Write script for a personal presentation for the following information

1. compare and contrast my leadership (authentic) orientation with that of your supervisor (transformational leadership)
2. Address the following two questions:

• How similar is your leadership style to your supervisor's? Provide examples of differences (3)
• Describe the level of leader-member exchange. Does that exchange foster self-efficacy in you?

3 site evidence from the following throughout the paper:

-Employee Trust in Supervisors and Affective Commitment: The Moderating Role of Authentic Leadership

Kehan Xiong

Department of Psychology, Peking University, Beijing, China

Weipeng Lin

Psychological Reports 2016, Vol. 118(3) 829-848 ! The Author(s) 2016 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/0033294116644370 prx.sagepub.com

Department of Human Resource Management, Business School, Nankai University, Tianjin, China

Jenny C. Li and Lei Wang

Department of Psychology, Peking University, Beijing, China

Abstract

Although previous research has examined the main/direct effects of subordinates' trust in their supervisors on the levels of subordinates' affective commitment towards the organizations, little attention has been paid to explore the boundary conditions of this relationship. Two studies were conducted to examine the moder- ating effect of authentic leadership on the relationship between subordinates' trust in supervisor and their levels of affective commitment towards the organization. In line with the hypothesized model, both Study 1 (cross-sectional design, n14138) and Study 2 (lagged design, n 14 154) demonstrated that authentic leadership moderated the relationship between employees' trust in supervisor and their levels of affective commitment towards their organizations. Specifically, the positive relationship was stronger for employees under higher levels of authentic leadership. The implications for theory and practice are discussed.

Keywords

trust in supervisor, trustworthiness, authentic leadership, affective commitment

Introduction

Trust refers to the positive attitudes and expectancies held by individuals toward others amid uncertainties (Lewis & Weigert, 1985; Rousseau, Sitkin,

Corresponding Author:

Weipeng Lin, Department of Human Resource Management, Business School, Nankai University, Tianjin, China.

830 Psychological Reports 118(3)

Burt, & Camerer, 1998; Schoorman, Mayer, & Davis, 2007). Modern work environments are becoming increasingly complex and uncertain (Sayles, 1979). Under such conditions, interpersonal trust between members within organizations is crucial to organizational effectiveness (Thompson, 1967). Thus, developing a relationship of trust is important in organizational settings (McAllister, 1995). Trust relationships, especially those developed by employees with significant others in the organization (e.g., their supervisors), have important implications for their work attitudes and behaviors. For example, previous research has found that trust emotionally connects leaders and their subordinates (Bennis & Nanus, 1985). Moreover, trust in supervisor has been demonstrated to affect a stream of organizational outcomes, such as employees' commitment to the organization (Dirks & Ferrin, 2002; Yang & Mossholder, 2010).

Previous research on employees' trust in supervisors has mainly focused on its direct effects on organizational outcomes. For example, Dirks and Ferrin (2002) found that employees' trust in supervisors has a positive relationship with their organizational commitment. Yang and Mossholder (2010) found that employ- ees' trust in supervisors is positively related to their affective commitment. However, little is known about the boundary conditions of these relationships, which limits understanding of the generalizability of trust theory. In the current research, the focus was on the possible moderating effect of authentic leadership, due to its inherently close connection to trust. As one of the crucial contextual factors that may affect the strength of the connection between trust in supervisor and affective commitment to the organization (Dasborough, 2003), authentic leadership may strengthen the relationship between trust in supervisor and affective commitment.

Trust in leadership and affective commitment

"Trust in supervisor" refers to the level of trust that subordinates hold toward their supervisors, and is related to a series of positive outcomes within organ- izations (e.g., job satisfaction, organizational commitment; Colquitt, Scott, & LePine, 2007; Dirks & Ferrin, 2002). In their integrative model of trust, Burke, Sims, Lazzarz, and Salas (2007) suggest that subordinates' perception of trust in supervisor and its effects on outcomes are based upon characteristics of the supervisor, such as integrity, accountability, transparency, openness, predictabil- ity, consistency, and so on. Without these qualities, subordinates may see the leader as untrustworthy and lower their organizational commitment.

As the core component of organizational commitment, affective commitment refers to the emotional attachment workers feel toward an organization, and is reflected through identification with and involvement in the organization, as well as happiness to be part of the organization (Meyer & Allen, 1984). High affective commitment is associated with reduced employee turnover and absenteeism, reduced work-family conflict, reduced stress, increased overall job performance,

Xiong et al. 831

and organizational citizenship behaviors (Meyer, Stanley, Herscovitch, & Topolnytsky, 2002). Affective commitment is posited to result from many ante- cedents, including demographic variables, individual differences, work experi- ences, and alternatives/investments (Meyer et al., 2002).

Drawing on the emotional contagion perspective, it is reasonable to argue that subordinates' trust in a supervisor will be positively related to their affective commitment. Emotional contagion is "a process in which a person or group influences the emotions or behavior of another person or group through the conscious or unconscious induction of emotion states and behavioral attitudes" (Schoenewolf, 1990, p. 50). As a dynamic process, emotional contagion has subtle but important ripple effects in groups and organizations (Barsade, 2002). Subordinates' trust in supervisor could foster a basic emotional link between employees and their work context (Rajah, Song, & Arvey, 2011). In an atmosphere of positive emotion, employees are driven by this positive affect (Barsade, 2002). Accordingly, they may more easily identify with and attach to their organization, thus showing higher affective commitment. Previous research provides support for this argument. For example, both Dirks and Ferrin's (2002) and Burke et al.'s (2007) studies demonstrate that subordinates' trust towards their leaders was positively related to their affective commitment towards the organization.

The moderating role of authentic leadership

It is worth noticing that the strength of emotion contagion is contingent upon contextual factors such as climate of authenticity (Grandey, Foo, Groth, & Goodwin, 2012). Thus, it is reasonable to believe that authentic leadership, which is the most important source of authenticity climate, would be a crucial factor that moderates the relationship between trust in leadership and affective commitment. Authenticity is defined as a self-referent term (Maslow, 1968; Sartre, 1943), and informed by the "true self" (Avolio & Gardner, 2005; Michie & Gooty, 2005). Authentic leadership is the expression of the "true self" which is described as "being your own person" (George, 2003, p. 12). Its key component and starting point is self-awareness (Avolio & Gardner, 2005; Avolio, Gardner, Walumbwa, Luthans, & May, 2004; Gardner, Avolio, Luthans, May, & Walumbwa, 2005; Kliuchnikov, 2011; Lord & Hall, 2005; Silvia & Duval, 2001), which is connected to self-regulation (Sparrowe, 2005). This helps leaders assess discrepancies between their internalized standards and their praxis (Avolio & Gardner, 2005), facilitate transparency and consistency in leadership where behavior reflects inner standards of a leader (Sparrowe, 2005), and to withstand external pressure and influence (Gardner, Fischer, & Hunt, 2009; Ilies, Morgeson, & Nahrgang, 2005). As such, authentic leadership is opposite to impression management (Avolio, 2005). Given the abovementioned evidence, authentic leaders are considered to promote morality and transparency by drawing upon and promoting positivity in

832 Psychological Reports 118(3)

both psychological capacity and ethical climate (Walumbwa, Avolio, Gardner, Wernsing, & Peterson, 2008). Due to the expression of their authentic inner beliefs, emotions, and attitudes, as well as their behavioral consistency (Bedian & Day, 2004; Tate, 2008), authentic leaders are better able to strengthen the relationship between subordinates' trust in supervisor and their affective commitment.

As one of the crucial contextual factors in the organization (Dasborough, 2003), authentic leadership may largely influence employees' reaction by setting an organization's psychological climate and cultural tone (Hassan et al., 2013). Authenticity is a shared belief between leader and follower (Gillespie & Mann, 2004). The authenticity of a leader provides a signal of relational transparency (Avolio et al., 2004), which indicates that such a leader would act in an authentic way and thus reduces the uncertainty and risk associated with interacting with them. This type of leadership style would create climate of authenticity (Grandey et al., 2012), under which subordinates would more likely perceive their leaders as integral, capable, and benevolent (Schoorman et al., 2007). Therefore, they would be more likely to transform their trust in supervisor into attachment to and identification with their organization, namely affective commitment. In contrast, if a leader acts in a way that is against the values or moral standards they espouse, employees would perceive their leaders as inauthentic. They may experience this as a breach of psychological contract (Schoorman et al., 2007) and thus have less motivation to transform their trust into affective commitment (Mayer, Davis, & Schoorman, 1995; Mayer & Gavin, 1999; Robinson & Rousseau, 1994).

Two studies were conducted. In Study 1, the hypothesized relationships were assessed using cross-sectional data from 14 Chinese companies. Given that the cross-sectional nature of Study 1 might induce common methods variance (CMV, Podsakoff, MacKenzie, Lee, & Podsakoff, 2003), a second study (Study 2) was conducted to replicate the findings from Study 1 and to do so in a way that mitigates the CMV threats. This study used a lagged design to collect data from 13 companies in China. The differing designs provide con- structive replication and enhance the validity of the findings.

Hypothesis. The relationship between subordinates' trust towards their leader and their affective commitment is moderated by their leader's authentic leadership, such that the relationship is stronger for subordinates under higher authentic leadership.

Study 1

In Study 1, the hypothesized moderation was tested using cross-sectional data from 14 Chinese companies, across nine industries including financial services, chemicals, information technology, etc.

Xiong et al. 833Participants

A total of 228 participants within the 14 companies were randomly selected as participants, among which 138 volunteers returned complete responses (60.53% response rate). Among these participants, 50.0% were men; the average age was 31.1 years (SD 14 7.2); the average length of tenure was 9.1 years (SD 14 7.7); and 74.6% of participants had an associate's or higher degree.

Procedures

With help of the human resource departments within the 14 companies, possible participants were chosen randomly. Participants were asked to rate their levels of trust in their supervisors, their perceptions of supervisor authentic leadership, and their levels of affective commitment. Before the data were collected, the nature of voluntary participation and anonymity was emphasized. Each partici- pant received a small gift (worth US$5) for his or her participation.

Measures

Trust in supervisor

The scale developed by Podsakoff and colleagues (Podsakoff, Mackenzie, Moorman, & Fetter, 1990) was used to measure employees' trust in supervisor. To limit the length of the questionnaire, four items with the highest factor load- ings in the original scale were selected (see Gillespie and Mann, 2004 for a similar treatment). Employees were asked to rate on a 7-point rating scale ranging from 1 (strongly disagree) to 7 (strongly agree). An example item is "I have complete faith in the integrity of my supervisor." Cronbach's a coefficient in this sample was .74.

Authentic leadership

The authentic leadership questionnaire (ALQ) was developed by Walumbwa et al. (2008) to measure authentic leadership. ALQ is one of the most widely used measures of authentic leadership (Gardner, Cogliser, Davis, & Dickens, 2011). The original ALQ includes four dimensions and 16 items. To limit the length of the questionnaire, the current study adopted the item selection approach for building a short-version scale used by Gillespie and Mann (2004) and Peterson, Park, and Seligman (2005); one item with the highest factor loading on each dimension of the original ALQ (Walumbwa et al., 2008) was chosen to build a 4-item short form. Participants were asked to report the perceptions of their direct supervisors' authentic leadership on a 5- point rating scale ranging from 1 (highly unlikely) to 5 (highly likely). A sample item is "My supervisor demonstrates beliefs that are consistent with actions." Cronbach's a of this 4-item version of the ALQ was .82 in the current sample.

834 Psychological Reports 118(3)Affective commitment

The 6-item scale developed by Meyer, Allen, and Smith (1993) was used to measure subordinates' affective commitment. This scale has been widely adopted (e.g., Chen & Francesco, 2003; Yao & Wang, 2006). Employees were asked to rate items on a 7-point scale ranging from 1 (strong disagree) to 7 (strongly agree). A sample item is "This organization has a great deal of personal meaning for me." In the present study, the Cronbach's a for this scale was .74.

Control variables

As previous research suggests that demographic variables may influence the accuracy of employees' perceptions and evaluations of supervisors in leadership research (Fleenor, Smither, Atwater, Braddy, & Sturm, 2010), participants' demographic variables, such as gender, age, education, and tenure were included as control variables in this study.

Results and discussion

The means, standard deviations, and correlations of the variables are presented in Table 1.

Testing measurement model

Confirmatory factor analysis (CFA) was conducted to examine whether the constructs were distinguishable from each other. As item parcels have advanta- geous psychometric properties (e.g., greater normality), and can also improve the sample size to parameter ratio as this ratio adversely affects the standard errors and stability of the estimates (see Landis, Beal, & Tesluk, 2000), parcels of

Table 1. Descriptive statistics for Study 1.

Time1Variable

1. Gender 2. Age 3. Education 4. Tenure (years) 5. Trust in supervisor 6. Authentic leadership 7. Affective commitment

M SD 1 2 3 4 5 6 7

1.50 0.50 31.06 7.22 .07 3.22 1.21 .01 9.12 7.65 .03 5.85 1.06 .15 5.74 0.92 .01 5.49 1.07 .08

.11 .93y.03 .00 .06 .04 (.74) .09 .07 .06 .49y(.82) .04 .05 .03 .54y.22* (.74)

Note. n 14 138. Cronbach's a coefficients for scales are presented in parentheses on the diagonal. *p < .05.yp < .01 (two-tailed).

Xiong et al. 835

items were created for affective commitment using a random assignment pro- cedure (Little, Cunningham, Shahar, & Widaman, 2002). The 6-item affective commitment measure was divided into three parcels, each including two items. Results showed that this three-factor measurement model (i.e., authentic lead- ership, trust in supervisor, and affective commitment) had acceptable fit indices,2(41) 14 98.27, p < .01, Comparative Fit Index (CFI) 14 .95, Tucker-Lewis Index (TLI)14.93, Root Mean Square Error of Approximation (RMSEA)14.10, Standardized Root Mean Square Residual (SRMR)14.08 (see Hu & Bentler, 1999; Nye & Drasgow, 2011). The measurement model was supported, although more weakly than hoped.

Convergent and discriminant validity of the measures was assessed. The results of CFA showed that each observable indicator loaded significantly (ps<.01) on its intended factor, supporting the convergent validity of scale items (Anderson & Gerbing, 1988). To examine discriminant validity, the uncon- strained measurement model was compared with three constrained models in which any two of the three factors were combined (see Goffin, 2007). A significant chi-square difference implies that the unconstrained model is a better fit for the data, thereby supporting the existence of discriminant validity (Anderson & Gerbing, 1988). Results showed that all chi-square differences were significant (119.442(df142)213.62, ps<.01), indicating high discrim- inant validity.

Testing moderation effects

Because participants were nested in 14 companies, the nested nature of the data set violates the data independence assumptions of ordinary least squares regres- sion models (Bliese, 2000; Klein & Kozlowski, 2000). Thus, multi-level modeling was used to test the hypothesis. Model estimation was conducted using Mplus 7 (Muthe n & Muthe n, 1998-2012). The model tested specified Level 1 (i.e., indi- vidual level) as fixed effects of control variables (i.e., gender, age, education, and tenure), trust in supervisor, authentic leadership, and the interaction terms between trust in supervisor and authentic leadership on affective commitment. At Level 2 (i.e., company level), all Level 2 intercepts of Level 1 focal variables (i.e., trust in supervisor, authentic leadership, and affective commitment) were set to freely correlate with each other. All exogenous variables at Level 1 were centered along their grand mean. This model explained 33% of the individual level variance in affective commitment.

Results are presented in Table 2. Controlling for the random effects of employing unit (Level 2), as well as the fixed effects of gender, age, education, and tenure, trust in supervisor was positively related to affective commitment (g 14 0.51, p < .01). Results also showed that authentic leadership moderated the relationship between trust in supervisor and affective commitment (g140.20, p<.01), such that the positive relationship between trust in supervisor and

836

Psychological Reports 118(3)

Table 2. Hierarchical linear modeling results. Affective commitment

Study 1 (Time 1) (Time1) g SE t p g SE t p

Variable

Study 2 (Time 2)

Intercept 5.24 Gender 0.06 Age 0.02 Education 0.02 Tenure 0.01 AC

TS 0.51 AL 0.03 TS AL 0.20 Level 1 R2.33

0.18 29.15 0.11 0.57 0.04 0.49 0.05 0.43 0.03 0.21

0.10 4.98 0.10 0.33 0.07 2.95

.08 4.21

<.001

<.001 5.21 0.09 56.97 .57 0.02 0.10 0.23 .82 .62 0.01 0.02 0.86 .39 .67 0.04 0.05 0.76 .45 .84 0.01 0.02 0.45 .65

<.001 <.001

0.49 0.10 5.00 <.001 0.13 0.08 1.54 .12 .74 0.01 0.09 0.17 .87

<.001 0.14 0.04 3.74 <.001 0.37 0.10 3.63

<.001

Note. Study 1: n 14 138. Study 2: n 14 154. Gender was coded 1 for men and 2 for women. Education was coded 1 for No degree, 2 for High school diploma, 3 for Associate's degree, 4 for Bachelor's degree, 5 for Master's degree, and 6 for Doctoral degree. Entries for g are unstandardized estimations of the fixed effects. AC: affective commitment; AL: authentic leadership; TS: trust in supervisor.

affective commitment was stronger for employees reporting higher authentic leadership, but was weaker for those reporting lower authentic leadership (Figure 1).

The online calculator developed by Preacher, Curran, and Bauer (2006) was used to estimate simple slopes, which describe the relationships between trust in supervisor and affective commitment at varying levels of authentic leadership. Specifically, high authentic leadership was designated as 1 SD above the mean, average authentic leadership was the mean, and low authentic leadership was 1 SD below the mean. Results showed that the slope was larger for individual under higher levels of authentic leadership (simple slope 14 0.69, p < .01) than for those under lower levels of authentic leadership (simple slope 14 0.33, p < .01).

In sum, the findings of Study 1 provided support for the hypothesis. However, there are several limitations in Study 1. First, the use of short-version scales (i.e., trust in supervisor and authentic leadership) might limit the validity of the findings. Second, data collected at one time point might induce common method bias. As such, a second study was conducted to mitigate these concerns and constructively replicate the findings. Specifically, in Study 2, the original and complete-version scales were used, as these have been demonstrated to have

Xiong et al. 837

Figure 1. The moderating effect of authentic leadership on the relationship between sub- ordinates' trust in supervisor and affective commitment: Study 1.

good reliability and validity to measure the variables. A lagged design reduced common method bias.

Study 2

Given that the cross-sectional nature of Study 1 might induce common method variance (Podsakoff et al., 2003), a second study (Study 2) was con- ducted to replicate the findings from Study 1 in a way that also mitigated the common method variance threats. A lagged design was used to collect data from 13 companies in China. The differing research designs provide constructive rep- lication and enhance the validity of the findings.

Participants

This study was conducted on participants from 13 companies in China across three industries, including information technology, financial services, and prop- erty management. Participants were 168 randomly selected employees, 154 of whom returned completed answers (91.7% response rate). Among these partici- pants, 49.4% were men; the average age was 29.0 years (SD 14 5.6); the average

838 Psychological Reports 118(3)length of tenure was 6.2 years (SD145.2); and 79.2% of participants had an

associate's or higher degree.

Procedures

Survey data were collected at two time points separated by one month to reduce common method variance (Podsakoff et al., 2003). At Time 1, employees reported their demographics (i.e., gender, age, education, and job tenure), levels of trust in supervisor, affective commitment, and their perceptions of supervisors' authentic leadership. One month later, at Time 2, employees again rated their levels of affective commitment. The nature of voluntary par- ticipation and confidentiality was emphasized. A code on each questionnaire ensured data confidentiality while linking employees' two-wave questionnaires. Each participant received a small gift (worth US$5) for his or her participation.

With the help of human resource personnel, 168 participants were randomly chosen within 13 companies. Among them, 165 (98.2%) responded to the first survey and of these, 154 (93.3%) responded to the second survey. Thus, the final sample included 154 participants, with a final response rate of 91.7%.

Measures

Participants were asked to indicate on a 7-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree) the extent to which they agreed with each item.

Trust in supervisor

The 6-item scale developed by Podsakoff et al. (1990) to measure trust in super- visor was used at Time 1. An example item is "I have complete faith in the integrity of my supervisor." Cronbach's a coefficient in this sample was .84.

Authentic leadership

The Authentic Leadership Inventory (ALI) developed by Neider and Schriesheim (2011) to measure authentic leadership was administered at Time 1. The ALI is a newly developed authentic leadership measure that has shown high reliability and validity. A sample item is "My supervisor shows consistency between his/her beliefs and actions." Cronbach's a of this scale was .93 in the current sample.

Affective commitment

Employees' affective commitment was measured at both Time 1 and Time 2 using the same measure as in Study 1 (Meyer et al., 1993). In the current sample, Cronbach's a for this scale was .77 at Time 1 and .76 at Time 2.

Xiong et al. 839Control variables

As in Study 1, participants' demographics including gender, age, education, and tenure were included as control variables. Because there was an interest in changes of affective commitment among employees, level of affective commit- ment at Time 1 was also controlled.

Result and discussion

The means, standard deviations, and correlations of the variables are presented in Table 3.

Testing measurement model

CFA was conducted to test the distinctiveness of the constructs. As in Study 1, parcels of items were created for all variables (i.e., Time 1 trust in supervisor, Time 1 authentic leadership, Time 1 affective commitment, and Time 2 affective commitment) to improve the ratio of sample size to parameters (see Landis et al., 2000). Following the internal-consistency approach (Kishton & Widaman, 1994), the items assessing four facets of Time 1 authentic leadership were each regrouped into four distinct parcels that were used as indicators of the latent construct of authentic leadership. Following the random assignment procedure (Little et al., 2002), three parcels of items were constructed for Time 1 trust in supervisor, Time 1 affective commitment, and Time 2 affective commitment,

Table 3. Descriptive statistics for Study 2.

Variable

Time 1 1. Gender 2. Age 3. Education 4. Tenure (years) 5. Trust in supervisor 6. Authentic leadership 7. Affective commitment Time 2 8. Affective commitment

M SD 1 2 3 4 5 6 7 8

1.51 0.50 28.97 5.57 .20* 3.25 0.97 .04 .12 6.20 5.15 .15 .76y 5.92 0.94 .13 .04 5.69 0.90 .25y.12 5.39 0.99 .08 .01 .00 .12 .52y.42y

5.23 0.97 .10 .05

.12 .03 .09 .09 .01 .52y(.93)

(.84)

(.77) .04 .07 .39y.31y.64y(.76)

Note. n 14 154. Cronbach's a coefficients for scales are presented in parentheses on the diagonal. *p < .05. yp < .01 (two-tailed).

840 Psychological Reports 118(3)

with each parcel including two items. Each parcel for affective commitment was created based on the same set of items across time.

Results showed that our four-factor measurement model (i.e., Time 1 authen- tic leadership, Time 1 trust in leader, Time 1 affective commitment, and Time 2 affective commitment) had acceptable fit indices,2(59) 14 104.42, p < .01, CFI 14 0.96, TLI 14 0.95, RMSEA 14 0.07, SRMR 14 0.04. All items loaded signifi- cantly on their corresponding factors. This measurement model fits the data better than all six constrained models in which any two of the four factors were combined, 32.93 2(df 14 3) 182.51, ps < .01. These results provided support for the hypothesized measurement model.

Testing moderation effects

As in Study 1, multi-level modeling was used to test the hypothesis. At Level 1 (the individual level), the fixed effects of control variables were specified (i.e., gender, age, education, tenure, and Time 1 affective commitment), Time 1 trust in supervisor, Time 1 authentic leadership, and the interaction terms between Time 1 trust in supervisor and Time 1 authentic leadership on Time 2 affective commitment. At Level 2 (the company level), all Level 2 intercepts of Level 1 variables (i.e., trust in supervisor, authentic leadership, and affective commit- ment) were set to freely correlate with each other. All exogenous variables at Level 1 were centred along their grand mean. This model explained 37% of the individual level variance in affective commitment.

As shown in Table 2, authentic leadership moderated the relationship between trust in supervisor and affective commitment, g140.04, p<.01 (Figure 2). Simple slope tests showed that the relationship between trust in supervisor and affective commitment was stronger for individuals under higher levels (i.e., 1 SD above the mean) of authentic leadership (simple slope 14 0.26, p < .05) rather than for those under lower levels (i.e., 1 SD below the mean) of authentic leadership (simple slope 14 0.00, p > .05). Therefore, the hypothesis was supported.

Supplementary analysis

Although two studies were conducted to constructively test and provide support for the hypothesis that authentic leadership moderates the relationship between trust in supervisor and affective commitment, one may argue that it is possible that trust in supervisor might mediate the relationship between authentic lead- ership and affective commitment. Therefore, this mediation model was tested using data from both Study 1 and Study 2. Results demonstrated that, after taking control variables into account, the mediation effect of trust in supervisor is only significant in the sample from Study 1, with an indirect effect14.23, p<.01, 95% bias-corrected bootstrap confidence interval (CI) of (.04, .41), but not significant in the Study 2 sample, indirect effect14.03, p>.05, with a

Xiong et al. 841

Figure 2. The moderating effect of authentic leadership on the relationship between sub- ordinates' trust in supervisor and affective commitment: Study 2.

95% bias-corrected bootstrap CI of (.04, .15). The results demonstrated the same pattern when the effects of control variables were not included in the models (results are available from the authors upon request). Considering the cross-sectional design of Study 1, the mediation effect of trust in supervisor might be due to common method bias. Nevertheless, given that the hypothesized moderation model was supported by both studies, it suggests that the moder- ation model is more stable and robust than the mediation model, providing support for the hypothesis.

General discussion

As suggested by previous theories and research, trust arises from uncertainties, with its core related to the willingness to take risks, to accept vulnerability, and to rely on someone (Rousseau et al., 1998). However, its premise is that the trustee deserves to be trusted (Mayer et al., 1995). Leaders and their sub- ordinates go through a process full of uncertainty, risk, and possible loss in forming a trust relationship. If a subordinate deems their leader to be trust- worthy, they will be willing to accept the vulnerability that goes with the risk of trusting, and demonstrate more affective commitment toward the organiza- tion (Burke et al., 2007).

842 Psychological Reports 118(3)

The present research examined the moderating effect of authentic leadership on the relationship between employees' trust in supervisor and their organiza- tional affective commitment. Multilevel modeling analyses from two studies revealed that the level of authentic leadership was a significant marginal condi- tion for the relationship between levels of trust in supervisor and organizational affective commitment. While a particular relationship exists between leaders winning subordinates' trust and subordinates' affective commitment toward the organization, the strength of the relationship is subject to subordinates' perceptions of leaders' authenticity.

-Psychology

Integrating Implicit

Leadership Theories,

Leader-Member

Exchange, Self-Efficacy,

and Attachment

Theory to Predict

Job Performance

Alireza Khorakian and

Mohammad Sadegh Sharifirad

Faculty of Economics and Administrative Sciences, Ferdowsi University of Mashhad

(FUM), Iran

Abstract

The impact of implicit leadership theories on performance and the mechanism linking

them have received insufficient theoretical and research attention. Drawing on

Bandura's social cognitive theory, the present study contributes theory through

examining the assertion that higher congruence between followers' implicit leader-

ship theory and the characteristics of supervisors enhance job performance through

higher quality of leader-member exchange and self-efficacy. Moreover, in the pro-

posed model, attachment insecurity was considered as the antecedent of the con-

gruence and leader-member exchange in addition to the moderator of the

relationship between them. Capitalizing upon Structural Equation Modeling (SEM),

this study tested the model in a field study using a sample of employees in knowl-

edge-oriented firms in Iran. The results suggest that the congruence between fol-

lowers' implicit leadership theory and the characteristics of supervisors does not

directly impact performance and leader-member exchange and self-efficacy are the

full mediators. The results also showed that attachment insecurity is the predictor of

neither the congruence nor the leader-member exchange. Additionally, attachment

insecurity moderates the relationship between these two variables in a way that

when attachment insecurity is high, the congruence has more positive impact on

leader-member exchange.

Psychological Reports

2019, Vol. 122(3) 1117-1144

!

The Author(s) 2018

Article reuse guidelines:

sagepub.com/journals-permissions

DOI: 10.1177/0033294118773400

journals.sagepub.com/home/prx

Corresponding Author:

Alireza Khorakian, Faculty of Economics and Administrative Sciences, Ferdowsi University of Mashhad

(FUM), Iran and Honorary lecturer, Stirling Management School, University of Stirling, Stirling, UK.

Email: a..n@um.ac.ir; a..n@stir.ac.uk

Keywords

Implicit leadership theories, leader-member exchange, self-efficacy, attachment

insecurity, job performance

Introduction

In recent years, there has been an increasing investigative interest in follower-

centered leadership (Shamir, 2007; Junker & van Dick, 2014) and hence, a

growing amount of research has studied implicit leadership theories (ILTs).

ILTs refer to some schemas and cognitive structures specifying the abilities

and traits characterizing a business leader (Epitropaki & Martin, 2005).

Witnessing and feeling the possession of ILTs in supervisors can impact an

array of work variables such as job satisfaction, well-being, affective commit-

ment, and leader-member exchange (LMX) (Engle & Lord, 1997; Epitropaki &

Martin, 2005; Topakas, 2011). As for the mechanisms linking ILTs to job out-

comes, a few studies have explored the antecedents, mediators, and moderators

(Epitropaki, Sy, Martin, Tram-Quon, & Topakas, 2013).

Despite the emerging literature, no empirical research has yet introduced self-

efficacy as the potential mediator bridging the relationship between the congru-

ence of followers' ILTs about a leader in general and their recognized ILTs for

their current supervisors and follower performance. Second, despite the provi-

sion of a theoretical framework connecting follower attachment styles (anxious,

avoidant, and secure) and ILTs by Keller (2003), no research has empirically

validated the proposed framework. Third, attachment theory, as a leading the-

oretical framework established to investigate individuals' differences in close

relationships and leadership (Gillath, Karantzas, & Fraley, 2016; Richards &

Hackett, 2012), has the power to function to moderate the relationship between

leaders and followers after the picture of a supervisor as a leader or nonleader is

solidified.

This study aims to void these gaps by, first, introducing attachment insecurity

as the antecedent of the supervisor-leader match (based on ILTs) and the mod-

erator on the relationship between this match and LMX and, second, investigat-

ing how the two mediators of LMX and self-efficacy mediate the relationship

between supervisor-leader congruence and follower job performance.

ILTs and job performance

Leadership is a prevailing social phenomenon; therefore, both experts and lay-

people create and develop mental characteristics distinguishing leaders. Explicit

theories (e.g., LMX, transformational leadership and transactional leadership)

are devised by scientists through scientific data and observations; whereas,

implicit theories (e.g., ILTs) are constructed in the minds of people regardless

of their education, knowledge, and expertise (Levy, Chiu, & Hong, 2006).

1118

Psychological Reports 122(3)

The knowledge structures defining leadership in the minds of people are called

ILTs (Eden, 1992; Lord & Maher, 1991).

Leadership categorization theory (Lord, Foti, & De Vader, 1984) is rooted in

social cognitive theory, positing that followers possess a mental depiction of a

leader via prototypes. This provides a background for the followers to make a

comparison between actual people and these prototypes to give a verdict if they

are (ideal) leaders or not (Lord & Maher, 1991; Shondrick & Lord, 2010).

The extent to which this match can be verified determines the level of leader

power (Maurer & Lord, 1991). Additionally, ILTs are some cognitive represen-

tations utilized for ''sensemaking'' (Weick, 1995); therefore, they create expect-

ations of individuals and also act as a simplification mechanism to free cognitive

resources to be employed for more important tasks such as problem-solving and

coordination (Dinh & Lord, 2010). Research has shown that the consideration

of leader as the owner of followers' ILTs can positively impact the quality and

amount of attributed influence and even charisma (Maurer & Lord, 1991).

One can sensibly think of these attributions of power due to the overlap between

employees' ILTs and managers' traits, characteristics and behavior as the vehicle

to increase their expectations of receiving positive outcomes and rewards due to

their association with such a manager.

From one point of view, heightened expectations are likely to alter employee

perceptions more pervasively regarding their work environment, job attitudes,

and performance. Taken from a different angle, when one is accepted as a

leader, the followers reciprocate via openness toward the leaders' actions

(van Quaquebeke & Brodbeck, 2008) and mental confirmation of leader's effect-

iveness (Giessner & van Knippenberg, 2008). Both, followers' reverence and

perception about the effectiveness of leader initiate the welcoming of leaders'

influence, which is the core to leadership effectiveness (Hollander, 2006; Yukl,

2010). Leaders' influence, in return, manifests itself in followers' behavior or

performance (van Knippenberg, 2011).

The mediating role of LMX

Based on LMX theory, leaders categorize followers into two groups of in-groups

and out-groups. For the former, the employees have high-quality relationships

with their leaders and are privileged to access a bigger share of leader's attention,

resources, and support than do the ones in the latter whose relationships of less

quality. From the perspective of leadership, the approach investigating the qual-

ity of the dyadic relationship between a leader and a follower (LMX) has been

popular.

Having roots in social exchange theory (Blau, 1964) and role theory (Merton,

1968), LMX concentrates on the dyadic exchange between leaders and followers

in addition to the process of relationship development. Graen and Uhl-Bien

(1995) pointed out that ''LMX clearly incorporates an operationalization of a

Khorakian and Sharifirad

1119

relation-based approach to leadership'' (p. 109), which is based on the benefits of

both sides (Graen & Uhl-Bien, 1995). Graen and colleagues pioneered the clas-

sification of exchange quality between leaders and followers (see Dansereau,

Graen, & Haga, 1975; Graen & Cashman, 1975), which differentiates in-group

members from out-group members. The former receives more support and

resources while the latter possess less freedom in role development and have

little power in decision making.

This dichotomy causes a wide array of work-related consequences. Gerstner

and Day's (1997) meta-analysis showed that LMX is correlated with perform-

ance, job satisfaction with one's supervisor, organizational commitment, overall

satisfaction, and follower's competence level. In the most recent meta-analysis,

Martin, Guillaume, Thomas, Lee, and Epitropaki (2016) demonstrated that

there is a positive relationship between LMX and objective task performance.

Some other positive job outcomes are employee well-being (Epitropaki &

Martin, 2005), self-efficacy (Murphy & Ensher, 1999), lower levels of followers'

stress (e.g., Bernas & Major, 2000), empowerment (Keller & Dansereau, 1995),

and perceptions of transformational leadership style (Howell & Hall-Merenda,

1999). High-quality relationship between leaders and followers can lead to

higher organizational citizenship behavior (Ilies, Nahrgang, & Morgeson,

2007), time and effort invested in the job (Liden & Graen, 1980), and innovation

(Tierney, Farmer, & Graen, 1999).

A plethora of research has explored the predictors of LMX quality. In one

classification, Martin, Epitropaki, Thomas, and Topokas (2010) put the

antecedents in four groups of leader characteristics, subordinate characteristics,

interactional variables, and contextual factors. As examples of subordinate

characteristics, follower's personality (Bernerth, Armenakis, Field, Giles, &

Walker, 2008; Lapierre & Hacket, 2007) is positively related to LMX.

In terms of interactional variables, leader and subordinate liking (Martin

et al., 2010) is positively associated with LMX. In a study, Nahrgang,

Morgeson, and Ilies (2009) found out that leader personality predicts LMX.

As a more pertinent study, researchers have shown that the congruence between

implicit and explicit leadership theories can positively impact LMX quality

(Epitropaki & Martin, 2005; Topakas, 2011).

The mediating role of self-efficacy

Self-efficacy is rooted in social cognitive theory, which mainly shows people have a

tendency to give a direction to their effort through some controllable behaviors

leading to desired outcomes (Bandura, 1997). The relationship between perform-

ance and self-efficacy has been explored by different researchers. It is widely proved

that there is a positive correlation between performance and self-efficacy (Stajkovic

& Luthans, 1998). Also, prior performance can boost up a person's self-efficacy

(e.g., Bandura, 1997; Mitchell, Hopper, Daniels, George-Falvy, & James, 1994).

1120

Psychological Reports 122(3)

Despite being the subject of scrutiny, it is also shown that self-efficacy can

positively influence performance (Bandura, 1997) via controlling distractions

and disruptive thoughts leading to more focus on tasks (Kanfer & Ackerman,

1996), standing against obstacles and failure (e.g., Cervone & Peake, 1986),

using strategies effectively (e.g., Bandura & Wood, 1989), higher goal

setting (Bandura, 1997; Bouffard-Bouchard, 1990), goal commitment

(Locke & Latham, 1990) in different realms of academic settings (Multon,

Brown, & Lent, 1991), health functioning (Holden, 1992), work organizations

(Stajkovic & Luthans, 1998), and athletics (Moritz, Feltz, Fahrbach, &

Mack, 2000).

Social cognitive theory posits that people do not function in isolation and

people's environment can shape performance expectations and direct individ-

uals' endeavor (Bandura, 1997). For instance, Eden, Ganzach, Flumin-Granat,

and Zigman (2010) demonstrated that admiring systems' superior technology

could decline the time of filing entitlement claims by 43.5%. Bearing this notion

in mind, in this research, we focus on interpersonal self-efficacy, identification

with leader, and the perception of transformational leadership to justify the

mediating role of self-efficacy.

Individuals influencing a focal task are the critical building blocks of a

task environment and therefore can boost up performance in conjunction with

self-efficacy (Emich, 2012). However, prior research done to explore the rela-

tionship between self-efficacy and performance shows contradictory results.

In interdependent task outcome, because of the interconnection of skills to do

the task, the performance is contingent upon the perception of partners' skills

(Kerr, Messe, Park, & Sambolec, 2005; Kerr & Fleming, 2007), whereas in

noninterdependent tasks, due to shared instrumentality, individuals' attempt

declines and as a result performance decreases (Karau & Williams, 1993).

Since leadership and followership are essentially interdependent (Baker,

Mathis, & Stites-Doe, 2011; Hollander, 1992), it is likely that self-efficacy is

improved by the perception of leaders' skills.

Bass and Avolio (1989) showed that leadership prototypes correlated the

characteristics of transformational leaders. In an international vein, GLOBE

project results also bolstered this idea and showed that culturally endorsed

implicit leadership theories associated with charismatic/transformational leader-

ship (Den Hartog et al., 1999).

Bandura (1997) posits that role modeling increases self-efficacy. As it was

mentioned, self-efficacy is a belief that one has the ability to fulfill a task

(Bandura, 1997), whereas relational identification is the extent to which one

defines oneself based on a given role-relationship (Sluss & Ashforth, 2007).

Modeling is a strategy to provide information on ''correct'' performance

which is not available otherwise (Gist & Mitchell, 1992). Identifying oneself

with the high-quality relationship with the leader enables employees to behave

aligned with their identity which provides the opportunity for the followers to

Khorakian and Sharifirad

1121

increase their knowledge, acquire new skills, learn from their leaders, and

thereby boost up self-efficacy (Latham, Winters, & Locke, 1994).

According to Sluss and Ashforth (2007), relational identification entails the

participation of two sides. This extension from identifying oneself with personal

qualities to the inclusion of another's quality is likely to be beneficial through

more empathy, mutual understanding, social support, cooperation, and loyalty

which can lead to trust as the predictor of knowledge sharing. Isaksen (1983)

has already shown that trust and information sharing can foster followers' self-

efficacy through escalating the ability to specify the correct course of action.

One may argue that high relationship quality with leader can benefit fol-

lowers through relational identification which can cause followers to incorporate

positive relationship into their self-concept and, as a consequence, increase self-

efficacy and performance.

Based on self-efficacy theory, individuals receive clue from the environment to

evaluate some factors pertinent to themselves and their situations to figure out

the availability of resources to accomplish tasks (Gist & Mitchell, 1992). One of

these factors is mastery experience or experiencing success repeatedly (Bandura,

1997). High levels of LMX quality increase mastery experience which can bolster

self-efficacy (Walumbwa & Hartnell, 2011). Self-efficacy is independently asso-

ciated with job performance (e.g., Liao & Chuang, 2007; Stajkovic & Luthans,

1998). Since it impacts an individual's goal setting and activates taking up a

challenge or activity, the task is believed to be successfully accomplished. Wise

selection of situations via the road map of self-efficacy plays a critical role in

strong commitment and high aspiration to fulfill expectation.

Attachment styles

Based on the attachment literature, two general types of attachment styles are

detectable: secure and insecure (anxious and avoidance attachment). Secure

attachment is the consequence of consistent caregiving. In this style, infants

are optimistic, have positive views of the self and others, have enough confidence

in the availability of support in the times of distress; hence, leading to optimal

functioning (Mikulincer, 1995). On the contrary, attachment anxiety features

negativity of self-image and overreliance on mutual relationships (Mikulincer

& Shaver, 2005) as well as hyper vigilance to social and emotional signals

received from others. This germinates from the seeds of distress from separation

or rejection and abandonment from the attachment figure.

Being categorized as insecure attachment style, attachment avoidance engen-

ders untrustworthiness and unavailability of others in the times of need

(Mikulincer & Shaver, 2005) which causes the ''deactivation of proximity seek-

ing, inhabitation of the quest for support and active attempts to handle distress

done.'' (Mikulincer & Shaver, 2005). Avoidance is the response to the termin-

ation of anticipating additional frustration emanating from attachment figure's

1122

Psychological