Abstract
We draw on the relevant extant literatures to examine the pathways to building trust through leader behaviors with three distinct emphases: the leader (personal leadership), the leader-follower relationship (relational leadership), and the situation (contextual leadership). We test this model using experimental data collected from experienced managers (Study 1) and field data collected from the peers and direct reports of business executives (Study 2). The results from these two studies both build on and challenge current views in the trust and leadership literatures about how leaders influence trust. Consistent with past literature, our findings indicate that various leadership behaviors appear to directly promote follower trust when analyzed independently. However, when these behaviors are analyzed jointly, relational leadership behaviors were found to mediate the effects of personal and contextual leadership behaviors on follower trust. The implications for theory and practice are discussed.
A leader’s capacity to cultivate follower trust affects his or her ability to direct followers. Follower trust in the leader has been found to facilitate the achievement of organizational goals, the implementation of organizational changes, and the ability to guide organizations through challenging situations (e.g., Barnard, 1938; Jung & Avolio, 2000; Selznick, 1957; Weber, 1947). Although meta-analytic studies have established direct links between a range of leadership behaviors and follower trust, it remains unclear how multiple types of leadership behaviors might have simultaneous and distinct influences on the development of follower trust (Colquitt, Scott, & LePine, 2007; Dirks & Ferrin, 2002). Because leadership behaviors typically do not occur independently, but rather contemporaneously and jointly, a more precise examination of how some leader behaviors might “have stronger relationships with trust than others” (Dirks & Ferrin, 2002, p. 615) is necessary. We examine this issue by theorizing and testing how follower trust can be influenced both directly and indirectly through distinguishable leadership behaviors.
Given the frequent assertion of the causal interdependence between leadership and trust, it would be reasonable to expect a well-developed theory of the relationship between these phenomena. However, as reflected in Kouzes and Posner’s (2002, p. 244) assertion that “without trust you cannot lead,” most of the scholarly and managerial attention has focused on trust as a determinant or precondition of leader effectiveness rather than exploring the effects of different leadership behaviors on trust. In only a few studies of trust has research explicitly and systematically examined how leaders might represent a distinct source of trust. Lind and Tyler (1988; Tyler & Degoey, 1995; Tyler & Lind, 1992), for example, argued that trust in authorities is based not only on the competence of the authority, but also on the exhibition of fair and respectful treatment that foster feelings of relational ties to the person in the authority position, and the legitimacy of the institutional context within which the authority operates. Similarly, in their meta-analysis, Dirks and Ferrin (2002) contended that when leaders possess desirable personal characteristics, demonstrate care and concern for their followers to strengthen relational ties, and make salient how the governing organizational context is dependable and salutary, they can create follower trust. Likewise, Sitkin (1995) argued that how leaders address aspects of the context (e.g., rules, frameworks, systems) can also influence follower trust. Synthesizing these themes, Sitkin and Lind (2007) posited that although personal characteristics of the leader and how the leader frames contextual characteristics for followers play important roles in creating follower trust, the leader’s relationship with his or her followers is crucial in this process.
Building on previous research, we posit that followers decide how much to trust a leader based on information about the leader, about the leader’s stance regarding his or her followers, and about the situation in which the leader acts. Accordingly, to frame our analyses we distinguish between the leader-, relationship-, and situation-focused influences on follower trust. We utilize the in-depth leadership review of Hernandez, Eberly, Avolio, and Johnson (2011) in which they examine the past century of leadership research to delineate a multifaceted perspective of leadership behaviors related to trust. We focus on three paths to follower trust creation based on Hernandez et al.’s (2011) different “loci” (i.e., sources) of leadership and use Sitkin and Lind’s (2007) nomenclature to identify each category of leadership. Originating from the leader locus, personal leadership behaviors are leader-focused leadership behaviors that convey to followers that the leader has personal qualities that merit trust. Originating from the leader-follower dyad locus, relational leadership behaviors are relationship-focused leadership behaviors that facilitate the connections between leaders and followers and demonstrate to followers that leaders will not take advantage if trust is conferred. Finally, originating from the collective and context loci, contextual leadership behaviors are situation-focused leadership behaviors that focus on interpreting the organizational dynamics and environment for followers.
Although each category has been explored by a wide variety of scholars, it is unclear whether all leader behaviors equally influence trust. Leaders are often called upon to exude competence, maintain relationships, and clarify situational priorities; however, it is not clear if leaders need to display all of these behaviors, or if one particular category of leader behavior is more critical than others. Accordingly, to test the relative influence of a multifaceted perspective on leader-follower trust relations, we juxtapose two alternative ways of modeling the leadership-trust link that have not yet been compared empirically in the literature.
Our first set of hypotheses aligns with the bulk of existing trust research and conceptualizes personal, relational, and contextual leadership behaviors as direct influences on follower trust. Trust scholars have demonstrated that a myriad of leader- and relationship-focused leadership behaviors, such as displaying integrity, dependability, concern and respect (Dirks & Ferrin, 2002), predict trust. Scholars have also shown that situation-focused leadership behaviors within groups that affect followers’ understanding of the context (Shamir & Lapidot, 2003; Sitkin, 1995; Williams, 2001) can affect trust. It is important to note that an underlying assumption in much of this past research is that each leadership behavior has an independent effect on trust.
In our second set of hypotheses, we posit that relational leadership behaviors represent a central conduit to follower trust creation. This prediction implies that leadership behaviors that communicate information about the leader (i.e., personal leadership) and situation (i.e., contextual leadership) are filtered through the lens of the leader-follower relationship (Sitkin & Lind, 2007). Indeed, past trust scholars have emphasized the crucial role that relationship-building leadership behaviors such as fairness can play in assisting leaders to establish and maintain trust (Levin, Whitener, & Cross, 2006; Lewicki, Tomlinson, & Gillespie, 2006; Lind, 2001). Building on this work, we propose that relational leadership behaviors mediate the effects of personal and contextual leadership on follower trust.
Our approach is distinctive in that we build on and empirically test both models to gauge how all three categories of leadership behaviors independently and jointly influence how leaders can foster the development of follower trust. In so doing, we answer calls from previous scholars to “empirically examine the mediating processes involved” (Dirks & Ferrin, 2002, p. 621), “identifying the mechanisms” (Colquitt et al., 2007, p. 921) between the leadership antecedents to follower trust. Across two studies with diverse participant samples, our research ultimately supports the centrality of relational leadership in a model of leader influence on follower trust. Our foremost contribution is our finding that relational leadership behaviors can account, in large part, for the effects of personal and contextual leadership behaviors on follower trust.
It is important to note that although past research by Colquitt and colleagues (2007) has examined several independent paths (ability, integrity, and benevolence) to trust creation, because their study was based exclusively within the trust literature, the antecedents to trust did not capture a comprehensive range of leadership behaviors. Their focus on ability, integrity, and benevolence represented only two loci of leadership (see Hernandez et al., 2011): the leader—the primary source from which ability and elements of integrity such as moral character originate—and the leader-follower dyad—the primary source from which benevolence and elements of integrity such as fairness originate. Leadership behavior originating from situational influences, such as collective interactions within social networks or contextual elements such as organizational systems, were not examined. Thus, our article offers a useful extension of Colquitt and colleagues’ study not only by examining the relative influence of different trust antecedents, but also by expanding the range of antecedent leadership behaviors to include situational influences.
Taken together, we demonstrate the centrality of relational leadership behaviors in creating follower trust within a comprehensive model of leadership determinants. Our finding that leadership behaviors do not necessarily affect follower trust independently, but rather that leadership behaviors are often filtered through the leader-follower relationship, enhances our understanding of current theory and practice. Specifically, our research calls for leaders to prioritize relational behaviors to effectively cultivate their followers’ trust. In the following sections, we begin by utilizing Hernandez et al.’s (2011) framework to distinguish between the distinct leadership determinants of follower trust; as such, we examine the multiple ways in which leaders can directly cultivate follower trust. We then empirically test the central function of relational leadership in determining how leader-, relationship-, and situation-focused behaviors can affect follower trust.
Leadership Determinants of Follower Trust
We adopt a broad, inclusive approach to conceptualizing how leadership behaviors can influence trust in order to identify a parsimonious yet sufficiently “comprehensive range of leadership” activities (Avolio, 2004, p. 1564) that can influence follower trust in the leader. Widely accepted definitions of leadership conceptualize the phenomenon as an interpersonal process through which one individual influences the attitudes, beliefs, and behavior of one or more other people (Dirks & Ferrin, 2002). Trust exists within interpersonal exchanges when individuals “accept vulnerability based upon positive expectations of the intentions or behavior of another” (Rousseau, Sitkin, Burt, & Camerer, 1998, p. 395) under conditions of uncertainty and vulnerability (Das & Teng, 1998; Lewis & Weigert, 1985) and interdependence (Sheppard & Sherman, 1998). For example, follower trust has been argued to result when the leader is seen as exhibiting the behaviors that lead his or her constituents to believe that he or she has their best interests at heart (Kouzes & Posner, 1993).
To begin explicating our theoretical model, it is important to first theoretically clarify trust as an outcome of leadership behaviors. Indeed, Mayer, Davis, and Schoorman (1995) posited that trust is determined by certain characteristics of the trustee (i.e., “trustworthiness”). In support of this distinction, meta-analytic findings by Colquitt and colleagues (2007) have shown that trust is a situational state caused by different types of trustee behaviors. We thus delineate the leadership behaviors of the trustee (our independent variables) and test their relative influence on follower trust (our dependent variable).
To build theory on the different paths for trust creation, we examine the vast literature on leadership by utilizing Hernandez et al.’s (2011) organizing framework. Their comprehensive review of the past century of research consists of identifying the loci (i.e., the source of leadership; leader, followers, dyads, collective, context) 1 and mechanisms (i.e., the mode of transmission; traits, cognition, affect, and behaviors) of leadership. Consistent with this framework, we identified leader-, relationship-, and situation-focused leadership behaviors that we classified as personal, relational, and contextual leadership. In the following sections, we discuss each leadership cluster’s direct influences on follower trust; we then delineate how these paths independently and jointly influence follower trust.
Leader-focused leadership behaviors: personal leadership
Leader-focused leadership behaviors have a long history in the leadership literature. This is understandable, considering that the locus of leadership was initially believed to reside solely in the leader (Hernandez et al., 2011). For more than a century, trait approaches examined the personality characteristics of the leader to distinguish between leaders and non-leaders (e.g., Bowden, 1926; Galton, 1869; Gibb, 1947; Jenkins, 1947; Judge, Bono, Ilies, & Gerhardt, 2002; Kirkpatrick & Locke, 1996; Kohs & Irle, 1920; Stogdill, 1948). Fiedler’s contingency theory later examined how leaders with different styles would differ in their behavior, and thus their effectiveness across situations (Fiedler, 1964, 1971, 1976). More recent advances derived from ethical leadership theory (Brown, Treviño, & Harrison, 2005) have also expanded beyond discussing traits to explicate how those traits might serve to orient behaviors and guide decision-making processes. For example, scholars have proposed that leadership behaviors can communicate the leader’s qualities, such as a willingness to adhere to high moral standards, and his or her corresponding personal values (Brown & Treviño, 2006; Ciulla, 2004).
Viewed collectively, this research describes how a leader can influence follower trust by communicating to followers a clear sense of who he or she is as a leader and what he or she brings to the leadership role. Such behaviors can demonstrate a leader’s task expertise, values, and passion for what he or she believes, which can enhance the leader’s perceived capacity to credibly and predictably lead. Accordingly, we classify leader influence on follower trust through sharing information about his or her individual competencies and values as personal leadership (Sitkin & Lind, 2007). Personal leadership provides followers with confidence that the leader is personally capable and committed to achieving the goals he or she espouses.
Following a large body of work on leadership and trust, we hypothesize that personal leadership behaviors affect follower trust because they demonstrate levels of competence and credibility that form the basis for the more calculative aspects of trust (e.g., Doney, Cannon, & Mullen, 1998; Lewicki & Bunker, 1995; Mayer et al., 1995; Williamson, 1993). When the leader authentically projects and reliably behaves in accordance with his or her values, the follower is able to more effectively gauge the leader’s capacity and inclination to act consistently and predictably (e.g., Whitener, Brodt, Korsgaard, & Werner, 1998). Through these actions, leaders establish that they are able to provide their followers with skilled and truthful guidance.
Hypothesis 1: The more that leaders exhibit personal leadership behaviors, the more followers will trust the leader.
Relationship-focused leadership behaviors: relational leadership
Hernandez and colleagues’ (2011) review highlights the variety and breadth of the numerous leadership perspectives that center on the role of the leader-follower relationship. These perspectives range from situational leadership theory that posits that leader effectiveness is contingent on the interaction between leader behaviors and follower level of maturity (Hersey & Blanchard, 1969; 1982), to leader member exchange theory (LMX) which focuses on the quality of the leader-follower interaction (Graen & Uhl-Bien, 1995), to transformational leadership theory that highlights the role of trust in spurring follower motivation to go above and beyond what is normally expected of them (Bass, 1985). Crucially, as illustrated by LMX theorists, leadership arises not from attributes of the leader or the follower alone, but rather from the unique aspects of the leader-follower relationship (Graen & Uhl-Bien, 1995). The centrality of the dyadic interactions between leaders and followers is a key feature to this work and thus the dyad is conceptualized as the locus of leadership (Hernandez et al., 2011).
We build on these insights to identify specific aspects of how interpersonal exchanges between leader and follower can affect trust as relational leadership behaviors. This dimension of leadership is based on notions of leadership that do not emphasize the personal qualities of the leader, but rather the leader-follower relationship. Through relational leadership behaviors, leaders can bolster their followers’ sense that they are valued contributors (Sluss & Ashforth, 2007). Strong relational leaders display both a concern for and willingness to respect the personal needs, interests, and priorities of their followers (Lind & Tyler, 1988; Sitkin & Lind, 2007). By actively displaying concern for their followers and a high level of respect for their actions, these leaders communicate their desire to maintain relationships that will not ignore, exclude, or unduly exploit their followers (Lind, 2001; Van den Bos, Lind, & Wilke, 2001).
We posit that relational leadership behaviors increase follower trust because they signal that the leader will take into account their followers’ objectives and that the leader will likely act in ways that promote follower well-being (Mayer et al., 1995). Specifically, actions that show concern, respect, and fair treatment for followers comprise important social heuristics (Lind, 2001) that, we argue, signal a positive, pro-social relationship between the leader and individual followers. Social heuristics, according to Lind (2001), are impressions that may not logically guarantee good treatment or a strong relationship but that nonetheless are widely used to predict the future actions of another person, especially an authority figure. Social heuristics provide followers with a “good enough” basis for assessing the advisability of acting pro-socially and, hence, facilitate decisions to trust without requiring either a long acquaintance with a particular leader or a deep evaluation of the personal characteristics of that leader (Sitkin & Bies, 1993).
Hypothesis 2: The more that leaders exhibit relational leadership behaviors, the more followers will trust the leader.
Situation-focused leadership behaviors: contextual leadership
Context can significantly influence the ways through which leadership is created and enacted (Hernandez et al., 2011). As exemplified in complexity leadership theory (Marion & Uhl-Bien, 2001; Uhl-Bien, Marion, & McKelvey, 2007), situations are rarely static; rather, the contexts within which individuals lead comprise complex systems that constantly evolve through interactive and dynamic processes. Situation-focused leadership behaviors thus originate from the context itself, comprised of the culture, norms, processes, and systems of an organization, and by collectives within an organization such as the teams and social networks. As such, following the classification of Hernandez and colleagues (2011), the source of contextual leadership behaviors encompasses both the context and collective loci of leadership. Taken together, these situational demands challenge leaders’ capacities to effectively interpret and respond to varying levels of behavioral, cognitive, and emotional complexity (Denison, Hooijberg, & Quinn, 1995; Hooijberg, Hunt, & Dodge, 1997). Consequently, it stands to reason that leaders often devote significant amounts of time and attention to the job of sensemaking (Weick, 1995).
Following this focus, we classify leader behaviors into contextual leadership if they create, define, or explain the situation and set of procedures and practices for individuals, teams or units. Leaders help followers make sense of organizational systems and the social dynamics therein. In exhibiting contextual leadership behaviors, the leader may simplify the context, which in turn provides their followers with a clear understanding of how the unit functions collectively and where it is headed (e.g., Sitkin, 1995; Sitkin & Lind, 2007; Weick, 1995; Weick, Sutcliffe, & Obstfeld, 2005). They put these plans into action by effectively explaining how the work of individuals is being coordinated within the unit using rules, procedures, practices, and norms that delineate accountability and outline the parameters of joint work efforts (Tetlock, 1998).
We propose that leaders create follower trust by assisting followers in navigating often uncertain and complex institutional environments (i.e. legal forms, social networks, and societal norms regarding conflict management and cooperation) to achieve their personal and organizational goals (House, 1971; House & Mitchell, 1974; Rousseau et al., 1998; Zucker, 1986). Indeed, trusted leaders often take pains to provide their followers with guidance on how to effectively and appropriately cope with and control such contexts (Shamir, House, & Arthur, 1993). Contextual leadership thus increases follower trust because it creates a shared situational understanding and a sense of inclusion for followers. That is, leaders generate follower trust by clarifying how and why the organization functions, and helping the follower interpret his or her role within a collective and, more generally, the organizational context. This is important because these leadership behaviors can provide followers with an increased sense of self-control and confidence that they are a meaningful part of an organization seeking to successfully accomplish its objectives (Sabel, 1993; Sitkin, 1995; Weick, 1995).
Hypothesis 3: The more that leaders exhibit contextual leadership behaviors, the more followers will trust the leader.
In addition to describing how three categories of leadership behaviors directly and independently enhance follower trust, our second research objective is to theorize how these behaviors might simultaneously influence trust development, and how we might model those effects more accurately. Hence, we turn our attention to outlining a hypothesized set of dynamics among the three clusters of leadership behaviors and how they jointly influence the development of follower trust. Specifically, we explore how follower perceptions of their leader’s relational leadership behaviors can mediate the relationships between follower trust and personal and contextual leadership behaviors.
The Mediating Role of Relational Leadership
The leader-follower relationship plays a central role in determining how followers experience the organization (e.g., Gerstner & Day, 1997; Lord, Brown, & Freiberg, 1999; Seers & Graen, 1984). We draw on this basic insight from the literature to propose that a crucial mechanism for trust creation is how followers perceive leader-follower relationships. These relationships determine how both a leader’s perceived personal credibility and competence, and a leader’s capacity to effectively interpret aspects of the situation, can affect follower trust. Although there are good reasons to predict that each of our three clusters of leadership actions will directly and independently enhance follower trust, we test the role that relational leadership behaviors play in forming a central conduit through which trust in leaders is formed.
Following the logic and tone of the social heuristics argument (Lind, 2001), we reason that followers may well base most of their trust in a leader on relational interactions. Rather than basing trust on a direct assessment of the leader’s competence, or his or her command of a situation, followers employ social heuristic judgements to quickly arrive at a rough assessment of the quality of their relationship with the leader. If this assessment, via heuristics like concern, respect and fairness, is positive, then the leader is more trusted; if this assessment is not positive, then the leader is less trusted.
In support of this proposition are meta-analytic findings (Colquitt et al., 2007; Dirks & Ferrin, 2002) which have consistently shown that specific types of relationally focused leader behaviors are strongly related to trust development. For example, consistent with Dirks and Ferrin (2002), we contend that issues of fairness and relational support are key antecedents of follower trust. Five of the seven factors that these scholars classified as primary antecedents to trust in leaders are subsumed within our relational leadership dimension. Specifically, they identify three fairness dimensions (i.e., distributive, procedural, and interactional), describe participative decision-making, and highlight leader behaviors that demonstrate sensitivity to follower needs and follower interests.
Additional support for these ideas is provided by Tyler (2003) who observed how the relational leadership behaviors we describe more strongly predict the extent to which individuals accept (i.e., trust) the decisions of their authorities than the extent to which they perceive those authorities to be competent or to act predictably. Sitkin and Stickel (1996) explain these dynamics by suggesting that relationship-based behaviors (e.g., sensitivity to others’ needs, a sense of understanding) positively influence trust creation by serving as a filter through which behaviors that are more directly related to personal characteristics (e.g., individuals’ competence, skills, values) are understood. Accordingly, we posit that relational leadership behaviors serve as the mechanism through which personal leadership behaviors influence follower trust creation.
Hypothesis 4: Relational leadership mediates the relationship between personal leadership and follower trust.
Relational leadership can also account for the link between contextual leadership and follower trust. Specifically, if contextual leadership actions are undertaken by leaders who do not exhibit relational leadership behaviors, attempts to institute key organizational design elements (e.g., structures, processes, procedures) may be viewed as coercive institutional control mechanisms, rather than enabling structures that mitigate risk and uncertainty (Adler & Borys, 1996; Long & Sitkin, 2006; Sitkin, 1995). That is, when leaders do not effectively develop follower perceptions of them as concerned and caring leaders, leader behaviors intended to assist followers in navigating organizational systems will be more likely seen as attempts to unduly restrict and exploit followers, perceptions that fundamentally undermine trust (e.g., Granovetter, 1985; Shapiro, 1987; Sitkin & Bies, 1993; Sitkin & Stickel, 1996; Yudof, 1981; Zucker, 1986).
Accordingly, although contextual leadership behaviors that mitigate uncertainty by establishing norms, processes, and structures can directly facilitate follower trust, we posit that trust is created when leaders enact contextual elements in ways that demonstrate that they care about their followers and understand their followers’ needs. Leaders’ displays of concern, respect, and fairness make it more likely that followers will accept or adopt the leader’s definition of the organization’s structural and social dynamics. Consistent with this rationale, we propose that relational leadership behaviors represent the mediating mechanism through which contextual leadership behaviors influence follower trust.
Hypothesis 5: Relational leadership mediates the relationship between contextual leadership and follower trust.
Overview of Empirical Tests
We examined the leadership determinants of follower trust in two studies. In Study 1, we examined the causal ordering of these relationships in a controlled experiment administered to experienced workers. We manipulated personal, relational, and contextual leadership behaviors in order to gauge their direct and indirect effects on follower trust. In Study 2, we evaluated these relationships using both peer and direct report ratings from a 360-degree field survey administered to business executives. Specifically, leaders’ peers rated their personal, relational, and contextual leadership behaviors while their direct reports rated the extent to which they trust their leader. Our results from these two studies show that although all three categories of leadership behaviors directly influence follower trust, relational leadership behaviors mediate the relationships of personal and contextual leadership on follower trust.
To effectively test Hypotheses 1 to 5 in these two studies, we first examined the factor structure of leaders’ personal, relational, and contextual leadership behaviors. To do this, we employed exploratory factor analysis (EFA) on data obtained from a sample of 848 direct reports who assessed 27 item-based descriptions of their leaders’ behaviors. All items were measured with 5-point Likert scales. Missing values were replaced by mean values.
Three factors obtaining eigenvalues over 1.0 were retained (alpha extraction; equimax rotation); two items were dropped because they loaded onto multiple factors. Two additional items were dropped because they did not load onto a factor at a level greater than .40. The items that were retained formed three factors that measure personal, relational, and contextual leader behaviors and obtained Cronbach alphas of .87, .90, and .91, respectively. Table 1 shows the results of the factor analyses and reliability tests on this data.
Exploratory Factor Analysis of Leadership Dimensions a .
Loadings over .40 are reported; retained items are in bold.
Personal leadership
Personal leadership behaviors convey an individual’s unique personal qualities as a leader. The items we used in these scales measured the extent to which a leader was seen as projecting these personal leadership qualities. Eight items tapped raters’ thoughts about the leader’s passion (e.g., “(__) is passionate about the work the unit does”), expertise (e.g., “(__) thoroughly understands the unit’s work”) and authenticity (i.e., “(__) lives his/her values”). The composite scale yielded a high level of reliability (Cronbach alpha = .87).
Relational leadership
When demonstrating relational leadership behaviors, the leader demonstrates concern for and respect to followers, and carries out fair practices. Seven items tapped into rater perceptions of leaders’ relational behaviors, including displays of concern (e.g., “(__) displays concern for those he/she leads”), respect (e.g., “(__) shows respect for people regardless of their level in the organization”), and fairness (e.g., “(__) is unbiased in making decisions”). The scale reliability of this measure was high (Cronbach alpha = .90).
Contextual leadership
Contextual leadership behaviors help followers to make sense of organizational structures, processes, and policies. Eight items described how leaders create a sense of coherence (e.g., “___makes clear how responsibilities are being divided”) for followers, simplifies and focuses their attention (e.g., “(__) cuts through complex or ambiguous problems to make them easier to understand”) and helps followers to effectively coordinate their activities (e.g., “(__) creates processes that facilitate work”). This scale obtained a high reliability score (Cronbach alpha = .91).
Study 1
Participants
In Study 1, we examined Hypotheses 1 to 5 in an experimental setting; 131 males and 158 females took part in this study. The work experience of participants ranged between 1 and 58 years, with a mean of 24.5 (s.d. 11.7) years. Their ages ranged from 22 to 73, with a mean of 45 (s.d. 11.5) years.
Methods
We used Qualtrics, a third-party online survey administration company, for our data collection efforts. Qualtrics is a company that provides researchers with a population of over one million individuals over 18 years old. These individuals have previously registered with Qualtrics, provided Qualtrics with their demographic information, and agreed that Qualtrics could contact them about survey opportunities for which they would be qualified. Once a study is initiated, Qualtrics sends information about a survey to potential participants. When an individual successfully completes a survey, Qualtrics compensates these participants with “survey cash,” which can be converted into monetary compensation after they complete a minimal number of research studies.
Based on the selection parameters we specified (i.e., individuals who work over 20 hours per week), Qualtrics estimates that our initial solicitation was received by 1,000 potential respondents. Of these, roughly 400 potential participants who met our employment restriction of working at least 20 hours per week received and viewed our solicitation. From this sample, 289 individuals completed our study, which produced a 72% response rate.
Experimental approach
Each participant was randomly assigned to one of eight conditions that described a set of behaviors of a leader of a hypothetical work team within a 2 × 2 × 2 between-subjects experimental design. The leader in each scenario directed a work team to which the participant belonged. We manipulated leader behaviors using three sets of vignettes describing one of two levels of personal leadership (high or low level), relational leadership (high or low level), and contextual leadership (high or low level). The text of these scenarios is provided in Appendix A.
All participants were first given some general background information about the organization (e.g., “Your company exists in a competitive industry but your organization is performing reasonably well”), their job function (e.g., “Your job is to examine and write reports about companies within the retail sector”) and their boss (e.g., “Your immediate boss is Pat Seeley. You have worked with Pat for a year and a half”). This was common information that each participant received. Participants were then told that their boss, Pat, hosts bi-weekly meetings and that each member of their 12-person unit attends these meetings. Participants were instructed to imagine they were part of this unit and to reflect on their job, boss, and the organizational situation. Next, participants were given narratives of actions that Pat takes in leading the work unit. Each narrative described the extent to which Pat engaged in personal, relational, and contextual leadership behaviors. Each vignette was pre-tested using a 65-person sample comprising students enrolled in an MBA program in the Southeast USA. These data confirmed that the narratives accurately conveyed the three dimensions of leader behaviors used as independent variables in our study.
When participants finished reading their study scenarios, they were asked to fill out evaluations of the leader’s personal, relational, and contextual leadership behaviors and provide a general assessment of their leader’s capabilities. Participants were then asked to judge the level of trust they would have in the leader described in the scenario. The entire study took participants approximately 15 minutes to complete.
Measures
Across all of the items, participants responded on seven-point Likert-type scales with anchors “1 = not at all likely,” “4 = somewhat likely,” and “7 = extremely likely.” We standardized all variables before analyzing them.
The measures of personal, relational, and contextual leadership we use in this study were consistent with those we developed through the factor analysis (a full list of items is provided in Appendix B). In each case, the wording of these items was only slightly modified to reflect the content described in the scenarios. The Cronbach alphas for personal, relational, and contextual leadership were .93, .97, and .95, respectively.
Follower trust
Our measure of trust is consistent with the notion that trust necessitates a willingness to be vulnerable based upon positive expectations of the intentions or behavior of another (e.g., Mayer & Gavin, 2005; Rousseau et al., 1998); thus, our trust scale captured trust-related risks identified by Sheppard and Sherman (1998). Indeed, past qualitative (Bijlsma & van de Bunt, 2003) and experimental (Malhotra, 2004) investigations have adopted a similar view of trust, centered on trust as related to risk and used a global measure of trust to capture this construct (e.g., Brockner, Siegel, Daly, Tyler, & Martin, 1997; Huff & Kelley, 2003). The following five questions measured follower trust in the leader (Cronbach alpha = .88): “I trust that Pat will honor confidences,” “I expect Pat will live up to his commitments,” “I feel like the team can always rely on Pat’s word,” “I would depend on Pat’s openness with information and ideas,” and “I expect Pat will share credit for successes.”
The fit statistics we obtained suggest that the factor structure which includes three leadership dimensions and our measure of follower trust fit with the data reasonably well: X2/df = 2.94, CFI = .93, IFI = .92, RMSEA = .082. In addition, chi-squared difference tests confirm that our three-factor model of leadership dimensions fit the data better (at p < .001) than the one-factor model (X2/df = 4.51, CFI = .86, IFI = .86, RMSEA = .111).
Results
Manipulation checks
We checked our manipulations in two ways. First, using independent t-tests we found that participants in the high personal leadership condition rated their leaders as exhibiting a higher level of personal leadership than participants in the low personal leadership condition (mean difference = 1.32, p < .001). This same pattern was consistent across relational leadership conditions (mean difference = 2.68, p < .001) and contextual leadership conditions (mean difference = 1.86, p < .001).
Second, because leadership dimensions were highly correlated, we also used paired samples t-tests to assess the extent to which scenario conditions produced higher levels of perceived leadership on targeted dimensions (e.g., personal leadership) when compared with perceptions of alternative forms of leadership (e.g., vs. relational and contextual leadership). Our manipulation checks confirmed that when participants encountered the high personal leadership condition, participants’ perceptual ratings of Pat’s personal leadership were higher than their ratings of Pat’s relational leadership (mean difference = .83, p < .001) and contextual leadership (mean difference = .61, p < .001). These tests also confirmed that when participants encountered the high relational leadership condition, participants’ perceptual ratings of Pat’s relational leadership were higher than their ratings of Pat’s personal leadership (mean difference = .53, p < .001) and contextual leadership (mean difference = .99, p < .001). Finally, our manipulation checks confirmed that when participants encountered the high contextual leadership condition, participants’ perceptual ratings of Pat’s contextual leadership were higher than their ratings of Pat’s personal leadership (mean difference = .18, p < .01) and relational leadership (mean difference = .60, p < .001).
Results
We present the descriptive statistics and correlations for this study in Table 2. Of substantive interest, follower trust is associated with the leadership antecedents in the manner that we hypothesized. That is, personal (r = .79, p < .001), relational (r = .73, p < .001), and contextual (r = .69, p < .001) leadership behaviors are significantly and positively related to follower trust, supporting Hypotheses 1 to 3.
Study 1: Correlations among Predictor and Criterion Variables.
p < 0.001.
Next, we tested whether relational leadership mediates the effects of personal and contextual leadership behaviors on follower trust (Hypotheses 4 and 5) using the SPSS module, INDIRECT (Preacher & Hayes, 2004, 2008). INDIRECT was chosen for this analysis because it allows mediators to be simultaneously evaluated with a bootstrapping technique that generates 95% confidence intervals to test the probability that mediation has occurred. As compared to Baron and Kenny’s (1986) mediation approach and the Sobel test, the Preacher and Hayes non-parametric bootstrapping method has become increasingly utilized because it permits a formal estimation of indirect effects and prevents the violation of normal distribution assumptions. We find support for our mediation hypotheses (Hypotheses 4 and 5) when we observe two results. First, that the coefficients of the association between the independent and the mediating variable and between a mediating variable and the criterion variable are significant. Second, that zero is not included in the 95% confidence interval that is generated to examine the strength of a mediated relationship (Hayes, 2013, 2014).
Accordingly, we observe that personal (β = .55, p < .001), and contextual (β = .19, p < .01) leadership behaviors are both significantly related to relational leadership behaviors. In addition, we observe that personal (β = .33, p < .001), contextual (β = .21, p < .001), and relational (β = .37, p < .001) leadership behaviors are positively related to follower trust. Using bootstrapping techniques to conduct the mediation analyses, our evaluation of the confidence intervals examining these relationships demonstrates that relational leadership behaviors mediate the influence of personal (95% CI .135, .289) and contextual (95% CI .019, .138) leadership behaviors on follower trust. These results are displayed in Table 3.
Study 1: Analyses Examining Relational Leadership Behaviors Mediating the Effects of Personal and Contextual Leadership Behaviors on Follower Trust (N = 289).
p < .01, ***p < .001.
Test of personal and contextual leadership predicting relational leadership.
Test of relational leadership mediating the effects of personal and contextual leadership on follower trust.
Study 1 Discussion
The results of Study 1 offer initial support for Hypotheses 1 to 5. By testing the effects of different types of leadership behaviors within leader-follower relations in a controlled scenario experiment, we were able to establish the direction of causality in our model. Study 1 also independently manipulated leader behaviors, thereby teasing apart their differential effects on follower trust.
Despite its strong support for our theory, this initial test of our hypotheses had limited external validity in that we examined followers’ perceptions of their leader’s behaviors in an experimental context. Although leader behaviors were independently manipulated, the absence of a tangible work context and repeated interactions with a leader could have limited the inferences drawn by participants. In our mediation test, for example, personal and contextual leadership behaviors maintain significant direct effects on follower trust (see Table 3). Although Preacher and Hayes (2004, 2008) argue that mediation can be established even if the relationship between independent variables and the dependent variable remains significant with the mediator in the regression equation and that the absence of zero between the upper and lower limit of a bootstrapped 95% CI is what determines mediation, a residual effect remains. We argue that this could be due to the inherent limitations participants have in interpreting leader behaviors from an artificial context. Thus, to increase our methodological rigor and the generalizability of our results, we wished to directly examine the untested external validity of our Study 1 findings. In Study 2, we sought to mitigate these limitations by examining how the behaviors of practicing managers influence the level of trust placed in them by their direct reports. Accordingly, we used assessments of actual leader behaviors to examine the extent to which they jointly and independently influence follower trust.
Study 2
Participants
This study examined the actions taken and outcomes produced by 266 practicing managers enrolled in a leadership class (87% attended the leadership course as part of their executive MBA program and 13% were enrolled in a non-degree version of this course). At the time the survey was administered, participants had been employed from 3 to 28 years, with a mean of 10.3 years; 82.7% of participants were male. Their ages ranged from 27 to 56, with a mean of 35.3 (s.d. 5.8) years of age; 76% described themselves as being from the United States and 24% resided in other countries.
The study was conducted as a part of a 360-degree feedback survey (SDLS, Lind & Sitkin, 2007) that participants completed in advance of the course. To ensure that participant responses were not contaminated by course lessons or readings, participants were contacted prior to the beginning of the course and were asked to submit the names of peers and direct reports who could provide unbiased, accurate assessments of their leadership approach and the experience working with them. Each participant’s peers and direct reports were contacted and were required to complete their evaluations before the course began. Although participation in the survey was mandatory for class attendees, allowing survey data to be used for research was not and five students opted not to allow their data to be used for research purposes, so their results were excluded from all analyses.
Measures
Personal, relational, and contextual leadership
The questionnaires used five-point Likert-type scales, with anchors at 1 (strongly disagree) and 5 (strongly agree). We used items that were consistent with those obtained through the factor analysis procedures described earlier but were modified to reflect the perspectives of referents (i.e., peers or direct reports). The Cronbach alphas for these personal, relational, and contextual leadership scales were .80, .92, and .93, respectively.
Follower trust
In addition, we measured follower trust using the same five-item scale used in Study 1. The items used were modified slightly to reflect the perspective of participants’ direct reports. For example, whereas in Study 1 the behaviors referred to “Pat” as the leader, in Study 2, the online 360-degree survey program utilized the leader’s name to customize the questions to each participant. The Cronbach alpha for this scale was .88.
Analyses
In Study 2, we used multiple sources of data for our independent and dependent variables, which allowed us to provide a more rigorous test of our theoretical model. Due to evidence in the social psychological literature which suggests that leader self-perceptions are often inaccurate (e.g., Bem, 1972; Lewis & Brooks-Gunn, 1979) and susceptible to egoistically biased judgements (e.g., Messick, Bloom, Boldizar, & Samuelson, 1985), we relied on ratings from their peers on the extent to which leaders engaged in personal, relational, and contextual leadership behaviors. In total we obtained 876 peer reports. A median of three peers rated each course participant. In addition, a total of 729 direct reports rated the extent to which they trusted their manager. A median of three direct reports reported on the extent to which they trusted their managers. Merged data from managers’ peers and direct reports resulted in a final sample of 206 managers.
To conduct our analyses, we first averaged the personal, relational, and contextual leadership ratings that were reported by each leader’s peers. We obtained rWG, ICC1, and ICC2 measures for personal (rWG = .95; ICC1 = .13; ICC2 = .33), relational (rWG = .94; ICC1 = .23; ICC2 = .49), and contextual leadership (rWG = .93; ICC1 = .15; ICC2 =.37). Differences in ratings of personal, contextual, and relational leadership behaviors between peer groups exceeded within peer-group differences (i.e., at p < .001). Our analyses also required that we aggregate direct report ratings of the extent to which they trusted their managers. The rWG, ICC1, and ICC2 values for follower trust (rWG = .96; ICC1 = .20; ICC2 = .54) and analyses indicating that the differences in follower trust ratings between direct report groups exceeded within direct report-group differences (i.e., at p < .001) suggested that aggregating these values was appropriate.
Further, we conducted a confirmatory factor analysis (CFA) to examine the overall factor structure of personal, relational, and contextual leadership measures. The fit statistics we obtained suggest that the factor structure describing three dimensions of leadership behaviors provides a good fit with the data: X2/df = 1.90, CFI = .94, IFI = .94, RMSEA = .058. In addition, chi-squared difference tests confirm that this three-factor model fit the data better (p < .001) than the one-factor model (X2/df = 2.29, CFI = .92, IFI = .91, RMSEA = .070).
Results
The means, standard deviations, and correlations between these variables are presented in Table 4. We examined the correlations between key variables to assess whether positive relationships exist between personal, contextual, and relational leadership behaviors and follower trust. Our results indicate that personal (r = .30, p < .001), relational (r = .36, p < .001), and contextual (r = .34, p < .001) leadership behaviors are positively related to the extent to which followers trust their leaders. These findings support Hypotheses 1 to 3.
Study 2: Correlations among Predictor and Criterion Variables.
p < 0.001.
As with Study 1, we examined Hypotheses 4 and 5 using the SPSS module, INDIRECT (Preacher & Hayes, 2004, 2008). In line with our predictions, our analyses indicate that relational leadership behaviors mediate the effects of personal and contextual leadership behaviors on follower trust. Specifically, in addition to the direct, positive effect of relational leadership behaviors on follower trust (β = .21, p < .05), we observe that personal (β = .29, p < .001) and contextual (β = .46, p < .01) leadership behaviors are both significant predictors of relational leadership behaviors. Our examination of the confidence intervals around these relationships confirm that relational leadership behaviors mediate the influence of personal (95% CI .005, .205) and contextual (95% CI .007, .170) leadership behaviors on follower trust (Table 5).
Study 2: Analyses Examining Relational Leadership Behaviors Mediating the Effects of Personal and Contextual Leadership Behaviors on Follower Trust (N = 206).
p < 0.001; ** p < 0.01; * p < .05.
Test of personal and contextual leadership predicting relational leadership.
Test of relational leadership mediating the effects of personal and contextual leadership on follower trust.
Study 2 Discussion
As predicted, the results of our field test suggest that relational leadership behaviors mediate the effects of personal and contextual leadership behaviors on follower trust. Consistent with the results of Study 1, which manipulated leader behavior within an experimental context and showed that relational leadership can account for a significant portion of these effects, the results of Study 2 lend strong support for the mediating effect of relational leadership. Hence, even though there are significant direct, independent influences of personal, relational, and contextual leadership behaviors on follower trust, our findings demonstrate that relational leadership behaviors are a central channel through which follower trust is created by leaders.
General Discussion
We theorized about the direct and indirect effects on follower trust of a variety of leadership behaviors that had been considered in past research. We examined the direct predictions of leader effects used in prior empirical research and modeled them simultaneously to allow for a more comprehensive assessment of the causal paths that influence leader-trust relations. Our findings across both experimental and field studies indicate that personal, relational, and contextual leadership behaviors have direct influences on follower trust. However, when modeled jointly, our results suggest that leader behaviors that communicate competence and other personal attributes, as well as leader structural or context-building leadership behaviors, influence follower trust through the relational features of follower treatment, such as leader behaviors that demonstrate concern, respect, and fairness. As such, this research presents important implications for both theory and practice.
Theoretical Implications
Our finding, that relational leadership behaviors mediate the effects of personal and contextual leadership behaviors on follower trust, challenges some of the received wisdom of the trust literature where researchers have largely emphasized the independent effects of various behaviors on follower trust (cf. Dirks & Ferrin, 2002). By examining the flow of causality from separate independent variables to follower trust, we advance a more nuanced understanding of leader-follower relations than explored in prior research. That is, our findings not only show that different pathways to trust exist, as predicted in the trust literature, but also empirically demonstrate that one pathway represents a central conduit to trust creation. The results demonstrate the centrality of relational leadership behaviors, underscoring the significance of interpersonal exchanges in the development of follower trust. These findings, therefore, lend strong support to the social exchange perspective of trust in leadership where followers perceive their relationship with their leader as holding mutual obligations, reciprocal communication, and displays of care and consideration (e.g., Blau, 1968; Graen & Uhl-Bien, 1995; Whitener et al., 1998).
Moreover, the current research deepens our understanding of the Mayer et al. (1995) model. We found that behaviors related to ability and the consistency-based part of integrity (i.e., personal leadership) are indeed important determinants of trust creation when examined in isolation. We also found that behaviors associated with benevolence and the fairness-based part of integrity (i.e., relational leadership) can account for this effect. These results imply that the various facets of trustworthiness once posited to similarly affect trust could have differential effects. We also advance the field beyond a focus on the personal and relational behaviors of leaders that might arise from different characteristics of the trustee, to build theory on leader behaviors associated with the situation.
Specifically, we shed light on the possible linkages between leader-follower dyadic interactions and the situation, answering the numerous calls to further examine the influence of organizational context on leadership (e.g., Johns 2006, Klein & Kozlowski 2000, Yammarino & Dansereau, 2009). Hernandez and colleagues (2011) asserted that the context plays an active role in shaping how leadership emerges. Through both proximal influences, such as a leader’s focus on increasing efficiency, or distal influences, such as the cultural composition of leaders and followers, “context can shape leadership by redefining the social and task-oriented interaction among individuals” (Eberly, Johnson, Hernandez, & Avolio, 2013, p. 9). In the current research, we have theoretically modeled and empirically tested one particular type of social interaction: interpersonal exchanges that create follower trust. Our findings empirically demonstrate that contextual leadership behaviors, which can interpret and communicate both proximal and distal contextual characteristics to followers, directly influence these exchanges.
More broadly, our study has implications for the relationship between trust and the institutional environment. There is a substantial history of research by sociologists examining the effect of institutional routinization on trust (e.g., Fox, 1974; Shapiro, 1987; Zucker, 1977). This work has found that the adoption of rules, procedures, and routinized cultural practices tends to have a negative effect on trust. In contrast, management scholars (e.g., Bachmann & Inkpen, 2011; Shamir & Lapidot, 2003; Sitkin, 1995; Sitkin & George, 2005) have found that the explanation of formal rules and procedures can have either a positive or negative effect on trust, depending upon specific mechanisms or contingencies. Our research suggests that this earlier sociological work should be reexamined to reflect that both contextual features and explanations of those features may have a different effect on trust when combined with relational leadership behaviors. Specifically, when leaders exhibit relational leadership behaviors, they may be able to reverse the trust-eroding effects of formalization found in the sociological literature and mediate the positive direct effects of contextual explanation found in the management literature.
In sum, we provide an integrative and parsimonious model for linking leadership to trust through specific influencing actions. By utilizing Hernandez et al.’s (2011) framework and Sitkin and Lind’s (2007) nomenclature, we advance the current understanding of trust-specific aspects of leader behavior. This research thus both offers clarity by organizing trust-inducing leader behaviors across multiple theoretical perspectives and builds complexity in our current theorizing about trust creation.
Practical implications
We have tested the role of relational leadership behavior as a central driver of the effects of other leadership behavior on trust. That is, personal and context-building leadership behaviors are effective to the extent they are coupled with a relational foundation that leaders create. This implies that leaders wishing to enhance follower trust should invest more time in crafting and maintaining positive leader-follower relationships, than in projecting their own personal characteristics or transmitting the organizational mission. Building trust through such relational avenues can, for instance, increase job performance and organizational citizenship behaviors (Konovsky & Pugh, 1994).
Our findings also underscore the importance of building trust through reciprocal interpersonal exchanges. Whereas personal leadership behaviors convey the individual leader’s attributes and contextual leadership behaviors communicate organizational characteristics, relational leadership behaviors require the leader to develop individualized knowledge of his or her followers. This implies that relational leadership involves taking an active role in understanding the needs, aspirations, challenges, and skills of their followers to effectively develop trust in the leader-follower relationship. Accordingly, managers should not only aim to develop a positive relationship with their followers, but also actively demonstrate concern and understanding by developing substantive knowledge of their followers through iterative social exchanges.
Drawing on the new insights we provide regarding the context-building elements of leadership behavior, our findings imply that making sense of organizational characteristics involves more than mere transmission of contextual information. To build follower trust, information on organizational characteristics should be tied to the relational knowledge leaders have of their followers. Rather than communicating that the organization values training and development, our findings imply that follower trust is optimized when the leader is able to connect this overarching emphasis to the follower’s specific needs. For instance, if the leader knows that acquiring further technical training is especially important to the follower’s professional aspirations, then highlighting opportunities within the organizational environment that would allow for the development of this specific skill set can enhance the follower’s trust in the leader. By highlighting such opportunities, the leader is not only clarifying available institutional resources, he or she is doing so while conveying concern and understanding for the follower.
Finally, the results also have practical implications for leadership development. Leaders who are attuned to their followers’ experiences and perceptions are more likely to be perceived by followers as exhibiting relational leadership. This can be facilitated both by leader self-awareness and leader awareness of follower responses. Training programs that utilize reflection, introspection, and social skills development may be especially efficacious in enhancing leaders’ capabilities to establish and maintain positive connection with followers. Helping leaders to see their own actions and their followers’ reactions more clearly can serve as a foundation for a better leader-follower relationship and, thus, enhanced follower trust in leaders.
Limitations and future research
Our model showed how our theory-driven approach to modeling leadership behavior effects on trust can enhance both trust and leadership research. Future scholars should seek to continue to test more complex, embedded models of how trust arises and is influenced by leaders. In particular, several avenues for future research arise out of the limitations of the current research. First, additional complexity can be captured by using multidimensional measures of trust. In both studies, we used a global measure of trust based on the willingness to accept vulnerability in the trustee-trustor relationship (e.g., Mayer & Gavin, 2005; Rousseau et al., 1998; Sheppard & Sherman, 1998). Although global measures of trust have been used in the literature (e.g., Brockner et al., 1997; Huff & Kelley, 2003), scholars have argued for trust measures to capture different components of trust (Dietz & Den Hartog, 2006). Thus, a more nuanced understanding of the effect of leadership behaviors on trust can be explored in future research by utilizing trust measures that gauge different types of trust. For instance, given the affective emphasis of relational leadership behaviors as compared to the more cognitive emphasis of personal leadership behaviors, future research should examine the distinct effects of these leadership behaviors on different types of trust, such as cognitive and affective trust (McAllister, 1995) and, more generally, explore how emotionality might play a role in determining the type of trust that leaders build (Sloan & Oliver, 2013).
Second, this article offers preliminary evidence that different “characteristics of the trustee” (Mayer et al., 1995) might have differential effects on trust. To more completely understand the implications of the current research, future scholars should explore the psychometric properties of personal and relational leadership behaviors with ability, integrity, benevolence, and trust. For example, although our measures of personal leadership might be closely associated with Mayer et al. (1995)’s ability and the consistency-based part of integrity, future research should empirically test these relationships. In Venn diagram terms, it could be the case that the “ability” circle is entirely within the “personal leadership” circle and that the “integrity” and “personal leadership” circles overlap. 2 That our laboratory and field studies produce consistent results lends strong preliminary evidence to support our conclusions; however, before we can call for scholars to rethink the Mayer and colleagues (1995) framework, additional attention to this measurement issue is needed.
Third, given that the leader-follower relationship evolves over time, future work should also continue to explore the causal links between leadership behaviors and trust within longitudinal designs. Our studies offer an important first glance at the directionality of leadership influence. Future research could further this exploration by examining if and how these effects might evolve across time periods.
Conclusion
We adopted a multidimensional model of leader-trust relations aimed to capture the richness of these organizational phenomena. While we were able to replicate the findings of the trust literature by showing that personal, relational, and contextual leadership behaviors are direct determinants of trust if examined independently, we also showed that followers’ perceptions of personal and contextual leadership are mediated by their perceptions of relational leadership. We conclude that, in cultivating follower trust, all leader behaviors are not equally influential. Our findings suggest that relational leadership behaviors are the central determinants of follower trust creation.
Footnotes
Appendix A
Appendix B
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.