Talk Matters at Work: The Effects of Leader-Member Conversational Quality and Communication Frequency on Work Role Stressors

Work Role Stressors

Work stress is composed of multiple sources (Rosen et al., 2010). Of particular interest in this study are the sources of stress related to work roles known as role stressors. The theory of role dynamics serves as the foundation for understanding role stressors (Kahn et al., 1964). According to the theory, a working individual has an implicitly or explicitly assigned role in an organization or a set of activities he or she is expected and/or demanded to perform. These role expectations and demands generate pressure or stress over the performing individual. According to their varied nature and quality, role stressors are differentiated into three types: role ambiguity, role conflict, and role overload. Kahn et al. (1964) and Rizzo, House, and Lirtzman (1970) refer to role ambiguity as the level of certainty and clarity an employee perceives with regard to his or her role’s responsibility, expectations, and evaluations. Role conflict refers to the contradictory assignments, inconsistent expectations, or competing obligations experienced by an employee. In Kahn et al.’s (1964) initial conceptualization, the third role stressor, role overload was considered a particular type of role conflict. However, later studies (Glazer & Beehr, 2005; Tordera, González-Romá, & Peiró, 2008) suggested role overload, although having a high correlation with role conflict, have its unique quality. Role overload is defined by the degree to which an employee perceives his or her assignments or expectations as exceeding what his or her ability affords to handle with available resources without compromising quality (Jian, 2014; Tordera et al., 2008).

Since Kahn et al.’s (1964) seminal work, an impressive amount of research on role stressors has accumulated over the years, identifying negative associations of role stressors with work performance (e.g., Rosen et al., 2010) and attitudinal outcomes, such as job satisfaction and organizational commitment (e.g., Antón, 2009). For example, a meta-analysis of 169 sample studies found that job performance is negatively correlated with all role stressors, and its relationship with role ambiguity is most pronounced (Gilboa, Shirom, Fried, & Cooper, 2008).

Another line of research made efforts to identify antecedents and correlates that may contribute to the perception of role stressors, focusing on variables related to alleviating stress and devising effective interventions. One consistent finding in this line of research suggests that social support from coworkers and supervisors, feedback, and participation in decision making are significantly associated with perceived levels of role stressors (Miller, Ellis, Zook, & Lyles, 1990; Ray & Miller, 1991).

In spite of these significant findings, we argue that more could be learned about role stressors by taking a closer look at communication in routine work settings through focusing on routine conversations and interactions that take place at work. Routine conversations at work are important, first, because both conceptual (Taylor & Van Every, 2000) and empirical research (Boden, 1994; Clifton, 2012) in organizational communication in recent years has gradually revealed the central role of conversations in the process of organizing. More important, according to the theory of role dynamics, role expectations and demands are communicated from role senders to their receivers largely through conversations. We argue that the quality of conversation may affect how role expectations are perceived. Second, knowing the precise mechanisms by which communication works in relation to stress could offer concrete behavioral guidance for practitioners in managing and alleviating stress. The present study focuses on work conversations between leaders and members. The following section offers theoretical rationale for our focus and proposes several hypotheses.

Leader-Member Conversation and Role Stressors

In their comprehensive review of research on occupational health, Macik-Frey et al. (2007) suggested that leadership be a promising area that may offer significant understanding regarding follower psychological health. The authoritative position a leader often occupies in providing resources, defining roles, and evaluating performance of their followers has led scholars to focus on leadership as a key explanation for follower psychological health (e.g., Britt, Davison, Bliese, & Castro, 2004). Studies have linked certain leadership styles to follower well-being. For example, in an experimental study, Lyons and Schneider (2009) found that participants exposed to a transformational leadership style would perceive a stressful task as less threatening than those exposed to transactional leaders. In a review of research conducted in military settings, Britt et al. (2004) observed that leader behaviors help reduce group conflict and followers’ perception of role ambiguity, and increase their perception of contribution to a mission.

In spite of these findings on the association between leadership roles and followers’ psychological health, Macik-Frey et al. (2007) critiqued in their review of this literature, “[t]here has been little study of the processes whereby leadership affects follower well-being in terms of specifying precise mechanisms” (p. 825). While an examination of an exhaustive set of mechanisms is beyond the scope of a single study, we contend that a detailed look at leader-member conversations would be an appropriate starting point in formulating the precise mechanisms that Macik-Frey et al. (2007) suggested.

First, we draw on the theory of leader-member exchange (LMX) as a framework for understanding leadership (Graen & Uhl-Bien, 1995). Although defining leadership is a contentious issue in the broader literature of leadership, the majority of empirical LMX studies defines and operationalizes leader-member relationship within the context of supervisor-subordinate relationship in organizational settings. Thus, the terms supervisor-subordinate and leader-member are used interchangeably and leadership is narrowly defined by the formal supervisory position. The present study follows this tradition. The LMX theory lends its focus on the relational dynamic in leader-member dyads. It postulates that leaders develop varied levels of relationship quality with their members. Particularly, meta-analyses of LMX studies have presented strong evidence that lower LMX quality is associated with higher levels of member perceived role ambiguity and role conflict (Dulebohn, Bommer, Liden, Brouer, & Ferris, 2012; Gerstner & Day, 1997). However, LMX, being a perceptual assessment of overall relational quality, lacks the specificity in pinpointing the communicative process that may affect role stressors (Jian, Shi, & Dalisay, 2014). In contrast, a focus on leader-member conversation reflects a more specified look at the communicative exchange. The focus on leader-member conversations also has its roots in the interaction richness theory, as articulated by Barry and Crant (2000). Interaction richness theory conceptualizes communication as both information exchange and meaning sharing. The theory argues that a rich interaction is one in which “communication within an interacting dyad at a given point in time is efficient (high in symbolic content), coordinated (characterized by synchronous interaction), and accurate (symbolic meaning is shared and appropriately interpreted)” (Barry & Crant, 2000, p. 651). In a quality conversation, more meaning can be conveyed with fewer words; the interactional movements are highly synchronized; intended meanings conveyed by one partner are precisely interpreted by the other. We extend this to the context of leader-member conversations. In assigning a work task, for example, a high-quality conversation in a leader-member dyad should have a high degree of “carrying capacity” (Barry & Crant, 2000, p. 651) in the sense that the interactional moment allows the sharing of this task assignment efficiently and accurately in a highly coordinated manner.

In addition, our focus on leader-member conversation is informed by recent studies on the effects of leader motivating language on follower outcomes. Drawing on speech act theory (Austin, 1962), Sullivan (1988) argues that leader language not only acts to reduce follower uncertainty through transmission of information but also to facilitate follower’s meaning making and to establish human bonding. It is argued that these leader speech acts exert motivational forces on follower performance and psychological outcomes. Subsequent empirical research has lent strong support to the theory. For example, in a series of studies, it was found that leader motivating language is strongly associated with the reduction of employee absenteeism (Mayfield & Mayfield, 2009) and with improvement in employee self-efficacy and performance (Mayfield & Mayfield, 2012) and job satisfaction (Mayfield & Mayfield, 2006). Sharbrough, Simmons, and Cantrill (2006) reported that leader motivating language is significantly correlated with member perception of job satisfaction, communication satisfaction, leader effectiveness, and leader’s communication competency.

Based on these different bodies of literature on leadership communication, we argue that the quality of leader-member conversations has significant implications on subordinate perceptions of role stressors. If conversations are of high quality, with other factors being equal, we would expect a lower level of perceived role ambiguity on the part of subordinates, because high-quality conversations should convey expectations with more clarity. In a leader-member dyad, a high-quality interaction is not simply about the flow of messages from supervisors to subordinates but also how messages flow the other way around. In other words, the leader-member dyad develops a synchronized co-orientation toward mutual understanding. Because of such co-orientation, a high-quality conversation would allow a robust feedback mechanism in the dyad so that problems and questions regarding conflict or excessive work assignments tend to be detected and brought to resolution. As a result, members who perceive to have had higher quality conversations with their supervisors tend to perceive lower levels of role ambiguity, role conflict, and role overload. Hence, we propose the following hypotheses:

In addition to conversational quality, communication frequency has drawn scholarly attention as an important communication factor that is associated with organizational and employee outcomes, especially within the leader-member context (Becerra & Gupta, 2003; Gajendran & Joshi, 2012; Kacmar, Witt, Zivnuska, & Gully, 2003). For example, in a study of globally distributed teams, Gajendran and Joshi (2012) observed that communication frequency between team leaders moderates the association between leader-member relationship and member influence on team decisions such that, when communication frequency increases, the association is strengthened. The moderating effect of communication frequency was also found in a study by Becerra and Gupta (2003) on trust in organizational settings. For instance, their study revealed that, as communication frequency increased, an organizational member’s attitudinal predispositions toward his or her colleagues became less important in determining his or her evaluation of their trustworthiness (Becerra & Gupta, 2003). Based on existing research, it is reasonable to expect a moderating effect of communication frequency on the association between LMCQ and members’ perception of role stressors. The rationale is that when communication frequency is low, the dyad may not be able to fully reap the benefit of quality conversations. However, when communication frequency increases, the effects of quality conversations become greater in lowering the perception of role stressors. Hence, we propose the following hypotheses. Figure 1 presents a conceptual diagram depicting the hypothesized moderating effects of communication frequency on the association of LMCQ with three role stressors.

OPEN IN VIEWER

Figure 1.

A conceptual diagram of the hypothesized moderating effect of communication frequency on the association between leader-member conversational quality (LMCQ) and role stressors.

Sample and Data Collection

The study was part of a larger study on foreign-born employees working in the United States. It employed an online survey method for data collection. We adopted the definition of “foreign born” by the U.S. Bureau of Labor Statistics (BLS, 2013), which refers to those who reside in the United States but were born “outside of the U.S. or one of its outlying areas such as Puerto Rico or Guam, to parents neither of whom was a U.S. citizen” (p. 5). Because of the challenges in reaching a large number of foreign-born employees at one time, we adopted a network sampling strategy. In exchange for extra credit, students, who were enrolled in the undergraduate classes of the school of communication in a public university in a large Midwestern metropolitan area, were asked to refer an online survey to potential participants in their social networks who were foreign-born and currently employed in a U.S. organization with a direct supervisor. We invited 263 students and received 262 survey responses in total. Following Jian et al. (2014), as a deterrent against referral protocol violation on the part of students, we asked students to supply their referral’s contact information, and, in the survey, participants were asked to supply the name of the student who referred them to the study. Also, to make sure that participants met the survey criteria, we included survey items to identify employment status and generational identity (foreign-born or later generation of immigrants). After deleting incomplete surveys and cases that did not meet the basic criteria for participation, 145 cases were retained for analysis.

In this sample, 55% were men and 45% women with a median age of 29 years (M = 34). The median organizational tenure of survey participants was about 3 years (M = 4.77) and median time with supervisor 2 years (M = 2.67). The employers that were represented ranged from small (less than 100 employees; 52%) and medium (between 100 and 500 employees; 17%) to large organizations (more than 500 employees; 31%). Four types of industry sectors had the highest representation: educational institution (16.8%), retail and wholesale (15.9%), health services (14.5%), and manufacturing (11%). Participants were from 57 countries and territories. Only four countries exceeded 4% in frequency: China (6.2%), India (5.5%), Russia (4.9%), and Mexico (4.8%). We compared the demographic characteristics of our sample with those of foreign-born workers in the United States as reported by the U.S. BLS in 2013. Although the gender makeup is similar, the comparison shows several key differences in other areas. Specifically, by gender, men accounted for 57.7% in the BLS data, compared with 55% in our sample. In the BLS data, Hispanics accounted for 47.8% of the foreign-born workers, compared with 11.8% in our sample. The educational level shown in the BLS data was lower than that in our sample. For example, 33.8% in the BLS data had a bachelor’s degree or higher, compared with 46.8% in our sample; 17.1% in the BLS data had some college, compared with 34.8% in our sample. In addition, by age, the proportion of the foreign-born labor force in the age bracket between 25 and 54 years (74.6%) in the BLS data was higher than that in our sample (60.2%). Thus, the differences in the demographic makeup have implications on the generalizability of our findings, which we address later in the study.

Measurement Instruments

Data Analysis

For preliminary analyses, we calculated univariate statistics and conducted correlational analyses. For primary analyses, we employed hierarchical regression analyses to test Hypotheses 1, 2, and 3. Specifically, we entered the control variables in Step 1, including age, sex, time with supervisor, perceived similarity, and perceived language proficiency. In Steps 2 and 3, we entered communication frequency and LMCQ, respectively.

To test Hypotheses 4, 5, and 6, we used Hayes’s (2014) PROCESS macro for SPSS to test the moderation effects and to further probe the moderation. Hayes’s PROCESS macro commands automatically calculate the product term (LMCQ × Communication Frequency) and estimate the moderation models. Using PROCESS has one additional advantage, which is the choice of using a percentile approach to generate visual representations of moderation effects and estimate simple slopes. In the past, the most common approaches include choosing one standard deviation below and above the mean, or median split or mean split, which may include values beyond the actual range of the observed values (Hayes, 2014). The percentile approach, instead, uses values of the moderator corresponding to the 10th, 25th, 50th, 75th, and 90th percentiles in the sample distribution of the moderator (Hayes, 2014). Simple slopes at these five percentile levels of the moderator were then estimated for statistical significance. Mean-centered values of the predictor variables were used for the tests of moderation.