Abstract

Computer-mediated communication (CMC) makes it possible to deliver evaluative feedback, an essential component of learning, over distance. This article presents a theoretical model of the CMC-based evaluative feedback process from the perspective of both senders and receivers of feedback. Hypotheses suggested by the model were tested in a quasi-laboratory experiment with part-time MBA students comparing email with voicemail. Within matched-pair dyads, email senders, but not voicemail senders, viewed their negative feedback as being significantly more negative than did their receivers. Voicemail senders, but not email senders, reported significantly lower comfort levels than did their receivers. No feedback effectiveness differences were found between media conditions, although determinants of feedback effectiveness differed significantly depending on the medium. These results are generally consistent with the theoretical model.

Introduction

Evaluative feedback is a fundamental component of human learning (Locke, Cartledge, & Koeppel, 1968). Without evaluative feedback, individuals and organizations cannot evolve in ways that meet the standards of others. For example, the number of students taking courses in online degree programs was almost one million in 2004 (Carnevale, 2005), and these students are all being evaluated at a distance. In organizations, virtual teams are evaluated by distant managers. Software developers in the United States evaluate the output of their outsourced computer programmers in India. As education and organizations become increasingly global, evaluation processes often must take place over distance. And while written evaluations have long been a part of most performance appraisal processes, new media such as email and voicemail are increasingly the means by which these distal evaluation processes are accomplished. Research lags practice in this area, and more research attention needs to be directed to the context of mediated performance evaluation (Illgen, Barnes-Farrell, & McKellin, 1993; Levy & Williams, 2004).

Computer-mediated learning can produce equivalent or better learning outcomes than face-to-face learning (Alavi, 1994; Benbunan-Fich & Hiltz, 1999). Yet there is a paucity of research on the micro-mechanisms underlying how use of new media affects the evaluative feedback process. This study aims to begin to fill this gap from two perspectives. First, I investigate differences in perceptions between senders and receivers of mediated evaluative feedback, to see if media have a differential affect on one or the other. Second, I look at receivers' experience of the mediated feedback process and their perceptions of the content delivered, to investigate ways that media might affect this process.

The study begins by developing a theoretical model of the evaluative feedback process when it transpires asynchronously by means of media that vary in their capacity to transmit social presence information. The model presents potential differences in perceptions across paired sender-receiver dyads. The model is empirically evaluated using a quasi-experiment in which participants were 65 part-time students in an Information Systems course in an MBA program in the United States. Students delivered and received both positive and negative feedback about final website projects to a classmate via either voicemail or email. The students then reported their perceptions on questionnaires immediately after delivering or receiving feedback. Analysis of the results indicates significant differences by medium for both within-dyad differences, and for receiver feedback processing.

The theoretical discussion below develops the study's hypotheses. This section is followed by a description of the quasi-experiment used to investigate the hypotheses. Next, the results are presented of the analyses of the data. In concluding, the implications of the results are discussed for theory and for those who deliver or receive evaluative feedback via new media.

Theoretical Background

Feedback refers to a response to a person's behavior. It influences the manner in which that behavior will be continued or not. Deliberate feedback is used as a change strategy in educational, occupational, therapeutic, and other settings. It can be primarily descriptive or evaluative (Jacobs, 1974). It is evaluative when it offers an assessment of behavior in relation to a performance criterion (Kluger & DeNisi, 1996). Evaluative feedback has been shown to improve the performance of receivers of the feedback (for a review of this literature, see Illgen, Fisher, & Taylor, 1979; Nadler, 1979). The feedback-performance link is typically characterized as a cybernetic system in which feedback serves as a measure indicating distance from desired goals (Locke, et al., 1968). These measures enable the receivers of feedback to make accurate decisions regarding their future performance. In this way, feedback satisfies individuals' information needs regarding the extent to which their goals are being met (Nadler, 1979). Evaluative feedback promotes change through interpersonal influence (Claiborn, 1986) and falls on a scale from negative to positive, referred to in the literature as valence (note that this is distinct from negative and positive valence as applied to an emotional response). Feedback is negative when it indicates that the target behavior has fallen short of the criterion, and it is positive when it identifies that the target behavior has met or exceeded the criterion.

From the perspective of communication theory, evaluative feedback is one type of asymmetrical communication task. People in organizations face asymmetrical communication tasks whenever they possess information that is of significance to others but not possessed by them, or vice versa. In addition, the performance appraisal process is generally characterized by power asymmetry. That is, most performance appraisal researchers view the process in terms of leader-member exchange, characterized by vertical dyad linkages between supervisors and supervisees, with consequent power inequities (for a review of this research, see Gerstner & Day, 1997). At the same time, peer appraisals are becoming increasing common in both practice and research (Druskat & Wolff, 1999). The focus of this research is on the effects of computer-mediated communication (CMC) on the evaluative feedback process. For this reason, the complicating factor of power asymmetry is excluded from the model, in order to understand media effects without additional confounds. Thus the context to which the findings of this study may be applied is limited to peer appraisals and not vertical dyads.

Much of the literature on feedback processes in organizations focuses on the supervisor-subordinate relationship (Gerstner & Day, 1997), and these studies almost always utilize the dyadic level of analysis (Mawhinney, 2005). Feedback also comes from peers or subordinates, in this case conceived of theoretically as an interpersonal influence process that transpires between individuals, which is why it is appropriately investigated at the dyadic level of analysis. Dyadic analysis is generally superior to individual and system-level analyses of organizational networks (Mizruchi & Marquis, 2006).

Evaluative feedback is increasingly being delivered via means other than traditional face-to-face contexts. An entire industry of online performance management software has sprung up to provide support to organizations that want to use the Web for their performance appraisal and evaluation processes. Online education support tools all come with features to provide evaluative feedback remotely. These tools have emerged in response to demand for such products from educational and for-profit organizations. This demand is evidence of just how widespread the phenomenon of delivering CMC-based evaluative feedback is. Yet research lags behind this phenomenon. While studies of the process of virtual learning are increasingly common (e.g., Heckman & Annabi, 2005; Schweizer, Paechter, & Weidenmann, 2000), there has been no systematic empirical investigation of the influence of media use on evaluative feedback delivery. This study seeks to fill this gap.

CMC, Voicemail, and Asymmetrical Communication

Much of the research on CMC has focused on media choice as the dependent variable, investigating how people make selections among available media. Various explanations of media choice have been put forth, broadly categorized as objective factors, social factors, and person/technology factors (Trevino, Webster, & Stein, 2000). Objective factors include receiver geographic distance and number of receivers (both are well supported via email) and how equivocal the particular message to be delivered is. Social factors include perceptions of how a medium is viewed by coworkers and supervisors and perceptions of the symbolism of its use. Of the various contextual factors affecting media perceptions, the time over which a mediated relationship transpires can enable hyperpersonal interactions despite the use of "lean" media (Walther, 1996).

Media richness theory has also been widely applied to the problem of media choice, falling into the category of person/technology factors. According to media richness theory, there is an optimal fit between the richness of a medium and the particular task at hand, and selection of the appropriate medium for the task will result in higher performance than a less optimal task-media fit (Daft & Lengel, 1986; Daft, Lengel, & Tevino, 1987). While clearly appropriate for understanding the selection of media for feedback delivery, none of these paradigms addresses the phenomenon of whether and how media affects the feedback delivery process. To the extent that geographic distance makes use of CMC for evaluative feedback delivery necessary, media choice is constrained. Nor does the media choice paradigm help us understand effects of the media once the selection has been made.

Social presence theory (Short, Williams, & Christie, 1976) addresses this issue by positing that media differ in the extent to which they make interpersonal cues salient. For example, email transmits fewer cues about the presence of the communicative partner than does video. Where these cues are attenuated, the social presence of the receiver—the feeling that the other person is involved in the communication exchange—is less salient (Rice, 1993; Short, et al., 1976). For example, the lack of social status cues appears to enable greater participation by peripheral members of online communities, and lack of identification cues enables the anonymity that seems to spur increased self-absorption and flaming in mediated communication (Kiesler & Sproull, 1992).

Most of the previous experimental research on CMC has used tasks characterized by impartial information rather than information with positive or negative consequences for the receiver. For the most part, these tasks have also been characterized by information symmetry; that is, all participants have equal information resources to contribute to the interaction and resulting performance. Such tasks include getting to know someone (Kiesler, Zubrow, Moses, & Geller, 1985), brainstorming (e.g., Connolly, Jessup, & Valacich, 1990; Dennis & Valacich, 1993; Gallupe, Dennis, Cooper, Valacich, Nunamaker, & Bastianutti, 1992; Valacich, Dennis, & Connolly, 1994), planning (Applegate, Konsynski, & Nunamaker, 1986), and decision making (e.g., Hiltz, Johnson, & Turoff, 1986; Kiesler & Sproull, 1992; Weisband, Schneider, & Connolly, 1995).

Evaluative feedback delivery, by contrast, is an asymmetric information task in which one person is the primary source and the other is the primary receiver of the information. With the exception of Sussman and Sproull (1999), asymmetric communication has not been the focus of CMC research. Considering the extent to which the Web is used for asymmetric communication tasks such as feedback delivery, however, asymmetric communication may be a fruitful direction for CMC researchers to take, particularly in light of the current emphasis on information asymmetries created by the Web.

This study focuses on two common communication technologies: email and voicemail. Email is widely used in organizations today. It is asynchronous, fast, and multipurpose, used for message creation, editing, storage, retrieval, and forwarding. It is used to send information to groups of people as well as individuals (Culnan & Markus, 1987) and can be used to communicate equivocal, complex information (Markus, 1994). Because it is text-based, it decreases social context cues as compared to face-to-face communication, resulting in self-absorption, status equalization, and communication disinhibition (Sproull & Kiesler, 1986). Organizational use of email has been studied for its effects on social context cues (Sproull & Kiesler, 1986), knowledge sharing (Sussman & Siegal, 2003), and interactions with media richness (Fulk, Schmitz, & Steinfeld, 1990; Markus, 1994; Rice, 1992).

Voicemail has also altered the landscape of options for communicating with people who are geographically distal or frequently unavailable (Straub & Karahanna, 1998). Research on the use of voicemail in organizations has taken one of two perspectives. The first of these focuses on determinants of its use, particularly why people choose to use it relative to other media (El Shinnawy & Markus, 1997; Rice & Tyler, 1995; Straub & Karahanna, 1998). The second perspective looks at perceptions of appropriateness and other characteristics of voicemail that affect acceptance of the technology within organizations (Caldwell, Uang, & Taha, 1995; Duthler, 2006; Epley & Kruger, 2005; Subramanian & Rohrer-Meek, 1998). Researchers are beginning to look at the effects of voicemail in particular contexts, such as language teaching (Volle, 2005). This study takes this third perspective, examining effects of the actual usage of voicemail in one particular context—the asymmetrical task of feedback evaluation.

Email and voicemail were selected for this study because both are widely used in most organizations and institutions and because both are asynchronous. It is important to compare technologies that are alike with respect to synchronicity, since doing otherwise would introduce the potential confound of differences due to this factor. The following sections present hypotheses based on social presence theory suggesting ways that use of email or voicemail might affect the evaluative feedback process.

Sender-Receiver Pair Differences

When evaluative feedback is negative, its delivery can be an emotionally-laden and discomforting task. People are reluctant to communicate undesirable information—a widely documented phenomenon first labeled the "Mum" effect by Rosen and Tesser (1970). Differences in transmission of bad news as opposed to good news have been demonstrated across a wide variety of cultures, settings, and relationships (O'Neal, Levine, & Frank, 1979; Tesser & Rosen, 1975). Example domains include a social work agency (Tesser, Rosen, & Tesser, 1971), organizational hierarchies (Fulk & Mani, 1986; Lee, 1993), psychiatry and psychotherapy practice (Kivlighan, 1985; Rice & Warner, 1994), personnel hiring (Rosen, Grandison, & Stewart, 1974), doctor-patient relationships (Seale, 1991; Waitzkin, 1984), test failure (Bond & Anderson, 1987), and the performance feedback context (Larson, 1986). The Mum effect occurs when the sender anticipates that the receiver will react with defensiveness and emotional distress. If the sender anticipates that the receiver will be distressed, he or she may anticipate having to deal with the receiver's negative emotional state.

Previous research indicates that feedback senders may be more comfortable using media that reduce social context cues, such as email, to deliver negative evaluative information than face-to-face or voice delivery (Sussman & Sproull, 1999). During feedback delivery, the use of email tends to increase the negative content of the feedback (Kurtzberg, Naquin, & Belkin, 2005). This is because "lean" media that are limited in their capacity to transmit social cues can buffer the receiver from cues about the affect of the sender and vice versa, essentially "filtering out" those cues. In this way CMC can decrease the sender's psychological discomfort throughout the process of delivering negative evaluative feedback.

Senders who use voicemail to deliver negative evaluative feedback are not so fortunate, however. They may expect that their voicemail messages will be scrutinized for social presence cues such as tone of voice and nervousness, as receivers of feedback seek additional information about the degree and impact of this information. This potential for scrutiny may create discomfort on the part of the sender. Since these kinds of cues are not likely to be revealed by a feedback sender using email, an email sender should be more comfortable with the delivery process. This effect may be augmented by the fact that email senders can compose and edit their messages in advance, while the voicemail sender cannot. For both these reasons, then, deliverers are likely to experience a differential level of comfort depending on the medium they are using to deliver their feedback. However, there is no theoretical or practical reason to believe that feedback receivers' comfort levels will vary according to medium. Therefore, I hypothesize asymmetrical experiences of comfort within deliverer-receiver dyads using voicemail. Voicemail senders will experience lower comfort levels than their communication partners, but this comfort differential will not be found in email dyads, since email senders will not experience heightened discomfort.

How negative feedback is perceived is also likely to differ within dyads depending on the social presence of the media. When people report on their own feelings and behavior, they tend to be more negative when they report electronically than when they report face-to-face (Kiesler, et al., 1985). Email has spawned the phenomenon of flaming—sending email messages intended to inflame or arouse the receiver's negative emotion. People tend to be more honest in delivering bad news via CMC than face-to-face or using the telephone (Sussman & Sproull, 1999). That is, they attend less to the cultural norms and intrapsychic processes that motivate the sugarcoating of negative feedback. This is because media that do not support the transmission of social presence cues enable people using them to ignore the social presence of those with whom they are communicating. Thus I expect that people delivering negative feedback via CMC will produce more negative feedback than those delivering feedback via voice, and will also perceive their feedback to be more negative.

In contrast, there is no reason to suspect that these forces would affect people using the "richer" medium of voicemail to deliver feedback. Relative to email, voicemail supports transmission of social presence cues that remind the communicator of social norms and processes. Thus voicemail senders are more likely to sugarcoat their negative feedback and, accordingly, perceive it to be less negative than are email senders.

Further, none of the theoretical reasons for differences in negativity perceptions by medium apply to the experience of the receiver. Email senders are likely to produce feedback that they anticipate will be viewed as highly negative, but receivers of this feedback are no more likely to see it as highly negative than are voicemail receivers. For this reason, perceptions of negativity are likely to be asymmetrical within email dyads but not within voicemail dyads.

Whereas negative feedback motivates sugarcoating when it is delivered via CMC, positive feedback does not. To the extent that perceptions of feedback align with production of it, the within-pair asymmetry of negativity perceptions discussed for H2a is not expected to be apparent for perceptions of content positivity. Thus:

The Receiver Experience

Thus far I have relied on a model of the sender's experience to characterize the nature of the dyadic experiences of sending and receiving mediated evaluative feedback. I now turn to the experience of the feedback receiver in order to understand media effects on feedback effectiveness. Ultimately, I am interested in how effective the feedback is for conveying to the receiver ways to improve, such that the receiver is motivated to do so.

The dual process theories of informational influence (Chaiken & Eagly, 1983; Petty & Cacioppo, 1986) are useful to investigate how feedback receivers process this content. This widely accepted body of theory for understanding how validity-seeking information receivers respond to the received information is increasingly being applied to the study of CMC in organizations (Robert & Dennis, 2005; Sussman & Siegal, 2003). The elaboration likelihood model (ELM) of persuasion (Petty & Cacioppo, 1986) is one of several theories of dual-process cognitive processing. According to the ELM, different message receivers will vary in the extent to which they elaborate cognitively on a particular message, and these variations in elaboration likelihood, along with other factors, affect the results of information transfer. Elaboration involves attending to the content of the message, scrutinizing and assessing its content, and reflecting on issues relevant to the message. Because of the cognitive effort involved, receivers do not elaborate on every message they receive, and some receivers elaborate on fewer messages than others do.

Informational influence can occur at any degree of receiver elaboration, but as the result of very different processes: High levels of elaboration represent a central cognitive processing route, while low levels represent a peripheral route. The central route occurs when receivers carefully consider (e.g., elaborate on) the issues presented by the message, whereas the peripheral route occurs when receivers use simple decision rules to evaluate the message rather than analyzing its content. The central and peripheral routes are viewed as the extremes of a single underlying elaboration dimension.

Peripheral cues are informational indicators that people use to assess content validity other than the arguments consisting of the content itself (Petty & Cacioppo, 1986). Peripheral route processing applies these informational cues during assessment of received information (Chaiken & Eagly, 1983). A potentially infinite number of these heuristics exist in interpersonal communication contexts (Gergen, 1982). For instance, in groups people are persuaded by consensus cues and attributes of the group leader such as charisma. Early research on the use of peripheral cues in CMC contexts found that the social presence of the communicator varies by media and affects persuasive communication by attenuating potential peripheral cues (Short, Williams, & Christie, 1976). Research on rich versus lean media (Andreoli & Worschel, 1978; Worschel, Andreoli, & Eason, 1975) and vividness effects (Pallak, 1983) supports the notion that lack of peripheral cues and heuristics for informing attitude judgments serve to force message receivers to elaborate on the nature of the arguments presented, especially when the arguments are complex (Chaiken & Eagly, 1983). Based on this logic, for peripheral cues that get filtered out by "lean" media, we are more likely to see effects of these cues on feedback effectiveness for participants in the richer media of voicemail condition than for those in the leaner email condition.

Out of the many cues that people use to inform their decision rules, previous research has identified cues pertaining to the message's source as important when people are unable or unwilling to expend the effort to elaborate on the message's content (Petty & Cacioppo, 1986). This body of research investigated source cues having to do with the appearance of the source, such as attractiveness and likeability, whereas appearance is attenuated in both the media investigated in this study. However, we know that email receivers do form impressions about the message author (Walther, 1992). Clearly some peripheral cues are operating in such interaction.

In voicemail we are aware of vocal tonalities and inflections that are not available via CMC. These tonalities may convey the valence of the information delivered—the negative or positive tone of the content—as a peripheral cue that may be influential in the context of outcome-based evaluative feedback. Valence is clearly distinct from the content of the message, yet is an attribute of asymmetrical messages that is known to influence the feedback effectiveness process (London, Larsen, & Thisted, 1999; Sorensen & Franks, 1972). If information valence does in fact serve as a peripheral cue in this context, one would expect it to be more salient in the voicemail media condition than in the email one, since voicemail has the capacity to carry tonal cues unavailable with text-based media such as email. And since peripheral cues affect attitude change in message receivers, this medium difference can be expected to influence the effectiveness of the feedback for the receiver. Hypothesis 3 below summarizes this line of thinking:

According to the ELM, we undertake the effort of central route processing when we are motivated to do so, or else when peripheral cues are not available. In this study I expect to find a greater reliance on peripheral cues in the medium condition least likely to attenuate them—voicemail. If dual information processing models are operating in this context, we should see a corresponding increase in central route processing by email feedback receivers. Relative to voicemail, CMC receivers should undertake more central route processing, since they do not have peripheral cues to inform their decision rules.

This raises the thorny issue of how to assess central route processing. ELM identifies central route processing as that which takes place when argument quality is attended to, but has "postponed the question of what specific qualities make arguments persuasive" (Petty & Cacioppo, 1986, p. 32). One approach to this problem has come from social cognition research. Kiesler and Sproull (1982) identify the relevance of a stimulus as the key factor determining whether managers ignore the stimulus or go on to interpret it during problem sensing. Information that is diagnostic and likely to produce useful task resolutions receives precedence and weight in judgment and choice processes (Feldman & Lynch, 1988). From this perspective, content that is perceived as relevant and useful is most likely to be fodder for central route processing. I call this construct Content Quality. As argued above, we can expect that the peripheral cue of feedback valence will be available to those receivers in the voicemail condition but to a large extent unavailable to those in the email condition. Therefore, we should find higher levels of central route processing in the email condition, relative to the amount of peripheral processing done by these CMC receivers.

Figure 1 describes the relationships hypothesized above in terms of the experience of the feedback receiver.

Method

The hypotheses were investigated in a quasi-experiment in which participants were randomly assigned to deliver feedback that they generated themselves, via either voicemail or email, to another participant. The level of analysis was dyadic, as is appropriate to the study of feedback processes, operationalized in this case as peer influence processes. Participants completed questionnaire items after both delivering feedback to another participant and receiving feedback from another participant. In this way, each participated in two dyadic interactions, one in which he or she delivered feedback, and the other in which he or she received it.

Study Participants

Participants were evening MBA students attending an introductory course in information systems at a large Midwestern U.S. university. Two out of the 97 participants were full-time students, and the rest were employed at the time of data collection. Participation was voluntary and required explicit permission. In order to achieve a reasonable sample size, it was necessary to collect data from three classes, one each during the spring, summer, and fall of 1999. No significant differences by cohort were found in either their demographics or in the model constructs. In total, 65 participants both delivered and received feedback; none reported their receiving experience without reporting on their delivery experience. Table 1 presents these demographic data by cohort. Out of the 65 participants, 47 matched sender-receiver dyads occurred. This was due to the fact that participation was entirely voluntary—participants were not receiving compensation or course credit—and the sample population was night students who mostly worked part-time and attended class in the evening. In order to generate matched-pair dyads, participants needed to both deliver and receive feedback, and only where both participants happened to do so were dyadic responses usable. For this reason, there was attrition from 67 initial participants to the 47 who completed both parts of the activity.

Task and Procedure

The task entailed composing and delivering both negative and positive feedback about a group project in which the feedback receiver had participated as a member. The group project was the production of a website about a technological topic, developed over the course of the semester. All students in each of the three classes were randomly assigned to a feedback condition, but reporting on the delivery and receipt of their feedback was optional. For this reason, the number of participants in each media condition is not equal, since only a percentage of each class chose to participate. Moreover, since the study of matched pairs required elimination of those participants whose dyadic other did not participate, the matched pairs are not distributed evenly by media condition.

Two versions of the questionnaire were designed—one for feedback senders, and one for feedback receivers. See the Appendix for the questionnaire items. This exercise took place at the end of the semester, so senders and receivers were familiar with one another. For obvious reasons, no participants were assigned to deliver feedback to their own group members.

All participants provided their email address and voicemail number and permission to release either of these to one other class member. Within each class cohort, each student was given the name of another student in his or her class to deliver feedback to, the media type that had been randomly assigned, and the email address or voicemail number through which to deliver the feedback. They were then given one week to review the website that their feedback receiver had helped to build as his or her group project. They were asked to write down the three most positive aspects of the website project, and the three most negative aspects of it. These items constituted the evaluative feedback to be delivered, consisting of both positive and negative valence information. All students were required to deliver feedback; however, only study participants completed the questionnaire used for this research.

Students delivered this content to their assigned classmate using the medium they had been randomly assigned. Those choosing to participate in the study then completed the sender's questionnaire immediately after they performed the feedback delivery task. If they received evaluative feedback about their own group's website and chose to participate, they then also completed the receiver's questionnaire about their experience of receiving positive and negative feedback about their group project. Participants were instructed to forward all emails and voicemails to the instructor after completing the appropriate forms. They delivered their completed questionnaires to the instructor in person (four were delivered by fax), including a list of the actual positive and negative feedback they had compiled for their project reviews.

Measures and Preliminary Analyses

The Appendix lists the items included in the questionnaire. The items for Content Quality are a subset of Bailey and Pearson's (1983) computer users' satisfaction instrument. As no previously validated Valence items were available from the literature, these were developed by the author and pre-tested on 12 Ph.D. students. Three items were trimmed as a result of this analysis and the resultant measure was piloted on a convenience sample of 16 MBA students who did not participate in the actual study. Analysis of these results led to the conclusion that the measure had stable psychometric properties. The construct of Comfort with the Communication Experience was measured with seven-point Likert items from Sussman and Sproull (1999). Since objective feedback effectiveness measures were likewise unavailable, Feedback Effectiveness was operationalized in terms of whether or not the feedback motivated the receivers to work harder in the future and the extent to which they viewed receiving feedback as a waste of time. This pragmatic view of feedback effectiveness reflects an evolutionary (Nelson & Winter, 1982) organizational perspective that is focused on the need to engender behavioral change and responsiveness. An item was included asking participants their extent of previous relationship with their assigned communication partner.

Tables 2 presents means, descriptive statistics and Cronbach alphas resulting from reliability tests of each construct for the full sample of 65 senders and receivers, minus one respondent who provided incomplete data. Table 3 presents these statistics for the 47 matched-pair dyads. Table 4 presents results of Pearson correlations between sender constructs, and Table 5 shows Pearson correlations between receiver constructs. Discriminant validity between constructs was confirmed using factor analyses. Measures are consistent with assumptions of normality underlying the analyses that follow.

In a power analysis, a commonly specified power level of 0.8 and an alpha value of 0.05 were used to derive a required sample size that matches the pre-identified level of effect size. I assumed that the theoretical model will generate an R² level of 0.26, which is reasonable for social-psychological studies (Cohen, 1988). Thus the required sample size is derived to be 32 for a model such as this that contains four independent variables. Because our smallest sample size was 47 for the matched pairs, the appropriate sample size requirement has been exceeded. The extent of prior relationships item was examined at this time to seek any dyads with extensive prior relationships; none were identified. This may be due to the fact that the participants were non-cohorted, part-time students.

Analyses and Results

To investigate H1, paired sample t-tests were performed on the dyadic pairs of sender and receiver, first for those in the email condition, and then for those in the voicemail condition. Table 6 summarizes the results of these analyses below. In support of hypothesis 1, comfort levels are fairly consistent among email senders and receivers, but not between voicemail senders and receivers, since voicemail senders report lower levels of comfort than voicemail receivers.

Hypothesis 2 investigates differences in perceptions of feedback valence within each dyadic pair. To assess H2a, paired sample t-tests were performed on the sender/receiver pairs, first for those in the email condition and then for those in the voicemail condition. Table 6 above also summarizes results of this analysis. The pattern revealed is consistent with H2a, such that perceptions of feedback negativity are fairly consistent among voicemail senders and receivers, but not between email senders and receivers, since email senders report higher levels of perceived negativity than voicemail receivers. One-way ANOVA confirmed that this effect is moderately significantly different across media (3.34, p=.074). Hypothesis 2b investigates differences in perceptions of feedback positivity within each dyadic pair. Again, paired sample t-tests were performed on the sender/receiver pairs, first for those in the email condition, and then for those in the voicemail condition. As hypothesized, no significant differences exist in pairs' perceptions of positivity, regardless of media.

Hypothesis 3 investigates associations between perceived feedback valence and feedback effectiveness for receivers. The sample of receivers was split according to medium, and feedback effectiveness was regressed first onto perceived negativity and then onto perceived positivity for the split samples. Table 7 below summarizes the results of these analyses. For email receivers, perceptions of content negativity are not associated with their reported feedback effectiveness. But for voicemail receivers, the more negative the content, the lower the reported feedback effectiveness. This analysis was repeated for perceptions of content positivity and is also shown in Table 7. Taken together, these results support H3: Content valence influences perceptions of feedback effectiveness for voicemail receivers, but not for email receivers.

Hypotheses 4a and 4b compare the relationship of central and peripheral processing to receivers' reported feedback effectiveness. Since the peripheral cue of negativity was minimally associated with feedback effectiveness for participants in both conditions as compared to the effect of positivity, hypotheses 4a and 4b were investigated using perceived positivity as the sole peripheral cue in the model. This is consistent with previous research findings that positive reinforcement and supportive comments are more strongly associated with feedback effectiveness than negative reinforcement or comments (London, Larsen, & Thisted, 1999; Sorensen & Franks, 1972).

For H4a, email receivers' feedback effectiveness was regressed onto both their perceived content quality and their perceived positivity simultaneously. The model was significant. As predicted, positivity was not significant while content quality was. This analysis was repeated for voicemail receivers, and again the overall model was highly significant. As hypothesized, positivity was significant, and surprisingly, so was content quality. Comparing the models of H4a and H4b, I find that they are significantly different from one another according to the Chow test (F=4.016, p=.012), as seems obvious from the large difference in variance explained (Chow, 1960). Table 8 shows the details of these analyses.

Exploratory Analyses

The above analyses are revealing about feedback as it is received via voicemail, but less so about feedback effectiveness for receivers of email feedback. Thus, an exploratory investigation was undertaken in an attempt to unearth determinants of feedback effectiveness for email receivers. The full receiver sample was split on negativity, and using independent samples t-tests, means were compared across positivity, feedback effectiveness, and content quality for only those receivers reporting above median negative feedback. No significant differences between email and voicemail recipients were found for positivity or content quality, but feedback effectiveness means were significantly higher for email receivers than for voicemail (t=2.21, p=.047, df=12). Means were then compared across the same three for the sample of below median negative feedback. No significant differences by medium were found.

Feedback effectiveness means for this sample are shown in Table 9 for comparison purposes. The results indicate that feedback is significantly more effective in the email condition for very negative feedback. For less negative feedback, effectiveness is higher in the voicemail condition, but not significantly so. These results contrast with the lack of means differences by medium for feedback effectiveness in general.

Discussion

This study presents a theoretical model describing ways that media can affect the process of evaluative feedback when it occurs via CMC. The results broadly support the hypotheses suggested by the model. From the results for H1 we learn that, as reported immediately following feedback delivery and receipt, voicemail senders are more uncomfortable with the interaction than the receivers of their messages are, significant to p<.10. While this significance level is low, there is some evidence to suggest an asymmetric experience of comfort level within voicemail dyads but not within email ones. Figure 2 visually depicts this finding.

This phenomenon bears further investigation, since asymmetrical discomfort during a communication experience may have unintended and potentially adverse consequences on a relationship over time. Conceivably a cycle could be established in which a feedback sender's discomfort increases over time as he or she anticipates future discomfort on the basis of his or her prior delivery experiences. In such cases, experienced discomfort could become self-reinforcing. Presumably, when a senders' discomfort level gets sufficiently high, the receiver may also begin to feel discomfort as he or she mirrors the experience of the communication partner, and this may impede the effectiveness of the delivered feedback. While this study did not investigate media choice, discrepant within-dyad comfort experiences imply that email may be a better medium for delivering evaluative feedback, since levels of comfort did not differ within email dyads.

Within-dyad differences were also found regarding how negative the feedback was perceived to be. Figure 3 presents this finding visually. Within pairs, email senders viewed the content of their feedback as being significantly more negative than did the receivers of their messages (H2a). As expected, this was not found to be the case for perceptions of feedback positivity (H2b).

We know that mediated communication decreases the tendency to sugarcoat negative information when delivering it (Sussman & Sproull, 1998); this was the reason for theorizing that email senders will view their feedback as negative, since they are not sugarcoating it. Because social norms dictate sugarcoating, feedback that is not sugarcoated may appear to be negative, when in reality it is simply bald. Another possibility is that senders accurately view their comments as quite negative. However, since message receivers do not share these negative perceptions, the former explanation of this phenomenon seems more likely.

As with discrepant comfort levels, the discrepancy in negativity perceptions within email dyads could affect the ongoing deliver/receiver relationship. When senders believe that their messages are quite negative, they may imagine that they have breached social norms by not sugarcoating. Such a breach of norms may affect how the sender interacts with the receiver during future communications. If this happens, and since receivers did not interpret their feedback as particularly negative, they may view any behavior changes in the sender as unexpected and unwarranted. In this way, the nature of future communication exchanges may potentially be affected by the email medium. Research is needed to investigate the effects of this discrepancy on ongoing relationships. As hypothesized, this within-pair discrepancy in negativity perceptions for those using email was not found within voicemail dyads.

It is interesting that the perceived negativity differences within email pairs were not accompanied by differences in experienced comfort levels, as one might expect. Note that the sample investigated here is small, and the statistical tests used are not designed to identify small effect sizes. For this reason, non-significant findings do not confirm that an effect does not exist; this limitation should be taken into account. However, for this sample, high perception of content negativity was not associated with discomfort for email pairs, as it might have if the communication medium supported more social presence information. Conversely, differences in experienced comfort levels in voicemail dyads were not accompanied by differences in negativity perceptions. For voicemail dyads, discomfort was not "caused" by the negativeness of the feedback, but rather, apparently, by the social presence information made available by the medium.

We now turn to the experience of the feedback receiver, acknowledging the difficulties inherent in measuring feedback effectiveness (Ausubel, 1963). These results need to be interpreted with caution in light of the pragmatic operationalization of feedback effectiveness used here, and are not generalizable to other conceptions of feedback effectiveness. Hypothesis 3 investigates the role of the medium in the relationship between feedback effectiveness and feedback valence—that is, perceptions of its negativity and positivity. For email receivers, there was no relationship between valence and feedback effectiveness, while for voicemail receivers the relationships between perceived negativity and positivity and feedback effectiveness were both significant for voicemail receivers (and especially so for negative valence cues). Figure 4 summarizes these results pictorially.

This medium-based difference suggests two possible explanations. The first, as suggested by theory, is that negativity and positivity operate as peripheral cues that are associated with aspects of the sender that do not come through on text, such as vocal intonation and emphasis. A second explanation is that these cues are available to email receivers, but for some reason these receivers do not attend to them. In the absence of any theoretical or practical reason to suggest the latter, I advance the former explanation as the underlying cause of this finding. In terms of dual-process theories such as the ELM, these results suggest that feedback valence can serve as a peripheral cue when it is available in the form of vocal tone and emphasis. However, since vocal tone and emphasis are not available to email receivers, these receivers have less access to this type of peripheral cue. Thus they are less able to use valence as a peripheral cue informing their attitude towards the feedback and its effectiveness. This is consistent with prior research indicating that positive feedback has a larger impact on feedback effectiveness than does negative feedback (Argyris, 1994). This implies that voicemail feedback delivery and face-to-face feedback delivery share the common characteristic of the salience of feedback valence.

Finally, we turn to the relationship of perceived feedback quality and feedback effectiveness, relative to the relationship of valence cues and feedback effectiveness, for both media conditions. Having found that positive feedback valence has a much stronger relationship to feedback effectiveness than negativity for all receivers, I used positive valence as the peripheral cue for this analysis. As illustrated in Figure 5 below, positivity explains 58% of the variance in feedback effectiveness for voicemail receivers, but is insignificantly related to feedback effectiveness for email receivers. These findings support the premise that feedback valence is more influential in the "richer" medium, since it enables transmission of vocal tonality and emphasis cues, and less influential in the "leaner" medium, which does not support transmission of these cues.

The ELM suggests that content quality should explain a significant amount of variance in feedback effectiveness perceptions for email receivers, since central processing should occur in the absence of cues for peripheral processing. This was found to be the case. Positivity was insignificant in this model for email receivers—content quality explained 15% of the variance in feedback effectiveness. Thus for email receivers, results support the application of the ELM to the evaluative feedback process. The voicemail results were not as consistent with theory, however. For voicemail receivers, content quality and perceived positivity were of roughly equal influence on feedback effectiveness, together explaining 62% of the variance in feedback effectiveness. This suggests that these voicemail participants, rather than making tradeoffs between peripheral and central processing, utilize all available cues to assess received evaluative information. This is not supported by ELM, according to which central processing is undertaken when peripheral processing is insufficient (Petty & Cacioppo, 1986). However, these results are supported by another widely accepted dual-process theory—the Heuristic Systematic Model. According to this model, proposed by Chaiken and Eagly (1983), central and peripheral processes occur simultaneously and are not mutually exclusive. The results reported here are thus consistent with this variant of a dual-processing approach to feedback delivery.

Conclusion

Further research is needed to unearth the mechanisms underlying feedback effectiveness for email receivers, since this model explains much less about feedback effectiveness via email. Research is needed to identify those peripheral cues and content variables that contribute to perceptions of feedback effectiveness for email receivers, since the valence cues investigated here are not readily available from email. Feedback effectiveness levels are similar across both media conditions. Interestingly, the exploratory analysis revealed that for receivers who view their feedback as highly negative, feedback effectiveness is significantly higher for those in the email condition than in the voicemail condition. I conjecture that for very bad news, lack of social presence cues in the email condition serves to reduce defensive reaction on the part of the receiver, with a corresponding increase in feedback effectiveness. For content perceived as less negative, feedback effectiveness is higher for those receiving feedback via voicemail, where positivity cues are more readily available. These findings are not conclusive, but are included here to suggest a possible avenue for future research.

One limitation of this study is that the evaluative feedback process studied here is outside of its naturally occurring context—an ongoing interpersonal relationship, and the relationship clearly impacts CMC use over time (Chidambaram, 1996). However, participants in this study were in a relationship with one another, since they evaluated a classmate at the end of the semester, after having sat in class with them over the course of three months. An important avenue for future research is how these results play out in the context of an ongoing relationship as it transpires over time. Another limitation of the study is the fact that evaluation materials did not contribute directly to the course grade of either sender or receiver, so the feedback delivered was less consequential than if it had real personal ramifications for the receiver. A more realistic feedback scenario would have more closely approximated the power asymmetry that tends to accompany performance appraisals in organizational contexts.

For practitioners engaged in knowledge work that depends on evaluative feedback, this study has several implications. Deliverers of email feedback should be aware of the heightened negativity perceptions of email senders but understand that feedback receivers may not share these perceptions of feedback negativity. However, there does not appear to be any differences in feedback effectiveness on the basis of medium alone. Senders should be aware of their potentially biased perceptual filters with regard to negativity, lest these unnecessarily affect the ongoing evaluative relationship over time. Voicemail senders should be aware that they have a predisposition to higher levels of discomfort than their communication partners, and look for implications of this as they relate to their feedback receivers over time. Also, since voicemail receivers perceptions’ of feedback effectiveness were associated with how positive they perceived the feedback to be, voicemail senders should use sugar instead of vinegar in their evaluations.

This study found empirical support for a theoretical model of the evaluative feedback process when it makes use of new media to overcome geographic distance. These findings are not intended to suggest that CMC or voice messages are preferable to face-to-face interaction for evaluative feedback. They do shed light on the feedback process when face-to-face interaction is not an option and communication must occur over distance, however. As education and work become increasingly global, many evaluative feedback processes will come to resemble the mediated phenomena that have been investigated here. The better these phenomena are understood, the better equipped researchers and practitioners will be to design evaluative processes around highly effective feedback delivery.

References

Alavi, M. (1994). Computer-mediated collaborative learning: An empirical evaluation. MIS Quarterly, 18 (2), 159-174.

Andreoli, V., & Worschel, S. (1978). Effects of media, communicator, and message position on attitude change. Public Opinion Quarterly, 42 (1), 59-70.

Applegate, L., Konsynski, B., & Nunamaker, J. (1986). A group decision support system for idea generation and issue analysis in organization planning. In H. Krasner & I. Greif (Eds.), Proceedings of the First Conference on Computer Supported Cooperative Work (pp. 16-34). Austin, TX: ACM Press.

Argyris, C. (1994). Good communication that blocks feedback effectiveness. Harvard Business Review, 72 (4), 77-85.

Ausubel, D. (1963). The Psychology of Meaningful Feedback Effectiveness. New York: Grune and Stratton.

Bailey, J., & Pearson, S. (1983). Development of a tool for measuring and analyzing computer user satisfaction. Management Science, 29 (5), 530-545.

Benbunan-Fich, R., & Hiltz, R. (1999). Impacts of asynchronous learning networks on individual and group problem solving: A field experiment. Group Decision and Negotiation, 8 (5), 409-426.

Bond, C. Jr., & Anderson, E. (1987). The reluctance to transmit bad news: Private discomfort or public display? Journal of Experimental Social Psychology, 23 (2), 176-187.

Caldwell, B., Uang, S., & Taha, L. (1995). Appropriateness of communications media use in organizations—Situation requirements and media characteristics. Behavior and Information Technology, 14 (4), 19-207.

Carnevale, D. (2005). Online courses continue to grow, report says. The Chronicle of Higher Education, July 8, p. 29.

Chaiken, S., & Eagly, A. (1983). Communication modality as a determinant of message persuasiveness and message comprehensibility. Journal of Personality and Social Psychology, 34, 605-614.

Chidambaram, L. (1996). Relational development in computer-supported groups, MIS Quarterly, 20 (2), 143-167.

Chow, G. (1960). Test of equality between sets of coefficients in two linear regressions. Econometrika, 28 (3), 591-605.

Claiborn, C. (1986). Social influence: Toward a general theory of change. In F. J. Dorn (Ed.), Social Influence Processes in Counseling and Psychotherapy (pp.31-42). Springfield, IL: Thomas.

Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences. Hillsdale, NJ: Lawrence Erlbaum Associates Inc.

Connolly, T., Jessup, L., & Valacich, J. (1990). Effects of anonymity and evaluation tone on idea generation in computer-mediated groups. Management Science, 36 (6), 689-703.

Culnan, M., & Markus, M. (1987). Information technologies. In F. Jablin, L. Putnam, K. Roberts, & L. Porter (Eds.), Handbook of Organizational Communication (pp. 420-444). Newbury Park, CA: Sage Publications.

Daft, R., & Lengel, R. (1986). Organizational information requirements, media richness, and structural design. Management Science, 32 (5), 554-571.

Daft, R., & Lengel, R., & Trevino, L. (1987). Message equivocality, media selection, and manager performance. MIS Quarterly, 11 (3), 355-366.

Dennis, A., & Valacich, J. (1993). Computer brainstorms: More heads are better than one. Journal of Applied Psychology, 78 (4), 531-537.

Druskat, V., & Wolff, S. (1999). Effects and timing of developmental peer appraisals in self-managing work groups. Journal of Applied Psychology, 84 (1), 58-74.

Duthler, K. (2006). The politeness of requests made via email and voicemail: Support for the hypersonal model. Journal of Computer-Mediated Communication, 11 (2), article 6. Retrieved December 26, 2006 from http://jcmc.indiana.edu/vol11/issue2/duthler.html

El Shinnawy, M., & Markus, M. (1997). The poverty of media richness theory: Explaining people's choice of electronic mail versus voicemail. International Journal of Human-Computer Studies, 46 (4), 443-467.

Epley, N., & Kruger, J. (2005). When what you type isn't what they read: The perseverance of stereotypes and expectancies over email. Journal of Experimental Social Psychology, 41 (4), 414-422.

Feldman, J., & Lynch, J. (1988). Self generated validity and other effects of measurement on belief, attitude, intention, and behavior. Journal of Applied Psychology, 73 (3), 421-435.

Fulk, J., & Mani, S. (1986). Distortion of communication in hierarchical relationships. In M. McLaughlin (Ed.), Communication Yearbook 9 (pp. 483-510). Newbury Park , CA: Sage.

Fulk, J., Schmitz, J., & Steinfeld, C. (1990). A social influence model of technology use. In J. Fulk & C. W. Steinfeld (Eds.), Organizations and Communications Technology (pp. 117-140). Newbury Park, CA: Sage Publications.

Gallupe, R., Dennis, A., Cooper, W., Valacich, J., Nunamaker, J. Jr., & Bastianutti, L. (1992). Electronic brainstorming and group size. Academy of Management Journal, 35 (2), 350-369.

Gergen, K. (1982). Toward Transformation of Social Knowledge. New York: Spriger-Verlag.

Gerstner, C., & Day, D. (1997). Meta-analytic review of leader-member exchange theory: Correlates and construct issues. Journal of Applied Psychology, 82 (6), 827-844.

Heckman, R., & Annabi, H. (2005). A content analytic comparison of learning processes in online and face-to-face case study discussions. Journal of Computer-Mediated Communication, 10 (2), article 7. Retrieved December 26, 2006 from http://jcmc.indiana.edu/vol10/issue2/heckman.html

Hiltz, S., Johnson, K., & Turoff, M. (1986). Experiments in group decision making. Human Communication Research, 13 (2), 225-252.

Ilgen, D., Barnes-Farrell, J., & McKellin, D. (1993). Performance appraisal process research in the 1980s: What has it contributed to appraisals in use. Organizational Behavior and Human Decision Processes, 54 (3), 321-368.

Ilgen, D. R., Fisher, C. D., & Taylor, M. S. (1979). Consequences of individual feedback on behavior in organizations. Journal of Applied Psychology, 64 (4), 349-371.

Jacobs, A. (1974). The use of feedback in groups. In A. Jacobs & W. E. Spradlin (Eds.), The Group as an Agent of Change (pp. 408-448). New York: Behavioral Publications.

Kiesler, S., & Sproull, L. (1982). Managerial response to changing environments: Perspectives on problem sensing from social cognition. Administrative Science Quaterly, 27 (4), 548-570.

Kiesler, S. & Sproull, L. (1992). Group decision making and communication technology. Organizational Behavior and Human Decision Processes, 52 (1), 96-123.

Kiesler, S., Zubrow, D., Moses, A., & Geller, V. (1985). Affect in computer-mediated communication: An experiment in synchronous terminal-to-terminal discussion. Human Computer Interaction, 1 (1), 77-104.

Kivlighan, D. M. (1985). Feedback in group psychotherapy: Review and implications. Small Group Behavior, 16 (3), 373-385.

Kluger, A., & DeNisi, A. (1996). The effects of feedback interventions on performance: A historical review, a meta-analysis, and a preliminary feedback intervention theory. Psychological Bulletin, 119 (2), 254-284.

Kurtzberg, T., Naquin, C., & Belkin, L. (2005). Electronic performance appraisals: The effects of email communication on peer ratings in actual and simulated environments. Organizational Behavior and Human Decision Processes, 98 (2), 216-226.

Larson, J. (1986). Supervisors' performance feedback to subordinates: The impact of subordinate performance valence and outcome dependence. Organizational Behavior and Human Decision Processes, 37 (3), 391-408.

Lee, F. (1993). Being polite and keeping MUM: How bad news is communicated in organizational hierarchies. Journal of Applied Psychology, 23 (14), 1124-1149.

Levy, P., & Williams, J. (2004). The social context of performance appraisal: A review and framework for the future. Journal of Management, 30 (6), 881-905.

Locke, E., Cartledge, N., & Koeppel, C. (1968). Motivational effects of knowledge of results: A goal-setting phenomenon. Psychological Bulletin, 70 (6), 474-485.

London, M., Larson, H., & Thisted, L. (1999). Relationships between feedback and self-development. Group and Organization Management, 24 (1), 5-27.

Markus, M. (1994). Electronic mail as the medium of managerial choice. Organization Science, 5 (4), 502-527.

Mawhinney, T. (2005). Effective leadership in superior-subordinate dyads: Theory and data. Journal of Organizational Behavior Management, 25 (4), 37-79.

Mizruchi, M., & Marquis, C. (2006). Egocentric, sociocentric, or dyadic? Identifying the appropriate level of analysis in the study of organizational networks. Social Networks, 28 (3), 187-208.

Nadler, D. (1979). The effects of feedback on task group 'margin-left:.5in;text-indent:-31.5pt'>Nelson, R., & Winter, S. (1982). An Evolutionary Theory of Economic Change. Cambridge, MA: Harvard University Press.

Nunnally, J. (1967). Psychometric Theory. New York: McGraw-Hill.

O'Neal, E., Levine, D., & Frank, J. (1979). Reluctance to transmit bad news when the receiver is unknown: Experiments in five nations. Social Behavior and Personality, 7 (1), 39-47.

Pallak, S. (1983). Salience of a communicator's physical attractiveness and persuasion: A heuristic versus systematic processing interpretation. Social Cognition, 2, 122-141.

Petty, R., & Cacioppo, J. (1986). Communication and Persuasion: Central and Peripheral Routes to Attitude Change. New York: Springer-Verlag.

Rice, R. (1992). Task analyzability, use of new media, and effectiveness: A multi-site exploration of media richness. Organization Science, 3 (4), 475-500.

Rice, R. (1993). Media appropriateness: Using social presence theory to compare traditional and new organizational media. Human Communication Research, 19, 451-484.

Rice, K., & Warner, N. (1994). Breaking the bad news: What do psychiatrists tell patients with dementia about their illness? International Journal of Geriatric Psychiatry, 9 (6), 467-471.

Rice, R., & Tyler, J. (1995). Individual and organizational influences on voicemail use and evaluation. Behavior and Information Technology, 14 (6), 329-341.

Robert, L., & Dennis, A. (2005). Paradox of richness: A cognitive model of media choice. IEEE Transactions on Professional Communication, 48 (1), 10-21.

Rosen, S., Grandison, R., & Stewart, J. (1974). Discriminatory buckpassing: Delegating transmission of bad news. Organizational Behavior and Human Performance, 12, 249-263.

Rosen, S., & Tesser, A. (1970). On the reluctance to communicate undesirable information: The Mum effect. Sociometry, 33 (3), 253-263.

Schweizer, K., Paechter, M., & Weidenmann, B. (2000). Distance eduation: Experiences of a virtual tutor. International Journal of Psychology, 35 (3-4), 136-145.

Seale, C. (1991). Communication and awareness about death: A study of a random sample of dying people. Social Science and Medicine, 32 (8), 943-952.

Short, J., Williams, F., & Christie, B. (1976). The Social Psychology of Telecommunications. New York: John Wiley.

Sorensen, J., & Franks, D. (1972). The relative contribution of ability, self-esteem, and evaluative feedback to performance. The Accounting Review, 47 (4), 735-746.

Sproull, L., & Kiesler, S. (1986). Reducing social context cues: Electronic mail in organizational communication. Management Science, 32, 1492-1512.

Stone, E. (1978). Research Methods in Organizational Behavior. Glenview, IL: Scott, Foresman, & Co.

Straub, D., & Karahanna, E. (1998). Knowledge worker communications and receiver availability: Toward a task closure explanation of media choice. Organization Science, 9 (2), 160-175.

Subramanian, G., & Rohrer-Meek, M. (1998). User perceptions of voicemail technology: A comparative study of internal and external users. Journal of Computer Information Systems, 39 (2), 48-53.

Sussman, S., & Siegal, W. (2003). Mediated knowledge transfer in organizations: The role of information utility in persuasion. Information Systems Research, 14 (1), 47-65.

Sussman, S., & Sproull, L. (1999). Straight talk: Delivering bad news through electronic communication. Information Systems Research, 10 (2), 150-166.

Tesser, A., & Rosen, S. (1975). The reluctance to transmit bad news. In L. Berkowitz (Ed.), Advances in Experimental Social Psychology 8 (pp. 193-232). New York: Academic Press.

Tesser, A., Rosen, S., & Tesser, M. (1971). On the reluctance to communicate undesirable messages (the Mum effect): A field study. Psychological Reports, 29 (2), 651-654.

Trevino, L., Webster, J., & Stein, E. (2000). Making connections: Complementary influences on communication choices, attitudes, and use. Organization Science, 11 (2), 163-182.

Valacich, J., Dennis, A., & Connolly, T. (1994). Idea generation in computer-based groups: A new ending to an old story. Organizational Behavior and Human Decision Processes, 57 (3), 448-467.

Volle, L. (2005). Analyzing oral skills in voice e-mail and online interviews. Language Learning and Technology, 9 (3), 146-163.

Waitzkin, H. (1984). Doctor-patient communication: Clinical implications of social scientific research. Journal of the American Medical Association, 252 (17), 2441-2446.

Walther, J. (1992). Interpersonal effects in computer-mediated interaction. Communication Research, 19 (1), 52-90.

Walther, J.B. (1996). Computer-mediated communication: Impersonal, interpersonal, and hyperpersonal interaction, Communication Research, 23 (1), 3-43. 

Weisband, S., Schneider, S., & Connolly, T. (1995). Computer-mediated communication and social information: Status salience and status differences. Academy of Management Journal, 38 (4), 1124-1151.

Worschel, S., Andreoli, V., & Eason, J. (1975). Is the medium the message? A study of the effects of media, communicator, and message characteristics on attitude change. Journal of Applied Social Psychology, 5 (2), 157-172.

Appendix: Questionnaire Items

Items completed immediately following feedback delivery:

Comfort with the experience:

Negativeness of negative feedback:

Postiveness of positive feedback:

Items completed immediately following receipt of feedback:

Negativeness of negative feedback:

Postiveness of positive feedback:

Feedback effectiveness:

Content Quality:

Please rate the quality of the negative feedback you just received according to the following adjectives:

About the Author

Stephanie A. Watts is an assistant professor of Information Systems at the Boston University School of Management. She was previously on the faculty at Case Western Reserve University. Her research focuses on the impact of computer-mediated communication on knowledge sharing in organizations.
Address: Department of Information Systems, Boston University School of Management, 595 Commonwealth Ave., Boston, MA 02215, USA