To what extent is the Mayer and Salovey (1997) model of emotional intelligence a useful predictor of leadership style and perceived leadership outcomes in Australian educational institutions?

Abstract

Researchers have found that transformational leadership is related to positive outcomes in educational institutions. Hence, it is important to explore constructs that may predict leadership style in order to identify potential transformational leaders in assessment and selection procedures. Several studies in non-educational settings have found that emotional intelligence is a useful predictor of transformational leadership, but these studies have generally lacked methodological rigor and contextual relevance. This project, set in Australian educational institutions, employed a more rigorous methodology to answer the question: to what extent is the Mayer and Salovey (1997) model of emotional intelligence a useful predictor of leadership style and perceived leadership outcomes? The project was designed to move research in the field forward by using valid and reliable instruments, controlling for other predictors, obtaining an adequately sized sample of current leaders and collecting multiple ratings of their leadership behaviours. The study (N = 144 leaders and 432 raters) results indicated that emotional intelligence was not a useful predictor of leadership style and perceived leadership outcomes. In contrast, several of the other predictors in the study were found to predict leadership style.

Keywords

Emotional intelligence leadership assessment and selection transformational leadership leadership style

Introduction

A body of international empirical research now exists which confirms the importance of effective school leadership (Leithwood and Jantzi, 2005; Leithwood et al., 2004, 2006). Influenced by this body of work, governments around the world are allocating considerable resources to the assessment, selection and development of school leaders. Christie and Lingard (2001) propose that an increase in the prominence of institutional performance has created a new discourse of effectiveness, efficiency and accountability. This emphasis on performance has placed school leaders under unprecedented pressure to meet performance targets as in the realms of the corporate world and has led to a growing interest in cross-disciplinary approaches to educational leadership. Hence, educational researchers have become more interested in leadership theories such as transformational leadership (e.g. Bass, 1985) which originated in the management literature. According to Leithwood and Sleegers (2006) future research which explores transformational leadership is especially timely, as questions about the relative value of various approaches to school leadership are being raised by researchers and human resource practitioners who are aiming to meet the mandates of reform-seeking policy makers.

Literature review

Leadership

Leadership is considered by many to be an essential function in organisations, including educational institutions, and as such it attracts ongoing research interest. Yukl (2002: 7) defines leadership as “the process of influencing others to understand and agree about what needs to be done and how it can be done effectively, and the process of facilitating individual and collective efforts to accomplish shared objectives". Many researchers argue that successful schools are headed by principals who are effective leaders who have a clear sense of direction for their schools (Waters et al., 2003). There is also a considerable amount of research which confirms that the achievements of students increase when effective leadership is practised in schools (Andrews and Soder, 1987; Waters et al., 2003). In addition to providing a clear direction for their schools, in the current climate of quality league tables, self-management and distributed leadership, educational leaders are now under unprecedented pressure to meet corporate managerial performance targets (Christie and Lingard, 2001). Therefore, it is important to be able to identify potentially effective leaders in assessment and selection procedures in order to ensure that more schools are led by effective leaders.

Several theories have been developed to conceptualise and explain leadership. Theories such as trait theory (e.g. Mann, 1959; Stogdill, 1948), contingency theory (e.g. Evans, 1970; Fiedler, 1967; Hersey and Blanchard, 1969; Vroom and Yetton, 1973) and transformational leadership theory (e.g. Avolio et al., 1995; Bass, 1985; Burns, 1978; Conger and Kanungo, 1987; Podsakoff et al., 1990; Yukl, 1989) have all made an important contribution to the field, each emphasising the importance of individual differences in the leadership process.

Transformational leadership theory (e.g. Avolio et al., 1995; Bass, 1985) led to the development of the full range leadership model (Avolio, 1999; Bass, 1999) which has generated a considerable amount of confirmatory research. Full range leadership (Avolio, 1999; Bass, 1999) identifies three contrasting leadership styles; transformational, transactional and passive/avoidant. Each leadership style, scales and characteristic leader behaviours are presented in Table 1.

Table 1.

Full range leadership styles, scales and example behaviours identified by Avolio (1999) and Bass (1999).

Leadership Style	Scales	Example Behaviours
Transformational leadership	Idealised attributes	Committed and trustworthy
	Idealised behaviours	Ethical consequences of decisions are considered important
	Inspirational motivation	Confident, articulates vision of future and encourages others
	Intellectual stimulation	Questions the norm and facilitates expression of ideas
	Individualised consideration	Considers individual abilities, needs and aspirations
Transactional leadership	Contingent reward	Negotiates for resources and rewards achievements
	Management-by-exception active	Takes action following mistakes
Passive/avoidant leadership	Management-by-exception passive	Does not take action until mistakes are noticed and problems escalate
	Laissez-faire	Unwilling to accept responsibilities and not present when needed

Traditionally, the implementation of instructional and transactional leadership methods has been popular in educational settings. Instructional leadership has been described as a top-down hierarchical style that focuses on the growth of students, but not on the growth of teachers (Liontos, 1992). Instructional leadership was originally conceived as a role carried out by school principals, and during the 1980s very little reference was made to teachers, department heads or even assistant principals (Hallinger, 2005), whereas transactional leadership involves an exchange of services for various rewards which are controlled, and may be manipulated, by the leader (Bass, 1985). The limitations of instructional and transactional leadership led to interest in the application of transformational leadership in schools. Transformational leaders (as depicted in Table 1) inspire and intellectually stimulate followers, encouraging their confidence and motivation to broaden and elevate their goals (Bono and Judge, 2004). Subsequently, educational leadership researchers have found that transformational leadership has considerable advantages when compared with instructional or transactional leadership (Leithwood, 1994; Liontos, 1992; Sergiovanni, 1984, 1990). Furthermore, many studies have found that transformational leadership is related to positive outcomes such as the satisfaction, motivation and performance of followers in organisations (Judge and Piccolo, 2004; Lowe et al., 1996), including in educational institutions (Chin, 2007; Leithwood and Jantzi, 2005).

Leithwood and Jantzi (2005) conducted a meta-analysis of 33 empirical studies that examined transformational leadership in schools. The nature of transformational leadership, its antecedents, and the variables that moderate and mediate its effects on students were analysed. Leithwood and Jantzi (2005) reached several important conclusions based on the meta-analysis. Firstly, the authors asserted that the effects of transformational leadership on perceptions of organisational effectiveness were large. They argued that the studies reviewed demonstrated the significant positive impact of transformational leadership on teacher commitment and job satisfaction as well as school culture, learning, planning and change strategies. Leithwood and Jantzi (2005) asserted that there was sufficient evidence to indicate that each of these individual and organisational variables contributed positively to student learning. Secondly, the effects of transformational leadership on objective measures of organisational effectiveness were less well known, but the existing studies reported positive effects which were modest in size. Thirdly, the authors reported that the effect of transformational leadership on independently measured student outcomes was promising but only a limited number of studies existed. Lastly, Leithwood and Jantzi (2005) reported that recent evidence about the effect of transformational leadership on student engagement was uniformly positive, but was limited by the small number of existing studies.

As transformational leadership is considered to be an effective form of school leadership, human resource practitioners need to seek out leaders whose dominant leadership style will be transformational. This may be achieved through effective selection and assessment processes, or by training. As training is costly, complex and has uncertain outcomes, identifying potential transformational leaders prior to their engagement is arguably the most efficient method of ensuring that the dominant leadership style of more educational leaders is transformational. In order to achieve this goal human resource practitioners need to be able to predict leadership style. Hence, it is necessary to identify the antecedents of transformational leadership behaviours.

Although many studies have assessed the outcomes of transformational leadership style in the workplace (Chin, 2007; Judge and Piccolo, 2004; Lowe et al., 1996), the antecedents of transformational leadership have received less attention from researchers. The antecedents of leadership behaviours can be divided into three domains: leader-focused (e.g. general mental ability, personality, integrity), follower-focused (e.g. follower efficacy) and situation-focused (e.g. organisational culture) (Rubin, 2003). Leader-focused leadership selection processes normally assess individual differences by including both cognitive and non-cognitive components of assessment. Many empirical studies have found that cognitive ability, also referred to as general mental ability (GMA), is a strong predictor of job performance (see review by Schmidt and Hunter, 1998). This is especially true for jobs requiring complex tasks, and leadership is considered to be a complex task (e.g. developing strategies, solving problems and motivating employees). However, Goldstein et al. (2002) argue that personnel selection processes often overemphasise the importance of GMA and suggest that researchers need to design and test alternative measures of intelligence whilst continuing to explore the predictive ability of non-cognitive measures. Other cognitive components under investigation include multiple intelligences (Gardner, 1983).

Research related to non-cognitive constructs and job performance has highlighted the importance of predictors such as personality factors (especially conscientiousness) (Judge et al., 2002) and integrity (Schmidt and Hunter, 1998). Although integrity is considered to be a useful predictor of job performance (Schmidt and Hunter, 1998), little is known about its impact on leadership behaviours. Hence, further research is required which clearly identifies the antecedents of leadership behaviours.

There is also considerable disagreement among researchers concerning the extent to which the leadership behaviours of men and women differ. Several studies in organisational settings have found that female leaders may be more transformational than males (Bass and Avolio, 1994; Bass et al., 1996; Eagly et al., 2003). However, Yukl (2002) has questioned the extent to which conclusions about the interaction of gender and leadership can be drawn as many leadership studies do not report the gender of leaders, which makes it difficult to conduct meta-analyses of this interaction. Therefore, there is still some disagreement regarding the impact of gender on leadership behaviours and more research is required in this area.

Emotional intelligence

Several researchers have proposed that emotional intelligence (EI) is one construct that may account for hitherto unexplained variance in transformational leadership (Ashkanasy and Daus, 2005; Mayer, 2001; Watkin, 2000). Ashkanasy and Daus (2005) propose that EI is a construct which is worthy of further investigation as a predictor of leadership style. Interest in EI generates from the possibility that it may account for aspects of workplace performance that cannot be accounted for by other constructs (Mayer, 2001; Watkin, 2000). According to Greenockle (2010), it has become important for academic leaders to understand the role EI plays in leadership effectiveness. Greenockle (2010) argues that in addition to inspiring others through their vision, academic leaders must now also be able to execute their vision successfully. Greenockle (2010) suggests that this requires a level of teamwork and cooperation which surpasses that required by the traditional top-down leadership approach, and that the EI of the academic leader is an important factor in determining whether or not the execution of the vision will succeed.

Alternative concepts of EI have been developed by Salovey and Mayer (1990), Goleman (1995, 2001) and Bar-On (1997). Mayer and Salovey presented an updated model in 1997 which separated the concept of EI from personality traits and confined it to a mental ability. Mayer and Salovey (1997: 10) believe EI involves “abilities to perceive, appraise, and express emotion; to access and/or generate feelings when they facilitate thought; to understand emotion and emotional knowledge; and to regulate emotions to promote emotional and intellectual growth”. In the Mayer and Salovey (1997) model, total EI is comprised of two areas: experiential and strategic. Each area is divided into two branches which consist of psychological processes ranging from basic to complex. The experiential area consists of branches related to perceiving emotions and using emotions whilst the strategic area consists of branches related to understanding emotions and managing emotions. Each stage of the model includes levels of abilities that are completed in sequence before progression to the next branch occurs (Mayer and Salovey, 1997). The dimensions of the Mayer and Salovey (1997) model and the emotional abilities they encompass are presented in Table 2.

Table 2.

Areas, branches and abilities of Mayer and Salovey’s (1997) model of emotional intelligence.

EI area and branch	Emotional abilities
Experiential area Perceiving emotions branch	The perception and appraisal of emotion. Assessed by how well emotions and emotional content can be identified
Experiential area Using emotions branch	The emotional facilitation of thinking. Describes emotional events that assist intellectual processing
Strategic area Understanding emotions branch	Assessed by how well emotions are understood and analysed. Requires the ability to recognise and interpret emotions
Strategic area Managing emotions branch	Requires conscious regulation of emotions to promote emotional and intellectual wellbeing

Note. Total EI comprises the two areas and their component branches.

The Mayer and Salovey (1997) model of EI is referred to as an ‘abilities’ model, meeting the criteria of a traditional intelligence focusing on emotions and their interactions with thought. The ‘mixed’ models developed by Goleman (1995, 2001) and Bar-On (1997) are considered to have considerable conceptual weaknesses, mainly because they draw too heavily on personality traits and dispositions (McCrae, 2000). Furthermore, some of the claims about the significance of EI made by Goleman (1995) have not been supported by empirical studies (Mayer et al., 2000, 2004). Mixed models emphasise how abilities and personality factors determine success. Convergent validity between the mixed models and the abilities model is low enough to suggest that they may be measuring different constructs altogether (Matthews et al., 2002). Critiques of the mixed models indicate that these models of EI extend beyond what is normally considered part of emotion or intelligence theory, drawing heavily on personality traits or dispositions (McCrae, 2000).

Momentum is now growing for the Mayer and Salovey (1997) abilities model to be considered the most useful model of EI (Antonakis and Dietz, 2010; Ashkanasy and Daus, 2005; Daus and Ashkanasy, 2005), and this model is considered to be worthy of further investigation (Van Rooy and Viswesvaran, 2004). There have been premature attempts to apply EI concepts in the workplace (Antonakis et al., 2009), but more empirical research needs to be undertaken to assess the EI construct prior to its further application (Ashkanasy and Daus, 2005; Daus and Ashkanasy, 2005).

Relationship between EI, leadership style and leadership outcomes

Most studies in the field have explored aspects of the relationship between EI, leadership style and leadership outcomes in non-educational settings. The existing studies in educational settings have usually employed a mixed model of EI rather than the ability-based Mayer and Salovey (1997) model.

Sivanathan and Fekken (2002) undertook an examination of the relationship between transformational leadership, EI, moral reasoning and leader effectiveness in a Canadian university. Fifty eight dons completed the Emotional Quotient Inventory (EQ-i) (Bar-On, 1997) and the Defining Issues Test (Rest, cited in Sivanathan and Fekken 2002). Raters were 12 supervisors and 232 residents who completed the Multifactor Leadership Questionnaire (MLQ) (Avolio et al., 1995). Leaders reporting greater EI were perceived by the residents to display more transformational leadership behaviours. They were also perceived to be more effective leaders.

In another Canadian study, Stone et al. (2005) reported on the Ontario Principals’ Council leadership study which explored the relationship between EI and school leadership. Principals and vice-principals (N = 464) from nine school boards completed an online version of the EQ-i (Bar-On,1997). Leadership skills were rated by the immediate supervisor, peers and followers of each participant using a 21-item leadership abilities questionnaire. Consistent with previous research using the EQ-i (Bar-On, 1997), women were found to score higher than men on the interpersonal dimension. However, no differences in EQ-i (Bar-On,1997) scores were found between principals and vice-principals. The authors found that men and women did not differ on any of the leadership ratings. However, principals were rated higher than vice-principals by their supervisors on task-oriented leadership, relationship-oriented leadership and total leadership. Vice-principals were rated higher by their staff on relationship-oriented leadership. The authors suggested that educational boards should consider the use of EI measures in the recruitment process for new school administrators and in succession planning.

In a Kenyan study, Ayiro (2009) investigated the relationship between the EI of school principals (N = 100) and their performance on annual targets as rated by their immediate supervisors. EI was measured using the Mayer–Salovey–Caruso Emotional Intelligence Test (MSCEIT) (Mayer et al., 2002). Ayiro (2009) found that total EI scores, and the perceiving emotions and using emotions branches of EI were positively related to performance ratings, whereas the understanding emotions and managing emotions branch scores were not significantly related to performance ratings. Ayiro (2009) recommended that further research which examined the relationship between EI and the performance of school principals should be conducted.

A meta-analysis of 48 studies with a total of 7343 participants was conducted by Mills (2009) to ascertain if there was enough empirical evidence to support the inclusion of EI as a component of leader effectiveness. The studies were based in various organisational contexts and represented the main conceptualisations of EI, including the Mayer and Salovey (1997) model. The meta-analysis yielded a combined effect of .38 which can be interpreted as a moderate relationship between EI and leader effectiveness. Consequently, Mills (2009) advocated the inclusion of EI in the curriculum of leadership preparation programmes.

Harms and Crede (2010) conducted a meta-analysis to assess claims that EI is significantly related to the full range of leadership behaviours (Avolio, 1999; Bass, 1999). In total, 62 independent studies consisting of data from 7145 leaders based in various organisational contexts were included in the meta-analysis. Several different measures had been used in the studies but the MSCEIT (Mayer et al., 2002) was the most frequently used measure of EI (k = 12) and the MLQ (Avolio et al., 1995) was the most frequently used measure of leadership behaviours (k = 39).

Harms and Crede (2010) reported that the relationship between EI and transformational leadership was strong (k = 62, N = 7145, ρ = .41). However, the analysis indicated that the validity estimate was much higher (k = 47, N = 4994, ρ = .59) when ratings of EI and leadership behaviours were provided by the same source (e.g. both self-report) compared with when ratings of the constructs were derived from different sources (e.g. self, peer, supervisor and follower) (k = 22, N = 2661, ρ = .12). Agreement between same-source and multi-source ratings was low for both transformational leadership (ρ = .14) and EI (ρ = .16).

Harms and Crede (2010) also performed separate analyses to assess the relationship between transformational leadership and the different EI measures. Trait, or mixed, measures of EI demonstrated higher validities than abilities measures of EI. Notably, the relationship between EI and transformational leadership was significantly weaker for the MSCEIT (Mayer et al., 2002) than for the other measures.

Harms and Crede (2010) conducted further meta-analyses for studies which had used same-source ratings and multi-source ratings. Both abilities and mixed measures of EI demonstrated lower validity estimates when multi-source ratings were used. Mixed measures of EI demonstrated a strong relationship between EI and transformational leadership when same-source ratings were used (k = 38, N = 4424, ρ = .66), and a weak relationship when multi-source ratings were used (k = 20, N = 2491, ρ = .11). Abilities measures of EI demonstrated lower validity estimates than mixed measures when same-source ratings were used (k = 10, N = 1066, ρ = .24) and had no relationship with transformational leadership when multi-source ratings were used (k = 4, N = 441, ρ = .05).

The authors also conducted meta-analyses of the studies which assessed the relationship between EI and the MLQ (Avolio et al., 1995). The relationship was moderate in strength for same-source ratings (k = 33, N = 3999, ρ = .54), and weak for multi-source ratings (k = 14, N = 1549, ρ = .09). Regarding transactional leadership, Harms and Crede (2010) reported that EI had a positive relationship with the contingent reward scale for same-source ratings (k = 12, N = 1272, ρ = .35) and a weak relationship for multi-source ratings (k = 6, N = 622, ρ = .13). There was no significant relationship between EI and the management-by-exception active scale. Both scales of passive/avoidant leadership were negatively related to EI. Notably, EI demonstrated a moderate negative relationship with the management-by-exception passive scale for same-source ratings (k = 10, N = 871, ρ = –.22) and a weak relationship for multi-source ratings (k = 3, N = 333, ρ = −.12). EI also demonstrated a moderately strong negative relationship with the laissez-faire scale of passive/avoidant leadership for same-source ratings (k = 14, N = 1304, ρ = −.36) and a weak relationship for multi-source ratings (k = 8, N = 617, ρ = −.17). Above all, this meta-analysis demonstrates the importance of using multi-source ratings of leadership behaviours, rather than same-source ratings, in order to avoid an overestimation of positive leadership behaviours caused by common method variance and self-serving bias.

Critique of previous research

Many previous studies which have examined the relationship between EI, leadership style and leadership outcomes have been lacking in methodological rigor. The choice of instruments used to operationalise EI has often been far from ideal. Some researchers (e.g. Mandell, 2003; Sivanathan and Fekken, 2002; Stone et al., 2003) have used instruments measuring mixed models of EI rather than the conceptually superior abilities model. Srivsastava and Bharamanaikar (2004) used a relatively unknown EI instrument, and Coetzee and Schaap (2005) used the outdated and outmoded Multifactor Emotional Intelligence Scale (Mayer et al., 1999). Also, there is a paucity of studies which have examined the relationship between EI and leadership styles using the Mayer and Salovey (1997) abilities model in educational settings.

Most studies (e.g. Coetzee and Schaap, 2005; Kerr et al., 2006; Leban, 2003; Stone et al., 2005; Weinberger, 2009) have not controlled for established predictors of leadership behaviours and leader effectiveness, such as GMA and personality factors. Hence, the amount of variance in leadership behaviours and leader effectiveness thought to be accounted for by EI may have been overestimated in some of these studies. Also, as EI has rarely been compared with other predictors in these studies, questions related to the divergent and incremental validity of EI remain unanswered.

Several studies have used different approximations of leader effectiveness (e.g. Kerr et al., 2006; Rosete and Ciarrochi, 2005). However, the construct validity of these approximations is questionable as it is not possible to tell whether or not they are really measuring leader effectiveness.

Schulte’s (2003) study demonstrates construct rigor with the use of valid and reliable instruments, including the MSCEIT (Mayer et al., 2002) and the MLQ (Avolio et al., 1995), to examine the relationship between EI, personality factors, GMA and leadership styles. Unfortunately, Schulte (2003) used a student participant sample instead of real leaders, yet the MLQ (Avolio et al., 1995) was designed to test real, practising leaders. Furthermore, the MLQ (Avolio et al., 1995) is designed to measure leadership style and perceived leadership outcomes by assessing the scores from multiple ratings obtained from different organisational levels. However, Schulte (2003) only used self-ratings of leadership.

Weinberger (2009) also used the MSCEIT (Mayer et al., 2002) and the MLQ (Avolio et al., 1995) instruments to explore the relationship between EI, leadership style and leadership effectiveness. However, Weinberger (2009) only obtained ratings of leadership style and leadership effectiveness from one level of the organisation. As Weinberger’s (2009) raters were all followers, the perceptions of peers, supervisors and the leaders themselves were not taken into account. Multiple ratings of leadership behaviours are considered to be more valid than self-ratings (Landy and Conte, 1996) but there is a shortage of research in the field which has obtained multiple ratings from different organisational levels, particularly in studies which have used the MSCEIT (Mayer et al., 2002) to operationalise EI. As almost all previous studies have only obtained self-ratings of leadership behaviours, rather than multiple independent ratings, self-serving bias is likely to have resulted in an overestimation of positive leadership behaviours and an inflation of the association between EI and self-rated leadership behaviours due to common method variance.

These weaknesses have limited the validity of the findings of previous studies. Therefore, there is scope for further research in educational institutions that applies a high level of methodological rigor to address the question: to what extent is the Mayer and Salovey (1997) model of EI a useful predictor of leadership style and perceived leadership outcomes? This overarching question may be broken down into a series of questions such as: is the Mayer and Salovey (1997) model of EI related to leadership style and leadership outcomes? Does the Mayer and Salovey (1997) model of EI have divergent validity from GMA and personality factors? Is the Mayer and Salovey (1997) model of EI able to predict leadership style and leadership outcomes when multiple ratings of leadership behaviours from different organisational levels are obtained? Does the Mayer and Salovey (1997) model of EI have incremental validity above other predictors of leadership style and leadership outcomes? These questions were further sub-divided and became the research questions and hypotheses examined in this study.

Method

The study replicated previous research in the field by examining the relationship between EI, leadership style and perceived leadership outcomes, and by assessing whether EI has discriminant validity from established predictors of job performance, and incremental validity above these constructs. Predictors were selected from individual difference variables commonly used to predict leadership behaviours and leader effectiveness (GMA and personality factors). The project advanced research in the field by obtaining multiple ratings of leadership behaviours from different organisational levels and by including integrity as an additional potential predictor. The impact of role and gender on leadership style and EI were also examined. The independent variables and dependent variables are presented in Table 3.

Table 3.

Independent variables and dependent variables.

Independent variables	Dependent variables^a
Total EI Experiential EI Strategic EI Perceiving emotions Using emotions Understanding emotions Managing emotions General mental ability Neuroticism Extraversion Openness Conscientiousness Agreeableness Integrity	Transformational leadership Contingent reward (transactional leadership) Management-by-exception active (transactional leadership) Passive/avoidant leadership Satisfaction (with leader) Extra effort (by followers) Effectiveness (of individual/group)

^aSelf-rated and rated by one follower, one peer and one supervisor per leader.

Each variable was operationalised using empirically validated instruments: leadership styles and perceived leadership outcomes (MLQ, Avolio et al., 1995), EI (MSCEIT, Mayer et al., 2002), personality factors (The Big Five Inventory, John et al., 1991), GMA (Wonderlic Personnel Test-Quicktest, Wonderlic, 2003) and integrity (Integrity Express, Vangent, 2002). The instruments were used in their standard online off-the-shelf form as they would be by human resource practitioners in the workplace. The project employed quantitative methods and a cross-sectional design. A pilot study preceded the main study. Table 4 provides a summary of the research questions and hypotheses.

Table 4.

Summary of research questions and hypotheses.

Research questions and hypotheses
Hypothesis 1. Total EI will have discriminant validity from personality factors.
Hypothesis 2. Total EI will have discriminant validity from GMA.
Hypothesis 3. Total EI scores will be significantly higher for females than for males.
Hypothesis 4. Transformational leadership scores will be significantly higher for females than for males.
Hypothesis 5. Scores for the contingent reward scale of transactional leadership will be significantly higher for females than for males.
Hypothesis 6. Scores for the management-by-exception active scale of transactional leadership will be significantly higher for males than for females.
Hypothesis 7. Passive/avoidant leadership scores will be significantly higher for males than for females.
Research question 1. Investigate whether scores of transformational leadership vary according to the role of the leader.
Research question 2. Investigate whether scores of total EI vary according to the role of the leader.
Research question 3. Investigate whether integrity has discriminant validity from personality factors.
Research question 4. Investigate whether EI predicts transformational leadership.
Research question 5. Investigate whether EI has incremental validity above GMA in predicting transformational leadership.
Research question 6. Investigate whether EI has incremental validity above personality factors in predicting transformational leadership.
Research question 7. Investigate whether EI has incremental validity above integrity in predicting transformational leadership.
Research question 8. Investigate whether EI predicts satisfaction (of followers).
Research question 9. Investigate whether EI predicts effectiveness (of leader/group).
Research question 10. Investigate whether EI predicts extra effort (by followers).
Research question 11. Investigate whether EI predicts the contingent reward scale of transactional leadership.
Research question 12. Investigate whether EI predicts the management-by-exception active scale of transactional leadership.
Research question 13. Investigate whether EI predicts passive/avoidant leadership.

In order to obtain participants for the study, project information was emailed to the Australian Council of Educational Leaders and then forwarded to its members. The final sample consisted of 144 educational leaders (52 male and 92 female) and 432 nominated peers, followers and supervisors. More specifically, the sample consisted of 66 principals, 51 vice-principals and 10 heads of departments employed in Australian schools. Twelve course and unit coordinators employed at an Australian university and five Department of Education and Training administrators from Queensland also participated. Each leader nominated one supervisor, one peer and one follower to rate them on their leadership behaviours. Participants completed the tests online.

Data entry and analysis was conducted using SPSS version 15.0 (SPSS, 2007). The data was coded and screened. Then, correlation, multiple regression and difference between the means procedures were undertaken to enable the statistical relationships between the variables to be analysed. In order to assess whether or not levels of interrater agreement for each of the MLQ (Avolio et al., 1995) scales were sufficient to justify the aggregation of ratings r_wg(j) (James et al., 1984) estimates were calculated. Aggregate scores from each of the four raters (self, peer, follower and supervisor) were calculated for each item from each MLQ (Avolio et al., 1995) scale. In accordance with the recommendations of James et al. (1984), negative values and values exceeding 1.00 were considered to be the result of sampling error and were reset to .00 to indicate a complete lack of agreement. Using the standards for interpreting interrater agreement estimates suggested by LeBreton and Senter (2008), the mean r_wg(j) estimates for each of the MLQ (Avolio et al., 1995) scales, except management-by-exception active, demonstrated strong (.71 to .90) or very strong (.91 to 1.00) levels of interrater agreement. The mean r_wg(j) estimate for management-by-exception active was .56 which is lower than the .70 cut-off often used to justify the aggregation of ratings (LeBreton and Senter, 2008). Therefore, data for the management-by-exception active scale was reported for each of the four rating levels for this scale. The data for all of the other MLQ (Avolio et al., 1995) scales was reported as the aggregate of ratings across the four rating levels as it resulted in a more parsimonious outcome. Tabachnick and Fidell (2007) state that the sample size for regression should be at least N > 50 + 8 m (where m is the number of independent variables). The regression models undertaken in the study used 11 independent variables. Therefore, a sample size of 138 (50+8[11] = 138) was required to test 11 predictors. The sample size for this study exceeded this requirement.

Results

The descriptive statistics for the study variables are presented in Table 5. The intercorrelations between the study variables are presented in Table 6.

Table 5.

Descriptive statistics for study variables.

Variable	Mean	Standard deviation	Minimum	Maximum	Range	Cronbach’s alpha	r_wg(j)	Sample size
Age	46.93	7.14	26.00	62.00	36.00			144
Total EI	101.61	11.16	69.03	130.04	61.01			144
Experiential EI	101.55	14.43	65.54	130.32	64.78			144
Strategic EI	100.49	7.79	78.81	116.40	37.59			144
Perceiving emotions	101.86	15.62	61.44	135.43	73.99			144
Understanding emotions	99.43	9.77	74.25	118.35	44.10			144
Managing emotions	100.30	7.81	72.96	115.44	42.48			144
Using emotions	101.25	12.61	69.49	128.49	59.00			144
Openness	4.08	.53	2.70	5.00	2.30	.76		144
Neuroticism	2.35	.80	1.00	4.63	3.63	.85		144
Extraversion	3.85	.74	2.25	5.00	2.75	.84		144
Conscientiousness	4.47	.52	2.56	5.00	2.44	.83		144
Agreeableness	4.26	.58	2.56	5.00	2.44	.79		144
GMA	25.81	3.76	13.00	33.00	20.00			144
Integrity	12.57	2.16	6.00	17.00	11.00			144
Transformational leadership^a	3.27	.27	2.34	3.80	1.46	.90	.98	576
Idealised attributes^b	3.28	.30	2.44	3.88	1.44	.61	.87	576
Idealised behaviours^b	3.39	.30	2.40	3.94	1.54	.68	.91	576
Inspirational motivation^b	3.39	.36	2.31	4.00	1.69	.83	.92	576
Intellectual stimulation^b	3.07	.36	1.81	2.06	3.88	.81	.88	576
Individualised consideration^b	3.21	.32	1.88	2.00	3.88	.53	.82	576
Contingent reward^b	3.13	.36	1.96	3.88	1.92	.67	.82	576
Management-by-exception active self-ratings	1.44	.76	.00	3.25	3.25	.71		144
Management-by-exception active supervisor ratings	1.86	.93	.00	4.00	4.00	.68		144
Management-by-exception active peer ratings	1.79	.96	.00	4.00	4.00	.75		144
Management-by-exception active follower ratings	1.56	.91	.00	3.75	3.75	.63		144
Passive/avoidant leadership^c	.53	.24	.09	1.34	1.25	.63	.92	576
Management-by-exception passive^b	.67	.32	.06	1.69	1.63	.55	.81	576
Laissez-faire^b	.38	.25	.00	1.31	1.31	.55	.87	576
Satisfaction (of followers) ^b	3.50	.32	2.38	4.00	1.63	.79	.89	576
Effectiveness (of individual/group) ^b	3.43	.28	2.69	3.92	1.23	.64	.91	576
Extra effort (of followers) ^b	2.85	.39	1.83	3.92	2.08	.59	.73	576

^aAll four rating levels and the five transformational scales combined. ^bAll four rating levels combined. ^cAll four rating levels and the two passive/avoidant scales combined.

Table 6.

Intercorrelations between predictor variables, transformational leadership, and perceived leadership outcomes.

	1.	2.	3.	4.	5.	6.	7.	8.	9.	10.	11.	12.	13.	14.	15.	16.	17.
1. Total EI
2. Experiential EI	.91**
3. Strategic EI	.68**	.33**
4. Perceiving emotions	.82**	.91**	.17**
5. Understanding emotions	.51**	.26**	.75**	.23**
6. Managing emotions	.48**	.24**	.67**	.17*	.03
7. Using emotions	.72**	.75**	.33**	.44**	.21*	.30**
8. Openness	.13	.12	.09	.03	.04	.19*	.11
9. Neuroticism	−.07	−.10	−.05	−.11	.00	−.06	−.06	.01
10. Extraversion	.06	.05	.07	.02	−.07	.17*	.04	.17*	−.27**
11. Conscientiousness	.05	.05	.01	.04	−.07	.09	.07	.17*	−.17*	.22**
12. Agreeableness	.07	.03	.11	.00	.00	.13	.09	.23**	−.46**	.32**	.27**
13. GMA	.26**	.19*	.26**	.22**	.25**	.11	.07	−.22**	.07	−.15	−.05	−.14
14. Integrity	.13	.09	.14	.05	.03	.19*	.11	.18*	.06	.01	.22**	.12	.00
15. Transformational^a	.06	.05	.08	.03	.07	.02	.02	.34**	−.29**	.14	.09	.30**	−.26**	.16
16. Satisfaction ^b	.00	−.04	.09	−.04	.10	.01	.00	.18**	−.22**	.00	.05	.30**	−.11	.02	.72**
17. Effectiveness ^b	−.05	−.06	−.02	−.09	.05	−.08	.03	.24**	−.19	.05	.20*	.16	−.16	.15	.72**	.70**
18. Extra effort ^b	.04	.06	−.02	.09	.04	−.04	.03	.27**	−.15	.04	.14	.12	−.18*	.15	.53**	.38**	.55**

^aAll four rating levels and the five transformational scales combined. ^bAll four rating levels combined.

Significant correlations found between each of the study variables are highlighted below. As expected, there was a significant relationship between total EI and the strategic (.68, p = < .01) and experiential (.91, p = < .01) areas, and the managing emotions (.48, p = <.01), understanding emotions (.51, p = <.01), using emotions (.72, p = < .01) and perceiving emotions (.82, p = < .01) branches. The strategic area of EI was also significantly related to the experiential area (.33, p = < .01), and the managing emotions (.67, p = <.01), understanding emotions (.75, p = <.01), using emotions (.33, p = < .01) and perceiving emotions (.27, p = < .01) branches. Further significant correlations were found between the experiential area of EI and the managing emotions (.24, p = <.01), understanding emotions (.26, p = <.01), using emotions (.75, p = < .01) and perceiving emotions (.91, p = < .01) branches. Regarding the relationships between the EI branches themselves, managing emotions shared significant correlations with using emotions (.30, p = < .01) and perceiving emotions (.17, p = < .05). Understanding emotions shared further correlations with using emotions (.21, p = < .05) and perceiving emotions (.23, p = < .01). Lastly, using emotions also shared a significant correlation with perceiving emotions (.44, p = < .01).

As expected, a strong positive relationship was found between transformational leadership and the three perceived leadership outcomes variables: satisfaction (.72, p = < .001), effectiveness (.72, p = < .01) and extra effort (.53, p = < .01). A moderately sized positive correlation was found between transformational leadership and the contingent reward scale of transactional leadership (.68, p = < .01), whilst a negative correlation was found between transformational leadership and passive/avoidant leadership (−.41, p = < .01). These findings help to confirm the validity of the MLQ (Avolio et al., 1995) regarding the relationship between transformational leadership style and perceived leadership outcomes. The findings also confirm the divergent validity of the three leadership styles measured by the MLQ (Avolio et al., 1995).

The results of the correlation analysis indicated that four correlations between transformational leadership and the predictor variables were statistically significant. Moderately sized significant correlations were found between transformational leadership and openness (.34, p = < .001), and transformational leadership and agreeableness (.30, p = < .001). Small negative correlations were found between transformational leadership and neuroticism (−.29, p = < .001), and transformational leadership and GMA (−.26, p = < .01).

The contingent reward scale of transactional leadership shared a moderately sized correlation with openness (.30, p = < .001). Self-ratings of the management-by-exception active scale of transactional leadership shared small correlations with neuroticism (.22, p = < .01) and agreeableness (.18, p = < .05), and small negative correlations with the strategic area of EI (−.18, p = < .05) and the managing emotions branch of EI (−.20, p = < .05). Correlations between EI and the other leadership style and perceived leadership outcomes variables were not significant. Passive-avoidant leadership shared a small negative correlation with conscientiousness (−.25, p = < .01).

Satisfaction shared a small correlation with openness (.18, p = < .05), a moderate correlation with agreeableness (.30, p = < .001) and a small negative correlation with neuroticism (−.22, p = < .01). Effectiveness shared small correlations with openness (.24, p = < .01) and conscientiousness (.20, p = < .05), and a small negative correlation with neuroticism (−.19, p = < .05). Finally, extra effort shared a small correlation with openness (.27, p = < .01) and a small negative correlation with GMA (−.18, p = < .05).

The study tested the divergent validity of the selected instruments by addressing the question: does the Mayer and Salovey (1997) model of EI have divergent validity from GMA and personality factors? Only small significant correlations existed between GMA and the understanding emotions (.25, p = < .01) and perceiving emotions (.22, p = < .01) branches of EI. Hence, the MSCEIT (Mayer et al., 2002) did demonstrate divergent validity from GMA and personality factors. A small significant correlation was found between integrity and the managing emotions branch of EI (.19, p = < .05). Integrity demonstrated divergent validity from the five factors of personality as it only shared small significant correlations with openness (.18, p = < .05) and conscientiousness (.22, p = < .01).

Regarding difference between the means procedures, the effect size, or proportion of variance in the dependent variable explained by the independent variable, was determined by Eta squared using the guidelines proposed by Cohen (1988): .01 = small effect, .06 = moderate effect, .14 = large effect. The impact of role and gender on leadership style and EI were examined. Results indicated that levels of transformational leadership F(4, 139) = .98, p = > .05 and EI F(4, 139) = .99, p = > .05 did not vary according to role for the five groups (principal, vice-principal, tertiary coordinator, head of department and administrator). Hence, those whose dominant leadership style is transformational are not necessarily fulfilling higher level leadership roles in Australian educational institutions.

Regarding gender, females were perceived to demonstrate more transformational leadership behaviours than males. The size of the difference in scores was small (Eta squared = .04). Females were also perceived to demonstrate more behaviours measured by the contingent reward scale of transactional leadership (Eta squared = .08) than males, whereas males were perceived to engage in more passive/avoidant leadership behaviours than females (Eta squared = .04). There was no significant difference between males and females for scores of EI t(142) = 1.75, p = .08.

The usefulness of EI as a predictor of leadership style and perceived leadership outcomes was examined to address the question: is the Mayer and Salovey (1997) model of EI able to predict leadership style and leadership outcomes when multiple ratings of leadership behaviours from different organisational levels are obtained? Furthermore, EI was assessed to determine whether or not it was a better predictor of transformational leadership and perceived leadership outcomes than other predictors by addressing the question: does the Mayer and Salovey (1997) model of EI have incremental validity above other predictors of leadership style and leadership outcomes? The two EI areas (experiential EI and strategic EI) and total EI were deleted from the multiple regression analysis to solve the problem of singularity as they had very high correlations with the branches of EI. The multiple regression analysis was then rerun without the deleted variables. All variance inflation factor (VIF) values were now below 10, indicating that the problem with multicollinearity had been resolved by deleting the variables. The results of multivariate analysis revealed that only a small amount of variance in transformational leadership (26.5%) could be accounted for by the predictors in regression Model 1. Openness (5.57%) and emotional stability (the inverse of neuroticism) (4.45%) were the most important predictors, followed by GMA (inversely) (3.17%). None of the EI variables were found to predict transformational leadership or the transactional leadership scales (contingent reward and management-by-exception active).

In regression Models 2–4, none of the EI variables were able to predict any of the perceived leadership outcomes variables (satisfaction, effectiveness and extra effort). Agreeableness contributed 4.24% to the variance of satisfaction (of followers). Openness contributed 3.31% to the variance of effectiveness (of individual/group) and neuroticism (inversely) contributed 3.24%. Openness also contributed 4.71% to the variance of extra effort (of followers).

In regression Model 5, none of the EI variables predicted the contingent reward scale of transactional leadership. Openness (6.20%) was found to predict the contingent reward scale of transactional leadership. Finally, in regression Model 6, the understanding emotions branch of EI was found to predict passive/avoidant leadership inversely (4.04%). Conscientiousness also predicted passive/avoidant leadership inversely (5.62%). A summary of the regression models is presented in Table 7.

Table 7.

Summary of regression models.

Model	1			2			3			4			5			6
Variable	B	SE B	β	B	SE B	β	B	SE B	β	B	SE B	β	B	SE B	β	B	SE B	β
Perceiving emotions	.00	.00	.02	.00	.00	−.05	.00	.00	−.13	.00	.00	.12	.00	.00	−.08	.00	.00	.13
Understanding emotions	.00	.00	.10	.00	.00	.13	.00	.00	.09	.00	.00	.06	.01	.00	.14	−.01	.00	−.22*
Managing emotions	.00	.00	−.04	.00	.00	.01	−.01	.00	−.13	.00	.00	−.08	.00	.00	−.01	.00	.00	−.06
Using emotions	.00	.00	.01	.00	.00	−.05	.00	.00	.04	.00	.00	−.07	.00	.00	.05	.00	.00	−.01
Agreeableness	.05	.04	.11	.14	.05	.25*	−.01	.05	−.20	−.02	.07	−.04	−.04	.06	−.06	.01	.04	.03
Neuroticism	−.08	.03	−.25**	−.06	.04	−.14	−.07	.03	−.21*	−.08	.05	−.17	−.04	.04	.08	.02	.03	.08
Openness	.13	.04	.26**	.09	.05	.15	.11	.05	.20*	.18	.07	.24**	.19	.06	.28**	−.01	.04	−.02
Extraversion	.00	.03	−.01	−.06	.04	−.14	−.02	.03	−.05	−.03	.05	−.06	−.04	.04	−.08	.01	.03	.04
Conscientiousness	−.03	.04	−.05	−.01	.05	−.20	.07	.05	.13	.05	.07	.07	.03	.06	.05	−.12	.04	−.26*
Integrity	.02	.01	.13	.00	.01	−.02	.02	.01	.12	.02	.02	.12	.01	.01	.07	.00	.01	.00
GMA	−.01	.01	−.20*	−.01	.01	−.07	−.01	.01	−.09	−.02	.01	−.15	−.01	.01	−0.13	.01	.01	.17

Note. R² for Model 1 Transformational Leadership = .27. R² for Model 2 Satisfaction (of followers) = .16. R² for Model 3 Effectiveness (of Individual/Group) = .17. R² for Model 4 Extra Effort (of Followers) = .15. R² for Model 5 Contingent Reward Scale of Transactional Leadership = .14. R² for Model 6 Passive/Avoidant = .14. *p < .05. **p < .01.

A summary of the research questions, hypotheses, findings and statistical analyses undertaken in the study is presented in Table 8.

Table 8.

Summary of research questions, hypotheses findings and statistical analyses undertaken.

Research questions and hypotheses	Findings^a	Statistical analysis undertaken
Hypothesis 1. Total EI will have discriminant validity from personality factors.	Total EI demonstrated discriminant validity from personality factors.	Correlation
Hypothesis 2. Total EI will have discriminant validity from GMA.	Total EI demonstrated discriminant validity from GMA.	Correlation
Hypothesis 3. Total EI scores will be significantly higher for females than for males.	Total EI scores were not significantly different for males and females.	Independent samples t-test
Hypothesis 4. Transformational leadership scores will be significantly higher for females than for males.	Transformational leadership scores were significantly higher for females than for males.	Independent samples t-test
Hypothesis 5. Scores for the contingent reward scale of transactional leadership will be significantly higher for females than for males.	Contingent reward scores were significantly higher for females than for males.	Independent samples t-test
Hypothesis 6. Scores for the management-by-exception active scale of transactional leadership will be significantly higher for males than for females.	Management-by-exception active scores were not significantly higher for males than for females.	Independent samples t-test
Hypothesis 7. Passive/avoidant leadership scores will be significantly higher for males than for females.	Passive/avoidant leadership scores were significantly higher for males than for females.	Independent samples t-test
Research question 1. Investigate whether scores of transformational leadership vary according to the role of the leader.	No significant difference for scores of transformational leadership according to the role of the leader.	One-way between groups ANOVA
Research question 2. Investigate whether scores of total EI vary according to the role of the leader.	No significant difference for scores of Total EI according to the role of the leader.	One-way between groups ANOVA
Research question 3. Investigate whether integrity has discriminant validity from personality factors.	Integrity demonstrated discriminant validity from personality factors.	Correlation
Research question 4. Investigate whether EI predicts transformational leadership.	None of the EI branches predicted transformational leadership. Openness, neuroticism (inversely) and GMA (inversely) were significant predictors.	Multiple regression
Research question 5. Investigate whether EI has incremental validity above GMA in predicting transformational leadership.	EI did not have incremental validity above GMA. None of the EI branches predicted transformational leadership whereas GMA was an inverse predictor.	Multiple regression
Research question 6. Investigate whether EI has incremental validity above personality factors in predicting transformational leadership.	EI did not have incremental validity above personality factors. None of the EI branches predicted transformational leadership whereas neuroticism and openness did.	Multiple regression
Research question 7. Investigate whether EI has incremental validity above integrity in predicting transformational leadership.	None of the EI branches or integrity predicted transformational leadership.	Multiple regression
Research question 8. Investigate whether EI predicts satisfaction (of followers).	None of the EI branches predicted satisfaction (of followers). Agreeableness was a significant predictor.	Multiple regression
Research question 9. Investigate whether EI predicts effectiveness (of leader/group).	None of the EI branches predicted effectiveness (of individual/group). Openness and neuroticism (inversely) were significant predictors.	Multiple regression
Research question 10. Investigate whether EI predicts extra effort (by followers).	None of the EI branches predicted extra effort (of followers). Openness was a significant predictor.	Multiple regression
Research question 11. Investigate whether EI predicts the contingent reward scale of transactional leadership.	None of the branches of EI predicted the contingent reward scale of transactional leadership. Openness was a significant predictor.	Multiple regression
Research question 12. Investigate whether EI predicts the management-by-exception active scale of transactional leadership.	None of the branches of EI predicted the management-by-exception active scale of transactional leadership.	Multiple regression
Research question 13. Investigate whether EI predicts passive/avoidant leadership.	The understanding emotions branch of EI predicted passive/avoidant leadership whereas the other EI branches did not.	Multiple regression

^a N = 144 leaders and 432 other raters of leadership behaviours.

Discussion

The study assessed the divergent validity of the MSCEIT (Mayer et al., 2002) and examined the usefulness of EI as a predictor of leadership style and perceived leadership outcomes. The impact of role and gender were also assessed. Although the MSCEIT (Mayer et al., 2002) demonstrated divergent validity from GMA and personality factors questions remain regarding what exactly the MSCEIT (Mayer et al., 2002) is measuring. However, this finding contributes to a body of work upon which the boundaries of the EI construct will ultimately be established.

As no significant differences in levels of transformational leadership were found according to the role of the leaders (principal, vice-principal, tertiary coordinator, head of department and administrator) it is possible that EI may not be an important factor in determining the position of leaders within the leadership hierarchy in Australian schools. Alternatively, the abilities of those who scored more highly on EI may not be recognised by the current frameworks for career progression.

Regarding gender, females were perceived to demonstrate more transformational behaviours than males. This replicates the findings of Bass and Avolio (1994), Bass et al. (1996) and Eagly et al. (2003), who reported that females were more transformational than males. Furthermore, females were perceived to demonstrate more contingent reward behaviours and less passive/avoidant behaviours than males in this study, which also replicates the findings of Eagly et al. (2003), and provides further support for the selection of females for leadership roles.

No significant difference was found between males and females for scores of EI. This is contrary to the findings of Mandell (2003), Mayer and Geher (1996), Mayer et al., (1999) and Mayer et al. (2004), all of whom found that females scored significantly higher than males on tests of EI. As the mean EI score for males (M = 99.47) in this study was lower than the mean score for females (M = 102.82), and lower than the mean score for males and females in the normative sample (M = 100), this finding cannot be explained by the males in this project having a particularly high level of EI. One explanation may be that differences between the scores for males and females on EI simply failed to reach significance.

One of the most surprising findings was the negative relationship between GMA and transformational leadership. This may have been the result of a restriction of range in the sample. Alternatively, GMA scores above a certain threshold may not be useful for predicting transformational leadership. Another possible explanation is that leaders with higher scores on GMA may focus on behaviours which are outside the boundaries of transformational leadership, such as the behaviours representative of instructional leadership. It is also worth recalling that although Schmidt and Hunter (1998) had reported the usefulness of GMA as a predictor of job performance, leadership and job performance are different constructs and as such they may require different levels of abilities, or different abilities altogether. Therefore, the findings from regression Model 1 support the position of Goldstein et al. (2002) who argued that personnel selection processes overemphasise the importance of GMA. This may well be the case when selecting leaders in Australian educational institutions and elsewhere.

Personality factors did have some impact in the study. Judge et al. (2002) suggested that different personality factors may be important in different vocational settings. As openness and emotional stability (the inverse of neuroticism) were found to predict transformational leadership in this project, they may be especially important for identifying transformational leaders in educational settings.

Regarding the usefulness of EI as a predictor of leadership style and perceived leadership outcomes, none of the EI branches were found to be related to transformational leadership. This finding is contrary to the findings of previous studies by Sivanathan and Fekken (2002), Leban (2003), Coetzee and Schaap (2005) and Srivsastava and Bharamanaikar (2004), all of whom found that EI predicted transformational leadership. However, it is important to note that, apart from the study by Leban (2003), these studies used different EI instruments from the one used in this study. The findings of this study are in line with the findings of Schulte (2003) and Weinberger (2009), who also used the MLQ5X (Avolio et al., 1995) and the MSCEIT (Mayer et al., 2002). Weinberger (2009) found that EI did not predict transformational leadership whilst Schulte (2003) found that EI did not account for additional variance in transformational leadership when GMA and personality were included in the regression. The meta-analysis by Harms and Crede (2010) also reported that when multiple ratings of leadership behaviours were obtained the MSCEIT (Mayer et al., 2002) did not predict transformational leadership style.

None of the EI branches predicted the contingent reward scale of transactional leadership, or the management-by-exception active scale of transactional leadership at any of the four rating levels at which it was assessed (self-ratings, supervisor ratings, peer ratings and follower ratings). This differs from the findings of Harms and Crede (2010) who reported that EI had a positive relationship with the contingent reward scale.

The understanding emotions branch of EI was found to be a negative predictor of passive/avoidant leadership in this study. This was an interesting finding as it is useful to be able to identify leaders who are likely to consistently engage in passive/avoidant leadership behaviours, if only to be aware of their limitations in a leadership role. As the amount of variance predicted in passive/avoidant leadership was small it is unlikely that human resource practitioners would use the MSCEIT (Mayer et al., 2002) for the sole purpose of identifying those who would consistently engage in this style of leadership.

None of the EI variables were able to predict any of the perceived leadership outcomes variables (satisfaction, effectiveness and extra effort) in this study. This underlines the limitations of using the Mayer and Salovey (1997) model of EI in this context when multiple ratings of leadership behaviours have been obtained. This finding is in line with the findings of Weinberger (2009). Although several other studies have found that EI is a useful predictor of leader effectiveness (Mills, 2009), the outcome measures used to assess leader effectiveness have varied which makes it difficult to compare the findings of previous research with the findings from this study.

Overall, EI was not found to be a useful predictor of leadership style and perceived leadership outcomes in this project. Hence, EI is not recommended for use by human resource practitioners as a means of predicting transformational leadership style in Australian educational institutions.

Limitations and recommendations

The project was subject to the usual limitations associated with quantitative methods, cross-sectional designs and psychological testing methods. The findings increase the theoretical understanding of the relationship between EI, leadership style and leadership outcomes, and contribute to a body of literary work assessing the usefulness of EI as a predictor of leadership style and leadership outcomes. Ultimately, the application of the findings may contribute to improved methods of leadership assessment and selection in Australian educational institutions.

Further research which takes a cross-disciplinary approach to educational leadership would be useful taking into account the discourse of effectiveness, efficiency and accountability which has emerged in the field (Christie and Lingard, 2001). As so much of the variance in transformational leadership style remained unexplained in this study, more research into the antecedents of transformational leadership style is clearly required. Although meta-analytic studies of organisational research (e.g. Bono and Judge, 2004) demonstrate some support for the relationship between personality and transformational leadership style, additional empirical research is required to confirm which personality factors are useful predictors in educational settings. More research is also required to clarify whether GMA scores above a certain threshold are useful for predicting transformational leadership style, or whether there is a ceiling beyond which their usefulness diminishes. Additionally, the relationship between overt integrity tests and leadership behaviours warrants further investigation to determine if integrity is a useful predictor of transformational leadership. There is also a need for more research which examines the relationship between leadership style and objective performance outcomes rather than perceived outcomes.

Regarding the future of EI research, questions remain unanswered regarding how EI should be defined and what exactly the MSCEIT (Mayer et al., 2002) is measuring. Although the abilities model is generally regarded as being the most suitable of the current EI models for further research in the field (Antonakis and Dietz, 2010), better assessment tools could be developed as self-reported ability is not the same as actual ability (Van Rooy et al., 2010). Hence, the measurement of EI could evolve to include abilities tests which use objective, correct answers, unlike the consensus and expert scoring methods used by the MSCEIT (Mayer et al., 2002). Van Rooy et al. (2010) propose that these tests could measure the emotions of participants using physiological or other biological means. Alternatively, Roberts, Matthews and Zeidner (2010) suggest that video-based situational judgment tests of emotional regulation and emotional understanding could be included in future objective performance measures. Furthermore, Van Rooy et al. (2010) propose that future EI research should seek to measure the shared variance between different response methods (e.g. self-reports, other reports, and performance items), different scoring methods (e.g. target, consensus, and expert) and different content domains. Roberts et al. (2010) go even further by proposing that a comprehensive EI factor model should be developed requiring a wider sampling of emotional abilities. Hence, there is considerable scope for researchers in the field to participate in the evolution of the EI construct and its measurement tools.

Conclusion

In summary, this research project examined the relationship between EI, leadership style and perceived leadership outcomes in Australian educational institutions. Overall, EI was not found to be a useful predictor of leadership style and leadership outcomes. Consequently, EI is not recommended for use by human resource practitioners as a means of predicting transformational leadership style in Australian educational institutions. The findings of this project go some way to confirming the limitations of using this conceptualisation of EI to predict leadership style and leader effectiveness and answer the main research question: to what extent is the Mayer and Salovey (1997) model of EI a useful predictor of leadership style and perceived leadership outcomes?

Footnotes

Conflicts of Interest

None declared.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

References

Andrews

Soder

(1987) Principal leadership and student achievement. Educational Leadership 44(6): 9–12.

Antonakis

Dietz

(2010) Emotional intelligence: On definitions, neuroscience, and marshmallows. Industrial and Organizational Psychology 3: 165–170.

Antonakis

Ashkanasy

Dasborough

(2009) Does leadership need emotional intelligence? The Leadership Quarterly 20: 247–261.

Ashkanasy

Daus

(2005) Rumors of the death of emotional intelligence in organizational behavior are vastly exaggerated. Journal of Organizational Behavior 26: 441–452.

Avolio

(1999) Full leadership development: Building the vital forces in organizations. Thousand Oaks, CA: Sage.

Avolio

Bass

Jung

(1995) MLQ: Multifactor Leadership Questionnaire: Technical Report. Palo Alto, CA: Mind Garden.

Ayiro

(2009) An analysis of emotional intelligence and the performance of principals in selected schools in Kenya. Advances in Developing Human Resources 11(6): 719–746.

Bar-On

(1997) BarOn Emotional Quotient Inventory: Technical Manual. Toronto, Ontario, Canada: Multi-Health Systems.

Bass

(1985) Leadership and Performance beyond Expectations. New York, NY: The Free Press.

10.

Bass

(1999) Two decades of research and development in transformational leadership. European Journal of Work and Organizational Psychology 8(1): 9–32.

11.

Bass

Avolio

(1994) Shatter the glass ceiling: Women may make better managers. Human Resource Management 33: 549–560.

12.

Bass

Avolio

Atwater

(1996) The transformational and transactional leadership of men and women. Applied Psychology: An International Review 45: 5–34.

13.

Bono

Judge

(2004) Personality and transformational and transactional leadership: A meta analysis. Journal of Applied Psychology 89(5): 901–910.

14.

Burns

(1978) Leadership. New York, NY: Harper and Row.

15.

Chin

(2007) Meta-analysis of transformational school leadership effects on school outcomes in Taiwan and the USA. Asia Pacific Education Review 8(2): 166–177.

16.

Christie

Lingard

(2001) Capturing complexity in educational leadership. Paper presented at the meeting of the American Educational Research Association, April 2001. Seattle, WA.

17.

Coetzee

Schaap

(2005) The relationship between leadership behaviour, outcomes of leadership and emotional intelligence. SA Journal of Industrial Psychology 31(3): 31–38.

18.

Cohen

(1988) Statistical Power Analysis for the Behavioral Sciences, 2nd edn. Hillsdale, NJ: Erlbaum.

19.

Conger

Kanungo

(1987) Toward a behavioral theory of charismatic leadership in organizational settings. Academy of Management Review 12: 637–647.

20.

Daus

Ashkanasy

(2005) The case for the ability-based model of emotional intelligence in organisational behavior. Journal of Organizational Behavior 26: 453–466.

21.

Evans

(1970) The effects of supervisory behavior on the path goal relationship. Organizational Behavior and Human Performance 5: 277–298.

22.

Eagly

Johannesen-Schmidt

van Engen

(2003) Transformational, transactional and laissez-faire leadership styles: A meta-analysis comparing men and women. Psychological Bulletin 129: 569–591.

23.

Fiedler

(1967) A Theory of Leadership Effectiveness. New York, NY: McGraw-Hill.

24.

Gardner

(1983) Frames of Mind. New York, NY: Basic Books.

25.

Goldstein

Zedeck

Goldstein

(2002) g: Is this your final answer? Human Performance 15(1/2): 123–142.

26.

Goleman

(1995) Emotional Intelligence: Why it can Matter More than IQ. New York, NY: Bantam.

27.

Goleman

(2001) An EI-based theory of performance. In: Cherniss

Goleman

(eds) The Emotionally Intelligent Workplace. San Francisco, CA: Jossey-Bass, pp.27–44.

28.

Greenockle

(2010) The new face in leadership: Emotional intelligence. Quest 62: 260–267.

29.

Hallinger

(2005) Instructional leadership and the School Principal: A passing fancy that refuses to fade away. Leadership and Policy in Schools 4(3): 221–239.

30.

Harms

Crede

(2010) Emotional intelligence and transformational and transactional leadership: A meta-analysis. Journal of Leadership & Organizational Studies 17(1): 5–17.

31.

Hersey

Blanchard

(1969) Life cycle theory of leadership. Training and Development Journal 23: 26–34.

32.

James

Demaree

Wolf

(1984) Estimating within-group interrater reliability with and without response bias. Journal of Applied Psychology 69: 85–98.

33.

John

Donahue

Kentle

(1991) The Big Five Inventory-versions 4a and 54. Berkeley, CA: University of California, Berkeley, Institute of Personality and Social Research.

34.

Judge

Piccolo

(2004) Transformational and transactional leadership: A meta-analytic test of their relative validity. Journal of Applied Psychology 89(5): 755–768.

35.

Judge

Bono

Ilies

Gerhardt

(2002) Personality and leadership: A qualitative and quantitative review. Journal of Applied Psychology 87(4): 765–780.

36.

Kerr

Garvin

Heaton

Boyle

(2006) Emotional intelligence and leadership effectiveness. Leadership and Organization Development Journal 27(4): 265–279.

37.

Leban

(2003) The relationship between leader behavior and emotional intelligence of the project manager and the success of complex projects. Dissertation Abstracts International Section A: Humanities & Social Sciences 64(5-A).

38.

Landy

Conte

(2006) Work in the 21st Century: An Introduction to Industrial and Organizational Psychology, 2nd edn. Malden, MA: Blackwell.

39.

LeBreton

Senter

(2008) Answers to 20 questions about interrater reliability and interrater agreement. Organizational Research Methods 11(4): 815–852.

40.

Leithwood

(1994) Leadership for school restructuring. Educational Administration Quarterly 30(4): 498–518.

41.

Leithwood

Jantzi

(2005) Review of transformational school leadership research 1996-2005. Leadership and Policy in Schools 4: 177–199.

42.

Leithwood

Sleegers

(2006) Transformational school leadership: Introduction. School Effectiveness and School Improvement 17(2): 143–144.

43.

Leithwood

Seashore Louis

Anderson

. (2004) How Leadership Influences Student Learning. New York, NY: The Wallace Foundation.

44.

Leithwood

Day

Sammons

. (2006) Successful School Leadership: What is it and How it Influences Pupil Learning. London, England: Department for Education and Skills.

45.

Liontos

(1992) Transformational leadership. ERIC Clearing House on Educational Management 72: Eugene, OR.

46.

Lowe

Kroeck

Sivasubramaniam

(1996) Effectiveness correlates of transformational and transactional leadership: A meta-analytic review of the MLQ literature. Leadership Quarterly 7(3): 385–425.

47.

Mandell

(2003) Relationship between emotional intelligence and transformational leadership style: A gender comparison. Journal of Business and Psychology 17(3): 387–404.

48.

Mann

(1959) A review of the relationship between personality and performance in small groups. Psychological Bulletin 56: 241–270.

49.

Matthews

Zeidner

Roberts

(2002) Emotional Intelligence: Science and Myth. Cambridge, MA: MIT Press.

50.

Mayer

(2001) Emotion, intelligence and emotional intelligence. In: Forgas

J P

(ed) Handbook of Affect and Social Cognition. Mahwah, NJ: Lawrence Erlbaum Associates, pp.410–431.

51.

Mayer

Geher

(1996) Emotional intelligence and the identification of emotion. Intelligence 22: 89–113.

52.

Mayer

Salovey

(1997) What is emotional intelligence? In: Salovey

Sluyter

(eds) Emotional Development and Emotional Intelligence: Implications for Educators. New York, NY: Basic Books, pp.3–31.

53.

Mayer

Caruso

Salovey

(1999) Emotional intelligence meets traditional standards for an intelligence. Intelligence 27: 267–298.

54.

Mayer

Salovey

Caruso

(2000) Models of emotional intelligence. In: Sternberg RJ (ed), Handbook of Intelligence. Cambridge: Cambridge University Press, pp.396–420.

55.

Mayer

Salovey

Caruso

(2002) Mayer-Salovey-Caruso Emotional Intelligence Test: Manual. Toronto, Ontario, Canada: Multi-Health Systems.

56.

Mayer

Salovey

Caruso

(2004) Emotional intelligence: Theory, findings, and implications. Psychological Inquiry 15: 197–215.

57.

McCrae

(2000) Emotional intelligence from the perspective of the five-factor model of personality. In: Bar-On

Parker

JDA

(eds) The Handbook of Emotional Intelligence. San Francisco, CA: Jossey-Bass, pp.263–276.

58.

Mills

(2009) A meta-analysis of the relationship between emotional intelligence and effective leadership. Journal of Curriculum and Instruction 3(2): 22–38.

59.

Podsakoff

MacKenzie

Moorman

. (1990) Transformational leader behaviors and their effects on followers’ trust in leader, satisfaction, and organizational citizenship behaviors. Leadership Quarterly 1(2): 107–142.

60.

Roberts

Matthews

Zeidner

(2010) Emotional intelligence: Muddling through theory and measurement. Industrial and Organizational Psychology 3: 140–144.

61.

Rosete

Ciarrochi

(2005) Emotional intelligence and its relationship to workplace performance outcomes of leadership effectiveness. Leadership & Organization Development 26(5): 388–399.

62.

Rubin

(2003) Paddling upstream in leadership research: Exploring antecedents of transformational leadership. Dissertations Abstract International Section A: Humanities & Social Sciences (UMI No. 3102927).

63.

Salovey

Mayer

(1990) Emotional intelligence. Imagination, Cognition, and Intelligence 9: 185–211.

64.

Schmidt

Hunter

(1998) The validity and utility of selection methods in personnel psychology: Practical and theoretical implications of 85 years of research findings. Psychological Bulletin 124: 262–274.

65.

Schulte

(2003) Emotional intelligence: A predictive or descriptive construct in ascertaining leadership style or a new name for old knowledge? Dissertation Abstracts International Section A: Humanities & Social Sciences 63(10-A).

66.

Sergiovanni

(1984) Leadership and excellence in schooling. Educational Leadership 41(5): 4–13.

67.

Sergiovanni

(1990) Value-Added Leadership. New York, NY: Harcourt Brace Jovanavich.

68.

Sivanathan

Fekken

(2002) Emotional intelligence, moral reasoning and transformational leadership. Leadership & Organization Development Journal 23(4): 198–204.

69.

SPSS (2007) Statistics Package for the Social Sciences Version 15.0.1. Chicago, IL: SPSS Inc.

70.

Srivsastava

KBL

Bharamanaikar

(2004) Emotional intelligence and effective leadership behaviour. Psychological Studies, India 49: 107–113.

71.

Stogdill

(1948) Personal factors associated with leadership: A survey of the literature. Journal of Psychology 25: 35–71.

72.

S tone

Parker

JDA

Wood

(2005) Report on the Ontario principals’ council leadership study. Issues in Emotional Intelligence 1: 3.

73.

Tabachnick

Fidell

(2007) Using Multivariate Statistics, 5th edn. Boston, MA: Allyn and Bacon.

74.

Vangent (2002) Integrity Express. Chicago, IL: Vangent, Inc.

75.

Van Rooy

Viswesvaran

(2004) Emotional intelligence: A meta-analytic investigation of predictive validity and nomological net. Journal of Vocational Behavior 65: 71–95.

76.

Van Rooy

Whitman

Viswesvaran

(2010) Emotional intelligence: Additional questions still unanswered. Industrial and Organizational Psychology 3: 149–153.

77.

Vroom

Yetton

(1973) Leadership and Decision Making. Pittsburgh, PA: University of Pittsburgh Press.

78.

Waters

Marzano

McNulty

(2003) Balanced Leadership: What 30 Years of Research tells us about the Effect of Leadership on Student Achievement. Aurora, CO: Mid-continent Research for Education and Learning.

79.

Watkin

(2000) Developing emotional intelligence. International Journal of Selection and Assessment 8: 89–92.

80.

Weinberger

(2009) Emotional intelligence, leadership style, and perceived leadership effectiveness. Advances in Developing Human Resources 11(6): 747–772.

81.

Wonderlic

(2003) Wonderlic Personnel Quicktest (WPT-Q) User's Guide. Libertyville, IL: EF Wonderlic.

82.

Yukl

(1989) Managerial leadership: A review of theory and research. Journal of Management 25(2): 251–289.

83.

Yukl

(2002) Leadership in Organizations, 5th edn. Upper Saddle River, NJ: Prentice Hall.