Original Research

Supervisory trust to be earned: The role of ethical leadership mediated by person-organisational fit

Anton Grobler, Maggie M.E. Holtzhausen

Received: 26 Apr. 2017; Accepted: 04 Sept. 2017; Published: 28 Mar. 2018

Copyright: © 2018. The Author(s). Licensee: AOSIS.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Background: The trust relationship between employees and their supervisors (called supervisory trust) has a definite impact on employee behaviour and attitudes. Furthermore, various studies found that ethical leadership impacts on supervisory trust, but in different contexts, and often with homogeneous or limited samples. The interactionist construct of person-organisational fit (P-O fit), consisting of a combination of supplementary fit (indirect fit or value congruence) and complementary fit (direct or person-job fit, as well as needs-supply fit), may however impact on the relationship between ethical leadership and supervisory trust. The unique permutations of these relationships are important not only for conceptualisation purposes, but also for intervention design to enhance the employees’ trust in their supervisors; this would contribute to positive employee behaviour and attitudes.

Aim: The purpose of this study was to determine whether a relationship exists between ethical leadership and supervisory trust, with possible mediation by P-O fit.

Setting: The research was conducted among ±60 employees from each of 17 private sector and 4 public sector organisations in South Africa.

Method: This study utilised a positivist methodology based on an empirical approach, while using a cross-sectional design and quantitative analysis. The sample is relatively representative (in terms of race, gender and the South African work force), as it consisted of 60 employees from each of the 21 South African organisations that participated in the study, with 1260 respondents in total.

Results: Significant, positive relationships were found between ethical leadership, P-O fit and supervisory trust. Additionally, it was found that P-O fit partially mediates the relationship between ethical leadership and supervisory trust, confirming the proposed model.

Conclusion: A strong, positive relationship exists between ethical leadership (consisting of morality and fairness, role clarification leadership and power-sharing leadership) and supervisory trust, which is partially mediated by P-O fit (consisting of supplementary fit and complementary fit).

Introduction

Blau (1964:98) argues that trust lies at the centre of all relationships and, therefore, influences the way parties behave towards each other. The same may be said for supervisory trust that has been found to influence constructive employee outcomes, holding much benefit for organisations. For example, supervisory trust was found to relate positively to both individual and organisationally directed organisational citizenship behaviour (OCB) and to task performance (Aryee, Budhwar & Chen 2002:276). In addition, Davis et al. (2000:564; 568–571) maintain that employees who perceive their supervisors as trustworthy feel safer, act more productively and indicate loyal behaviour towards their organisation. Fair and caring supervisory treatment and leadership create higher levels of trustworthiness, which potentially fosters social exchange relationships, resulting in employees who are more likely to perform beyond their normal call of duty (Brown & Treviño 2006:597; Ruiz-Palomino & Martınez-Canas 2014:103). When there is supervisory trust, a relationship develops which is characterised by a positive reciprocal concern, just behaviour and dedication (Chen, Aryee & Lee 2005:465–466). The importance of supervisory trust for organisational success is evident from the above and the reason why further research on the topic is necessary.

Ethical leadership contributes to higher levels of supervisory trust as employees believe ethical leaders will remain true to their promises and will fulfil their obligations (Kalshoven, Den Hartog & De Hoogh 2011:65). Perceived ethical leadership and behaviour heightens employee engagement through intensified feelings of vigour, dedication and absorption at work (Den Hartog & Belschak 2012:42–43). Ethical leadership directly and indirectly affects employees’ level of commitment and turnover intention (Shin et al. 2015:54). This article therefore focuses on the relationship between supervisory trust and ethical leadership.

The emphasis of the article lies in the relationship between supervisory trust and ethical leadership, but with the addition of person-organisational fit (P-O fit), a relationship that has not been widely researched. This is an important contribution as the significance of P-O fit as an interactionistic concept has been widely accepted, but not in combination with supervisory trust and ethical leadership. For instance, drawing on Blau’s Social Exchange Theory of 1964, Farzaneh, Farashah and Kazemi (2014:683–684) found that the perceived fit employees experience between their needs and capabilities, versus the benefits provided by their job or organisation, prompts employees to commit to their organisations with increased levels of OCB and to perform tasks given to them.

Purpose and research objectives of the study

The purpose of this study was to determine whether a relationship exists between ethical leadership and supervisory trust, with possible mediation by P-O fit. Aligned to this purpose, four research objectives were formulated. The first two research objectives are based on the existing literature available, and the last two are based on the empirical investigation. The research objectives are: (1) to define the constructs ethical leadership, P-O fit and supervisory trust; (2) to identify and report on previous studies where the relationship between ethical leadership, P-O fit and supervisory trust and related concepts were found; (3) to determine the relationship between ethical leadership, P-O fit and supervisory trust; and lastly, (4) to test the mediating effect of P-O fit on the relationship between ethical leadership and supervisory trust.

First research objective: Defining the constructs

The first research objective entails defining the constructs of supervisory trust, ethical leadership and P-O fit from literature.

Defining supervisory trust

Rousseau et al. (1998:395) define the general construct of trust as ‘a psychological state comprising the intention to accept vulnerability based upon positive expectations of the intentions or behaviour of another’ – thus emphasising two aspects central to most descriptions of trust: the positive expectations and beliefs of the person placing his trust in another that the other person is competent, honest and caring; and secondly, a willingness to accept vulnerability, reflecting the intent to be dependable on another. Mayer, Davis and Schoorman (1995) define trust as:

the willingness of one party to be vulnerable to the actions of another party based on the expectation that the other party will perform a particular action important to the trustor, irrespective of the ability to monitor or control that other party. (p. 712)

Similarly, McAllister (1995:25) describes interpersonal trust as the extent of confidence a person has to willingly act on the basis of the words, actions and decisions taken by another.

More specifically to supervisory trust, Sanders and Schyns (2006:514–515) support the aspect of mutual dependency, stating that according to the Leader-Member-Exchange (LMX) theory, subordinates will trust their leaders when they experience supervisors as trusting them. Mulki, Jaramillo and Locaander (2006:20) conclude that most explanations of supervisory trust do, in fact, have this one common element of an individual’s behaviour that reflects his or her vulnerability to another person in an exchange relationship.

Defining ethical leadership

Ethical leadership is defined by the pivotal work of Brown, Treviño and Harrison (2005:120) as ‘the demonstration of normatively appropriate conduct through personal actions and interpersonal relationships, and the promotion of such conduct to followers through two-way communication, reinforcement, and decision-making’. Brown et al. (2005:119–121) thus recognise the two dimensions of ethical leadership as reported by the seminal work of Treviño, Hartman and Brown (2000:28), namely the moral person and the moral manager.

The reference to a moral person (Treviño et al. 2000:28) touches on the characteristics and traits of the leader, e.g. being caring, honest, a principled decision maker, and behaving ethically – both professionally and personally. Further characteristics of ethical leadership include fairness, integrity and consistent behaviour that, in turn, encourages ethical behaviour (Brown et al. 2005:130; De Hoogh & Den Hartog 2008:298; Kalshoven et al. 2011:51; Treviño, Brown & Hartman 2003:5;14–18). Ethical leaders foster good relationships and behave courteously, politely and in a friendly manner towards employees and other stakeholders (Frisch & Huppenbauer 2014:32–38), thereby showing moral and legal conduct (De Hoogh & Den Hartog 2008:298). They have an internal sense of duty and operate in ways conducive to the greater good (Treviño et al. 2003:19).

These characteristics result in ethical leaders being viewed by their employees as more credible and trustworthy (Ruiz-Palomino & Martınez-Canas 2014:97). In fact, according to research (Brown & Treviño 2006:597; Frisch & Huppenbauer 2014:34–35; Treviño et al. 2003:14) ethical leaders regard it as imperative that their behaviour results in their employees considering them as ethical role models and are often influenced by ethical role models themselves. Den Hartog (2015:418–428) argues that, should leaders act as ethical role models, ethical conduct will increase in the total organisation and not only among the immediate followers. Indeed, Mayer et al. (2009:7–9) put forward a trickle-down model suggesting that ethical leadership flows down from top-level management through supervisors, to employees. Demirtas and Akdogan (2015:64) confirm that leaders modelling ethical behaviour achieve moral authority which positively affects organisational members; similarly, the ethical and moral behaviour of top management influences ethical supervisory behaviour (Choi 2014:4776).

The moral manager (Treviño et al. 2000:28) refers to ethical leaders who actively endeavour to inspire change in their followers by demonstrating ethical conduct; they also encourage employee voice. While holding followers accountable for ethical behaviour and actions, they remain accountable for their own actions (Treviño et al. 2003:18). Ethical leaders communicate ethical standards to their followers by sending clear messages about ethical values, while also being clearly noticeable as ethical and honourable leaders (Brown et al. 2005:120, 130; Brown & Treviño 2006:597; De Hoogh & Den Hartog 2008:298; Kalshoven et al. 2011:51; Treviño et al. 2003:5,14–18).

Ethical managers and leaders consider fairness in all they do (e.g. in considering just compensation systems and providing job-related training or other suitable development opportunities), while also demonstrating trust in their employees through, for instance, increased decision-making powers (Frisch & Huppenbauer 2014:32–38; Kim & Kim 2013:157). Ethical supervisors act in the best interests of all their followers (Brown et al. 2005:130) – for example, employees, customers, society and the natural environment – thereby enhancing the well-being of several stakeholders (Frisch & Huppenbauer 2014:39). Ethical leadership diminishes employees’ anxiety about job uncertainty or workplace behaviours (Treviño et al. 2003:18). They are transparent and open in their communication with employees regarding their expectations and responsibilities (De Hoogh & Den Hartog 2008:298), and they encourage a transparent organisational culture (Huhtala et al. 2013:264).

Notwithstanding the many writings on ethical leadership, scholars such as Den Hartog (2015:426–428) and Yukl et al. (2013:38) point out that the exact conceptual meaning of ethical leadership still causes confusion. Contributing meaningfully to the research and findings on ethical leadership, Den Hartog (2015) and Yukl et al. (2013) summarise the most important aspects as set out below.

Den Hartog (2015:421–422) recaps ethical leadership as being comprised of leader behaviour that shows integrity and cognisance of moral values. In addition, ethical leadership is characterised as honest, fair, respectful, caring and trustworthy, with no indication of favouritism evident. Ethical leaders use power in a socially responsible manner and will also share power. They engage in social responsibility initiatives and make principled choices, showing a concern for sustainability issues. Furthermore, ethical leaders act in a transparent manner, engaging in open communication with followers, and encouraging employee voice. Finally, ethical leaders give ethical guidance to employees and clarify roles and responsibilities so that employees comprehend what is expected of them (Den Hartog 2015).

Similarly, and based on both theory and research, Yukl et al. (2013:40–41) summarise the most relevant aspects of this concept as including integrity (including consistent actions to advocate adopted values) and honesty; actions taken to communicate or administer ethical principles; fair decision-making and reward distribution (including no preferential treatment or use of rewards to motivate unsuitable actions); and conduct indicative of kindness, compassion and concern for others (altruism). Yukl et al. (2013:47) furthermore conclude that ethical leadership significantly relates to LMX and leader effectiveness – concluding also that ‘ethical leadership is not only commendable but also effective’.

Defining P-O fit

P-O fit is a complex and multidimensional construct (Farzaneh et al. 2014:674), aptly summarised by Ruiz-Palomino and Martınez-Canas (2014:97) as the level of similarity between an employee’s and an organisation’s beliefs, norms, values (Chatman 1989:339) and goals (Kristof-Brown, Zimmerman & Johnson 2005:318). Kristof (1996), one of the original researchers on P-O fit, defined it as:

the compatibility between people (employees) and organisations that occurs when (a) at least one entity provides what the other needs, or (b) they share similar fundamental characteristics, or (c) both. (pp. 4–5)

Similarly, Liu, Liu and Hu (2010:616–617) emphasise three components to P-O fit, namely the congruence between an organisation’s characteristics and the employee’s personality; the compatibility between the goals of the employee and the organisation; and lastly, the extent to which an employee’s values and the organisational culture coincide. Mitchel et al. (2001:1104–1105) concisely describe overall P-O fit as the way employees are perceived to be compatible and at ease with an organisation.

Various dimensions have developed regarding this broad concept, and scholars distinguish between supplementary and complementary fit (Kristof 1996:3), needs-supplies versus demands-abilities (Kristof 1996:3) and, lastly, the perceived (subjective) dimension contrasted with an actual (objective) dimension (Kristof-Brown & Jansen 2007:132–135).

Kristof (1996:3) explained the complementary fit as occurring when an employee’s characteristics complete an environment by adding that which was missing, and supplementary fit as occurring when an employee supplements or holds characteristics comparable to other employees. To enhance the possibility of a fit between the organisation and the employee, the employee’s personal values, goals and plans should fit the organisational culture, values, norms and goals, and the demands of the employee’s job (Kristof 1996:4).

A second theory by Kristof (1996:3) refers to the demands-abilities (occurring when an employee meets organisational demands) versus needs-supplies (occurring when the organisation fulfils the employee’s needs, requirements and preferences) theory. P-O fit will be higher when employees meet the demands of the organisation because the employees demonstrate the necessary abilities to do so, and the employee’s needs are also met by the organisation (Kristof 1996:4).

As stated above, Kristof-Brown and Jansen (2007:132–135) described two further dimensions of overall P-O fit – the subjective (perceived) fit and the objective (actual) fit between the organisation and the employee. It is especially the subjective fit which often determines behaviour (Kristof-Brown & Jansen 2007:132–135). Other authors have offered similar dimensions, for instance Cable and Edwards (2004:822; 829–831) identified a supplementary fit, closely associated with the subjective fit of Kristof-Brown and Jansen (2007:132–135). The subjective P-O fit occurs when similarity of personal and organisational values, beliefs, norms and goals is found, resulting in a feeling of being involved in the broader mission of the organisation (Cable & DeRue 2002:876). The objective fit is associated with the complimentary fit (Cable & Edwards 2004:822; 829–831) and refers to aspects such as the match between the skills of the employee as measured by the performance management of the organisation (Grobler 2014:5); or, differently put, the fit between the employee’s skills and abilities and the job or organisational demands (Grobler 2016:1421).

Second research objective: Reporting and explaining the relationship between the three variables

The second research objective relating to the literature review focused on the relationship between the antecedent of ethical leadership and the outcome of supervisory trust, and how this relationship is mediated by P-O fit. Research indicates that ethical leadership significantly impacts on both supervisory trust and P-O fit; however, the possible mediating effect of P-O fit on the relationship between supervisory trust and ethical leadership does not seem to be well researched. Some of the important work that are based on the proposed theoretical model (see Figure 1) will be discussed next.

FIGURE 1: A proposed model depicting the role of ethical leadership mediated by P-O fit in earning supervisory trust.

This model assumes a three-variable system such that there are two causal paths feeding into the outcome variable: the direct impact of the independent variable (Path c) and the impact of the mediator (Path b). There is also a path from the independent variable to the mediator (Path a).

The relationship between ethical leadership and supervisory trust

Trust is central to both institutional and interpersonal relationships (McAllister 1995:25, 53–55). Brown and Treviño (2006:597) confirm that ethical leaders are regarded as trustworthy, and ethical leadership contributes significantly to enhanced supervisory trust (Choi 2014:4771, 4776; Kalshoven et al. 2011:65), aiding supervisors and their underlings to develop trust-based relationships (Brown et al. 2005:122; 130). Sallee and Flaherty (2003:306) found in their research that, when a supervisor reflects a value system which portrays characteristics such as concern for their employees (noted as an ethical leadership quality), it benefits the formation of supervisory trust.

Ethical leadership enhances both affective and cognitive trust (Newman et al. 2014:119). McAllister (1995:25–26) described affective trust as strong emotional ties that are formed between leaders and their followers when a process of reciprocated social exchange is evident. Cognitive trust is described as trust-based on performance-related perceptions such as proficiencies, responsibility and dependability (Schaubroeck, Lam & Peng 2011:864). Brown and Treviño (2006:597) confirm that ethical leadership is positively related to affective trust, while being negatively related to abusive supervision.

Chughtai, Byrne and Flood (2015:655) argue that Blau’s Social Exchange Theory (1964) supports the idea that ethical leadership enhances supervisory trust, as constructive and valuable actions by supervisors lead to high-quality exchange relationships with subordinates, who, in turn, will reciprocate in equally positive ways. According to the seminal work of Blau (1964:94), employees will reciprocate through positive contributions towards the organisation, as social exchange relationships tend to engender feelings of personal obligations, gratitude and trust. Aryee et al. (2002:267, 276) used a social exchange model to suggest positive relationships between interactional justice perceptions and supervisory trust. Supervisory trust thus holds the perspective of a mutual exchange resulting from integrating both honesty and emotionally oriented activities (Sallee & Flaherty 2003:307).

Another significant finding relates to the impact of ethical leadership and supervisory trust on organisational culture. Top management, supervisors and peers need to continuously display moral virtuous behaviour combined with formal ethics mechanisms to build a strong ethical culture (Ruiz-Palomino, Martınez-Canas & Fontrodona 2013:184). A perceived ethical climate (where supervisors and managers are seen to honour ethical behaviour, remain true to their promises and act as ethical role models) enhances supervisory trust (Treviño & Brown 2004:72–73, 78–80). In fact, Mulki et al. (2006:23–24) determined that an ethical climate is regarded as an important precursor to supervisory trust.

Moreover, Ruiz-Palomino et al. (2013:181) argue that validating ethical behaviour within the organisation’s culture holds potential for ensuring satisfied and committed employees who want to remain with the organisation, and that this is a likely outcome of solid trust-based human relationships. Research has found that, even in situations where employees foresee a short-term relationship with the organisation and purely economic in nature, higher levels of emotional attachment to the organisation will exist, resulting in a bigger sense of obligation to it (Philipp & Lopez 2013:307; 310–311). Ethical leadership encourages OCB on both individual and group levels (Brown et al. 2005:130; Kalshoven et al. 2011:65; Mayer et al. 2009:8–11), with trustworthiness in supervisory ethical leadership being especially important in relation to OCB (Brown & Treviño 2006:612; Ruiz-Palomino, Ruiz-Amaya & Knörr 2011:252–253). Furthermore, ethical leadership significantly contributes towards an ethical climate, enhancing an environment of procedural justice which mediates the effects of top management ethical leadership on OCB (Ruiz-Palomino et al. 2013:175; 184; Shin et al. 2015:52–54).

Additionally, ethical leaders contribute to higher levels of employee engagement, commitment and trust among followers, heightening desired behaviour among followers (Brown et al. 2005:129–130; Choi 2014:4771, 4776; Den Hartog and De Hoogh 2009:220–223; Engelbrecht, Heine & Mahembe 2014:6–8; Hansen, Brown & Dunford 2013:444). Chughtai et al. (2015:659–660) found that supervisory trust is fully mediated by the effects of ethical leadership on, respectively, work engagement and emotional exhaustion, and thus employee well-being. Additionally, ethical leadership enhances trust among employees during organisational change, aiding the change process as employees trust their supervisors to take appropriate decisions and are therefore more likely to engage in OCBs (Sharif & Scandura 2014:186; 192).

The mediating and moderating effects of P-O fit on ethical leadership and supervisory trust

The mediating effect of P-O fit on the relationship between supervisory trust and ethical leadership was not extensively explored in the past. Boon and Biron (2016:2190–2102) found that the relationship between supervisors and their employees impacts on the perception of fit and subsequent behaviour, arguing that high-quality LMX relationships between supervisors and their employees affect the employee’s attitude and behaviour, and that the quality of these relationships is, among other things, based on the experience of high levels of trust between these two parties.

Kristof-Brown et al. (2005:311) regard Person-Supervisor fit as a specific type of P-O fit and found a moderate relationship between P-O fit and supervisory trust. However, their study also suggested that employees do not regard their supervisors as ‘isomorphic representations of the organization’ (Kristof-Brown et al. 2015:316). Kim and Kim (2013:158–161) found that, as Person-Supervisor fit increases, moral competence becomes stronger among leaders and employees; however, they caution that the construct of moral competence is distinctive from ethical leadership. Nonetheless, the finding is interesting and worthy of mentioning as moral competence is viewed as a multidimensional construct relating to integrity, responsibility, compassion and forgiveness (Kim & Kim 2013:156), and thus related to ethical leadership.

Ruiz-Palomino and Martınez-Canas (2014:104–105) found overall P-O fit has a moderating effect on the relationship of ethical culture of which ethical leadership is a component (Ruiz-Palomino et al. 2013:173) and ethical intent. Nonetheless, they also found that even when P-O fit is relatively poor, ethical culture is still strongly associated with ethical intent. In addition, Ruiz-Palomino and Martınez-Canas (2014:104–105) found that P-O fit mediated ethical culture and OCB.

Clearly, not much research exists on the mediating effect of P-O fit between ethical leadership and supervisory trust. Subsequently, this study aimed to support two empirical research objectives, namely ‘What is the nature of the statistical inter-relationships between ethical leadership and supervisory trust?’ and ‘Does P-O fit significantly mediate the relationship between ethical leadership and supervisory trust?’

Method

Research design

This study utilised a typical positivist methodology based on an empirical approach, while using a cross-sectional design and quantitative analysis. This approach enables the researcher to collect the required data across all participating demographic groups at the same time.

Sample

The participants consisted of employees of a convenience sample of 21 organisations in South Africa, with 17 organisations being from the private sector, including the Medical, Engineering, Retail, Construction, Financial, Telecommunication, Pharmaceutical and Information Technology industries. The public sector is less represented, with four organisations, consisting of National and Provincial Departments, as well Local Government. In each organisation, 60 employees were selected to participate in the study. The pooled data could therefore be considered to be a convenience sample (because of the convenience sampling of the participating organisations). The fieldwork was conducted by 21 co-researchers working on a larger project; ethical clearance was granted by the institution.

Although race, gender and age were used as grouping variables to distinguish between groups, it is still important to reflect on the composition of the sample in terms of these demographic variables. The majority of the participants were Africans (50%), followed by white people (31%), Indians (11%) and mixed race (8%). The representation of the gender groups was higher for males at 58% compared to 42% for females. The racial and gender distribution of the sample seems to be relatively representative of the South African workforce in general, taking into consideration that the distribution of the workforce as indicated by Statistics South Africa (2016:Appendix A) was 79.77% Africans, 8.36% white people (over-represented in the sample), 2.49% Indians and 9.17% mixed race. According to the same source, the proportion of males in employment is 49.36%, while the proportion for females stands at 50.64%.

The mean age of the respondents was 37.26 years (s.d. = 9.29), and the mean tenure in the specific organisation 7.24 years (s.d. = 8.27). The assumption can thus be made that the sample is well representative of the general work force, and that the participants, in terms of age and tenure, would be able to provide an accurate assessment of their perceptions of the constructs being measured.

Measuring instruments

The Ethical Leadership Questionnaire (ELQ) developed by De Hoogh and Den Hartog (2008:297–311), consisting of three sub-factors, was used. This instrument was designed to elicit respondents’ reports of behaviour of leaders with whom they are familiar and uses a 7-point Likert scale. The first factor is morality and fairness (six items), and the first item reads ‘The leaders in my organisation make sure that their actions are always ethical’. The second factor is role clarification leadership, and a typical item reads ‘The leaders in my organisation explain who is responsible for what’. The last factor is power-sharing leadership (six items), and one of the items reads ‘The leaders in my organisation allow subordinates to have influence on critical decisions’. De Hoogh and Den Hartog (2008:303) reported Cronbach’s alpha coefficients of 0.81, 0.88 and 0.78 for the respective factors.

P-O fit was measured using an instrument developed by Cable and Judge (1996:294–311). The initial instrument was multidimensional and consisted of three factors, with three items on each of the factors. However, Grobler (2016:1419–1434) developed a reconfigured factor structure (still with all of the original nine items), but with two factors. The first factor, supplementary fit (organisation fit as values congruence), consists of three items with one of the items reading ‘The things that I value in life are very similar to the things that my organisation values’. The second factor, called complementary fit (that includes needs–supplies fit, as well as demands–abilities fit), consists of six items. Two of the items read, for instance, ‘There is a good fit between what my job offers me and what I am looking for in a job’ (needs–supplies fit) and ‘The match is very good between the demands of my job and my personal skills’ (demands–abilities fit). The instrument uses a five-point Likert scale. Cronbach’s alpha coefficients of 0.92 and 0.89 were reported for the two factors respectively by Grobler (2016:1428).

Supervisory trust was measured using Robinson’s (1996:574–599) seven-item scale. This scale integrates both cognitive and affective views of trust between individuals, and a typical item reads ‘I believe my employer’s motives and intentions are good’. Mulki et al. (2006:22) reported an acceptable Cronbach’s alpha coefficient (α > 0.70) for this unidimensional instrument.

Statistical analysis

The statistical analysis was performed by using the Statistical Package for the Social Sciences (SPSS version 24), supported by Analysis of a Moment Structures (AMOS) version 24. Descriptive statistics were calculated to provide information on the distribution, with the mean score as either the average, or as the precise centre of the amalgamated values, with the standard deviation as the measure of variability. Skewness and kurtosis were also calculated to investigate the distribution of the data. The critical values for these two statistics are two and seven respectively (West, Finch & Curran 1995:74).

The Cronbach’s alpha coefficient (α) was calculated to test the proportional variance error and the internal consistency of the instrument. A score α = 0.70 or higher is considered by Clark and Watson (1995:309) as acceptable. Multicollinearity (tolerance and variance inflation factor [VIF]) of the items was also determined (with the main construct as dependant variable) to test a possible inflation of the reliability coefficient.

Correlations between the constructs were calculated by means of Pearson’s product moment correlations. Stepwise multiple regression analysis was used to determine the amount of variance explained by ethical leadership construct in supervisory trust, when P-O fit is forced into the analysis. The rationale for this forced inclusion of P-O fit is to determine whether it improves the model or not. The multiple regression was also utilised to determine the relative strength of the two independent variables in the prediction of the dependent variable (in terms of the beta values). The tolerance, as well as VIF values, was also calculated to test for possible multicollinearity (Pallant 2010:136–138). Tolerance is an indicator of the amount of variance not explained by the other independent variables (in this case item) in the model, and should preferably be larger than 0.10. VIF on the other hand is the inverse of tolerance, and values should be below 10.

In addition to the multiple regressions, structural equations analysis by means of structural equation modelling (SEM) was used because it presents some advantages over traditional multivariate techniques (Haenlein & Kaplan 2004:285). In order to perform SEM, the missing values were deleted case-wise. The missing values were accounted for less than 2% of the total sample and thus did not impact on the analysis. AMOS, which is statistical software, was used to perform the SEM to analyse the model from a theoretical perspective. To clarify the meaning of mediation, a path diagram was introduced as a model for depicting a causal chain. The basic causal chain involved in mediation is shown in Figure 1. To assess the model fit, several fit indexes were used, including the comparative fit index (CFI), the root mean square error of approximation (RMSEA), chi-square (χ²) and the ratio of the differences in chi-square to the differences in degrees of freedom (χ²/df). Given that there is no one acceptable cut-off value of what constitutes adequate fit, it was elected to evaluate each model and to recommend the model closest to the CFI value of 0.90, an RMSEA value of 0.05 and χ²/df, a ratio of less than 5.00 (Byrne 2010).

The Baron and Kenny procedure was also performed to confirm the relationship between the variables. The aim of this procedure is to determine whether the independent variable affects the dependent variable through a mediating variable (Zhao, Lynch & Chen 2010:205). In addition to Baron and Kenny’s procedure, the Sobel test, which is considered to be suitable for large samples (Preacher, Rucker & Hayes 2007:200), was applied. Preacher and Hayes (2004:718) regard the Sobel test as sufficient in terms of its power and intuitive appeal. They also indicate that the rough critical value for the Sobel test is ±1.96 (p < 0.05) for a significant mediation effect.

Ethical consideration

Ethical clearance was obtained for the research from the relevant institution (School of Business Leadership, University of South Africa).

Findings

The descriptive and collinearity statistics of each of the items across the three constructs were calculated, and the majority of the values on the skewness and kurtosis scales reported were negative. This is an indication that the distribution tails off to the left. Subsequently, the negative skewness contributes to the relatively high mean scores on ethical leadership (ranging from 4.05 to 5.10 on a seven-point Likert scale), P-O fit (ranging from 3.27 to 3.96 on a five-point Likert scale) and Supervisor trust (ranging from 4.45 to 5.46 on a seven-point Likert scale). The skewness and kurtosis values for the items do, however, not exceed the critical values of 2.00 and 7.00 respectively, which is an indication of a normal distribution of the data. The tolerance as well as the VIF values on each of the items further suggests that there is no violation of the multicollinearity assumption within the respective constructs. The descriptive statistics and the psychometric properties of the instruments are reported in Table 1.

The descriptive statistics in Table 1 show that the skewness and kurtosis values of the factors do not exceed the critical values of 2.00 and 7.00 respectively, which is an indication that the data are normally distributed. The majority of the values of the ethical leadership, P-O fit and supervisory trust constructs and sub-factors on the skewness scale were negative, which is an indication that the distribution has relatively few small values and tails off to the left. The Cronbach’s alpha coefficients of the factors are acceptable if the guideline of α > 0.70 is applied. It would thus appear that the factors possess acceptable levels of internal consistency, providing some evidence that the items measure the same general construct, which assures the overall reliability of the instruments.

Third research objective: To determine the relationship between ethical leadership, P-O fit and supervisory trust

In order to investigate the third research objective of this study, namely to determine the relationship between ethical leadership, P-O fit and supervisory trust, a correlational analysis and a stepwise multiple regression analysis were performed. The strength and the direction of the linear relationship between the factors (and total score) of ethical leadership, P-O fit and supervisory trust are reported in Table 2.

The correlations coefficients reported in Table 2 indicate a strong positive and statistical significant (p ≤ 0.001) relationship between the sub-factors of ethical leadership, P-O fit and supervisory trust. The purpose of this study is, however, to determine the relationship between the overall (or total scores) of the constructs. Supervisory trust and ethical leadership are correlated (r = 0.71), while supervisory trust is also correlated with P-O fit (r = 0.58) (both p ≤ 0.001). There is thus a strong, positive relationship or association between supervisory trust and both ethical leadership and P-O fit. The practical implication thus is that a change in P-O fit as well as ethical leadership will impact (in the same direction) on supervisory trust. An improvement in P-O fit and ethical leadership will thus also lead to an improvement in supervisory trust.

In order to analyse these relationships further (in support of the third research objective), a stepwise multiple regression analysis was performed. The results are reported in Tables 3 and 4.

The stepwise multiple regression analysis, with supervisory trust as dependent variable and the two independent variables as unidimensional constructs (using the total scores of ethical leadership and P-O fit), yielded significant results with ethical leadership explaining 51% of the variance in supervisory trust (F(1, 1221) = 1261.60, p < 0.001). With the addition of P-O fit, the model improves by 4% (total variance explained is 55%), with F(2, 1220) = 748.50 (p < 0.001).

The results of the stepwise multiple regression support the third research objective, as it shows that both ethical leadership and P-O fit explain the variance in supervisory trust. Subsequently, in order to determine the relative strength of the independent variables (ethical leadership and P-O fit) in the prediction of supervisory trust, the beta coefficients (β) were determined. The results are reported in Table 4.

The tolerance value is 0.67 and the VIF value is 1.49, indicating non-multicollinearity between the two independent variables, in this instance ethical leadership and P-O fit. Thus, although ethical leadership and P-O fit are highly correlated (see Table 2), it can still be used independently from each other with a substantial degree of accuracy.

In support of research objective three, as well as previous results reported, it was found that the strongest predictor of supervisory trust is ethical leadership (β = 0.57, ≤0.001), followed by P-O fit (β = 0.25, ≤0.001). Thus, in order to improve supervisory trust, the main intervention should be focused on the perceptions of ethical leadership. Supervisory trust can further be enhanced with the addition of interventions to improve P-O fit. Together, ethical leadership and P-O fit can contribute up to 55% in the improvement of supervisory trust.

Fourth research objective: To test the mediating effect of P-O fit on the relationship between ethical leadership and supervisory trust

The fourth research objective of this study was to determine the possible mediating effect of P-O fit on the relationship between ethical leadership and supervisory trust. SEM and specifically the Baron and Kenny’s procedure was performed to test the possible mediating effect, and the results are reported in Table 5.

After inspection of the results reported in Table 5, it is clear that a mediated effect and direct effect exist and point in the same direction. The estimated beta value of the relationship between ethical leadership (independent variable) and supervisory trust as dependent variable is 1.18, but decreased to 0.86 after P-O fit was introduced as mediating variable. This is also referred to as complementary partial mediations or as ‘consistent’ or ‘positive confounding’ models (Zhao et al. 2010:200). This finding is supported by Sobel z score of 2.48 (p =0.01).

Finally, in order to confirm the model (as shown in Figure 1), a SEM was conducted. The results indicated a good fit (chi-square = 246.14, df = 44, p < 0.001, IFI = 0.97, TLI = 0.96, CFI = 0.97, RMSEA=0.06). This result supports the fourth research objective, as it was found that P-O fit causes mediation in the supervisory trust and ethical leadership. It thus explains the relationship between the supervisory trust and ethical leadership; it can be regarded as an intervening variable. Practically, it means that P-O fit could be used as an intervening (or intervention) variable to improve the relationship between ethical leadership (the perceptions thereof) and supervisory trust.

Conclusion

From the literature it is clear that ethical leadership significantly impacts on the level of trust evident in organisations, also between employees and supervisors. Ethical leadership is evident through a trickle-down effect (Mayer et al. 2009:7–9), impacting from the ethical leader at top management level, to the supervisor, to the employee. Ethical leadership also enhances supervisory trust (Choi 2014:4771–4776; Kalshoven et al. 2011:65), resulting in high-quality social exchange relationships (e.g. Chughtai et al. 2015:655). The advantages of both ethical leadership and enhanced supervisory trust are multifold, including for instance strong OCB and increased employee engagement. The literature further points out that a high-quality LMX relationship based on trust between supervisors and employees impacts positively on employees’ outlooks and behaviour (Boon & Biron 2016:2190–2102), enhancing P-O fit (Ruiz-Palomino et al. 2013:183).

This empirical study also confirmed a strong, positive relationship between ethical leadership (consisting of morality and fairness, role clarification leadership and power-sharing leadership), P-O fit (consisting of supplementary fit and complementary fit) and the unidimensional supervisory trust.

It was found that ethical leadership and P-O fit in combination explain 55% of the variance in supervisory trust, with ethical leadership explaining 51%. P-O fit was found to be a mediating or confounding variable, mediating the relationship between ethical leadership and supervisory trust. This implies that, if the targeted variable is supervisory trust in terms of intervention, the organisation should enhance the perception of ethical leadership through direct, honest and transparent communication, as well as the actual ethical conduct of leaders. This could be supported by interventions to improve overall P-O fit that consists of two components. Overall P-O fit can therefore be enhanced through constant value-based interaction (to ensure value congruence between the employees and the organisation – component one), as well as direct job, performance and reward-related discussions (to ensure need-supply and demands-abilities fit – component two).

However, this research has certain limitations – mainly in terms of the methodology. All three instruments are based on self-reporting – a method which may lead to method bias. This may be a reality, even with the assurance provided to participants during the briefing regarding anonymity and confidentiality. Social desirability and subsequent response bias will always remain a concern and a limitation in studies such as this one, while self-reporting may be seen as a one-sided report from the respondents’ side. An additional possible limitation is that the wording of the initial scales was used ‘as is’, without adapting it to the South African (multilingual) context. A further limitation of this study is the drawback of a cross-sectional design which might have artificially increased the relationship between the three components.

A recommendation for further studies is to investigate the relationship between the three constructs over a period of time through a longitudinal study. Another recommendation is to analyse results further with the possible addition of the effect of membership of specific demographic groups (e.g. difference between sectors) and to include other work attitudes and organisation behavioural constructs in the analysis.

Acknowledgement

The authors declare that they have no financial or personal relationship(s) that may have inappropriately influenced them in writing this article.

The article was co-written by both authors.

References