Abstract
Perceptions of justice are an important consideration for organizations as past research has shown that when employees feel they are treated fairly, positive outcomes result and negative outcomes can be avoided. The current study examined the effects of private and public delivery of positive and negative feedback on perceptions of procedural and interactional justice. Although past research found no differences for channel when comparing face-to-face to email, current results indicate that differences existed across channel characteristic such that private delivery was viewed as more procedurally and interactionally just than public delivery. Positive feedback was also viewed as more procedurally and interactionally just than negative feedback. This may indicate to practitioners that a private channel is more appropriate for feedback and that giving positive feedback may help engender perceptions of just treatment in employees.
Introduction
One of the authors of this research worked for an organization that gave out commendations and monetary awards for “going above and beyond” in secret. That is, there were never organizational announcements about who had worked hard enough to receive one of these awards, although they were being given out regularly, and some awardees were even advised to keep their award secret from other employees. This led to the question: why would an organization not publicize positive feedback? One possible reason could be a concern about employees’ perceptions of justice. This is a legitimate concern for organizations, as perceptions of fair treatment have been shown to affect important outcomes in ways such as increasing organizational citizenship behavior (Moorman, 1991) and outcome satisfaction, job satisfaction, trust, and organizational commitment (Colquitt, Conlon, Wesson, Porter, & Ng, 2001) and decreasing police misconduct (Wolfe & Piquero, 2011). Researchers and practitioners alike can be well served by learning more about the conditions that engender the perceptions of fairness, which increase positive and minimize negative outcomes. This research sought to determine some of the communication factors that may lead to perceptions of procedural and interactional justice in organizations. These factors may include valence of feedback delivered to a worker in an organization and the level of privacy with which the message was delivered. A review of procedural and interactional justice, positive and negative feedback, and channel privacy follow, as well as method, results, and discussion of the results.
Organizational Justice
When employees perceive their supervisors’ actions as fair, other positive organizational outcomes, such as altruism, courtesy, sportsmanship, and conscientiousness (Moorman, 1991), and lower turnover (Jones & Skarlicki, 2003), follow. Organizational justice has become an increasingly popular topic for researchers over the past two decades (Byrne & Cropanzano, 2001; Colquitt & Greenberg, 2003). Researchers have conceptualized and studied three different types of justice: distributive, procedural, and interactional justice. Distributive justice consists of perceptions of fairness pertaining to outcomes such as pay, rewards, or promotions (Colquitt, Greenberg, & Zapata-Phelan, 2005), whereas procedural justice consists of perceptions of fairness pertaining to the procedures used to reach the outcomes (Colquitt et al., 2005). Interactional justice consists of justice pertaining to interpersonal treatment (Colquitt et al., 2005). Although much research has been conducted on all three types of organizational justice, Cropanzano, Byrne, Bobocel, and Rupp (2001) noted that procedural and interactional have received more attention than distributive justice. This may be because some studies examining distributive justice have shown that when paired with either procedural or interactional justice, distributive justice is only associated with negative outcomes such as organizational retaliation (Skarlicki & Folger, 1997) or turnover (Jones & Skarlicki, 2003) when the other types of justice are low. In other words, if procedural or interactional justice is perceived to be high, then distributive justice seems to be a moot point. In addition, both procedural and interactional justice are more easily modified by organizations than is distributive justice. Especially in poor economic times, organizations may not be able to give employees more money or other incentives, but they can attempt to use fair procedures and to treat their employees fairly during interpersonal interactions. Because of this, the current study focuses specifically on procedural and interactional justice.
Procedural justice
Procedural justice consists of judgments about the fairness of the process used to reach an outcome (Cropanzano et al., 2001). Numerous positive effects have been attributed to perceptions of procedural justice in organizations. For example, fair procedures have been correlated with trust in management (Barling & Phillips, 1993) and job satisfaction and organizational commitment (Pettijohn, Pettijohn, Taylor, & Keillor, 2001). Fair procedures were found to ease acceptance of smoking bans (Greenberg, 1994) and pay systems (Miceli, 1993; Miceli & Lane, 1991). Even disciplinary actions are easier to accept if they are done using fair procedures (Ball, Trevino, & Sims, 1994). Given this research evidence, it seems natural that organizations would strive to ensure the use of procedures perceived to be fair, particularly when addressing sensitive organizational issues such as innovation, performance appraisal, and termination. However, it may be challenging for organizations to ensure fair procedures because it is not clear what characteristics of the procedure lead to the perception of fairness. Yamaguchi (2005) found that “soft” (e.g., flattery, praise) and rational interpersonal tactics were positively correlated with perceptions of procedural justice, but there is still more work to be done in this area. For example, changes in the perceptions of fairness could stem from the position of the person delivering a message (e.g., immediate supervisor, coworker, head of organization, external consultant), the timing of the delivery of the message, the transparency of the decision-making process, the valence of the message, and/or the relative publicity or privacy of the delivery of the message. If organizations cannot identify the source(s) of perceptions of fairness, it may be difficult for them to modify procedures successfully.
Interactional justice
Interactional justice (Bies, 2001) is defined as how fairly treated one feels during an interpersonal interaction and is also referred to as “interpersonal justice.” Chory and Kingsley Westerman (2009) cited prior research that denoted that interactional fairness occurs in conjunction with positive organizational outcomes and behaviors whereas the absence of interactional fairness occurs in conjunction with negative organizational outcomes and behaviors. This provides one reason why organizations may want to learn how to engender perceptions of interactional fairness. A second reason interactional justice may be useful to supervisors is that they may have more control over their interaction with employees than over the procedures used to deliver feedback or the type of outcomes that employees experience (Skarlicki & Folger, 1997). The interaction provides supervisors with an opportunity to be seen as fair when delivering feedback. Similar to procedural justice, the antecedents to perceptions of interactional fairness are not entirely clear. For example, one potential antecedent could be the employee’s impression of the person who is delivering a message to them. In many cases, this will be their supervisor.
Both types of justice seem important for maintaining a productive workplace, and as discussed above, the supervisor may be able to control some antecedents of justice, but others may be beyond the supervisor’s control. The current article examines two potential factors for perceptions of procedural and interactional justice: message valence of a piece of feedback, which may be beyond the supervisor’s control, and choice of private or public medium to deliver the message, which may be within the supervisor’s control.
Message Valence
The current study focused on how a receiver perceives positive and negative feedback in an organizational context. Feedback is a report of information on whether a standard has been met or missed by the receiver of the feedback (Westerman, 2008). Some organizations deliver this type of information once or twice a year in a performance appraisal. However, other organizations may deliver feedback less formally, such as whenever a worker has a success or a failure. For example, when a worker presents an idea for innovation to his/her boss, if it is a good idea, the worker may receive praise, and if it is an unworkable idea, the worker may be discouraged from suggesting ideas in the future.
Whether the feedback is positive or negative can have an effect on the receivers and their perceptions of the feedback and the sender. For example, Jacobs, Jacobs, Feldman, and Cavior (1973) found that positive feedback is seen as more credible, more desirable, and as having more impact than negative feedback. Kacmar, Wayne, and Wright (1996) found that positive feedback elicits more positive supervisor impressions than negative feedback. Fedor, Eder, and Buckley (1989) found employees had more motivation to improve, higher satisfaction with feedback, and higher perceptions of feedback utility when the feedback they received was positive versus negative. These studies speak to the general tendency of people to respond better to positive than negative feedback. Positive feedback is likely preferred to negative feedback because people do not like to hear bad news (Sussman & Sproull, 1999) or deliver it (Timmerman & Harrison, 2005; see also see MUM effect; Dibble & Levine, 2010; Rosen & Tesser, 1970).
Westerman and Westerman (2010) found that both distributive and interpersonal (also known as interactional) justice were perceived to be higher when a feedback message was delivered with face-saving content rather than without it. It is our contention that the preference for positive feedback over negative feedback and the finding that perceptions of justice were higher with relatively more positive feedback (as found by Westerman & Westerman, 2010) will translate into perceptions of justice as follows:
Hypothesis 1: Positive messages will lead to perceptions of more procedural justice than negative messages. Hypothesis 2: Positive messages will lead to perceptions of more interactional justice than negative messages.
Privacy and Publicity
Westerman and Westerman (2010) also previously examined the differences between face-to-face (FtF) and email delivery of negative feedback messages and found no differences in receivers’ impressions of justice based on channel. However, as Walther and Ramirez (2010) recently pointed out, advancement in the field of computer-mediated communication is more likely to come by consideration of theoretically relevant, general characteristics of channels, rather than simply comparing one channel with another. This highlights a weakness in Westerman and Westerman’s previous study of perceptions of justice: The two channels examined were seemingly chosen simply because FtF is considered a “gold standard” channel and email is considered a more limited channel. Given the knowledge of people’s ability to communicate on more limited channels (i.e., research on Social Information Processing Theory; Walther, 1992), it may be that Westerman and Westerman found no differences for channel because they did not examine a relevant difference between their channel choices. To that end, the current study examines a characteristic of communication channels that seems likely to impact how messages sent through them are processed: whether the channel is public or private.
O’Sullivan (2005) suggested that technology has outpaced classic distinctions between mass and interpersonal (IP) communication. He pointed out that we now frequently use mass channels for IP messages (e.g., marriage proposals that take place on large screens at sporting events) and supposedly IP channels for mass messages (e.g., so-called robo-calls and spam emails). He also suggested that it might be more advantageous to look at new distinctions and classifications of communication, and one major classification he suggested was message access exclusivity (i.e., public vs. private messages). He argued that this is likely a variable that will affect communication greatly and that consideration of this variable helps separate message and function from channel. Thus, rather than assuming that richer channels allow for more just delivery of messages as did Timmerman and Harrison (2005), this viewpoint examines how the private or public delivery of a message affects perceptions of justice, allowing for consideration of broader outcomes that will help stimulate growth in the field of computer-mediated communication as suggested by Walther and Ramirez (2010).
Other research that also suggests that public versus private disclosure is likely to make a difference is Petronio’s (2002) work on Communication Privacy Management Theory. Petronio suggests that the management of private information has a major impact on relationships. Petronio also suggests that people come to manage privacy in relationships by forming rules about the disclosure of information within their relationships. This creation of boundaries suggests that people like to be in control of that information. Notably, Petronio centers the navigation of these rules on self-disclosure and suggests that there are shared boundaries for information that is “co-owned”—that is, “private” information that must be kept so by multiple individuals. If another person chooses to disclose co-owned information about you, this may be a violation of the rules or expectations of rules for the relationship. When someone else (e.g., your supervisor) decides to disclose information about you publicly, it takes the decision rule out of your hands, and this may lead to reduction in the perceived justice of the dissemination of the information. This leads to the third and fourth hypotheses of the current article:
Hypothesis 3: Private messages will lead to perceptions of more procedural justice than public messages. Hypothesis 4: Private messages will lead to perceptions of more interactional justice than public messages.
It is possible that the valence of the message will interact with the privacy of channel used for judgments about justice. However, it is not clear exactly what that pattern will look like. Sussman and Sproull (1999) suggest that senders prefer to deliver bad news electronically because of the “social buffer” provided by the channel as compared to FtF delivery. They found that participants (who were sending the information) had higher levels of satisfaction and comfort when delivering negative information via computer-mediated communication than when using a telephone or FtF delivery. O’Sullivan (2000) also found that people prefer using mediated channels for negatively valenced messages, especially when the message threatens their own face (rather than another’s face). Both these studies suggest interactions are possible between channel and valence. Sussman and Sproull (1999) point out their work focuses only on senders to the exclusion of receivers, whom we investigate here. This leads to the first research question of the article.
Research Question 1: How do privacy of channel and valence of message interact to impact perceptions of procedural and interactional justice?
Method
Overview
A 2 (positive valence vs. negative valence) × 2 (private vs. public delivery) fully crossed experimental design was used to examine the relationship of valence of message and publicity of the message on the perceived procedural and interactional justice of the scenario.
Participants
Data were collected from both students and working adults. To be considered working adults, participants had to meet two requirements: they had to work more than 35 hours per week and they had to have worked at their current position for more than 1 year. Overall, the sample comprised 214 people: 124 students and 90 full-time working adults. For the student sample, the average age was 20.94 (SD = 1.66) with a range of 19 to 32. Sixty-three (50.8%) of the students were male. One hundred and twelve (90.3%) reported Caucasian, with 7 (5.6%) African Americans, 5 (2.4%) Hispanic, and 2 (1.6%) Other.
For the working adults sample, the average age was 27.12 (SD = 10.61) with a range of 19 to 54. Fifty-seven (63.3%) of the working adults were male. Seventy-eight (86.7%) reported Caucasian, with 2 (2.2%) African American, 2 (2.2%) Native-American, 1 (1.1%) Pacific Islander, 5 (5.6%) Middle Eastern, and 2 (2.2%) Other. Among the 90 working adults, the years worked in their current job ranged from 1 to 30, M = 4.01, SD = 5.28. The weekly hours worked ranged from 35 to 70, M = 42.09, SD = 5.42.
The two samples were compared in terms of the dependent variables in all tests; no significant differences were found between the students and working adults. Therefore, the two samples were collapsed into one group for the remaining analyses.
Procedure
Participants were handed a paper-and-pencil survey and were asked to complete it themselves or to request participation from an acquaintance who works more than 35 hours a week and had been at their job for at least 1 year.
Channel and Message Content
Participants were randomly assigned to one of four conditions. They received one of four possible messages, crossing valence (positive/negative) with channel (public/private). The messages were introduced with the following text:
Imagine you had submitted a proposal for a project you would like to do at work in response to an organizational request for proposals from all employees at your workplace. The proposals were all due at the same time and were competitively evaluated because only the top 10% would be able to be funded. The date for the announcement of the winners arrives and you receive word about your project . . .
Following this text were two more sentences with the private/public manipulation, which was achieved by the message being sent via email (private) or posted on a company website (public). For example, in the private condition, it stated, “Your supervisor sends out emails notifying individuals about their projects. You check your email and receive the following message.” Following those two sentences was the message manipulation. Regardless of the delivery channel, the acceptance message (positive valence) read as follows: “Congratulations! I am happy to let you know that your project was approved for funding!” and the rejection message (negative valence) read as follows: “Unfortunately, your project was rejected and will not be funded.”
The messages were designed for their brevity and parallel structure. The brevity of the messages served to ensure that any differences in our dependent variables could be attributed to the specific intended manipulations rather than to extraneous phrases that might otherwise be added to increase mundane realism. The parallel structure of the messages also provides a tight experimental design, which ensures that differences in our dependent variables can be attributed to the manipulations rather than differences in the structure of the messages.
Measures
Procedural justice
Participants answered five semantic differential items to indicate whether they thought the procedure used to notify them about their project success was fair. First participants read this lead-in: “The procedure used to notify me about my project was . . . ” and then a list of semantic differentials including “Unfair . . . Fair” and “Unjust . . . Just.” Higher scores indicated things were perceived as more fair. Confirmatory factor analysis (CFA) was performed on this scale and all five items formed a unidimensional solution with a reliability (Cronbach’s α) of .89.
Interactional justice
Participants also answered five semantic differential items to indicate whether they thought the supervisor acted fairly. First participants read the lead-in statement, “In communicating the information to me, the supervisor was . . . ” and then same list of semantic differentials as were listed for procedural justice. CFA was performed on the scale and all five items formed a unidimensional solution with a reliability (Cronbach’s α) of .92. See Appendix A for all justice items.
Manipulation Checks
Scenario realism
Scenario realism was measured with a set of 5 Likert-type items on a scale of 1 (strongly disagree) to 5 (strongly agree). Sample items include, “This situation could happen, or has happened” and “I can imagine being in a situation like this one.” CFA was performed on this scale, resulting in a solution with 4 items. The reliability (Cronbach’s α) of this scale was .71. Overall, the mean level of scenario realism was 3.56 (SD = 0.76). This is significantly above the midpoint based on a one-sample t test, t(213) = 10.90, p < .001, suggesting that this scenario was seen as realistic. There also were no significant differences between the students and the working adults, F(1, 211) = .01, p = .94.
Message expectation
Expectation of actually receiving this type of message was measured with a set of 4 Likert-type items on a scale of 1 (strongly disagree) to 5 (strongly agree). Sample items include, “I consider this type of message normal” and “I would never expect to hear this kind of message.” CFA was performed on the scale, resulting in a unidimensional solution with 4 items. The reliability (Cronbach’s α) of this scale was .80. Overall, the mean level of message realism was 3.32 (SD = 0.83). This is significantly above the midpoint based on a one-sample t test, t(212) = 5.66, p < .001, suggesting that people could expect to receive messages like the ones presented in the current study. There also were no significant differences between the students and the working adults, F(1, 211) = 1.35, p = .25.
Valence
Perceived valence of the message was measured with a 7-point semantic differential scale preceded by the statement, “The information contained in this message was . . . ” and followed by pairs of words such as, “Positive . . . Negative,” “Happy . . . Unhappy,” and “Favorable . . . Unfavorable.” CFA was performed on this 4-item scale, showing a unidimensional solution with a reliability (Cronbach’s α) of .94. To test whether or not the message valence manipulation worked as expected, a 2 (valence) × 2 (channel) analysis of variance (ANOVA) was performed on perceived valence. A main effect for valence of message was found for perceived valence, F(3, 208) = 160.59, p < .001, η2 = .43, such that acceptance messages were seen as more positive (M = 5.41, SD = 1.23) than rejection messages (M = 3.13, SD = 1.39). Also, acceptance messages were above the midpoint, suggesting they were seen as positive, and rejection messages were below the midpoint, suggesting they were seen as negative. There were no significant differences for channel on perceived valence, F(3, 208) = 1.74, p = .19, η2 = .00, nor for the interaction of message valence and channel on perceived valence, F(3, 208) = .91, p = .34, η2 = .00. Thus, the message manipulation worked as expected.
Privacy
The degree of privacy of the message was measured with a 7-point semantic differential scale preceded by the statement, “The information contained in this message was . . . ” and followed by pairs of terms such as “Private . . . Public,” “Exclusive . . . Unexclusive,” and “Confidential . . . Unconfidential.” There were 7 pairs of words, which were subjected to a CFA, which resulted in a unidimensional solution. The reliability (Cronbach’s α) of this 7-item scale was .94. To test whether or not the channel manipulation induced perceptions of privacy as expected, a 2 (valence) × 2 (channel) ANOVA was performed on perceived privacy. A main effect for channel was found for perceived privacy, F(3, 207) = 46.86, p < .001, η2 = .18, such that a company website was seen as more public (M = 5.02, SD = 1.43) than email (M = 3.69, SD = 1.36). There were no significant differences for message type on perceived privacy, F(3, 207) = .00, p = .99, η2 = .00, nor for the interaction of message valence and channel on perceived valence, F(3, 208) = .05, p = .83, η2 = .00. Thus, the channel manipulation worked as expected. See Appendix B for all manipulation check items.
Results
To test the hypotheses and research question, a 2 (valence) × 2 (channel) ANOVA was conducted on both procedural justice and interactional justice. Main effects for valence and channel were examined to test the four hypotheses. Interactions between valence and channel were examined to answer the Research Question 1. Post hoc power analysis was conducted for nonsignificant findings using the program G*Power 3 (Faul, Erdfelder, Lang, & Buchner, 2007). Using the conventional power level of .8, the current study was sufficiently powered to find an effect size (f2) of .037, which is considered a relatively small effect size (Cohen, 1988). Thus, the current study seems well-powered.
To test Hypotheses 1 and 3, a 2 × 2 ANOVA was performed on perceptions of procedural justice. First, a statistically significant effect was found for valence of the message, F(1, 213) = 17.48, p < .001, η2 = .07, such that positively valenced acceptance messages (M = 5.13, SD = 1.22) were seen as more just than the negatively valenced rejections (M = 4.38, SD = 1.33). Thus, the data were consistent with Hypothesis 1. Second, a statistically significant effect was found for channel of the message, F(1, 213) = 6.82, p = .01, η2= .03, such that the private channel (M = 5.02, SD = 1.30) was seen as more just than the public channel (M = 4.50, SD = 1.31). Thus, the data were consistent with Hypothesis 3.
To test Hypotheses 2 and 4, a 2 × 2 ANOVA was performed on perceptions of interactional justice. First, a statistically significant effect was found for valence of the message, F(1, 213) = 14.44, p < .001, η2 = .06, such that positively valenced acceptance messages (M = 5.22, SD = 1.23) were seen as more interactionally just than negatively valenced rejections (M = 4.50, SD = 1.41). Thus, the data were consistent with Hypothesis 2. Second, a statistically significant effect was found for channel of the message, F(1, 213) = 7.59, p = .006, η2 = .03, such that the private channel (M = 5.14, SD = 1.34) was seen as more interactionally just than the public channel (M = 4.59, SD = 1.34). Thus, the data were consistent with Hypothesis 4.
Research Question 1 asked about the possible interaction between valence and privacy of the channel on perceptions of justice. For procedural justice, no statistically significant interaction was found, F(1, 213) = .002, p = .97, η2 = .00. For interactional justice, no statistically significant interaction was found, F(1, 213) = 1.57, p = .21, η2 = .006.
Discussion
The current study was designed to answer questions regarding the impact of message valence and privacy/publicity of channel on perceptions of justice in an organizational setting. The findings reveal additional ways organizations may be able to engender perceptions of fairness in their employees: through feedback valence and by paying attention to the privacy of the channel used to deliver feedback. Effects were found for both valence and channel. Consistent with predictions, positively valenced messages were seen as more just, as were messages sent using a private channel (rather than a public one). The study found no interaction effect between valence and channel on perceptions of justice. Each of these findings will be discussed in greater detail, as will limitations and directions for future research.
First, the channel characteristic of private/public had an impact on perceptions of both procedural and interactional justice, such that messages sent through a more private channel were perceived as more just. This sheds new light on the findings of Westerman and Westerman (2010), who found no differences for channel on perceptions of justice. The current study hypothesized that differences would occur because of a general characteristic difference between the channels chosen, unlike Westerman and Westerman (2010) and others (e.g., Timmerman & Harrison, 2005) who have focused on channel richness rather than the way a channel is used (publicly vs. privately). The fact that differences were found in the current study using the general difference of private/public rather than simply comparing email with FtF delivery suggests two things. First, consistent with Walther and Ramirez (2010), the field of computer-mediated communication is likely better served by consideration and theorizing about general channel characteristics rather than simply comparing specific channels. Second, the specific difference of public versus private channel is one that future studies should continue to examine because it seems likely to make an impact on communication outcomes. It is possible that differences were not found in the Westerman and Westerman (2010) study because both email and FtF were seen as private channels. This possibility, combined with the results of the current study, provides evidence for the notion that channel characteristics are more fruitful to examine than simple channel to channel comparisons, and also that public/private is a useful specific characteristic to examine in future research.
The valence of a message was found to affect justice perceptions such that positively valenced messages were seen as more procedurally and interactionally just than negatively valenced messages. Other research on organizational feedback and justice focuses on the delivery of negative feedback in different ways, often to the exclusion of any positive feedback (e.g., Westerman & Westerman, 2010; Chory & Kingsley Westerman, 2009; Steelman & Rutkowski, 2004). For example, Steelman and Rutkowski (2004) found that there was a stronger motivation to use feedback when it was delivered in a considerate manner than when it was less considerate, but both conditions used negative feedback. The current findings add to the body of empirical evidence by examining responses to both positive and negative feedback and showing that positive information leads to more positive responses in terms of perceptions of both procedures and the deliverer of the information.
Interestingly, there were no interaction effects among valence and the level of privacy of the delivery. This seems to indicate that the selection of a public or private channel operates separately from the valence of feedback information delivered. Although we had no theoretical reason to predict this, it would seem intuitive that, for negative feedback, people might not want public delivery but for positive feedback they might feel the opposite. In fact, perhaps the delivery of positive feedback publicly would be welcomed and appreciated. It could be that because the study focused on perceptions of justice that it did not capture any sense of honor or pride associated with positively valenced feedback delivered publicly. Further research is needed to determine if this effect will continue to hold with different dependent variables.
Also, these findings may indicate “boundary turbulence” (Petronio, 2002)—when the rules for managing co-owned information are not in synch (i.e., the sender and receiver are unable to coordinate their boundary management). That is, there may be an expectation from a worker that a supervisor would deliver the acceptance or rejection news in a certain way to maintain a co-owned boundary about a work proposal, similar to how a teacher would maintain a co-owned boundary of grade information shared with a student. Perhaps the more public delivery of the information violates an expectation of privacy, whether the information is positive or negative, similar to that of the expectation of grade privacy, and thus creates a lack of synchronization about the sharing of the information. Future research can examine this possible boundary turbulence as it relates to channel usage and message valence.
Practical Implications
One major implication of this study is that channel characteristics are important when delivering feedback to employees. Supervisors cannot assume that richer channels are always better; rather, characteristics that transcend a particular channel are important to consider when determining how to deliver a message. Along with the privacy of the channel, these may include the synchronicity of a channel, and the permanence of messages sent through a channel. This study provides some evidence that the level of privacy provided by a channel does have an impact when a supervisor is delivering feedback to employees. More specifically, based on these findings, supervisors should deliver feedback privately regardless of whether it is positive or negative. However, there may be mitigating factors that this study did not reveal, and future research must investigate, as there may be circumstances under which supervisors want to publicly deliver feedback, be it positive or negative.
Positive feedback leads to perceptions that the procedures used by a supervisor are just and that the supervisor is just. The authors know of some individuals who profess that they categorically “do not give positive feedback.” It may be wise for those who work in a supervisory capacity to consider whether changing their communication patterns with respect to feedback (i.e., giving positive in addition to negative feedback) may be worth the effort to increase their employees’ perceptions of justice.
Although this study did not explicitly test the violation of co-owned boundaries for feedback, it may be worthwhile for supervisors to talk with their employees to get an idea of the expectations employees have for how performance feedback should be shared. This may differ across organizations, or even for different employees in the same organization. Learning more about employees’ expectations and sharing their own expectations can help supervisors more successfully manage information in a way that is perceived as just by their employees.
The perception of justice has been shown to be a worthwhile pursuit for organizations as a whole, as it is associated with a host of positive organizational variables, such as perceptions of career stability (Yamaguchi, 2005) and outcome satisfaction, job satisfaction, trust, and organizational commitment (Colquitt et al., 2001). In addition, perceptions of injustice are associated with negative organizational outcomes (see Chory & Kingsley Westerman, 2009). This study reveals two important methods organizations may use to engender perceptions of both procedural and interactional justice in their employees. The use of positive feedback and private channels are simple elements that organizations can implement without much difficulty.
Limitations/Future Directions
One potential limitation of this study is the balance between external validity and experimental control. In this case, we chose to use an experimental message at the expense of studying existing relationships in organizations. Although this allows for a clearer determination of the source of any significant differences, it simultaneously excludes contextual variables such as organizational culture or power distance. As Sussman and Sproull (1999) suggest, it is possible that workers in organizations could receive bad news from individuals in their organization with whom they have no relationship (e.g., they cite change initiatives, reorganization, cross-division information delivery). Similarly, feedback such as that used in this study (acceptance/rejection of a proposal) could be delivered by someone much higher up in the organization rather than one’s direct supervisor, or it could be delivered by a new boss who has just been placed in the position. This is not to discount the importance of studying existing, ongoing relationships in organizations but to point out that the type of information delivery simulated in this study is not completely divorced from reality, as further evidenced by participants’ responses to perceived realism measures.
Also, there was not much difference between the response to procedural and interactional justice—that is, the same pattern of results held across both types of justice. Future studies should attempt to tease out whether either type of justice would operate without the other. Similar to previous findings that distributive justice does not have much of an effect when either procedural or interactional justice is present, perhaps procedural justice does not have much of an effect when interactional justice is present, or vice versa. Of particular interest to supervisors would be whether interactional justice in the absence of procedural and distributive justice would be enough to garner positive responses from employees. Because the interaction is the most controllable factor for supervisors, learning more about if and how procedural and interactional justice may work separately is worthwhile.
Future research may be designed to examine communication responses to messages delivered through public and private channels, such as voice and silence of the receiver of both positive and negative feedback. Van Dyne, Ang, and Botero (2003) define voice as expressing ideas and silence as intentionally withholding ideas. Whether individuals respond to positive and negative feedback with silence or voice may have important organizational impacts. Another interesting question would be whether the receiver accommodates the sender by using the same channel for the response or if the receiver chooses a different channel. For example, if the message comes through a public channel and the receiver believes that is unfair or inappropriate somehow, does the receiver send a response through the public channel or choose a more private channel to respond?
The channel characteristic method suggested by O’Sullivan (2005) and Walther and Ramirez (2010) and used in this study shows promise. Future studies would do well to continue investigating channels based on their characteristics, including the public/private dimension explored in the current study. Other relevant characteristics such as message personalization (O’Sullivan, 2005) and disentrainment (Walther, 1996) have also received attention and should continue to be focused on. This method of study will help improve the relevance of current studies in the future if and when current channels of choice (e.g., Facebook, Twitter) become obsolete. Channels may change, and people may use channels in different ways (e.g., Facebook can be used both publicly and privately), but learning about relevant characteristics that affect communication behaviors and outcomes allows for continued study relevance even after newer channels are adopted, as long as they retain those relevant characteristics.
Conclusion
Perceptions of justice can be instrumental to organizations, particularly because of the negative responses to perceived unfairness in organizations. If supervisors can know the correct channel characteristics to choose for a specific message, organizations may be more successful in delivering feedback to their employees. As delivery of messages is something over which managers most likely will have control, choosing a private channel for the delivery of feedback information would seem to be a wise choice. Finally, knowing the likely response to positively and negatively valenced messages can also prove useful in terms of preparation for delivering a given message.
Footnotes
Appendix A
| Procedural Justice. Semantic differential items with opening phrase |
| The procedure used to notify me about my project was . . . |
| Unfair-Fair |
| Unjust-Just |
| Biased-Neutral |
| Unwarranted-Warranted |
| Unjustified-Justified |
| Interactional Justice. Semantic differential items with opening phrase |
| In communicating the feedback to me, the superior was . . . |
| Unfair-Fair |
| Unjust-Just |
| Biased-Neutral |
| Unwarranted-Warranted |
| Unjustified-Justified |
Appendix B
Manipulation Check Scales
| Scenario Realism. Likert-type items |
| I didn’t have any problem with the realism of this situation. |
| It was difficult to make myself feel that this situation was real. a |
| This situation could happen, or has happened. |
| A situation like this could develop in real life. |
| I can imagine being in a situation like this one. |
| Message Expectation. Likert-type items |
| I consider this type of feedback normal. |
| I would be shocked to receive this feedback. a |
| I would never expect to hear this kind of message. a |
| I don’t know what it would be like to hear this kind of message. a |
| Valence Scale. Semantic differential items |
| Positive-Negative |
| Happy-Unhappy |
| Pleasant-Unpleasant |
| Favorable-Unfavorable |
| Disconfirming-Confirming |
| Privacy Scale. Semantic differential items |
| Public-Private |
| Closed-Open |
| Available-Unavailable |
| Exclusive-Unexclusive |
| Restricted-Unrestricted |
| Shared-Individualized |
| Confidential-Unconfidential |
| Secret-Well-known |
Recoded item.
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.