THE MEDIATING EFFECT OF LEADERS’ IDEALISED INFLUENCE ON THE RELATIONSHIP BETWEEN LEADERS’ EMOTIONAL INTELLIGENCE AND JOB PERFORMANCE AMONG ACADEMICS IN MALAYSIAN PUBLIC RESEARCH UNIVERSITIES

Academics exert themselves tremendously to advance in higher education institutions, and their leaders’ emotional intelligence and idealised influence are fundamental to responsiveness regarding performance. The effects of transformational leadership on job performance have been established, but a single component, idealised influence, has been neglected. Hence, this study aims to investigate the potential mediating effects of academic leaders’ idealised influence between academic leaders’ emotional intelligence and subordinates’ http://e-journal.uum.edu.my/index.php/ijms INTERNATIONAL JOURNAL OF MANAGEMENT STUDIES


INTRODUCTION
The Ministry of Higher Education has launched the National Strategic Plan for Higher Education for 2010-2020 and the 11th Malaysia Plan Malaysia, Higher Education Blueprint (2015)(2016)(2017)(2018)(2019)(2020)(2021)(2022)(2023)(2024)(2025). These strategic plans focus on teaching and learning quality, expanding access to education and quality improvement, enhancing research and innovation, embracing technology and improving service delivery efficiency at the ministry level. To welcome this new era, Malaysia's higher education and especially public universities should focus on strategies which provide greater sustainability, visibility management and talents towards education 4.0. The Malaysian government provided RM 400 million to empower five research universities in the 2019 budget (Basir, 2018). This was to ensure that each research university could continue to produce high-quality inventions and research results. This is consistent with the government's desire for research universities to produce high-impact research outcomes (Ithnin et al., 2018). The Malaysian government has allocated an additional RM524 million in the 2020 budget to various ministries and public agencies to intensify research and development in the public sector, signifying the importance of Malaysian research universities in research and development.
Academics experience stress when they are required to get funding for research, carry out post-doctoral publication and consulting projects, or obtain industry cooperation (Azman & Kutty, 2016;Ghasemy et al., 2018;Kassim et al., 2018). Thus, the academic staff's role in today's era has become more difficult and varied, leading to a new workaholic phenomenon (Torp et al., 2017). However, research into the problems of higher education in Malaysia remains scarce (Ghasemy et al., 2018;Ibrahim et al., 2018). It is essential to explore not only the issues encountered by leaders, but also higher education and universities (Ghasemy et al., 2018). Furthermore, it is difficult for organisations to move forward in achieving their goals without productive and committed employees (Darus et al., 2016). In March 2020, Universitas 21 (U21) released findings on 50 countries' university rankings by measuring education spending plan/resources, including grants with expected results, for example, publication, research, and employability among universities graduates. In terms of resources, Malaysia was eleventh position in 2016, twelfth in 2017, seventeenth in 2018 and 2019, and fifteenth in 2020. Malaysia's investment levels are higher than other nations, such as, New Zealand and Australia. In the context of university performance, Malaysia has recorded a diminishing performance, at thirty-ninth in 2018, fortysecond in 2019 and forty-fifth in 2020. The situation of Malaysian research universities does not mirror academics' performance, as the yields referenced do not correspond with the rankings. Truth be told, query about on the issues and magnitudes of higher education in Malaysia remains vague (Raman et al., 2020;Ghasemy et al., 2018& Ibrahim et al., 2018. Therefore, as the gap between ranking and outcomes in studies into higher education remains, this study aims to determine methods to improve ranking and performance as a whole, ideally through leadership.

Theoretical Perspectives
The theory of attribution explains to how an individual perceives a circumstance and how his/her understanding influences comprehension on a phenomenon. This prompted an explanation that an individual wishes to know why things occur, particularly unpleasant incidents (Weiner, 2018;Gendolla & Koller, 2001). Taking the subordinates' recognition towards academic leaders' emotional intelligence, for instance, this research epitomised in emotional responses of the leaders towards the subordinates might be reasonably like the evaluation perspectives of subordinates and social intentions of tolerating or dismissing their leaders' emotional abilities. In this manner, this study adapted attribution theory as it embraces and displays that emotions and affective do mediate attributions and practices (Becker et al., 2018). Incidentally, in 1970s attribution theory was applied in leadership model (Mitchell et al., 1977;Pfeffer, 1977). Unfortunately, existing research has been done a little to research attribution theory particularly on academic leader's emotional intelligence towards subordinates' performance. The quintessence of social exchange theory (SET) is the "returns for both revelries in it is distort if only one parties cherish the expected" (Blau, 1964;Rasoolimanesh et al., 2015). For subordinates who perceived positively academic leaders idealised influence will build up a positive thought regarding the leaders' influence and build up the contemplations towards job performance. In which, subordinates perform in a successful way with an inspirational demeanour (Cropanzano & Mitchell, 2005). From the SET perspective, academic leaders will manage their subordinates toward performance with establishment of support and influence. At the point when leaders give strong and caring subordinates feel an ethical commitment to reward their leaders. In such circumstances, subordinates' likewise feel committed to take part in extra-job tasks (Blau, 1964;Homans, 1961). It is assumed that the social exchange works effectively when each exchange procedures is perceived positively by both parties (Coyle-Shapiro & Shore 2007). Along these lines, this study believes that a leader will invigorate idealised influence in them and support subordinates towards performance.

Emotional Intelligence and Job Performance
Emotional intelligence is defined as the ability to process information on the recognition, understanding and management of emotion in oneself and others (Checa & Fernández-Berrocal, 2019). Emotional intelligence leaders may likewise have propelled their subordinates' motivation as a result of their experiences into human instinct and their capacity to influence others' feelings (Miao et al., 2018). Furthermore, emotional intelligence is associated with work performance (O'Boyle et al., 2011). Along these lines, academic leader's emotional intelligence is a profound psychological catalyst with a significant effect on academic ability and performance. Overall, any such attributes, whether positive or negative, will have an effect on the ability or inability of academic leaders to perceive their subordinates' emotions. The theory of attribution is suited to inferring the emotions observed in others (Becker et al., 2018) when the attributes found during observation are more inter-relational than interpersonal (Weiner, 2000). The immaculate performance of subordinates evolved around leadership's ability to develop relationships and supports (Ibrahim & Ahamat, 2019;Rumaropen et al., 2019). As shown by Mohamad, Ismail and Shariff (2020), leaders play a critical role in achieving organisational success. Furthermore, Sakiru et al. (2014) pointed out that the role of leadership is vital in higher learning organisation performance, especially in the leadership role in Malaysian higher education institutions.
Consequently, Johnson (2016) added that leaders with high emotional intelligence will be capable of using their positive emotions to play a major role in organisational success. In particular, Coetzee and Harry (2013) stressed that abilities related to emotional intelligence are vital, as they set a foundation of their profession and serve as the potential for performance, though the relationship between emotional intelligence and academic leaders is indiscernible in the literature (Parrish, 2013). In fact, as mentioned by Anthony and Antony (2017), academic leadership is a relatively new field in leadership. Thus, the following hypothesis is proposed.
H 1 : Leaders' emotional intelligence positively influences job performance among subordinates.

Emotional Intelligence and Idealised Influence
Initially, emotional intelligence was characterised by Salovey and Mayer (1990) as the subset of social intelligence that includes the capacity to monitor one's and others' sentiments and feelings, to segregate among them and to utilise this information to control one's reasoning and activities. As defined by Bass (1985), transformational leadership has four parts: idealised influence, inspiration motivation, intellectual stimulation, and individualized consideration. Dionne et al. (2004) portrayed idealised influence as a leadership with enunciation of vision and spurring employees to work past their personal circumstance to accomplish the shared objectives. The idealised influence is how much a leader is seen as a good example who is sure and can affect representatives (Gomes, 2014). As indicated by Hughes (2014), the idealised influence part of the transformational leadership has the accompanying attributes ingrains pride in subordinates, surpasses his personal circumstance to benefit the organisation, presents a feeling of dependable and certainty and conveys the most significant convictions and qualities. This study focuses on idealised influence behaviour because idealised influence is a prominent factor in influencing and inspiring subordinates towards achieving organisational goals through values, vision and mission (Conger & Kanungo, 1998;Judge & Piccolo, 2004;Koveshnikov & Ehrnrooth, 2018;Lowe et al., 1996;). As Judge and Piccolo (2004) stated, the idealised influence reveals the strongest most valid relationships. As such, leaders who possess certain qualities will gain trust and respect, eventually producing subordinates with high performance, further indicating that they are responsible and responsive to their subordinates (Kim & Kim, 2017). The interactions that arise between leaders and subordinates are marked by the influence of leaders to change the behaviours of their subordinates to be able to feel capable and highly motivated and strive to achieve quality work performance (Zainuddin et al., 2019). Matjie (2018) has expressed that components of transformational leadership and emotional intelligence should be investigated, in fact, there is progressing debate in regard to the association between emotional intelligence and transformational leadership (Kim, 2018). In light of the discoveries of potential connections among the emotional intelligence and idealised influence, the following hypothesis is proposed.
H 2 : Leaders' emotional intelligence is positively associated with leaders' idealised influence.

Idealised Influence and Job Performance
The relationship between idealised influence and job performance can be operationalised using the key theoretical foundation of social exchange theory developed in this study's hypothesis. According to social exchange theory (Cropanzano & Mitchell, 2005), reciprocation refers to the needs in a relationship between two parties. Subordinates who positively perceive academic leaders' idealised influence will develop positive ideas about the leaders' idealised influence and the intention to perform. Subordinates then reciprocate by wanting to perform in an effective manner and with a positive attitude (Cropanzano & Mitchell, 2005). That is, to develop the rationale for formulating this hypothesis the research utilises the theory's key theoretical arguments, such as has been commonly embraced in research (O'Leary, 2016). Abbas, Iqbal, Waheed and Riaz (2012) cited the important link between idealised influence and employee's positive contribution, while Hayati, Charkabi and Naami (2014) found that idealised influence has a significantly positive relationship with employee engagement. Furthermore, idealised influence leaders are trusted to make good decisions, as they have profound emotional connections with and command over their subordinates (Kellner et al., 2018;Puni et al., 2018). In addition, Egheosasa et al. (2018) found that idealised influence behaviours significantly affect job satisfaction among employees. Furthermore, Dalal, Bhave and Biset (2014) mentioned that understanding the antecedents/causes of performance variances among individuals is vital. Finally, Jerobon et al. (2016) found that idealised influence significantly predicts employee performance among employees. As studies on idealised influence towards job performance in terms of subordinates' work performance are limited, the following hypothesis is proposed.
H 3 : Leaders' idealised influence is positively related to subordinates' job performance.

Emotional Intelligence, Idealised Influence and Job Performance
Idealised influence has been proposed as the mediator between emotional intelligence and job performance between leaders and subordinates. Since there have been few studies of idealised influence as a mediation construct between emotional intelligence and job performance, the inference approach can be used. For instance, various studies examining the effects of emotional intelligence and outcomes can be inferred in explaining a mediating relationship (Sood & Kaushal, 2018). Therefore, this current research inferred that emotional intelligence could increase positive emotions and improved performance. With limited studies and comprehension on the mediation effect of idealised influence between emotional intelligence and job performance, there is evidence that leaders' emotional intelligence and idealised influence have an impact on performance. On the contrary, evidence towards job performance is still elusive. With the inferred approach as suggested, the following hypothesis is proposed.
H 4 : Leaders' idealised influence mediates the relationship between leaders' emotional intelligence and job performance among academics.

Participants
The data were collected from five Malaysian public research universities: Universiti Malaya, Universiti Sains Malaysia, Universiti Kebangsaan Malaysia, Universiti Putra Malaysia and Universiti Teknologi Malaysia. Respondents were academicians, including department heads, senior lecturers and lecturers. A sample size of 370 was determined and 386 usable questionnaires were collected out of the 750 questionnaires distributed. A cover letter was provided explaining the objectives of the research and assuring the highest integrity and confidentiality of the data (Mukesh et al., 2013). Quota sampling as a non-probability sampling was used in this research to ensure that the identified group of academics from five public research universities and the populations were adequately represented.

Measurement
The measurements scales for the emotional intelligence, idealised influence and job performance were adapted from various researches and had high in reliability and validity in testing the theories. The emotional intelligence scale was adapted from 16 items from Wong and Law (2002). The present research was interested in examining idealised influence, therefore, Edwards et al.'s (2010) 19 items were adapted. Meanwhile, the job performance measurement scale was adapted from Raman et al. (2020). It had three items that were rated with response categories of above expectation, meet expectation and below expectation.

DATA ANALYSIS
Structural equation modelling (SEM) has become the most popular statistical method to analyse multivariate data. The research intention was to determine the affiliation between construct of emotional intelligence (EI) and job performance (JP), mediated by idealised influence (II) and the emotional intelligence as a construct is formative in nature. Using PLS-SEM is appropriate in this research, as emotional intelligence is a reflective-formative higher-order construct, in which all four dimensions of emotional intelligence are conceptually necessary to form emotional intelligence (Agnihotri et al., 2014). Meanwhile, idealised influence and job performance are reflective constructs, as the items are not dependent on each other. As such, the measurement model testing for reflective indicators was initiated because the research model has reflective and formative indicators. The measurement items related to the constructs were tested by examining the convergent and discriminant validity. In addition, the construct validity was also tested. Secondly, the measurement model for formative indicators was tested using multicollinearity and R 2 statistics. Finally, the structural model was established by analysing all path linkages.

RESULTS
The common method bias in this research reached 30.53 percent of total variance, lower than the recommended cut-off of 50 percent (Podsakoff et al., 2003). As such, common method bias has been overruled in this research. Table 1 indicates the values obtained from the measurement model with all loadings higher than 0.70, which is the threshold proposed by Hair et al. (2014). The average variance extracted (AVE) of all constructs surpasses 0.5 (Bagozzi & Yi, 1988), while the composite reliability score (CR) is higher than 0.7. Thus, we can reason that the convergent validity is achieved. Meanwhile, multicollinearity can be determined through the variance inflation factor (VIF). Hair et al. (2018) stated that a variance inflation factor (VIF) value of five or lower shows no collinearity problem. A high multicollinearity among constructs would create an unstable evaluation and would make it hard to isolate the impact of the individual dimensions on the construct (Petter et al., 2007).   The results of reliability test have shown that all variables are greater than 0.7. Ghozali (2014) recommended that the reliability tests have values of more than 0.7 to meet reliability requirements. The results meet the requirement and considered as reliable. Discriminant validity was performed to establish that all the constructs are distinct to each other. Discriminant validity is a problem when the HTMT value is greater than 0.85 (Kline, 2011) or 0.90 (Gold et al., 2001).

Direct Relationship and Mediation Effect
In Table 4, the path from emotional intelligence to idealised influence is significant (β = 0.85, p < 0.00 Thus, Hypothesis 1 was supported. Emotional intelligence was significantly related to job performance (β = 0.17, p < 0.001), thereby lending support to Hypothesis 2. Idealised influence was not significantly related to job performance (β = -0.01, p > 0.001). Thus, Hypothesis 3 was not supported. The coefficients path estimates of this structural model are shown in Table 4. Next, this research evaluates the effect size (f 2 ) determines the relative contributions or the strength of the exogenous constructs in producing R 2 value for the endogenous construct. Cohen's (1988) guidelines were used to determine the magnitudes of the effect size. The magnitudes effects are 0.02 (small), 0.15 (medium), and 0.35 (large). In this research, job performance is predicted by emotional intelligence and idealised influence. According to Table 4, emotional intelligence has a very small effect in generating the R 2 for job performance (f 2 = 0.00). Meanwhile, the effect size between idealised influence and emotional intelligence effect size is f 2 = 2.75, while the effect size between idealised influence and job performance is very small, that is, at f 2 = 0.00. Notwithstanding the R-square value, Chin et al. (2008) mentioned that the predictive sample reuse method (Q 2 ) may be adequately utilised as a predictive relevance measurement. This research adapted Q 2 procedures suggested by Chin (2010), using cross-validated redundancy. A Q 2 value higher than 0 implies that the model has prescient significance while Q 2 under 0 imply model needs prescient pertinence (Akter et al., 2011). Table 4 indicates the predictive relevance (Q 2 ) of idealised influence and job performance are 0.72 and 0.02, respectively. As such, the value of Q 2 above zero provides evidence that the model has predictive relevance. The results of hypothesis testing in Figure 2 show that only two hypotheses are supported, which are emotional intelligence to idealised influence and emotional intelligence to job performance. Meanwhile, hypothesis idealised influence to job performance is not supported as the result is not significant (β = -0.00, p > 0.001). As for the mediating role of idealised influence in the emotional intelligence-job performance relationship, it is evident from the results that the relationship is not significant (β = -0.00, p > 0.001). This result does not support Hypothesis H 4 .

Figure 1. Second order measurement model.
Meanwhile, the R-square (R 2 ) value interprets the proportion of dependent variable variance that is explained by independent variable. R-square infers the coefficient for determination in criterion constructs. Hair et al. (2017) concluded that an R-square strength of 0.75 is strong, 0.5 is moderate and 0.25 is weak. Figure 1 indicates that the R-square value of emotional intelligence is 0.02. This shows Note: EI= Emotional intelligence, II = Idealised influence, JP = Job performance, OEA=Others' emotional appraisal, ROE= Regulation of emotion, SEL= Self-emotional of immediate superior's, UOE= Use of emotion

DISCUSSION
Partial least square analysis indicates a positive link between the academic leaders' emotional intelligence and the subordinates' job performance, which supports Hypothesis 1 in this research. This highlight pertinent role of attribution theory, as individuals will interpret why people behave in a given manner and understand the attributes of one or more causes for this behaviour (Heider, 1958).
In addition, attribution theory can infer emotions in others (Becker et al., 2018;Sherry, 2018). Thus, the attribution process begins with observing academic leaders' emotional intelligence abilities 11 Note: EI= Emotional intelligence, II = Idealised influence, JP = Job performance, OEA=Others' emotional appraisal, ROE= Regulation of emotion, SEL= Self-emotional of immediate superior's, UOE= Use of emotion

DISCUSSION
Partial least square analysis indicates a positive link between the academic leaders' emotional intelligence and the subordinates' job performance, which supports Hypothesis 1 in this among subordinates and reflecting a positive or negative perception towards their leaders' emotional intelligence. Meanwhile, Hypothesis 2 revealed that the academic leaders' emotional intelligence is positively linked to their idealised influence, in which leaders with emotional intelligence demonstrate high response in understanding their subordinate's emotions and also being a responsible leader with pride and caring. Notwithstanding, this finding is supported with the leaders' universal norms of responsive and responsible leadership (Kim & Kim, 2017). Furthermore, idealised influence behaviour influences and inspires followers towards reaching certain aspirations and directions (Koveshnikov & Ehrnrooth, 2018). Hypothesis 3 proposes that idealised influence is positively correlated to job performance. However, this study indicates that academic leaders' idealised influence does not influence the subordinates' job performance. This is further supported by Koveshnikov and Ehrnrooth, (2018) in the Korean, Russia and Finland contexts, as idealised influence have different effects on followers. Therefore, leaders' idealised influence in Malaysian public research universities does not influence their job performance, even though leaders' idealised influence does predict academics' intentions to perform (Raman et al., 2020). This relationship occurred because the subordinates' perceptions towards their leaders are dominated by the norms of respect and autonomy and undoubtedly, leadership is a social practice framed by social norms values (Dickson et al., 2003).
Finally, statistical evidence shows that Hypothesis 4 is not supported, as idealised influence does not mediate the relationship between leaders' emotional intelligence and subordinates' job performance. In the case of university academics, there is little connection between emotional intelligence and leadership (Zurita-Ortega et al., 2020). Though emotional intelligence and idealised influence investigations are limited, the evidence shows that effective leaders can influence followers' commitment and involvement (Lee et al., 2014;Mitchell et al., 2001). However, emotional intelligence was found to be less related to idealised influence (Kim & Kim, 2017). As such, academic leaders' idealised influence does not encapsulate their own emotional intelligence towards subordinates' job performance.
This study presents some possible implications. From institution/ policymakers' perspectives, the results of this study show that leaders' emotional intelligence predicts subordinates' performance with substantial confirmation on this inference. This is useful for research universities in acknowledging the fact that leaders' greater emotional intelligence might spur in the subordinate's performance. Therefore, greater leaders' emotional intelligence may enable the subordinates to perform better in their job performance (Ingram et al., 2019). Hence, it is suggested that research universities increase awareness of emotional intelligence abilities at events and conferences dedicated to increasing the knowledge and importance of leaders' emotional intelligence. Human resources strategies and profession improvement plans are being implemented in accordance with emotional intelligence environment in various international organisations (Richardson & Norgate, 2015). Therefore, with emotional intelligence abilities, leaders will then have the option to alter their personality attributes and to curb their emotions in raising subordinates job performance. In addition, training is another approach that public research universities can enhance emotional intelligence abilities and idealised influence characteristics of their leaders.

CONCLUSION
This investigation presumes that emotional intelligence is a key predictor of leadership styles, especially in idealised influence and job performance. The outcomes showed a critical and direct connection between academic leaders' emotional intelligence and job performance as perceived by the subordinates. Along these lines, when applied to the leadership setting, this study proposes that leaders' emotional intelligence is an important construct in the new era of 4 th IR, specifically Education 4.0. Human resource department should discuss what the 4 th IR brings to the table over the different elements of the field. The elements are evolving drastically and human resource department should keep pace with these quick changes and progressions.