THE ROLE OF PATERNALISTIC LEADERSHIP IN EMPLOYEE VOICE AND SILENCE MOTIVES: DOES SURFACE SIMILARITY MATTERS?

Purpose of the study: This study compared the effect of paternalistic leadership on employee voice and silence motives based upon gender-based similarity attraction account among employee-leader dyads in Higher Educational Institutions (HEI). The study was conducted in public sector universities of Balochistan, Pakistan. Methodology: The Study used Partial least squares–structural equation modelling, along with advanced methods for multi-group analysis, to assess and compare the proposed relationships between the gender similar and dissimilar dyads. Main findings: The results of this study revealed significant differences between groups for the effect of the authoritative dimension of paternalism on pro-social silence, benevolence dimension of paternalism on quiescence silence and quiescence voice, and morality dimension on opportunistic silence. Application of the study: The current study might help HEI authorities in understanding the effects of paternalistic leadership and diversity management. Novelty/Originality of the study: This study makes a significant theoretical contribution by comparing the effect of paternalistic leadership on voicing motives of employees based on similarity attraction account between gender similar and dissimilar leader-subordinate dyads.


INTRODUCTION
The role of Higher Educational Institutions (HEI) in the development of the country is vital. But these institutions need to be managed differently from business organizations. For this, the leadership of HEIs needs to be developed in a way that can facilitate faculty to meet their targets effectively rather than traditionally controlling them. Although, leadership has been researched from a long time, but most of the researchers are using western developed leadership theories in the eastern context without considering differences among norms embedded in individualistic and collectivistic cultures (Zhang, Tsui & Wang, 2011). So, there is a need to investigate culturally embedded theories to develop more relevant policies. Therefore, the study intended to fill this gap by investigating the influence of paternalistic leadership style on employee intention to speak out that is embedded in cultural norms of the Pakistani context (Aycan, Kanungo, Mendonca, Yu, Deller, Stahl & Kurshid, 2000). The research revealed that employees' voices have a diverse relationship with power concentration or reconcentration in the institutional procedure and process including allocation of resources and determining input related with the main strategies and practices. In this connection, the organizational conducive environment for promoting employees' voice can be dependent upon the leadership wherein paternalistic leadership is widely explored as more effective which have been widely recommended in previous literature suggests that leadership is the main antecedents of employee voice.
Does the question arise that in organizations whether the employees have the right to claim the share in decision-making that might mark the organizational activities under the dynamic leadership? The employees normally face circumstances wherein they decide to express their views or stay silent about work-related important issues (Chen, & Hou 2016). Here, the role of leadership becomes important in considering the views of the employees about the important issues related with the needs at working place which in turn helps in augmenting their behaviour leading to the higher performances (Liu, Zhu & Yang, 2010). In the higher educational context, the employees' voices are considered as a leading issue that remained the focus of interest for researchers during the past decades that denotes the employees' participation in decision-making which may ultimately influence group innovation, job attitudes as well as organizational development. In this regard, the role of leadership has been widely recognized as an influential phenomenon in response to the employees' voices and silence motives. However, there is a dire need to examine that among many leadership styles (Detert & Burris, 2007), which leader trait is more effective whether ethical leadership (Chen & Hou, 2016) or paternalistic leadership style (Chen, 2017). Thus, this study aimed to examine the role of paternalistic leadership style in determining the employees voice and silence motives in the higher education institutional context. withhold information to sustain knowledge advantage or to avoid any resultant workload. Knoll and van Dick (2013) calls it opportunistic silence where an employee intends to achieve some hidden opportunistic objectives. This type of silence is also pervasive in academic culture (Campbell et al., 2002;Hernaus et al., 2019). So, this study followed the classification of silence by Knoll and van Dick (2013) by measuring silence motives along with acquiescent, quiescent, prosocial, and opportunistic silence.

Demographic Match: Surface Similarity
Individuals tend to categorize themselves and others based on some salient attributes such as demographic characteristics (Edwards et al., 2009;Tajfel & Turner, 2004). This process of categorization results in a comparison between "us" versus "them" scenarios. People, based on these comparisons, form their preferences for social interactions (Abrams et al., 1990;Lau & Murnighan, 2005). This theory provides researchers a ground to develop an understanding of human interactions and to predict their behaviours in social settings.

Leader and Employee Voice/Silence Behavior
Research literature highlights that leadership styles of leaders may have a direct influence over employee's response behaviour (Liu et al., 2010). Specially in organizations with traditional hierarchical systems, leaders have the power to plan and assign employee's job-related activities and even have influence over outcomes such as rewards or punishments (Chen and Hou 2016; Morrison, Wheeler-Smith, and Kamdar 2011). Resource dependency theory supports this claim (Emerson 1962;Pfeffer and Salancik 2003;Detert and Burris 2007). According to this theory leaders' control over resources and outcomes give them power. Such leaders consider unsolicited voices from their subordinates either in form of questions or suggestions for improvements as challenging the status quo (Detert & Burris, 2007;Milliken et al., 2003). Due to this, employees often opt to remain silent even when they have something important to share.
Additionally, leaders dominating in the authoritative behaviour prefers to have absolute authority and the control over organizational affairs, and employee's behaviour, may negatively reward subordinates when they do not follow his or her rules (Chan et al., 2013). Such leaders perceive voice as a challenge-oriented behaviour (Maynes & Podsakoff, 2014) and try to induce fear and create an oppressed environment (Wu et al., 2011). Such leaders follow the traditional paradigm and expect subordinates to perform tasks according to given instructions. This may undermine employees' self-efficacy to perform their jobs (Zhang et al., 2011). Further, it may increase the inclination of employees to withhold additional efforts in performing tasks or refuse to exhibit discretionary behaviours that would benefit the organization (Ferris et al., 2009). Therefore, it can be presumed that employees who encounter authoritarian leaders are less likely to speak up.
Thus, it provides the basis for the first hypothesis that:

Hypothesis 1:
A paternalistic leader having a high score on the authoritarian dimension is positively associated with silence behaviour. Furthermore, the benevolence dimension of paternalistic behaviour is related with protection; support, and care provided by paternalistic leaders toward subordinates (Aycan et al., 1999). Social exchange theory provides the justification for the direct influence of benevolent leadership on employee voice behaviour (Emerson 1976). According to this, it can be claimed that a leader's benevolent behaviour supports employees even in personal spheres by enhancing work-life balance and in return, the leader is expected to be reciprocated by the constructive voice from the employee side. Hence it can be proposed that:

Hypothesis 2:
Paternalistic leaders having a high score on the benevolence dimension is positively associated with employee voice. Furthermore, the moral dimension of a paternalistic leader demonstrates high ethical standards (Cheng et al., 2004). Morality in the leader is demonstrated by high self-discipline and unselfish behaviour (Chan, 2014). These leaders demonstrate this behaviour by raising concerns and taking action against unethical acts (van Gils et al., 2015). They serve as a protagonist for their subordinates (Chan, 2014). According to social learning theory (Brown et al., 2005), subordinates learn behaviour from their superiors and try to replicate them. So employees feel encouraged to follow these high moral standards and feel safe to share opinions and suggestions regarding their work, work-related processes, and organizational context (Chen & Hou, 2016). Hence it can be proposed that:

Hypothesis 3:
Moral leadership is positively related to employee voice.

Demographic Match as A Moderator Between Paternalistic Leadership and Employee Voice/ Silence Behavior
Individuals in a social context are attracted toward others based upon some attributes. Social identity theory provides an explanation of how this process occurs. According to this, individuals tend to categorize themselves and others based on attributions such as demographics (Tajfel & Turner, 2004). This creates the categorization of individuals among "us" Individual's inclination toward in-group members develops an environment of trust among in-group members (Whitener et al., 1998). Additionally Whitener et al., (1998). propose that perceived similarity among managers and subordinates may influence trustworthiness among them (Duffy & Ferrier, 2003). This perceived similarity has an impact on communication behaviour and the integration of members within a group (Jackson et al., 1991;O'Reilly et al., 1989;Zenger & Lawrence, 1989). Members of the team who share similar characteristics are expected to communicate more as compared to others (Zenger & Lawrence, 1989). Further, this demographical similarity attraction account influences performance evaluation of subordinates, correspondence by the supervisor, and ambiguity levels regarding task assignments and quality of social exchanges at the workplace (Judge & Ferris, 1993;Liden et al., 1993;Tsui & O'Reilly, 1989).
McAllister (1995) also found that leaders are perceived to be more benevolent when they share more similar characteristics with subordinates. So, based upon the above findings it can be proposed that individuals having high demographic similarity with the leader will feel more trust and favourability. Such subordinates feel less reluctance while communicating their concerns. So, based upon all of this it can be proposed that:

Hypothesis 4:
A paternalistic leader having a high score on the authoritarian dimension will foster more silence in dyadic relations having gender dissimilarity as compared to those having gender similarity.

Hypothesis 5:
Paternalistic leaders having a high score on benevolence dimension will be more positively associated with employee voice and negatively associated with employee silence for the dyadic relations having gender dissimilarity as compared to those having gender similarity.

Hypothesis 6:
Paternalistic leaders having a high score on morality dimension will be more positively associated with employee voice and more negatively associated with employee silence for the dyadic relations having gender dissimilarity as compared to those having gender similarity.

Participants and Procedures
The population of the study was university employees of Balochistan, Pakistan. Data was collected from the targeted sample by using the convenience sampling technique. For determining sample size recent research literature recommends conducting priori test for the expected power of the study (Chuan, 2006;Cohen, 1992;Prajapati et al., 2010). Conventional levels of effect size, alpha value, and power values provided by Cohen (1992) was used to determine the minimum sample size. Based upon these values minimum required sample size for this study was 119. Although this sample size is enough to yield a power of .80.

Demographic match
Similarity-Attraction Account based on Gender: Avery et al., (2008) was followed to collect responses on gender, respondents were asked to mention their own and immediate supervisor's gender. The results have been provided in table 1.

Analysis Strategy
The study used the Statistical Package for the Social Sciences (SPSS) for data management, and preliminary analysis. The main analysis was performed by following the structural equation modelling (SEM) technique by using SMART PLS software. SMART PLS is assumed to be a more preferred technique for complex models (Kline, 2011;Rigdon, 2016).

Respondent Profile
The results show that 65% of respondents were male and 35% were female. The respondents were categorized into five age groups: Less than 25 years (6.8%), 25

Model Assessment
For assessing the models for gender similarity and dissimilarity dyads and comparison, the study used a three-step approach guided by Henseler et al., (2016) and Sarstedt et al., (2011).

Measurement Model Assessment
First, the acceptability of the measurement models for gender similarity and dissimilarity was confirmed (Hair et al. 2014). The measurement model used in this study included constructs: paternalistic leadership style, employee voice, and silence motives. First, factor loadings of each indicator on the respective latent variable (LV) were assessed. For acceptability, it is required to be higher than 0.7 (Hair, Ringle, and Sarstedt 2011). However, values below 0.4 should be considered for the removal of the item from the measurement model, while values in between 0.4 and 0.7 can be considered for removal from the model only if doing the same increases the CR and AVE above the threshold (Chin, 2010;Hair et al., 2011). Factor loadings for both measurement models are presented in Table 1.

Assessment of the Structural Model
After conducting measurement analysis, structural models for gender similarity and gender dissimilarity dyads were assessed. The explanatory power of the model was tested by using R 2 values (Hair et al. 2014) (results are presented in Table 4). In the behavioural research field, the R 2 value of 0.2 is considered acceptable (Hair et al. 2014). Following the same, only the R 2 values of acquiescence silence, quiescence voice for gender similarity and pro-social silence for gender dissimilarity can be considered low. Further, model fit was assessed by using SRMR recommended by Henseler et al., (2016) as an approximate model fit measure and it is required to be less than 0.08. The results revealed SRMR model fit values of 0.073 and 0.078 for gender similarity and gender dissimilarity respectively. Hence met the established criteria.

Multigroup Analysis (MGA)
Measurement invariance of composites (MICOM) for both models was performed as a precondition for conducting MGA (Hair et al. 2014;Henseler, Ringle, and Sarstedt 2016). MICOM is a three-step process including, assessing configural invariance, ensuring compositional invariance, and assessing equality of means and variances (Henseler et al., 2016). Results revealed the partial measurement invariance of both groups (results are presented in Table 5), which is a prerequisite for comparing and interpreting the MGA (Henseler et al., 2016). But failed to establish full measurement invariance, so we were unable to pool the data for analysing basic hypotheses from H1-H3. But the establishment of the partial measurement invariance of both groups is sufficient for analysis of similarity attraction account (Henseler et al., 2016).    (Henseler et al., 2009) and the permutation test (Chin & Dibbern, 2010). To increase the credibility of results, the study had used both recommended methods of MGA to confirm the significance/no significance of the results for identifying differences among gender similar and dissimilar dyads. Results of MGA and structural model assessment are presented in Table 6. Results of both methods shown the significant differences in the effects of authoritarian on pro-social silence, benevolence on quiescence silence, benevolence on quiescence voice, and morality on opportunistic silence for both groups. Further, results alluded to the nonsignificant differences between other path coefficients and relationships.

DISCUSSION & CONCLUSION
The researchers attempted to analyse and compare the effect of paternalism leadership dimensions on employee voice and silence behaviour along with gender similar and dissimilar dyadic relationships among supervisor and subordinate. Results revealed that authoritarian behaviour of leader fosters silence behaviour both in gender similar/dissimilar dyadic. The current study also supports that Authoritarian leadership induces quiescent silence among subordinates, supported by Guo et al., (2018) for both genders similar and dissimilar dyads. Morsch et al., (2020), also reported a significant positive relationship between negative abusive supervision and employee quiescent silence behaviour. Additionally, authoritative behaviour encourages opportunistic silence among subordinates in gender dissimilar dyads. That is a nonproductive behaviour among university staff. This might limit knowledge sharing and the development of new knowledge. Further, results showed that authoritative leader behaviour fosters acquiescence voice that is an ignorant behaviour, and quiescence voice that is a defensive behaviour in gender similar groups. Chen, (2017), also found a positive but insignificant relationship between authoritative leadership and acquiescence voice and quiescence voice.
Moreover, the benevolence of the leader showed a negative significant effect on acquiescent silence among both groups. It reveals that a leader's benevolence toward subordinates will promote individual interest in the current situation and employees might encourage to participate and provoke themselves to found new ways. Further, benevolence also showed a significant negative relation with opportunistic silence implying that leader's benevolence toward subordinates encourages them to share their knowledge and use their expertise to sort out problems and bring new opportunities. Furthermore, the benevolence of a leader has also been proven to be negatively associated with prosocial silence and quiescent voice for gender dissimilar groups. It indicates that leaders' perceived benevolence discourages employees to opt for defensive behaviour that is also supported by Chen (2017). This perceived benevolence may also encourage employees to discuss matters rather than keeping silent for short-term benefits as they expect this will not harm their colleague rather supervisor will handle the issue with personal care. Additionally, the morality of the leader revealed a positive significant effect on the quiescent voice. A quiescent voice is backed by the defensive motive of the employee.
Here context and issues need further investigation as it might be possible that employees expect the fair investigation of sensitive issues and might try to raise voices to protect or to justify their own perspective.
The main contribution of this study is conducting MGA for gender similar dyads and dissimilar dyads. The results revealed significant differences across both groups for the effect of authoritarian on pro-social silence, benevolence on quiescence silence and quiescence voice, and morality on opportunistic silence. Although the relationships were not strongly identified but the results gave a glimpse about the direction of relationships. The strong positive effect of authoritative leadership behaviour on pro-social silence for gender similar groups turned into the insignificant relationship for gender dissimilar group revealing the power of similarity attraction account. Similarly, the weaker positive effect of benevolence on defensive voice for gender similar groups turned into a significantly negative relationship for gender dissimilar groups. It revealed the role of supervisor's benevolence on employee's defensive behaviour that might produce positive effects beyond similarity attraction boundaries

THEORETICAL AND PRACTICAL IMPLICATIONS
The results of the study contribute both practically and theoretically. From the practical side, it will help leaders and policymakers to understand the effect of culturally embedded and least researched leaders' behaviour on employee voice and silence behaviour especially in the presence of gender diversity.
Further, the context of the study is universities that are known to generate knowledge and inputs to policymakers. So, developing an understanding of the interactions of culturally embedded leadership style, gender diversity, and employee voice/silence behaviour is vital to investigate and understand for better policy development. Further, leaders of higher educational institutions need to understand how their behaviour impacts subordinate's decision to speak or remain silent at the workplace.
On a theoretical level, it is the contribution of the study to investigate the role of similarity attraction account among the interaction of leader behaviour and employee voice/silence behaviour. As it is among the pioneering study to investigate the impact of similarity attraction account on employee voice or silence behaviour in the presence of culturally embedded leadership style.

LIMITATIONS AND RECOMMENDATIONS
The study also has some limitations. First, the study used a cross-sectional design to collect data, which limits the researcher's ability to claim causal effects. Future studies can work on longitudinal and experimental designs to test causal relationships. Second, the study relied on self-response from employees, other researchers can collect data from supervisor and subordinates to reduce response biasness.
Third, this study was conducted in public sector universities of Balochistan, which limited the generalizability of results to other university settings or other types of organizations. To enhance the generalizability other researcher can replicate the study in other sector to validate results. Fourth, there can be many other variables that can interact with the paternalistic dimension to determine voice or silence behaviour, such as the organization's structure, climate, and/or culture voice mechanisms and nature of the issue, etc. that can be considered for further enhancing the concepts.

AUTHORS' CONTRIBUTION
Saadia Babur is the main contributor who developed the main theme related to the concept under consideration. Zainab Bibi performed the drafting and revising of the entire manuscript. Jahanvash Karim performed the statistical procedures and interpretation