Do You Want the Good News or the Bad News First? The Nature and Consequences of News Order Preferences

Abstract

Information often comes as a mix of good and bad news, prompting the question, “Do you want the good news or the bad news first?” In such cases, news-givers and news-recipients differ in their concerns and considerations, thus creating an obstacle to ideal communication. In three studies, we examined order preferences of news-givers and news-recipients and the consequences of these preferences. Study 1 confirmed that news-givers and news-recipients differ in their news order preferences. Study 2 tested two solutions to close the preference gap between news-givers and recipients and found that both perspective-taking and priming emotion-protection goals shift news-givers’ delivery patterns to the preferred order of news-recipients. Study 3 provided evidence that news order has consequences for recipients, such that opening with bad news (as recipients prefer) reduces worry, but this emotional benefit undermines motivation to change behavior.

Keywords

giving bad news communication bad news delivery news-givers news-recipients news order

The process of giving or getting bad news is difficult for most people (Rosen & Tesser, 1970; Sweeny & Shepperd, 2007). News-givers face a number of concerns when giving bad news, including how best to give the news without hurting the news-recipient, how to decrease their own anxiety, and what exactly they should say during the interaction (Maynard, 2003; Sweeny & Shepperd, 2007). Unsurprisingly, then, the difficulty of delivering bad news has inspired extensive popular media articles that prescribe “best” practices for giving bad news (Bies, 2012; Lickerman, 2013; Sun, 2011), but these prescriptions remain largely anecdotal rather than empirically based.

This lack of scientific attention is particularly problematic in light of evidence suggesting that news-recipients have strong preferences for how good and bad news should be delivered (Marshall & Kidd, 1981). Some evidence suggests, however, that news-givers and news-recipients may disagree on the best order in which to relay bad and good news. In the present studies, we compare the news order preferences of news-givers and news-recipients, test two solutions for bringing disparate news order preferences into line between news-givers and news-recipients, and examine whether news order has consequences for news-recipients’ reactions to news.

News Order Preferences

Despite some attention given to bad news delivery in the medical literature, almost no research examines preferences for the order of receiving good and bad news. In other words, the typical answer to the question “Which do you want first, the good news or the bad news?” remains largely unanswered. In general, people prefer to begin with a loss or negative outcome and ultimately end with a gain or positive outcome, rather than the reverse (Do, Rupert, & Wolford, 2008; Ross & Simonson, 1991). That is, people tend to prefer improving sequences of events, sequences that increase in positivity or decrease in negativity. This pattern emerges in preferences for how people experience hypothetical headache (Chapman, 2000), spend their weekend time (Lowenstein & Prelac, 1993), experience comfort or pain during dental procedures (Kaakko et al., 2003), and sequences of financial loss and gain (Ross & Simonson, 1991). These findings suggest that people may prefer to learn bad news first and then proceed to good news when given a choice of news order. Indeed, the only studies of news order preferences found that the vast majority (up to 88%) of people wanted to hear bad news first when given a choice (Marshall & Kidd, 1981). In fact, news-recipients are astute in their news order preferences: People who experience improving sequences report greater positive affect following the experience (Baumgartner, Sujan, & Padgett, 1997; Kaakko et al., 2003; Kahneman, Fredrickson, Schreiber, & Redelmeier, 1993).

Receiving bad news is difficult, but giving bad news is not easy, either. Although people who frequently deliver bad news, such as physicians, report a primary focus on patients’ needs in bad news delivery (McKee & Ptacek, 2001; Sweeny, Shepperd, & Han, 2011), they must also mitigate their own stress and discomfort to avoid long-term consequences such as burnout (Baile et al., 2000; Sweeny & Shepperd, 2007). Thus, news-givers face competing goals in bad news delivery that may lead them to use communication strategies that are not in news-recipients’ best interests. Due to the double burden faced by news-givers (i.e., concern for their own and the news-recipient’s discomfort), we doubt that preferences for giving good and bad news align with preferences for receiving good and bad news. That is, the very same people who prefer to “get the bad news out of the way” when they are news-recipients may prefer to lead with good news when they are news-givers. Despite the preliminary finding that people typically prefer to get the bad over with (Marshall & Kidd, 1981), news-givers likely experience another type of motivation that may overpower their sensitivity to this preference. Research suggests that people are reluctant to relay negative information, a phenomenon called the “MUM effect” (Rosen & Tesser, 1970). Research on the MUM effect (Mum about Undesirable Messages) reveals that people engage in a variety of strategies to postpone or forego the delivery of negative information. Thus, when a news-giver must convey both good and bad news, delivering the good news first can serve to delay the unpleasant experience of giving bad news.

Further exacerbating news-givers’ insensitivity to news-recipients’ order preferences are egocentric biases that may prompt news-givers to predominantly focus on their discomfort with giving bad news and overestimate their ability to communicate well, while underestimating the perspective and emotions experienced by news-recipients (e.g., Keysar, 2007; Kruger, Epley, Parker, & Ng, 2005; Newton, 1990). These findings suggest that news-givers, unlike news-recipients, may prefer to lead with good news and end with bad news and may be largely oblivious to news-recipients’ conflicting preference for receiving the bad news first.

To the extent that news-givers focus on their own anxieties while news-recipients wait anxiously for bad news, one possible solution for the resulting divergence in news order preferences is for news-givers to adopt the news-recipient’s perspective. Perspective-taking by considering the emotions, thoughts, and contributions of others has reduced egocentric biases in several domains, including formations of social bonds (Galinsky, Ku, & Wang, 2005), reductions in intergroup conflict (Galinsky, 2002), and decreased stereotyping (Galinsky & Moskowitz, 2000). Thus, perhaps perspective taking can trigger news-givers’ awareness that news-recipients prefer to hear bad news first, thus decreasing the recipients’ anxious anticipation of the bad news while news-givers tip-toe into conversations with good news.

Does News Order Matter?

Perspective-taking may offer one solution to align news order preferences between news-givers and news-recipients. However, a second and perhaps equally important question is whether giving news in one order or the other determines recipients’ affective and behavioral responses to the news. Perhaps the most common contexts in which people deliver bad news are health-care settings and workplaces. In these environments, health-care professionals and employers often deliver bad news in hopes that recipients will take action to improve their health or increase productive workplace activities. But as previously discussed, news-givers and news-recipients prioritize their emotional responses during and after bad news delivery (Maynard, 2003; Sweeny & Shepperd, 2007). In the case of news-recipients, their preference to receive bad news first places a premium on their affective response to the news (Baumgartner et al., 1997; Kaakko et al., 2003; Kahneman et al., 1993). Thus, although perspective-taking may offer one way to align news order preferences and thereby improve recipients’ emotional state, it may do so at the expense of failing to catalyze behavior change.

News-givers may intend to minimize negative affect and promote behavior change in recipients, but it may be difficult to achieve both goals simultaneously. In fact, negative affect, and particularly worry, is positively associated with preventive health behavior (Consedine, Adjei, Ramirez, & McKiernan, 2008; Croyle & Lerman, 1993; Diefenbach, Miller, & Daly, 1999; McCaul, Schroeder, & Reid, 1996). These findings suggest that any news-giving strategy that successfully reassures recipients may simultaneously undermine their motivation to subsequently make beneficial changes to their behavior.

Overview and Hypotheses

In three studies, we compared news order preferences between news-givers and news-recipients, tested two possible solutions for bringing news order preferences into alignment between news-givers and news-recipients, and examined whether varying news orders (good followed by bad vs. bad followed by good) yield different outcomes for recipients. Study 1 replicated and extended previous research by comparing news order preferences of givers and recipients. We hypothesized that news-recipients would report a strong preference for receiving bad news first, whereas news-givers would prefer to give good news first.

Study 2 provided a test of a perspective-taking manipulation and compared its effectiveness with a communication-goal manipulation. News-givers in Study 2 also delivered good and bad news face-to-face with a confederate, thus amplifying the mundane and experimental realism of the procedures. We hypothesized that news-givers generally would prefer to give good news first. However, we further hypothesized that taking the perspective of the news-recipient would bring givers’ news order preference in line with recipients’ preferences, such that news-givers would more often give bad news first. Finally, we hypothesized that priming emotion-protection goals would also shift the order of news delivery toward recipients’ preference for bad news first, thus providing two possible routes to aligning news order preferences.

Finally, Study 3 examined news-recipients’ affective and behavioral responses to good and bad news in an effort to determine whether one news order is “best.” We hypothesized that news-recipients would prefer to hear bad news first and that receiving news in this preferred order would mitigate their negative emotional responses to the news. However, we also hypothesized that violating news-recipients’ order preferences by opening with good news would lead to greater behavior change and that rising worry throughout the news delivery would mediate the relationship between news order and behavioral responses.

Study 1

Method

Undergraduate participants (N = 121; 71% female; 31% Hispanic/Latino, 34% Asian, 12% White/Caucasian, 12% Black/African American, and 11% Other or Multicultural) came to a lab in pairs, although they signed up separately for the study. The researcher immediately seated the participants in separate rooms. Participants were randomly assigned to one of two roles: news-giver or news-recipient. Although the experiences of givers and recipients were similar in some ways, participants in each role essentially participated in separate studies from this point on. The experimenter told news-recipients that they would take a personality test and receive feedback on the test at the end of the session. Recipients then completed a questionnaire that included the “personality test” (actually the Big Five Inventory; John, Donahue, & Kentle, 1991) and demographic items. When recipients completed the questionnaire, the experimenter returned and said the following: “I have scored your personality test. You have both good and bad results. Which would you like to hear first?” The experimenter recorded the recipient’s choice (either “good news first” or “bad news first”) and then asked why the news-recipient made that choice.

In contrast, the experimenter told news-givers that although they would complete a personality test during the session, they would not receive feedback on their test but would instead be responsible for delivering test feedback to the other participant in the study. After givers completed the same questionnaire completed by news-recipients, the experimenter returned to the room and said the following: “I have scored the personality test for the other participant. S/he has both good and bad results. Which would you like to deliver first?” The experimenter recorded the news-giver’s choice and then asked why the giver made that choice.

Two coders independently categorized the open-ended explanations. Categories for news-recipients’ responses were emotional statements, behavior change statements, and comments about news order colloquialisms or improving sequences. Categories for news-givers’ responses were self-focused statements (e.g., explanations referring to the self and any anxiety experienced or how to make the news-giving task easier for oneself) and other-focused statements (e.g., explanations referring to the recipient, their emotions, or how to deliver the news in a way that is easiest for the recipient). Reliability between the two raters was high (92% agreement), and on the responses for which disagreement occurred, the two raters discussed the items and came to an agreement.

Results

Results supported our hypothesis that news-recipients and news-givers would differ in their news order preferences, χ² (1, N = 121) = 12.16, p < .001, ϕ = .33. As hypothesized, the majority (78%) of recipients wanted to hear bad news first, whereas news-givers were split in their order preferences: 54% preferred to give the good news first, and 46% preferred to give the bad news first. Participants’ reasons for their choices provided further insight into news order preferences and were consistent with our hypotheses. We excluded 15 participants because their responses could not be coded into any category (e.g., “Just because,” “It was the first thing that popped into my head,” or “I don’t know”). News-recipients’ explanations confirmed that most recipients prefer to end with good news to reduce negative emotions: 53.3% (n = 32) of news-recipients’ open-ended explanations addressed emotional concerns (e.g., “I’d rather hear the good news [last] to make me feel better”). Further support for our hypotheses came from a noteworthy minority of news-recipients (10%, n = 6) who described their preference to end with bad news as a way to motivate behavior change (e.g., “I need to hear the bad news last so I know what to change.”). Several news-recipients (18%, n = 11) also reported colloquialisms that reflect preferences for improving sequences by saying, for example, “It’s always better to save the best for last” or “I like to end on a high note.”¹

A closer look at news-givers’ explanations for their preferences paint a picture that is also consistent with our hypotheses. News-givers’ explanations largely fell into two categories: self-focused and other-focused. We excluded 14 participants because their explanations could not be coded as self- or other-focused (e.g., “I don’t know”). As an example of a self-focused explanation, one giver defended her choice of leading with good news by saying that “it makes me uncomfortable to give bad news first.” In contrast, an other-focused participant who chose to lead with bad news noted that “it’s best to give the good news before she leaves so that she can remember the good on her way out.” In fact, participants who provided other-focused explanations were more likely to choose to give bad news first (67% chose this order), whereas participants who gave self-oriented explanations were far more likely to choose to give good news first (75% chose this order), χ²(1, N = 47) = 4.79, p = .04, ϕ = .32.

Study 2

Study 2 tested and compared two potential interventions to increase the likelihood that news-givers will lead with bad news, as recipients prefer. We introduced a perspective-taking manipulation that prompted participants to focus on the news-recipient. The goal of this manipulation was to counteract the egocentric biases that may lead news-givers to focus on their own conversational concerns to the detriment of news-recipients’ preferences (Keysar, 2007; Kruger et al., 2005; Newton, 1990). In addition, we introduced a manipulation of news-givers’ goals during the news delivery task. As discussed earlier, two broad goals compete during bad news delivery: protecting the emotions of the news-recipient and easing the process for the news-giver (Sweeny & Shepperd, 2007; Sweeny et al., 2011). We hypothesize that similar to a broad perspective-taking manipulation, prompting news-givers with the goal of protecting the recipient’s emotions would bring their news order choices in line with recipients’ preferences.

Method

Undergraduate participants (N = 190, 62% female, M_age = 19.96, SD_age = 2.49) received partial course credit for participating in this study. The sample, again, was ethnically diverse: 37% of participants identified as Asian, 33% identified as Hispanic/Latino, 11% identified as White/Caucasian, 5% identified as Black/African American, and 14% identified as other/multiple. Participants arrived at the lab and met a confederate posing as another student participating in the study. The experimenter brought the participant and confederate into the lab and explained that the purpose of this study was to assess how people respond to personality tests results. By drawing out of a hat, the experimenter ostensibly randomly assigned the participant and confederate to one of two roles: news-giver or news-recipient (both sheets of paper said “news-giver,” and the experimenter always assigned the real participant to his or her condition first). Thus, the confederate was always assigned the news-recipient role. The experimenter then seated the participant and the confederate in separate rooms. The experimenter then introduced the cover story: “Our lab has developed a new personality assessment that will be used in job selection processes at businesses across the United States.” While the confederate supposedly took the personality test, the participant completed demographic measures and other filler questionnaires.

After the participant completed the first questionnaire, the experimenter entered the room and presented two sheets of paper, ostensibly reporting the personality test results of the confederate. One sheet listed five positive personality trait results (indicating above-average organizational aptitude, ambitiousness, detail-orientation, creativeness, and innovativeness), and the other sheet listed five negative personality trait results (indicating below-average leadership abilities, cooperativeness, self-confidence, patience, and work ethic). The experimenter presented the two sheets to the participant in counterbalanced order (one on the left, one on the right) and said,

The other participant has completed the personality assessment. I printed out their personality test results for you to look over before you discuss them with the other participant. Because these personality assessments were developed to be used by businesses, we cannot give the participant the printed out results. Instead, you will need to prepare to give the other participant the results of his/her assessment. You can communicate these results to the other participant in any way you see fit.

At this point, the experimenter introduced the manipulation (randomly assigned prior to the start of the session). In the control condition, the experimenter simply told the participant, “You can communicate these results to the other participant in any way you see fit.” In the perspective-taking condition, the experimenter read the following: “It can be very anxiety-inducing to receive this type of feedback. Please place yourself in the other person’s shoes and think about how you would like to hear the results delivered, if you were the one receiving them.” In the emotion-protection condition, the experimenter read the following:

Please keep in mind that the other person may feel hurt, sad, or upset about hearing their negative personality characteristics. It is your goal to communicate their results in a way that minimizes how bad they might feel. Please look over their results and take a moment to think about how best to deliver these results in a way that protects the other person’s emotions and helps to keep them from feeling sad or upset.

Following the introduction of the manipulation, the experimenter told the participant to take a few moments to review the confederate’s results and to let the experimenter know when she or he was ready to go to the other room to deliver the news. At this point, the experimenter escorted the participant into the confederate’s room and prompted the participant to deliver the results, a process that was video-recorded with the knowledge of the participant, although the participant believed that the camera was directed at the confederate to record reactions to the news.

After delivering the results with no further guidance from the experimenter, the participant returned to the other room to complete a second questionnaire that included a manipulation check to ensure that participants understood the perspective-taking or emotion-protection prompt (“What did the researcher ask you to think about before you gave the results to the other participant?” “I was supposed to think about protecting the other person’s feelings” or “I was supposed to think about the other person receiving the results.”), and an item asking why they presented the results in the order they chose (“Please describe why you delivered the results in the way you did.”). Participants were then fully debriefed, a process that began by revealing the confederate to be a research assistant and not another participant, and thanked for their participation.

Results

Two independent coders watched the videos of the news-delivery and recorded whether the participant started with positive test results or negative test results. The coders agreed in all cases. Participants in the control, perspective-taking, and emotion-protection conditions differed from each other, χ²(2, N = 190) = 10.29, p = .006, ϕ = .23. As hypothesized and consistent with Studies 1 and 2, 68% of participants in the control condition chose to deliver the good news first (see Table 1 for frequencies). However, the control participants delivered the personality test results differently from participants in the perspective-taking condition, χ²(1, N = 128) = 4.49, p = .03, ϕ = .20, and the emotion-protection condition, χ²(1, N = 127) = 8.51, p = .004, ϕ = .27. In fact, both interventions were successful at aligning news-givers’ behavior with recipients’ preferences: A majority of participants in the perspective-taking (52%) and emotion-protection conditions (60%) delivered the bad news first.² The interventions were equally effective, such that participants in the perspective-taking and emotion-protection conditions did not differ from each other in their news order choices, χ²(1, N = 125) = .41, p = .52, ϕ = .07.

Table 1.

Correlations Among Condition, Change in Worry, Change in Mood, and Behavioral Response (Study 3).

	1	2	3	4
1. Condition	—
2. Change in worry	−.52*	—
3. Change in mood	−.57*	.50*	—
4. Choice between watching videos and stapling papers	.24*	−.33*	−.02	—

Note. Condition coded such that 0 = good news first, 1 = bad news first. Positive change scores indicate increases in worry and worsening mood. Choice coded such that 0 = watching videos, 1 = stapling papers.

Study 3

The findings from Study 1 confirmed the strong preference of news-recipients to hear bad news first and extends those findings to document the egocentric biases driving news-givers’ preference to lead with good news. These explanations point to a clear cause of news order misalignment: News-recipients are focused on their own emotional considerations, and many news-givers give little thought to news-recipients’ concerns. Study 2 demonstrated that perspective-taking and priming emotion-protection goals result in more frequent alignment in news order preferences (i.e., to deliver news in bad-then-good patterns). However, both of these approaches focus heavily on mitigating the emotional reactions of news-recipients to bad news. A conflicting goal of bad news delivery is to promote behavior change, and it is unlikely that pacifying news-recipients’ negative emotions will motivate behavior change (e.g., Consedine et al., 2008; Croyle & Lerman, 1993; Diefenbach et al., 1999; Hay, McCaul, & Magnan, 2006; McCaul et al., 1996). Thus, Study 3 provides an empirical investigation to the question of whether giving bad news first always leads to the most positive outcomes, either emotional or behavioral.

We hypothesized that news-recipients would feel better (i.e., positive emotional outcomes) following a bad-to-good order of news delivery, but that they would be more likely to engage in productive behavioral responses to the news following a good-to-bad order of news delivery (i.e., detrimental emotional outcomes). We further hypothesized that rising worry would be the catalyst for behavioral responses to the news. The research on improving sequences provides evidence that the trajectory of experience matters more than the total or average experience (e.g., Baumgartner et al., 1997; Kaakko et al., 2003; Kahneman et al., 1993; Ross & Simonson, 1991), and thus we hypothesized that people who experience a downward trajectory (increasing worry) would be motivated to respond productively to the news, whereas people who experience an upward trajectory (diminishing worry) would be relatively unmotivated. We also included a measure of general mood to evaluate the specificity of worry’s mediating role.

Method

Undergraduate students (N = 79; 58% female; 38% Hispanic/Latino, 36% Asian, 10% White/Caucasian, 4% Black/African American, 2% Native Hawaiian or Pacific Islander, 4% Middle Eastern, 5% Other or Multicultural, and 1% unreported) participated for partial course credit. Participants were randomly assigned to one of two conditions: bad news last (n = 42) or good news last (n = 37). When participants arrived at the lab, the experimenter informed them that the study examined college students’ personalities. Participants then completed a computer-based personality test.

After completion of the baseline questionnaire, participants received their (bogus) personality test results. In the good news first condition, participants received good test results followed by bad test results, and in the bad news first condition, participants received bad test results followed by good test results. All participants received the same set of contrived personality test results (the personality test was never actually scored). The experimenter verbally led participants through their scores on each personality category while participants viewed their results on a computer monitor. Participants viewed good results and bad results on separate successive screens, and the experimenter determined when to move forward to the next set of results. The good test results consisted of five desirable traits (altruistic motivations, intuitive instincts, self-monitoring, friendliness, and idealism) for which participants scored in the 81st to 96th percentile, depending on the trait. Bad test results consisted of five undesirable traits (rigid thinking, lack of self-confidence, unmotivated, neuroticism, and unreliability) for which participants scored in the 84th to 96th percentile.

Following the first set of test results, either good or bad, participants completed a questionnaire that included one item assessing general mood (“Please indicate your current mood”; 1 = extremely good mood, 7 = extremely bad mood) and two items assessing worry (“I am worried about what the results of this test could mean,” “I am concerned about the personality test results I received”; 1 = strongly disagree, 7 = strongly agree, α = .76).

Following the second set of test results, participants completed another questionnaire that included the mood and worry items from the previous questionnaire, as well as a validation measure for the forthcoming behavioral measure (“I plan on taking steps to improve my personality”; 1 = strongly disagree, 7 = strongly agree). The experimenter then explained to participants that they still had some time remaining for their experimental session and gave participants the option to watch a set of interactive videos on improving aspects of their personality or to help the lab by stapling sets of questionnaires, which served as our behavioral measure.³ Participants indicated whether they wanted to watch the personality improvement videos or staple papers. In actuality, no personality improvement videos or paper-stapling task existed, and so, regardless of participants’ choice, the experimenter began the debriefing process after recording the participants’ decision. Participants indicated in the debriefing process that they truly believed they would have to either watch videos or staple papers, depending on their choice.

Results

Preliminary analyses

We first examined gender or ethnicity (once again using the largest two subsets of Asians and Hispanic/Latinos) to identify important covariates. A 2 (Condition: bad news first vs. good news first) × 2 (Ethnicity: Asian vs. Hispanic/Latino) analysis of variance for each dependent variable of interest revealed no main effects of ethnicity and no interaction effects, all Fs < 1.78 ps > .19. The same was true for gender, all Fs < 2.74 ps > .10. Thus, we did not retain gender or ethnicity as covariates in our primary analyses or conduct further analyses with demographic variables.

Affective reactions

We examined the trajectories of mood and worry over the course of receiving the news by conducting a mixed 2 (Condition: bad news first vs. good news first) × 2 (Time: between results vs. after results) ANOVA.⁴ A main effect of time emerged for worry but not mood, such that participants became more worried over the course of news delivery, F(1, 77) = 6.21, p = .01, r_es = .28, but their mood did not change on average, F(1, 74) = 0.03, p = .85, r_es = .02. No main effects of order condition emerged, Fs < .25, ps > .62. However, as expected, the main effect of time was qualified by an interaction between time and order condition for worry, F(1, 77) = 28.13, p < .0001, r_es = .60, and an interaction also emerged for mood, F(1, 74) = 36.17, p < .0001, r_es = .57. Participants in the good news first condition became more worried (M = 1.01, SD = 1.34), t(41) = 4.91, p < .0001, r_es = .61, and developed a worse mood, t(39) = 4.76, p < .0001, r_es = .60, over the course of receiving the news. In contrast, participants in the bad news first condition became less worried (M = −.36, SD = .89), t(36) = −2.48, p = .02, r_es = .38, and developed a better mood, t(35) = −3.80, p < .001, r_es = .54, over the course of news delivery.

Behavioral responses

We conducted a goodness-of-fit chi-square analysis to compare participants’ choices of watching personality videos versus stapling papers across conditions. As hypothesized, participants in the good news first condition were more likely to opt to watch the personality videos than were participants in the bad news first condition, χ²(1, N = 79) = 4.37, p = .04, ϕ = .24. The majority (66.67%, n = 28) of participants in the good news first condition chose to watch the videos, whereas fewer than half (43.24%, n = 16) of participants in the bad news first condition chose to watch the videos. Participants’ self-reported intentions to improve their personality validated our behavioral measure, such that participants in the good news first condition reported stronger intentions (M = 4.79, SD = 1.7) than participants in the bad news first condition (M = 3.76, SD =1.6), t(77) = 2.79, p = .007, r_es = .30.

Mediation analyses

Using methods developed and recommended by Preacher and Hayes (2008), we estimated the path coefficients in a multiple mediator model and generated bootstrap bias-corrected confidence intervals (2,000 bootstrapped samples) for specific indirect effects of order condition on behavior through change in worry and change in mood. These procedures are preferable to a Sobel test, which does not allow for multiple mediators. See Table 2 for bivariate correlations.

Table 2.

Study 2 News Order Preference Percentages and Frequencies by Condition.

	Control	Perspective taking	Emotion protection
Bad news first	32% (21)	52% (33)	60% (37)
Good news first	68% (44)	48% (30)	40% (25)

The direct effects of change in worry, b = −.13, p = .01, and change in mood, b = .13, p = .03, on behavior were statistically significant, and the bias-corrected 95% confidence intervals (CI) for the indirect effects of order condition on behavior through change in worry and change in mood did not contain zero (for change in worry, CI = .05 to .32; for change in mood, CI = −.34 to −.02). However, a pairwise contrast between the indirect effects for change in worry and change in mood indicated that as predicted, change in worry was a far stronger mediator than change in mood (95% bias-corrected CI = .12 to .58).⁵

General Discussion

We conducted three studies to examine whether news-recipients and news-givers would provide similar answers to the question, “Which do you want first, the good news or the bad news?”, whether prompting perspective-taking or emotion-protection goals in news-givers would align news order preferences, and whether news order alignment promotes beneficial responses for news-recipients. The findings support our hypotheses that news-givers and news-recipients differ in their news order preferences but can be brought into alignment with recipients’ preferences by reducing egocentric biases and drawing attention to recipients’ concerns, and that each news order has distinct and consequential outcomes for recipients.

As hypothesized, news-recipients in Study 1 preferred to hear bad news first, and the preferences of news-givers differed from the preferences of news-recipients, such that givers more often preferred to deliver good news first. In fact, news-givers in Study 2 showed an even stronger preference to lead with good news, presumably due to the face-to-face nature of news delivery in that study. Participants in Study 1 provided explanations for their preferences that further supported the theoretical underpinnings for our hypotheses. News-recipients unsurprisingly provided explanations that put their own concerns at the forefront, and as predicted, these concerns reflected a desire to mitigate unpleasant emotional reactions to the news or alluded to preferences for improving sequences. On the other hand, although many news-givers provided explanations consistent with the MUM effect and thus led with good news, other news-givers overcame their egocentric biases to take the perspective of the news-recipient and thus led with bad news, just as recipients prefer.

Furthermore, the success of Study 2’s perspective-taking and emotion-protection interventions point convincingly toward the cause of misalignment between news-givers’ and news-recipients’ order preferences. That is, it seems that people have accurate insight into the best way to receive good and bad news, at least with regard to protecting their emotions, but this insight is hampered in news-givers by immediate and thus more salient concerns about the discomfort of delivering bad news. These findings suggest that a simple reminder to news-givers to remember how it feels to be the recipient of bad news may be sufficient to reduce egocentric biases and thus align news-givers’ delivery with recipients’ order preferences even in stressful news delivery contexts.

Our findings from Study 3 revealed that news-recipients have good insight into their emotional responses to bad news: People who got the bad news out of the way, thus ending on a high note as they preferred, were in a better mood and less worried than recipients who heard bad news last. However, news-recipients seem to have less insight regarding how they should receive bad news that motivates them to take action. That is, the emotional benefits of a bad-then-good pattern may ultimately be to news-recipients’ detriment when they need to change some behavior. Study 3 demonstrated that news-recipients who received bad news first were less likely to take an easy opportunity to improve, choosing instead to engage in a boring and personally unproductive task (i.e., stapling papers for the researcher). Furthermore, the results from Study 3 suggest that the very negative emotions news-recipients hope to avoid by getting the bad news over with actually promote beneficial behavior. As hypothesized, the unpleasant roller-coaster of worry that accompanies a good-then-bad news pattern explained the relationship between this news order and increased willingness to watch a personality improvement video. These findings support previous research indicating a relationship between worry and behavior change (Consedine et al., 2008; Diefenbach et al., 1999; Hay et al., 2006; McCaul et al., 1996), although our findings go a step further by identifying trajectories of worry as key predictors of behavior following bad news delivery.

Although news-givers’ explanations in Study 1 suggest that their intention is to selfishly delay bad news delivery, Study 3 indicates that their preference for a good-then-bad news order is beneficial when they want to encourage the recipient to respond to the news with behavior change. That is, news-givers may inadvertently motivate news-recipients as a side effect of their effort to delay bad news delivery. However, to the extent that news-givers are focused on selfish concerns regarding the difficulty of news delivery, they may use other communication strategies that undermine the motivating effect of a good-then-bad news order. For example, news-givers may find it difficult to leave bad news hanging and instead tack on some form of encouragement to the end of the conversation, creating a “bad news sandwich” that is likely to mitigate the recipient’s emotional response and undermine motivation to change.

Of course, in some cases the goal of bad news delivery is to minimize recipients’ emotional reactions. For example, editors rejecting a manuscript, professors notifying students of a failing grade, and spouses announcing their desire for a divorce may hope to avoid active efforts on the part of the news-recipient to alter the situation. In these and many other cases, news-givers want to relay news in ways that ultimately protects recipients’ emotions, or perhaps more selfishly, that avoids an emotional outburst (e.g., Buckman, 1984), which requires giving news in the order preferred by news-recipients (leading with bad news).

Implications

These findings have important implications for communication in many domains. Doctors must give good and bad health news to patients, teachers must give good and bad academic news to students, and romantic partners may at times give good and bad relationship news to each other (“It’s not you, it’s me”). Our findings suggest that the doctors, teachers, and partners in these examples might do a poor job of giving good and bad news because they forget for a moment how they want to hear the news when they are the patients, students, and spouses, respectively. News-givers attempt to delay the unpleasant experience of giving bad news by leading with good news (Rosen & Tesser, 1970), while recipients grow anxious knowing that the bad news is yet to come. This tension can erode communication and result in poor outcomes for both news-recipients and news-givers. For example, people who feel that a news-giver did a poor job of communicating bad news report less satisfaction with the communication, more anxiety about the news, and poorer adjustment (Ellis & Tattersall, 1999), and news-givers who are unsure of the best way to give bad news experience anxiety, distress, poor mental health, and distrust from news-recipients (Ambuel & Mazzone, 2001; Holland, 1989; Ramirez et al., 1995). Our finding that news-givers can improve their communication simply by remembering to take the recipient’s perspective provides a potentially powerful antidote to some of the negative consequences of poor news delivery.

Of course, news-givers have more options for giving good and bad news than simply presenting news in one of two orders. As mentioned earlier, news-givers might attempt to further soften the blow of bad news by delivering bad news encased within a “bad news sandwich” or good-bad-good pattern. Numerous websites and management handbooks recommend the use of this strategy (see Buron & McDonald-Mann, 2007; Dohrenwend, 2002; Kay, Meyer, & French, 1965), but our findings suggest that the primary beneficiary of the bad news sandwich is news-givers, not news-recipients. Although recipients may be pleased to end on a high note, they are unlikely to enjoy anxiously waiting for the other shoe to drop during the initial good news.

Under some circumstances, people may welcome bad news, particularly when the news can guide decision-making or motivate behavior (Trope, 1986). The usefulness of the news in our studies (personality feedback) is somewhat ambiguous; although some participants may have welcomed the opportunity to learn how they could improve, others may have perceived personality to be unchangeable, thus undermining the usefulness of good or bad news in that domain. Although bad news may be less holistically “bad” when it is useful, we suspect that our findings regarding news order preferences generalize across news that varies in usefulness. News-givers likely have little insight into the potential usefulness of bad news for recipients, and recipients may if anything be even more eager to hear the bad news first if they believe it will serve a useful purpose. However, our findings, paired with related research demonstrating that success feedback increases openness to subsequent negative feedback (Trope & Neter, 1994), suggests that news-recipients would particularly benefit from a good-then-bad news order when the bad news is useful to recipients.

Finally, although our studies addressed related good and bad news, news-givers can integrate unrelated good news into bad news delivery if related good news is unavailable. For example, a husband might compliment his wife’s outfit before revealing that he forgot to pay an important bill. Future studies can examine the generalizability of our findings to situations in which good news is unrelated to the bad news, but some evidence indicates that even unrelated positive feedback provides coping resources that increase openness and responsiveness to negative feedback (Trope & Neter, 1994). However, if news-recipients suspect that news-givers provided good news simply to make the conversation easier for themselves or with other selfish or disingenuous motives, this strategy may ultimately be ineffective.

Conclusion

Delivering bad news is a difficult and uncomfortable task, made even worse when news-givers feel unsure about how to proceed with the conversation (Ramirez et al., 1995). Fortunately, our findings provide two “take-home messages” that can provide news-givers with confidence and may ease the communication process. First, news-givers should be aware that left to their own devices, they are likely to focus on their own emotional concerns at the expense of the news-recipients’ communication preferences. This inward focus translates to a preference for easing into bad news delivery by opening with good news, despite a clear preference on the part of news-recipients to get the bad news over with quickly. Second, news-recipients should be aware that left to their own devices, they also are likely to focus on emotional concerns, in this case at the expense of their motivation to change their behavior in response to bad news.

Footnotes

Acknowledgements

The authors thank David Funder and Robin DiMatteo for their feedback on an earlier version of this manuscript and Ann T. Nguyen, Christina Armenta, and Max Barranti for their help conducting these studies. We also thank the reviewers for their time and comments on earlier versions of this manuscript. A portion of these data were presented at the 2012 Society for Personality and Social Psychology conference in San Diego, CA and the 2011 Society for Personality and Social Psychology conference in San Antonio, TX.

Declaration of Conflicting Interests

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

Funding

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

Notes

References

Ambuel

Mazzone

M. F.

(2001). Breaking bad news and discussing death. Primary Care, 28, 249-267.

Baile

W. F.

Buckman

Lenzi

Glober

Beale

E. A.

Kudelka

A. P.

(2000). SPIKES—A six-step protocol for delivering bad news: Application to the patient with cancer. The Oncologist, 5, 302-311.

Baumgartner

Sujan

Padgett

(1997). Patterns of affective reactions to advertisements: The integration of moment-to-moment responses into overall judgments. Journal of Marketing Research, 34, 219-232.

Bies

(2012, May 30). The 10 commandments for delivering bad news. Forbes Leadership Forum Online. Retrieved from http://www.forbes.com/sites/forbesleadershipforum/2012/05/30/10-commandments-for-delivering-bad-news/

Buckman

(1984). Breaking bad news: Why is it still so difficult? British Medical Journal, 288, 1597-1599.

Buron

R. J.

McDonald-Mann

(2007). Giving feedback to subordinates (Vol. 86). Greensboro, NC: Center for Creative Leadership.

Chapman

(2000). Preferences for improving and declining sequences of health outcomes. Journal of Behavioral Decision Making, 13, 203-218.

Consedine

N. S.

Adjei

B. A.

Ramirez

P. M.

McKiernan

J. M.

(2008). An object lesson: Source determines the relations that trait anxiety, prostate cancer worry, and screening fear hold with prostate screening frequency. Cancer Epidemiology, Biomarkers & Prevention, 17, 1631-1639.

Croyle

R. T.

Lerman

(1993). Interest in genetic testing for color cancer susceptibility: Cognitive and emotional correlates. Preventive Medicine, 22, 284-292.

10.

Diefenbach

M. A.

Miller

S. M.

Daly

M. B.

(1999). Specific worry about breast cancer predicts mammography use in women at risk for breast and ovarian cancer. Health Psychology, 18, 532-536.

11.

A. M.

Rupert

A. V.

Wolford

(2008). Evaluations of pleasurable experiences: The peak-end rule. Psychonomic Bulletin & Review, 15, 96-98.

12.

Dohrenwend

(2002). Serving up the feedback sandwich. Family Practice Management, 9, 43-46.

13.

Ellis

P. M.

Tattersall

M. H. N.

(1999). How should doctors communicate the diagnosis of cancer to patients? Annals of Medicine, 31, 336-341.

14.

Galinsky

A. D.

(2002). Creating and reducing intergroup conflict: The role of perspective-taking in affecting out-group evaluations. Research on Managing Groups and Teams, 4, 85-113.

15.

Galinsky

A. D.

Wang

C. S.

(2005). Perspective-taking and self-other overlap: Fostering social bonds and facilitating social coordination. Group Processes & Intergroup Relations, 8, 109-124.

16.

Galinsky

A. D.

Moskowitz

G. B.

(2000). Perspective-taking: Decreasing stereotype expression, stereotype accessibility, and in-group favoritism. Journal of Personality and Social Psychology, 78, 708-724.

17.

Hay

J. L.

McCaul

K. D.

Magnan

R. E.

(2006). Does worry about breast cancer predict screening behaviors? A meta-analysis of the prospective evidence. Preventive Medicine, 42, 401-408.

18.

Holland

J. C.

(1989). Now we tell, but how well? Journal of Clinical Oncology, 7, 557-559.

19.

John

O. P.

Donahue

E. M.

Kentle

R. L.

(1991). The Big Five Inventory–Versions 4a and 54. Berkeley: Institute of Personality and Social Research, University of California, Berkeley.

20.

Kaakko

Horn

M. T.

Weinstein

Kaufman

Leggott

Coldwell

S. E.

(2003). The influence of sequence of impressions on children’s anxiety and discomfort. Pediatric Dentistry, 25, 357-364.

21.

Kahneman

Fredrickson

B. L.

Schreiber

C. A.

Redelmeier

D. A.

(1993). When more pain is preferred to less: Adding a better end. Psychological Science, 4, 401-405.

22.

Kay

Meyer

H. H.

French

J. R. P.

(1965). Effects of threat in a performance appraisal interview. Journal of Applied Psychology, 49, 311-317.

23.

Keysar

(2007). Communication and miscommunication: The role of egocentric processes. Intercultural Pragmatics, 4, 71-84.

24.

Kruger

Epley

Parker

(2005). Egocentrism over e-mail: Can we communicate as well as we think? Journal of Personality and Social Psychology, 89, 925-936.

25.

Lickerman

(2013, January). How to deliver bad news to patients: 9 tips to do it better. MedPage Today’s Kevin MD. Retrieved from http://www.kevinmd.com/blog/2013/01/deliver-bad-news-patients-9-tips.html

26.

Lowenstein

Prelac

(1993). Preferences for sequences of outcomes. Psychological Review, 100, 91-108.

27.

Marshall

L. L.

Kidd

R. F.

(1981). Good news or bad news first? Social Behavior and Personality, 9, 223-226.

28.

Maynard

(2003). Bad news, good news: Conversational order in everyday talk and clinical settings. Chicago, IL: University of Chicago Press.

29.

McCaul

K. D.

Schroeder

D. M.

Reid

P. A.

(1996). Breast cancer worry and screening: Some prospective data. Health Psychology, 15, 430-433.

30.

McKee

T. L. E.

Ptacek

J. T.

(2001). I’m afraid I have something bad to tell you: Breaking bad news from the perspective of the giver. Journal of Applied Social Psychology, 31, 246-273.

31.

Newton

(1990). Overconfidence in the communication of intent: Head and unheard melodies (Unpublished doctoral dissertation). Stanford University, Stanford, CA.

32.

Nguyen

Legg

A. M.

Sweeny

(2011). Do you want the good news or bad news first? News order influences recipients’ mood, perceptions, and behaviors. UCR Undergraduate Research Journal, 5, 31-36.

33.

Preacher

K. J.

Hayes

A. F.

(2008). Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods, 40, 879-891.

34.

Ramirez

A. J.

Graham

Richards

M. A.

Cull

Gregory

W. M.

Leaning

M. S.

. . . Timothy

A. R.

(1995). Burnout and psychiatric disorder among cancer clinicians. British Journal of Cancer, 71, 1263-1269.

35.

Rosen

Tesser

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

36.

Ross

W. T.

Simonson

(1991). Evaluations of pairs of experiences: A preference for happy endings. Journal of Behavioral Decision Making, 4, 273-282.

37.

Sun

(2011, April 6). 10 tips for delivering bad news. TechRepublic. Retrieved from http://www.techrepublic.com/blog/10things/10-tips-for-delivering-bad-news/2396

38.

Sweeny

Shepperd

J. A.

(2007). Being the best bearer of bad tidings. Review of General Psychology, 11(3), 235-257.

39.

Sweeny

Shepperd

J. A.

Han

P. K. J.

(2011). The goals of communicating bad news in health care: Do physicians and patients agree? Health Expectations, 16, 230-238. doi:10.1111/j.1369-7625.2011.00709.x

40.

Trope

(1986). Identification and inferential processes in dispositional attribution. Psychological Review, 93, 239-257.

41.

Trope

Neter

(1994). Reconciling competing motives in self-evaluation: The role of self-construal in feedback seeking. Journal of Personality and Social Personality, 66, 646-657.