!
thesis&
SUPPORTIVE LEADERSHIP IN CONTEXT
:
IS ITS EFFECT ON
JOB PERFORMANCE AND COMMITMENT CONSTRAINED BY
FOLLOWERS
'
PROPENSITY TO TRUST
?
EPMS- TRACK LEADERSHIP AND MANAGEMENT
31-03-2017- FINAL VERSION AUTHOR:ROMAIN BODDAERT STUDENT NUMBER:10899561
“N
OBODY CARES HOW MUCH YOU KNOW,
UNTIL THEY KNOW HOW MUCH!
!
STATEMENT OF ORIGINALITY
This document is written by Romain Boddaert who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents
table!of!contents!
table!of!content!...!4! list!of!figures!and!tables!...!5! abstract!...!6! preface!...!7! introduction!...!9! theoretical!review!&!hypotheses!...!12! job!performance!...!12! supportive!leadership!and!its!effect!on!job!performance!...!13! the!mediating!effect!of!affective!organizational!commitment!...!16! the!moderating!effect!of!propensity!to!trust!...!18! method!...!20! research!method!...!20! sample!...!21! procedure!...!22! measurement!of!variables!...!23! job$performance$...$24! supportive$leadership$...$24! affective$organizational$commitment$...$25! follower$propensity$to$trust$...$25! control$variables$...$25! results!...!26! validity!and!reliability!tests!...!27! hypothesis!testing!...!29! hypothesis$1$...$29! hypothesis$2$...$30! hypothesis$3$...$31! additional$analyses$...$33! discussion!...!38! theoretical!implications!...!39! practical!implications!...!42! limitations!and!suggestions!for!future!research!...!43! conclusions!...!47! references!...!48! appendices!...!53! appendix!A:!questionnaires!...!53! appendix!B:!Sobel!test!results!...!65!list!of!figures!and!tables!
Figure!1:!conceptual!model!...!12! Table!1:!!company!overview!...!21! Table!2:!!means,!SD!and!correlation!table!...!28!! Table!3:!!hierachical!regression:!job!performance!...!30! Table!4:!!hierarchical!regression:!mediating!effect!...!31! Table!5:!!hierarchical!regression:!moderating!effect!...!32! Table!6:!!hierarchical!regression:!trust!in!supervisor!...!34! Table!7:!!hierarchical!regression:!moderating!effect!...!35! Table!8:!!hierarchical!regression!effect:!moderating!effect!on!POS!...!37! Table!9:!!moderating!effects!on!AOC!...!38!abstract!
This study examined the effect of supportive leadership on follower job performance, whether this effect was mediated by follower affective organizational commitment and if these direct and indirect effects were moderated by follower propensity to trust. It suggested that
supportive leadership can have positive contributions to the development of affective organizational commitment and to follower job performance, provided that these followers have higher levels of propensity to trust. A dyadic survey was conducted within 11 companies active in different business environments in the Netherlands and Belgium and provided 72 usable dyads. Data analysis did not result in significant support for either the main effect on job performance, the mediating effect through affective organizational commitment or a potential moderating effect on direct and indirect relationships of propensity to trust. Additional analyses provided support for the positive effects of supportive leadership and trust in supervisor on perceived organizational support, thus adding to our understanding of the antecedents that can predict affective organizational commitment and organizational citizenship behavior. These findings underwrite the notion that particularly in a time when work is increasingly organized in team-based settings, while at the same time employees are expected to work autonomously, supportive leadership can contribute to fostering affective organizational commitment from followers and help them to become valued organizational citizens.
!
preface!
It is with great pleasure and a healthy dose of relief that I am presenting my Msc-thesis to you. This thesis concludes nearly three years of studying Leadership and Management at the University of Amsterdam, a period of time in which I have voluntarily given up part of life-as-I-knew-it, for the benefit of gathering new knowledge, meeting new people and, if only a little bit, reliving those careless days when I was still a full-time student. Looking back, I have not even an inch of regret for having decided to go back to school, yet at the same time I am grateful that it is now completed.
Supportive leadership, the main character of this thesis, is a subject that felt like a natural choice to me. I was convinced that by far the most important element of effective leadership is displaying a genuine interest in followers and optimizing working conditions for them. As time went by I came to realize, as so often was the case during these past 2,5 years, that in reality things are more nuanced. "It depends" is a commonplace of the highest order and at the same time an unavoidable response to every question that wants to explain
if/how/when a certain phenomenon can be effective or beneficial. Admitting that supportive leadership is not the answer to all leadership challenges, this more nuanced stance will undoubtedly help me to be a better manager, even if I will continue to believe that a supportive leader (especially in today's turbulent business times) will in fact promote individual, team and company performance in many occasions.
I am thankful to a number of people that have been important in the process of finishing my study/thesis. First of all, I am very much indebted to Rien Slagter, who in his position as CEO of Five 4 U has always encouraged me and given me the opportunity to develop myself. Secondly, my gratitude goes out to Claudia Buengeler for her work as my thesis supervisor. I have tried to write my thesis as independently as possible, and even though in retrospective I
feel like I have bumped into nearly every foreseeable obstacle that I could possibly encounter on the road, she has always been there to prevent me from a real crash.
Femke, Sil, Kaat and Freek: you are the most important persons in my life and over the last years, you have had to live with a partner or father who was pretty much always studying, or at least tired and distracted in those few occasions when he was not studying. I hope I have not fallen you short too much, and I am so looking forward to spending more relaxing times with you! And last but not least, a final word goes out to everybody else that is either family or that I can hopefully still call a friend: the days when I could forget about your birthdays, not return your phone calls or miss out on parties or other get-togethers with impunity are finally over. I will not promise you immediate recovery just yet (will need to sleep off my study-period for the next 6 months or so), but let's say I am on my way back!
Romain Boddaert Utrecht, February 2017
introduction!
In a business environment that seems to get ever more dynamic, demanding and short-cycled, our workplace is rapidly changing (Van der Smissen, Schalk & Freesen, 2013). These changes occur with important aspects of work, such as compensation, social atmosphere, career opportunities, job security and HR practices (Turnley & Feldman, 1998), and can potentially result in employees feeling less committed to their organization (Nishii, Lepak & Schneider, 2008). The level of this organizational commitment, and in particular of affective organizational commitment, however, is considered to affect in-role and extra-role facets of job performance (Meyer, Stanley, Herscovitch & Topolnytsky, 2002). Yet, despite this consensus on the importance of a committed workforce for organizational success, our understanding of how affective organizational commitment can be deliberately and purposefully managed is still underdeveloped (Morrow, 2011). This study proposes supportive leadership as an appropriate leadership style to foster affective organizational commitment and to improve job performance.
Although a positive effect of supportive leadership on job performance has long been disputed, (Yukl, Gordon & Taber, 2002), this discussion has been settled after more recent publications demonstrated the existence of such a relationship (Judge, Piccolo & Ilies, 2004; Derue, Nahrgang, Wellman & Humphrey, 2011). Yet, a clear understanding of the exact steps through which this positive effect takes place is still missing. There is a general understanding that supportive leader behavior has a positive contribution to both followers' satisfaction and identification with the leader (Bass & Avolio, 1990), as well as it is considered to help establishing and maintaining a supportive work environment (Eisenberger, Huntington & Hutchison, 1986). These aspects will positively impact a follower's perception of the
organization and are likely to lead to followers becoming more affectively committed to the organization. However, while a positive relation between affective organizational
commitment and job performance has been described already (Meyer & Stanley, 2012), there is no definite support for affective organizational commitment's mediating role between supportive leadership and job performance. Hence, the first goal of this study is to replicate the findings of a positive relationship between supportive leadership and job performance, and demonstrate how this effect is at least partially mediated by AOC.
Much like any other leadership style, the effectiveness of supportive leadership depends on many different situational factors (Yukl et al, 2002). Consequently, a second goal of this study is to examine if trust propensity of followers influences the positive effects of
supportive leadership. Although an increase in the level of trust in the leader is considered to be one of the outcomes of effective supportive leadership (Fairholm & Fairholm, 2000), not everybody is capable of trusting others in the same way or to the same degree. The extent to which an individual is inclined to believe in the good intentions of others is reflected by the individual's propensity to trust (Rotter, 1980). According to Colquitt, Scott and LePine (2007), this personality trait denotes a trustor's perception of the good intentions of the trustee. Considering that supportive leadership is directly focused at the follower, providing encouragement, recognition, development opportunities and seeking his/her feedback, it can be argued that a follower's ability to trust others, or lack thereof, will be an influential factor on a supportive leader's effectiveness. The same study by Colquitt et al (2007) also indicated that this perception of benevolence is a significant predictor of AOC. This is important, as higher levels of propensity to trust will generally lead to a more positive perception of benevolence (and vice versa). In this way, considering that followers assume their leaders to be "proximal representatives of the organization" (Meierhans, Rietmann & Jonas, 2008), it can be theorized that the previously described mediating role of affective organizational commitment between supportive leadership and job performance will also be influenced by followers' propensity to trust.
Concluding, I aim to advance our understanding of the context in which supportive leadership can be effective. This study adds to the literature in two different ways. First, it aims to provide additional support for the argument that supportive leadership improves job performance, both in a direct relationship and through the mediating effect of AOC. Secondly, it examines the moderating role of follower propensity to trust on these effects, by answering the following research question:
How does follower propensity to trust influence the effect of supportive leadership on affective organizational commitment and job performance ?
A better understanding of these relationships is important, as the rapidly changing business environment that was described in the first paragraph is demanding not only for employees, but for the leaders of the organization as well. Leaders are no longer dominant bosses that can consider their job a transactional activity that will ultimately return a healthy profit and a satisfied workforce anyway. Instead, they have increasingly become creators and guardians of company culture, organizing individuals and teams around the same norms and beliefs, expecting individuals and teams to organize their work autonomously (Wendt, Euwema & van Emmerik, 2009). Granting that supportive leadership is only one of the instruments in the toolbox that today's leaders will need to be able to use, this study considers it to be an important behavioral style to create affective organizational commitment with followers and improve job performance. I argue that a deeper insight into the relationship between supportive leadership and follower trust propensity will help managers to better understand what behaviors are effective in the specific context in which they are operating. Additionally, improved knowledge about the relationship between supportive leadership and follower trust propensity can be helpful to answer questions in the field of HRM practices, like recruitment, team composition or work design.
theoretical!review!&!hypotheses!
The following conceptual model summarizes the key relationships proposed in this study. In the following sections I will provide a theoretical review of the different elements it contains.
job!performance!
Narrowly defined, job performance indicates whether a subordinate is adequately fulfilling the tasks and responsibilities that are part of the job and company goals. A theory developed by Campbell et al (1996) proposes three determinants of job performance:
declarative knowledge (factual knowledge and knowledge about principles and procedures), procedural knowledge and skill (knowing what to do and how to do it), and motivation (how much effort is put into the things that need to be done). These aspects include all activities that are part of the "in-role" job performance.
Notwithstanding the convention to include formal job requirements in the measure of job performance, this study applies a broader definition of it. Organizational Citizenship Behavior represents all behavior that is discretionary and not directly recognized by the
formal reward system. It is "extra-role" behavior that employees voluntarily engage in, to the benefit of the organization, and consists of five different dimensions: altruism,
conscientiousness, sportsmanship, courtesy and civic virtue (Organ, 1988). There is convincing empirical evidence that supportive leadership has a positive relationship with every dimension of OCB (Podsakoff, MacKenzie, Paine and Bachrach, 2000). With reference to the previously described shift from hierarchical work structures to work design that
focusses on job autonomy and teamwork, organizations become increasingly dependent on the organizational citizenship behaviors of their workforce (Meierhans, Rietmann & Jonas, 2008).
The following sections elaborate on if and how supportive leadership contributes to both the in- and extra-role definitions of job performance.
supportive!leadership!and!its!effect!on!job!performance!
Over the years, leader behavior has been described and classified in many different ways. As indicated by Fleishman, Zaccaro and Mumford (1991), these classifications can in most occasions be broken into one of two categories: they are either person-focused or task focused. Supportive leader behavior was originally introduced in the Path-Goal theory by House (1971), and is seen as part of the person-focused leader behavior classifications. The Path-Goal theory builds upon the Ohio State Leadership Behavior studies (see Stogdill, 1950), which, in a response to the limitations of studying leadership as a trait, focused on behavior and introduced consideration (person-focused) and initiating structure (task-focused) as the two main concepts that comprise leader behavior. Consideration encompasses all leader behavior that is directed at developing and maintaining close relationships with team members. It promotes two-way communication between leaders and followers and aims to develop a cooperation between them that is built on mutual respect and trust (Burke et al,
2006). The concepts of supportive leadership and consideration are very similar, if not identical, and since consideration seems to be used more frequently to construe person-focused leadership studies than supportive leadership in House's Path Goal Theory, I have adopted consideration as the concept to define supportive leader behavior.
Notwithstanding the fact that supportive leadership has been positively linked to job performance (Judge et al, 2004), much remains to be desired about our knowledge about the exact content of this relationship. A prerequisite for obtaining a better understanding of these linkages, which this study aims to do, is to gain a deeper insight into what supportive
leadership is and what it does to followers.
In short, supportive leadership can be characterized as showing a genuine concern for followers. Supportive leaders will be observant to problems that followers might have, will treat followers equally, contribute to their personal development and inspire them to give input. This behavior allows followers to operate in a friendly and psychologically supportive work environment, a context that has been shown to have different beneficial consequences for both followers and, indirectly, the organization itself. Followers that experience supportive leadership will usually be more socially involved with their colleagues, which means there is more potential for employees giving/receiving feedback and for sharing information within the organization. In this way, followers may have a better idea about the formal and informal expectations that colleagues may have for their work and, likewise, for the formal and informal rules and procedures of the organization (Dale & Fox, 2008). If employees have a better understanding of these aspects, they are likely to show both improved in-role job performance and extra-role job performance (organizational citizenship behavior).
Furthermore, in line with Eisenberger et al (2010), employees will positively value the support, personal attention and encouragement they receive from their supervisor. More specifically, because followers have a general tendency to personify the organization
(Eisenberger et al, 1986), their perception of the supervisor will "spill-over" to their
perception of the organization. Hence, supportive leadership is positively related to perceived organizational support (Kim, Eisenberger & Baik, 2016). As has been described in social exchange theory (Blau, 1964), employees will identify themselves with the supervisor and the organization, and will be likely to reciprocate the good treatment they receive from their supervisor and the organization and improve their performance. As a result, this extra effort can then lead to better job performance, both on the level of the formal requirements of the job (in-role job performance) as well as in the area of organizational citizenship behavior (extra-role job performance) like civic virtue and conscientiousness (Podsakoff et al, 2009). It should be noticed that when comparing in-role and extra-role behavior, especially the effect on organizational citizenship behavior seems particularly important in today's rapidly changing work environment. As traditional and more hierarchical work structures are being gradually replaced by autonomous and team oriented job designs, organizational citizenship behavior has become increasingly important for colleagues to meet the requirements of their job and to be able to contribute to organizational success (Meierhans, 2008). These findings are also supported by Oldham and Cummings (1996), who argue that supportive leadership is particularly effective when job incumbents have more complex jobs with high levels of job autonomy. Keeping with the aforementioned reciprocal effect, and considering that many elements of supportive leader behavior can be compared to extra-role behavior, it might be argued that supportive leadership has a stronger effect on organizational citizenship behavior than on in-role job performance. However, this relationship has not been adequately
In accordance with the above argumentation, hypotheses are proposed as follows: H1A: There is a positive relation between supportive leadership and in-role job performance. H1B: There is a positive relation between supportive leadership and extra-role performance
(organizational citizenship behavior).
H1C: The positive relation between supportive leadership and extra-role job performance (organizational citizenship behavior) is stronger than the relation between supportive leadership and in-role job performance.
the!mediating!effect!of!affective!organizational!commitment!
Organizational commitment is a concept that has been intensively studied for many years already. Many of these studies are grounded in the three-component model of Allen and Meyer (1990), which distinguishes commitment from necessity (continuous commitment), commitment from obligation (normative commitment), and affective organizational commitment. Of these three, affective organizational commitment has received the most research attention, which can be explained by the notion that affective organizational commitment is assumed to benefit many different behaviors that are beneficial to the organization. Affective organizational commitment has also been demonstrated to have a larger effect on job performance than the other 2 components of organizational commitment (Meyer et al, 2002). Strong affective organizational commitment means that employees want to continue being a member of their organization because of their own desire, rather than through some cognitive reasoning. Among many different variables that have been shown to influence affective organizational commitment, an important characteristic of some of these variables is that they are connected to a supportive work environment, such as perceived organizational support (Meyer et al, 2002), perceived organizational justice (Colquitt & Murphy, 2001) and organizational citizenship (Meyer et al, 1989). Keeping with its connection to supportive leadership and job performance, especially the relationship of
perceived organizational support with affective organizational commitment seems important to consider in this study. Albeit similar, the 2 variables are considered to be part of different constructs. The most notable difference between the two is that perceived organizational support concerns a cognitive assessment of the support provided by the organization, whereas affective organizational commitment is an affect-laden state of mind (Kim et al, 2016). Rhoades et al (2001) have provided evidence that perceived organizational support acts as a predictor of affective organizational commitment. In accordance with the above, I argue that supervisors play a crucial role in promoting affective organizational commitment in followers. In essence, supervisors (at least partly) define how their followers experience their workplace (Eisenberger et al, 2002). When leaders express concern for followers’ needs and preferences, facilitate the best possible working conditions and develop a supporting climate that builds confidence, these individuals are likely to feel a sense of approval and increased affiliation with both their leader and the organization. Enhanced attraction to the leader will then lead to incorporation of the organizational membership into the employee’s identity (Rhoades, Eisenberger, & Armeli, 2001). Once this affective state is present, it can subsequently lead to followers being more likely to spend extra effort on behalf of the goals of the organization (Mowday et al, 1979; Poon, 2012), which in turn should lead to improved job performance. Indeed, Meyer & Stanley (2012) have already reported a positive effect of affective
organizational commitment on both in-role and extra-role behavior. Summarizing, this
mediating role of affective organizational commitment between supportive leadership and job performance, is examined through the following hypotheses:
H2A: There is a positive relation between supportive leadership behavior and affective organizational commitment.
H2B: There is a positive relation between affective organizational commitment and job performance.
H2C: Affective organizational commitment mediates the relationship between supportive leadership and job performance.
the!moderating!effect!of!propensity!to!trust!
Like any other leader behavior, supportive leadership is no individual activity but involves a relationship between a leader and his/her followers. Successful leadership implies that leaders and followers are able to work together in a climate that is characterized by common values and common vision (Fairholm, 1991). This means that both actors have a good understanding about what is considered to be an appropriate return in the purpose of their relationship. This exchange can be of either a social or an economic nature (Blau, 1964). Economic exchange comprises the explicit returns (wages, job performance, good appraisal) that one can expect from the other, whereas social exchange taps on the implicit levels of reciprocity that have been previously described in the context of affective organizational commitment as well. It considers how feelings of obligation towards the other are instilled through gratitude for and recognition of his/her actions. But without any guarantee that one's input will be equally returned by the other, social exchange brings about a certain level of uncertainty. The central aspect that helps people to deal with this uncertainty is trust. Trust is a variable that can either be considered as a behavioral intention or a personality trait, and is described as the extent to which one person is deliberately vulnerable towards the intentions or actions of another individual (Colquitt et al, 2007). As it has been identified as an
trust is an important aspect to include in a study that aims to contribute to our understanding how a purposeful relationship between leader and follower can result in an improved job performance.
With respect to trust as a personality trait, not everybody is capable to trust other people in the same way and to the same extent. This propensity to trust denotes the general
expectations an individual has about the trustworthiness of other people (Rotter 1967, 1971). It is an aspect of somebody's character that remains more or less stable over time, and is considered to be both genetically determined and based on early life experiences. Propensity to trust is an influential factor, especially when novel situations or new people are involved (Gill, Boies, Finegan, & McNally, 2005). An individual with low propensity to trust does not think other people are generally trustworthy and will probably be more cynical, avoiding and "closed" towards the intentions and actions of others, whereas somebody who easily trusts other people will allow himself/herself to be more vulnerable in the early stages of a new relationship already (Butler, 1999). The above implies that subordinates with low propensity to trust will be less receptive towards a supportive leader, which could subsequently mean that the subordinate will not identify with either the leader or the organization as strongly as would be the case for somebody with a higher propensity to trust. As they would ultimately feel less obligated to return the positive inputs of the supportive leader, I expect the influence of supportive leadership on a follower's job performance to be reduced or even neutralized if an employee has low propensity to trust. Furthermore, in consideration of the previously described argumentation on the positive relation between supportive leadership and a follower's perceptions of the organization, I also assume that trust propensity will affect the relationship between supportive leadership and affective organizational commitment. These assumptions lead to the following (and final) hypotheses:
H3A: The positive relation between supportive leadership and job performance is moderated by a follower's propensity to trust, such that the effect will be weaker if a follower's propensity to trust is lower as compared with higher . H3B: The mediating effect of affective organizational commitment between
supportive leadership behavior and job performance is moderated by propensity to trust, such that the effect of supportive leadership on job performance via affective organizational commitment will be weaker if a follower's propensity to trust is lower as compared with higher.
method!
The following sections provide a detailed overview of the way this study has been conducted. I will consecutively discuss the fundamentals of the approach, the data collection procedures, the operationalization of the constructs and a description of the analysis itself.
research!method!
This study concerns testing a number of hypotheses that were developed based on existing theories in the fields of leadership and management. As such, the study follows a quantitative-deductive approach. This involves testing the relationship between different variables, with the goal in mind to confirm the hypotheses that were developed to answer the original research question. In order to be able to test the relationship between variables, the appropriate data is collected and analyzed using various statistical techniques and procedures, which will be outlined in the following sections.
sample!
The data was collected in a cross-sectional fashion, through a dyadic survey that was administered to supervisors and followers from different teams in 11 different companies. The dyadic nature of the survey was adopted in order to avoid common source bias. The vast majority of participants were sourced either within my employer, or within a company from the network of companies that surrounds my employer. The remaining participants were found through my fellow students or personal network. With this non-probability sampling technique, the respondents have been collected in a broad range of different industries and within different levels of management. A short description of the companies involved in the data is included in table 1.
With a 68% proportion in the sample of followers and a 85% proportion in the sample of supervisors, the sample of respondents was predominantly male. The age among
respondents ranged from 23 to 60 for followers (SDage = 8.13) and from 25 to 60 for supervisors (SDsup-age = 9.98). On average, followers had been a member of their current
Company Name Description #
1 Distributor of Apple computers 35 2 Manufacturer of smart LED-light solutions 8
3 Accounting and tax 4
4 Warehouse Software Developing Company 8 5 ERP system Software Developing Company 5
6 Transporting Company 3
7 RF-scanner equipment distributor and solution provider 1
8 Hospital 1
9 Dutch Ministry of Defense 1
10 Tank Container Operator 4
11 Public Utility Company (energy metering) 2
organization for 7 years (SDtenure = 6.79), whereas for supervisors this was slightly longer: 9 years (SDsup-tenure = 6.51). The average number of years that supervisors and followers had been working together was 4 ((SDtog = 3.89). There were 10 followers (or 14%) and 14 supervisors (or 19%) who had the Belgian nationality, all other participants had the Dutch nationality. As for functional backgrounds, more than 45% of all respondents were active in sales and marketing, 28% were involved in ICT, 8% in logistics, 6% in purchasing, 5% had administrative tasks and the remaining 8% came from other functional backgrounds.
procedure!
In order to be able to find participants, I followed a two-step approach. As a first step, team leaders were approached by email with a general description of the study and a request for their help. As soon as their support was secured, an email was sent to their team members and/or a posting on the company's intranet, explaining the purpose of the study and kindly requesting followers to participate as well. In some occasions, the sending of the letter to team members or the posting on the intranet was carried out by the supervisors because I had no direct access to these team members myself. All communication to both team leaders and team members explicitly stressed the fact that participation was completely voluntarily and that all data obtained would be treated completely anonymously. As for the supervisors, they were allowed to be part of several dyads, as long as they completed a separate survey for each dyad.
The questionnaires were administered through the online research platform Qualtrics. Each questionnaire contained statements and questions about either the participant, the company or the counterpart in the dyad. In order to allow merging the responses from supervisors and followers, both questionnaires from each dyad contained the same unique
code word that supervisor and follower defined before filling in the questionnaires. Sample questionnaires for both leader and follower have been included in appendix A.
As a first step, the data from Qualtrics was exported as two separate SPSS data files (one for each survey). For all cases, I checked whether the code words were spelled exactly in the same way in both versions. Issues with upper case/lower case and obvious typos were corrected in order to ensure a proper match between the code words. Also, in order to function as covariates, I standardized the free-text responses from the variables nationality, functional background and company name and subsequently recoded them into numerical variables. In both data sets, the variables which are automatically created by Qualtrics were removed when they contained no data (such as "first name respondent" or "last name respondent") or, for the eventual purpose of merging the files, were renamed to ensure that only the code word variable would have the same variable name in both surveys. As a next step, the follower and supervisor files were merged, such that each case contained both the responses from
supervisors and followers. Eight cases were dropped because they were incomplete or because there was no unique code word to link supervisor and follower survey, leaving 72 dyads for analysis. There were no missing values in these cases. A total of 29 items were reverse scaled and were therefore recoded into new, scale type items. Finally, all items that pertained to a specific variable were subjected to the Cronbach Alpha reliability estimate. The scores of this test are included in the correlation table that is part of the analysis section.
measurement!of!variables!
All variables of the conceptual model except one have been measured using existing measurements from English publications. As the survey needed to be carried out in Dutch in order to secure a maximum response rate, all English items were translated into Dutch. Considering that every translated item would need to bring across the exact same question as the original English version, these items were back translated by a bilingual person. After
comparing both versions, some discrepancies between the original English version and the back translated version were corrected. One variable, AOC, was measured using an existing Dutch translation of the measurement.
job!performance!
As I am theorizing about the different effects of supportive leadership on in-role and extra-role performance, these sublevels of job performance were separately measured. In order to get the supervisor's perspective on follower job performance, the measure developed by Williams and Anderson (1991), which contains subscales for in-role job performance and OCB was included in the questionnaire for supervisors. This measure contains twenty items in total and was applied using a five-point Likert scale, ranging from (1) completely disagree to (5) fully agree. Sample items are "My follower engages in activities that will directly affect his/her performance evaluation" and "My follower takes time to listen to co-workers' problems and worries". The Cronbach Alpha reliability estimate of in-role job performance was .88, and .80 for extra-role job performance.
supportive!leadership!
Supportive leadership was operationalized by consideration (Stogdill, 1950). It was measured from the perspective of followers, using 10 items from the revised version of the Leader Behavior Description Questionnaire (LBDQ-form XII; Stogdill 1963). Responses were recorded on a 5-point Likert scale, ranging from (1) completely disagree to (5) fully agree. Sample items are "My manager looks out for the personal welfare of individual group members" and "My manager does little things to make it pleasant to be a member of the group". The Cronbach Alpha reliability estimate of this variable was .74.
affective!organizational!commitment!
For affective organizational commitment, the affective commitment scale by Meyer & Allen (1990) was used in the revised and translated version by Jak and Evers (2010). It contains 6 items that were included in the questionnaire for followers. Sample items include "This organization has a great deal of personal meaning for me" and "I do not feel a ‘strong’ sense of belonging to my organization". The Cronbach Alpha reliability estimate of this variable was .82.
follower!propensity!to!trust!
Propensity to trust was measured from the perspective of followers, using the
Predisposition to Trust scale by Mayer & Davis (1999). Sample items of this eight-item scale include "Most repair people will not overcharge people who are ignorant of their specialty" and "These days, you must be alert or someone is likely to take advantage of you". The Cronbach Alpha reliability estimate of this variable was .69, which is at the low end of what is considered to be the minimum score but still deemed acceptable.
control!variables!
Both questionnaires included items to record demographic information that were positioned at the end of the surveys and referred to age, gender, nationality, tenure, company name, years of working together (between supervisor and follower), functional background and level of education. Prior studies have revealed that demographics are potential predictors of my study variables, such as organizational commitment (Avolio et al, 2004) or trust propensity (Bernerth & Walker, 2008).
Furthermore, each questionnaire included a number of additional variables in order to minimize the potential confounding effects that could be present in the relationships proposed in the conceptual model. More specifically, the questionnaire for both supervisors and
followers contained 7 items that measured trust as a state (Robinson, 1996), and the
questionnaire for followers included 7 items that measured leader-member exchange (Graen & Uhl-Bien, 1995), mainly to reduce the potential confounding effect on trust propensity, which is a trait rather than a state. Additionally, the questionnaire for followers contained 10 items that measured initiating structure from the LBDQ XII questionnaire (Stogdill, 1963) to rule out alternative explanations on the effect of supportive leadership; 8 items to measure perceived organizational support (Eisenberger et al, 1986) to reduce confounding effects on affective organizational commitment, 9 items from the Work Design Questionnaire by Morgeson & Humphrey (2006) to measure job autonomy and 4 items from the same publication to measure job complexity. These last variables were included to rule out alternative explanations through situational factors that affect leader behavior effectiveness and stem from the substitutes for leadership theory (Kerr & Jermier, 1978). The Cronbach Alpha reliability estimate of these additional variables are part of the correlation table in the results section.
results!
In this section, I will provide insight into the analysis that was carried out on the 72 cases that the dyadic survey provided. I will explain what analytical procedures were followed and how the different hypotheses were tested. As a preliminary step before the actual
analyses, which were all carried out in the statistical software application SPSS, I took all the items that belonged to one of the variables which were explained in the previous section and computed means for each of them.
validity!and!reliability!tests!
As a first step, descriptive statistics were generated to gain a better insight into the distribution of the data. Results showed how quite some variables, notably trust in follower, perceived organizational support, job complexity and both in- and extra-role job performance were left-skewed. Besides job complexity, for which an average score of above 3 indicates that the average follower perceives the job as complex, it means that respondents had a (very) favorable opinion about their counterpart and the organization. The skewed variables were corrected using a logarithmic transformation. Despite the fact that in most cases this improved the distribution of the data, most variables were still not in compliance with the normality assumptions of parametric tests. Still, the analysis will be conducted through both hierarchical regression and bootstrapping and any potential difference in results will be discussed in the appropriate section of this report.
Based on the statistic for kurtosis of the trust in follower variable, I visually checked the distribution of the responses through a boxplot. One outlier, which looking at the other
answers of that particular respondent may be attributable to a reversed interpretation of the scale, was identified. Based on the fact that it only impacted the main descriptive statistics in the third decimal place, I decided to keep this particular dyad in the analysis.
As a next step, the correlations between the different variables were calculated. As becomes clear from the correlation table in table 2, the main study variables are not correlated. Based on the correlations between the study variables and some of the control variables, a multicollinearity test was conducted. However, testing for confounding effects between supportive leadership and either of the variables initiating structure, LMX, trust in supervisor, trust in follower, affective organizational commitment or perceived organizational support provided VIF-figures that were well below 3 in each instant.
Table 2: Means, SD & Correlations Mean SD 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 1. Company(Name 3,76 3,74 &&& 2. Nationality 1,19 0,40 .,347** &&& 3. Gender(Follower 1,32 0,47 .,077 ,266* &&& 4. Gender(Supervisor 1,15 0,36 .,233* .,014 ,123 &&& 5. Tenure(Follower 7,29 6,84 ,024 .,011 .,126 ,152 &&& 6. Tenure(Supervisor 8,85 6,57 ,226 ,173 .,189 .,315** ,155 &&& 7. Age(Follower 21,07 8,19 ,009 ,177 .,204 ,129 ,587** ,217 &&& 8. Age(Supervisor 26,24 10,17 .,095 ,308** .,158 .,358** ,186 ,431** ,347** &&& 9. Years(of(Working(Together 5,07 3,94 ,035 ,242* .,119 .,195 ,328** ,448** ,268* ,297* &&& 10.Trust(in(Supervisor 4,32 0,57 .,111 .,004 ,092 ,295* ,087 .,063 ,303** .,076 .,088 0.852 11.Trust(in(Follower 4,51 0,62 ,048 .,018 ,066 ,148 ,135 ,008 ,162 .,110 ,044 ,095 0.856 12.LMX 3,73 0,51 .,171 ,096 ,117 ,174 ,177 ,065 ,294* ,063 ,088 ,668** .,038 .751 13.Perceived(Organizational(Trust 3,95 0,70 ,049 ,173 ,064 ,071 ,170 ,072 ,322** ,140 ,063 ,531** .,020 ,454** .884 14.Job(Autonomy 4,08 0,55 ,058 .,009 .,214 .,199 ,382** ,214 ,392** ,295* ,046 ,243* ,081 ,239* ,328** .889 15.Job(Complexity 4,08 0,75 .,020 .,017 .,194 .,215 ,321** ,174 ,287* ,347** ,133 ,118 ,034 ,137 ,140 ,413** .706 16.Supportive(Leadership 4,10 0,48 .,158 ,100 ,246* ,282* ,041 .,153 ,132 .,186 .,067 ,638** ,105 ,678** ,426** ,147 .,123 .739 17.Initiating(Structure 3,86 0,61 .,409** ,320** ,098 ,453** ,038 .,095 ,216 .,009 ,075 ,420** ,090 ,479** ,325** .,089 .,078 ,538** .856 18.Supervisor(Propensity(to(Trust 3,24 0,50 ,027 .,080 ,132 ,201 ,071 ,142 ,100 .,135 .,143 ,131 ,244* .,007 ,049 .,148 .,065 ,178 ,083 .652 19.Affective(Organizational(Commitment 3,87 0,67 .,185 ,010 .,245* ,291* ,326** ,117 ,321** ,066 ,152 ,286* ,017 ,368** ,393** ,246* ,095 ,243* ,344** .,058 .820 20.Follower(Trust(Propensity 3,02 0,50 ,123 ,002 .,051 .,167 ,044 ,283* ,202 ,202 ,094 ,145 ,062 ,074 ,310** ,369** .,039 ,033 ,010 .,026 ,117 .694 21.In.role(Job(Performance 4,22 0,74 .,012 ,025 ,116 ,077 ,244* ,269* ,079 ,068 ,122 ,072 ,415** ,024 .,117 ,018 ,056 ,046 .,092 ,361** ,043 .,121 .875 22.Extra.Role(Job(Performance 4,30 0,46 ,062 .,004 ,108 ,154 ,096 ,244* ,147 .,057 ,084 ,120 ,516** ,090 ,040 .,094 ,042 ,093 ,103 ,360** ,037 .,125 ,409** .796 Variable N=072.0Coefficient0alpha0reliability0estimates0are0on0the0diagonal0in0italics.0*.0Correlation0is0significant0at0the00.050level0(2&tailed).0**.0Correlation0is0significant0at0the00.010level0(2&tailed).
hypothesis)testing)
A hierarchical regression analysis was performed to explore the relationship between the different variables of the conceptual model. For this purpose, all but binary and dependent variables were z-standardized. As a first step, the control variables company name, gender of follower, years of working together, job complexity, trust in supervisor, trust in follower and initiating structure were entered in the regression. Based on the results of the correlation tests, the other control variables that were included in the data-set were dropped. The second step in each regression analysis was to add the respective predictor(s).
hypothesis)1)
Hypothesis 1 states that supportive leadership is positively related to both in-role job performance (1A) and extra-role job performance (1B). Also, a stronger effect on extra-role job performance compared to role job performance is expected (1C). For the effect on in-role job performance, the control variables explained 15,5% of total variance (F [7,64] = 2.864; p = .01), but adding supportive leadership returned insignificant results. With respect to extra-role behavior as an outcome, the regression showed comparable results. Control variables now explained 20,9% of total variance (F [7,64] = 3.672; p =.00), but adding supportive leadership did not significantly increase the amount of variance explained. Thus, hypotheses 1A and 1B are rejected. Hypothesis 1C cannot be assessed without a significant result on 1A and 1B and therefore also rejected.
hypothesis)2)
Hypothesis 2 considers a mediating effect of affective organizational commitment between supportive leadership and both in-role job performance and extra-role job
performance. According to the procedure by Baron and Kenny (1986), a mediating effect is present if a number of conditions are met. First of all, the independent variable should have a significant effect on the dependent variable. Then, the independent variable should also be significantly related to the mediator variable. Third, the mediating variable should have a significant relationship with the dependent variable as well, while controlling for the independent variable. Finally, the independent variable should not predict the dependent variable, while controlling for the mediator. Considering that hypothesis one has been rejected, the first condition of the Baron and Kenny procedure cannot be met. This would render the rest of the analysis irrelevant. Nevertheless, in accordance with Hayes (2009), it is possible for independent variables to exert an effect on a dependent variable though a
Direct'Effects
Step%1 Step%2 Step%1 Step%2
Step%1 Company%Name /.118 /.120 .072 .073 Gender%of%Follower .111 .104 .085 .092 Years%of%Working%Together .151 .152 .066 .065 Job%Complexity .005 .012 .033 .027 Trust%in%Supervisor .129 .104 .052 .077 Trust%in%Follower .417** .416** .493** .494** Initiating%Structure /.254 /.269 .056 .071 Step%2 Supportive%Leadership .048 /.049 R2 .239 .240 .287 .288 R2%adjusted .155 .143 .209 .197 %R2 .001 .001 F 2.864 2.481 3.672 3.180 % F .086 .098 NOTES:' N=72.%Standardized%Regression%Coefficients%are%reported.%%%%%*:%p%<%.05%%%%%**:%p%<%.01 In2role'JP Extra2role'JP
mediating variable, without a direct effect being present. These examples particularly concern multiple influential variables in opposite directions, within or outside the scope of the model. The causal steps approach by Baron and Kenny was therefore completed, but returned insignificant results, as can be inferred from table 4.
Additionally, a Sobel test was conducted to determine if the indirect effect through the mediating variable significantly differs from 0. Both for in-role job performance and extra-role job performance, the test returned insignificant results (in-extra-role job performance: z = 0.46, p = .64, extra-role job performance: z = .16, p = .87). Details for the Sobel test are included in appendix B. Since none of the conditions of the causal steps method by Baron and Kenny have been met, and without a significant result for the Sobel test, hypotheses 2A, 2B and 2C are rejected.
hypothesis)3)
The last hypothesis that was tested is whether follower trust propensity moderates the effects of supportive leadership on both in-role and extra-role job performance, as well as on affective organizational commitment. Follower trust propensity and the interaction term (supportive leadership x follower trust propensity) were added to the regression equation. For in-role job performance, the control variables included in step 1 of the regression explained
Baron&&&Kenny
Condition&1 Condition&1 Condition&2 Condition&3 Condition&3 Condition&4 Condition&4
In#role(JP Extra#Role(JP AOC In#role(JP Extra#role(JP In#role(JP Extra#Role(JP
Company(Name #.120 .073 #.092 #.110 .073 #.111 .076 Gender(of(Follower .104 .092 #.290* .136 .091 .130 .099 Years(of(Working(Together .152 .065 .122 .140 .064 .141 .063 Job(Complexity .012 .027 .031 .004 .033 .009 .026 Trust(in(Supervisor .104 .077 .159 .110 .047 .090 .073 Trust(in(Follower .416** .494** #.010 .417** .493** .461** .494** Initiating(Structure #.269 .071 .209 #.276 .051 #.287 .066 Supportive(Leadership .048 #.049 .099 .039 #.052 AOC .090 .020 .088 .024 R2 .240 .288 .251 .245 .287 .245 .288 R2(adjusted .143 .197 .156 .149 .196 .136 .185 F 2.481 3.180 2.635 2.551 3.168 2.240 2.787 NOTES:& N=72.(Standardized(Regression(Coefficients(are(reported.(((((*:(p(<(.05(((((**:(p(<(.01
15,5% of total variance (F [7,64] = 2.864; p = .01), but no additional variance was explained when follower trust propensity and the interaction term (supportive leadership x follower trust propensity) were added to the equation. Comparable results were reported for extra-role job performance, with control variables explaining 20,9% of variance (F[7,64] = 3.672; p = .00), a result that was not significantly improved by adding follower trust propensity and the interaction term to the equation. Hence, H3A and H3B are also rejected.
Besides the hierarchical regression approach and the Sobel test, the conceptual model was also entered into PROCESS, a statistical analysis tool developed by Hayes (2012) that applies bootstrapping techniques to measure the effects between variables. As a non-parametric test tool, PROCESS is particularly useful for samples with low statistical power (like mine), because normality assumptions do not apply. The analysis was conducted for both in-role and extra-role behavior, using a regular .05 confidence interval and 50000
Moderation
Step%1 Step%2 Step%1 Step%2 Step%1 Step%2
Step%1 Company%Name /.118 /.114 .072 .075 /.088 /.117 Gender%of%Follower .111 .126 .085 .127 /.276* /.255* Years%of%Working%Together .151 .169 .066 .085 .121 .112 Job%Complexity .005 .002 .033 .018 .018 .047 Trust%in%Supervisor .129 .148 .052 .126 .210 .142 Trust%in%Follower .417** .426** .493** .506** /.008 /.011 Initiating%Structure /.254 /.290 .056 .040 .240 .188 Step%2 Supportive%Leadership .024 /.079 .099 Follower%Trust%Propensity /.155 /.172 .092 SL%*%FTP /.102 /.151 /,108 R2 .239 .274 .287 .341 .246 .267 R2%adjusted .155 .155 .209 .232 .164 .147 %R2 .036 .054 .021 F 2.864 2.306 3.672 3.149 2.987 2.221 % F 1.004 1.664 .574 NOTES:0 N=72.%Standardized%Regression%Coefficients%are%reported.%%%%%*:%p%<%.05%%%%%**:%p%<%.01
In2role0JP Extra2Role0JP AOC
bootstrap samples . In each instance, the results were consistent with the initial tests for main effects, mediating effects or moderation.
additional)analyses)
Additional analyses were conducted to get a better understanding of the previously obtained results and to examine potential alternative explanations. These analyses were based on the same data-set and all have an explorative character. All patterns or relationships that could emerge from these analyses would have to be confirmed in follow-up research that follows the appropriate methodology. Unless otherwise stated, the same control variables were used as for the hypotheses tests, plus 2 extra variables that were added because of their correlation with other control variables: supervisor age and supervisor gender. First, I
examined the relationship between supportive leadership and a follower's level of trust in the supervisor. Theoretically, the effect of supportive behavior should fulfill socio-emotional needs of the follower, which in consequence influences the perception of his/her supervisor and, ultimately, builds trust (Fairholm & Fairholm, 2000). The first block of the regression, which included all control variables, accounted for 15,5% of total variance (F [8, 63] = 2.629; p =.02), which could be accounted for by the significant result of initiating structure: (ß = .421, p = .02). An additional 23,10% of explained variance was added when supportive leader behavior was entered into the model (∆R2 = .221, ∆F [1,62] = 27.531; p =.00). Hence, supportive leadership was significantly positively related to trust in supervisor (ß = .600, p = .00), which means an increase in supportive leader behavior by 1 unit would theoretically lead to an increase of trust in supervisor by .600 units.
In line with the setup of my initial conceptual model, I subsequently performed a follow-up regression analysis to see if this effect was perhaps moderated by follower trust propensity. Step 1 of the regression was identical to previous tests and explained 15,5% of total variance (F [8, 63] = 2.629; p = .02). The second step of the test included supportive leadership, propensity to trust and the interaction term [supportive leadership * follower propensity to trust], but despite the fact that the model was still significant and now explained 41,8% of total variance (∆R2 = .228, ∆F [3,60] = 10.497; p =.00), the interaction term
showed no significant result (ß = .059; p = .57). As for the trust propensity variable itself, a significance at a .10 confidence interval was reported, but deemed inconclusive to explore any further. 1 Step%1 Step%2 Step%1 Company%Name .097 .059 Gender%of%Follower .064 =.038 Years%of%Working%Together =.106 =.070 Job%Complexity .224 .222* Trust%in%Follower .021 .001 Initiating%Structure .421** .078 Supervisor%Age =.052 .034 Supervisor%Gender .124 .154 Step%2 Supportive%Leadership .600** R2 .250 .481 R2%adjusted .155 .405 %R2 .231 F 2.629 6.381 % F 27.531 NOTES:( N=72.%Standardized%Regression%Coefficients%are%reported.%%%%% *:%p%<%.05%%%%%**:%p%<%.01 Trust(in(Supervisor
As a next step, I examined the relationship between supportive leadership and perceived organizational support. There is empirical evidence that the positive effects of supportive leadership on a follower's perception of the leader are mirrored in that follower's perception of the organization in general. In this way, followers with supportive superiors may experience more support from the organization (Dirks & Ferrin, 2002). Hence, a regression of perceived organizational support on supportive leadership was conducted, with trust in supervisor now back in the list of control variables. However, whereas step 1 of the model explained 28,4% of total variance (F [9, 62] = 4.133; p < .00), there was no significant additional variance explained when supportive leadership was added to the model in step 2 (∆R2 = .011, ∆F [1,61] = 1.096; p = .30), despite a significant result for the trust in supervisor variable (ß =
2 Step&1 Step&2 Step%1 Company&Name .097 .041 Gender&of&Follower .064 <.047 Years&of&Working&Together <.106 <.063 Job&Complexity .224 .249* Trust&in&Follower .021 <.020 Initiating&Structure .421** .073 Supervisor&Age& <.052 <.022 Supervisor&Gender .124 .190 Step&2 Supportive&Leadership .586** Follower&Trust&Propensity .166 SL*FTP .059 R2 .250 .508 R2&adjusted .155 .418 &R2 .258 F 2.629 5.640 & F 10.497 NOTES:' N=72.&Standardized&Regression&Coefficients&are&reported. *:&p&<&.05&&&&&**:&p&<&.01 Trust'in'Supervisor
.399; p = .00). The fact that supportive leadership is not significant when it serves as a predictor variable together with trust in supervisor, poses the question if these two variables are potentially confounding. A multicollinearity test was therefore performed in SPSS and returned VIF-scores under three. Although there is no "official" number that marks the line between acceptable and inacceptable multicollinearity, this result is generally considered to be no reason for concern. Once more, the follow-up analysis that was performed considered supportive leadership as the independent variable and perceived organizational support as the dependent variable, but now also included trust in follower as a moderator. Step 1 of the regression did not provide significant results (F [8, 63] = 2.001; p = .06), but adding
supportive leadership, trust in supervisor and the interaction variable [supportive leadership * trust in supervisor] explained 31% of total variance (∆R2 = .021, ∆F [3,60] = 7.359; p = .00). More specifically, although a direct effect of supportive leadership on perceived
organizational support cannot be reported, there is, in addition to the effect of the trust in supervisor variable (ß = .485; p = .00), an interaction effect present (ß = .199; p = .05).
Interestingly, this moderating effect of trust cannot be replicated when initiating structure, the "other" behavioral element of Ohio State Leadership Studies, replaces supportive leadership as the predictor variable. This may be attributable to the fact that initiating structure concerns behavior that is associated with effective management, developing clear standards and focusing on goal setting and attainment. This behavior is more connected to perceived organizational competence than support (Kim et al, 2016).
In line with the aforegoing, and considering that perceived organizational support has been identified as an important predictor of AOC (Eisenberger et al, 1986), I examined whether this study can provide support for an argument that supportive leadership contributes to this relationship as well. Therefore, the final step of these additional analyses was to look at a potential moderating effect of supportive leadership and/or trust in supervisor on the
relationship between perceived organizational support and affective organizational
commitment. Control variables explained 16,9% of total variance (F [8, 63] = 2.801; p = .01), and adding POS to the regression equation added another 10,6% of variance (∆R2 = .106, ∆F [1,62] = 10.395; p = .00). No moderating effects of either supportive leadership or trust in supervisor emerged from the data.
3 Step&1 Step&2 Step%1 Company&Name .257 .207 Gender&of&Follower .093 .018 Years&of&Working&Together E.024 .013 Job&Complexity .160 .040 Initiating&Structure .440** .193 Supervisor&Age& .130 .152 Supervisor&Gender .007 E.088 Step%2 Supportive&Leadership .152 Trust&in&Supervisor .485** SL*Trust&in&Supervisor .199* R2 .203 .417 R2&adjusted .101 .310 &R2 .214 F 2.001 3.903 & F 7.359 NOTES:' N=72.&Standardized&Regression&Coefficients&are&reported. *:&p&<&.05&&&&&**:&p&<&.01 Perceived'Organizational'Support
discussion)
The rapid changes in today's business environment can be challenging for both supervisors and followers, and might have detrimental effects for followers' affective
organizational commitment and job performance. Effective leadership is able to prevent this from happening, and this study examined the valuable role of supportive leader behavior in this regard. The study also included a contextual variable which potentially puts a constraint on leader effectiveness: followers' propensity to trust. In accordance with prior research, it
3
Step&1 Step&2 Step&3 Step&4
Step%1 Company&Name 1.045 1.139 1.138 1.093 Gender&of&Follower 1.262* 1.296** 1.302** 1.325** Years&of&Working&Together .135 .143 .142 .127 Job&Complexity .087 .029 .023 .026 Initiating&Structure .225 .064 .069 .078 Supervisor&Age& .049 .002 .015 .056 Supervisor&Gender .278* .275* .274* .266* Step%2 Perceived&Organizational&Support .364** Step%3 Trust&in&Supervisor .002 TiS&*&POS 1.042 Step%4 Supportive&Leadership .121 SL&*&POS .132 R2 .262 .368 .370 .388 R2&adjusted .169 .277 .254 .251 &R2 .106 .001 .019 F 2.801 4.016 3.200 2.832 & F 10.395 .069 .880 NOTES:' N=72.&Standardized&Regression&Coefficients&are&reported.&&*:&p&<&.05&&&&&**:&p&<&.01 Affective'Organizational'Commitment
was expected that, provided that followers are capable of trusting others, supportive leadership has a positive effect on both in-role job performance and extra-role job performance of followers, and that this effect would be mediated by followers' affective organizational commitment. However, the obtained results do not empirically support any of the proposed hypotheses. In the following sections the reasons that were identified as possible explanations for this lack of support are exposed, and some interesting findings that resulted from the additional analyses will be discussed for their theoretical or practical relevance. Limitations and suggestions for further research are provided as well.
theoretical)implications)
Albeit some controversy existed about the proposed effectiveness of supportive
leadership on job performance, this discussion has been settled to the advantage of supportive leadership with the publication of the meta-analysis by Judge et al (2004). Since then, other studies have provided additional evidence for the existence of such a relationship (Derue et al, 2011). Notwithstanding the fact that this current study has not been able to demonstrate a main effect of supportive leadership on either in-role or extra-role job performance, the additional analyses provided some valuable insights that might serve as an interesting proposition for future research.
First, the positive relationship shown in the regression analysis of affective
organizational commitment on perceived organizational support, together with the identified relationship between supportive leadership, trust in supervisor and, again, perceived
organizational support provides some substantiation for the claim by Yukl et al (2002) that supportive leader behavior indirectly contributes to job performance. Followers with
supportive leaders may generalize the treatment of their supervisor to the organization, hence experiencing more support from an organizational level as well. These followers will feel a
stronger obligation to reciprocate the support they perceive from both supervisor and organization and therefore, given the similarities between supportive leadership and
organizational citizenship behavior, are more likely to engage in extra-role job performance. Indirect support for this argument can be found in the study by Shore and Wayne (1993), who demonstrated that organizational citizenship behavior and affective organizational
commitment both have perceived organizational support as a predictor. This would mean that these variables have a common cause, rather than affective organizational commitment acting as a mediating variable between perceived organizational support and extra-role job
performance.
Secondly, the previously mentioned taxonomy by Yukl et al (2002) also proposed that supportive leadership is particularly effective in combination with other leadership behaviors. Admitting there is no universal solution for the many different contexts in which leadership needs to be effective, and with reference to the fact that my additional analyses revealed how both initiating structure and supportive leader behavior have a significant effect on the level of trust in supervisor, it makes sense to consider the role of this latter variable in predicting perceived organizational support and (ultimately) affective organizational commitment. In this regard it is important to acknowledge the work of Schaubroeck, Peng and Lam (2011), who observed that the sublevels of trust, cognition-based and affect-based trust, have both different antecedents and a different outcome. Cognition-based trust is associated with the follower's perceptions of the competence, responsibility, reliability and dependability of the supervisor (McAllister, 1995), whereas affect-based trust stems from the interpersonal
relationship between 2 individuals and the amount of empathy, care and concern each of these individuals has for the other. In this way, cognition-based trust is more connected to initiating structure, while affect-based trust "is more salient for relational aspects of supervisor-follower interactions" (Yang, Mossholder & Peng, 2009) and can therefore be connected to supportive
leadership. As for this study, the applied scale by Robinson (1996) does not distinguish between these two sublevels, and it is therefore not possible to assess how the data relates to this line of reasoning. Nevertheless, the additional analyses proposed a moderating effect of trust in supervisor when a regression of perceived organizational support on supportive leadership was performed, but not when initiating structure was used as a predictor. This could be considered as an indication that affect-based trust was measured in the data-set and reinforces the support for a positive effect of supportive leadership on perceived
organizational support.
Finally, the hypothesized moderating effect of propensity to trust was supposed to provide further insight into the boundary conditions for the effectiveness of supportive leadership. Although an effect of follower trust propensity on perceived organizational support could be reported at the .10 level of significance, there were no true direct or
moderating effects present. This was surprising for 2 reasons. First of all, propensity to trust is considered to be a "driver" of trust from one individual towards another (Mayer et al, 1999) and secondly, results of the analyses have revealed a central role of trust (as a state) between leader behavior and certain desirable outcomes like perceived organizational support and affective commitment. Consistent with Colquitt, Scott and Lepine (2007), these results should have been antecedents for a minor but nevertheless significant positive effect on both task performance (in-role job performance) and organizational citizenship behavior (extra-role job performance). I assume the absence of this relationship is more likely to be explained by the limitations of the data-set than based on any theoretical argumentation. One potential issue could have been the tenure of the followers, as only 14 out of 72 respondents had been with their employer for less than 2 years. Disregarding the argument that trust propensity can still have an impact on trust, even after trustworthiness of the other can be inferred (Govier, 1994), there is a general understanding that the magnitude of trust propensity is stronger when the
