Telling ’em How It Will Be

Abstract

Implementers need to decide the degree to which to preview the challenges and possible downsides of a change process. Scant research has explored the announcement of planned change—especially regarding the previewing of potential painful or stressful effects of the process. This study uses a pen-and-paper experimental method with a sample of 218 working adults to examine the extent to which acknowledging potentially negative aspects of change in announcements heightens perceptions of honesty and trustworthiness of implementers. Also, we sought to explore the effects of these negative previews on initial favorability toward changes and on stakeholders’ subsequent communication targets and purposes for communication. We found that previews of possible negatives did not increase initial favorability or judgments of credibility of implementers. We found high-risk change creates a challenging context. Future research needs to consider whether refutational messages are necessary for high-risk change announcements.

Keywords

change management change announcement sidedness realistic previews bad news trust risk

Change implementers face a critical practical question each time they start a change process—how to introduce the change stakeholders are about to experience. Although change in organizations is often preceded by rumors and leaks about the nature, rationale, and likely consequences of change, a formal announcement of the change is nearly always crafted and disseminated by implementers. Announcing change is an important role of change agents (Lewis, Schmisseur, Stephens, & Weir, 2006). Changes can differ significantly in terms of magnitude, risk, and potential disruption. Therefore, implementers have to make decisions about ways to introduce changes of various types. That is, the ways in which messages about change are initially framed for stakeholders important change outcomes. Changes that are likely to be controversial or perceived negatively may suggest to implementers a different message strategy than those changes that are likely to be perceived as benign or minor in nature. However, we have little direct evidence of the better announcement strategies in either situation. We know very little of how stakeholders respond to change announcements of various types, especially messages that preview negative experiences that will come with change.

To this point, scant research has explored initial formal messages about planned change. Smeltzer and Zener have published a handful of articles on the subject of effective announcement of planned changes (Smeltzer & Zener, 1992, 1994, 1995; see also Smeltzer, 1991, 1995). In Smeltzer’s (1991) article, he concludes, based on interviews with implementers of changes in 43 large organizations, with a description of commonly perceived predictors of failure of change announcements. Furthermore, we do not have a strong empirical database to explain if message framing promotes specific stakeholder reactions to change and what, if any, affects those initial messages have on their subsequent attitudes and communication.

Understanding which message elements create what effects in stakeholders will be helpful in predicting reactions and crafting persuasive messages. This study focuses on assessing the impact of previewing negative elements of change in initial announcements and whether perceptions of riskiness of change made a difference in how these messages were perceived by stakeholders. We were interested in understanding stakeholders’ assessments of implementers as well as their subsequent communicative activity after receiving announcements of change programs. Our goal was to determine the extent to which making reference to potentially negative aspects of change at the start of implementation might heighten perceptions of honesty and trustworthiness of implementers. Also, we sought to explore the effects of these negative previews on initial favorability toward changes and on stakeholders’ subsequent communication targets and purposes for communication. We next discuss the issue of trust in the context of organizational change.

Trust is a critical issue in organizations in general, and communication plays an important role in development of trust (Thomas, Zolin, & Hartman, 2009). Lines, Selart, Espedal, and Johansen (2005) argue that organizational change is a “critical trust building or trust destroying episode in a long term and ongoing relationship between the organization, represented by its management, and non managerial employees” (p. 222). Lines et al. argue that the level of uncertainty and vulnerability experienced by organizational members during change lead to an active processing of trust-relevant information. They go on to suggest that among other factors, openness matters in the determination of trust. In the Schoorman, Mayer, and Davis (2007) model of organizational trust, the authors argue that trust leads to risk taking. For these authors and others (cf. Thomas et al., 2009), trust is, in part, a willingness to be vulnerable to another party. Thomas et al. (2009) found in their study of trust that both quality and quantity of information available in the organization enhances levels of employee trust.

The consequences of trusting relationships for subsequent communication behaviors, especially during a stressful change are not well understood. For example, we do not know what effects stakeholders’ perceptions of implementers’ trustworthiness has on assertive advocacy behaviors (e.g., lobbying for a change to be altered or stopped) or on more cooperative, but engaged behaviors such as requesting more information. However, Thomas et al. (2009) found that those reporting higher levels of trust were more “involved” in their organizations, suggesting that trust has important implications for organizations.

Review of Literature

There are several important bodies of literature that provide a backdrop for our study of previewing potential negative aspects of change. We review here key findings and theory concerning risky change, delivering bad news, realistic previews, message sidedness, and responses to bad news. Each of these distinct literatures provides evidence of how we might expect change announcements that preview pain to be received by stakeholders. In some cases, hearing bad news or early previews of possible negative downsides of events and circumstances can enhance a person’s ability to cope. In other cases, it might not be appreciated or well received. We look to the current empirical literature for clues to how previewing pain in the context of organizational change might be received.

Risky Change

The concept of risk during organizational change is rarely discussed in the management or organizational literature. However, we do have both conceptual and empirical scholarship that helps describe the magnitude of changes. For example, Bartunek and Moch (1987) outline three different types of change referred to as orders of change magnitude: first-, second-, and third-order changes. However, as Lewis (2011) observes about this means of assessing magnitude of change, stakeholders often view changes in different ways. That is, change may be viewed as relatively minor by some and as significant for others because individuals’ tolerances for change vary as do their perceptions of size and scope. She argues:

Our individual assessments of the size and scope of change are effected by how directly the change effects [sic] us; how profound the change to our own lives may be; what we value in our organizational lives; our own history with change in our personal and organizational life; and perhaps most profoundly, the interactions we have with others about the change. (Lewis, 2011, p. 39)

One significant way to understand magnitude of change and the risk change represents for various stakeholders is to consider the threats that particular changes represent. Threats during change can range from increases in stress, workload, hassles, and job insecurity to threats to personal relationships, political and personal capital, and real or perceived loss of competence in one’s job. As these potential threats increase, one would assume stakeholders would consider the change more “risky.” However, little is known in organizational science about how individuals assess these sorts of risks and how assessments of “risky situations” affect behaviors or the reception of managers’ messages.

Delivering Bad News

People are generally more hesitant to share bad news than good news (Dibble & Levine, 2010; McGlone & Batchelor, 2003). This is referred to as the “mum effect.” As McGlone and Batchelor argue, reference to a negative object can be a threat to the face of the recipient as well as the communicator. The communicator in this situation often avoids the delivery of bad news, or selects a euphemistic means of delivery to soften the blow (“rightsizing” rather than “layoffs”). According to Dibble and Levine’s (2010) review of this effect, researchers have found that senders, who anticipate the need to deliver bad news, will find ways to delay negative feedback, distort information to appear less negative, and seek less immediate channels to deliver news. Their own research suggests that senders with bad news to deliver might be both reluctant to share bad news as well as eager to share good news. Thus, their construction of messages may be equally driven by a desire to stay upbeat as well as to avoid negative information.

For example, in the medical context, research has explored the needs of both doctors and patients in the delivery of bad news. Del Vento, Bavelas, Healing, MacLean, and Kirk (2009) suggest that difficulties in delivering bad news in this context arise from the joint goals doctors have to be clear about diagnosis and prognosis while being kind. Results of their study found that in delivering bad news, doctors tended to use alternative terms for diagnosis, qualify their evaluation, underemphasize certainty, and subtly separate the patient from the disease. These results are consistent with earlier research (Bavelas, Black, Chovil, & Mullet, 1990) suggesting that when all communicative options lead to negative outcomes (e.g., a double-bind of either lying or delivery of harsh news), senders frequently use implicit language whose meaning has to be inferred by the recipient rather than being explicitly stated by the speaker.

The medical context differs from the context of implementation of organizational change in significant ways. Doctors rarely, if ever, have a choice of what news they deliver (since they cannot determine the disease or prognosis of their patients), whereas in the case of organizational change, implementers (or decision makers that they represent) have usually selected the future course for the organization and then announce it. Thus, implementers have a level of accountability for the situation that doctors do not. Furthermore, doctor-patient communication is typically dyadic in nature whereas implementer-stakeholder communication is typically one-to-many. Implementers have to weigh the reactions of multiple audiences in announcing change, whereas doctors have the ability to tailor messages for particular patients.

Another difference may be that in doctor-patient contexts or even general interpersonal contexts, euphemism, overly positive assessments, and avoidance of bluntness may be appreciated and thought of as kind and/or polite. However, evidence suggests that although this behavior is commonplace in organizations (Kline, Simunich, & Weber, 2009), it is likely to lead to ridicule and negative assessments by receivers (Smeltzer, 1991).

Realistic Previews

A robust literature in several contexts suggests the benefits of realistic previews of future experience. Realistic previews have been found to have positive effects in contexts of job interviews (Brett, Gustello, & Aderman, 1982; Faller, Grabarek, & Ortega, 2010; Phillips, 1998; Richardson, McBey, & McKenna, 2008; Wanous, 1980), education (Brinthaupt, 2004), and medical procedures (Demarest, Hook, & Erickson, 1984) among others. Realistic previews are likely to bring about more positive outcomes across contexts because they help receivers avoid negative surprises and create an environment of transparency between management and employees. Contrastingly, “broken promises” and unexpected failures during adjustments to new situations predict more negative outcomes such as reduction of trust and possibly voluntary termination.

The evidence of realistic job previews (RJPs) is mixed (Breaugh, 1983; Dugoni & Ilgen, 1981). As Breaugh (1983) notes, RJPs are intended to produce positive results (e.g., decrease voluntary and involuntary turnover) because they increase met expectations, ability to cope, and create an air of transparency, which is often equated with honesty in the organization. Forewarning of potential downsides of a job provides applicants the opportunity to self-select out of the job, to cope with the downsides better, and to trust the hiring organization. However, as Wanous (1980) and others have pointed out, RJPs will be less or not effective in cases where job applicants have no other job prospects. Additionally, RJPs are suggested to have little effect if a candidate is overly confident about his/her skills (Rynes, 1990). Furthermore, a variety of methodological shortcomings of the RJP literature have made it difficult to determine both the underlying psychological processes accounting for success of RJPs and the manner and timing of delivery of RJPs that is most effective (Breaugh, 1983).

In the change literature, we find some evidence suggesting that honesty and openness on the part of implementers is both preferred by stakeholders and tends to produce more subsequent cooperativeness in engaging the change as well as enhanced ability of stakeholders to cope with change. A rare field experiment (Schweiger & DeNisi, 1991) examined extensive realistic previews in the case of the implementation of a merger. In the experimental plant, the plant manager met frequently with employees, listened to and answered questions, and provided information about rationale, expected layoffs, detailed implications of the merger, and background. However, in the control plant employees received only notification that the merger would take place. The researchers found that providing early previews of the details of the merger and its process in the experimental plant significantly lowered uncertainty and increased job satisfaction, commitment, perceived trustworthiness, honesty, caring, and self-reported performance. These effects endured over time as the merger progressed.

Furthermore, Laster’s (2008) study of a merger that was followed by multiple additional changes found that forewarning of the complexity of change helped employees cope. Coping efficacy increased for the employees who reported receiving a multifaceted communication message about the merger. In her conceptualization, multifaceted change occurs when more than one change occurs within the same time frame. Certainly, multifaceted change presents a greater burden on stakeholders experiencing the change than changes that are more singular and/or of a smaller order. In her study, the key difference in how individual employees made sense of the size and scope of the change was how implementers introduced and discussed it in the initial stages. This had a huge effect on what these employees were expecting at the outset of change and colored their experiences of change in terms of setting them up for surprises, additional stress, and disappointment. Thus, the results suggest that employees who receive a more realistic announcement of a large, perhaps risky change will respond more favorably then employees who do not.

It is important to note that some of the realistic preview literature has been applied to circumstances where the receiver of the preview has a choice about whether to proceed or not. The preview then acts as input to a decision for the receiver. In the case of most planned organizational change, stakeholders do not have the option to avoid participation. Typically, change is considered mandatory and announcements are informational regarding what will take place to implement change rather than invitations for individuals to participate. In these contexts, previewing uncomfortable or undesirable experiences that likely will be encountered as change proceeds may not enhance recipients’ feelings of control or commitment to the change. They are not being forewarned of downsides and then being asked if they want to proceed. They are merely being told of possible future pain. Therefore, the potential benefits of providing a realistic preview of the change may not enhance outcomes, but as a result of a stakeholder’s inability to have much agency about it, may have an opposite effect, and create greater concerns during this stage of the change. However, in a few studies referenced here where change was mandatory, we see positive evidence that previews produce greater trust and coping.

Message Sidedness

Persuasion researchers dating back at least to the 1940s have explored the impact of messages that highlight only the persuader’s point of view (one-sided) versus those that acknowledge opposing arguments as well (two-sided; see, for reviews, Allen, 1998; O’Keefe, 1999). There are two sorts of two-sided messages, refutational and nonrefutational. In refutational arguments, the opposing arguments are raised and refuted (e.g., argued against). As O’Keefe (2002) describes the difference, “refutational two-sided messages characteristically attempt to undermine opposing arguments (by refuting them directly), whereas nonrefutational two-sided messages characteristically attempt to overwhelm opposing arguments (using supportive ones)” (p. 220, italics added).

This research has largely been conducted in the contexts of public policy campaigns (e.g., gun control), political messaging, organ donation, or consumer advertising. There are greatly mixed findings. Sidedness research has been premised in part on the argument that two-sided messages will increase recipients’ confidence and trust in message senders (Chebat, Filiatrault, Laroche, & Watson, 1988; Eisend, 2006), trigger active processing of the message because it is assessed as “novel” by the receiver (Eisend, 2006; Petty & Cacioppo, 1986), as well as “inoculate” them against counterarguments (Eisend, 2006; McGuire, 1961). This has been a highly complex line of research exploring many intervening variables that have been proposed to influence the relationship between sidedness and outcomes. These have included audience characteristics such as education level, self-acceptance, original position on issue, and involvement. Studies have also examined message characteristics such as numbers of negative arguments and order effects.

In his book on persuasion, O’Keefe (2002) concludes “there appears to be no general difference in persuasiveness between one-sided and two-sided messages” (p. 219). However, he concludes when considering refutational messages, two-sided refutational messages are much more persuasive than one-sided messages. Furthermore, two-sided nonrefutational messages are much less persuasive than one-sided messages. He notes an exception to this general trend. In the context of consumer advertising, credibility seems to mitigate these effects somewhat. In the consumer-advertising context, two-sided nonrefutational messages appear to increase the credibility of the sender and, at least in some cases, that credibility has enhanced the persuasion effects of the message in the desired direction (O’Keefe, 2002). O’Keefe (2002) also cautions that in some cases the increases in credibility may make the receiver pay attention to the counterarguments as well and result in decreased persuasion effects. In light of the mixed results of persuasion in this context, he concludes that senders are better off using refutation.

In contrast, in a meta-analysis of marketing sidedness research, Eisend (2006) suggests that “there are situations when two-sided advertising can be applied as a useful advertising technique” (p. 187). He argues that such an approach can be helpful when consumers hold negative beliefs, when they will be exposed to attacking arguments, or when they will receive negative publicity about a product or brand. In his analysis of 29 independent studies, he found that two-sided nonrefutational messages are more effective than one-sided messages in these contexts. Chebat and Picard (1988) argue from their review of the literature that two-sided nonrefutational messages are most effective when the audience is highly involved with regard to the object of persuasion. We also have evidence that nonrefutational two-sided messages work better for audiences who are initially negatively inclined (Pratt, 2004).

One key to success with two-sided nonrefutational messages is to balance the amount of negative information carefully. Eisend (2006) argues, “A trade-off occurs if the proportion of negative information exceeds a certain level” (p. 188). He further argues that “extreme amounts of negative information could also lead to low credibility . . . and inclusion of a refutation statement may reduce the credibility [of the sender]” (p. 188). In his conclusions, based on the meta-analysis, he suggests “a curvilinear relationship emerges depending on the proportion of negative information included in the two-sided message such that up to fifty percent of negative information with rather moderate importance does not diminish the positive credibility effects of two-sided messages” (p. 195).

O’Keefe’s (1999) review and meta-analysis found that two-sided messages led to significantly greater credibility than one-sided messages in the cases where the two-sided messages discussed supporting arguments first, followed by opposing arguments, or where the two sets of arguments were interwoven. He further argued that, in the context of advertising messages, nonrefutational two-sided arguments tend to enhance credibility because of initial audience skepticism:

Consumer advertising is likely to be met with a good deal of skepticism. . . . Given this general skepticism, consumers may well anticipate that advertisers will give a one-sided depiction of the advertised object. When instead an advertiser freely acknowledges the opposing considerations, the advertiser’s credibility will naturally be enhanced. (p. 237)

In the broader literature concerning organizational change, we find a good deal of research that suggests stakeholders have great needs for information (cf. Bordia, Hobman, Jones, Gallois, & Callan, 2004; Kramer, 1999; Kramer, Dougherty, & Pierce, 2004; Schweiger & DeNisi, 1991). Stakeholders have needs to reduce uncertainty about strategic aims, structural issues, and job-related concerns. In fact, advice books concerning the introduction of change “frequently advocate that implementers provide information that will clarify roles, tasks, responsibilities, and procedures as well as remind stakeholders of the rationale and goals of change” (Lewis, 2011, p. 60). Such advice and research findings would suggest that detailed previews are a good strategy for implementers to employ.

In the organizational change literature, we find only a few studies that examine message effects similar to sidedness. A study by Griffith and Northcraft (1996) found that “balanced” (two-sided) messages about a new technology significantly affected performance of new users in a positive direction. It is however uncertain if employees in this study perceived the technology change as a threat or a natural part of organizational evolution. In a field experiment by Miller and Monge (1985), employees so valued information during change that they found even negative information more helpful than no information. These findings would suggest that stakeholders might appreciate previews of “pain.”

In sum, people do not like to tell bad news, and perhaps strategically avoid it because it may reflect negatively on the sender, it makes people uncomfortable in future interactions, and because people prefer to highlight good news. However, presenting bad news is sometimes a strategic way to preview and highlight potential negatives to audiences who will likely discover them later (e.g., Chebat & Picard, 1988; Griffith & Northcraft, 1996). Such previews may equip receivers with a forewarning that may help them better cope (e.g., Laster, 2008), and it may enhance their views of senders as more honest and trustworthy (e.g., Schweiger & DeNisi, 1991). However, the context of implementation of planned change is much different from the contexts in which these dynamics have been investigated, leaving open the possibility that previews of negative experiences may not always be a good idea in the early stages of a change.

Subsequent Communication After Bad News

To this point there is very little empirical research on stakeholders’ responses to change announcements. The Bad News Response Model (BNRM; Sweeny & Shepperd, 2009) was developed to account for responses to negative health information. This model suggests that three aspects of bad news (the controllability of outcomes, likelihood of negative outcomes, and severity of negative consequences) predict which response (watchful waiting, active change, or acceptance) people are likely to choose. Watchful waiting involves responding to bad news in a relatively passive or conservative manner (monitoring but no aggressive action) similar to emotional coping (Folkman & Lazarus, 1980). Active change involves responding to bad news with vigorous efforts to improve the situation or preventing it from getting worse. This approach often involves information seeking in an effort to learn more about the situation and to connect with others who have experienced similar situations, preventative behaviors and treatment behaviors. This is similar to problem-focused coping (Folkman & Lazarus, 1980). In these responses, individuals are focused on attempts to change the situation rather than just manage one’s emotions. Acceptance involves directly engaging in moving forward and dealing with the consequences of the bad news rather than trying to change the outcomes of the news. “People who engage in acceptance tell others about the bad news to gain social support and a better understanding of the experience” (Sweeny & Shepperd, 2009, p. 896). Acceptance also involves accommodation of the new circumstances into the person’s life.

These authors found that people prefer watchful waiting when controllability, likelihood of negative outcomes, and severity of negative consequences are low. They prefer active change when controllability and likelihood are high. They prefer acceptance when controllability is low. This model is built on the presumption that responses to bad news are based on assessments of the situation as serious or benign.

Context of Previous Research

Most of the research on message receptiveness whether it is in mum effect or sidedness has used undergraduate students as subjects (O’Keefe, 1999). When role-play experimental methods are used, they typically involve dyadic communication where one person has to tell another news of some sort. In sidedness research, the sender of messages is frequently unknown to the receiver (as in a marketing campaign). Neither of these situations is like organizational contexts in which one-to-many announcements are made and where speaker and audiences are familiar (at least by reputation if not personally) and have preexisting histories and are engaged in ongoing relationships. Furthermore, as noted above, organizational change announcements are not typically a context in which the receiver is being asked to make a choice (preference for policy, accepting a job offer). Thus, it is unclear how well these findings about message dynamics will apply in the organizational change context.

Hypotheses and Research Questions

Given what we know about the delivery of bad news, and people’s reluctance to do so, as well as the potential positives of realistically engaging stakeholders and message recipients in transparent communication about known negatives, we argue that it likely most wise for change implementers to preview the negatives that an upcoming change may involve. Although there may certainly be merit in a refutational two-sided message (as is supported in the general sidedness literature), we consider an advertising model most relevant to this context and thus did not examine refutational messages. Much like an advertising campaign, implementers’ initial announcements of change are often an attempt to “sell” the change—despite potential difficulties that will be encountered during implementation. As in the advertising context, we expect that stakeholders receiving negative information about an upcoming change will attribute higher credibility and higher initial trust to an implementer willing to admit to the downsides of their own plan. In this model, refutation is not required in order to obtain the positive evaluation of the message sender. Additionally, we chose not to examine the effects of refutational messages as a point of comparison with one-sided and nonrefutational two-sided messages because of the complexity of testing a variety of refutational approaches (e.g., comparative advantage, tolerable loss/pain, questioning validity of counterarguments) in separate conditions.

Our study further examined cases where the riskiness of change, and potential negatives that stakeholders will encounter, is relatively low. We argue that it may be unwise for implementers to raise downsides early in the change effort. Such a strategy may give rise to unnecessary concerns that will never be encountered. This strategy may also damage the credibility of the message sender in that recipients may feel calling attention to negatives indicates a lack of competence on the part of the sender. The need for complete transparency in such cases seems unwise.

In an effort to examine how stakeholders receive change announcements that preview future negative change experiences, we have developed several hypotheses and research questions. We sought to examine (a) effects of message sidedness (one- and two-sided nonrefutational messages), in the context of low- and high-risk change, on assessments of favorability, trustworthiness, honesty; and (b) effects of risk and message sidedness on recipients’ subsequent communication (i.e., targets for communication, primary target of communication, and purpose of communication).

In alignment with our arguments above, we investigated whether one-sided and two-sided messages would produce similar levels of favorability in this context. We also predicted that two-sided messages would enhance trust in implementers and beliefs that implementers are honest.

Research Question 1: Will one-sided and two-sided messages produce similar levels of favorability toward change programs?

Hypothesis 1: One-sided and two-sided messages will produce different levels of trust and honesty assessments of implementers such that stakeholders receiving two-sided messages will assess implementers as more trustworthy and honest.

Risky change—changes that are perceived as potentially personally threatening (e.g., in terms of job security, personal relationships, and work productivity)—is more likely to result in stress for stakeholders. Thus, we would expect that those stakeholders are likely to hold implementers accountable for the selection of the specific change that is announced. As a result, we expect, in general, that perceived high-risk changes would yield lower levels of trust and honesty assessments of implementers as well as lower favorability of the change. Therefore, we propose the following:

Hypothesis 2: Perceived high- and low-risk change will produce different levels of trust and honesty assessments of implementers such that stakeholders in the high-risk change condition will assess implementers as less trustworthy and honest and be less favorable to the change.

A key interest of this study is to determine whether message sidedness would affect favorability and honesty/trust assessments differently depending on the level of risk in the type of change. We argue that in perceived high-risk situations stakeholders need the reassurance that comes with a fully transparent implementer who admits to future possible negative experience. However, in the case of low-risk change, stakeholders will consider two-sided messages as “too much information” that leads them to question the trustworthiness of the implementer (“Does he/she know what she is doing?”) and raises negatives too early. Therefore, we propose the following:

Hypothesis 3: There is an interaction between message sidedness and perceived level of risk such that in high-risk change, two-sided messages would yield higher assessments of trust and honesty and favorability, whereas in low risk change, one-sided messages would yield higher assessments of trust and honesty and favorability.

We further posed two central research questions concerning the subsequent communication behavior of stakeholders’ receiving change announcements. Research Question 2 concerns the announcement recipients’ communication targets following change announcement. Research Question 3 concerns recipients’ communication purposes of subsequent communication following change announcements.

Research Question 2: Who do stakeholders most and first want to communicate with following an announcement of change?

Research Question 3: For what purposes do stakeholders communicate following the announcement of change?

Given the expectations and findings associated with the BNRM, we predicted that in the high-risk change scenario (where the potential for negative outcomes were high and the possibility of avoiding the change—controllability—was at least moderate) stakeholders’ ratings of the importance of communication purposes associated with strong Active Change (ACS) would be higher than for the recipients of the low-risk scenario.

Hypothesis 4: Recipients of a high-risk change announcement, regardless of message sidedness, will place more importance on purposes of communication associated with ACS than will those receiving the low-risk change announcement.

Finally, we were interested in understanding what effects message sidedness will have on how recipients’ select targets for communication subsequent to change announcements, in the context of perceived high and low risk. Although we did not think that the current literature would support specific predictions as to differences in these subsequent communication behaviors, and reported purposes for communication, we sought to discern different patterns.

Research Question 4: How does message sidedness, in the context of perceived high and low risk, affect the selection of communication targets and communication purposes?

Method

To examine our hypotheses and research questions, we designed a 2 × 2 experiment manipulating perceptions of riskiness of change (high/low) and sidedness of announcement (one-sided/two-sided nonrefutational messages) in which participants were randomly assigned to read a memo introducing an organizational change and then respond to the questionnaire as if this were a real work situation in which they received this memo from an employer. Four conditions were created (a) high-risk change with one-sided announcement, (b) low-risk change with one-sided announcement, (c) high-risk change with two-sided announcement, and (d) low-risk change with two-sided announcement (see the appendix for announcement texts). We matched the word count in the memos as closely as possible to prevent length effects and attempted to add only information that altered the sidedness of the message.

Memos announcing the change framed the message either in a positive manner only (one sided) or framed the message in a balanced manner (two sided), informing the participants of both the advantages and challenges of the proposed change. In addition, there were two types of change situations created. The first change type presented low degree of risk to the participants by announcing a change in database technology of the organization, whereas the second one presented higher degree of risk to the participants by announcing the restructuration of the organization. Manipulation checks were performed to ensure we had created the intended manipulations. At the end of the questionnaire, respondents were asked to indicate differences in sidedness of the message and differences in change riskiness.

Manipulation Checks

For sidedness, respondents were asked on a 5-point Likert-type scale to indicate agreement with the statement “the memo acknowledged the upsides and the downsides of the message.” A significant difference was found between subjects assigned a one-sided and two-sided message, F(1, 213) = 29.514, p < .000. The two-sided message recipients assessed this item significantly higher (M = 2.90, SD = 1.162) than did recipients of the one-sided message (M = 2.10, SD = 1.009).

To ensure we had effectively manipulated risk of change, respondents were asked to indicate on a 5-point Likert-type scale 6 items measuring the risk of the change. Some of the items included “highly stressful to me,” “likely to cause major problems,” “threatening to my job security,” and “threatening to my ability to do my job well.” A significant difference, F(1, 214) = 36.24, p < .000, was found between responses on this scale by recipients of the high-risk (restructuring) change (M = 3.11, SD = 0.964) and those receiving the low-risk (technology) change (M = 2.34, SD = 0.890).

Procedure

To access a convenience sample of working adults as subjects, we distributed up to two questionnaires each to undergraduates enrolled in communication courses at two large state universities, one in New Jersey and one in Kansas. The students were asked to distribute the questionnaires to working adults they personally knew who had at least 3 years of work experience and were older than 21 years. Students were offered extra credit to deliver and pick up completed questionnaires from participants. The students returned the completed questionnaires in sealed envelopes with contact details of the respondents in a separate sheet attached to the envelope. For verification purposes, we selected a random set of envelopes representing 25% of the total number of returns and contacted respondents of these questionnaires to ensure authentic responses. All questionnaires we checked were verified as completed by working adults. The sample achieved is not a scientifically random sample and thus generalizability to the general population of working adults is not strictly possible. However, the sample achieved is demographically representative of our target population. Furthermore, the sample is arguably superior to other samples of organizational change that are typically restricted to a single organization and superior to samples in message effects research that have typically included only students.

Measures

Six dependent variables were assessed (see Table 1 for item sets). We constructed the initial favorability measure following Lewis and Seibold’s (1996) similar measure using a five-item scale (M = 3.32; SD = 1.001). The scale reliability was high (α = .88). We created a scale to assess perceived trustworthiness of the implementers (M = 3.20, SD = 0.9778). This construct was assessed with four items and also achieved high scale reliability (α = .84). To assess perceived honesty we created a four-item scale (M = 3.93, SD = 0.992) and achieved a moderately high reliability (α = .78).

Table 1.

Item Sets for Dependent Variables.

Item		Reliability
Favorability	I am in favor of this change	.88
	I think this is a positive change for the organization
	I am happy this change is coming to our organization
	This change is not a good thing
Perceived Trustworthiness	I trust the people who are introducing this change	.84
	I feel confident that the people introducing this change will keep everyone’s best interests in mind
	I think the people introducing this change are not so trustworthy
	The people introducing this change can be counted on to do what is right
Perceived Honesty	I think that people introducing this change are lying about the change	.78
	I am concerned that the people introducing this change are not being completely honest about it
	I am being told the whole truth about the change
	I think important information about the change is being withheld

We asked respondents to report their targets of communication with the question, “How likely are you to communicate with the following people after receiving the memo?” Participants’ perception regarding the primary communication targets was measured by asking the question, “How likely would it be that each of the following people would be the first one you would want to talk to about the change?” For both these questions, participants were provided a list of individuals whom they could potentially contact (see Tables 2 and 3), such as the person or department that sent the memo, their immediate supervisor, their peers at work, their nonwork friends or family members, the people responsible for the memo, colleagues considered to be allies. The response scale was anchored with “not very important target” and “very important target.”

Table 2.

Communication Targets.

	One sided			Two sided			Overall
Description of items	n	M	SD	n	M	SD	n	M	SD
The person/department that sent the memo	113	2.81	1.379	105	3.01	1.213	218	2.91	1.303
My immediate supervisor	112	4.21	1.067	105	4.27	0.858	217	4.24	0.969
My peer(s) at work	113	4.44	0.812	105	4.47	0.833	218	4.45	0.820
My personal nonwork friends/family	113	3.23	1.402	105	3.52	1.557	218	3.37	1.482
The person(s) responsible for this decision	113	2.75	1.672	105	2.71	1.426	218	2.73	1.555
People who were nearest to me when I first heard about the change	113	3.86	1.194	105	3.68	1.173	218	3.77	1.185
My boss’s boss	113	2.31	1.254	105	2.35	1.536	218	2.33	1.408
Friends/colleagues that work else with experience with this change	110	3.59	1.370	104	3.76	1.370	104	3.76	1.692
People in my organization whom I consider to be allies	113	3.58	1.230	104	3.65	1.121	104	3.62	1.177
People in my organization whom I consider to be powerful	113	3.27	1.232	105	3.06	1.027	218	3.17	1.140
People in my organization whom I consider likely to agree with me	113	3.80	1.010	105	3.61	0.915	218	3.71	0.968
People in my organization whom I consider to be experts	113	3.71	1.066	105	3.69	0.974	218	3.70	1.021
People in my organization whom I think will give a point of view different from my own	113	3.35	1.043	105	3.38	1.095	218	3.37	1.066
People in my organization whom I think will disagree with me about this change	113	2.74	1.301	105	2.64	1.066	218	2.69	1.192

Table 3.

Primary Targets of Communication.

	One sided			Two sided			Overall
Targets	n	M	SD	n	M	SD	n	M	SD
The person/department that sent the memo	111	2.63	1.452	105	2.75	1.568	216	2.69	1.507
My immediate supervisor	113	4.23	1.217	105	4.07	1.211	218	4.15	1.214
My peer(s) at work	111	4.09	1.005	105	4.10	1.189	216	4.09	1.096
My personal nonwork friends and family	112	2.70	1.413	105	2.93	1.595	217	2.81	1.505
Person(s) responsible for this decision	111	2.86	1.387	105	2.88	1.542	216	2.87	1.461
People who were nearest to me when I first heard about this change	112	3.50	1.315	104	3.29	1.492	216	3.40	1.404
My boss’s boss	112	2.19	1.219	105	2.40	1.668	217	2.29	1.454
Friends/colleagues that work elsewhere but have experienced this change	109	3.07	1.412	105	3.50	1.749	214	3.28	1.597
People in my organization whom I consider to be allies	112	3.31	1.201	105	3.33	1.284	217	3.32	1.239
People in my organization whom I consider to be powerful	112	3.21	1.126	105	3.05	1.368	217	3.13	1.249
People in my organization whom I consider likely to agree with me	112	3.49	1.170	105	3.51	1.309	217	3.50	1.236
People in my organization whom I consider to be experts	112	3.66	1.087	105	3.51	1.249	217	3.59	1.168
People in my organization with point of view different from my own	112	2.79	1.202	105	2.86	1.348	217	2.82	1.272
People in my organization whom I think disagree with me	112	2.29	1.128	105	2.33	1.298	217	2.31	1.211

Note. Highest means in each col are bolded.

Participants’ purpose for communication was assessed with the question, “To what extent would you be seeking each of the following in talking with people about this change?” The question included options such as clarification of information, gaining social support, gaining more information, expressing concerns, and sharing opinions. The response scale was anchored with “not very important goal” and “very important goal.” In order to compare these reported purposes of communication with the BNRM, we grouped them into sets that represented ACS, moderate Active Change (ACM), and Acceptance (ACC; see Table 4). We found modest to high reliabilities for these scales (ACS α = .850; ACM α = .775; ACC α = .590). We then used these sets as dependent variables in some of our analyses.

Table 4.

Purposes of Communication (Organized Into BNRM Categories) After receiving One- and Two-Sided Announcements of Change.

		One sided			Two sided			Overall
BNRM categories	Purposes	n	M	SD	n	M	SD	n	M	SD
Active Change Strong	Request to have this reconsidered/dropped	113	2.65	1.194	105	2.67	1.459	218	2.66	1.325
	Expressing my anger	113	2.18	1.174	105	2.37	1.533	218	2.27	1.359
	Suggesting an alternative to this change	112	2.94	1.232	105	3.05	1.396	217	2.99	1.312
	Making an argument that this is not the way to go	113	2.58	1.201	105	2.58	1.433	218	2.58	1.315
	Expressing my dissent	113	2.60	1.098	105	2.94	1.537	218	2.77	1.336
	Finding allies who would support my position	113	3.04	1.224	105	3.27	1.346	218	3.15	1.287
	Expressing excitement about this change	113	2.86	1.060	105	2.95	1.228	218	2.90	1.142
	Expressing sense that it will cause trouble	113	2.41	1.147	104	2.68	1.381	217	2.54	1.269
	Convincing others that this is a good idea	113	2.74	1.163	105	2.74	1.352	218	2.74	1.255
	Convince others that this is a bad idea	113	2.50	1.181	105	2.61	1.362	218	2.55	1.270
Active Change Moderate	Clarification of Information	113	4.43	0.865	105	4.55	0.843	218	4.49	0.855
	Gaining more details of what is happening	113	4.42	0.821	105	4.46	0.832	218	4.44	0.825
	Locate additional expertise	113	3.25	1.313	105	3.78	1.538	218	3.50	1.447
	To know what opinions others have	113	3.48	1.095	105	3.30	1.255	218	3.39	1.176
	To know what reactions others have	113	3.48	1.135	105	3.32	1.229	218	3.40	1.181
	Venting my complaints	113	2.84	1.214	105	3.05	1.553	218	2.94	1.388
	Sharing my opinions	113	3.58	0.998	104	3.53	1.114	217	3.56	1.053
	Sharing expertise	113	3.60	1.057	105	3.60	1.106	218	3.60	1.078
	Identify how many others are in agreement with me about the change	113	3.07	1.273	105	3.17	1.274	218	3.12	1.272
	Discovering what others’ opinions/reactions are to this change	113	3.19	1.177	105	3.38	1.180	218	3.28	1.180
	Discovering what others think about the implementers of this change	113	2.88	1.276	104	2.89	1.321	217	2.88	1.295
Acceptance	Gaining social support (to help ME feel better about it)	113	3.09	1.162	105	3.20	1.369	218	3.14	1.264
	Provide social support (to help OTHERS feel better about it)	113	3.10	1.009	105	3.21	1.412	218	3.15	1.218
	Detecting if others have strong negative or positive reaction to it	113	3.02	1.203	105	3.11	1.163	218	3.06	1.182
	Expressing my concerns	113	3.60	1.098	105	3.75	1.336	218	3.67	1.218

Note. BNRM = Bad News Response Model.

Sample

A total of 218 respondents completed the survey, 103 males and 115 females, and the average education level was completion of bachelor’s degree. Current roles included executive management (10%), middle management (12%), front-line supervisor (6%), administrative (11%), nonsupervisory/general employee (25%), specialists (17%), and other (19%). No significant differences on variables of interest were detected between the New Jersey and Kansas subsamples, so they were combined for all analyses.

We asked respondents to report their level of experience with change at their current job and in their career thus far. Our respondents reported an average of 3.85 on a 5-point scale (very few changes to many changes). We also asked our respondents to report how much experience they have in planning and introducing change. Our respondents reported mean of 3.11 on a 5-point scale (very little experience to a great deal of experience). These both indicate a moderate level of experience with change.

Results

The data were analyzed using analysis of variance (ANOVA) to detect differences between groups. Although some of our analyses could have been reported as t tests, we have some analyses with three groups. To enhance comparability, we used ANOVA for all analyses.

Message Sidedness and Favorability to Change

Our first research question asked whether initial reports of favorability toward the change would differ between stakeholders receiving the one-sided message and the two-sided message. An ANOVA was performed to determine if differences in favorability occurred. Analysis revealed a significant difference for favorability, F(1, 215) = 8.681, p = .004. Respondents who received the one-sided message were significantly more in favor of the change (M = 3.53, SD = 1.01) than those who received the two-sided message (M = 3.13, SD = 0.961).

Message Sidedness and Trust and Honesty

Hypothesis 1 predicted that stakeholders receiving two-sided messages would report higher assessments of implementers’ trustworthiness and honesty. To examine this hypothesis, ANOVAs were performed for each of these dependent variables. Analyses revealed no significant differences between the two groups for trust, F(1, 216) = 1.550, p = .214. The trust assessments of those who received the one-sided message (M = 3.27, SD = 0.956) did not differ significantly from the trust perceptions of those receiving the two-sided messages (M = 3.11, SD = 0.999). Furthermore, assessments of implementers’ honesty did not differ between the groups, F(1, 5216) = .344, p = .588. Those receiving the one-sided (M = 3.43, SD = 1.00) and two-sided (M = 3.35, SD = 0.980) change messages evaluated implementers’ honesty nearly the same. Therefore, Hypothesis 1 is not supported.

Risky Change and Favorability, Trust, and Honesty

Hypothesis 2 posited that perceived high-risk change would stimulate lower assessments of implementers’ honesty and trustworthiness as well as favorability than would low-risk change. ANOVAs were run to test this hypothesis. The results indicated a significant main effect for trust, F(1, 216) = 8.59, p = .004, for honesty, F(1, 216) = 12.48, p = .001, and favorability, F(1, 215) = 8.681, p = .004, suggesting that respondents’ perceptions of trustworthiness, honesty, and favorability differed based on whether they received a message in the context of high- or low-risk change. Planned comparisons revealed that trust perceptions of those who received the high-risk change (M = 3.01, SD = 0.962) were significantly lower than those receiving the low-risk change (M = 3.31, SD = 0.961). Similarly, it was revealed that honesty perceptions of those who received the high-risk (M = 3.17, SD = 1.025) change message were also significantly lower than those receiving the low-risk change message (M = 3.63, SD = 0.899). Favorability was lower in the high-risk change (M = 3.13) than the low-risk change conditions (M = 3.53). These results are consistent with our predictions. Thus, Hypothesis 2 was supported.

Interaction Effects: Message Sidedness and Risk on Favorability, Trust, and Honesty

Hypothesis 3 posited that there would be an interaction between risk of change and sidedness of messages on the dependent variables assessing implementers’ trustworthiness, honesty, and favorability. More specifically, we predicted that in high-risk change, two-sided messages would yield higher assessments of trust, honesty, and favorability, whereas in low-risk change, one-sided messages would yield higher assessments of trustworthiness, honesty, and favorability. After testing this hypothesis with an ANOVA, we found no evidence of an interaction between sidedness and perceived risk in accounting for assessments of trust, honesty, or favorability. In the high-risk conditions, we saw no significant differences between one- and two-sided message recipients’ assessments of trust, F(1, 111) = .073, p = .788, honesty, F(1, 111) = .426, p = .515, or favorability, F(1, 110) = .014, p = .907. This was true for the low-risk change scenarios for trust, F(1, 103) = 1.585, p = .211, honesty, F(1, 103) = 1.967, p = .164, and favorability, F(1, 103) = 1.156, p = .285, as well (see Table 5 for means). Given these results, Hypothesis 3 is not supported.

Table 5.

Trust, Honesty, and Favorability in Context of Risk and Sidedness.

	Mean values
	Favorability	Honesty	Trust
One-sided message recipients
High risk	3.12	3.15	3.07
Low risk	3.63	3.76	3.27
Two-sided message recipients
High risk	3.14	3.19	2.59
Low risk	3.42	3.51	3.51

Communication Targets and Purposes of Communication

Two of our research questions concerned the initial communication targets and purposes that stakeholders have following a change announcement.

Communication Targets

The respondents’ top five communication targets overall (see Table 2) are (a) peers at work, (b) immediate supervisor, (c) people nearest to me when I first heard about the change, (d) friends/colleagues that work elsewhere with experience with this change, and (e) people in my organization whom I consider likely to agree with me.

Primary Target

Overall, stakeholders’ primary targets (see Table 3) for communication are (a) immediate supervisor, (b) peers at work, (c) people in the organization considered to be experts in this area, (d) people in the organization likely to agree with me about the change, and (e) people who are nearest to me when I first heard.

Purpose of Communication

The respondents’ top five communication purposes (see Table 4) for talking to their targets, regardless of risk, are (a) clarification of information (ACM), (b) gaining more details of what is happening (ACM), (c) expressing concerns (ACC), (d) sharing opinions (ACM), and (e) sharing expertise (ACM). Overall, these subjects reported that they would seek communication for the purposes associated with ACM (M = 3.51; SD = 0.758) and less so for purposes associated with ACC (M = 3.26, SD = 0.876) or ACS (M = 2.72, SD = 0.978).

Risk and Purposes of Communication

We sought to understand the effects of risk on recipients’ ratings of the importance of purposes for communication subsequent to change announcements. Hypothesis 4 proposed that recipients of the high-risk change announcements rate the importance of purposes for their communication that reflect ACS higher than did those receiving the low-risk change announcements. After an ANOVA analysis, we found that recipients of the high-risk change announcement placed more importance of communication associated with ACS than did low-risk message recipients, but were more likely to report higher importance of purposes of communication for all three sets of purposes (ACS, F[1, 214] = 5.263, p = .023), (ACM, F[1, 212]= 6.706, p = .010), (ACC, F[1, 214] = 15.138, p < .000) as compared with those in the low-risk condition (see Table 6 for a complete list of means). Thus, Hypothesis 4 is supported.

Table 6.

Purposes of Communication During High- and Low-Risk Change.

	Strong active change	Moderate active change	Acceptance
Overall	2.72 (SD = 0.978)	3.51 (SD = 0.758)	3.26 (SD = 0.876)
High risk	2.86 (SD = 1.142)	3.64 (SD = 0.866)	3.48 (SD = 0.927)
Low risk	2.56 (SD = 0.73)	3.37 (SD = 0.593)	3.03 (SD = 0.759)

Risk and Message Sidedness Effects on Targets and Purposes of Communication

We also sought to understand the effects of message sidedness, in the context of high- and low-risk change, on the selection of communication targets and communication purposes (Research Question 4). A series of ANOVAs were run to test for significant differences in these contexts. In the context of low-risk change, no differences were found between recipients of one-sided and two-sided messages for targets or primary targets.

For targets in general, in the context of high-risk change, respondents receiving the two-sided announcement (M = 4.25, SD = 1.175) were more likely, F(1, 111) = 7.218, p = .008, to seek out the target of “personal nonwork family and friends” than those receiving the one-sided announcement (M = 3.60, SD = 1.355). In addition, respondents receiving the high-risk change announcement and a two-sided message (M = 4.15, SD = 1.669) reported the intention to talk to “friend and colleagues that work elsewhere that have had experience with this change” more, F(1, 111) = 5.534, p = .020, than respondents receiving the one-sided message (M = 3.51, SD = 1.167), as a target of communication.

For primary targets of communication, in a high-risk change, people receiving the two-sided message (M = 3.91, SD = 1.342) were more likely, F(1, 108) = 7.407, p = .008, to communicate with “friends or colleagues that work elsewhere with experience with the change” than respondents receiving the one-sided message (M = 3.05, SD = 1.914).

Purposes of Communication

We were also interested in understanding if sidedness, in the context of high- and low-risk change, affects the purpose of communication. The communication purposes included talking for ACS, for ACM, and for ACC. An interaction was run to test for significant differences in these contexts. For both contexts, high- and low-risk change, no differences were found between recipients of one-sided and two-sided messages for communication purpose (ACS, F[1, 212] = .007, p = .932; ACM, F[1, 212] = .363, p = .547; ACC, F[1, 214] = .418, p = .519), thus suggesting that in the context of risk, sidedness did not affect communication purpose. For a full list of the means see Table 7.

Table 7.

Purposes of Communication During High- and Low-Risk Change After Receipt of One- or Two-Sided Messages.

	Strong active change	Moderate active change	Acceptance
One-sided message recipients
High risk	2.79 (SD = 0.841)	3.57 (SD = 0.711)	3.38 (SD = 0.747)
Low risk	2.50 (SD = 0.785)	3.36 (SD = 0.651)	3.00 (SD = 0.711)
Two-sided message recipients
High risk	2.94 (SD = 1.413)	3.71 (SD =1.02)	3.59 (SD = 1.09)
Low risk	2.62 (SD = 0.6717)	3.37 (SD = 0.533)	3.05 (SD = 0.798)

In examining individual differences of specific purposes, we found one case within the context of high-risk change, where recipients of the two-sided and one-sided message differed in ratings of the importance of purposes of communication. Recipients of the high-risk change announcement and the two-sided message (M = 3.94, SD = 1.63) reported more importance in “locating additional expertise about the change” than respondents receiving the one-sided message (M = 3.30, SD = 1.36), F(2, 162) = 4.28, p = .015.

Discussion

Reporting bad news to stakeholders is part of the organizational change territory. Implementers of change have to unfortunately, at times, announce bad news. They must, at times, announce a change that will likely be viewed as risky or threatening to some, if not all, stakeholders. In previous literature about delivery of bad news, we have learned that people are reluctant to do so and that their default tendency is to withhold or downplay negative information in favor of highlighting positive spin. This study sought to investigate this context of message effects as a way to add to our knowledge of communication during organizational change at a very practical level. Furthermore, this study helps to account for message effects of sidedness in a very distinct context, different from others studied in the literature to date.

Message Sidedness and Risk Effects on Favorability, Trust, and Honesty

Our first set of questions in this study dealt with the wisdom of this strategy in the contexts of high- and low-risk change. Given the potential for positive reactions to early honesty about downsides of a change perceived to be risky, including increases in the credibility of implementers, preparing stakeholders to grapple with downsides, and avoidance of unpleasant surprises, we argued that it might be wiser to be transparent in initial communication about upcoming change. However, we questioned the wisdom of such transparency about possible downsides in the context of less risky change. In those cases, we suggested, it might be wiser to say less about downsides rather than raise concerns that were less likely to be hugely problematic.

Without considering risk, we found that one-sided messages were superior to two-sided nonrefutational messages in terms of producing favorable responses and that there was no difference between two-sided nonrefutational messages and one-sided messages in terms of recipients’ assessments of trustworthiness and honesty of implementers. In the context of risky change, there was no difference in the use of one- or two-sided nonrefutational messages in producing higher levels of trust and honesty assessments.

Our results mirror O’Keefe’s (2002) conclusions that two-sided nonrefutational messages are significantly less persuasive than one-sided messages. Our results are dissimilar to the findings in the advertising context where nonrefutational two-sided messages were observed to increase credibility of the speaker. In our study, we were not able to produce enhanced credibility assessments with the nonrefutational two-sided messages.

However, riskiness of change does appear to matter in determining the relative effects of the sidedness of the message in terms of favorability, trust, and honesty assessments. Although we did not find an interaction effect between message sidedness and perceived riskiness of change as we predicted, our data do suggest that both one- and two-sided nonrefutational messages do well in terms of honesty and trust assessments as well as favorability in low-risk situations. We found that in high-risk situations neither one- or two-sided nonrefutational messages do as well in terms of these outcomes as they do in perceived low-risk situations. The means for all three dependent variables in high-risk situations for both message types are lower than they are for the low-risk situations. Essentially, perceived high-risk situations are more challenging in terms of the messaging strategies for announcements.

We are not surprised that our two-sided nonrefutational messages were not as persuasive in general. We were also not surprised that high-risk situations lessened the assessments of trust, honesty, and favorability. However, we are surprised that the credibility (trust and honesty) of implementers was not assessed more positively in the two-sided conditions given the findings in the advertising literature. There are a number of ways to account for these findings.

First, it may be that our two-sided nonrefutational messages were not entirely “balanced” with positive and negative information. It may be that we over-did the important negative information such that the credibility and persuasion effects achieved in some advertising contexts were not found here. Eisend (2006) argues that both the importance of negative information and the amount of negative information will affect the assessments of credibility of the advertisement. We may have overemphasized the negative information to a point of getting the negative credibility assessments.

It is also possible that our two-sided message provoked a different sort of skepticism than what O’Keefe (1999) argued would work to improve credibility in nonrefutational two-sided advertising messages. O’Keefe argued that consumers expect to get a one-sided (all positive) message from advertisers (as one would expect in the announcement of change in a workplace) and that the presence of an admission of the opposing position might enhance the speaker’s credibility. What may have happened in this context is that people with experience with change introductions may have expected to get an all-positive message that was essentially concealing the worst news. It is possible that they thought the two-sided nonrefutational message was still hiding something worse. So, rather than feeling pleased that the implementer revealed true downsides, they may have been thinking that the change would be even worse than the implementer revealed. This may have to do with the difference in sample populations between the advertising research and our study. The typical undergraduate student may have been impressed with the admission of downsides in product advertising enough to increase assessments of credibility of the advertiser. However, the skepticism of the working adults in our sample may have led to a heightened concern about the change and the implementer in the two-sided conditions.

Third, we may have failed to get the enhanced effects of credibility in the two-sided nonrefutational conditions because this is not a situation of choice. In the previous research on consumer advertising, respondents are in a choice scenario where they are being asked if they like a product and/or intend to purchase the product. They have maximal control in such a circumstance. However, our scenario was not a choice context—which is a novel context for examining the effects of sidedness. These change announcements were informational previews rather than a context where the audience had control. In such a context, especially in the high-risk scenario, we might expect to see different dynamics. Although one-sided and two-sided nonrefutational messages worked about evenly well (above the midpoint on the scale for all three dependent variables) in the low-risk condition, the high-risk condition produced less successful outcomes for the implementer. This may be a circumstance in which refutation is necessary. In the context of high risk for the audience, where stakes such as job security and high stress are in play, it may be necessary to directly confront the downsides of the change rather than merely be willing to acknowledge that pain is coming. This is likely particularly true since this nonchoice context places more accountability on the implementer for causing the pain.

Sidedness, Risk, and Effects on Communication Targets and Purposes of Communication

The literature on change announcements, as well as persuasion literature concerning message effects, has yet to examine the subsequent communication behaviors of recipients of messages. Although our method did not permit us to capture reports of actual behavior, we were able to capture behavioral intentions, as well as ratings of the relative importance of various communication purposes of these participants, as indicative of how individuals might plan to respond. We sought to understand how the different levels of risk and message sidedness might prompt recipients to look to different targets for communication about the change and/or see different purposes as more or less critical.

Our results suggest that, in general, recipients tend to seek convenient, familiar, and agreeable targets. Similarly, primary targets are those who are convenient, familiar, agreeable, and expert. The importance of the purpose of their communication, in terms of the BNRM, suggest that in the context of risky change, recipients are likely to more highly value communication associated with ACS than recipients in the low-risk context. We also found that those in the high-risk scenario reported an increased importance of communication for purposes associated with both ACM and ACC.

It appears that risky change prompts higher importance of all communication purposes for recipients. They view communication as increasingly important to try to improve or alter the situation (ACS and ACM); however, they also view increasing importance in coping with the situation without attempting to influence outcomes and progression of the change (ACC). In the low-risk situation, recipients see less importance for all of these purposes of communication, although all three categories are still moderately important.

Message sidedness effects focus on various targets only in the high-risk scenarios. In this context, respondents receiving the two-sided announcements were more likely to seek out friends and family, particularly those with some experience with the change. In light of the data about sided messages in the context of change in general, these findings suggest that the two-sided message may have raised enough concerns that the recipients felt the need to seek a strong network tie in order to sort out the situation and/or decide how best to act.

Sidedness effects on the importance of various purposes for communication were modest. When examining the three sets of communication purposes—ACS, ACM, and ACC—it appears that sidedness did not have different effects on importance ratings for purposes of communication for situations of high and low risk. We observed a significant difference between two-sided and one-sided message recipients in only one individual case during high-risk change—“locating additional expertise about the change.” Those in the two-sided condition tended indicate this was a significantly more important purpose of their subsequent communication behavior than those in the one-sided condition. Overall, given these results, we think it unlikely that sidedness of change announcements, relative to perceived risk of change, makes an important difference in how recipients differently value communication subsequent to change.

Applied Conclusions

There are a few lessons that can be derived from these data regarding the practical question of whether negative aspects of a change program should be addressed in initial announcements. First, it appears to be equally successful to use a two-sided nonrefutational or one-sided message during either high-risk or low-risk change. Neither sort of announcement strategy held a significant advantage over the other in producing more favorable responses or higher assessments of implementer trustworthiness or honesty. The advantages of selecting the two-sided nonrefutational message may only be realized in a delayed effect after the message recipients actually begin to confront some of the pain involved in the change. If the realistic preview literature is applicable in this context, it is at that point that benefits of the preview are likely to be realized. Implementers should not expect to see immediate positive reactions from previews of negatives. However, initial favorability toward the change was not harmed by these previews during perceived high-risk change.

Implementers should also consider that unlike medical contexts, perhaps, where messengers’ equivocation about negatives may be viewed as kind, in organizational contexts (especially where the messenger is held accountable for the pain about to be endured), such equivocation might spark angry reactions or ridicule. In any case, it could cause long-term erosion of trust. Given that our data suggest no difference between one- and two-sided nonrefutational messages in initial reactions, and the previous research in realistic previews, it is likely best to opt to be more forthcoming up front.

Second, in the case of high-risk change, where stakes are very high and the circumstances perceived as potentially very threatening for the stakeholders who receive change announcements, implementers should be mindful that although neither one- or two-sided nonrefutational messages are consistently better, neither do particularly well in this setting. Initial strong reactions to announcements of change, and the context of lowered trust and honesty assessments driven by the context of perceived high risk, combine to make this a potentially problematic situation for implementers. Our data are not clear in terms of a recommended announcement strategy, but the likelihood is higher that a two-sided refutational message is necessary in these circumstances.

Regarding the subsequent communication of stakeholders after receiving messages, recipients of all change announcements will likely seek convenient, familiar, and agreeable targets for communication. They are also likely to see importance in engaging in communication for purposes both of accepting as well as challenging or altering change. In cases of high risk they will see added necessity in questioning or improving the situation. These results bolster those found in the health context in which the BNRM (Sweeny & Shepperd, 2009) has been developed. It is important for implementers to realize that forewarning of potential downsides or pain related to the change does not appear to affect these patterns.

Limitations

This study is limited by the sample achieved and the degree of realism associated with the scenarios we provided to subjects. Our sample may have had unusual characteristics from a general working population to which we had hoped to extend the findings of this study. The demographic data would suggest that we achieved a demographically diverse sample in terms of the variables we assessed, but there could be bias if the individuals participating in this study were more attitudinally disposed against or for change in general or more or less skeptical of management than the general population. There is no reason to believe that our convenience sample would be more or less inclined toward change than a general population. Furthermore, our sample was more diverse than the typical sample achieved in studies of organizational change—wherein a set of respondents from an organization’s population are assessed.

Our scenarios were constructed around two specific types of change—a technology change and a restructuring. The same results may not have been obtained for other sorts of policy, procedural, cultural, or production changes.

Future Directions

It will be important to explore the role that two-sided refutational messages can play in both low- and high-risk change situations. Our data would suggest that these types of messages could be necessary in high-risk change where all positive messages (one-sided) and those that raise but do not refute negatives (two-sided) are not very predictive of strong outcomes. A part of that research will need to include exploration of different refutational approaches in this context. It might involve, for example, suggesting that the counterarguments are exaggerated, inaccurate, or misapplied as they might be in another context. One strategy might be to refute the counterarguments through promoting the offsetting advantages of the change (e.g., indicating that the change will be difficult but that the company will gain in market share, profits, long-term job security for all employees). Some of these refutational strategies may be successful in raising trust, honesty, and favorability ratings, but some may make the implementer sound defensive.

In other future research, it will be important to examine the delayed effects on trust and honesty over time. As we have suggested above, stakeholders may only recognize and reward the honesty of implementers when they confront the negative aspects of the change themselves. It may be at that point that they realize the benefits of having been forewarned and may feel better prepared. As a result they may then raise their assessment of implementers’ credibility. The effects of the realistic preview appear not to be immediate given our data. Therefore, future research should not only conduct longitudinal field studies to determine if there is a lag effect but should also examine these constructs in ways that increase examination of individuals’ involvement and risk of consequences in change.

Footnotes

Appendix

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.

Author Biographies

Laurie K. Lewis is a professor in the Rutgers School of Communication and Information. She teaches and conducts research in areas of organizational change, nonprofit organizations, and interorganizational collaboration. She recently published a book titled, Organizational Change: Creating Change Through Strategic Communication (Wiley, 2011).

Nicole Laster is currently a Senior Social Scientist at CGI Group, Inc. She assists military leaders in cultivating a holistic understanding of their operating environments by conducting population-centric, socio-cultural research. She spent twelve years in academia prior to joining the U.S. Army as a civilian where she deployed to Afghanistan in support of Operation Enduring Freedom XII. She lives and works in San Antonio, Texas.

Vaibhavi Kulkarni is a doctoral candidate at the Rutgers School of Communication and Information. Her research interests include the examination of communication during organizational change processes.

References

Allen

(1998). Comparing the persuasive effectiveness of one- and two-sided messages. In Allen

Priess

R. W.

(Eds.), Persuasion: Advances through meta-analysis (pp. 87-98). Cresskill, NJ: Hampton.

Bartunek

J. M.

Moch

(1987). First order, second order, and third order change and organization development interventions: A cognitive approach. Journal of Applied Behavioral Science, 23, 483-500.

Bavelas

Black

Chovil

Mullet

(1990). Equivocal communication. Newbury Park, CA: Sage.

Bordia

Hobman

Jones

Gallois

Callan

(2004). Uncertainty during organizational change: Types, consequences, and management strategies. Journal of Business and Psychology, 18, 507-532.

Breaugh

J. A.

(1983). Realistic job previews: A critical appraisal and future research directions. Academy of Management Review, 8, 612-619.

Brett

Gustello

S. J.

Aderman

(1982). The effect of realistic job preview on expectancy and voluntary versus involuntary turnover. Journal of Psychology, 111, 101-108.

Brinthaupt

(2004). Providing a realistic course preview to students. Teaching of Psychology, 31, 104-106.

Chebat

Filiatrault

Laroche

Watson

(1988). Compensatory effects of cognitive characteristics of source, the message, and the receiver upon attitude change. Journal of Psychology, 122, 609-621.

Chebat

Picard

(1988). Receivers’ self-acceptance and the effectiveness of two-sided messages. Journal of Social Psychology, 128, 353-362.

10.

Del Vento

Bavelas

Healing

MacLean

Kirk

(2009). An experimental investigation of the dilemma of delivering bad news. Patient Education and Counseling, 77, 443-449.

11.

Demarest

D. S.

Hooke

J. F.

Erickson

M. T.

(1984). Preoperative intervention for the reduction of anxiety in pediatric surgery patients. Children’s Health Care, 12, 179-184.

12.

Dibble

J. L.

Levine

T. R.

(2010). Breaking good and bad news: Direction of the MUM effect and senders’ cognitive representations of news valence. Communication Research, 37, 703-722.

13.

Dugoni

B. L.

Ilgen

D. R.

(1981). Realistic job previews and the adjustment of new employees. Academy of Management Journal, 24, 579-591.

14.

Eisend

(2006). Two-sided advertising: A meta-analysis. International Journal of Research in Marketing, 23, 187-198.

15.

Faller

Grabarek

Ortega

R. M.

(2010). Commitment to child welfare work: What predicts leaving and staying? Children & Youth Services Review, 32, 840-846.

16.

Folkman

Lazarus

R. S.

(1980). An analysis of coping in a middle-aged community sample. Journal of Health and Social Behavior, 21, 219-239.

17.

Griffith

T. L.

Northcraft

G. B.

(1996). Cognitive elements in the implementation of new technology: Can less information provide more benefits? MIS Quarterly, 20, 99-109.

18.

Kline

S. L.

Simunich

Weber

(2009). The use of equivocal messages in responding to corporate challenges. Journal of Applied Communication Research, 37(1), 40-58.

19.

Kramer

(1999). Motivation to reduce uncertainty: Reconceptualizing uncertainty reduction theory. Management Communication Quarterly, 13, 305-316.

20.

Kramer

Dougherty

D. S.

Pierce

T. A.

(2004). Managing uncertainty during a corporate acquisition: A longitudinal study of communication during an airline acquisition. Human Communication Research, 30(1), 71-101.

21.

Laster

N. M.

(2008). Communicating multiple change: Understanding the impact of change messages on stakeholder perceptions. Dissertation Abstracts International. (UMI No. AAT 3342339)

22.

Lewis

L. K.

(2011). Organizational change: Creating change through strategic communication. Chichester, England: Wiley-Blackwell.

23.

Lewis

L. K.

Schmisseur

A. M.

Stephens

K. K.

Weir

K. E.

(2006). Advice on communicating during organizational change: The content of popular press books. Journal of Business Communication, 43(2), 1-25.

24.

Lewis

L. K.

Seibold

D. R.

(1996). Communication during intraorganizational innovation adoption: Predicting users’ behavioral coping responses to innovations in organizations. Communication Monographs, 63, 131-157.

25.

Lines

Selart

Espedal

Johansen

S. T.

(2005). The production of trust during organizational change. Journal of Change Management, 5, 221-245.

26.

McGlone

M. S.

Batchelor

J. A.

(2003). Looking out for number one: Euphemism and face. Journal of Communication, 53, 251-264.

27.

McGuire

(1961). Persistence of the resistance to persuasion induced by various types of prior belief defenses. Journal of Abnormal and Social Psychology, 64, 241-248.

28.

Miller

K. I.

Monge

P. R.

(1985). Social information and employee anxiety about organizational change. Human Communication Research, 11, 365-386.

29.

O’Keefe

D. J.

(1999). How to handle opposing arguments in persuasion messages: A meta-analytic review of the effects of one-sided and two-sided messages. Communication Yearbook, 22, 209-249.

30.

O’Keefe

D. J.

(2002). Persuasion: Theory and research (2nd ed.). Thousand Oaks, CA: Sage.

31.

Petty

Cacioppo

(1986). Communication persuasion: Classic and contemporary approaches. Dubuque, IA: Brown.

32.

Phillips

J. M.

(1998). Effects of realistic job previews on multiple organizational outcomes: A meta-analysis. Academy of Management Journal, 41, 673-690.

33.

Pratt

C. B.

(2004). Crafting key messages and talking points: Or grounding them in what research tells us. Public Relations Quarterly, 49(3), 15-20.

34.

Richardson

McBey

McKenna

(2008). Integrating realistic job previews and realistic living conditions previews: Realistic recruitment for internationally mobile knowledge workers. Personnel Review, 37, 490-508.

35.

Rynes

S. L.

(1990). Recruitment, job choice, and post-hire consequences: A call for research directions. In Dundee

M. D.

Hugh

(Eds.), Handbook of industrial and organizational psychology (2nd ed., pp. 399-444). Palo Alto, CA: Consulting Psychologists Press.

36.

Schoorman

F. D.

Mayer

R. C.

Davis

J. H.

(2007). An integrative model of organizational trust: Past, present, and future. Academy of Management Review, 32, 344-354.

37.

Schweiger

D. M.

DeNisi

A. S.

(1991). Communication with employees following a merger: A longitudinal field experiment. Academy of Management Journal, 34, 110-135.

38.

Smeltzer

L. R.

(1991). An analysis of strategies for announcing organization-wide change. Group and Organization Studies, 16(1), 5-24.

39.

Smeltzer

L. R.

(1995). Organization-wide change: Planning for an effective announcement. Journal of General Management, 20, 31-43.

40.

Smeltzer

L. R.

Zener

M. F.

(1992). Development of a model for announcing major layoffs. Group and Organization Management, 17, 445-472.

41.

Smeltzer

L. R.

Zener

M. F.

(1994). Minimizing the negative effect of employee layoffs through effective announcements. Employee Counseling Today, 6(4), 3-9.

42.

Smeltzer

L. R.

Zener

M. F.

(1995). Organization-wide change: Planning for an effective announcement. Journal of General Management, 20(3), 31-43.

43.

Sweeny

Shepperd

(2009). Responding to negative health events: A test of the bad news response model. Psychology and Health, 24, 895-907.

44.

Thomas

G. F.

Zolin

Hartman

J. L.

(2009). The central role of communication in developing trust and its effect on employee involvement. Journal of Business Communication, 46, 287-310.

45.

Wanous

J. P.

(1980). Organizational entry: Recruitment, selection, and socialization of newcomers. Reading, MA: Addison-Wesley.