Predicting Employee Dissent Expression in Organizations

Abstract

Employee dissent is understood as resulting from unsatisfying organizational states. Dissatisfaction can be felt but not expressed; dissent is the actual performance of disagreement with immediate circumstances. Constructive expression of dissent can have positive consequences, and therefore, much scholarly attention has been drawn to the phenomenon. In this article, we theorize that argumentativeness and verbal aggressiveness influence employee dissent behavior. Our particular contribution is that we hypothesize that these effects are mediated by assessments of the costs and benefits of dissent. We test whether the previously reported direct effects of argumentativeness and verbal aggressiveness are mediated by this cost–benefit analysis. A survey was conducted with undergraduate students and Amazon’s Mechanical Turk participants (N = 817). Structural equation modeling showed that the apparent direct effects of verbal aggressiveness and argumentativeness became insignificant after the mediators were included in the model. We conclude that cost and benefit analysis mediates the main effects.

Keywords

dissent voice argumentativeness verbal aggressiveness

When employees “feel apart from their organizations” (Kassing, 1997, p. 312), they may either endure the experience of feeling apart or speak up. Should they give voice to their thoughts, they may try to maintain the status quo, change the situation, or vent their emotions. Dissent can be potentially beneficial to organizations, and scholars recognize the value of employee dissent, challenging the traditional view that “management knows best” (Kassing, 2011).

Previous research working with the theory of independent mindedness suggested that Americans are disposed to express their dissent, and they express dissent to different audiences (e.g., Dozier & Miceli, 1985; Sprague & Ruud, 1988). Kassing (1997, 1998) proposed the theoretical model of organizational dissent that we expand on here. According to this model, once a triggering event exceeds an employee’s tolerance, he or she is likely to consider dissent expression. The expression of dissent can follow three paths: articulated, latent, and displaced dissent. Articulated dissent is the expression of dissent directly to supervisors, management, or people who have the ability to change the situation. Latent dissent, complaining to coworkers, is given its name because latent dissent usually is not seen by the organization’s management teams (Kassing, 1998). Other scholars use the term lateral dissent to refer to this type. Finally, displaced dissent is expressed to external parties (e.g., a spouse or a friend who does not work for the same organization).

We theorize that people will consciously or unconsciously consider the potential benefits and costs of expressing dissent. Possible benefits include having the situation improved or gaining respect as an assertive person (Morrison, 2011). However, for many, the potential costs are more forceful considerations. Expressing a complaint to a superior, coworker, or spouse might result in some sort of conflict. The other person might refute one’s remarks or might cross-complain in response (Gottman, Markman, & Notarius, 1977). The possibility of escalation may be salient (Paglieri, 2009), creating the further chance of being personally attacked. In other words, people understand that an organizational complaint could well be the first move in an interpersonal argument. Being anxious or optimistic about the prospect of having such an argument should be predictive of whether or not people will give voice to their concerns.

Several studies (Cionea, Hample, & Paglieri, 2011; Hample & Irions, 2014; Hample, Paglieri, & Na, 2012; Richards & Cionea, 2015) considered a cost–benefit formulation to predict whether or not people will choose to engage in an argument. Given the conceptual and subjective similarities between complaining and engaging in a reasonably clearly defined argument, we extend that cost–benefit model to the domain of dissent in the workplace. Argument engagement research has generated a list of subjectively estimated costs and benefits that have predicted argument engagement at substantial levels of accuracy. These studies also tested for the effects of argumentativeness and verbal aggressiveness on engagement decisions, with inconsistent results. Significant effects for personality traits have generally been mediated by the cost and benefit measures. In organizational contexts, costs and benefits of dissent expression are comparable in many subjective respects to arguments with friends or family members. Occupational costs can be as high as losing one’s job, and the benefits can be as high as promotion or salary increase in addition to the sense of winning.

Researchers have explored how individual dispositions influence how employees express their dissent (e.g., Sprague & Ruud, 1988), with special attention given to the role of aggressive communication traits (Kassing & Avtgis, 1999). Argumentativeness is an aggressive communication trait that refers to people’s tendency to argue over controversial issues, focusing on the substance of the argument (Infante & Rancer, 1982). Argumentativeness predicted articulated and latent dissent, but not displaced dissent, indicating that articulated dissent “represents a constructive attempt to voice dissent that entails some degree of argument” (Kassing & Avtgis, 1999, p. 110). Verbal aggressiveness is another type of aggressive communication trait that involves people’s tendency to attack the other’s self-concept rather than the argument’s merits (Infante & Wigley, 1986). In the Kassing and Avtgis (1999) study, verbal aggressiveness was found to predict latent dissent.

However, stable individual differences such as argumentativeness and verbal aggressiveness can only affect situated behavior—with a particular person, in a particular situation, at a particular time (Rancer & Avtgis, 2014). Contextual circumstances might facilitate personal expression or discourage it. Personality traits are thresholds for behavior, and the subjectively perceived situation determines whether the threshold for aggressive communication has been reached. We propose that subjectively estimated costs and benefits of expression are salient elements of the immediate situation and therefore should affect the likelihood and nature of an employee’s expression of dissent. Because we conceive of costs and benefits as situational features, we propose that they will mediate the effects of traits such as argumentativeness and verbal aggressiveness.

A Model of Employee Dissent

We propose that the effects of traits are mediated by cost–benefit estimates, and only through that mediation, do they affect the likelihood of expressing dissent in an organization. However, to support the case for our mediated model, we begin by expressing the hypotheses for direct effects between traits and dissent. We expect to discover these when the mediators are omitted from the analysis, but that they will disappear when the mediators are included in our model. Although confirming these main effects will replicate prior work, discovering that the effects are mediated will modify earlier conclusions.

The employee dissent literature suggested that dissent behavior happens when the incongruence between an employee’s desired and current state of affairs is elevated to the extent that it surpasses the person’s threshold for voice (Kassing, 1997; Redding, 1985). However, dissent behavior is risky because it could bring negative personal consequences, such as being viewed negatively as troublemaker, and in turn damage valued relationships (Milliken, Morrison, & Hewlin, 2003). Therefore, the decision-making process of voicing dissent is complex and is influenced by individual, relational, and organizational factors.

Aggressive Communication Traits

Kassing (1997) proposed that independent individual predispositions (i.e., values or characteristics that are imported or that existed outside members’ organizations) influence how people behave within organizations. For example, Sprague and Ruud (1988) found that people’s tendency to avoid conflict inhibited their willingness to perform dissent within organizations. Aggressive communication traits, argumentativeness and verbal aggressiveness, were significant predictors of articulated dissent (Infante & Gorden, 1985; Kassing & Avtgis, 1999).

Argumentativeness and verbal aggressiveness are two aggressive communication traits that, at high levels, predispose employees to engage in forceful communication behaviors (Infante & Gorden, 1985; Infante & Rancer, 1982; Infante & Wigley, 1986). Prior studies on both organizational dissent and argument engagement have been somewhat inconsistent in their results for argumentativeness and verbal aggressiveness, possibly suggesting that the direct effects of the two traits are minor or are somehow mediated (e.g., Cionea et al., 2011; Croucher et al., 2009; Goodboy & Myers, 2012; Hample & Irions, 2014; Hample et al., 2012).

Argumentativeness

Argumentativeness is constructive in nature. It is a personality trait that predisposes people to propose a controversial case, defend their positions, and refute others’ arguments on controversial issues (Infante & Rancer, 1982). Argumentativeness is a balance between impulses to approach or avoid conflict. Specifically, individuals who are high on tendency to approach argument and low on tendency to avoid argument are high argumentatives, and low argumentatives’ avoidant impulses exceed their motivation to approach arguments. In the context of organizational dissent, the natural prediction has been that high argumentatives, when they experience incongruence between the current and desired state of affairs, will exhibit more dissent expressive behaviors. The more subtle question has been whether they will tend to defend their positions and seek status quo changes by dissenting to potential change makers, or express themselves to people who cannot change conditions.

High argumentatives were found to be highly effective subordinates (Infante & Gorden, 1989), perhaps because those people actively engage in dissent behaviors that aim at changing unsatisfying situations. Similarly, high argumentativeness was associated with larger amounts of verbal behavior (Mortensen, Arntson, & Lustig, 1977). Assuming that complaints to superiors are seen as having the greatest potential for negative consequences, when frustration levels do not reach a level justifying these risks (e.g., of retaliation; Kassing, 1998), high argumentatives may complain to colleagues. In this way, they can voice their dissent, maintain psychological well-being (Finkenauer, Kuback, Engels, & Kerkhof, 2009), and possibly receive constructive suggestions from their colleagues. At the same time, they are protected against significant negative consequence because colleagues seldom have the motivation or ability to initiate negative actions toward the dissenter (Kassing, 1998). People with high levels of tendency to avoid arguments would naturally voice their dissent more to members outside the organizations, if they voice it at all. Non-organizational members such as family and friends may agree with the dissenter, meaning that he or she would not actually be initiating an argument by complaining to them. These people would also be able to provide emotional support, but they would not be able to change the unsatisfying situation.

Therefore, we propose,

Hypothesis 1: Argumentativeness will be positively related to articulated dissent.

Hypothesis 2: Argumentativeness will be positively associated with latent dissent.

Hypothesis 3: Tendency to avoid arguments will be positively related to displaced dissent.

Verbal aggressiveness

Verbal aggressiveness is a hostile, destructive interpersonal trait that involves hurting the other person psychologically by attacking his or her self-concept (Infante & Wigley, 1986). For example, people high in verbal aggressiveness attack others’ character or competence, and they insult, demonize, or tease other people (Infante, 1987). In the workplace, aggression is manifested more often in verbal or passive forms than physical ones. Increased organizational changes can be upsetting, and this could make verbal aggressiveness more salient and readier to activate. This produces more workplace aggression (Baron & Neuman, 1996). Infante and Wigley theorized that verbal aggression is especially likely to appear when people lack argumentative skills.

Moving these ideas to the domain of organizational voice, Kassing and Avtgis (1999) argued that employees express latent dissent (to coworkers) out of frustration, likely due to lack of argumentative skills. Prosociality (i.e., scores on the Prosocial Verbal Aggressiveness subscale) should contribute to articulated dissent, because the benevolently worded items represent prosocial cooperation, other-esteem, and support (Levine et al., 2004). These considerations should be weightier when applied to people who can change the unsatisfying situation (e.g., the dissenter’s immediate boss). Verbal aggressiveness should have no effect on displaced dissent, because people normally have no intention to attack family and friends’ self-concept with regard to their work issues.

Kassing and Kava (2013) found that prosocial upward dissent (i.e., the constructive type of articulated dissent) was positively correlated with open upward influence. People who are high on prosocial verbal aggressiveness are cooperative, and tend to use the confirming and “supportive, ego-enhancing” type of communication (Levine et al., 2004, p. 259), and may also be likely to make their concerns clear to their supervisors.

Therefore, we expect,

Hypothesis 4: Verbal aggressiveness will be positively associated with latent dissent.

Hypothesis 5: Prosocial verbal aggressiveness will be positively associated with articulated dissent.

Employee Dissent Expression: A Cost and Benefit Approach

Hypotheses 1 to 5 specify direct effects from traits to dissent behaviors. Most of these are replication hypotheses. Here, we propose a model that specifies how those direct effects are mediated by cost–benefit assessments (see Figure 1). We expect the direct effects to disappear or dissipate in this more fully specified model.

Figure 1.

The proposed mediated model.

We propose that individual predispositions will not influence employee dissent behavior invariably across situations. We conceptualize the situation as being subjectively constructed (Hample, 2015). As other scholars suggest, relational factors, which deal with the quality of the type of relationship employees maintained, should also contribute to employees’ dissent decisions (Kassing, 1997, 2008). These subjectively experienced elements of the organizational moment, many of which have implications for perceived costs and benefits, will make dissent expression a localized, momentary decision, not one dictated only by general traits.

Articulated dissent is change-oriented behavior that is essential and necessary for the healthy operation and growth of organizations (Redding, 1985), but it is also a complicated and risky decision that could bring negative personal consequences (Kassing, 1997, 1998, 2011; Milliken et al., 2003). Kassing (2011) argued that people’s tolerance for dissent expression is influenced by risk of retaliation, issue seriousness, and the likelihood that the issue is going to be addressed. This is consistent with our model because it posits that people assess costs (possible retaliation) and benefit (benefits brought by the resolved issue) when making dissent decisions. Moreover, like other human decision-making processes under uncertainty, speaking up about dissatisfying conditions in organizations can lead to gains and/or losses. As discussed below, many of the decisions are affected by perceived relational partners’ characteristics (e.g., their projected civility, their reasonability), which in turn influence the anticipated cost and benefit of the dissent behavior.

Our understanding of how costs and benefits are implicated in the decision to engage in dissent expression is based on Hample et al.’s (2012) work. In their cost and benefit model of the decision to engage in an argument, cost assessment includes effort, which involves psychological effort, emotional exposure, and anticipated negative relational consequences; (in)appropriateness, people’s evaluation of whether the possible argument should be undertaken with that person, in that situation, at that time; and whether the dissent target would be reasonable and expected to behave with civility. These subjective projections indicate whether the possible interaction would be demanding, unsuitable, or unproductive. In contrast, people may project potential gains: Benefit assessment involves employees’ evaluation of what might be gained, such as the possibilities that arguing might be satisfying, pleasant, and productive, another consideration is the perceived likelihood of winning (i.e., succeeding with the complaint), and finally, an estimate of whether or not the issue is resolvable. This list of cost–benefit factors is comparable in many respects with the list of reasons for avoiding conflict developed by Roloff and Wright (2009). Research indicates that cost–benefit assessments explain one third to one half the variance in people’s engagement intentions (e.g., Hample et al., 2012). However, individuals’ aggressive communication traits can influence their evaluations of the costs and benefits of dissent expression. In the following section, we address connections between communication traits and cost–benefit assessments.

Argumentativeness and assessment of costs and benefits

Employees’ argumentativeness may contribute to increased evaluations of the benefits and decreased evaluations of the costs of dissent expression. For example, Richmond, McCroskey, and McCroskey (1989) found that argumentativeness was positively correlated with self-perceived communication competence. Thus, highly argumentative employees may think they are more capable of giving information to others and making themselves understood. When people are confident about their ability to communicate, their evaluations of resolvability, the benefits, and likelihood of success in the dissent expression are likely to be higher than for people with less confidence. Relatedly, they may believe their ability to express themselves will make the conversation a civil and reasonable one. This reasoning is supported by Boster, Levine, and Kazoleas’s (1993) finding that people high in argumentativeness were more skillful in providing multiple reasons and arguing logically. These skills may also result in higher levels of perceived resolvability, likelihood of success, appropriateness, and benefits of dissent expression. This is also consistent with Kassing and Avtgis (1999), who found that high argumentatives had better argumentative skills.

Thus, people who rate higher in argumentativeness will expect dissent expression process to be less costly than those who are less argumentative. In other words, high argumentatives engage in and enjoy defending their positions on controversial issues (Boster et al., 1993), so that dissent expression is less effortful for them. Therefore, we propose,

Hypothesis 6: Argumentativeness will influence people’s evaluation of costs and benefits of dissent expression, such that argumentativeness will be positively related to resolvability, civility, reasonability, likelihood of success, appropriateness, and benefits, and negatively related to effort.

Verbal aggressiveness and assessment of costs and benefits

Verbal aggression is not competent communication behavior, and consequently should generate higher costs and fewer benefits than more skillful communication. Verbally aggressive people attack others’ characters, and tend to be regarded as less friendly, less likable, and more hostile (Infante, 1995; Martin & Anderson, 1996). Such people lack argumentative skills (Infante & Wigley, 1986), and so resort to ad hominem attacks. Without strong arguments and sound reasoning, those high in verbal aggressiveness should find and perhaps recognize that their advocated positions are not likely to be accepted by receivers, so that the perceived appropriateness, resolvability, likelihood of success, and benefits of the dissent expression are expected to be low. In addition, verbal aggressiveness was negatively associated with self-esteem (Rill, Baiocchi, Hopper, Denker, & Olson, 2009), which also implies that verbally aggressive people’s negative evaluations of their ability to argue should lead to low levels of perceived resolvability, success, and getting benefits via dissent expression. The essential nastiness of verbal aggression can become reciprocal. People who were attacked experienced psychological discomfort and communication dissatisfaction (Madlock & Kennedy-Lightsey, 2010), encouraging them to be unreasonable and uncivil. Thus, it is proposed,

Hypothesis 7: Verbal aggressiveness will influence people’s evaluation of costs and benefits of dissent expression, such that prosocial verbal aggressiveness will be positively associated with estimates that the episode will be resolvable, civil, reasonable, low in effort, likely to succeed, appropriate, and beneficial, whereas the opposite results will be negatively associated with the Antisocial subscale of verbal aggressiveness.

Hypotheses 6 and 7 posit links between the traits of argumentativeness and verbal aggressiveness on one hand, and cost–benefit estimates on the other. The following section considers the links between cost–benefit appraisals and dissent expression.

Cost and benefit assessments and dissent

Cost–benefit estimates involve situated relational considerations. For example, perceiving one’s dissent target as civil and reasonable will likely increase the subjective probability of establishing and maintaining a high-quality interpersonal relationship, should be thought to buffer against possible retaliations, should lower the expectations for negative consequences from dissent, and should consequently foster open and direct communication (Krone, 1992). Similarly, people’s evaluation of perceived appropriateness is also connected to the type and quality of the relationship maintained. In general, the cost and benefit assessments include elements that involve both relational concerns and some aspects of the triggering event itself, such as resolvability (can this matter be settled with this person in this moment?), effort (how hard will it be to express myself properly to this person right now?), likelihood of success (how good is my case, and how open to listening is the other person?), and anticipation of benefits (is there really any point in my pursuing this, here and now?). In other words, when employees decide whether or not they should express their dissent, they evaluate expected costs and benefits, both in regard to situated relational others and characteristics of the dissent behavior itself. Van Dyne and LePine (1998) argued that employees evaluate costs and benefits before speaking up, but did not actually test a cost and benefit model.

We propose that employees will be more likely to decide to engage in dissent expression when they expect substantial benefits, and less likely when costs are anticipated to increase (Hample et al., 2012). Synthesizing Hample et al.’s (2012) model of argument engagement and Kassing’s (1997) model of employee dissent, we hypothesize,

Hypothesis 8: Employees’ subjective assessments of costs and benefits determine their decision to engage in dissent expression, such that the tendency to voice dissent is highest when employees perceive the dissent expression process to be resolvable, civil, reasonable, low in effort, likely to be successful, appropriate, and beneficial.

The mediated model

Finally, we specify our overall model (Figure 1). We expect the direct effects reported to test Hypotheses 1 to 5 to disappear or weaken, but that the effects predicted in Hypotheses 6 to 8 will be present.

Hypothesis 9: Argumentativeness and verbal aggressiveness will affect estimates of costs and benefits (Hypotheses 6 and 7), cost and benefit estimates will predict dissent decisions (Hypothesis 8), and the effects of argumentativeness and verbal aggressiveness on dissent decisions will be fully mediated by the cost and benefit expectations.

In the course of testing Hypothesis 9, we will also investigate the falsifying possibility of direct effects between aggressive communication traits and dissent, effects that are not (or not fully) mediated by cost–benefit considerations.

Method

Respondents

A total of 457 undergraduate students from a large public mid-Atlantic university and another 421 Amazon Mechanical Turk (MTurk) workers were recruited to complete an online survey. Students were compensated with extra credit points and MTurk workers were compensated with US$0.50. Amazon’s MTurk workers have been found to be equally or more diverse and representative than other samples (e.g., student sample, traditional, or Internet samples), and the quality of data has commonly been shown to meet the psychometric standards for published research (Buhrmester, Kwang, & Gosling, 2011; Paolacci, Chandler, & Ipeirotis, 2010). Both samples were included in our analyses to yield a more diverse sample that had experiences of difference work settings (e.g., full-time jobs, internships, part-time jobs) and to have better statistical power.

Most of the sample had work experience, with 97.6% of the participants indicating they had either worked full-time, part-time, or had had an internship. Twenty-one participants (14 students and seven MTurk workers) did not have work experience, and 40 cases had significant amount of missing data. We excluded those cases in our analyses. The total sample size was therefore 817. Specifically, 43.1% of them had full-time jobs, 24.6% reported that they had part-time jobs, 17.7% of them had summer jobs, and 12.3% of them had internships (6.5% paid and 5.8% unpaid).

More females than males responded to the survey (40.9% males and 59.1% females), and their ages ranged from 18 to 73 (M = 27.59, median = 22, SD = 11.73). Exactly half of the respondents reported that they had supervised one or more people and were supervised by other people, 40.1% reported they worked only as subordinates, and 5.6% of them supervised one or more people and only answered to themselves.

Procedures

Respondents received a link that directed them to the questionnaire. They were first asked several demographic questions, including information about their previous working experience. After that, we asked them to respond to the measures for argumentativeness and verbal aggressiveness. Following the trait measures, three organizational change scenarios were described, in which respondents (i.e., the employees in the scenarios) had disagreement and may have felt that they needed to express dissent (see the appendix). The vignettes invited dissent, but did not require it. We randomly assigned participants to scenarios. After reading one of the scenarios, respondents were asked about their willingness to dissent and their evaluation of the costs and benefits of the behavior with the dissent target they indicated in mind.

Measures

All measures in the study used a 10-point Likert-type scale that ranges from 1 (almost never true) to 10 (almost always true). All of the Cronbach’s alphas were calculated based on the raw data for all items. Later, we reduced the number of items used in the structural equation models (SEM). Some items were reworded from their original phrasing so that they referred to complaints at work.

Argumentativeness

We measured argumentativeness using Infante and Rancer’s (1982) 20-item scale. By means of its two subscales, we separately measured individuals’ tendencies to approach or avoid argument (e.g., “I enjoy defending my point of view on an issue,” “I try to avoid getting into arguments”). Cronbach’s alphas were .87 for tendency to avoid argument and .90 for tendency to approach argument.

Verbal aggressiveness

We used Infante and Wigley’s (1986) 20-item Verbal Aggressiveness Measure. The scale measured the “personality trait that predisposes persons to attack the self-concepts of other people instead of, or in addition to, their positions on topics of communication” (Infante & Wigley, 1986, p. 61). Sample items were, “I try very hard to avoid having other people feel bad about themselves when I try to influence them” and “I try to make people feel good about themselves even when their ideas are stupid.” Cronbach’s alphas were .87 for the Prosocial subscale and .88 for Antisocial Verbal Aggressiveness.

Resolvability

We used Hample et al.’s (2012) six-item Resolvability Measure (Cronbach’s α = .84). The scale measured the degree to which the individual thinks the issue can be resolved. Sample items included, “I wouldn’t think we would ever agree on the issue” and “I wouldn’t see much chance that it would be resolved.”

Civility

We used Hample et al.’s (2012) 10-item civility measure (Cronbach’s α = .90). The scale measured the degree to which the individual thinks the issue can be discussed in a civil way. Sample items were, “If I express my complaint: the conversation would involve loud and negative voices” and “the conversation would involve cooperation by parties.”

Reasonability

We used Hample et al.’s (2012) six-item projected other’s reasonability measure (Cronbach’s α = .86). The scale measured the degree to which the other individual is perceived by the respondent to be reasonable. Sample items included, “he/she would be reasonable if I express my complaint” and “he/she would be mature about it if I express my complaint.”

Effort

We used Hample et al.’s (2012) 10-item Effort Measure (Cronbach’sα = .81). The scale measured how much effort the individual thinks the conversation would involve. Sample items included, “if I express my complaint, it would take a long time to settle anything” and “if I express my complaint, I would have to go to a lot of effort.”

Likelihood of success

We used Hample et al.’s (2012) eight-item Likelihood of Success Measure (Cronbach’s α = .70). The scale measured the degree to which the individual thinks he or she will have the ability to change the current situation into a satisfying one. Sample items were, “if I express my complaint, I think I would win” and “if I express my complaint, I think the one who I express my complaint to would see that my position was better than his/hers.”

Appropriateness

We used Hample et al.’s (2012) seven-item Appropriateness Measure (Cronbach’s α = .87). The scale measured the degree to which the individual thinks the issue is appropriate to discuss with the other person. Sample items were, “it would be appropriate to express my complaint—with this person, at this time, and in this place” and “it would be inappropriate to express my complaint with this person.”

Beneficial

We used Hample et al.’s (2012) six-item Benefits Measure (Cronbach’s α = .89). The scale measured the degree to which the individual thinks bringing up the issue will bring benefits to him or her. Sample items included, “I would be better off if I expressed my complaint” and “I would regret it if expressed my complaint.”

Articulated dissent

We used Kassing’s (1998) nine-item questionnaire to measure the individual’s willingness to voice articulated dissent. The measure was reworded to express behavioral intentions (Cronbach’s α = .90). The scale measured the extent to which the individual is willing to “share concerns directly and openly with management, supervisors, and corporate officers” (Kassing, 1998, p. 207). Sample items were, “I would be hesitant to raise questions or contradictory opinions in my organization” and “I would not question management.”

Displaced dissent

We used Kassing’s (1998) six-item questionnaire to measure willingness to voice displaced dissent (Cronbach’s α = .82). The scale measured the extent to which the individual is willing to “express dissent to audiences that are exclusively external to organizations” (Kassing, 1998, p. 213). Sample items were, “I would refuse to discuss work concerns at home” and “I would make it a habit not to complain about work in front of my family.”

Latent dissent

We used Kassing’s (1998) nine-item questionnaire to measure willingness to voice latent dissent (Cronbach’s α = .86). The scale measured the extent to which the individual is willing to discuss an issue “that involves complaining to coworkers and voicing criticism openly within organizations” (Kassing, 1998, p. 211). Sample items included, “I would criticize inefficiency in this organization in front of everyone” and “I would join in when other employees complain about organizational changes.”

Data Preparation

To prepare data for further data analysis, 21 cases without work experience were deleted, and missing data were deleted listwise, resulting in an effective sample size of 817. Data screening procedures were performed to examine possible violations of the assumptions of SEM. At the univariate level, all items had skewness and kurtosis statistics within the absolute value of 2 (Hancock & Mueller, 2010), so original data were used without transformation. In addition, residual plots of indicators were examined, and the results showed homoscedasticity and constant variance.

Data Analyses

All of the measures showed internal consistency by having alphas above .70 based on all items collected. Two-step SEM (Kline, 2010) was then used to test the proposed hypotheses and the structural model. Due to the requirement of model convergence and getting stable parameter estimates in SEM (i.e., maintaining a high ratio of sample size to free parameters), we chose the three best indicators for each latent variable (using their R² values with their latents) and retained them in subsequent analyses. Cronbach’s alphas were calculated again, and all passed the .70 threshold.

Mplus 7.0 was used to conduct the analysis. First, the initial measurement model with all the indicators was run. Based on the results of the all-item confirmatory factor analysis (CFA), the three best indicators were retained for further analyses. The reduced-item CFA model was respecified and evaluated. Second, the structural relationships were specified in the program. To evaluate the fit of the hypothesized structural model, we used the joint-criteria approach (Hu & Bentler, 1999), using four major model evaluation indices: chi-square test (χ² with degree of freedom), Comparative Fit Index (CFI), Standardized Root Mean Square Residual (SRMR), and the value of Root Mean Square Error of Approximation (RMSEA).

Results

Means, standard deviations, and correlations among the variables (measured by summing the reduced set of indicators for each variable) are summarized in Table 1.

Table 1.

Means, Standard Deviations, and Correlations (N = 817).

	M	SD	1	2	3	4	5	6	7	8	9	10	11	12	13
1. Tendency to avoid argumentativeness	6.47	2.06	—
2. Tendency to approach argumentativeness	5.16	2.08	−.50**	—
3. Prosocial verbal aggressiveness	4.02	1.92	.35**	.05	—
4. Antisocial verbal aggressiveness	3.56	1.92	−.17**	.26**	−.39**	—
5. Resolvability	5.71	1.01	−.05	.05	.05	−.09**	—
6. Civility	7.48	1.93	.07	−.10**	.28**	−.43**	.19**	—
7. Reasonability	6.52	1.90	.08*	.10**	.35**	−.17**	.16**	.44**	—
8. Effort	4.59	1.77	.11**	.02	−.11**	.26**	−.27**	−.58**	−.41**	—
9. Likelihood to success	6.30	1.55	.00	.16**	.27**	−.09*	−.02	.14**	.33**	−.09*	—
10. Appropriateness	7.05	1.99	−.05	.05	.18**	−.25**	.13**	.42**	.31**	−.47**	.30	—
11. Benefits	6.16	1.91	.01	.19**	.26**	−.09**	.16**	.26**	.57**	−.32**	.42**	.40**	—
12. Articulated dissent	6.00	1.97	−.05	.23**	.20**	−.05	.11**	.16**	.34**	−.13**	.39**	.32**	.47**	—
13. Displaced dissent	5.99	1.89	.09*	.05	.13**	.07	−.10**	−.06	.17**	.09*	.16**	.01	.08*	.02	—
14. Latent dissent	5.39	1.87	−.04	.15**	.00	.22**	−.08*	−.16**	.05	.15**	.19**	.02	.09**	.36**	.29**

Note. Correlations are based on latent variables of reduced items.

p < .05. **p < .01, two-tailed.

Measurement Model Testing

Fit indices for measurement and structural models are summarized in Table 2. The initial measurement model was based on all items collected. Fit indices of the initial measurement model showed that the model-implied covariance matrix did not fit the observed data covariance matrix very well, χ²(6578,N = 817) = 22,961.12, p < .001, RMSEA = .06, CFI = .71, SRMR = .09. In addition, although all of the items loaded on their intended latent factor, some items were not well explained by that latent factor (i.e., latent factors explained less than 50% of the variances in items). Following previous research (Park, Kim, & Krishna, 2014), to improve the measurement model and the ratio of sample size to free parameters, three items for each latent factor were retained for further analyses (see Table 3 for a list of items retained). To ensure that the three indicators retained were still caused by the same latent factors compared with the original indicators (i.e., the theorized latent factors did not change due to the trimming of items), the means of the three items were correlated with the means of the original items, and they were highly correlated, with Pearson’s rs ranging from .83 to .94.

Table 2.

Fit Statistics for Confirmatory Factor Analyses and Structural Models (N = 817).

	χ² (df)	RMSEA [95% CI]	CFI	SRMR
Measurement models
All items	22,961.12 (6,578)***	.055 [.054, .056]	.71	.09
Reduced items	1,612.276 (728)***	.039 [.036, .041]	.96	.04
Structural models
Main effect only model	455.94 (168)***	.046 [.041, .051]	.97	.04
Full mediation model	1,657.31 (740)***	.039 [.036, .041]	.95	.05
Partial mediation model with LD on VAA	1,633.69 (739)***	.038 [.036, .041]	.96	.04
Partial mediation model with all direct effects	1,612.276 (728)***	.039 [.036, .041]	.96	.04

Note. RMSEA = root mean square error of approximation; CI = confidence interval; CFI = comparative fit index; SRMR = standardized root mean square residual; LD = latent dissent; VAA = antisocial verbal aggressiveness.

***

p < .001, two-tailed.

Table 3.

Items Retained, in Standard Orderings.

	Combined sample
Tendency to avoid argument	3, 12, 19
Tendency to approach argument	9, 15, 17
Prosocial verbal aggressiveness	1, 3, 15
Antisocial verbal aggressiveness	2, 6, 7
Articulated dissent	5, 6, 7
Displaced dissent	3, 5, 6
Latent dissent	4, 6, 9
Resolvability	2, 3, 5
Civility	6, 7, 8
Reasonability	1, 3, 6
Effort	1, 4, 9
Likelihood to success	1, 7, 8
Appropriateness	2, 3, 4
Benefit	1, 3, 4

After the respecification, the measurement model became tenable based on Hu and Bentler’s (1999) criteria, χ²(728, N = 817) = 1,612.28, p < .001, RMSEA = .04, CFI = .96, SRMR = .04. In addition, all of the indicators were associated with the intended latent factors, with the latents accounting for more than half of the variance in each of their associated indicators. In summary, the CFA showed that the respecified measurement model fit the data well.

Structural Models

The hypothesized fully mediated structural model (Figure 1) was then run. In addition to the hypothesized structural relationships among the latent factors, the errors among the mediators (i.e., the costs and benefits factors) were allowed to covary. It is expected that the mediators could be caused by other exogenous variables other than the ones included in the model (i.e., verbal aggressiveness and argumentativeness), and the common themes of cost and benefit make correlations among these measures theoretically sensible.

Hypothesis Testing

Main effects model

To test Hypotheses 1 to 5, a structural model that allowed all direct paths from argumentativeness and verbal aggressiveness to the three types of dissent was run, aiming at further explorations of previously inconsistent results. This analysis, of course, omits the mediators specified in Figure 1. The fit indices of the model were good, χ²(168, N = 817) = 455.94, p < .001, RMSEA = .05, CFI = .97, SRMR = .04. The results (see Figure 2) indicated that people who were high in tendency to approach arguments and high in prosociality engaged in more articulated dissent (R² = .12). Both latent dissent and displaced dissent were predicted by antisocial verbal aggressiveness and prosocial verbal aggressiveness (R² = .06 for latent dissent; R² = .10 for displaced dissent). So Hypothesis 1 was partially supported, Hypotheses 4 and 5 were supported, and Hypotheses 2 and 3 were not supported.

Figure 2.

The structural diagram for main effects only model.

Fully mediated model

To test Hypotheses 6 to 9 (see Figure 1), the hypothesized fully mediated model with no direct effects was run. Results are in Figure 3. The data fit the model well, χ²(740, N = 817) = 1,657.31, p < .001, RMSEA = .04, CFI = .95, SRMR = .05. Significant amounts of variance were explained. More than a third of the variance—36%—was explained in articulated dissent, 14% was explained in displaced dissent, and 12% was explained in latent dissent. The amount of variance explained in articulated dissent was 3 times more than in the main effects model, and the amounts of variance explained in latent and displaced dissent were also larger than in the main effects model. However, we determined that the fully mediated model was actually not the best one to test Hypotheses 6 to 9.

Figure 3.

Structural diagram for full mediation model.

Partially mediated model with latent dissent on antisocial verbal aggressiveness

Modification indices of the full mediation model suggested that the path between antisocial verbal aggressiveness and latent dissent should be freed. This was consistent with the results of previous research (Kassing & Avtgis, 1999).

After freeing the path between antisocial verbal aggressiveness and latent dissent, the data fit the model better, Δχ²(1) = 23.62, p < .001, and the Akaike Information Criterion (AIC) difference between the two models (ΔAIC = 21.6) exceeded the straightforward decision rule of 10 (Burnham & Anderson, 2004). Variance explained in latent dissent changed a little: 16.8% of the variance was explained by the modified model, compared with 11.5% without this additional direct path. The path coefficient from antisocial verbal aggressiveness to latent dissent was .21, and the other significant paths did not change much (the biggest change was the path from civility to latent dissent: from .18 [p < .01] to .06 [p > .05]). In other words, adding this path made the model fit slightly better, but did not change the results to any great degree, compared with the fully mediated model. Therefore, as suggested by the modification indices and fit indices, the full mediation model plus the path between antisocial verbal aggressiveness and latent dissent was used to test the second set of hypotheses.

Hypothesis testing: Paths in the partially mediated model with latent dissent on antisocial verbal aggressiveness

The results partially supported Hypotheses 6 to 9 (see Figure 4), including our expectations about the direction of significant relationships.

Figure 4.

Structural diagram for partial mediation model with LD on VAA.

Specifically, in terms of aggressive communication traits’ influences on perceived costs and benefits (Hypothesis 6), tendency to approach arguments was found to predict perceived reasonability (b = .15), likelihood of success (b = .07), and benefits (b = .26). Tendency to avoid arguments was a significant predictor of perceived effort (b = .22) and appropriateness (b = .14). All these results were as predicted, except that tendency to approach and avoid arguments was not a significant predictor of resolvability and civility. Hypothesis 6 was partially supported.

Hypothesis 7 was also partially supported. Antisocial verbal aggressiveness predicted perceived resolvability (b = −.35), civility (b = −.46), effort (b = .29), and appropriateness (b = −.31), and prosocial verbal aggressiveness predicted perceived civility (b = .18), reasonability (b = .41), effort (b = .11), success (b = .22), appropriateness (b = .17), and benefits (b = .29). Most of these results were as predicted, except that high prosocial verbal aggressiveness did not result in estimates that the episode will be resolvable and low in effort, and high antisocial verbal aggressiveness did not result in estimates that the episode will be reasonable, likely to succeed, and appropriate.

Hypothesis 8, which concerned the relationships between cost–benefit estimates and dissent, was partially supported. Articulated dissent was predicted by resolvability (b = .10), reasonability (b = .11), effort (b = .34), perceived likelihood of success (b = .33), appropriateness (b = .16), and benefits (b = .32). Latent dissent was predicted by perceived reasonability (b = .15), effort (b = .35), appropriateness (b = .16), and likelihood of success (b = .22). Finally, displaced dissent was predicted by perceived reasonability (b = .29), effort (b = .29), and likelihood of success (b = .25). We expected a negative path from effort to the three types of dissent rather than the positive path we found.

In sum, Hypothesis 9 was partially supported here. The direct path between the Antisocial subscale of verbal aggressiveness and latent dissent (b = .21) was not predicted, and was inconsistent with our prediction of full mediation. However, the general pattern of results was that the trait effects on dissent were otherwise mediated by cost and benefit estimates, and this was the essential element of our theorizing.¹

Discussion

The study proposed and tested a model in which the effects of argumentativeness and verbal aggressiveness on organizational dissent were fully mediated by people’s assessments of costs and benefits. The goal of the study was to test whether the previously reported effects of argumentativeness and verbal aggressiveness on dissent behaviors were mediated by cost–benefit assessments. Our analysis provides support for this conclusion.

Theoretical Implications

Dissent is prompted by dissatisfaction with circumstances at work (Kassing, 1997). The immediate behavioral tendency resulting from that dissatisfaction derives from people’s worldviews and personality traits such as argumentativeness and verbal aggressiveness (Croucher et al., 2009; Kassing & Avtgis, 1999). Although dissatisfaction combined with aggressive impulses should lead to expression of dissent, people do not always act on their impulses. This study indicates that the impulse to complain was not merely affected by, but was almost completely mediated by, people’s cost–benefit analyses. In other words, situational features are important twice in this process: once in stimulating dissatisfaction, and then again in giving rise to a more nuanced appreciation of whether overt dissent would be a good or bad idea. The second process disciplines and perhaps temporarily reshapes a person’s inclination to pursue things aggressively.

We replicated many, but not all, of the direct effects between traits and dissent previously reported. However, when we introduced our cost–benefit panel of mediators, all but one of those direct effects disappeared. The predictive power of those traits was captured in our analysis by those traits’ effects on cost–benefit estimates. It was the costs and benefits that had the proximate effects on the decision to perform dissent. Our results provided evidence that aggressive communication traits influence dissent behaviors through people’s evaluations of costs and benefits, and implied that dissent expression is influenced by situated factors, not influenced by personality traits invariably. Previous work that has found significant relationships between individual differences and dissent behaviors should probably be reinterpreted to accommodate these situated mediating factors (e.g., Kassing & Avtgis, 1999; LePine & Van Dyne, 1998).

The unanticipated direct effect of antisocial verbal aggressiveness on latent dissent was statistically significant, but substantively and perhaps even statistically unimportant to our mediation model. However, it is interesting that in other work, antisocial verbal aggressiveness had by far the largest positive association with blurting, compared with the other three trait measures used here (Hample, Richards, & Skubisz, 2013). Blurters, by their nature, do not attend very much to situational elements, and simply say spontaneously what is on their mind (Hample et al., 2013). In statistical terms, this means that blurting may well be unmediated by any situational features. Thus, we may have found indirect evidence that latent dissent is characteristic of blurters. We interpret this negligible direct effect as pointing toward the possibility that blurting proclivity might be interesting to assess in an organizational context. For example, blurters might be the very people most likely to dissent without forecasting possible repercussions. We are not optimistic about finding generous effect sizes for blurting’s impact on dissent when the proper mediators are specified in the model, but perhaps other risky communication behaviors might be more immediately affected by naked impulse.

Limitations and Future Directions

Our use of cross-sectional data limited our ability to make causal claims. Our design did not permit experimental manipulations (e.g., of costs or benefits), nor do we have the sort of longitudinal data that permits changes at one point in time to be traced backward to earlier changes. SEM is essentially correlational, and so our conclusions about causality are based as much on our theories as on our empirical results. Correlations are perhaps more decisive in falsifying causal theories than in directly proving them. However, some of our time ordering is obvious: A personality trait has to exist prior to a person’s situated evaluations of cost–benefits and immediate decisions regarding dissent expression or suppression, for example. Therefore, we think that our generating theories are sound, and the research that supports them gives confidence in their most basic causal claims. Still, future research studies that use experimental and longitudinal design on these topics would be extremely valuable.

There are also operationalization issues. First, we used hypothetical scenarios as dissent-triggering events. Although we wrote the vignettes based on previous research (Kassing, 2011) and tried to cover events that were relevant to various parties (i.e., employee himself or herself, colleagues, or the organization), the scenarios may or may not be relatable to participants. Participants may react differently compared with the actual dissent-triggering events they experience at work. Second, dissent expression was measured based on self-reports of intentions, which may be mainly indicative of participants’ self-perceived behavioral intentions. Future research using observational methods that examine whether and how employees will actually express dissent when they experience unsatisfactory events at work might be helpful in exploring dissent expression.

Along with previous research, our results provided more evidence that dissent expression is situated behavior (Kassing & Avtgis, 1999). Employees’ aggressive communication traits (except the effect of antisocial verbal aggressiveness on latent dissent) did not influence dissent expression behaviors directly, but only through people’s estimates of cost and benefits. To get a more comprehensive understanding of dissent expression behaviors, future research could explore whether other individual differences found to be significant predictors (e.g., self-esteem) influence dissent expression behavior in similar ways (i.e., mediated by cost–benefit estimates). In addition, exploring potential moderators might be fruitful for dissent research. For example, is it possible that evaluations of potential cost and benefits only influence some people’s dissent expression decisions (e.g., individual differences of locomotors and assessors)? Do the relationships between employees’ estimated costs and benefits and dissent expression decisions depend on organizational climate (Morrison, Wheeler-Smith, & Kamdar, 2011), or the relationship between the dissenter and the dissent target (Li & Liao, 2014; Li & Tan, 2013)?

Practical Implications

Organizational dissent expression is essential in helping identify problems in organizations and increasing employee satisfaction (Kassing, 2011). Employees’ expression of perceived problematic operations or management procedures sometimes provides insightful opinions for the management, because these employees are the people who operationalize work procedures and interact with customers (Morrison, 2011). Moreover, employees who suppress their disagreement are found to be unhappy—keeping secrets is harmful to psychological health (Finkenauer et al., 2009), and will be harmful to organizational productivity.

Our results suggest that employee-perceived costs and benefits play powerful roles in the decision-making process bearing on dissent expression. Organizations that want to encourage dissent expression could do so by changing the distribution of true costs and benefits to employees. Making dissent actually beneficial will presumably lead workers to perceive it as such. Because (based on the present set of results) the balance of potential gains and losses is the proximal cause of dissent expression, managers can thereby encourage employee voice. As a practical matter, we cannot help but observe that real costs and benefits could also be manipulated in the other direction to make the workplace “quieter” for managers. This often natural impulse needs to be guarded against. Due to the importance of dissent, organizations should promote overt dissent by making employees feel their costs to express disagreement are low and the benefits are high. For example, likelihood of success, as the most powerful mediator, could be promoted by creating an organizational culture that readily accepts employee suggestions. Managers should also react to employee dissent with open minds, encouraging future dissent expressions.

Footnotes

Appendix

[EACH RESPONDENT READS ONE SCENARIO]

We want you to think about a particular event that you have encountered at work. We can only give an outline of the sort of thing we would like you to remember and think about. Please read our little sketch, and then fill it in with an actual experience you have had, as best you can. All the rest of the questions in this survey will concern themselves with that experience.

Dissent scenarios

[1. Organization change/dissent about immediate boss related to yourself]

Your department has gone over a lot of changes recently, including new colleagues, new department manager, and some new work procedures. Your new manager has a working style that’s totally different from the previous one, and he or she has very high and strict expectations. Specifically, he or she needs you to launch a new training program for the new employees, but it is the busiest and most stressful season of the year, you need to spend much extra after work hours to get the work all done. You think it’s impossible to finish launching the program at this time of the year.

[2. Organization change/dissent about company procedures of employee treatment]

Your company has gone over many changes recently, including new working procedures, new colleagues, and a new CEO. The new CEO has a totally different working style from the previous one, and made some new changes to the paid leave program. Specifically, the total paid leave days are cut from 25 days to 15 days for the marketing team. Although you are not a member of the marketing team and your paid leave program will be the same as before, you think it’s unfair to have the cut for the marketing team.

[3. Organizational change/dissent about inefficiency procedures]

Your department has gone over a lot of changes recently, including new colleagues, a new department manager, and some new work procedures. Your new manager has a working style that’s totally different from the previous one, and he or she likes to schedule more meetings than you think are necessary. Most of the time, the meetings are boring and pointless. Although it doesn’t influence how much time you spend at work or your salary, you think it doesn’t make sense to have so many meetings that takes away the time that can be used to accomplish tasks, which is extremely inefficient.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biographies

Mengqi (Monica) Zhan (MA, University of Miami) is a doctoral candidate in the Department of Communication at University of Maryland, College Park, the United States. Her main research interests include workplace dissent and multicultural work teams.

Dale Hample (PhD, University of Illinois) is an associate professor in the Department of Communication at University of Maryland, College Park, the United States. His main research interests center on interpersonal arguing.

References

Baron

R. A.

Neuman

J. H.

(1996). Workplace violence and workplace aggression: Evidence on their relative frequency and potential causes. Aggressive Behavior, 22, 161-173. doi:10.1002/(sici)1098-2337(1996)22

Boster

F. J.

Levine

Kazoleas

(1993). The impact of argumentativeness and verbal aggressiveness on strategic diversity and persistence in compliance-gaining behavior. Communication Quarterly, 41, 405-414. doi:10.1080/01463379309369901

Buhrmester

Kwang

Gosling

S. D.

(2011). Amazon’s mechanical turk: A new source of inexpensive, yet high-quality, data? Perspectives on Psychological Science, 6, 3-5. doi:10.1177/1745691610393980

Burnham

K. P.

Anderson

D. R.

(2004). Multimodel inference: Understanding AIC and BIC in model selection. Sociological Methods & Research, 33, 261-304. doi:10.1177/0049124104268644

Cionea

I. A.

Hample

Paglieri

(2011). A test of the argument engagement model in Romania. In Zenker

(Ed.), Argumentation: Cognition and community: Proceedings of the 9th international conference of the Ontario Society for the Study of Argumentation. Retrieved from http://scholar.uwindsor.ca/ossaarchive/OSSA9/papersandcommentaries/6/

Croucher

S. M.

Braziunaite

Homsey

Pillai

Saxena

Saldanha

. . . Agarwal

(2009). Organizational dissent and argumentativeness: A comparative analysis between American and Indian organizations. Journal of Intercultural Communication Research, 38, 175-191. doi:10.1080/17475759.2009.505057

Dozier

J. B.

Miceli

M. P.

(1985). Potential predictors of whistle-blowing: A prosocial behavior perspective. Academy of Management Review, 10, 823-836. doi:10.5465/amr.1985.4279105

Finkenauer

Kubacka

Engels

R. C. M. E.

Kerkhof

(2009). Secrecy in close relationships: Investigating its intrapersonal and interpersonal effects. In Afifi

Afifi

(Eds.), Uncertainty, information management, and disclosure decisions (pp. 301-319). New York, NY: Routledge.

Goodboy

A. K.

Myers

S. A.

(2012). Instructional dissent as an expression of students’ verbal aggressiveness and argumentativeness traits. Communication Education, 61, 448-458. doi:10.1080/03634523.2012.699635

10.

Gottman

Markman

Notarius

(1977). The topography of marital conflict: A sequential analysis of verbal and nonverbal behavior. Journal of Marriage and Family, 39, 461-477. doi:10.2307/350902

11.

Hample

(2015). A theory of interpersonal goals and situations. Communication Research. Advance online publication. doi:10.1177/0093650214565899

12.

Hample

Irions

A. L.

(2014). Argument engagement under invitational versus demanding conditions. In Palczewski

C. H.

(Ed.), Disturbing argument (pp. 149-154). London, England: Taylor & Francis.

13.

Hample

Paglieri

(2012). The costs and benefits of arguing: Predicting the decision whether to engage or not. In van Eemeren

F. H.

Garssen

(Eds.), Topical themes in argumentation theory: Twenty exploratory studies (pp. 307-322). New York, NY: Springer. doi:10.1007/978-94-007-4041-9_20

14.

Hample

Richards

A. S.

Skubisz

(2013). Blurting. Communication Monographs, 80, 503-532. doi:10.1080%2F03637751.2013.830316

15.

Hancock

Mueller

R. O.

(Eds.). (2010). The reviewer’s guide to quantitative methods in the social sciences. New York, NY: Routledge.

16.

L. T.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6, 1-55. doi:10.1080/10705519909540118

17.

Infante

D. A.

(1987). Aggressiveness. In McCroskey

J. C.

Daly

J. A.

(Eds.), Personality and interpersonal communication (pp. 157-192). Newbury Park, CA: Sage.

18.

Infante

D. A.

(1995). Teaching students to understand and control verbal aggression. Communication Education, 44, 51-64. doi:10.1080/03634529509378997

19.

Infante

D. A.

Gorden

W. I.

(1985). Superiors’ argumentativeness and verbal aggressiveness as predictors of subordinates’ satisfaction. Human Communication Research, 12, 117-125. doi:10.1111/j.1468-2958.1985.tb00069.x

20.

Infante

D. A.

Gorden

W. I.

(1989). Argumentativeness and affirming communicator style as predictors of satisfaction/dissatisfaction with subordinates. Communication Quarterly, 37, 81-90. doi:10.1080/01463378909385529

21.

Infante

D. A.

Rancer

A. S.

(1982). A conceptualization and measure of argumentativeness. Journal of Personality Assessment, 46, 72-80. doi:10.1207/s15327752jpa4601_13

22.

Infante

D. A.

Wigley

C. J.

(1986). Verbal aggressiveness: An interpersonal model and measure. Communications Monographs, 53, 61-69. doi:10.1080/03637758609376126

23.

Kassing

J. W.

(1997). Articulating, antagonizing, and displacing: A model of employee dissent. Communication Studies, 48, 311-332. doi:10.1080/10510979709368510

24.

Kassing

J. W.

(1998). Development and validation of the organizational dissent scale. Management Communication Quarterly, 12, 183-229. doi:10.1177/0893318998122002

25.

Kassing

J. W.

(2008). Consider this: A comparison of factors contributing to employees’ expressions of dissent. Communication Quarterly, 56, 342-355. doi:10.1080/01463370802240825

26.

Kassing

J. W.

(2011). Dissent in organizations. Cambridge, UK: Polity Press.

27.

Kassing

J. W.

Avtgis

T. A.

(1999). Examining the relationship between organizational dissent and aggressive communication. Management Communication Quarterly, 13, 100-115. doi:10.1177/0893318999131004

28.

Kassing

J. W.

Kava

(2013). Assessing disagreement expressed to management: Development of the upward dissent scale. Communication Research Reports, 30, 46-56. doi:10.1080/08824096.2012.746225

29.

Kline

R. B.

(2010). Principles and practice of structural equation modeling (3rd ed.). New York, NY: Guilford Press.

30.

Krone

K. J.

(1992). A comparison of organizational, structural, and relationship effects on subordinates’ upward influence choices. Communication Quarterly, 40, 1-15. doi:10.1080/01463379209369816

31.

LePine

J. A.

Van Dyne

(1998). Predicting voice behavior in work groups. Journal of Applied Psychology, 83, 853-868. doi:10.1037//0021-9010.83.6.853

32.

Levine

T. R.

Beatty

M. J.

Limon

Hamilton

M. A.

Buck

Chory-Assad

R. M.

(2004). The dimensionality of the Verbal Aggressiveness Scale. Communication Monographs, 71, 245-268. doi:10.1080/0363452042000299911

33.

A. N.

Liao

(2014). How do leader–member exchange quality and differentiation affect performance in teams? An integrated multilevel dual process model. Journal of Applied Psychology, 99, 847-866. doi:10.1037/a0037233

34.

A. N.

Tan

H. H.

(2013). What happens when you trust your supervisor? Mediators of individual performance in trust relationships. Journal of Organizational Behavior, 34, 407-425. doi:10.1002/job.1812

35.

Madlock

P. E.

Kennedy-Lightsey

(2010). The effects of supervisors’ verbal aggressiveness and mentoring on their subordinates. Journal of Business Communication, 47, 42-62. doi:10.1177/0021943609353511

36.

Martin

M. M.

Anderson

C. M.

(1996). Argumentativeness and verbal aggressiveness. Journal of Social Behavior and Personality, 11, 547-554. doi:10.1177/0021943609353511

37.

Milliken

F. J.

Morrison

E. W.

Hewlin

P. F.

(2003). An exploratory study of employee silence: Issues that employees don’t communicate upward and why. Journal of Management Studies, 40, 1453-1476. doi:10.1111/1467-6486.00387

38.

Morrison

E. W.

(2011). Employee voice behavior: Integration and directions for future research. Academy of Management Annals, 5, 373-412. doi:10.1080/19416520.2011.574506

39.

Morrison

E. W.

Wheeler-Smith

S. L.

Kamdar

(2011). Speaking up in groups: A cross-level study of group voice climate and voice. Journal of Applied Psychology, 96, 183-191. doi:10.1037/a0020744

40.

Mortensen

C. D.

Arntson

P. H.

Lustig

(1977). The measurement of verbal predispositions: Scale development and application. Human Communication Research, 3, 146-158. doi:10.1111/j.1468-2958.1977.tb00513.x

41.

Paglieri

(2009). Ruinous arguments: Escalation of disagreement and the dangers of arguing. In Ritola

(Ed.), Argument cultures: Proceedings of OSSA 2009. Retrieved from http://scholar.uwindsor.ca/ossaarchive/OSSA8/papersandcommentaries/121/

42.

Paolacci

Chandler

Ipeirotis

P. G.

(2010). Running experiments on Amazon Mechanical Turk. Judgment and Decision Making, 5, 411-419. doi:10.2139/ssrn.2100631

43.

Park

S. H.

Kim

J. N.

Krishna

(2014). Bottom-up building of an innovative organization: Motivating employee intrapreneurship and scouting and their strategic value. Management Communication Quarterly, 28, 531-560. doi:10.1177/0893318914541667

44.

Rancer

A. S.

Avtgis

T. A.

(2014). Argumentative and aggressive communication (2nd ed.). New York, NY: Peter Lang.

45.

Redding

W. C.

(1985). Rocking boats, blowing whistles, and teaching speech communication. Communication Education, 34, 245-259. doi:10.1080/03634528509378613

46.

Richards

A. S.

Cionea

I. A.

(2015). Extending the argument engagement model: Expected utility and interacting traits as predictors of the intend to argue with friends. Journal of Argumentation in Context, 4, 110-133. doi:10.1075/jaic.4.06ric

47.

Richmond

V. P.

McCroskey

J. C.

McCroskey

L. L.

(1989). An investigation of self-perceived communication competence and personality orientations. Communication Research Reports, 6, 28-36. doi:10.1080/08824098909359829

48.

Rill

Baiocchi

Hopper

Denker

Olson

L. N.

(2009). Exploration of the relationship between self-esteem, commitment, and verbal aggressiveness in romantic dating relationships. Communication Reports, 22, 102-113. doi:10.1080/08934210903061587

49.

Roloff

M. E.

Wright

C. N.

(2009). Conflict avoidance: A functional analysis. In Afifi

Afifi

(Eds.), Uncertainty, information management, and disclosure decisions (pp. 320-340). New York, NY: Routledge.

50.

Sprague

J. A.

Ruud

G. L.

(1988). Boat-rocking in the high technology culture. American Behavioral Scientist, 32, 169-193. doi:10.1177/0002764288032002009

51.

Van Dyne

LePine

J. A.

(1998). Helping and voice extra-role behaviors: Evidence of construct and predictive validity. Academy of Management journal, 41, 108-119. doi:10.2307/256902