birdcage
Transformational leadership and leader moral orientation:Contrasting an ethic of justice and an ethic of care
Sheldene K Simola a,⁎,Julian Barling b ,Nick Turner c
a
Business Administration Program,Trent University,1600West Bank Drive,Peterborough,ON,Canada K9J 7B8b
Queen's School of Business,Queen's University,Canada c Asper School of Business,University of Manitoba,Canada
a r t i c l e i n f o a
b s t r a
c t
Previous rearch on the moral foundations of transformational leadership has focud on a
Kohlbergian (1969,1976)ethic of justice.However,propod associations between level of
justice reasoning and transformational leadership have received only partial support.We
reasoned that an ethic of care would be more consistent with the nature of transformational
leadership than would be an ethic of justice.Multilevel regression analys on data obtained
from a sample of leaders (N =55)and followers (N =391)at a Canadian university supported
our predictions.Speci fically,leader propensity toward using an ethic of care was signi ficantly,
positively related to follower perceptions of transformational (but not transactional)
leadership.Leader propensity toward an ethic of justice was signi ficantly,positively related
to follower perceptions of transactional (but not transformational)leadership.Conceptual,
rearch,and practical implications are discusd.
©2009Elvier Inc.All rights rerved.Keywords:Ethics Ethic of care Ethic of justice Leadership Transformational leadership
1.Introduction
During the last decade,there has been a growing interest in the interction of leadership and ethics (e.g.,Banerji &Krishnan,2000;Bass &Steidlmeier,1999;Ciulla,1998a,b;Keeley,1998;Krishnan,2001;Wren,1998).However,despite the suppod centrality of ethics in effective leadership,there remains with few exceptions (e.g.,Brown,Treviño,&Harrison,2005;Treviño,Brown,&Hartman,2003)little empirical rearch in this area.Continued rearch could have important implications for leadership lection,development,and training.
The primary purpo of this study was to contrast two alternative accounts of the relationship between leadership behaviors and moral problem solving orientation.While previous rearch has focud almost exclusively on a Kohlbergian (1969,1976)ethic of justice when asssing the moral foundations of transformational leadership (e.g.,Sivanathan &Fekken,2002;Turner,Barling,Epitropaki,Butcher,&Milner,2002),we contrast two moral reasoning orientations,namely an ethic of care (Gilligan,1982)and an ethic of justice.We argue that an ethic of care would be more consistent with the nature of transformational leadership than would an ethic of justice (Kohlberg,1969,1976),and that this distinction could explain inconsistencies in previous findin
gs showing partial (Turner et al.,2002)or no (Sivanathan &Fekken,2002)support for hypothesized relationships between transformational leadership and an ethic of justice.
The Leadership Quarterly 21(2010)179–188
⁎Corresponding author.
E-mail address:ssimola@trentu.ca (S.K.
Simola).
1048-9843/$–e front matter ©2009Elvier Inc.All rights rerved.
Contents lists available at ScienceDirect
The Leadership Quarterly
j o u r n a l h o m e p a g e :w ww.e l s e v i e r.c o m /l o c a t e /l e a q u a
180S.K Simola et al./The Leadership Quarterly21(2010)179–188
1.1.Transformational leadership,transactional leadership,and moral reasoning
Burns(1978)initially differentiated between“transactional”and“transformational”leadership.Transactional leadership refers to exchanges that advance the purpos of each party in economic,political,or psychological ways.In contrast, transformational leadership goes beyond benefits that accrue to each individual through social exchange,and reflects a relationship in which leaders and followers engage with each other through a shared purpo in ways that transform and elevate their motivation,conduct,and ethical aspirations.
Transformational leadership compris four behavioral dimensions(Bass,1985; Bass&Riggio,2006).The include“idealized influence,”in which leaders demonstrate vision and mission,and rve as role models to followers;“inspirational motivation,”characterized by the inspiration of a shared vision and team spirit directed toward achievement of group goals;“intellectual stimulation,”which reflects the process through which leaders rou followers toward creativity,innovation,and careful problem solving;and,“individualized consideration,”which is manifested when leaders establish a supportive environment in which they attend carefully to the individual and unique needs of followers.The four dimensions of transformational leadership can be differentiated from transactional leadership style,which itlf compris two components,namely“contingent reward”(an exchange in which rewards are contingent upon actions)and“management by exception”(which involves the u of constructive criticism and negative reinforcement).
Two previous studies have examined the relationship between transformational leadership and moral reasoning.Turner et al. (2002)assd whether transformational and transactional leadership were associated with different levels of Kohlbergian (1969,1976)cognitive moral reasoning.Within this framework,individuals are thought to develop through three levels of moral reasoning.In the prec
onventional stage,lf-interest is dominant,and obedience to authority takes place to avoid punishment.The conventional level emphasizes a shared understanding of societal norms and values in decision-making.Postconventional moral reasoning is the third and highest level,in which moral decisions occur bad on universal moral ,life is more important than property).
聚会的英文Turner et al.(2002)argued that individuals with more complex forms of Kohlbergian moral reasoning will be able to draw on more cognitively sophisticated conceptualizations of interpersonal situations.Therefore,they will be more likely to draw on a larger repertoire of ways to respond to life dilemmas,and to identify benefits inhering in tho choices that meet collective as oppod to individual needs.Turner et al.predicted that leaders with higher levels of Kohlbergian reasoning would exhibit more transformational leadership than leaders with lower levels of Kohlbergian reasoning.In addition,becau transactional leadership relies on leader-follower exchange,it does not require an ability to identify a wider range of choices that would facilitate group (as oppod to individual)lf-interest.Thus,level of moral reasoning was predicted to have no relationship to transactional behaviors.The data showed that leaders with preconventional(lowest)levels of moral reasoning demonstrated less trans-formational leadership behaviors than tho with postconventional(highest)levels of moral reasoning.However,no difference in transformational leader
ship was found between leaders at the conventional(moderate)level of moral reasoning relative to leaders at either the preconventional or postconventional levels.Their cond hypothesis,that no differences in moral reasoning would be found among transactional leaders,was supported.
Sivanathan and Fekken(2002)also considered the relationship between transformational leadership and level of Kohlbergian moral reasoning,but they found that transformational leadership was not related to Kohlbergian moral reasoning level.Thus, whereas Turner et al.(2002)found partial support for their hypothes about the relationship between transformational leadership and Kohlbergian level of moral reasoning,Sivanathan and Fekken(2002)found no relationship between the two variables.
Although authors of both of the studies identified potential measurement issues that could account for the discrepancies between predicted and obrved outcomes,1we suggest that specification of the nature of moral problem solving orientation might help account for the relationship between transformational leadership and moral reasoning.Specifically,the ethic of justice (Kohlberg,1969,1976),which was ud in the two aforementioned studies,and an ethic of care(Gilligan,1982)reflect two different approaches to understanding moral reasoning.In the current study,we develop and test different hypothes concerning the relationships between transformation
al and transactional leadership and the two moral orientations of justice and care.
1.2.Comparing moral orientations
Kohlberg's(1969,1976)model of moral reasoning reflects a justice orientation,and is characterized by a focus on adjudicating between individual interests or rights in solving moral dilemmas.This orientation is predicated on impartiality,fairness, reciprocity,and the application of universal moral principles to abstract features of ethical situations.In the preconventional level of moral reasoning,individuals are primarily egocentric in choosing the behaviors that will aid them in avoiding punishment and maximizing lf-interest(Kohlberg,1969,1976).However,as individuals develop cognitively,equality and fairness assume greater importance in moral decision making.In the conventional stage,fairness is evident in a shared understanding of societal norms and respect for conventions and laws,whereas in thefinal,post-conventional stage,fairness is related to equality of persons and reciprocity toward one another.Thus,the postconventional stage is characterized by reliance on universal moral principles that transcend laws.
美国留学学校
1Turner et al.(2002)identified that normative cut-off points dividing scores on their measure of moral reasoning might have been inadequate to differentiate levels of reasoning.Sivanathan and Fekken(2
002)hypothesized that weak internal consistency of their measure of Kohlbergian reasoning combined with a
In contrast,Gilligan's (1982)focus on the ethic of care emerged in respon to methodological concerns related to Kohlberg's (1969,1976)rearch.Speci fically,Kohlberg only studied males,and hypothetical dilemmas rather than actual ethical dilemmas experienced by the respondents themlves.Gilligan studied women confronted with actual moral dilemmas and in doing so,found evidence of an alternative moral orientation that was characterized by authentic relationships re flecting concern with understanding the subjective experiences and needs of others,and by being genuinely responsive to the.Within Gilligan's perspective,individuals demonstrating a care orientation would not focus on adjudicating between competing rights as would be the ca in a Kohlbergian perspective.Instead,a care orientation would focus on identifying creative ways of simultaneously ful filling competing responsibilities to others.Although Gilligan's rearch focud on the moral reasoning of women,subquent meta-analysis has shown that u of an ethic of care is not strongly gender differentiated (Jaffee &Shibley-Hyde,2000).
In addressing the ethics of transformational and transactional leadership,Bass and Steidlmeier (1999)noted that although both forms of leadership have clear philosophical foundations and compo
nents,the exact nature of the differs.Speci fically,the ethical issues that have salience to the transformational leadership construct as a whole are tho that re flect a concept of lf that is “connected …(wherein)one's moral obligations …are grounded in a …conceptualization of individuals within community ”(Bass &Steidlmeier,1999,p.186).This characteristic of interconnection,en as ethically central to transformational leadership,is also central to an ethic of care,as oppod to an ethic of justice in which notions of paration and autonomy would be critical.
In terms of the four components comprising transformational leadership,“idealized in fluence ”is associated with the notion of a “universal brotherhood ”(Bass &Steidlmeier,1999,p.187),as oppod to a n of community bad on “we-they ”value differences.This universal brotherhood also connotes the importance of interconnection central to a care approach,rather than the “we-they ”distinctions re flecting the underlying justice characteristic of autonomy or paration from others.This n of interconnection also involves being true to both lf and others (Price,2003)in a way that is consistent with a care approach.Similarly,the “inspirational motivation ”component of transformational leadership implies the development of a shared vision,rather than pursuit of individual goals.The n of community and connection implied by a shared vision is consistent with an ethic of care,whereas the primacy of individual rights and lf-interest in pursuing individual g
oals is more re flective on an ethic of justice.Additionally,the third component of transformational leadership,individualized consideration,implies responsiveness to the unique,subjective needs of followers,which again is consistent with an ethic of care.Finally,the “intellectual stimulation ”component of transformational leadership emphasizes creativity in the arch for ideals.The u of creative,win –win problem solving approaches is also consistent with Gilligan's (1982)ethic of care.Within a care perspective,creativity speci fically refers to the tendency toward identifying ways to simultaneously ful fill competing responsibilities (Gilligan,1982;Reiter,1996),as oppod to choosing between the con flicting rights of different individuals.That is,care-bad solutions are creative in that they rely on identifying and emphasizing the underlying interests of each party,in order to expand the array of options available,thereby facilitating non-zero sum solutions in which each party can win.Tendencies within a care approach toward simul-taneously ful filling competing responsibilities contrast with tendencies within a justice approach toward arbitrating between con flicting rights.As detailed by Reiter (1996),decision-making within the justice framework tends not to focus on elucidation of underlying interests to attain wins for each party.Rather,in emphasizing the adjudication of individual rights,the justice approach tends to support the rights of one party only.It is therefore less likely to demonstrate creativity in the arch for ideals that would be more characteristic of the intellectual stimulation component of transformational leadership.Bad on the nature of transformational leadership,and the characteristics of an ethic of care,we propo:warned
Hypothesis 1.Leaders with a higher propensity toward using an ethic of care will be perceived by their followers as being more transformational than tho with a lower propensity toward using an ethic of care.
In contrast,Bass and Steidlmeier (1999)argued that the ethical values salient to transactional leadership as an overall construct are tho that are associated with individualist philosophies,wherein the primacy of lf-interest among autonomous individuals leads to exchange-bad transactions.Moreover,the moral legitimacy of the transactions requires fairness in all such transactions.As indicated in previous discussion,the characteristics are all consistent with Kohlberg's (1969,1976)justice orientation.In terms of the components comprising transactional leadership,contingent reward implies that individual rewards given by the leader will be conditional upon the enactment of certain tasks or behaviors by the follower.Similarly,active management-by-exception implies that punishment will be implemented by the leader if followers fail to demonstrate required behaviors or complete certain tasks.Both of the dimensions re flect a form of exchange-bad transaction consistent with the underlying assumptions of a justice approach,namely paration,autonomy,and principles of fairness in exchange.Thus,
Hypothesis 2.Leaders with a higher propensity toward using an ethic of justice will be perceived by t
heir followers as being more transactional than tho with a lower propensity toward using an ethic of justice.
2.Method
2.1.Procedure and participants
The sample comprid 55employees holding leadership positions in a Canadian university,along with 391of their followers.All leaders were residence life staff with responsibilities in a number of areas,including identi fication and management of 181
S.K Simola et al./The Leadership Quarterly 21(2010)179–188
182S.K Simola et al./The Leadership Quarterly21(2010)179–188
of social and learning opportunities for followers;creation of a n of community;and,leadership related to fundraising activities.Followers were students living in residence,each of whom reported to a given leader,and had responsibilities in accordance with university policies and regulations.The respon rate for leaders was57%.Leaders(74.5%females)ranged in age from18to23years(M=20.04years,SD=1.14years).The respon rate for followers was21.7%.Universit
y policy precluded collection of follower age and gender information.Follower gender and age were not salient to the hypothes of the current study.For the population of all residence members,the average age was18.9years,and61.3%were women.The number of followers per leader ranged between2and17(median=6).
Each eligible leader and follower received an email invitation to participate.The invitation contained a confidential and individualized password that allowed access to the online survey,as well as the subquent linkage of follower respons to their respective leaders.Leaders completed their survey as part of a larger group of questionnaires that were administered for other purpos,and were paid ten dollars per hour for completing the entire package.Followers also completed their survey as part of a larger group of questionnaires.They were offered the option of having their names entered into a draw for one of six$300cash prizes distributed across residence units,such that the probability of winning in any draw would be3.33%.
2.2.Measures
2.2.1.Measure of moral orientation
Leader propensities toward using ethics of justice and care were assd using the nine moral dile
mmas comprising thefirst component of the Measure of Moral Orientation[MMO](Liddell&Davis,1996;Liddell,Halpin,&Halpin,1992).2Each of the nine dilemmas was accompanied by between six and nine statements reprenting possible respons to the dilemma,half of which were from a justice ,“I would treat the administrator like everyone el,regardless of his personal circumstances”), and half of which were from a care ,“I would want to talk with her andfind out more about her life before making a decision”).Following each dilemma,participants indicated on a4-point Likert scale(1=strongly agree,4=strongly disagree)the extent to which they agreed with each of the possible respons.The mean score across all dilemmas on respons reflecting a justice orientation yielded a propensity toward justice score,and the mean score across all dilemmas on respons reflecting a care orientation yielded a propensity toward care score.
The MMO has previously shown adequate internal consistency for both the justice(.73)and care(.84)components(Liddell et al.,1992).The justice and care components have also shown evidence of convergent and discriminant validity.For example, whereas the care component of the MMO was significantly,positively related(r=.27,p b.01)to the World View Questionnaire [WVQ](Strander&Jenn,1993),which asss the extent to which individuals prefer a worldview that enco
mpass care for others,the justice component of the MMO was not significantly related to WVQ care scores(Liddell&Davis,1996;Liddell,2006). Similarly,whereas MMO justice scores were significantly,positively related(r=.41,p b.001)to conventional levels of justice reasoning as measured by Rest's(1986)Defining Issues Test,MMO care scores were unrelated to the scores(Liddell&Davis, 1996;Liddell,2006).Internal consistencies from the current study were.64and.77for the justice and care scales,respectively.
2.2.2.Multifactor leadership questionnaire
Follower perceptions of transformational and transactional leadership were measured using the45-item Multifactor Leadership Questionnaire Form5×—Short Instrument[MLQ](Bass&Avolio,2000).3Followers rated the frequency(0=not at all, 4=always)with which their leaders demonstrated behaviors associated with both leadership styles.Five subscales measure transformational leadership.They include attributed idealized infl,“Goes beyond his/her own lf-interest for the good of the group”),behavioral idealized infl,“Specifies the importance of having a strong n of purpo”),inspirational ,“Articulates a compelling vision of the future”),intellectual ,“Seeks differing perspectives when solving problems”),and,individualized ,“Treats each of us as individuals with differe
nt needs,abilities,and aspirations”).An additional two subscales measure transactional leadership:contingent reward and active management by exception.We did not asss the passive-avoidant/laisz-faire leadership factor that was identified by Bass and Avolio(2000)as non-transformational and non-transactional,nor did we asss the leadership outcomes assd by the MLQ such as leader effectiveness and satisfaction with leadership.
Within-group agreement with respect to follower perceptions of transformational leadership as indicated by the intraclass correlation,ICC(1),was.08,and the reliability of the group mean scores as reflected by the ICC(2)value was.38.Similarly,within-group agreement with respect to perception of transactional leadership as indicated by ICC(1)was.07,with the reliability of group mean scores as indicated by the ICC(2)being.34,supporting the u of multilevel analysis,as oppod to upward aggregation of data.
Some concern exists about the appropriate factor structure of the MLQ(Bass&Avolio,2000).Although this might suggest the need for a confirmatory factor analysis,such analys are problematic when follower groups are of unequal(unbalanced)sizes. This is becau when using of a Full Information Maximum Likelihood estimator,a different between-group model is specified
2MMO items are reproduced here by special permission of Deborah Liddell.
3MLQ items are reproduced here by special permission of the Distributor,Mind Garden,Inc.,855Oak Grove Ave.,Suite215,Menlo Park,CA94025,USA www. from the Multifactor Leadership Questionnaire by Bernard M.Bass and Bruce J.Avolio.Copyright1995by Bernard M.Bass and Bruce J.Avolio.
for each group size.Although the Muthén Maximum Likelihood (Muthén,1989,1994)estimation algorithm provides a good maximum likelihood approximation,simulation studies show it to be effective only for group sample sizes of at least 100(Hox &Mass,2001);smaller group sample sizes have been associated with biad standard errors.Given the limitations,it was not possible to conduct a con firmatory factor analysis for the current sample.Rather,we relied on Bass and Avolio's (2000)factor model bad on 3,786respondents from fourteen parate samples.
2.2.
3.U of transformational and transactional full-scales versus constituent dimensions
When lecting predictor variables for the current study,we ud the overall scores that leaders received for transformational and transactional leadership,rather than the scores leaders received on the components comprising transformational and transactional leadership for two reasons.First,our g
oal was to address previous findings of partial (Turner et al.,2002)or no (Sivanathan &Fekken,2002)support for propod relationships between level of Kohlbergian justice reasoning and transformational leadership.The previous studies ud overall transformational and transactional leadership scores when asssing propod relationships with level of Kohlbergian justice reasoning.To be consistent and facilitate comparison with previous rearch studies,we also opted to ud full-scale leadership scores.Second,the constituent dimensions of transformational leadership tend to be very highly correlated with one another,making it dif ficult to isolate unique associations with other variables.For this reason rearchers often combine the constituent dimensions into an overall scale (Judge &Piccolo,2004;Judge,Woolf,Hurst,&Livingston,2008).
2.2.4.Statistical controls
First,becau prior rearch (e.g.,Bono &Judge,2004)has shown consistent relations between transformational and transactional leadership,we controlled for transactional leadership when transformational leadership was the outcome,and vice versa.Second,previous rearch has demonstrated a relationship between gender and transformational leadership (Bass,Avolio,&Atwater,1996;Eagly,Johannen,Schmidt,&van Engen,2003),as well as age and moral reasoning (Gilligan,Rogers,&Tolman,1991;Rest,1994).Consistent with Turner et al.(2002),therefore,w
e included both leader age and gender as control variables.
3.Results
Table 1prents the means,standard deviations,and intercorrelations among variables.Note,however,that the current study was multilevel in nature,with each cluster of followers being unique to and nested within a particular leader.We therefore ud hierarchical linear modeling (Raudenbush &Bryk,2002)to address dependencies within such clusters that might otherwi bias tests of signi ficance (Muthén &Satorra,1989).
3.1.Multilevel regression analys
Two multilevel regression analys were computed using HLM6(Raudenbush,Bryk,&Congdon,2000).First,follower perceptions of transformational leadership (level-1)were regresd onto follower perception of transactional leadership (level-1),leader gender and age (both level-2),and leader propensities toward using ethics of care and justice (both level-2).Second,follower perceptions of transactional leadership (level-1)were regresd onto transformational leadership (level-1),leader gender and age (both level-2),and leader propensities toward using ethics of justice and care (both level-2).
For each of the two regressions,we fitted three quential models.The first of the was a one-way random effects analysis of variance (ANOVA)model to test whether there was signi ficant variance in leadership styles between groups.The cond of the was a random-coef ficient regression (i.e.,one-way random effects ANCOVA)ud to test both our assumption of signi ficant variation in intercepts across groups,and our assumption that the mean slope between level-1variables was signi ficantly different from zero.The third model fitted for each regression was an “intercepts-as-outcome ”model ud to asss the actual hypothes for the current study,as oppod to the assumptions underlying them.This model therefore assd whether differences in moral problem solving orientation among leaders were related to follower perceptions of leadership style.We had no a priori conceptual grounds on which to expect differences among the slopes relating level-1variables,and ultimately were predicting and including variables with which to explain variation across groups only with respect to intercepts.Therefore,we held the slopes constant Table 1
Means,standard deviations,and intercorrelations among variables.
Variable
M SD 12341.
Care 3.05.21(.77)2.
Justice 2.88.19.08(.64)3.
TRF 2.57.82.08−.06(.94)4.TRA 2.10.77.01.04.75⁎⁎(.77)Note.N =391.Internal consistency reliabilities appear in parenthes along the diagonal.The correlations among care,justice and all other variables were computed using N =391.Therefore,care and justice scores for each group were assigned down to individual followers within tho groups.Thus,the effective N for justice and care is 55.TRF =transformational leadership.TRA=transactional leadership.
183
S.K Simola et al./The Leadership Quarterly 21(2010)179–188
across groups for the analys.4Bad on the conceptual models pod,grand-mean centering was ud in all analys (Hofmann &Gavin,1998;Kreft,De Leeuw,&Aiken,1995).The results of the three models ud for each of the two regressions are summarized in Table 2.
3.2.Relationship between transformational leadership and an ethic of care
The chi-square test associated with the one-way random effects ANOVA model indicated that the between-leader variance was signi ficantly different from zero,[χ2(54)=87.40,p b .01].Therefore,we proceeded to fit the cond model.
sbhResults of the random-coef ficient regression model demonstrated that the mean across groups for the slopes relating follower perceptions of transformational leadership to follower perceptions of transactional leadership was statistically signi ficant
[slope=.79;t (389)=18.85,p b .001].The R 2value for the level-1model was 54%.As indicated by Hofmann,Grif fin,and Gavin (2000),this value re flects the proportion of variance accounted for by transactional leadership to the total within-group variance
Table 2
Results of the estimated models.a Models for TRF and TRA c as L1dependent Variables Parameter estimates b
γ00γ01
to 04γ10γ11σ2π00π11One-way ANOVA L1:Transformational leadership ij =β0j +r ij
2.55–––.6
3.06–L2:β0j =γ00+U 0j
Random-coef ficient
freeze的过去式L1:TRF Ldrshp ij =β0j +β1j (TRA Ldrshp ij )+r ij
L2:β0j =γ00+U 0j
2.56–.79–.29.02–β1j =γ10
Intercepts-as-outcome
L1:TRF Ldrshp ij =β0j +β1j (TRA Ldrshp ij )+r ij
γ01=0.13L2:β0j =γ00+γ01(Gender j )+γ02(Age j )+
bye for now2.56γ02=0.0
3.79–.29.01–γ03(Care j )+γ04(Justice j )+U 0j
γ03=0.01β1j =γ10
英语教学案例范文
γ04=−0.01One-way ANOVA
会计培训学校L1:Transactional Leadership ij =β0j +r ij
2.09–––.56.04–L2:β0j =γ00+U 0j
Random-coef ficient
L1:TRA Ldrshp ij =β0j +β1j (TRF Ldrshp ij )+r ij
L2:β0j =γ00+U 0j
2.10–.70–.25.01–β1j =γ10
Intercepts-as-outcomes
L1:TRF Ldrshp ij =β0j +β1j (TRA Ldrshp ij )+r ij
γ01=−.02γ02=.01L2:β0j =γ00+γ01(Gender j )+γ02(Age j )+
2.10γ03=−.01.71–.25.01–
γ03(Care j )+γ04(Justice j )+U 0j
γ04=.01β1j =γ10γ01to 04=Slopes of level-2regression predicting β0j .
保护水资源作文γ10=Intercept of level-2regression predicting β1j .
γ11=Intercept of level-2regression predicting β1j .
σ2=Variance in level-1residual (variance in r ij ).
π00=Variance in level-2residual for models predicting β0j .
π11=Variance in level-2residual for models predicting β1j .
U 0j =Variance in intercepts.
a Organization of Table 2bad on Hofmann et al.(2000).
b γ00=Intercept of level-2regression predicting β0j .
c TRF Ldrshp =Transformational Leadership.TRA Ldrshp =Transactional Leadership.
4Becau we ud an “intercepts-as-outcomes ”model in which we neither predicted nor included variables to account for variability in slopes across groups,we made an a priori conceptual decision to constrain slopes across groups.When transformational leadership was the criterion,results indicated no signi ficant variation in slopes [χ2(54)=68.02,p =.10].When transactional leadership was the criterion,there was some variation in slopes across groups [χ2(54)=73.25,p b .05].Therefore,we investigated what would happen if the constraint holding slopes constant for the cond regression was removed.The key result is degraded slightly,now at p =.06.With a signi ficance level of .06,the result did not change signi ficantly.Therefore,the result can be considered suf ficiently 184
S.K Simola et al./The Leadership Quarterly 21(2010)179–188
