
The leade c ea es us in him/he sel and builds con idence in he uni .
A icula ing an e hical ision and enlis ing ollowe s (A i is)
A icula es an e hical ision o he u u e ha p o ides di ec ion. Inspi es con idence in he ision and ob ains ollowe
commi men o he ision.
Concep ualising e hical s a egy (CoS a )
De ines s a egic e hical issues clea ly. Builds s a egies and plans based on ho ough p oblem analysis and b oad-based ac -
inding. Conside s consequences o decisions.
2. Enabling he leade and he uni o implemen he e hical ision
Enabling he leade (Enablel)
Iden i ies challenging oppo uni ies o sel -de elopmen and is commi ed o con inuous lea ning. App ecia es eedback and
has good insigh in o own his/he own e hical iden i y, capabili ies and beha iou . Is commi ed o con inuous lea ning.
Empowe ing ollowe s (Enable )
Encou ages ollowe s o accep esponsibili y o hei own e hical lea ning and g ow h. C ea es condi ions which allow hem
he oppo uni y o ake meaning ul decisions.
Fo mula ing and implemen ing e hical s uc u es and sys ems (S usys)
Adap s s uc u es, p ocesses and p ocedu es o suppo implemen a ion o e hical s a egy in a changing en i onmen .
Implemen s e hical s uc u es and sys ems, o example a code o e hics, an ombudsman, e hics commi ee, and e hics aining
p og amme.
Building an e hical cul u e and clima e (E hcul )
Builds a cul u e ha e lec s sha ed belie s, alues and no ms; sha ed pe cep ions o e hically co ec beha iou ;and guidance
o handling di icul e hical issues.
3. Implemen ing he e hical ision
3.1. Leading wi h cou age, in eg i y and sensi i i y
Ac ing hones ly and wi h in eg i y
Hones ly manages he o ganisa ional uni and consis en ly li es ou he alues embedded in he ision (Hoin eg)
Conside s e hical implica ions o decisions, assu es ag eed upon alues a e adhe ed o and deals hones ly wi h all s akeholde s.
Decisi eness and ha diness (Ac deci)
Ac s decisi ely and makes ough e hical decisions. Pe o ms e ec i ely unde s ess and eac s posi i ely o change and
unce ain y.
Demons a ing in e pe sonal sensi i i y (Demsens)
Conside s he needs, eelings and digni y o o he s. Wo ks owa ds p oduc i e in e pe sonal ela ions..
3.2. Encou aging e hical beha iou
Challenging cu en eali y and s imula ing lea ning (Challen)
Challenges cu en hinking abou e hics, econside s and imp o es cu en p ac ices on an e hical basis. P omo es con inuous
e hical lea ning.
Inspi ing people owa ds e hical beha iou (Inspi e)
Raises he aspi a ions o ollowe s and builds con idence in hem o pe o m e ec i ely and e hically. A icula es e hical issues
clea ly.
3.3. S imula ing e hics ac oss bounda ies
Facili a ing in e depa men al co-o dina ion (In coo )
Facili a es c oss- unc ional collabo a ion and eamwo k. Helps people o see he e hical big pic u e.
In luencing ex e nal s akeholde s (In s ak)
Main ains p oduc i e ela ionships wi h ex e nal s akeholde s and builds he e hical image o he o ganisa ion.
3.4. Leading e hical ini ia i es and ewa ding e hical con ibu ions
Planning and implemen ing e hical ini ia i es (Planimp)
Ensu es ha e hical expec a ions o he uni and i s membe s a e cla i ied, and ha e hical ini ia i es a e designed and aligned
wi h e hical and business s a egies.
Re iewing e hical ini ia i es and beha iou (Re iewe)
Re iews he ou comes o uni , eam and employee e hical ini ia i es. P o ides speci ic eedback o ollowe s in o de o help
hem assess hei own con ibu ion o hese ini ia i es.
Rewa ding e hical con ibu ions and beha iou (Rewa de)
Gi es ecogni ion o accomplishing e hical ini ia i es as well as o exempla y wo k- ela ed a i udes and beha iou ;
celeb a es e hical success.
S.A .J.Bus.Manage.2005,36(2) 5
Field es ing o he ques ionnai e
In o de o ensu e ha he sample e lec s di e si y o
cul u e and o ganisa ional ype, a wide ange o
o ganisa ions we e app oached o pa icipa ion in
ieldwo k. O ganisa ional co-o dina o s acili a ed he
nomina ion o candida es (pa icipan s) and esponden s.
Fi s ly, hey iden i ied colleagues who managed wo k uni s,
which we e de ined as consis ing o h ee o mo e membe s,
wi h no uppe limi . Co-o dina o s we e speci ically asked o
include uni manage s om all cul u es, male and emale, in
o de o ob ain a ep esen a i e sample. Candida es
comple ed a Candida e Ques ionnai e and we e p omised o
ecei e pe sonal eedback on eques .
Secondly, he candida e, in conjunc ion wi h he
o ganisa ional co-o dina o , ideally iden i ied wo pee s and
wo subo dina es o ollowe s as esponden s. The supe io
o he candida e se ed as i h esponden . Responden s
comple ed a Responden Ques ionnai e o he candida e.
Again, co-o dina o s we e speci ically asked o include a
ep esen a i e sample o esponden s. In o de o conduc
es bias s udies on he ques ionnai e, bo h candida es and
esponden s we e asked o p o ide essen ial biog aphical
in o ma ion.
Sample
The da a used o he pu pose o his s udy was ob ained by
means o a se ies o non-p obabili y samples o uni leade s
selec ed om a non-p obabili y sample o o ganiza ions. A
o al o six y uni leade s we e selec ed and a ed 360° by
hei supe io s, pee s and subo dina es. The samples o
esponden s ha a ed each selec ed uni leade again we e
selec ed h ough non-p obabili y sampling p ocedu es.
Selec ed uni leade s also a ed hemsel es, which esul ed
in a o al o 315 comple ed ELI ques ionnai es a he ime o
he analysis. Al hough he objec i e ini ially was o ob ain
a ings om wo subo dina es, wo pee s and a single
supe io , he need o a la ge as possible sample size, in
conjunc ion wi h he di icul ies encoun e ed when ying o
apply a ques ionnai e o his leng h and na u e o
esponden s on his high a job le el, necessi a ed a de ia ion
om his ideal in a numbe o cases.
The sample o o ganiza ions and he numbe o uni leade s
sampled om each o ganiza ion is depic ed in Table 2.
Table 2: dis ibu ion o uni leade s ac oss sampled
o ganisa ions
Company F equency Pe cen Valid
Pe cen Cumula i e
Pe cen
Anglo Gold 3 5,0 5,0 5,0
Dis ell 4 6,7 6,7 11,7
Medi- Clinic 4 6,7 6,7 18,3
Peninsula
Technikon
7 11,7 11,7 30,0
Powe G oup 7 11,7 11,7 41,7
Rand Wa e 10 16,7 16,7 58,3
SA Rese e Bank 10 16,7 16,7 75,0
SAPS 14 23,3 23,3 98,3
Telkom S A 1 1,7 1,7 100,0
To al 60 100,0 100,0
The dis ibu ion o uni leade s in e ms o posi ion le el is
po ayed in Table 3.
The b eakdown o he sample o uni leade s in e ms o ace
is gi en in Table 4. A pleasing ea u e o he ace
composi ion o he sample is ha almos 50% o he uni
leade s belong o p e iously disad an aged g oups. Uni
leade s om p e iously disad an aged g oups do, howe e ,
end o occupy he ela i ely lowe posi ion le els (c oss
abula ion no shown).
A s onge ep esen a ion o uni leade s in he op
managemen ca ego y ( ha is uni leade s esponsible o
la ge , composi e uni s) would ha e been p e e able. In he
inal analysis i is hese highe echelon leade s ha should
be held accoun able o he e hical beha iou o all hei
ollowe s. I is hei e hical ision ha should be cascading
down h ough he o ganiza ional hie a chy, ca ied by he
walk and alk o lowe le el uni leade s.
The desc ip i e s a is ics cha ac e izing he age dis ibu ion
o he sample o uni manage s a e shown in Table 5.
Resul s indica e an app oxima ely symme ic, mesoku ic
age dis ibu ion.
Due o he sampling me hodology and he sample size, he
sample can clea ly no claim o cons i u e a ep esen a i e
sec ion o he popula ion o Sou h A ican uni leade s who
could be held accoun able o he e hical beha iou o he
membe s o hei uni s. This p ecludes he possibili y o
eaching any de ini i e conclusion on he compe ence o
Sou h A ican uni leade s on he leade ship compe encies
se ing he o mula ion and ealiza ion o an e hical uni
ision. This also ules ou he possibili y o eaching any
de ini i e conclusion on he abili y o he ELI o success ully
measu e he leade ship o e hics cons uc in acco dance
wi h he ins umen a chi ec u e in he a ge popula ion.
None heless, i he measu emen model implied by he
ins umen design would i he sample da a well, i would
cons i u e ele an , albei limi ed e idence ha he ELI can
be used o assess e hical leade ship compe encies o
de elopmen in he a ge popula ion.
Missing alues
Missing alues p esen ed a p oblem ha had o be add essed
be o e analysis could p oceed. The same a gumen s as
hose p esen ed in The on and Spangenbe g (2004) we e
conside ed wi h ega ds o he missing alues p oblem.
Pai -wise dele ing o cases p esen ed i sel as a possible
solu ion in his case since i would no esul in a co ela ion
ma ix wi h ex eme a ia ion in N- alues (a maximum o
307 and a minimum o 221 in his pa icula case). The
PRELIS p og am (Jö eskog & Sö bom, 1996a) was used o
impu e missing alues. The PRELIS un p o ed o be
easonably e ec i e in coun e ing he missing alue
p oblem in compa ison o he esul s ha would ha e been
achie ed unde lis wise dele ion, bu less use ul in
compa ison o he esul s ha would ha e been ob ained
unde pai -wise dele ion. By de aul , cases wi h missing
alues a e impu a ion we e elimina ed. A e impu a ion,
264 cases wi h obse a ions on all 103 i ems emained in he
alida ion sample.
6 S.A .J.Bus.Manage.2005,36(2)
Table 3: Dis ibu ion o uni leade posi ion le el
Table 4: Dis ibu ion o uni leade ace
Race F equency Pe cen Valid Pe cen Cumula i e Pe cen
Valid Asian 3 5,0 5,9 5,9
Black 13 21,7 25,5 31,4
Colou ed 8 13,3 15,7 47,1
Whi e 27 45,0 52,9 100,0
To al 51 85,0 100,0
Missing Sys em 9 15,0
To al 60 100,0
Table 5: Desc ip i e s a is ics cha ac e izing he uni leade
age dis ibu ion
Desc ip i e s a is ics Value
N
Vali
d
52
N
Missing 8
Mean 43,25
Median 43,00
Modea 41
S d, De ia ion 7,987
Va iance 63,799
Skewness ,061
S d, E o o Skewness ,330
Ku osis -,138
S d, E o o Ku osis ,650
Minimum 25
Maximum 62
a Mul iple modes exis . The smalles alue is shown
The ela i e na ow band in which he e ec i e sample sizes
a y unde pai -wise dele ion, combined wi h he ac ha
he median e ec i e sample size is 275 and he modal
e ec i e sample size 278, swayed he decision owa ds he
use o pai -wise dele ion o missing alues. The e ec i e
sample size ha has been ob ained h ough impu a ion by
ma ching alls below he wen y- i h pe cen ile (P25 = 267).
I em analysis
The a chi ec u e o he ELI e lec s he in en ion o cons uc
essen ially one-dimensional se s o i ems o e lec a iance
in each o he nine een la en a iables collec i ely
comp ising he e hical leade ship domain. The i ems a e
mean o unc ion as homogenous s imulus se s o which
a e s espond wi h beha iou ha is p ima ily a ela i ely
uncon amina ed exp ession o a speci ic unde lying e hical
leade ship la en a iable. Each o he nine een ELI sub-
scales we e consequen ly i em analysed h ough he SPSS
Reliabili y P ocedu e (SPSS 11.0 o Windows, 2004) o
iden i y and elimina e i ems no con ibu ing o an in e nally
consis en desc ip ion o he uni pe o mance ace in
ques ion. The summa y esul s o he i em analyses a e
shown in Table 6.
Two i ems we e lagged as p oblema ic. I em 3 o he
Building us in he leade and he uni subscale and i em
48 o he Ac ing hones ly and wi h in eg i y subscale we e
iden i ied as i ems ha lowe he homogenei y o he scales.
The magni ude o he co ec ed i em- o al co ela ions
(0,2091 and 0,3598) and he inc ease in alpha a ec ed by
he emo al o he i ems (0,1391 and 0,0309) jus i ied he
dele ion o hese i ems. Fou u he i ems (i ems 1, 4, 13
and 54) we e lagged as suspec . The magni ude o he
co ec ed i em- o al co ela ion (0, 4639; 0,4628; 0,4990;
and 0,4510``) as well as he a he modes inc ease in alpha
a ec ed by he emo al o he i ems (0,0056; 0,0055; 0,0001
and 0,006), howe e , a gued agains dele ing hese i ems.
Eigh een o he subscales e u ned C onbach alpha alues
g ea e han 0,80, wi h only he Inspi ing ollowe s owa ds
e hical beha iou subscale alling ma ginally below his cu -
o alue (0,79). The ela i ely high i em homogenei y
ound o each sub-scale, as indica ed by he C onbach alpha
alues in Table 6, is ega ded as sa is ac o y.
Dimensionali y analysis
Un es ic ed p incipal componen analyses wi h Va imax
o a ion we e subsequen ly pe o med on each o he
nine een ELI subscales, each ep esen ing a ace o he
mul i-dimensional e hical leade ship cons uc , o u he
e alua e he success wi h which each i em accomplishes i s
in ended unc ion. The objec i e o hese analyses
u he mo e was o con i m he uni-dimensionali y o each
sub-scale and o emo e i ems wi h inadequa e ac o
loadings (and e un he i em analysis) and/o spli
he e ogeneous sub-scales in o wo o mo e homogenous
subse s o i ems i necessa y. In he la e case concomi an
adjus men s o he unde lying ELI measu emen model
would ha e o be made and he i em analysis epea ed on he
newly c ea ed subscales. Hulin, D asg ow and Pa sons
Posi ion le el F equency Pe cen Valid Pe cen Cumula i e Pe cen
Valid Top managemen 9 15,0 18,8 18,8
Senio managemen /
senio p o essional
23 38,3 47,9 66,7
Middle managemen /
p o essional
16 26,7 33,3 100,0
To al 48 80,0 100,0
Missing Sys em 12 20,0
To al 60 100,0
S.A .J.Bus.Manage.2005,36(2) 7
(1983), howe e , cau ion ha ac o analysis as pe o med
he e on a ma ix o p oduc momen co ela ions migh no
be he mos app op ia e p ocedu e o es ablishing he uni-
dimensionali y o a scale due o he dange o ex ac ing
a e ac ac o s e lec ing di e ences in i em di icul y alue
o a iance only. Desc ip i e s a is ics we e he e o e also
calcula ed o he i ems o each subscale o examine he
possibili y o mul iple ac o s appea ing as an a e ac o
di e en ial i em cha ac e is ics like skewness (Schepe s,
1992). The analyses we e pe o med u ilising SPSS 11.0 o
Windows (2004). The eigen alue-g ea e - han-uni y ule o
humb was used o de e mine he numbe o ac o s o
ex ac . All nine een subscales passed he uni-
dimensionali y es . Fac o loadings gene ally we e
sa is ac o y, a ying be ween 0,565 and 0,870 wi h a median
loading o 0,779. Se en y i e pe cen o he loadings
exceeded 0,747. All ou i ems lagged as suspec du ing
he i em analyses e u ned loading in excess o 0,60, which
on he one hand con i ms hei bo de -line s a us bu
simul aneously seems o jus i y he decision o e ain hese
i ems. I should, none heless, be no ed ha o all subscales
he single p inciple componen solu ion ailed o
sa is ac o ily ep oduce he obse ed co ela ion ma ices.
The esul s o he p inciple componen analyses a e
summa ized in Table 7.
The majo i y o i ems ollowed a signi ican ly (p<0.05)
nega i ely skewed and lep oku ic dis ibu ion. Due o he
absence o nega i e i ems and he consis ency in
dis ibu ional o m ac oss i ems his did, howe e , no esul
in he eme gence o a e ac ac o s.
The esul s o he o egoing analyses seem o sugges ha
he ELI i ems gene ally do sys ema ically e lec hei
designa ed la en e hical leade ship dimension wi h
easonable success. Al hough no conclusi e e idence in his
ega d was de i ed om he cu en da a se , i is
none heless assumed ha he scales do e lec he in ended
la en a iables. Resul s on he i o he i s -o de
measu emen model epo ed below end inc ease he
con idence in his posi ion. Expanding he measu emen
model in o a ully ledged heo y d i en s uc u al model
and con on ing hese wi h he cu en da a se ia a se ies o
con i ma o y model i ing analyses u ilizing Lis el would,
howe e , be needed o gi e signi ican c edibili y o his
claim. The mode a ely high componen loadings, he
mode a e pe cen age o he a iance in he da a explained by
he single componen s, and he gene al inabili y o he single
componen solu ions o ep oduce he obse ed co ela ion
ma ices do, howe e , sugges ha he ELI i ems gene ally
also e lec a ai amoun o non- ele an in o ma ion.
Va iable ype
S uc u al equa ion modeling on he ELI in which each
indi idual i em se es as a mani es o indica o a iable o
he a ious la en leade ship ace s would ha e esul ed in an
ex emely cumbe some and ex ensi e exe cise simply due o
he numbe o i ems in ol ed. The o dinal na u e o he
da a would ha e equi ed he calcula ion o he asymp o ic
co a iance o asymp o ic a iance ma ices which end o
demand la ge amoun s o memo y and p ocessing ime
when he numbe o a iables a e la ge (Jö eskog &
Sö bom, 1996a; 1996b). Consequen ly, wo mani es
a iables we e c ea ed om each sub-scale by calcula ing
he unweigh ed a e age o he odd numbe ed i ems and he
e en numbe ed i ems o each scale. Apa om simpli ying
he logis ics o i ing he model, he c ea ion o wo linea
composi e indica o a iables o each la en a iable has he
added ad an age o c ea ing mo e eliable indica o
a iables (Nunnally, 1978). Ma sh, Hau, Balla and G ayson
(1998), on he o he hand, epo ha solu ions in
con i ma o y ac o analysis end o imp o e wi h inc easing
numbe o indica o s pe ac o . The complexi y o he
comp ehensi e Lis el model ha would ha e esul ed om
ea ing he indi idual i ems as indica o a iables swayed
he decision owa ds pa celling. The composi e indica o
a iables we e ea ed as con inuous a iables. The analysis
o he co a iance ma ix ins ead o he polycho ic
co ela ion ma ix was he e o e ega ded as pe missible
(Jö eskog & Sö bom, 1996a; 1996b; Mels, 2003).
Uni a ia e and mul i a ia e no mali y
The de aul me hod o es ima ion when i ing measu emen
models o con inuous da a (maximum likelihood), assumes
mul i a ia e no mali y. This is also ue o gene alized leas
squa es (GLS) and ull in o ma ion maximum likelihood
(FIML) as possible al e na i e es ima ion me hods o
s uc u al equa ion modeling wi h con inuous da a (Mels,
2003). The inapp op ia e analysis o con inuous non-no mal
a iables in s uc u al equa ion models can esul in
inco ec s anda d e o s and chi-squa e es ima es (Du Toi
& Du Toi , 2001; Mels, 2003). The uni a ia e and
mul i a ia e no mali y o he composi e indica o a iables
we e consequen ly e alua ed ia PRELIS (Jö eskog &
Sö bom, 1996a). The null hypo hesis o uni a ia e
no mali y had o be ejec ed in he case o all hi y-eigh
composi e indica o a iables, co obo a ing he skewness
indings epo ed ea lie . The esul s o he es o
mul i a ia e no mali y a e gi en in Table 8. The assump ion
o mul i a ia e no mali y clea ly is no enable.
Two possible solu ions o he lack o no mali y in he da a
we e in es iga ed. The i s was o no malize he composi e
indica o a iables. Al hough he PRELIS no maliza ion had
he e ec o imp o ing he symme y and ku osis o he
indica o a iable dis ibu ions, he null hypo hesis o
mul i a ia e no mali y s ill had o be ejec ed. The dilemma
is ha no malizing occu s o each a iable sepa a ely
(Mels, 2003).
Since he no maliza ion op ion ailed o achie e mul i a ia e
no mali y, he use o an al e na i e me hod o es ima ion
mo e sui ed o da a no ollowing a mul i a ia e no mal
dis ibu ion was conside ed ins ead. This app oach has he
ad an age o e he use o no mal sco es ha he solu ion
need no be in e p e ed in e ms o ans o med alues.
Weigh ed leas squa es (WLS), diagonally weigh ed leas
squa es (DWLS) and obus maximum likelihood (RML) a e
sugges ed o i s uc u al equa ion models o non-no mal
da a (Du Toi & Du Toi , 2001; Jö eskog, Sö bom, Du Toi ,
& Du Toi , 2000; Mels, 2003). Mels (2003) ecommends
he use o obus maximum likelihood es ima ion i he
assump ion o a mul i a ia e no mal dis ibu ion does no
hold.
8 S.A .J.Bus.Manage.2005,36(2)
Table 6: Reliabili y o ELI sub-scale measu es
Subscale Sample size (n) Alpha Mean Va iance Numbe o
i ems in inal
subscale
Numbe o
i ems dele ed
En i on 249 0,82 20,17 12,63 5 0
E ision 231 0,89 28,08 30,77 7 0
Bu us 269 0,85 21,15 13,57 5 1
A i is 253 0,83 20,43 12,98 5 0
Cos a 259 0,81 20,91 12,58 5 0
Enablel 262 0,82 21,17 12,16 5 0
Enable 207 0,82 23,44 20,55 6 0
S usys 229 0,89 28,73 25,43 7 0
E hcul 266 0,81 17,07 8,99 4 0
Hoin eg 262 0,84 21,12 12,32 5 1
Ac deci 271 0,83 21,26 11,84 5 0
Demsens 247 0,85 24,27 18,31 6 0
Challen 248 0,85 20,17 13,82 5 0
Inspi e 261 0,79 20,66 12,38 5 0
In coo 252 0,85 20,22 14,15 5 0
In s ak 262 0,81 20,97 11,92 5 0
Planimp 241 0,85 20,12 14,76 5 0
Re iewe 244 0,83 19,75 15,23 5 0
Rewa de 223 0,82 20,13 13,25 5 0
Table 7: P inciple componen analyses o ELI sub-scale measu es
Subscale De e minan KMO % Va iance
explained Max λ Min λ % Residual >
0,05
En i on 0,18 0,81 58,44 0,829 0,636 80
E ision 0,02 0,89 61,87 0,844 0,708 71
Bu us 0,13 0,83 62,46 0,847 0,740 70
A i is 0,11 0,81 61,79 0,870 0.629 70
Cos a 0,18 0,81 58,76 0,816 0,735 80
Enablel 0,19 0,84 58,89 0,832 0,710 90
Enable 0,08 0,88 60,29 0,826 0,730 80
S usys 0,05 0,90 64,67 0,842 0,747 66
E hcul 0,03 0,89 60,48 0,822 0,744 61
Hoin eg 0,25 0,78 64,49 0,847 0,749 100
Ac deci 0,17 0,85 60,08 0,827 0,712 100
Demsens 0,18 0,83 59,71 0,805 0,755 100
Challen 0,09 0,86 58,30 0,805 0,646 73
Inspi e 0,23 0,83 56,13 0,831 0,693 90
In coo 0,15 0,86 61,90 0,837 0,748 100
In s ak 0,16 0,78 58,62 0,827 0,649 80
Planimp 0,10 0,78 63,52 0,852 0,747 70
Re iewe 0,18 0,82 59,01 0,810 0,725 80
Rewa de 0,19 0,83 57,95 0,821 0,618 90
Table 8: Tes o mul i a ia e no mali y o con inuous a iables
Skewness Ku osis Skewness and Ku osis
Value Z-Sco e P-Value Value Z-Sco e P-Value Chi-Squa e P-Value
57, 624 77,496 0,000 2014,618 23,885 0,000 6576,108 0,000
E alua ion o he p ima y measu emen model
The ELI was de eloped o measu e he mul i ace ed
leade ship o e hics cons uc o which a speci ic
cons i u i e meaning has been a ached as shown in Table 1.
Ope a ional deno a ions we e explici ly and in en ionally
p oduced o e lec speci ic ace s o his cons uc . Speci ic
ELI i ems we e w i en o unc ion as homogenous s imulus
se s o which a e s would espond wi h beha iou which
would be ela i ely uncon amina ed beha iou al exp essions
o speci ic la en leade ship dimensions. The ques ion ha
needs o be answe ed is o wha ex en his p emedi a ed
ope a ional design did succeed in p o iding a
comp ehensi e and uncon amina ed empi ical g asp on he
cons uc as de ined. Gi en he in en ion o he ELI o
measu e speci ic, p e iously de ined leade ship
compe encies by means o a speci ic ope a ional
a chi ec u e, a hypo hesis es ing, es ic ed, con i ma o y
app oach was ollowed. In e ms o his app oach, speci ic
s uc u al assump ions a e made wi h ega ds o he numbe
o la en a iables unde lying he ELI, he ela ions among
he la en a iables, and he speci ic pa e n o loadings o
indica o a iables on hese la en a iables (Fe ando &
Lo enzo-Se a, 2000; Jö eskog & Sö bom, 1993). To he
ex en o which a measu emen model e lec ing hese
S.A .J.Bus.Manage.2005,36(2) 15
Powe assessmen
The s a is ical powe associa ed wi h he es s o exac and
close i was es ima ed (Diaman opoulos & Siguaw, 2000).
To de e mine he powe o hese wo hypo hesis es s, a
alue o RMSEA = 0,05 was assumed unde he exac i
al e na i e hypo hesis, while a alue o RMSEA = 0,08 was
assumed unde he close i al e na i e hypo hesis.
Powe ables compiled by MacCallum, B owne and
Sugawa a (1996) only make p o ision o ≤100 and
N≤500. A SPSS ansla ion o he SAS syn ax4 p o ided by
MacCallum e al. (1996) was consequen ly used o de i e
powe es ima es o he es s o exac and close i , gi en he
e ec sizes assumed abo e, a signi icance le el (α) o 0,05
and a sample size o 315. The deg ees o eedom (ν) in he
model is (½[(p+q][p+q+1]- )=741 - 247=494. Powe alues
o 1,00 we e ob ained o bo h he es s o exac and close
i . Rejec ing he null hypo hesis o exac i unde he ue
condi ion o close i hus is a ce ain y. Resul s indica e
ha e en i he ac ual model i would be as close as
RMSEA=0,03, he p obabili y o ejec ing he null
hypo hesis o exac i would s ill be as high as 0,989068.
The p obabili y o ejec ing he null hypo hesis o close i i
he ue model i was medioc e is also uni y. E en i he
ac ual model i would be a close as RMSEA=0,06, he
p obabili y o ejec ing he null hypo hesis o close i would
s ill be as high as 0,938511. These powe es ima es, aken
in conjunc ion wi h he decisions o ejec he null
hypo heses o exac and close i , sugges ha he conclusion
o poo model i could be con en ious in ha he es s o
exac and close i we e highly sensi i e o e en i ial
misspeci ica ions in he model.
Summa y and manage ial implica ions
E hical o ganiza ional uni beha iou is desi ed i s and
o emos because o ganiza ional beha iou con o ming o
e hical bes p ac ice is in and by i sel highly alued.
Mo eo e , o ganiza ional beha iou con o ming o e hical
bes p ac ice is highly alued, in pa , because i is
associa ed wi h ou comes o consequences bene icial o and
alued by he indi idual leade , s akeholde s, and he
physical, social and economic en i onmen s in which he
o ganiza ional uni ope a es. Consequen ly, such
o ganiza ional uni beha iou can pay subs an ial di idends
bo h in e ms o p e en ing nega i e sys emic esponses and
p omo ing posi i e sys emic eac ions. Seen om bo h
hese pe spec i es, a need he e o e exis s o inc ease he
e hical beha iou s anda d demons a ed by he membe s o
an o ganiza ional uni .5
4The help o Klay Ma ens o SPSS Sou h A ica is g a e ully
acknowledged in inding an i e a i e p ocedu e o de e mine he
c i ical chi-squa e ha would cu o an a ea alpha in he uppe ail o
he non-cen al chi-squa e dis ibu ion wi h d deg ees o eedom and
non-cen ali y pa ame e ncp.
5This aises he impo an ques ion whe he e hical uni beha iou
should be e lec ed in he cons i u i e in e p e a ion o uni
pe o mance, and consequen ly whe he ope a ional measu es o uni
pe o mance like he Pe o mance Index (Spangenbe g & The on,
2004) should accommoda e i as a sepa a e dimension o uni
pe o mance?
The e hical beha iou s anda d demons a ed by he
membe s o any gi en o ganiza ional uni is, howe e , no a
chance e en bu a he is sys ema ically de e mined by an
in ica e nomological ne wo k o la en a iables. The
abili y o a ionally and in en ionally imp o e he e hical
beha iou o he membe s o an o ganiza ional uni
consequen ly depends on he ex en o which he iden i y o
he la en a iables comp ising his nomological ne wo k a e
known as well as he manne in which hey combine o
shape he e hical beha iou o uni membe s. Leade ship
cha ac e is ics and beha iou mos p obably play a i al ole
in such an explana o y model. I does no seem
un easonable o hold uni leade s esponsible o he e hical
quali y o he beha iou o he membe s o he uni . The
e ec o leade ship on he e hical beha iou s anda d
demons a ed by he membe s o an o ganiza ional uni ,
howe e , mos p obably would no be di ec , bu a he
would be media ed h ough a laye o leade ship ou come
la en a iables. E hical clima e (Engelb ech e al., 2004)
and e hical cul u e a e la en a iables ha qui e possibly
could media e he e ec o leade ship on e hical membe
beha iou . Speci ic leade ship beha iou s (leade ship
compe encies) a e equi ed o build an app op ia e uni
e hical clima e and e hical cul u e and ul ima ely o achie e
e hical uni beha iou .
Would a uni leade , ac ing in acco dance wi h e ec i e
leade ship p esc ibed by he LBI, while simul aneously
complying wi h sound e hical s anda ds, bu wi hou
explici ly p omo ing an e hical ision, con ibu e owa ds
building e hical uni clima e and cul u e? P obably yes,
because o he desi e o ollowe s o emula e hei
( ans o ma ional) leade s (Bass, 1985; Bass & A olio,
1993). This s udy would, howe e , wan o a gue ha he
building o e hical uni clima e and e hical uni cul u e
equi es mo e han leade s who beha e e hically. To
pu pose ully build app op ia e uni e hical clima e and
e hical cul u e and ul ima ely o achie e e hical uni
beha iou addi ionally equi es leade s o e hics. To
achie e e ec i e leade ship o e hics, uni leade s ha e o
egis e and be conce ned abou signi ican disc epancies
be ween he e hical beha iou cu en ly being displayed in
he uni and an e hical beha iou al no m e lec ing
p e ailing e hical s anda ds. E ec i e uni leade s o e hics
ha e o ha e a clea e hical uni beha iou ision, which
an icipa es a eali y d ama ically di e en om adi ional
and con en ional beha iou al p ac ices. Mo eo e , o
achie e e ec i e leade ship o e hics, uni leade s ha e o be
mo i a ed o achie e he e hical ision. This esul s in he
impo an and ascina ing ques ion o which a ainmen s and
disposi ions mo e he leade o e hics o be dis u bed by
such e hical disc epancies and no o passi ely accep (like
mos o he s) he e hical sho comings o uni membe s, bu
a he o be passiona ely commi ed o and ac i e wo k
owa ds he ealiza ion o a adically di e en e hical
ision? The leade beha iou s equi ed o ealize he e hical
ision a e essen ially he same han hose in ol ed in he
ealiza ion o any o he ision. The undamen al posi ion
unde lying he ELI is ha he leade ship compe encies
equi ed o lead ollowe s o e hical excellence in e ms o
he e hical ision a e essen ially he same as hose in ol ed
in achie ing uni pe o mance excellence.
16 S.A .J.Bus.Manage.2005,36(2)
The ELI, like he LBI, hus in e p e s leade ship, as a
complex, con inuous p ocess exp essing i sel in an
ex ensi e a ay o in e -dependen beha iou al ac ions. The
p ocess essen ially en ails h ee sequen ially linked phases
(a) he de elopmen and selling o an app op ia e ye
challenging e hical ision o he uni , based on an
assessmen o he in e nal and ex e nal en i onmen o he
uni , (b) he p epa a ion o he uni o he implemen a ion o
he e hical ision, and inally (c) he bold ye hones
implemen a ion o he ision by con inually moni o ing,
e i alizing, ine- uning and o ches a ing a mul i ude o
p e equisi es o ealizing he e hical uni ision. I should,
howe e , be s essed ha leade ship is he eby no educed
o a ini e, linea ly s ic ly o wa d-mo ing p ocess.
Ac ually, he possibili y aised ea lie o iden i ying causal
in luences amongs i s -o de leade ship dimensions,
con esses ha he exis ence o a complex (Cillie s, 1998),
ichly connec ed leade ship s uc u e is accep ed as a basic
p emise.
F om he pe spec i e o he leade ship model unde lying he
ELI, a o al o 19 dis inc la en leade ship dimensions can
be dis inguished wi hin he o egoing b oad p ocedu al
s uc u e, cons i u ing he abs ac hemes common o
bundles o leade ship beha iou al ac ions. Leading
o ganiza ional uni s o build uni e hical clima e and e hical
cul u e and ul ima ely o achie e e hical uni beha iou hus
ep esen s a o midable ask. Uni leade s e y seldom
would be p o icien in all 19 o he a o emen ioned la en
leade ship dimensions. To imp o e leade e ec i eness, and
ul ima ely he e hical quali y o uni membe beha iou
h ough leade ship assessmen and de elopmen , equi es
he iden i ica ion o hose la en leade ship dimensions on
which a leade pe o ms ela i ely less well. The E hical
Leade ship In en o y was de eloped speci ically o he
Sou h A ican con ex o se e his pu pose.
The esul s epo ed he e p o ide easonable suppo o he
use o he 101-i em ELI. The possibili y o causal
in luences exis ing amongs speci ic i s -o de leade ship
dimensions, howe e , needs o be in es iga ed. The cu en
measu emen model hus needs o be expanded in o a ully-
ledged e hical leade ship s uc u al model. Once
sa is ac o y close i has been achie ed, he model will in
addi ion ha e o be c oss- alida ed (Diaman opoulos &
Siguaw, 2000) on an independen sample o p o ide
su icien e idence o eally use he ELI wi h con idence o
he de elopmen o leade ship o e hics.
Gi en he assumed pi o al ole o leade ship in
o ganiza ional uni pe o mance, he na u e o his p esumed
ela ionship should u he mo e be cap u ed in a
comp ehensi e leade ship-e hical uni beha iou s uc u al
model ha would explain he manne in which he a ious
la en leade ship dimensions, media ed by (amongs o he s)
e hical uni cul u e and e hical uni clima e, a ec he
quali y o e hical uni beha iou . A he same ime, he
ques ion should be esol ed whe he e hical clima e, e hical
cul u e and e hical uni beha iou should be included in he
cons i u i e de ini ion o o ganiza ional uni pe o mance
(Spangenbe g, & The on, 2004; The on, Spangenbe g, &
Henning, 2004; The on, & Spangenbe g, 2002).
Re e ences
And ews, K.R. 1989. ‘E hics in p ac ice’, Ha a d Business
Re iew, 89(5): 99 – 104.
Bass, B.M. 1985. Leade ship and pe o mance beyond
expec a ions. New Yo k: F ee P ess.
Bass, B.M. & A olio, B.J. 1993. ‘T ans o ma ional leade ship:
A esponse o c i iques.’ In Chemme s, M. & Ayman, R.
(Eds.). Leade ship heo y and esea ch pe spec i es and
di ec ions. O lando: Academic P ess.
Bass, B.M. & A olio, B. 1997. Manual o he mul i ac o
leade ship ques ionnai e. Redwood, CA: Mind Ga den, Inc.
Bass, B.M. & S eidlmeie , P. 1999. ‘E hics, cha ac e , and
au hen ic ans o ma ional leade ship beha iou ’,
Leade ship Qua e ly, 10(2):181 – 217.
Ben le , P.M. 1990. ‘Compa a i e i indexes in s uc u al
models’, Psychological Bulle in, 107(2):238 - 246.
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’,
Psychological Bulle in, 88(3):588 - 606.
Bollen, K.A. & Long, J.S. 1993. Tes ing s uc u al equa ion
models. Newbu y Pa k: Sage Publica ions.
B ickley, J.A., Smi h, C.W. & Zimme man, J.L. 1994.
‘E hics, incen i es, and o ganisa ional design’, Jou nal o
Applied Co po a e Finance, 7(2):20-30.
B ien, A. 1998. ‘P o essional e hics and a cul u e o us ’,
Jou nal o Business E hics, 17:391-409.
B owne, M.W. & Cudeck, R. 1993. ‘Al e na i e ways o
assessing model i ’, In Bollen, K.A. & Long, J.S. (Eds.).
Tes ing s uc u al equa ion models. Newbu y Pa k: Sage
Publica ions.
By ne, B.M. 1998. S uc u al equa ion modelling wi h
LISREL, PRELIS, and SIMPLIS: Basic concep s,
applica ions and p og amming. Mahwah, New Je sey:
Law ence E lbaum Associa es.
Chambe lain, L. 2004. ‘The in luence o ans o ma ional
leade ship on o ganisa ional ci izenship beha iou s’.
Mas e s Thesis. Uni e si y o S ellenbosch.
Cillie s, P. 1998. Complexi y and pos mode nism:
Unde s anding complex sys ems. London: Rou ledge.
Conge , J.A. & Kanungo, R.N. 1998. Cha isma ic
leade ship in o ganisa ions. Thousand Oaks, CA: Sage
Publica ions.
Cullen, J.B., Vic o , B. & B onson, J.W. 1993. ‘The E hical
Clima e Ques ionnai e: An assessmen o i s de elopmen
and alidi y’, Psychological Repo s, 73:667-674.
S.A .J.Bus.Manage.2005,36(2) 17
Deshpande, S. P. 1996. ‘E hical clima e and he link
be ween success and e hical beha iou : an empi ical
in es iga ion o a non-p o i o ganisa ion’, Jou nal o
Business E hics, 15:315-320.
Diaman opoulos, A. & Siguaw, J.A. 2000. In oducing
LISREL. London & New Delhi: Sage Publica ions.
Du Toi , M. & Du Toi , S.H.C. 2001. In e ac i e LISREL
use ’s guide. Lincolnwood, IL: Scien i ic So wa e
In e na ional.
Ende le, G. 1987. ‘Some pe spec i es o manage ial e hical
leade ship’, Jou nal o Business E hics, 6:657 – 663.
Engelb ech , A.S., Scheps, A. & The on, C.C. 2004. ‘The
in luence o e hical alues on ans o ma ional leade ship
and e hical clima e in o ganiza ions’. Pape p esen ed a he
Annual Eu opean Academy o Managemen (EURAM)
Con e ence, Uni e si y o S And ews, Sco land.
Fe ando, P.J. & Lo enzo-Se a, U. 2000. ‘Un es ic ed
e sus es ic ed ac o analysis o mul idimensional es
i ems: some aspec s o he p oblem and some sugges ions’,
Psicológica, 21:301-323.
Hai , J.F., Ande son, R.E., Ta ham, R.L. & Black, W.C. 1995.
Mul i a ia e da a analysis wi h eadings. New Je sey:
P en ice-Hall.
Hosme , L.T. 1987. ‘The ins i u ionalisa ion o une hical
beha iou ’, Jou nal o Business E hics, 6:439 – 447.
House, R.J. 1995. ‘Leade ship in he wen y- i s cen u y’.
In Howa d, A. (Ed.). The changing na u e o wo k. San
F ancisco: Jossey-Bass, pp. 411 – 450.
Hu, L.T. & Ben le , P.M. 1995. ‘E alua ing model i ’, In
Hoyle, R.C. (Ed.). S uc u al equa ion modelling: Concep s,
issues and applica ions. Thousand Oaks: Sage Publica ions.
Hulin, C.L., D asgow, F. & Pa sons, C.K. 1983. I em
esponse heo y: Aapplica ion o psychological
measu emen . Homewood, Illinois: Dow Jones-I win.
Hummels, H. 1998. ‘O ganising e hics: A s akeholde
deba e’, Jou nal o Business E hics, 17(13):1403 – 1419.
Hus ed, B.W. 1998. ‘O ganisa ional jus ice and he
managemen o s akeholde ela ions’, Jou nal o Business
E hics. 17:643 – 651.
James, H.S. 2000. ‘Rein o cing e hical decision making
h ough o ganisa ional s uc u e’, Jou nal o Business
E hics, 28:43-58.
Jö eskog, K.G. & Sö bom, D. 1993. LISREL 8: S uc u al
equa ion modelling wi h SIMPLIS command language.
Chicago: Scien i ic So wa e In e na ional.
Jö eskog, K.G. & Sö bom, D. 1996a. PRELIS 2: Use ’s
e e ence guide. Chicago: Scien i ic So wa e In e na ional.
Jö eskog, K.G. & Sö bom, D. 1996b. LISREL 8: Use ’s
e e ence guide. Chicago: Scien i ic So wa e In e na ional.
Jö eskog, K.G. & Sö bom, D. 1998. S uc u al equa ion
modelling wi h he SIMPLIS command language. Chicago:
Scien i ic So wa e In e na ional.
Jö eskog, K.G., Sö bom, D., Du Toi , S. & Du Toi , M.
2000. LISREL 8: New s a is ical ea u es. Chicago:
Scien i ic So wa e In e na ional.
Jose, A. & Thibodeaux, M.A. 1999. ‘Ins i u ionalisa ion o
e hics: The pe spec i e o manage s’, Jou nal o Business
E hics, 22:125 – 140.
Kanungo, R.N. & Medonca, M. 1996. Dimensions o e hical
leade ship. Thousand Oaks: Sage Publica ions.
Kelloway, E.K. 1998. Using LISREL o s uc u al equa ion
modelling; a esea che ’s guide. Thousand Oaks: Sage
Publica ions.
King Commi ee on Co po a e Go e nance. 2002. King
epo on co po a e go e nance o Sou h A ica.
Johannesbu g: Ins i u e o Di ec o s in Sou h A ica.
K a , P, Engelb ech , A.S. & The on, C.C. 2004. ‘The
in luence o ans o ma ional and ansac ional leade ship on
dyadic us ela ionships h ough pe cep ions o ai ness’,
Sou h A ican Jou nal o Indus ial Psychology, 30(1):10-
18.
Lindsay, R.M., Lindsay, L.M. & I ine, V.B. 1996.
‘Ins illing e hical beha iou in o ganisa ions: A su ey o
Canadian companies’, Jou nal o Business E hics, 15: 393 –
407.
MacCallum, R.C. 1995. ‘Model speci ica ion: p ocedu es,
s a egies and ela ed issues’. In Hoyle, R.H. (Ed.).
S uc u al equa ion modelling: Concep s, issues and
applica ions. Thousand Oaks, Cali o nia: Sage Publica ions.
MacCallum, R.C., B owne, M.W. & Sugawa a, H.M. 1996.
‘Powe analysis and de e mina ion o sample size o
co a iance s uc u e modelling’, Psychological Me hods,
1(2):130-149.
Maignan, I. & Fe ell, O.C. 2000. ‘Measu ing co po a e
ci izenship in wo coun ies: The case o he Uni ed S a es
and F ance’, Jou nal o Business E hics, 23:283 – 297.
Ma sh, H.W., Hau, K.T., Balla, J.R. & G ayson, D. 1998. ‘Is
mo e e e oo much? The numbe o indica o s pe ac o in
con i ma o y ac o analysis’, Mul i a ia e Beha iou al
Resea ch, 33(2):181-220.
Mels, G. 2003. A wo kshop on s uc u al equa ion modeling
wi h LISREL 8.54 o Windows. Chicago: Scien i ic
So wa e In e na ional.
Nunnally, J.C. 1978. Psychome ic heo y. McG aw-Hill:
New Yo k.
18 S.A .J.Bus.Manage.2005,36(2)
O`Boyle, E.J. & Dawson, L.E. J . 1992. ‘The Ame ican
Ma ke ing Associa ion code o e hics: Ins uc ion o
ma ke e s’, Jou nal o Business E hics, 11:921 – 932.
Ohmann, O.A. 1989. ‘Skyhooks’. In And ews, K.R. (Ed.).
E hics in p ac ice: managing he mo al co po a ion (pp.58-
69). Bos on: Ha a d Business School P ess.
Schein, E.H. 1992. O ganiza ional cul u e and leade ship.
San F ancisco: Jossey-Bass.
Schepe s, J.M. 1992. Toe skons uksie, eo ie en p ak yk.
Johannesbu g: RAU D ukpe s.
Schumacke , R.E. & Lomax, R.G. 1996. A beginne ’s guide
o s uc u al equa ion modeling. Mahaw, New Je sey:
Law ence E lbaum Associa es, Publishe s.
Spangenbe g, H.H. & The on, C.C. 1997. ‘De eloping a
pe o mance managemen audi ques ionnai e’, Sou h
A ican Jou nal o Psychology, 27(3):143-50.
Spangenbe g, H.H. & The on, C.C. 2002. ‘De elopmen o
a uniquely Sou h A ican leade ship ques ionnai e’, Sou h
A ican Jou nal o Psychology, 32(2):9-25.
Spangenbe g, H.H. & The on, C.C. 2004. ‘De elopmen o
a pe o mance measu emen ques ionnai e o assessing
o ganisa ional wo k uni e ec i eness’, Sou h A ican
Jou nal o Indus ial Psychology, 30(1):19-28.
‘SPSS 11.0 o Windows’. 2004. SPSS Inc.
h p://www.spss.com/
S ong, K.C., Ringe , R.C. & Taylo , S.A. 2001. ‘The ules
o s akeholde sa is ac ion’, Jou nal o Business E hics,
32:219-230.
Tabachnick, B.G. & Fidell, L.S. 1989. Using mul i a ia e
s a is ics (2nd Edi ion). New Yo k: Ha pe Collins
Publishe s.
The on, C.C. & Spangenbe g, H.H. 2002. ‘De elopmen o
a pe o mance measu emen ques ionnai e o assessing
o ganisa ional wo k uni e ec i eness’. Pape p esen ed a
he Global Con e ence on Business and Economics, Pa is.
The on, C.C. & Spangenbe g, H.H. 2004. ‘Towa ds a
comp ehensi e leade ship - uni pe o mance s uc u al
model: The de elopmen o second-o de ac o s o he
leade ship beha iou in en o y (LBI)’. Manusc ip accep ed
o publica ion in Managemen Dynamics.
The on, C.C., Spangenbe g, H.H. & Henning, R. 2004. ‘An
elabo a ion o he in e nal s uc u e o he uni pe o mance
cons uc as measu ed by he pe o mance index (PI)’,
Managemen Dynamics, 13(2):35-52.
T e ino, L.K. & Bu e ield, K.D. 1998. ‘The e hical con ex
in o ganisa ions: In luences on employee a i udes and
beha iou s’, Business E hics Qua e ly, 8:447-476.
T e ino, L.K. & Nelson, K.A. 1995. Managing business
e hics: S aigh alk abou how o do i igh . New Yo k: J
Wiley & Sons.
T e ino, L.K., Ha man, L.P. & B own, M. 2000. ‘Mo al
pe son and mo al manage : How execu i es de elop a
epu a ion o e hical leade ship’, Cali o nia Managemen
Re iew, 42(4):128-142.
Upchu ch, R.S. & Ruhland, S.K. 1996. ‘The o ganisa ional
bases o e hical wo k clima es in lodging ope a ions as
pe cei ed by gene al manage s’, Jou nal o Business E hics,
15:1083-1097.
Vaicys, C., Ba ne , T. & B own, G. 1996. ‘An analysis o
he ac o s uc u e o he E hical Clima e Ques ionnai e’,
Psychological Repo s, 79:115-120.
Va di, Y. 2001. ‘The e ec s o o ganisa ional and e hical
clima es on misconduc a wo k’, Jou nal o Business
E hics, 29:325-337.
Vic o , B. & Cullen, J.B. 1987. ‘A heo y and measu e o
e hical clima e in o ganisa ions’, Resea ch in co po a e
Social Pe o mances and Policy, 9:51-57.
Vic o , B. & Cullen, J.B. 1990. ‘A heo y and measu e o
e hical clima e in o ganisa ions’. In F ede ick, W.C. &
P es on, L.E. (Eds.). Business E hics: esea ch issues and
empi ical s udies. G eenwich CT: JAI P ess, pp. 77-97.
