Displaying Fairness in Bad News Delivery (PDF)

Journal of Applied Psychology © 2016 American Psychological Association 2016, V...

Journal of Applied Psychology © 2016 American Psychological Association 2016, Vol. 101, No. 6, 779 –792 0021-9010/16/$12.00 http://dx.doi.org/10.1037/apl0000087 Displaying Fairness While Delivering Bad News: Testing the Effectiveness of Organizational Bad News Training in the Layoff Context Manuela Richter, Cornelius J. König, Christopher Koppermann, and Michael Schilling Universität des Saarlandes Although giving bad news at work is a stressful experience, managers are often underprepared for this challenging task. As a solution, we introduce organizational bad news training that integrates (a) principles of delivering bad news from the context of health care (i.e., bad news delivery component), and (b) principles of organizational justice theory (i.e., fairness component). We argue that both the formal This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. and fair delivery of bad news at work can be enhanced with the help of training to mitigate distress both This document is copyrighted by the American Psychological Association or one of its allied publishers. for the messenger and the recipient. We tested the effectiveness of training for the delivery of a layoff as a typical bad news event at work. In 2 studies, we compared the performance of a training group (receiving both components of training) with that of a control group (Study 1, Study 2) and a basics group (receiving the bad news delivery component only; Study 2) during a simulated dismissal notification meeting. In general, the results supported our hypotheses: Training improved the formal delivery of bad news and predicted indicators of procedural fairness during the conversation in both studies. In Study 2, we also considered layoff victims’ negativity after the layoff and found that training significantly reduced negative responses. This relationship was fully mediated by layoff victims’ fairness perceptions. Despite preparation, however, giving bad news remained a challenging task in both studies. In summary, we recommend that organizations provide managers with organizational bad news training in order to promote professional and fair bad news conversations at work. Keywords: delivering bad news, training, organizational justice, procedural fairness, layoff Supplemental materials: http://dx.doi.org/10.1037/apl0000087.supp “So although I wish I were here with better news, the fact is that you esteem and creates uncertainty about their future (e.g., Baumeister, and I are sitting here today because this will be your last week of Bratslavsky, Finkenauer, & Vohs, 2001), and managers feel un- employment at this company.” comfortable with their responsibility for giving this news and thus —(George Clooney alias Ryan Bingham in the motion picture Up doing harm to the employee (e.g., Molinsky & Margolis, 2005). in the Air by Dubiecki, Clifford, Reitman, & Reitman, 2009) Furthermore, self-presentation concerns, feelings of guilt, or an Giving bad news to an employee is as much a regular task for anticipation of negative employee reactions can contribute to man- managers as it is a difficult one (Bies, 2013). Managers have to agers’ reluctance to give bad news (Rosen & Tesser, 1970). communicate not only organizational downsizing and layoffs Unfortunately, managers’ concerns often become reality, espe- (Clair & Dufresne, 2004), but also negative performance feedback cially if bad news is given in an unfair and insensitive way. (Ilgen & Davis, 2000), pay cuts (Greenberg, 1990), negative hiring Organizational justice research has widely demonstrated that em- (Lavelle, Folger, & Manegold, 2014) or promotion decisions ployees respond adversely to unfair treatment while learning about (Lemons & Jones, 2001), or disciplinary warnings (Cole & bad news, be it job applicants receiving rejection (Gilliland, 1994), Latham, 1997). What all these conversations have in common is employees experiencing negative performance appraisal (Hol- the stress they arouse in managers and employees alike: Employ- brook, 1999), or employees being given notice of a layoff (Kon- ees feel threatened by bad news because it impairs their self- ovsky & Folger, 1991). Nevertheless, it has barely been explored how managers should be prepared for the challenge of giving bad news in a fair way. The present research therefore addresses whether training can be de- This article was published Online First February 15, 2016. veloped that is useful for improving managers’ performance in a Manuela Richter, Cornelius J. König, Christopher Koppermann, and bad news conversation with an employee and, as a result, for Michael Schilling, Department of Psychology, Universität des Saarlandes. reducing the negative impact of the delivery or receipt, respec- We thank William McKinley for his helpful feedback on an early tively, of bad news for managers as the messengers and employees version of this paper. We also thank Scott Tonidandel for his helpful advice as the recipients. For this purpose, we developed organizational and support on applying relative weight analysis to MANOVA using the bad news training, building upon principles of delivering bad statistical package R. Correspondence concerning this article should be addressed to Manuela health news from the context of health care (Baile et al., 2000; Richter, Universität des Saarlandes, Arbeits- & Organisationspsychologie, Rosenbaum, Ferguson, & Lobas, 2004) and integrating principles Campus A1 3, 66123 Saarbrücken, Germany. E-mail: [email protected] of organizational justice theory (Colquitt, 2001; Leventhal, 1980). saarland.de We then conducted first empirical tests of the effectiveness of 779  780 RICHTER, KÖNIG, KOPPERMANN, AND SCHILLING organizational bad news training while applying it to a bad news 2008; Baile et al., 1999; Bonnaud-Antignac, Campion, Pottier, & event that is both prototypical and one of the most challenging a Supiot, 2010; Rosenbaum et al., 2004). Such training usually manager might face in working life— delivering layoff news to an conveys knowledge about the systematic structuring of a bad news employee. conversation with a patient. A prominent example is the SPIKES protocol, which describes step-by-step strategic guidelines for Introducing Organizational Bad News Training delivering bad health news (Baile et al., 2000; Buckman, 1992). In particular, physicians should arrange the setting before the bad The Bad News Delivery Component of Training news conversation (setting up), assess the patient’s awareness of the problem (perception), inquire about the patient’s desire for Encountering bad news is an undesired and unpleasant event for information disclosure (invitation), deliver bad health news recipients and messengers alike. In particular, messengers exhibit (knowledge), address the emotions expressed (empathy), and ar- an aversion to giving bad news that hinders them from carrying out range follow-up steps (e.g., treatment plan) and summarize the the task properly, a phenomenon referred to as the MUM effect discussion (strategy and summary). This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. (“keeping mum about undesirable messages”; Rosen & Tesser, Given the positive effects of such training in the context of This document is copyrighted by the American Psychological Association or one of its allied publishers. 1970, p. 254). Research has shown that messengers’ concerns health care, organizational bad news training should include a bad about giving bad news are manifold, and include feelings of guilt news delivery component that clarifies the manager’s role and toward those suffering from bad news (Tesser & Rosen, 1972), provides knowledge about the formal delivery of bad news to an fears of negative evaluations and self-presentation concerns of employee, using a similar step-by-step protocol: First, managers being associated with bad news (Bond & Anderson, 1987), and should arrange the setting; second, they should deliver the bad anticipation of negative reactions of the recipients (Rosen & Tes- news immediately at the beginning of the meeting; third, they ser, 1970). In line with these findings, giving bad news to an should provide a detailed explanation for the bad news; fourth, employee can create considerable stress in managers, whether it they should deal with the emotions expressed by the employee; pertains to the communication of negative hiring decisions or to fifth, they should provide information about follow-up measures to the delivery of layoff news (e.g., Folger & Skarlicki, 1998; Lavelle promote planning for the future; and, sixth, they should summarize et al., 2014). In some cases, for instance if managers do not support the discussion. the decisions they have to convey (e.g., conducting a layoff due to downsizing rather than performance deficits), doing harm to an The Fairness Component of Training employee may also contradict a manager’s role expectation of being a “good” supervisor who aspires to promote and support his As much as giving bad news is a challenging task for managers, or her employees (Kets de Vries & Balazs, 1997). This situation implementing an unfavorable outcome with interpersonal sensitiv- may also create uncertainty about which behaviors are appropriate ity and fairness is all the more demanding (Folger & Skarlicki, to implement this task (Kahn, Wolfe, Quinn, & Snoek, 1964). Such 1998; Molinsky & Margolis, 2005). Nevertheless, organizational uncertainties about the role, together with a lack of critical knowl- justice research has widely demonstrated the beneficial effects of edge and mastery experience in giving bad news, may relate to fairness at work on the establishment of positive work outputs and managers’ self-efficacy concerns regarding their ability to deal relationships (for an overview, see Colquitt, Conlon, Wesson, with the challenging task successfully (Bandura, 1997), and this Porter, & Ng, 2001). In particular, if employees have to deal with may in turn affect their performance (Stajkovic & Luthans, 1998). negative work events, procedural fairness seems to be crucial for To reduce managers’ stress and to increase their performance in their favorable reactions to the organization and its agents (e.g., giving bad news, organizational bad news training needs to clarify Cohen-Charash & Spector, 2001; Colquitt et al., 2001). Procedural the manager’s role as a leader whose task is sometimes to give bad fairness refers to the processes and procedures used to make or news to an employee in order to achieve a “greater good or implement decisions (e.g., Leventhal, 1980). Research has shown, purpose” (Molinsky & Margolis, 2005, p. 245). Furthermore, it is for example, that fair performance appraisal procedures were as- deemed necessary that training conveys knowledge about the sociated with employees’ motivation to improve their performance appropriate behaviors for performing this task, thus providing (Elicker, Levy, & Hall, 2006), whereas unfair procedures in pro- managers with a sense of predictability and personal control of the motion decisions reduced employees’ commitment to their em- situation. In practice, behaviors crucial for giving bad news have ployer (Lemons & Jones, 2001). Similarly, in the context of usually been examined in health care professionals (Rosenbaum et reorganization, surviving employees reported more commitment al., 2004). Nevertheless, physicians and managers may have sim- and fewer turnover intentions if the reorganization process had ilar goals; for instance, to facilitate recipients’ acceptance of a been fair (Kernan & Hanges, 2002). Laid-off employees, on the negative outcome and to preserve their positive attitudes, and they other hand, were less angry (Barclay, Skarlicki, & Pugh, 2005) and also seem to be exposed to similar challenges. Although the nature less likely to complain and to take legal action against the em- of physicians’ jobs—working with people with physical or mental ployer (Konovsky & Folger, 1991; Wanberg, Gavin, & Bunce, illnesses—implies a constant exposure to giving bad health news, 1999) if the layoff procedure had been fair. they often report stress and concerns as well as a lack of confi- Given the benefits of procedural fairness, organizational bad news dence and competence in delivering a diagnosis (e.g., Cohen et al., training should include a fairness component that provides managers 2003; Orgel, McCarter, & Jacobs, 2010). Training in delivering with knowledge about procedural fairness and its enactment in order bad health news has been found to improve medical students’ and to improve the perceived fairness of a bad news conversation and, as residents’ performance and confidence in delivering a diagnosis a consequence, to reduce employees’ negativity toward their super- during role-playing scenarios with peers or actors (e.g., Baer et al., visor and their employer afterward. Specifically, procedural fairness  TESTING BAD NEWS TRAINING 781 can be increased by implementing the principles postulated by Lev- Margolis, 2005). As indicated by interviews with layoff agents enthal (1980): Procedures are fair if, for example, they are used conducted by Kets de Vries and Balazs (1997), undertaking the consistently across persons and time (consistency) and without any role of a layoff agent can violate a manager’s role perception as a bias or self-interest (bias suppression), if they are based on accurate supportive leader, arouse feelings of role ambiguity, and impair information (accuracy), represent the needs of all parties involved confidence in one’s ability to conduct this task. Furthermore, (representativeness), and follow moral and ethical standards (ethical- managers may also feel conflicted between the company’s busi- ity). To implement the consistency principle while giving bad news, ness objectives and employees’ well-being, that is, the opposing managers should communicate the news in an unambiguous and expectations of the two parties. Accordingly, being a layoff agent coherent manner throughout the conversation, and they should dem- has been related to managers distancing themselves from the onstrate bias suppression by appealing to the facts instead of attrib- laid-off employees (Clair & Dufresne, 2004; Folger & Skarlicki, uting the bad news to the employee’s personality. To promote repre- 1998) in order to avoid feelings of emotional discomfort and sentativeness, managers should offer two-way communication and confrontation with negative employee reactions. Unfortunately, give employees the opportunity to voice their views; accuracy should managers’ concerns often hinder them from giving the bad news of This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. be fostered by providing adequate and reasonable explanations of the a layoff in a fair and sensitive way (Folger & Skarlicki, 1998). This document is copyrighted by the American Psychological Association or one of its allied publishers. bad news; and the principle of ethicality should be met by treating Research has shown that employees and their representatives often employees with politeness, dignity, and respect; for instance, by consider it necessary to take legal steps against the employer mentioning their positive attributes and contributions (as suggested by following unfair layoff procedures (Konovsky & Folger, 1991; Wood & Karau, 2009). Previous research has already shown that Wanberg et al., 1999). Such organizational justice deficits in leaders can be trained to be fairer in their interactions with their conducting layoffs are also reflected in German labor court statis- employees (for an overview, see Skarlicki & Latham, 2005). Fairness tics (Destatis, 2015): Since 2010, about 400,000 labor court pro- training (vs. no training) increased not only subordinates’ perceptions ceedings have been completed every year, around 50% of which of their leaders’ procedural fairness (Cole & Latham, 1997; Skarlicki were brought against the employer for layoff reasons. In 2014, for & Latham, 1996, 1997), but also employees’ organizational citizen- instance, 201,354 of 392,061 (51%) completed proceedings were ship behavior (Skarlicki & Latham, 1996, 1997). submitted for layoff reasons. Taken together, organizational bad news training needs to in- To summarize, the first component of organizational bad news clude (a) a bad news delivery component to improve managers’ training (i.e., the bad news delivery component) should provide the formal delivery of bad news and (b) a fairness component to layoff agent with knowledge about the formal delivery of layoff improve their display of fairness during this procedure. Whereas news by using the step-by-step protocol described previously. the former should influence managers’ outcomes (e.g., reduce These systematic guidelines should improve their performance stress), the latter should influence employee outcomes (e.g., reduce during a dismissal notification meeting. Furthermore, information negativity toward employer). about their role and about ways to manage critical employee reactions should give layoff agents an idea about what might Applying Organizational Bad News Training to the happen during the bad news conversation. This should provide them with a sense of personal control, which should in turn Layoff Context mitigate their feelings of stress and emotional discomfort in giving A layoff can be considered as both a typical and one of the most bad news (Tetrick & LaRocco, 1987). The second component of challenging bad news events at work. Therefore, it was deemed an organizational bad news training (i.e., the fairness component) appropriate field of application for testing the effectiveness of should teach layoff agents ways in which to enact procedural organizational bad news training on messengers’ performance in a fairness principles (Leventhal, 1980) while delivering layoff news. bad news conversation. For many years, organizational downsiz- This should improve the perceived fairness of the notification ing has been discussed as a prevalent phenomenon in both the procedure and, as a result, mitigate negative employee reactions, psychology and management literature, although it has often been given the findings that procedural fairness has a positive impact on related to negative outcomes for both organizations and humans laid-off employees’ emotional reactions (e.g., anger; Barclay et al., (e.g., Datta, Guthrie, Basuil, & Pandey, 2010; McKee-Ryan, Song, 2005) and attitudes (e.g., desire to complain or to take legal action; Wanberg, & Kinicki, 2005; van Dierendonck & Jacobs, 2012). Konovsky & Folger, 1991; Wanberg et al., 1999). Layoffs impair not only the physiological and psychological well- Study 1 was designed to test the overall effectiveness of orga- being of the employees who lose their jobs, the layoff victims nizational bad news training. For this purpose, a training group (McKee-Ryan et al., 2005; Paul & Moser, 2009), and the employ- was provided with complete organizational bad news training, ees remaining at the company, the layoff survivors (Grunberg, which included both the bad news delivery and the fairness com- Moore, & Greenberg, 2001; van Dierendonck & Jacobs, 2012), but ponents, and compared with a no-training control group. Follow- also the well-being of the managers who have to communicate the ing this, Study 2 was designed to identify the specific impact of the dismissal messages, the layoff agents (Grunberg, Moore, & Green- two components. For this purpose, three experimental groups were berg, 2006; Kets de Vries & Balazs, 1997; Parker & McKinley, needed: (a) a training group that was provided with both the bad 2008). news delivery and the fairness components of organizational bad Specifically, having to conduct layoffs is a stressful task for news training, (b) a basics group that was provided with the managers because they have to harm their employees by commu- bad news delivery component only, and (c) a control group that nicating a job loss for economic or strategic reasons that are was provided with neither of the components. Given the expected beyond an employee’s individual control and usually independent effect of the training components, the formal delivery of layoff of performance deficits (Folger & Skarlicki, 1998; Molinsky & news should improve, and feelings of emotional discomfort should  782 RICHTER, KÖNIG, KOPPERMANN, AND SCHILLING decrease for the training group and the basics group compared to PhD course at a German university (72% studied psychology), and the control group. However, the enactment of procedural fairness eight were professionals from start-up companies located at the should improve only for the training group, and layoff victims’ campus. Thirty-eight participants (74%) worked at least part-time, negative reactions should also decrease only for the training group and a considerable number of respondents had some kind of layoff as compared to both the basics group and the control group. experience: Eight (16%) reported that they had been laid off in the Finally, given that practical rehearsal is an important means to past, 25 (49%) had witnessed at least one layoff in a close rela- create proficiency and confidence in being able to perform a task tionship (e.g., family member or close friend), and 28 (55%) had successfully (Bandura, 1997), layoff agents’ confidence in their witnessed at least one layoff in a more distant relationship. ability to deliver layoff news should improve for the training group All participants had to formally register for a training session if only the training group is given the opportunity to exercise the and were randomly chosen for the training group or the control task as compared to the basics group and the control group. Based group, respectively. Training was announced as a workshop to on the above discussion, we therefore make the following hypoth- practice conduct in critical leader–member interactions. Partici- eses: pants in the training group received training in a traditional class- This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. room setting (see training intervention) and performed a dismissal This document is copyrighted by the American Psychological Association or one of its allied publishers. Hypothesis 1 (H1): Layoff agents’ formal delivery of layoff notification meeting in a face-to-face role-play session about five news improves for (a) a training group and (b) a basics group, days later (see testing scenario). Participants in the control group as compared to a control group. performed the dismissal meeting without training. Hypothesis 2 (H2): Layoff agents’ feelings of emotional dis- Training intervention. Classroom training consisted of a comfort in delivering layoff news decrease for (a) a training half-day workshop and comprised five learning modules (for de- group and (b) a basics group, as compared to a control group. tails, see Table 1). In Module 1, trainees were provided with information about their role as a layoff agent and the challenges of Hypothesis 3 (H3): Layoff agents’ confidence in their ability giving bad news. In Module 2, trainees were taught how to enact to deliver layoff news improves for a training group as com- procedural fairness principles while delivering layoff news (Lev- pared to (a) a control group and (b) a basics group. enthal, 1980). For example, they learned how to provide adequate and reasonable explanations for the layoff reasons in order to Hypothesis 4 (H4): Layoff agents’ procedural fairness in de- fulfill the accuracy principle. In Module 3, we explained the livering layoff news improves for a training group as com- step-by-step protocol of giving bad news at work (e.g., delivering pared to (a) a control group and (b) a basics group. the bad news immediately). Module 4 described emotional reac- Hypothesis 5 (H5): (a) Layoff victims dismissed by a training tions that might be expressed by employees in response to bad group report less negativity toward the employer than those news (i.e., shock, anger, negotiation) and how to deal with them. dismissed by a basics group and a control group, and (b) this In Module 5, trainees were asked to take the perspective of either effect is mediated by layoff victims’ perceptions of procedural the manager or the employee in a dismissal meeting and to act fairness. according to these roles in two peer role-plays (i.e., rehearsal). Testing scenario.1 All participants were assigned to the role Study 1 of the leading manager of the customer support division of a mobile telephone provider. They were informed that due to chang- In Study 1, we compared the performance of a training group ing market conditions, the company had reported declines in sales with a no-training control group in a simulated bad news conver- and that top management had decided upon strategic restructuring sation (i.e., dismissal notification meeting) in order to test H1a, and headcount reduction. Participants then were asked to conduct H2a, H3a, and H4a. We chose a laboratory setting to evaluate the a dismissal meeting with Mrs. (or Mr.) Brauer, a 29-year-old effectiveness of training for three reasons. First, it allowed us to employee who had been employed at the company for 5 years. A randomly assign participants to the training conditions. It would small conference table had been prepared and participants were hardly be possible, and would also be ethically problematic, to given some time to plan the conversation. They were also advised withhold training from a sample of managers conducting opera- to conduct the meeting professionally because they would receive tional layoffs. Second, laboratory settings and role-playing exer- feedback afterward. cises allow trainees to practice new skills without risking harm due Mrs. (or Mr.) Brauer was represented by one of five role- to improper treatment (Skarlicki & Latham, 2005), which is also players, henceforth referred to as the layoff victim. We chose both the reason why health care researchers typically simulate doctor– male and female victims to account for any differential reactions of patient interviews using actors or student peers as role-players participants toward men or women losing their jobs. In a prelim- (e.g., Baer et al., 2008; Bonnaud-Antignac et al., 2010). Third, as inary training session, layoff victims had been informed about their organizational bad news training has not yet been studied, we role and trained to play a shocked and stunned employee. To decided to begin this research in a laboratory setting to gain an realize semistandardized interviews, they had been taught a pro- impression about its effectiveness and applicability. tocol of predetermined statements which had to be made in each dismissal meeting (i.e., “You can’t be serious!”, “What did I do Method wrong?”, “Why me?”, “But we have just taken out a loan. I Participants and design. The sample consisted of 51 partic- ipants (30 females, 21 males) with a mean age of 27.18 years 1 The scenario used in Studies 1 and 2 is available online as supplemen- (SD ⫽ 6.33). Forty-three were students on a Bachelor, Master, or tal materials (Suppl. A).  TESTING BAD NEWS TRAINING 783 Table 1 Training Modules for Studies 1 and 2 Condition No. Module Description Study 1 Study 2 1 Role of a layoff agent Information about layoffs, the role and responsibilities of a leader and a Training Training layoff agent Basics 2 Fairness and Importance and enactment of procedural fairness principles: Training Training communication Consistency (e.g., be coherent, be unambiguous) Bias suppression (e.g., be objective, be impartial) Accuracy (e.g., provide reasonable explanations) Representativeness (e.g., allow voice, active listening) Ethicality (e.g., mention contributions, be polite) 3 Formal delivery of Step-by-step protocol of giving bad news: Training Training This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. bad news 1. Arranging the setting (e.g., private room) Basics This document is copyrighted by the American Psychological Association or one of its allied publishers. 2. Delivering the bad news immediately 3. Explaining the reasons for the decision in detail 4. Managing the employee’s emotions 5. Future planning/follow-up measures (e.g., job coaching) 6. Summary/finishing the meeting 4 Employee reactions Coping with employee reactions (i.e., shock, anger, negotiation) Training Training Basics 5 Rehearsal Practicing a dismissal meeting in a role-playing exercise Training Training Behavioral feedback from a trainer Note. Classroom training (Study 1) included lectures, group discussions, and a peer role-play (i.e., rehearsal). Web-based training (Study 2) included online lectures, video aids/exercises (e.g., learning game), and a virtual role-play (i.e., rehearsal). In both studies, the rehearsal during the training intervention was different from the simulated dismissal meeting during the testing session. In Study 2, the basics group was only provided with Modules 1, 3, and 4 and received no video aids/exercises. thought we had a good relationship!”, “What shall I do now?”, “I dismissal meeting referring to the step-by-step protocol of giving can’t manage this!”, “This is too much for me!”, “What will this do bad news (e.g., “Delivered the layoff message within the first five to me?”, “And if I kill myself?”).2 sentences”); on 5-point Likert scales, six items assessed the flex- During the conversation, an observer monitored the participants’ ibility in applying this protocol (e.g., “Responded to the employ- performance. The observer was hidden in the background, invisi- ee’s behavior flexibly”). Additionally, the observer evaluated par- ble to the participants and thus unable to unwittingly influence or ticipants’ enactment of procedural fairness (Leventhal, 1980). A coach their performance through nonverbal communications (e.g., multi-item measure was developed for the purpose of this study: facial expressions). Both the observer and the layoff victims were Consistency was measured with three items (e.g., “Remained bind- blind to the participants’ training condition; the participants them- ing”), bias suppression (e.g., “Based the conversation on occupa- selves were also unaware of the existence of different training tional grounds only”), ethicality (e.g., “Behaved in a polite and conditions. Immediately after the dismissal meeting, the dependent respectful manner”) and representativeness (e.g., “Facilitated the variables were measured. Finally, participants received feedback employee to express his/her views and feelings”) with four items about their performance, were debriefed about their experiences each, and accuracy (e.g., “Tailored the explanations to the em- during the simulation and the purpose of the study, and offered a follow-up talk if necessary. The whole procedure lasted for ap- ployee’s specific needs”) with six items. proximately 30 min. Measures.3 All dependent measures were collected after the Results and Discussion dismissal meeting. Unless otherwise specified, all scales used 5-point Likert scales ranging from 1 ⫽ strongly disagree to 5 ⫽ Means, standard deviations, and correlations among the vari- strongly agree. ables are displayed in Table 2. We used independent samples t Data from self-reports. Participants’ feelings of emotional discomfort were measured in terms of negative affect using a 2 subscale of the Positive and Negative Affect Schedule (PANAS; Layoff victims’ statements were inspired from practical reports of Watson, Clark, & Tellegen, 1988). Negative affect was assessed managers and consultants (Andrzejewski & Refisch, 2015; Richter & on 10 adjectives (e.g., “distressed”), using a 5-point scale (1 ⫽ not König, 2013). Although the reaction “And if I kill myself?” might seem very challenging to the reader, Andrzejewski and Refisch caution managers at all to 5 ⫽ extremely). Participants’ confidence in their ability to to take suicidal intentions seriously, and this problem is also addressed in deliver layoff news was assessed with six items developed for the the movie Up in the Air (Dubiecki et al., 2009), cited at the beginning of purpose of this study (e.g., “I felt capable of conducting the the article. 3 dismissal meeting”). More detailed information about the items developed for Studies 1 and 2 is available online as supplemental materials (Suppl. B). Unfortunately, Data from the layoff observer. The observer indicated partic- we had to exclude the layoff victims’ ratings of procedural fairness in ipants’ formal delivery of bad news on 12 items; on dichotomous Study 1 because of poor scale quality. As a consequence, we reduced the scales (1 ⫽ yes, 2 ⫽ no), six items measured the elements of the number of role-players in Study 2 to improve rating quality.  784 RICHTER, KÖNIG, KOPPERMANN, AND SCHILLING Table 2 Means, Standard Deviations, and Zero-Order Correlations (Study 1) Variable M SD 1 2 3 4 5 6 7 8 9 10 1. Condition 0.51 0.50 — Layoff agent self-reports 2. Negative affect 2.34 0.64 ⫺.09 (.84) 3. Confidence 3.19 0.71.27 ⫺.36ⴱⴱ (.85) Layoff observer ratings 4. Elements 4.53 1.29.62ⴱⴱ ⫺.13.31ⴱ — 5. Flexibility 3.56 0.66.64ⴱⴱ ⫺.26.20.83ⴱⴱ (.74) 6. Consistency 3.77 0.78.68ⴱⴱ ⫺.31ⴱ.18.60ⴱⴱ.73ⴱⴱ (.82) This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 7. Bias suppression 3.75 0.83.64ⴱⴱ ⫺.20.04.64ⴱⴱ.74ⴱⴱ.77ⴱⴱ (.81).60ⴱⴱ ⫺.04.78ⴱⴱ.75ⴱⴱ.72ⴱⴱ.72ⴱⴱ This document is copyrighted by the American Psychological Association or one of its allied publishers. 8. Accuracy 3.58 0.65.20 (.77) 9. Representativeness 3.48 0.73.48ⴱⴱ ⫺.08.21.55ⴱⴱ.58ⴱⴱ.47ⴱⴱ.46ⴱⴱ.57ⴱⴱ (.74) 10. Ethicality 3.82 0.83.47ⴱⴱ ⫺.29ⴱ.08.52ⴱⴱ.58ⴱⴱ.63ⴱⴱ.63ⴱⴱ.68ⴱⴱ.52ⴱⴱ (.79) Note. N ⫽ 51. Condition: 0 ⫽ control group, 1 ⫽ training group. Elements ⫽ whether participants complied with the elements of the step-by-step protocol of giving bad news; Flexibility ⫽ whether participants used the protocol in a flexible way. Cronbach’s alpha coefficients are displayed in parentheses where applicable. ⴱ p ⬍.05. ⴱⴱ p ⬍.01. tests to test H1a, H2a, and H3a (see Table 3 for results).4 In terms these correlations into account. The highest relative weights were of layoff agents’ formal delivery of bad news, we analyzed the found for consistency and bias suppression, with 21% and 14% of observer’s perception of participants’ compliance with the ele- variance accounted for by training condition, respectively. Thus, ments of the step-by-step protocol of giving bad news and their layoff agents’ enactment of consistency and bias suppression seem flexibility in applying this protocol. As predicted, analyses re- to be the most important factors in determining the perceived vealed significant effects of training condition on elements and procedural fairness of a layoff. flexibility, indicating that the training group complied better with In summary, although training was not effective in reducing the elements of the step-by-step protocol and was also more participants’ self-reported negative affect or in increasing their flexible in applying the protocol than the control group. H1a was confidence, Study 1 demonstrated the overall effectiveness of therefore supported. However, training did not reduce participants’ organizational bad news training on participants’ performance negative affect and also did not improve their confidence in their from an observer’s perspective: Training improved not only layoff ability to deliver layoff news (all ps ⬎.05); H2a and H3a were agents’ formal delivery of bad news, but also their enactment of therefore not supported. procedural fairness principles (Leventhal, 1980) while delivering Due to theoretical and methodological relationships among the procedural fairness variables, we used multivariate analyses of layoff news from an observer’s viewpoint. Nevertheless, it re- variance (MANOVA) to test H4a. MANOVA results for the ob- mained unclear which underlying mechanism produced the posi- server data revealed a significant multivariate effect of training tive effects of training, that is, whether it was the bad news condition on the combined procedural fairness principles, Wilks’ delivery aspect or the fairness aspect of organizational bad news ⌳ ⫽.48, F(5, 45) ⫽ 9.81, p ⬍.01, ␩2 ⫽.52, indicating that the training. More specifically, did training work because of partici- two groups differed significantly in terms of their enactment of the pants’ increase in knowledge about the formal delivery of bad procedural fairness principles. Follow-up independent samples t news at work or because of their enactment of procedural fairness tests revealed significant effects of training condition on each while communicating the bad news? Study 2 was designed to procedural fairness principle (see Table 3 for results). From the address this question. observer’s perspective, training improved layoff agents’ enactment of fairness during the dismissal meeting procedure: Trainees de- livered the layoff more consistently (consistency) and impartially (bias suppression) than nontrainees. The training group also out- 4 performed the control group in providing adequate explanations In Study 1, we also computed all analyses adding participants’ past (accuracy), allowing layoff victims to voice their views and feel- layoff experiences and gender as covariates because experiences with ings (representativeness), and treating them with respect (ethical- layoffs might have shaped their attitudes toward downsizing and therefore their performance in the dismissal meeting (Sronce & McKinley, 2006), ity), thus fully supporting H4a. However, multiple t tests as and because women might have been more empathic and supportive than follow-up tests to a MANOVA suffer from the methodological men toward the layoff victim (Trobst, Collins, & Embree, 1994). We also limitation of ignoring correlations among dependent variables, included layoff victims’ gender as a covariate because participants might unlike relative weight analysis applied to MANOVA (Tonidandel have responded differentially toward a man or a woman losing his/her job. We did not find any significant changes in our results considering the & LeBreton, 2013). Relative weight analysis allowed us to deter- covariates, except for participants’ confidence which turned significant mine the relative contribution of each fairness variable to the (p ⫽.04). In Study 2, considering covariates did not change the results at overall multivariate effect of training (again see Table 3), taking all.  TESTING BAD NEWS TRAINING 785 Table 3 Results of t Tests and Relative Weights (Study 1) Condition Control Training Confidence (n ⫽ 25) (n ⫽ 26) t test interval (CI) Relative Lower Upper M SD M SD t(49) d weight 95% CI 95% CI Layoff agent self-reports Negative affect 2.40 0.69 2.28 0.59 ⫺0.66 ⫺0.19 Confidencea 3.00 0.70 3.39 0.69 1.98 0.56 Layoff observer ratings Formal delivery Elements 3.72 1.17 5.31 0.84 5.58ⴱⴱ 1.56 This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. Flexibility 3.13 0.61 3.97 0.39 5.85ⴱⴱ 1.64 This document is copyrighted by the American Psychological Association or one of its allied publishers. Procedural fairness Consistency 3.24 0.70 4.28 0.42 6.46ⴱⴱ 1.80.205.079.359 Bias suppression 3.21 0.84 4.26 0.38 5.81ⴱⴱ 1.61.138.032.298 Accuracy 3.18 0.56 3.95 0.50 5.22ⴱⴱ 1.45.088.014.209 Representativeness 3.13 0.68 3.82 0.61 3.79ⴱⴱ 1.07.070.002.189 Ethicality 3.43 0.90 4.20 0.56 3.71ⴱⴱ 1.03.021 ⬍.001.121 Note. Elements ⫽ whether participants complied with the elements of the step-by-step protocol of giving bad news; Flexibility ⫽ whether participants used the protocol in a flexible way. Relative weight analysis was only computed for the procedural fairness variables. Raw weights and 95% confidence interval around the raw weights are displayed. a One participant did not provide confidence information, resulting in df ⫽ 48 for this variable. ⴱ p ⬍.05. ⴱⴱ p ⬍.01. Study 2 training group and the basics group, the role-play took place about