Personal Outcomes and Moral Responsibility as Motives for News Transmission

Abstract

In two experiments we compared contrasting findings on bad news transmission likelihood between literature on rumors and the MUM-effect in order to contribute to the development of a more general theory of news transmission. We argued that several contextual differences account for the contrasting findings between these research conditions. We predicted that fate similarity and fate uncertainty (both present in many rumor contexts and absent in most MUM-contexts) enhance the anticipated personal outcomes of bad news transmission for communicators and hence increase bad news transmission. Supporting our argument, we found that fate uncertainty and fate similarity each increased the likelihood of bad news transmission up to the level of good news transmission. Furthermore, these effects were mediated by communicators’ anticipated personal outcomes of transmission. In addition, Experiment 2 demonstrated anticipated personal outcomes to be only an important motive for news transmission decisions in superficial relationships; for close relationships, experienced moral responsibility appeared to be the paramount motive for transmission.

Keywords

good and bad news transmission decisions personal outcomes moral responsibility

An important question in communication is whether or not we are willing to communicate good and bad news to others. Two contrasting patterns of research findings prevail in the literature, one showing that bad news is communicated less often than good news (experimental research on the MUM-effect), the other indicating that bad news is diffused more often than good news (field research on rumors). Both lines of research have been performed in different settings. The first aim of the present study is to integrate findings from these two research traditions in order to take a first step to the development of a more general theory of news transmission. We examine whether findings from rumor research may shed some light on which contextual factors affect the reluctance to transmit bad news in MUM-settings. More specifically, we examine whether contextual differences such as fate (dis)similarity, fate (un)certainty, and interpersonal closeness can account for the contrasting findings between the two research traditions. We argue that these three factors may affect the anticipated personal outcomes of transmission for communicators and hence transmission likelihood, especially for bad news. In addition, we examine how people balance their anticipated personal outcomes against another motive for news transmission decisions, namely, their experienced moral responsibility to transmit the news (Weenig, Groenenboom, & Wilke, 2001).

Contrasting Findings Concerning Bad News Transmission

Tesser and Rosen (1975; Rosen & Tesser, 1970) have demonstrated that news that implies personal negative consequences for the target person, such as “You will receive electrical shocks,” is transmitted considerably less often to the person it concerns than good news. They named this the “MUM-effect”: the tendency to keep mum about unpleasant messages.

In contrast to this predominantly experimental research on the MUM-effect, results from mostly field-research on rumors has shown that the majority of rumors spread is “black,” often containing unverified but plausible information from an unofficial source about a negative event or consequence, such as potential layoffs in an organization, or the possible death of a well-known person (e.g., Kapferer, 1990; Rosnow, 1991). Often, rumors emerge in situations of general anxiety or uncertainty concerning an important issue, such as a major reorganization (Bordia & Rosnow, 1998; DiFonzo, Bordia, & Rosnow, 1994; Rosnow, 1991). We posit that “black” rumors can also be considered as a kind of bad news. After all, bad news is news with a general negative evaluative meaning to both communicators and targets because of the potential negative events or consequences involved. From that perspective, findings from rumor research indicate that at least in some cases bad news is transmitted more often than good news.

In sum, the MUM literature has shown that bad news transmission is less likely than good news transmission, whereas findings from the rumor literature indicate the opposite. Until now, this difference in findings has hardly been addressed. In the present paper we examine how findings from rumor research can shed some light on how contextual factors may affect the MUM-effect. By comparing findings from both lines of research, we aim to contribute to the development of a more general theory of news transmission than proposed until now in the two research settings. We argue that qualitatively differential weighing processes take place for decisions to transmit good versus bad news. For decisions of good news transmission, we posit that a quick, superficial motivational scan is sufficient for communicators. After all, good news transmission usually has positive consequences for the target and is not potentially harmful for communicators; it may even improve their relationship with the target. Thus, good news transmission may well result in positive personal outcomes for both communicator and target, and is not likely to result in negative personal outcomes. By contrast, bad news transmission often has negative consequences, not only for the target but also for communicators themselves. In Western societies, communicators may fear that transmitting bad news will harm their face (cf. Goffman, 1967; Wilson, Aleman, & Leatham, 1998) and their self-presentation to the target by being associated with the bad news (Uysal & Öner-Özkan, 2007). Bad news transmission even may deteriorate their relationship with the target. A classic illustration of the potentially problematic nature of communicating bad news is the beheading of the messenger of bad news. In all cases, bad news transmission implies negative outcomes for communicators.

We posit that, at least in Western societies, actors may have good reason not to transmit bad news because of the negative consequences that are involved for both themselves and the target and that an actor only will transmit if he or she can find an additional motivation to transmit. In line with this proposition, Dibble and Levine (2010) found that response delays were more than 1.5 times as long for bad news than for good news. Such need for an additional motivation is less necessary in case of good news as there are no negative consequences involved. Experimental research by Weenig et al. (2001) indicates that experienced moral responsibility, that is, the extent to which a communicator feels morally obliged to transmit the news, can serve as such additional motivation. Weenig and colleagues found that bad news transmission becomes more likely when communicators feel more morally responsible to transmit.

While Weenig et al. (2001) focused on moral responsibility, the present research focuses on another possible additional motive for transmission, namely, on potential benefits for communicators that may compensate for the high personal costs that are usually involved in bad news transmission. Findings from the MUM-literature suggest that people often refrain from bad news transmission because they expect high personal costs for themselves, such as unpleasant mood changes, feelings of guilt, strong negative emotional reactions from the target (cf. Tesser & Rosen, 1975), negative face implications, and a less positive self-presentation (Uysal & Öner-Özkan, 2007). By contrast, Kapferer (1990) contends that black rumor transmission can be beneficial for communicators: Transmission may result in a reduction of communicators’ uncertainty or anxiety if they obtain additional information in return. Furthermore, bad news transmission enables communicators to share their emotions with the target. Therefore, based on Kapferer, we predict that the likelihood of bad news transmission will increase when communicators anticipate more benefits from transmission, as this increases the anticipated cost-benefit ratio (i.e., anticipated personal outcomes) of transmission for communicators. This possibility has never been explicitly tested.

In the present research we aim to contribute to the existing literature on good and bad news transmission by examining two contextual factors that we expect to affect communicators’ anticipated personal outcomes. More specifically, we examine whether fate uncertainty and fate similarity between target and communicator each enhances the anticipated personal benefits of bad news transmission for communicators and, in this way, each increases the likelihood of bad news transmission. Both contextual factors are present in many rumor research contexts but absent in most MUM experiments.

Fate Uncertainty

An important difference between MUM and rumor research is that in MUM-settings the news always implies some certain (bad) fate for the target person (e.g., “You should call home about some very bad news,” or “The research assistant has some bad news for you”). In contrast, in many rumor settings such as the reorganization context mentioned before, the (bad) fate implied by the news is often uncertain for both target and communicator. For example, the rumor that “The upcoming reorganization will result in many layoffs, perhaps even one or more entire departments must be closed” implies that it is uncertain who will be affected by the bad fate and that almost everyone’s fate in the organization is uncertain. Conveying information in the latter situation may bring communicators the important benefit of obtaining some useful additional information themselves in return from the target (cf. Kapferer, 1990). This additional information is potentially beneficial as it may reduce communicators’ own uncertainty or anxiety (e.g., DiFonzo & Bordia, 2002; Festinger, 1954; Schachter, 1959). By contrast, in most MUM-contexts there presumably is no such uncertainty of fate and hence the transmission of bad news typically does not yield many benefits for communicators. Instead, in MUM-contexts transmission brings several psychological costs for communicators such as face threats or fear for a strong negative emotional reaction from the target (Tesser & Rosen, 1975). Accordingly, it can be expected that communicators will perceive bad news transmission as more beneficial for themselves (i.e., will be expected to result in higher personal outcomes) when it is uncertain who will be affected by the bad fate (like in many rumors), compared to the situation in which the news certainly implies a bad fate for the target (such as in most MUM-contexts).

Hence we propose the following hypothesis:

Hypothesis 1a: Communicators will anticipate more benefits and thus higher personal outcomes of bad news transmission and as a result will be more likely to transmit bad news in case of uncertain fate compared to certain bad fate.

In addition, we predict that fate uncertainty will have less or no impact on the likelihood of good news transmission, given that in case of good news communicators will not anticipate many costs and therefore do not need an additional motive to transmit. That is, we propose:

Hypothesis 1b: Fate uncertainty will have less impact on good news transmission than on bad news transmission.

Fate Similarity

In many rumor settings, the news contains information that can be consequential not only for the target but also for other people, including the communicator. For instance, rumors such as “Probably most employees from our department will be discharged” potentially involve a similar fate for most if not all members of the relevant community, including the communicator. In contrast, such a similarity of fate between communicator and target is absent in most of the research on the MUM-effect: In most of these studies, the news itself (e.g., “You will receive electrical shocks”) only has negative consequences for the target person and not for the communicator.

Important exceptions are two studies by Tesser and colleagues (Johnson, Conlee, & Tesser, 1974; Tesser & Rosen, 1972), which showed that fate similarity increases the likelihood of bad news transmission. However, in both studies the underlying motives for news transmission remained unclear, possibly because of the authors’ exclusive focus on the personal costs of bad news transmission. We believe that an increase in bad news transmission in case of similar fate will most likely be caused by a major personal benefit that shared fate brings for communicators: the opportunity to share one’s own emotions with the target. After all, the well-known sayings “A sorrow shared is a sorrow halved,” and “Company in distress makes sorrow less” indicate that sharing one’s negative emotions with another person (such as the target) may offer a considerable amount of comfort to a communicator. In addition, in case of similar fate, bad news transmission may also offer communicators some anxiety reduction because they may receive some additional reassuring information from the target in return. Hence, we propose the following hypothesis:

Hypothesis 2a: Communicators will anticipate higher benefits and therefore higher personal outcomes of bad news transmission and as a result will be more likely to transmit bad news in case of similar fate than with dissimilar fate.

In addition, based on the same argument as for Hypothesis 1b, we propose as follows:

Hypothesis 2b: Fate similarity will have less impact on good news transmission than on bad news transmission.

Overview of Experiments

We report two experiments. In Experiment 1 we tested whether fate uncertainty and fate similarity enhance the anticipated personal outcomes of bad news transmission for communicators and in this way increase the likelihood of bad news transmission.

In Experiment 2 we aimed to replicate and extend the findings of Experiment 1. More specifically, in Experiment 2 we examined how communicators balance the more hedonistic consideration of anticipated personal outcomes against their more social consideration of experienced moral responsibility to transmit the news, which is another known motive for bad news transmission (Weenig et al., 2001). We also examined whether this balancing process differs depending on whether communicators have a close versus superficial relationship with the target.

Experiment 1

Method

Participants and Design

One hundred and sixty predominantly White undergraduates from a West-European university participated in the study (45 male, 115 female, mean age = 21.64). Participants were randomly allocated to one of the eight conditions of a 2 (certain vs. uncertain fate for target) × 2 (similar vs. dissimilar fate for communicator and target) × 2 (good vs. bad news for target) factorial design.

Procedure

Participants were recruited from various classes. They participated in the experiment either during the break or after class in another room. Participation was rewarded with a snack. Each participant read one of the eight versions of the scenario. They then completed a questionnaire containing the dependent variables and the manipulation checks. Participants remained seated until all other participants present in the room had finished the questionnaire. Next, the experiment was explained to the participants and they were paid and thanked for their cooperation.

Stimulus Materials

Participants were given one of eight versions of a scenario and were asked to imagine being in the situation described in the scenario. The scenario described a situation that was very familiar to each participant, namely, obtaining a minimum number of course credits in order to avoid being expelled from the curriculum. Participants were asked to imagine they had received some information concerning an important exam that other classmates had not received:

You are a freshman at the university. This means that at the end of the first year’s course, you must have obtained at least 30 course credits. If not, you will be expelled from the curriculum. The end of the first year’s course is approaching, and you have not yet obtained your 30 credits. A week ago, you had your final exam of the year. In addition to the lectures, you also attended tutorials for this final part of the curriculum. Your tutor-class consists of 12 students, some of whom you only know by face, while you know several others much better, some even very well. You are not the only one who has not yet obtained 30 course credits. On the contrary, no one from your tutor-class has obtained the 30 credits yet. This also holds for several other tutor-classes. Therefore, everybody is apprehensively waiting for the results of the final exam, through which each student may reach the 30 credits norm. Failing the exam implies for you and for every other student in your class that the study at this university must be ended, whereas passing implies for every student in your class that the study can be continued. One afternoon, you leave the faculty building early because you have an appointment with your general practitioner (GP). While you are leaving, you run across your tutor. Although you know the official results will not be made public until next week, you ask him whether he has any information about the results of the exam. He replies that . . .

[hereafter follows the text for the four bad news conditions (the specific bad news condition is indicated between brackets); in the four good news conditions, the word “failed” was replaced by “passed”; all other texts were identical to the four respective bad news conditions]

. . . everybody from your tutor-class has failed, including you. (certain, similar fate)

. . . someone from your tutor-class has failed, but he does not remember who it is, although he is pretty sure that it is not you. (uncertain, dissimilar fate)

. . . from your tutor-class only Sacha has failed. (certain, dissimilar fate)

. . . in one of the tutor-classes everybody has failed, but he does not remember which class it is, yours or another. (uncertain, similar fate)

Immediately after you have thanked him for this information, you leave the faculty building, because otherwise you may be too late for your appointment with the GP. You enter the waiting room shortly before the end of the surgery hours for that day. The waiting room is empty. After a few minutes, Sacha enters the waiting room. Sacha belongs to your tutor-class, but you do not know Sacha much better than by face. With Sacha, the GP’s assistant also enters the room with the announcement that Sacha and you are the last patients for that day, but that the two of you have to wait for at least half an hour because of an emergency. After she has left, no one else is in the waiting room except you and Sacha. Sacha takes a seat next to you, and starts to talk. Soon the conversation shifts to the last exam. You get the impression that Sacha does not know the results yet.

Dependent Measures

Manipulation checks

Similarity of fate was checked by means of the question, “Which student(s) from your tutor-group did [fail/pass] the exam, according to your tutor?” (1 = only Sacha, 2 = one student, possibly Sacha, but not you, 3 = the entire tutor-group, including you and Sacha). The perceived valence of the news about the results of the final exam for Sacha was checked by means of the question, “How positive or negative do you consider the news for Sacha?” (1 = very negative to 7 = very positive).¹ Uncertainty of fate was checked as follows: “What can you conclude from the information you received from your tutor for Sacha?” (1 = that Sacha had definitely [failed/passed] the exam, 2 = that Sacha had possibly [failed/passed] the exam).

Transmission likelihood

Transmission likelihood was the main dependent variable. It was measured by means of the question, “Would you tell Sacha the news you have heard about the results of the exam while the both of you are waiting for the practitioner?” (1 = certainly not to 7 = certainly would). Participants were also asked to write down as comprehensively as possible what exactly they would tell Sacha if they would tell at least something about the exam.

Anticipated costs

The anticipated costs of transmitting the news were measured by six items that were all derived from Tesser and Rosen’s (1975) research and that were identical to the items used by Weenig et al. (2001): “To what extent would you, if you would give the message to Sacha, be afraid of an emotional reaction from Sacha?” “. . . an angry reaction from Sacha?” “. . . a negative change in your own mood?” “. . . becoming less respected by Sacha?” “. . . becoming less liked by Sacha for telling the news?” and “. . . feel guilty towards Sacha?” (1 = not at all to 7 = very much). Cronbach’s alpha of these six items was 0.69. An examination of interitem correlations did not reveal items that had problematic associations with other items and removing items did not improve the alpha. A factor-analysis further provided no indications of multidimensionality, whereupon it was decided to compute for each participant an anticipated costs score by averaging the responses to these six items, with higher scores indicating higher anticipated costs of transmission.

Anticipated benefits

The anticipated benefits of transmitting the news were measured by means of eight items. Six of these items were derived from Weenig et al. (2001) and based on Tesser and Rosen’s research (1975): “To what extent would you feel relieved, if you would tell Sacha the message?” “. . . expect a positive change in your own mood?” “. . . hope to become more respected by Sacha?” “. . . hope that Sacha would be grateful to you?” “. . . hope for a favor from Sacha in return?” and “. . . hope for being more liked by Sacha?” (1 = not at all to 7 = very much). Two additional items measured the two benefits posited by Kapferer (1990): “To what extent would you hope to get some additional information from Sacha about the results of the final exam?” and “To what extent would you wish to share your emotions concerning the results of the exam with Sacha?” (1 = not at all to 7 = very much). Cronbach’s alpha of the eight items was .72. None of these items had problematic interitem correlations and removing items did not improve the alpha. A factor-analysis showed some evidence of two-dimensionality, but separation of the items into two scales did not significantly improve the alpha or change the results.² Anticipated benefits were therefore computed by averaging the responses to the eight items, with higher scores indicating higher anticipated benefits of transmission.

Results

Manipulation Checks

The manipulation of fate similarity was understood correctly by 98.1% of all participants and the manipulation of fate uncertainty by 96.3% of the participants. One participant had misunderstood both manipulations and was therefore excluded from all further analyses. A 2 × 2 × 2 ANOVA with perceived valence of the message for the target person (Sacha) as dependent variable showed a strong significant main effect of valence, F(1, 151) = 414.87, p < .001, η² = .95: On average, and as intended, participants perceived the bad news as much more negative for the target person (M = 2.51, SD = 1.51) than the good news (M = 5.58, SD = 1.45).³ Thus the manipulation of all three experimental factors was successful.

Tests of Hypotheses

The influence of fate similarity, fate uncertainty, and news valence on transmission likelihood was tested by means of a 2 × 2 × 2 ANOVA. The cell means are presented in Table 1. The analysis showed significant main effects for all three independent variables. Overall, transmission likelihood was higher for good news than for bad news, F(1, 151) = 15.09, p < .001, η² = .09; in case of similar fate between communicator and target than in case of dissimilar fate, F(1, 151) = 16.70, p < .001, η² = .10, and if it was uncertain rather than certain to whom the news applied, F(1, 151) = 19.26, p < .001, η² = .11. The main effects of fate uncertainty and fate similarity were each qualified by the valence of the news: The Valence × Fate Uncertainty interaction was significant, F(1, 151) = 12.84, p < .001, η² = .08, and the Valence × Fate Similarity interaction was marginally significant as well, F(1, 151) = 3.50, p = .063, η² = .02. Post hoc tests showed that for bad news the differences in transmission likelihood between conditions were significant (ps < .01) and quite large, as predicted in Hypotheses 1a and 2a. For good news, there were no significant differences in transmission likelihood between experimental conditions on transmission likelihood, which is consistent with Hypotheses 1b and 2b. This means that all our hypotheses were supported. Fate uncertainty and fate similarity each significantly increased bad news transmission but did not affect good news transmission.⁴

Table 1.

Transmission Likelihood as a Function of News Valence, Fate Uncertainty and Fate Similarity (Experiment 1).

	Dissimilar fate		Similar fate		Total
News Valence/fate uncertainty	M	SD	M	SD	M	SD
Bad news	4.21	(2.31)	5.65	(1.59)	4.94	(2.10)
Certain	2.84_a	(1.98)	5.00_b	(1.56)	3.95	(2.06)
Uncertain	5.50b_c	(1.82)	6.30_c	(1.38)	5.90	(1.65)
Good news	5.60	(1.61)	6.15	(1.19)	5.88	(1.44)
Certain	5.35_b^a	(1.66)	6.20_c	(1.01)	5.78	(1.42)
Uncertain	5.85_bc	(1.57)	6.10_c	(1.37)	5.98	(1.46)
Total (bad & good)	4.91	(2.10)	5.90	(1.42)	5.41	(1.85)
Certain	4.13	(2.20)	5.60	(1.43)	4.87	(1.98)
Uncertain	5.68	(1.69)	6.20	(1.36)	5.94	(1.55)

Note: Transmission likelihood was measured on a scale ranging from 1 (certainly not) to 7 (certainly would). Cell means that do not share a common subscript differ significantly at p < .05.

The difference with uncertain, similar bad news (M = 6.30) is significant at p = .057.

Finally, the three-way interaction of Fate Similarity × Fate uncertainty × Valence did not reach significance, F(1, 151) < 1. However, when we concentrate on the bad news conditions, we can see that the likelihood of bad news transmission was lowest in case of certain dissimilar fate—the typical MUM context—and highest in case of uncertain similar fate, the condition that most closely resembles a “black rumor” context (see Table 1). In fact, the likelihood of news transmission in this most “black rumor” resembling condition (bad uncertain similar fate) was even higher (p = .057) than in the “standard” (i.e., certain, dissimilar) good news condition of the MUM studies!

Mediation Analyses

Next, we tested whether the effects of fate similarity and fate uncertainty on transmission decisions would be mediated by the (net) anticipated personal outcomes of transmission (benefits minus costs). Since we found that fate similarity and fate uncertainty only had significant effects on bad news transmission, we performed these analyses on bad news transmission only. The zero-order correlations between transmission decisions and personal outcomes are presented in Table 2.

Table 2.

Correlations Between News Transmission and Potential Mediators (Experiment 1).

	Personal outcomes	Costs	Benefits
News transmission	.60***	−.40***	.38***
Transmission bad news	.77***	−.45***	.57***
Transmission good news	.14	−.10	.01

***

p ≤ .001

For each respondent, a net total anticipated personal outcomes score was computed by subtracting the mean score on all anticipated personal costs items from the mean score on all anticipated personal benefit items, with higher scores indicating higher anticipated personal outcomes. We first performed a 2-way ANOVA with fate uncertainty and fate similarity as independent variables and respondents’ personal outcomes scores as the dependent variable (cf. Baron & Kenny, 1986). The analysis showed that, as predicted in Hypotheses 1a and 2a, respondents anticipated higher personal outcomes from bad news transmission in case of similar fate between communicator and target (M = 0.46, SD = 1.04) than in case of dissimilar fate (M = −0.11, SD = 1.69), F(1, 76) = 5.01, p = .028, η² = .06, and also in case of uncertain fate (M = 0.87, SD = 1.17) as compared to certain fate (M = −0.52, SD = 1.33), F(1, 76) = 29.80, p < .001, η² = .28.⁵

We subsequently tested by means of an ANCOVA whether differences in bad news transmission likelihood as a function of fate similarity and fate uncertainty were due to differences in anticipated personal outcomes. This analysis showed that the main effect of fate uncertainty on bad news transmission was significantly reduced (Sobel Z = 4.39, p < .001) and became nonsignificant after controlling for anticipated personal outcomes. The main effect of fate similarity on bad news transmission was also significantly reduced (Sobel Z = 2.14, p < .05) after controlling for anticipated personal outcomes. These results corroborate Hypotheses 1a and 2a: Bad news transmission was higher in case of similar fate compared to dissimilar fate because actors expected higher personal outcomes from transmission in case of similar fate. Similarly, bad news transmission was higher in case of uncertain fate compared to certain fate because people expected higher personal outcomes from transmission in case of certain fate.

Discussion

Overall, results of Experiment 1 are consistent with our hypotheses: We predicted and found that bad news transmission is more likely in case of uncertain fate than in case of certain fate and that bad news transmission is also more likely in case of similar fate compared to dissimilar fate, both due to higher anticipated personal outcomes (Hypotheses 1a and 2a). Thus, in this experiment, people’s anticipated personal outcomes were found to be an important motive for bad news transmission. Good news transmission, by contrast, was not affected by fate uncertainty and similarity (Hypothesis 1b and 2b), also as predicted.

Two remarks should be made about Experiment 1. The first concerns the manipulation of fate similarity. In order to resemble differences between MUM and rumor research settings as closely as possible, fate similarity was operationalized as similarity with the entire group to which target person and communicator belonged. A drawback of this approach is that it is not clear whether the higher transmission likelihood that we found in case of fate similarity is the result of fate similarity at the dyadic level (i.e., between communicator and target per se), the result of the fate similarity at the group level (i.e., between target and group), or both. This issue was addressed in an auxiliary experiment (Weenig & Knuttel, 2001) in which the fate similarity factor of Experiment 1 was split into two factors: dyadic similarity (present vs. absent) and group similarity (present vs. absent). Of these two factors, only the effect of dyadic fate similarity on transmission was significant: On a 7-point transmission likelihood scale that was identical to the one used in Experiment 1, participants rated (good and bad) news transmission as more likely when interpersonal fate similarity was present (M = 5.00, SD = 1.76) than when it was absent (M = 3.89, SD = 2.04), F(1, 160) = 14.48, p < .001, η² = .08. This means that the findings of Experiment 1 regarding the impact of fate similarity on bad news transmission should be attributed to the similarity of fate between communicator and target at the dyadic level and not to the similarity of fate with the group.

The second remark about Experiment 1concerns the relationship between communicator (i.e., the respondent) and target person. In all conditions of Experiment 1 this relationship was described as superficial. However, news transmission is by no means limited to superficial acquaintances. On the contrary, rumors appear to be more common in tightly knit communities than in communities in which people do not have that many close relationships with each another (Kapferer, 1990). This is illustrated by research from Weenig et al. (2001) who found that bad news transmission is more likely between close friends than between superficial acquaintances. In this study it was further shown that this higher transmission likelihood of bad news primarily occurs because close friends feel more morally responsible to transmit the bad news than superficial acquaintances. In Experiment 2 we examined how this finding relates to the importance of anticipated personal outcomes as a motive for bad news transmission that we found in Experiment 1.

Experiment 2

The main purpose of Experiment 2 was to extend findings from Experiment 1 by addressing the question whether the relationship a communicator has with the target person (close vs. superficial) affects the level of anticipated personal outcomes and subsequently the impact of anticipated personal outcomes on transmission decisions. The broader aim of Experiment 2 was to take a step forward in the development of a model of good and bad news transmission decisions. We argued that especially for bad news transmission, two broad motives would be important: anticipated personal outcomes (Experiment 1) and experienced moral responsibility (Weenig et al, 2001). The first motive represents the more personal, hedonistic considerations of news transmission, while the second motive represents a more social, normative consideration, in other words, what other persons think that communicators ought to do. In Experiment 2 we examined the relative importance of each of these motives for transmission decisions. We explored the possibility that these motives differentially facilitate bad news transmission, depending on the closeness of the relationship between communicator and target. To date, the issue of how people balance moral considerations and anticipated personal outcomes has received little attention in both MUM and rumor research. Experiment 2 contributes to previous research on news transmission by addressing the question of how communicators balance the motive of anticipated personal outcomes of transmission against the motive of experienced moral responsibility for transmission decisions.

Clark (1984) and Fiske (1992) make a distinction between “exchange” and “communal” relationships that is relevant here. Exchange relationships are based on the exchange of costs and benefits, which should be in balance over time (cf. exchange theory, Homans, 1958, 1974; Thibaut & Kelley, 1959). Relationships between superficial acquaintances (such as in Experiment 1) will mostly fall into this category, at least in Western societies. Close friends, on the other hand, belong more to the category of communal relationships, which are much less based on the exchange of costs and benefits, but more on mutual involvement. In close relationships, people keep less track of their input than in exchange relationships and behave more altruistically (and less in their self-interest), sharing everything they have with each other (Clark, 1984; Fiske, 1992). Applied to the domain of transmission decisions, this suggests that in Western societies transmission decisions between close friends will be based less on hedonistic costs-benefits considerations than between superficial acquaintances. This suggestion was tested in Experiment 2.

As in Experiment 1, in Experiment 2 we varied fate similarity (similar vs. dissimilar) and the valence of the news (good vs. bad). The consequences of the news for the target were described as certain in all experimental conditions. Furthermore, in extension of Experiment 1, in Experiment 2 we varied the closeness of the relationship between communicator and target (friends vs. acquaintances). As in Experiment 1, we predicted that fate similarity would affect the anticipated personal outcomes component (with a higher anticipated personal outcomes in case of similar fate compared to dissimilar fate) and in turn would have a positive effect on the transmission of bad news (Hypothesis 2a). Furthermore, in extension of Experiment 1 we predicted as follows:

Hypothesis 3a: The closeness of the interpersonal relationship between communicator and target would affect the level of experienced moral responsibility (with higher experienced moral responsibility in case of a close relationship compared to a superficial relationship) and for that reason would also have a positive effect on bad news transmission.

By contrast, we did not assume that the closeness of the interpersonal relationship would affect the level of anticipated personal outcomes. That is, we did not anticipate that the level of anticipated personal outcomes would be different for close relationships compared to superficial relationships. In addition, in line with Experiment 1, we expected that good news transmission would not require much additional motivation. Hence, we expected smaller effects of fate similarity (see Hypothesis 2b) and relationship closeness on good news transmission than on bad news transmission:

Finally, we predicted as follows:

Hypothesis 4: The anticipated personal outcomes would affect bad news transmission decisions more in superficial relationships than in close relationships.

Method

Participants and Design

One hundred and fifty-two predominantly White undergraduates from a West-European university participated in the study (32 male, 120 female, mean age = 20.04). Participants were randomly allocated to one of the eight conditions of a 2 (close vs. superficial relationship) × 2 (similar vs. dissimilar fate between communicator and target) × 2 (good vs. bad news for the target) factorial design.

Procedure

The procedure was similar to that in Experiment 1.

Stimulus Materials

Participants were given one of eight slightly adapted versions of the scenario used in Experiment 3 of the Weenig et al. study (2001). They were asked how they would respond to the situation described. The participant was asked to imagine that he or she was working as a higher employee in a middle-large organization and would be confronted with the following situation:

(The text hereafter is from the condition with bad news, dissimilar fate, and a superficial relationship with target. Adaptations in the texts that were made for the other conditions are represented between brackets.)

In the report of the weekly meeting between the director and the chief of the personnel department, you read that the quality of the work of a colleague [and you (similar fate)] has been discussed. The colleague is from another department, but you know the person superficially, in fact not much better than by face [very well, you even are close friends (close relationship)]. The director is very dissatisfied [satisfied (good news)] about your colleague [and you (similar fate)] and your colleague [and you] will be dismissed [promoted to a higher rank (good news)]. You also read who should inform [you and (similar fate)] your colleague of the dismissal [promotion (good news)], but you know that that person was taken to hospital this morning and that it is unlikely that he will be able to inform your colleague [and you (similar fate)] in the near future. Hence, it is not clear who should inform your colleague [and you (similar fate)]. You remain seated for a while to absorb this information, but then you realize you have to leave the office at once and catch the 15h30 bus. However, you miss the bus, and have to wait about half an hour for the next one. After a few minutes, the colleague about whose dismissal [promotion (good news)] you just read, and who you [know superficially (superficial relationship)/are closely befriended with (close relationship)], joins you at the bus stop. While you both wait for the bus, the colleague starts to talk, and soon the two of you talk about work. From the conversation you understand that your colleague is still ignorant about the dismissal [promotion (good news)].

Dependent Measures

Manipulation checks

Fate similarity was checked by means of the question, “To whom did the news about the [dismissal/promotion] concern: to your colleague (1), to you (2), or to both you and your colleague (3)?” The manipulation of relationship closeness was checked by means of the question, “How close did you image your relationship with the colleague involved?” (1 = very superficial to 7 = very close). The perceived valence of the news was measured with the question, “How good or bad do you think the news is for your colleague?” (1 = very bad to 7 = very good).

Transmission likelihood

Transmission likelihood, the main dependent variable, was measured as follows: “Would you tell your colleague the news while the two of you are waiting for the bus?” (1 = certainly not to 7 = certainly would).

Anticipated personal outcomes

The anticipated personal costs and benefits of transmission were measured identical to that in Experiment 1. Cronbach’s alpha was .68 for both scales and could not be improved by removing any items. As in Experiment 1, the net anticipated personal outcomes for each participant was computed by subtracting the mean score on all anticipated cost items from the mean score on all anticipated benefit items, with higher scores indicating higher anticipated personal outcomes.

Moral responsibility

Experienced moral responsibility to transmit the news was assessed by the one-item measure employed by Weenig et al. (2001): “To what extent would you feel morally responsible to transmit the news to your colleague?” (1 = not at all to 7 = very much).

Results and Discussion

Manipulation checks

The manipulation check of fate similarity was answered correctly by all but 1 participant. This participant was excluded from the analyses because responses of this participant on other manipulation checks also indicated a misunderstanding of the situation described.

A 2 (relationship) × 2 (similarity) × 2 (valence) ANOVA with perceived relationship closeness with the target person as the dependent variable, showed only the expected main effect of relationship closeness, F(1, 143) = 741.67, p < .001, η² = .84: When the target person was described as a close friend, participants imagined their relationship with this person to be much closer (M = 5.60, SD = 0.79) than when the target person was described as superficial acquaintance (M = 1.87, SD = 0.88).

Finally, a 2 × 2 × 2 ANOVA with perceived valence of the news as the dependent variable, demonstrated only a main effect of valence, F(1, 143) = 3477.98, p < .001, η² = .96: As intended, participants perceived the news about the dismissal as much more negative (M = 1.23, SD = 0.45) than the news about the promotion (M = 6.63, SD = 0.65). Thus all manipulations were successful.

News Transmission

A 2 × 2 × 2 ANOVA on news transmission showed significant main effects for relationship closeness, F(1, 143) = 8.64, p < .01, η² = .06, and fate similarity, F(1, 143) = 17.92, p < .001, η² = .11. Participants were more likely to transmit the news when the target person was a close friend rather than a superficial acquaintance and in case of similar rather than dissimilar fate (see Table 3). Both main effects were qualified by valence: F(1, 143) = 4.22, p < .05, η² = .03 for the Valence × Relationship interaction, and F(1, 143) = 10.05, p < .01, η² = .07 for the Valence × Fate similarity interaction. For bad news, the differences in transmission were quite large and significant: Bad news transmission was more likely if the target was a friend than if the target was an acquaintance, and in case of similar fate compared to dissimilar fate. For good news, there were no significant differences between experimental conditions. Thus, the results of Experiment 2 provide support for Hypotheses 2ab and 3ab, as good news was transmitted irrespective of contextual circumstances, whereas bad news transmission did depend on the two contextual factors we investigated.

Table 3.

Transmission Likelihood as a Function of News Valence, Relationship Closeness, and Fate Similarity (Experiment 2).

	Dissimilar fate		Similar fate		Total
News valence/relationship	M	SD	M	SD	M	SD
Bad news	3.11	(2.13)	5.34	(1.38)	4.24	(2.10)
Superficial relation	1.89_a	(1.29)	5.11_bc	(1.59)	3.50	(2.17)
Close relation	4.39_b	(2.12)	5.58_c	(1.12)	5.00	(1.76)
Good news	4.45	(2.21)	4.76	(1.82)	4.61	(2.02)
Superficial relation	4.21_b	(2.30)	4.74_bc	(1.88)	4.47	(2.09)
Close relation	4.68_bc	(2.16)	4.79_bc	(1.81)	4.74	(1.97)
Total (bad & good)	3.79	(2.26)	5.05	(1.63)	4.42	(2.06)
Total superficial	3.05	(2.18)	4.92	(1.73)	3.99	(2.17)
Total close	4.54	(2.12)	5.18	(1.54)	4.87	(1.86)

Note: Transmission likelihood was measured on a scale ranging from 1 (certainly not) to 7 (certainly would). Cell means that do not share a common subscript differ significantly at p < .05.

We also found a significant Relationship × Similarity interaction on news transmission, F(1, 143) = 4.22, p < .05, η² = .03. For friends, there was no significant difference in overall (i.e., good and bad) news transmission between similar and dissimilar fate, whereas it was significant (p < .01) and quite large for superficial acquaintances. In other words, superficial acquaintances transmitted the news significantly more often in case of similar fate compared to dissimilar fate, whereas for friends this difference was much smaller. This 2-way interaction was significant for bad news, F(1, 71) = 7.78, p < .01, η² = .10, but not for good news, F(1, 72) < 1. This can be regarded as an indication for the correctness of Hypothesis 4 that higher personal outcomes are mainly important for superficial relationships, a topic that we will address further in the next paragraph. Finally, if we look at the cell means of Table 3, we can see that bad news transmission to a friend in case of similar fate (M = 5.58) was significantly more likely than good news transmission to an acquaintance in case of dissimilar fate (M = 4.21), which is the “standard” good news condition in the MUM-studies.

Mediation Analyses

Next, we investigated whether the effects of relationship closeness and fate similarity on bad news transmission were mediated by people’s anticipated outcomes of bad news transmission (i.e., benefits minus costs), and their experienced moral responsibility to transmit the news. These mediation analyses were only performed on bad news transmission because fate similarity and relationship closeness only affected bad news transmission, as predicted. The zero-order correlations between news transmission and the potential mediators are presented in Table 4.

Table 4.

Correlations Between News Transmission and Potential Mediators (Experiment 2).

	Moral responsibility	Personal outcomes
News transmission	.70***	.38***
Transmission bad news	.65***	.48***
Transmission good news	.78***	.31**

p < .01. ***p < .001.

We first performed two two-way ANOVAs with relationship closeness and fate similarity as independent variables, and the potential mediators (i.e., anticipated personal outcomes and experienced moral responsibility) as the respective dependent variable. As predicted in Hypothesis 2a, and consistent with Experiment 1, participants anticipated higher personal outcomes from bad news transmission in case of similar fate (M = −.02, SD = 1.10) compared to dissimilar fate (M = −1.49, SD = 1.30), F(1, 71) = 26.95, p < .001, η² = .28. As assumed, we found that the relationship between communicator and target did not affect communicators’ anticipated personal outcomes of transmission, F(1,71) < 1, ns. This means that the level of anticipated personal outcomes of transmission was approximately the same regardless of whether the target was a close friend or a superficial acquaintance. The interaction of Fate similarity × Relationship was not significant either, F(1,71) < 1, which means that fate similarity enhanced the anticipated personal outcomes of transmission about as much in close as in superficial relationships, which also corroborates our assumption.

Furthermore, as predicted in Hypothesis 3a, participants regarded themselves as more morally responsible to transmit the bad news when the target person was a close friend (M = 5.68, SD = 1.31) than when the target was a superficial acquaintance (M = 4.37, SD = 1.82), F(1, 71) = 15.02, p < .001, η² = .18 Participants also regarded themselves as more morally responsible to transmit the news in case of similar fate (M = 5.66, SD = 1.21) compared to dissimilar fate (M = 4.35, SD = 1.21), F(1, 71) = 14.68, p < .001, η² = .17. Finally, we did not find a significant Relationship × Fate similarity interaction on moral responsibility, F(1, 71) = 2.30, ns. This means that fate similarity increased experienced moral responsibility about as much in close relationships as in superficial relationships.

Next, we performed three ANCOVAs on bad news transmission in which we entered the two potential mediators (anticipated personal outcomes and experienced moral responsibility) and the interaction of both variables as separate covariates, and another ANCOVA in which we entered these three covariates simultaneously. The results from these analyses are summarized in Table 5. They showed that anticipated personal outcomes and experienced moral responsibility were each a significant covariate of bad news transmission likelihood, both when entered separately as well as when entered simultaneously. The main effects of relationship closeness and fate similarity on bad news transmission were each affected by at least one of these covariates, but the interaction effect of fate similarity and relationship closeness was not affected by them. More specifically, the main effect of relationship closeness on bad news transmission likelihood was significantly reduced after controlling for moral responsibility (Sobel test significant at p < .01) and did not depend on personal outcomes or the interaction between personal outcomes and moral responsibility. Thus the greater bad news transmission for friends than acquaintances could be explained by the fact that people felt more morally responsible to transmit the news in close compared to superficial relationships.

Table 5.

F-values of ANCOVAs with Relationship Closeness and Fate Similarity as Independent Variables, Moral Responsibility and Personal Outcomes as Covariates, and Bad News Transmission as Dependent Variable (Experiment 2), F(1, 70).

Potential mediator	Covariate	Relationship (R)	Similarity (S)	R × S
(ANOVA without covariates)	—	16.79***	36.92***	7.78**
Separate
Moral responsibility (MR)	17.99***	5.63*	19.37***	5.32*
Personal outcomes (PO)	6.62**	16.94***	16.19***	8.33**
MR × PO	0.01	16.31***	36.23***	7.08**
Simultaneous
Moral responsibility (MR)	17.85***
Personal outcomes (PO)	6.09*	6.00*	8.19**	6.18*
MR × PO	0.40

p ≤ .01. ***p ≤ .001.

Furthermore, the main effect of fate similarity on transmission likelihood was significantly reduced after controlling for personal outcomes and also after controlling for moral responsibility (both Sobel tests significant at p ≤ .01). Thus the greater news transmission in case of similar fate compared to dissimilar fate could be explained by the fact that people feel more morally responsible to transmit the news and also expect more positive outcomes in the case of similar compared to dissimilar fate.

The interaction term of both mediators (PO × MR) was not a significant covariate and therefore could not be a significant mediator. Further inspection of the data revealed that this last mentioned covariate interacted significantly with relationship closeness, thus violating one of the assumptions of the analysis: The regression of the covariate on transmission was significantly different (Z = 2.38, p < .05) for close relationships (β = .19, ns) compared to superficial relationships (β = −.36, p < .05).

Relative Importance of Experienced Moral Responsibility and Anticipated Personal Outcomes

The significant difference between the regression coefficients of PO × MR on transmission likelihood for close and superficial relationships described in the previous paragraph can be regarded as first indication of the correctness of our reasoning that the weighing of each motive would be different for superficial relationships compared to close relationships. We subsequently tested whether the relative importance of experienced moral responsibility and anticipated personal outcomes for bad news transmission decisions would vary for close versus superficial relationships. We predicted (Hypothesis 4) that anticipated personal outcomes would affect bad news transmission decisions more in superficial relationships than in close relationships.

First, we computed for each experimental subgroup the simple correlations between transmission likelihood and anticipated personal outcomes and experienced moral responsibility. Results showed that the correlation between anticipated personal outcomes and transmission likelihood was somewhat higher (Z = 1.48, p = .07, one-tailed) for superficial relationships (r = .48, p < .01) than for close relationships (r = .27, p < .05), which corroborates Hypothesis 4. By contrast, the correlation between moral responsibility and transmission likelihood was not significantly different for both groups (r = .64, p < .001 for superficial relationships vs. r = .74, p < .001 for close relationships, Z = 1.15, ns).

Next, we created two groups of anticipated personal outcomes (high vs. low), as well as two groups of experienced moral responsibility (high vs. low), both by means of median split. A three-way ANOVA with relationship closeness, anticipated personal outcomes, and experienced moral responsibility as independent variables and bad news transmission likelihood as the dependent variable yielded significant main effects for anticipated personal outcomes, F(1, 143) = 4.91, p < .05, η² = .03, and experienced moral responsibility, F(1, 143) = 60.08, p < .001, η² = .30: News transmission was more likely in case of high anticipated personal outcomes or high experienced moral responsibility than with low anticipated personal outcomes or low moral responsibility (see Table 6). The interaction effect of relationship closeness and anticipated personal outcomes was not significant, F(1, 143) = 1.78, ns, but the three-way interaction did reach significance⁶, F(1, 143) = 6.67, p = .01, η² = .05. For superficial relationships, the anticipated personal outcomes significantly increased news transmission when experienced moral responsibility was low and did not significantly affect transmission when experienced moral responsibility was high. By contrast, for close relationships, high anticipated personal outcomes neither significantly increased transmission likelihood when experienced moral responsibility was low, nor when experienced moral responsibility was high. These results corroborate Hypothesis 4, and demonstrate that the anticipated personal outcomes of news transmission only increase transmission decisions when communicators have a superficial relationship with the target and when experienced moral responsibility to transmit the news is low.

Table 6.

Bad News Transmission as a Function of Personal Outcomes, Moral Responsibility, and Relationship Closeness (Experiment 2).

	Personal outcomes
	Low			High			Total
Relationship closeness + Moral responsibility	M	SD	N	M	SD	N	M	SD	N
Superficial relationship	3.10	2.11	39	4.92	1.83	37	3.99	2.17	76
Low responsibility	2.52_a	1.73	31	4.11_bc	1.88	19	3.12	1.93	50
High responsibility	5.38_cd	2.00	8	5.78_d	1.35	18	5.42	1.84	26
Close relationship	4.75	1.80	36	4.97	1.94	39	4.87	1.86	75
Low responsibility	4.00_b	1.88	14	3.39_ab	1.75	18	3.66	1.81	32
High responsibility	5.23_cd	1.60	22	6.33_d	0.58	21	5.77	1.32	43
Total (superficial & close)	3.89	2.12	75	4.95	1.88	76	4.42	2.06	151
Total low responsibility	2.98	1.89	45	3.76	1.83	37	3.33	1.89	82
Total high responsibility	5.27	1.68	30	6.08	1.04	39	5.72	1.40	69

Note: Transmission likelihood was measured on a scale ranging from 1 (certainly not) to 7 (certainly would). Cell means that do not share a common subscript differ significantly at p < .05. Levine’s test significant at p = .001.

In sum, the results of Experiment 2 provide support for Hypotheses 2ab, 3ab, and 4. Most importantly, they offer support for our prediction that the higher anticipated personal outcomes that result from fate similarity mainly affect bad news transmission decisions in superficial relationships and not so much in close relationships.

General Discussion

The first aim of the present work was to integrate the contrasting findings from research on the MUM-effect and on rumors concerning the likelihood of good versus bad news transmission in order to take a first step in the development of a more general theory of news transmission. We examined whether two contextual differences between these two lines of research—fate (dis)similarity and fate (un)certainty—could account for the contrasting findings. We argued that fate similarity and fate uncertainty are present in many rumor settings and absent in most MUM-settings and that each aspect would make bad news transmission more beneficial to communicators than in a typical MUM-setting.

Main Findings and Implications

Consistent with our hypotheses, results showed that when it was uncertain to whom the consequences of the news would apply or when the consequences of the news were similar for communicator and target, the likelihood of bad news transmission increased up to the level of good news transmission. By contrast, fate uncertainty and similarity of fate did not affect good news transmission. Combined, fate uncertainty and similarity increased the likelihood of bad news transmission slightly beyond the transmission likelihood of certain good news. In other words, bad news transmission becomes even more attractive than good news transmission when it is uncertain who will be affected by the bad consequences implied by the news, and when communicator and target are faced with a similar fate.

As predicted, the positive effects of fate similarity and fate uncertainty on bad news transmission were at least in part mediated by the higher anticipated personal outcomes of transmission for the communicator (benefits minus costs) in case of fate similarity and fate uncertainty compared to fate dissimilarity and fate certainty. Thus the results provide an explanation for why so many rumors are black, as rumors occur most frequently (though not exclusively) in situations of general anxiety or uncertainty (Kapferer, 1990; Rosnow, 1991), and when there is a certain degree of fate similarity among group members. The present study shows that under such conditions the transmission of bad news offers higher personal outcomes to communicators than in a typical MUM context.

Experiment 2 demonstrated that within a context of certain consequences for the target, the positive effect of fate similarity on bad news transmission is largely restricted to superficial acquaintances. In more close relationships such as between close friends, fate similarity did not significantly enhance bad news transmission. Remarkably, we found in Experiment 2 that bad news transmission to a friend who faces the same bad fate as the communicator is significantly more likely than good news transmission in the standard good news condition of the MUM-experiments (i.e., between superficial acquaintances who do not share a similar fate). Thus it is probably the combination of the three factors (fate uncertainty, fate similarity, and more close relationships) that causes the more rapid diffusion of bad news in many rumor settings compared to bad news transmission in a typical MUM-setting.

The broader aim of Experiment 2 was to get a better understanding of the underlying motives for transmission decisions. In Experiment 1 we found that the increased likelihood of bad news transmission due to fate uncertainty and fate similarity was at least in part mediated by the increased personal outcomes that communicators anticipated from transmission. In Experiment 2 we replicated these findings for fate similarity. In addition, consistent with findings from Weenig et al. (2001), Experiment 2 showed that communicators also consider whether they are morally responsible in their decision to transmit the bad news. Experiment 2 showed that moral responsibility and personal outcomes both mediate bad news transmission decisions, but that the relative importance of these two motives differs for close relations (such as close friends) compared to superficial relationships. In close relationships, the anticipated personal outcomes of bad news transmission appear to be almost irrelevant for transmission decisions, which converges with Clark’s (1984) and Fiske’s (1992) distinction between exchange and communal relationships. Instead, in close relationships transmission decisions are predominantly based on experienced moral responsibility, which appears to be relatively high even in MUM-situations (i.e., in situations in which the news has certain bad consequences for the target only). By contrast, in superficial relationships the anticipated personal outcomes of transmission are important for bad news transmission decisions, but only when experienced moral responsibility is low. When percieved moral responsibility is high, the anticipated personal outcomes of transmission do not increase bad news transmission likelihood any further, just like in close relationships.

Thus the results of the present study indicate that high anticipated personal outcomes can compensate for the negative effect that a low experienced moral responsibility has on bad news transmission, but only in superficial relationships. In close relationships, low moral responsibility is relatively rare, but its negative effect on bad news transmission decisions cannot be compensated by high anticipated personal outcomes. The practical implication of these findings is that when there is no close relationship between communicator and target, such as in many professional settings (e.g., the physician–patient setting), bad news transmission can be enhanced by increasing the anticipated personal outcomes of transmission of professional communicators. This might be accomplished by training techniques that make negative personal outcomes less likely and/or positive personal outcomes more likely. For instance, such trainings techniques can help communicators to cope with emotional and angry reactions from the target and teach communicators to bring the news in a way that will be respected by the target (e.g., by bringing the news very politely; cf. Brown & Levinson, 1987).

Limitations of the Present Research

The fact that we decided to use vignettes in both experiments might be considered as a possible limitation of the present study. After all, vignette studies do not measure actual behavior and may suffer from social desirable answers to the questions posed. However, it stands to reason that in our case results from behavioral experiments would not have differed much from our vignette studies on news transmission, as Weenig et al. (2001) found similar results for their vignette studies in comparison with their behavioral study on news transmission. This suggests that in the present domain of transmission decisions vignette studies basically reveal similar processes as behavioral studies.

It might be argued that the manipulation of the valence of the news to some extent could have been affected by the severity of the news. After all, being expelled from the curriculum (Experiment 1), and losing one’s job (Experiment 2), could have been perceived as more serious than being allowed to continue the study or receiving a promotion to a higher rank. However, the means from the manipulation checks suggest that the good news was perceived as good as the bad news was perceived bad.

Another issue is that we measured experienced moral responsibility with a single item measure in the present study. One of the most evident psychometric shortcomings of one-item measures is that they cannot yield estimates of internal consistency (Wanous, Reichers, & Hudy, 1997). However, studies on other constructs (e.g., job satisfaction, self-esteem, attitudes, beliefs, etc.) that compared single-item to multiple-item measures demonstrate satisfactory correlations between the different measures (see, for example, Gardner, Cummings, Dunham, & Pierce, 1998; Wanous & Reichers, 1996). Furthermore, single item measures are considered useful when a holistic impression is informative (cf. Youngblut & Casper, 1993), which clearly is the case in our study. That is, the single item employed by Weenig et al. (2001) taps the construct of experienced moral responsibility unambiguously and the most directly. Moreover, the use of this item allows for a comparison between the findings of the present study and results from Weenig et al (2001).

A final limitation of the present study is that the results that we found only tell us that bad news transmission is correlated with moral responsibility and personal outcomes. They tell us nothing about the causal direction of these relationships as they are based on mediation analyses (cf. Sigall & Mills, 1998). Nevertheless, for now we are confident that experienced moral responsibility and anticipated personal outcomes can both be important motives for bad news transmission.

Directions for Future Research

We think that the two motives model for transmission decisions as we presented here has a high potential for further research. A first suggestion for such future research would be to establish the causal direction of the relationships that we found by experimentally inducing various levels of experienced moral responsibility and anticipated personal outcomes and measure their effects on good and bad news transmission decisions. Furthermore, the issue of an additional motivation that seems to be required for bad news transmission needs to be investigated in more detail. In the present research we assumed and found that transmission of bad news demands more additional motivation than good news. It could well be that this additional motivation in case of bad news transmission decisions involves more elaboration compared to good news transmission decisions. Future studies might investigate the correctness of this assumption. Dibble and Levine (2010) already found longer reaction times for bad news than for good news transmission decisions. Further evidence might be obtained through thought listing techniques. We would expect people to report more thoughts in case of bad news transmission decisions than for good news transmission decisions. We also would expect more face-related thoughts in case of bad news transmission.

Finally, it would be relevant to investigate whether the mediating role of personal outcomes and experienced moral responsibility on bad news transmission differs between more collectivistic cultures (such as Japan) as opposed to the more individualistic Western culture in which our studies were performed. It could well be that in more collectivistic cultures the impact of anticipated personal outcomes on transmission is weaker than in more individualistic cultures, and/or that the impact of experienced moral responsibility stronger.

Conclusion

In sum, the present research shows that the preference for good news over bad news transmission (i.e., the MUM effect) disappears and even slightly reverses under conditions of high fate uncertainty and fate similarity, and that this in part is due to higher perceived personal outcomes. As these conditions are present in many rumor situations and absent in MUM contexts, the present research provides an explanation for the inconsistent findings regarding good and bad news transmission likelihood between the two research traditions. Finally, our work indicates that anticipated personal outcomes are only an important motive for news transmission decisions in superficial relationships. For close relationships, experienced moral responsibility appears to be the paramount motive for transmission.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interests with respect to the authorship and/or publication of this article.

Funding

The authors received no financial support for the research and/or authorship of this article.

Notes

Author Biographies

Mieneke W.H. Weenig, PhD (Leiden University, 1991), is an assistant professor in the Department of Social and Organizational Psychology at Leiden University, The Netherlands. Her research interests include news transmission, information processing, communication networks, and persuasion.

Henk A.M. Wilke, PhD (Utrecht University, The Netherlands, 1968) was professor in experimental social psychology. He was connected with Dartmouth College (NH), Stellenbosch University (SA), University of California (SB), Groningen University (NL) and Leiden University (NL). His current interests involve social decision making and consultancy.

Emma ter Mors, PhD (Leiden University, 2009),) is currently a postdoctoral researcher at the Department of Social and Organizational Psychology, Leiden University, The Netherlands. Her research interests include selective exposure, information processing, and attitude formation.

References

Baron

R. M.

Kenny

D. A.

(1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173-1182.

Bordia

Rosnow

R. L.

(1998). Rumor rest stops on the information highway—Transmission patterns in a computer-mediated model. Human Communication Research, 25, 163-179.

Brown

Levinson

S. C.

(1987). Politeness: Some universals in language usage. Cambridge, UK: Cambridge University Press.

Clark

M. S.

(1984). Record keeping in two types of relationships. Journal of Personality and Social Psychology, 47, 549-557.

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. Advance online publication. http://crx.sagepub.com

DiFonzo

Bordia

(2002). Corporate rumor activity, belief and accuracy. Public Relations Review, 28, 1-19.

DiFonzo

Bordia

Rosnow

R. L.

(1994). Reining in rumors. Organizational Dynamics, 23, 47-62.

Festinger

(1954). A theory of social comparison processes. Human Relations, 7, 117-140.

Fiske

A. P.

(1992). The four elementary forms of sociality: Framework for a unified theory of social relations. Psychological Review, 99, 689-723.

10.

Gardner

D. G.

Cummings

L. L;

Dunham

R. B.

Pierce

J. L.

(1998). Single-item versus multiple-item measurement scales: An empirical comparison. Educational and Psychological Measurement, 58, 898-915.

11.

Goffman

(1967). Interaction ritual: Essays in face-to-face behavior. Chicago: Adrine.

12.

Homans

G. C.

(1958). Social behavior as exchange. American Journal of Sociology, 63, 597-606.

13.

Homans

G. C.

(1974). Social behavior: Its elementary forms (2nd ed.). New York, NY: Harcourt Brace Jovanovich.

14.

Johnson

R. E.

Conlee

M. C.

Tesser

(1974). Effects of similarity of fate on bad news transmission: A reexamination. Journal of Personality and Social Psychology, 29, 644-648.

15.

Kapferer

(1990). Rumors: Uses, interpretations, and images. New Brunswick, NJ: Transaction.

16.

Rosen

Tesser

(1970). On reluctance to communicate undesirable information: The MUM-effect. Sociometry, 33, 253-263.

17.

Rosnow

R. L.

(1991). Inside rumor. American Psychologist, 46, 484-496.

18.

Schachter

(1959). The psychology of affiliation. Stanford, CA: Stanford University Press.

19.

Sigall

Mills

(1998). Measures of independent variables and mediators are useful in social psychology experiments: But are they necessary? Personality and Social Psychology Review, 2, 218-226.

20.

Tesser

Rosen

(1972). Similarity of objective fate as a determinant of the reluctance to transmit unpleasant information: The MUM effect. Journal of Personality and Social Psychology, 23, 46-54.

21.

Tesser

Rosen

(1975). The reluctance to transmit bad news. In Berkowitz

(Ed.), Advances in experimental social psychology (Vol. 8, pp. 193-232). New York: Academic Press.

22.

Thibaut

J. W.

Kelley

H. H.

(1959). The social psychology of groups. New York, NY: Wiley.

23.

Uysal

Öner-Özkan

(2007). A self-presentation approach to transmission of good and bad news. Social Behavior and Personality, 35, 63-78.

24.

Wanous

J. P.

Reichers

A. E.

(1996). Estimating the reliability of a single-item measure. Psychological Reports, 78, 631-634.

25.

Wanous

J. P.

Reichers

A. E.

Hudy

M. J.

(1997). Overall job satisfaction: How good are single-item measures? Journal of Applied Psychology, 82, 247-252.

26.

Weenig

M. W. H.

Groenenboom

A. C. W. J.

Wilke

H. A. M.

(2001). Bad news transmission as a function of the definitiveness of consequences and the relationship between communicator and recipient. Journal of Personality and Social Psychology, 80, 449-461.

27.

Weenig

M. W. H.

Knuttel

(2001). [The impact of dyadic and group similarity of fate on good and bad news transmission]. Unpublished raw data.

28.

Wilson

S. R.

Aleman

C. G.

Leatham

G. B.

(1998). Identity implications of influence goals. A revised anaysis of face-threatening acts and application to seeking compliance with same-sex friends. Human Communication Research, 25, 64-96.

29.

Youngblut

J. M.

Casper

G. R.

(1993). Single-item indicators in nursing research. Research in Nursing and Health, 16, 459-465.