Ethical Leadership Measures for Public Service in Sri Lanka

A lack of concern on ethics in public administration may be considered the biggest barrier to quality service delivery. Many efforts have been made to promote ethical standards between public servants. Among those efforts, improving leadership skills can prove to be effective in promoting ethics in public service (Hart 2001).Therefore, ethical leadership behavior is very vital as that ensures ethical guidance. In order to identify the ethical leadership behavior of the public sector, customized measures are essential. This study describes the process of development of a 30-item instrument for measuring ethical leadership behavior in the public service with reference to Divisional Secretariats in Sri Lanka. According to Hinkin's (1998) recommendations in generating items both qualitative and quantitative methods were utilized. Firstly, it observed articles which were published related to ethical leadership and examined to find items and dimensions. A focus group discussion was conducted with 20 in-depth interviews to generate additional items to determine whether additional dimensions of ethical leadership would emerge. Based on these two steps, the researcher found 30-items with five dimensions. They were fairness, employee orientation, role clarification, integrity and ethical guidance. These 30-items were included in a questionnaire and a quantitative study was undertaken with 100 respondents employed in the Divisional Secretariats within Gampatha District. Reliability test and factor analysis were used to ensure reliability and validity of this measures. Finally, these 30-items were loaded to four dimensions. Such as people orientation, integrity, ethical guidance and role clarification.


Introduction
Ethics is a must for public administrators because ethics are rules that define moral conduct according to the thought of a specific group. Though, the importance of ethics in public administration has long been neglected until recently (Haq 2011).The code of ethics ensures that the public receives what it needs in a fair manner. It also gives the administration guidelines for integrity in their operations. Ensuring ethics in the public service is an essential matter. Ethics are statements, written or oral, that prescribe or proscribe certain behaviours under specified conditions (Nigro & Nigro, 1989, p.37). Rosenbloom (1989) mentioned that ethics can be considered a form of self-accountability or an "inner check" of the conduct of public administrators. According to the United Nations Department of Economic and Social Affairs (1999 cited Haq 2011), public service ethics contains a broad range of principles and values. Those are objectivity, impartiality, fairness, sensitivity, compassion, responsiveness, integrity, accountability, transparency, selfless devotion to duty and protection of public interest. Ethical dilemmas usually take place around administrative discretion, accountability, corruption, nepotism, interest group pressure, information secrecy, policy complicacies etc. Some researchers found recent ethical scandals in business have raised important questions about how the role of leadership influences ethical conduct (Colvin, 2003;Mehta, 2003;Revell, 2003 cited Haq;2011). Most employees look outside themselves to significant others for ethical guidance (Kohlberg, 1969;Trevino, 1986). Therefore, leader can play a vital role of ethical guidance in the work place. Amaradasa (2012) a former President of the Sri Lanka Association for Quality mentioned that, the non-existence of a strong Sri Lankan work ethic, may be considered the biggest barrier to national quality and productivity improvement. Also, he mentioned the ability of the leadership to become perfect role models, will certainly pave the way for building a strong work ethic in Sri Lankan public sector in particular. Therefore, Public Institutions can be developed ethical behavior by providing an ethical leadership. According to the existing literature, no studies have been done to measure the ethical leadership provided by public institutions. Several measures have been developed to measure the ethical leadership, but they all have limitations (Yukl et al., 2013). Hence, there is a need of a research to develop customized ethical leadership measures for public service in Sri Lanka because Sri Lankan public service has not given attention for developing customized ethical leadership measures. Thus, the purpose of this study is to develop measures to examine employee perceptions on ethical leadership over the public sector organizations with reference to Divisional Secretariats. Divisional Secretariats are the institutes at the base level of the hierarchy, through which more than 90 percent of the government related services provided to general public (Herath,2008).The services are Civil Registration, Issuing of Permits/Licenses, Payment of Pensions, Samurdhi Program, Social welfare, Social Benefits and development programs.
The objective of this study was two-fold; (1) to develop customized items to measure ethical leadership of public service in Sri Lanka (2) to evaluate the reliability and validity of this new measures.

Literature
Most researchers have started to consider ethical leadership as a set of behaviors or a separate leadership style in itself rather than focusing only on the ethical components of other leadership styles (Brown et al., 2005;De Hoogh & Den Hartog, 2008, 2009Kanungo, 2001). According to the Webster dictionary the basics of ethics are dealing with what is good and bad, moral duty and moral obligation based on this Kanungo (2001) conceptualizes the ethical leadership using altruism approach and mentioned ethical leadership as a tension between altruistic and egoistic motives (e.g., Kanungo, 2001;Turner, Barling, Epitropaki, Butcher, & Milder, 2002). This approach shows that an ethical leader is driven by a system of accepted beliefs and proper judgments rather than self-interest, which is beneficial for followers, organizations and society. Following this approach, Kanungo (2001) and Aronson (2001) said the effect of leader's actions on followers as a major concern in ethical leadership. Brown et al. (2005) said ethical leadership is a separate style and defines ethical leadership 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" (p. 120). Ethical leader is a role model of appropriate behavior and use reward and punishment to stimulate ethical conduct (Brown et al., 2005;Trevino et al., 2003). Brown et al. (2005) address ethical leadership according to the social learning perspective and mentioned that followers will come to behave similar to their leader through imitation and observational learning (cf., Bandura, 1986). As well as this approach, some researchers view the ethical leadership from a social exchange approach (e.g., Mayer, Kuenzi, Greenbaum, Bardes, & Salvador, 2009;Turner et al., 2002). The social exchange approach focuses more on the norm for reciprocity (Cropanzano & Mitchell, 2005) and holds that followers are willing to respond when they treated fairly and with concern by their leaders (e.g., Mayer et al., 2009). Both of these approaches help to understand how individuals' react to the ethical leader behavior. Brown et al. (2005) developed a questionnaire to measure this construct called the Ethical Leadership Scale (ELS). This scale has 10 items, each with a 5-point Likert-type response format (1 = strongly disagree, 5 = strongly agree). According to the literature most researches used this scale to measure the ethical leadership behavior. However, Kalshoven, Den Hartog, and De Hoogh (2011) argued, this type of short scale is useful for certain research purposes because theoretically the core behaviors of ethical leadership seem rather different and they may have different antecedents and consequences. Also, combining different behaviors into a single undifferentiated construct would make it harder to expose the different mechanisms through which ethical leadership develops and may be effective. Yukl et al. (2013) also, mentioned some limitations of this scale. That were some relevant aspects like honest communication, behavior consistent with espoused values, fair allocation of assignments and rewards of ethical leadership not clearly included. Another limitation was that two of the items; listens to what employees have to say and has the best interests of employees in mind, were valued more representative of consideration than of ethical leadership. This ELS scale combines different leader behaviors, including allowing followers' voice, acting fairly and honestly, and rewarding ethical conduct in a single scale.
According to some ethical leadership literature, the researchers suggest several behavioral dimensions of ethical leadership in organizations. De Hough and Den Hartog (2008), distinguished the first three dimensions of ethical leadership. These were fairness, power sharing and role clarification. These behaviors are indicated as well in the work done by Brown et al. (2005). First, fairness is seen as a significant form of ethical leader behavior since ethical leader perform with integrity and treat others fairly. He makes principled and fair choices is trustworthy and honest, does not practice favoritism, and takes responsibility for his own actions (Brown et al., 2005;De Hoogh & Den Hartog, 2008;Trevino et al., 2003). Power sharing is the second ethical dimension. De Hoogh and Den Hartog (2009) argue that an ethical leader permits subordinates in decision making and listen to their ideas. Researchers discussed these empowering feature of ethical leadership (Resick et al.,2006). Likewise, Brown et al. (2005) proposed that ethical leaders provide followers with voice. Yukl, (2006) said power sharing allows followers more control and makes them less dependent on their leaders. Third, ethical leader is transparent and always engage in open communication (Brown et al., 2005).
Based on this, De Hoogh and Den Hartog (2008) pointed out that transparency is a component of ethical leadership because it clarifies performance goals, expectations and differentiate role clarifications. An ethical leader explains responsibilities, performance goals and expectations therefore subordinates know what is expected from them and the level of performance is up to the individual. Also subordinates do not worry unreasonably about uncertain expectations and they know how they can meaningfully contribute to achieve organization's goals. Some theoretical works having a true concern for people (people orientation) is a very important characteristic in an ethical leader's behaviors (Trevino et al.'s 2003;Resick et al. 2006). This people orientation dimension of ethical leadership reflects that ethical leaders are genuinely caring, respecting, supporting subordinates and where possible confirming that their needs are met (Kanungo & Conger, 1993;Trevino et al., 2003cited Kalshoven 2011).Then, Trevino et al. (2003 mentioned that ethical leaders clearly convey standards regarding ethical conduct. Top management of the organizations set rules, principles and codes of conduct which gives the guidelines for ethical behavior and leaders make the subordinates aware of these guidelines (Beu & Buckley, 2001cited Kalshoven 2011. The ethical leaders use rewards and punishments to grasp subordinates responsible for their actions (Trevino et al., 2003). Likewise ethical leaders guide subordinates in setting priorities and ethical dilemmas they experience (Brown et al.,2005). According to the behavioral integrity literature, integrity behaviors are defined as word-deed alignment or the extent to which what one says is in line with what one does (e.g., Dineen, Lewicki, & Tomlinson, 2006;Palanski & Yammarino, 2007,2009). Simons (2002) said that leaders who keep promises and behave consistently can be trusted because they work or behave as expected. In another way, Yukl (2006) says that the leaders as being ethical when they keep promises and behave consistently.
Based on these dimensions, Kalshoven, Den Hartog, and De Hoogh (2011) developed, revised and extended measures called the Ethical Leadership Work Questionnaire (ELW). The scale contains 38 items under seven dimensions like fairness, integrity, ethical guidance, people orientation, power sharing, role clarification, and concern for sustainability. Each item had a 7-point Likert-type response format (1=strongly disagree, 7 = strongly agree). But, Yukl et al.(2013) argue these behaviors are not inherently ethical, and they can be used for unethical purposes too. The dimension of sustainability involves social issues, and it is only one of many social issues that leaders may determine to endorse and support. Yukl et al.(2013) developed the Ethical Leadership Questionnaire (ELQ) which have 15 items, and each item had a 6-point Likert-type response format (1=strongly disagree, 6=strongly agree).The ELQ was contained with the most important elements of ethical leadership like integrity, honesty, communication of ethical values, fairness, ethical guidance, and altruism. However, no attention has been paid to development a customized measures of ethical leadership in public services since public service nature is different than the natures of profit oriented organizations. According to Ozer (1999 cited Erdogan & Bavik 2008) development of industry specific measurements is essential for a better fit to the nature of the industry. Therefore, objective of this study is to fill the gap in the relevant literature.

Methodology
According to Hinkin's (1998) recommendations in item generation qualitative and quantitative methods were utilized. For item-generation, the researcher first searched publish articles related to ethical leadership to find items and to generate dimensions. Then, the researcher carried out a focus group interview with six senior officer in public sector who were reading their PhD and 20 in-depth interviews were conducted with employees who work in divisional secretariats to generate additional items and to determine whether additional dimensions of ethical leadership would emerge. Participants were asked to describe a boss they regarded as an ethical leader. Example what made the boss an ethical leader, what kinds of behaviors were important to reveal the ethical leadership. Based on the literature and participants' explanation of ethical leader behaviors identifies 30-items with five factors emerged. They were fairness, integrity, role clarification, employee orientation and ethical guidance. Then a quantitative study was carried out to develop customized ethical leadership measurement for public service in Sri Lanka.

Analysis and Results
Resulting 30-items put into a questionnaire and it was used to collect data for first stage validation. Employees were asked to evaluate their immediate boss according to their perception. Chu and Murrmann (2006 cited Wijesekera & Fernando 2017) said, this stage was mainly serving the confirmation purpose of newly developed scales' psychometric properties. A fve-point Likert scale (Likert, 1932) ranging from (1) 'strongly disagree' to (5) 'strongly agree' was used. Since Brown et al. (2005) used this fve-point Likert scale for their ELS measure since most researchers heavily applied this scale to measure ethical leadership. The sample of the pilot study contained of 100 employees in five divisional secretariats in Gampaha District. The 30 items questionnaire was translated into Sinhala and both Sinhala & English questionnaires distributed separately as required by the respondents. To qualify for the study, respondents had to have worked in the current Divisional Secretariats during the past six months and they should evaluate their immediate boss. Figure 1 shows the service period of current Divisional Secretariat. Hundred questionnaires were distributed using non-probability judgmental sampling technique to respondents and they were asked to fill out the questionnaires in a self-administered manner. Ninety five questionnaires were returned and out of them only ninety three found to be useful representing a 93% response rate. 41% of the respondents have passed the GCE Advance Level examination and 27% of the respondents have completed the degree and post graduate degree. (Figure 2) The Cronbach's Alpha value for the 30 items was .96. There was no item to be deleted. Corrected Item-Total Correlation is the correlations among each item and the total score from the questionnaire. In a reliable scale, all items should correlate with the total.
Thus, it should be looked for items that don't correlate with the overall score from the scale: if any of these values are less than about .3 then there are a problem, because it means that a particular item does not correlate very well with the scale overall. Nunnally (1970) suggested omission of the items (<.3) with low corrected item-to-total correlations. The Cronbach's Alpha of these thirty items was 0.96. Cortina (1993) stated that many items are pooled, internal consistency estimates are relatively large. Corrected item-to-total correlations of these items were above .3. All thirty items had strong loadings on the construct, therefore they were supposed to measure indicating uni-dimensionality and construct validity. Total Variance Explained was 69.66%. The first stages of this scale development, there was no item to be deleted from the scale.
Second stage, Factor loadings obtained from EFA with Varimax rotation were further considered to test the factors and eliminate the poor performing items. All items were laded to four factors and Table 1 indicates the summary of 30 items loaded to four factors. The third stage of this scale development process, reliability and validity were tested for new four factors separately. The reliability of the data set was ensured with a Cronbach's Alpha value of more than .7 (Field, 2009, p. 648) the reliability of the instrument was ensured in terms of consistency. Next step of the instrument development was to examine whether the deletion of any items could increase the Cronbach's Alpha value. Exploratory Factor Analysis with Principal Component Analysis should be carried-out to ensuring construct validity. Also to examine whether items in the scale measures this theoretical construct (Ethical Leadership), convergent and discriminant validity have to be ensued. If an item loads significantly <.5 (Field, 2009, p  The Cronbach's Alpha value for the fourteen items in factor 1 was .953. So, there was no item to be deleted and the values of Corrected Item-Total Correlation labeled column were above .6 except item Q5. But it also >.3 which is good. The Cronbach's Alpha value for the five items included in factor 1 was .880. If item Q14 deleted the Cronbach's Alpha value can be increased up to .883. Already the data set was ensured with a Cronbach's Alpha value of more than .7. Therefore no need to be deleted. The Corrected Item-Total Correlation values were above .5. Total Variance Explained 68.19% All five items had strong loadings on the construct and they were supposed to measure uni-dimensionality and construct validity. Total Variance Explained was 68.19%.
Factor 3 The Cronbach's Alpha value for the five items included in factor 1 was .911. There was no item to be deleted. The values in the column labeled Corrected Item-Total Correlation were above .5. The Cronbach's Alpha value for the five items included in factor 1 was .711. There was no item to be deleted and the values in the column labeled Corrected Item-Total Correlation were above .4. Total Variance Explained 63.85% All five items had strong loadings on the construct, they were supposed to measure indicating uni-dimensionality and construct validity. Total Variance Explained was 63.85%.
To ensure more reliability of this measure, Split -half reliability was concerned. This SPSS out-put indicates the all these data were supportive of the reliability of the measurement. Over to ensure the reliability of this measure Average Varian Extracted (AVE) and Composite Reliability (CR) were calculated using following equations. The Average Variance Extracted shows the average percentage of variation explained by the measuring items for a latent construct. AVE >0.5 (Fornell & Larker, 1981) is required for every construct.The Composite Reliability indicates the reliability and internal consistency of a latent construct. A value of CR>0.6 (Fornell & Larker, 1981) is required in order to achieve composite reliability for a construct.  All AVE and CR values included in Table 15 indicate that there is a good reliability of these measures.

AVE= ∑Қ
In order to provide support for discriminant validity, Pearson correlations were computed. For this purpose, composite scores for each factor were calculated by averaging scores representing that dimension. Table 16 shows the significant correlations among the factors. The highest correlation occurred between F1 and F2 (0.791) and reversely, the lowest correlation was found between F2 and F4 (0.554) Bauer,et.al (2006) evaluated their newly developed scales' discriminant validity by utilizing conservative Fornell/Larcker test. Fornell and Larcker (1981) recommended that shared variance (i.e., square of the correlation) among any two constructs should be less than the average variance extracted (AVE) of each factor (Table 17). .791 ** 1 F3 .700 ** .767 ** 1 F4 .499 ** .554 ** .603 ** 1 **. Correlation is significant at the 0.01 level (2-tailed).  AVE vs. SMC significantly shows the discriminant validity of this measurement. Finally, in this scale development process, there was no item deleted and it remained all the 30 -items developed in the qualitative study. But these 30 items loaded to four dimensions. There were identified as people orientation, integrity, ethical guidance and role clarification. The developed new scale was mentioned in Table 19. Gives the promotions/rewards base on the performance 2 He listens very carefully to the ideas of others before make decisions. 3 Delegates responsibilities to his subordinates 4 My boss organized the unit/section suitably for the tasks it has to perform. 5 He respects for employees 6 He has ability to understand subordinates feelings 7 My boss should never too busy to respond to subordinates requests without delay 8 He has best interests of employees in mind 9 Make the fair and balanced decisions 10 Acknowledge mistake of subordinates 11 Conducts his personal life in an ethical manner 12 Keep his actions consistent with his stated values 13 He expresses his opinion honestly and openly to subordinates 14 He is honest and always tell the truth F2 -Integrity

15
He listens to what employees have to say 16 Don't practice favoritism 17 Unbiased when assign tasks to members 18 My boss engage, his work with full confidence 19 His work in a way which naturally enables employees to see what he is doing (transparency) F3 -Ethical guidance 20 Conflict among subordinates handled promptly and effectively 21 My boss gives explanations and instructions to us on a friendly way 22 Communicate the ethical standard for subordinates 23 Oppose the use of unethical practice to increase performance 24 My boss use punishments to hold subordinates responsible for their actions 25 He always use illustrates the ethical behavior in his decision and actions 26 Advice subordinates, how to do things the right way in term of ethics 27 Can be trusted to carry out promises F4 -Role clarification 28 My boss clarifies the responsibilities and his expectations from the employees 29 Discuss the success and failure at work often, and advice how to avoid failures in future 30 My boss gives employees individual attention

Discussion and Conclusion
The corruption, fraud, illegal conduct are considered as unethical behavior of public officers. It is a doubt that different types of unethical behaviour have led to significantly to increased cynicism and skepticism by the larger public (Wijesiri, 2016). It has also ensued in a resurgence of interest in ethics in Sri Lankan public sector. This paper aims to develop a measurement scale to measure ethical leadership of Divisional Secretariats. Divisional Secretariats are the institutes which more than 90 percent of the government related services provided to public. Therefore, the ethical leadership is an important topic that has useful for academic researchers. Measuring ethical leadership is very important since organizations want to know how to select, develop and retain ethical leaders.
To do so scale development steps recommended by Hinkin's (1998) followed. Qualitative study was under taken to develop 30-items were emerged five factors. They are role clarification, people orientation, fairness, integrity and ethical guidance. A quantitative study was done to purify the scale items, examine uni dimensionality, reliability, factor structure and validity. Finally, all 30-item were loaded to 4 factors such as people orientation, ethical guidance, integrity and role clarification. This study contributed to conceptual and methodological advancement of ethical leadership and public sector literature by developing new scale to measure employee perception on ethical leadership of Divisional Secretariats.

Comparison between ELS, ELW and ELQ Measures and New Scale for Public Service
In ELS scale (Annex 1) combined such different behaviors (10 items

Limitations and Future Studies
First, like any survey study, the results reflect correlations rather than causation. It is difficult to handling ethical leadership in field setting (Brown et al.2005).Secondly, the rating of leader ethical behaviors may be prejudiced by a subordinate's general evaluation of the leader. Employees were evaluated their boss according to their perception. The third limitation in the study is use of judgmental sampling technique. It is one of the nonprobabilistic sampling techniques and the use of one of the probabilistic techniques would provide more confidently the chance of generalizing the results. Also the sample size was 100 and it was selected only from Gampaha District. The original questionnaire was translated into Sinhala. Sometimes the real meanings expected from the items should be subject to change. As a closing note, another study can be suggested with a large sample size covering the all island using the newly developed scale to measure the ethical leadership of Divisional Secretariats and repeat the study with other public organizations for further generalizations of the newly developed scale.