Week 3 Question

Public Relations Review 36 (2010) 256–262

Contents lists available at ScienceDirect

Public Relations Review

Restoring reputations in times of crisis: An experimental study of the Situational Crisis Communication Theory and the moderating effects of locus of control

An-Sofie Claeys ∗, Verolien Cauberghe 1, Patrick Vyncke 2

Universiteit Gent, Korte Meer 7-11, 9000 Gent, Belgium

a r t i c l e i n f o

Article history: Received 17 December 2009 Received in revised form 19 April 2010 Accepted 20 May 2010

Keywords: Crisis communication Corporate reputation Crisis responsibility Crisis response strategy Locus of control

a b s t r a c t

This study attempts to provide empirical evidence for Coombs’ (2007) Situational Crisis Communication Theory (SCCT), which provides guidelines for matching crisis response strategies to crisis types to best restore organizational reputations in times of crisis. The impact of crisis type and crisis response strategies on perceptions of corporate reputation is measured for 316 consumers participating in a 3 (crisis type: victim crisis, accidental crisis, preventable crisis) × 3 (crisis response: deny strategy, diminish strategy, rebuild strategy) between-subjects experimental design. The results show that preventable crises have the most negative effects on organizational reputation and that the rebuild strategy leads to the most positive reputational restoration. Moreover, the more severe people judge a crisis to be, the more negative are their perceptions of the organization’s reputation. The inter- action effect between crisis type and crisis response strategies on corporate reputation is not significant. However, a person’s locus of control has a moderating impact on the rela- tionship between crisis response strategy and organizational reputation. Specifically, the results show that people with an external locus of control prefer the use of deny strategies more than people with an internal locus of control.

© 2010 Elsevier Inc. All rights reserved.

1. Introduction

The Situational Crisis Communication Theory (SCCT) was developed recently to investigate which crisis response strategy managers should apply in specific crisis situations to restore an organization’s reputation in the best possible way (Coombs, 2007). This is essential, considering that no organization is spared having to go through crises during its lifetime (Spillan, 2003). These crises challenge organizational reputations and credibility (Arpan, 2002; Weiner, 2006). Prior research has rarely applied experimental, quantitative designs to investigate the topic of crisis response strategies (Benoit, 1995). Therefore, the main goal of this study is to test the matches between crisis types and crisis response strategies, as recommended by the SCCT, using a between-subjects 3 × 3 factorial experimental design.

Until now, research on crisis communication has paid little attention to the moderating role of personality traits that may affect respondents’ reactions either to a crisis or to the crisis response strategy used to manage the crisis (Coombs, 2007). Therefore, the current study investigates the moderating influence of locus of control—a personality trait that refers

∗ Corresponding author. Tel.: +32 494 26 69 32. E-mail addresses: [email protected] (A.-S. Claeys), [email protected] (V. Cauberghe), [email protected] (P. Vyncke).

1 Tel.: +32 472 23 07 87. 2 Tel.: +32 09 264 68 91.

0363-8111/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.pubrev.2010.05.004

A.-S. Claeys et al. / Public Relations Review 36 (2010) 256–262 257

Table 1 Match between crisis types and crisis response strategies.

Crisis types Crisis response strategies

Victim cluster Deny strategies Natural disaster Attack the accuser Rumor Denial Workplace violence Scapegoat Product tampering/Malevolence

Accidental cluster Diminish strategies Challenges Excuse Technical-error accidents Justification Technical-error product harm

Preventable cluster Rebuild strategies Human-error accidents Compensation Human-error product harm Apology Organizational misdeed with no injuries Organizational misdeed management misconduct Organizational misdeed with injuries

Source: adapted from: Coombs (2007), p. 168 and 170.

to the attributions that people make for behavioural consequences and reinforcements (Duffy, Downey, & Shiflett, 1977). A person’s locus of control may influence the value he or she places on either internal or external control in organizational crisis responses (Rotter, 1966). Because crisis response strategies differ in the way they attempt to convince an audience that a crisis is attributed to either internal or external factors, it stands to reason that people with an internal locus of control prefer the use of different crisis response strategies than people with an external locus of control.

In what follows, we present a theoretical overview and develop the hypotheses. Then, we discuss the method of the experiment and the results.

2. The SCCT

The SCCT divides crisis types into three crisis clusters (Coombs, 2007; Coombs & Holladay, 2002). The victim cluster is defined as crises with weak attributions of organizational responsibility. The accidental cluster involves crises in which a certain, but low, level of responsibility are attributed to the organization. The preventable cluster includes crises for which the organization is perceived as being responsible (see Table 1).

The more responsibility that is attributed to the organization with respect to the crisis, the more negative is the impact on the organizational reputation (Coombs, 1998). Therefore, different types of crisis inflict different amounts of reputational damage. This leads to the following hypothesis:

H1. The victim crisis leads to a less negative effect on organizational reputation than the accidental or preventable crisis. The accidental crisis leads to a moderately negative effect on organizational reputation compared with the victim or preventable crisis. The preventable crisis leads to the most negative impact on organizational reputation compared with the victim or accidental crisis.

Within the range of response strategies, several clusters have also been detected (Coombs, 2004, 2007; Huang, Lin, & Su, 2005) (see Table 1). Deny strategies claim that no crisis exists (denial) or declare that the organization has no responsibility for it (scapegoat) (Coombs, 2007). Diminish strategies argue that a crisis is not as serious as people think (justification) or minimize organizational responsibility (excuse). Rebuild crisis strategies offer compensation for the crisis or apologize. Research has shown that offering an apology—a rebuild crisis response strategy—leads to more effective reputation restoration than the more defensive deny or diminish response strategies (Coombs & Holladay, 2008). Thus:

H2. The reputation of an organization using rebuild crisis response strategies is more positive than the reputation of an organization using either deny or diminish crisis response strategies.

The reputational threat is determined by the crisis type, which is defined in terms of the perceived responsibility attributed to the organization (Coombs, 2004, 2007). This idea is based on attribution theory (Coombs, 2007). Crisis managers should select crisis response strategies that are appropriate for the potential extent of reputational damage a certain crisis may inflict (Coombs & Holladay, 2002). In an experimental study, Coombs and Holladay (1996) find that when crisis communication responses match the crisis type in terms of responsibility attribution, this leads to a more positive reputation perception than either no response or a mismatched response. Their research finds that organizational reputation benefits when diminish strategies are used in response to accidental crises and when rebuild strategies are matched with preventable crises.

Previous research has examined the match between crisis type and a single crisis response strategy (Coombs & Holladay, 1996). However, Benoit (1997) suggests that the use of a combination of strategies can increase the effectiveness of image restoration. The SCCT advises crisis managers to combine crisis response strategies from the same cluster (Coombs, 2007). In line with the SCCT, the best option appears to be a combination of several strategies from the same response cluster into

258 A.-S. Claeys et al. / Public Relations Review 36 (2010) 256–262

a crisis response, depending on the crisis type. On the basis of previous SCCT-related research, we expect deny strategies to match with victim crises, diminish strategies to match with accidental crises, and rebuild strategies to match with preventable crises (Coombs, 2007). Thus:

H3. Matching crisis type and crisis response strategy leads to a less negative effect on organizational reputation than mismatches between crisis type and crisis response strategy.

Although Coombs (1998) has not been able to empirically prove that the severity of a crisis has an incrementally negative effect on reputational damage, the author asserts the likelihood that more severe crises indeed have a greater negative impact on organizational reputation than crises that lead to trivial damage (Coombs, 1998; Coombs & Holladay, 2002). Coombs (1998) tests his assumption by operationalizing crisis severity as the amount of property damage and seriousness of related injuries. An alternative to this approach would be simply to use perceived crisis severity. Therefore, we use the latter operationalization in the current study, which may help further uncover the impact of crisis severity on reputational damage.

H4. The greater the perceived severity of a crisis, the greater is the negative impact on organizational reputation.

3. Moderating influence of locus of control

A theory that informs the SCCT is attribution theory, which focuses on the causal inferences people make (Coombs & Holladay, 1996; Folkes, 1988). Organizations in crisis situations attempt to influence these causal inferences by using certain crisis response strategies. Most research has focused on environmental factors that lead an observer to make internal and external attributions about behaviour (Collins, 1974). However, little research has been conducted on observer characteristics that might affect causal attribution.

According to Collins (1974), many of the conclusions an observer reaches are influenced more by the observer him- or herself than by the external stimulus, which is represented herein as the organization. One of these important personality traits is locus of control. People believe that events are controlled either by internal, personal forces of the actor or by external, situational forces of the environment (Lefcourt, 1966). For example, when a student has bad test results, he or she might think either that the teacher asked the wrong questions (external locus of control) or that he or she did not study enough (internal locus of control). Rotter (1966) states that although a person’s locus of control is not immediately connected to a preference for internal or external control in others, it may be correlated with the value the person places on either internal or external control.

If locus of control implies the preference for internality or externality, various strategies will be more or less preferred by internally versus externally focused people. For example, internally focused people would be likely to assign more value to organizations that take full responsibility for their own actions (rebuild strategies), a response indicating that the firm controls its own actions. Likewise, externally focused people would assign more value than internally focused people to organizations that reject responsibility for what happened (deny strategies), a response indicating that the firm believes that the environment controlled what happened more than the actions of the organization itself. Because diminish strategies are somewhere in between in terms of taking responsibility, we posit that such strategies are equally preferred by both externally and internally focused people. This leads to the following hypotheses:

H5a. When an organization uses deny strategies, people with an external locus of control perceive the firm’s reputation in a more positive light than people with an internal locus of control.

H5b. When an organization uses diminish strategies, people with an external locus of control and those with an internal locus of control have similar perceptions of the firm’s reputation.

H5c. When an organization uses rebuild strategies, people with an internal locus of control perceive the firm’s reputation in a more positive light than people with an external locus of control.

4. Methods

4.1. Design and stimuli

We used a 3 (crisis type) × 3 (crisis response) between-subjects factorial experimental design to investigate the hypothe- ses. We manipulated crisis type and crisis response strategy using various scenarios.

More specifically, we manipulated crisis type by the selection of one crisis from each of the three clusters (victim cluster, accidental cluster, preventable cluster). We chose the crisis types on the basis of a practical consideration—namely, being able to describe each crisis type realistically. From the victim cluster, we selected product tampering (Coombs, 2007; Coombs & Holladay, 2002). From the accidental cluster, we used technical-error product harm. From the preventable cluster, we chose organizational misdeed with injuries. We used a fictitious organization in the three scenarios to avoid potential confounding effects of pre-existing knowledge. We held the objective crisis damage stable in all situations, namely, by mentioning the death of two adults.

A.-S. Claeys et al. / Public Relations Review 36 (2010) 256–262 259

To manipulate the response strategy, two strategies were selected from each cluster. From the diminish cluster, we combined excuse and justification into one reaction (Coombs, 2007). From the rebuild cluster, we combined compensation and apology. We used only one crisis response strategy from the deny cluster because all three deny strategies tend to conflict. Combining strategies will only enhance their individual impacts when the responses are compatible (Huang, 2006). Therefore, we describe the scapegoat reaction for the deny scenario.

4.2. Participants and procedure

We conducted two pre-tests to ensure that the manipulations were effective. For the actual study, we collected data from 316 respondents using an online questionnaire. We randomly divided the respondents across the nine conditions and instructed them to read a scenario containing a combination of a crisis type with a crisis response. After reading the scenario, respondents filled in the questionnaire. Participants were Dutch-speaking Belgian men and women with an average age of 35 years (SD = 14.46; range = 13–70 years). Approximately 47% were male, and 53% were female.

4.3. Measures

We measured organizational responsibility for a crisis using the four-item 10-point Likert scale from the work of Griffin et al. (1992) (e.g., “How responsible was the organization with respect to the crisis?”) (˛ = .81).

We measured organizational reputation with a combination of the Reputation Quotient scale from the work of Fombrun, Gardberg, and Sever (2000) and McCroskey’s (cited in Coombs & Holladay, 1996) scale of credibility (˛ = .97).

We held crisis damage constant across scenarios by telling respondents that each crisis caused the death of two adults. To measure the possible effects of the perceived crisis severity, respondents answered one question on a 10-point scale, ranging from 1 (not at all severe) to 10 (very severe): “How severe do you consider the damage caused by this crisis?”

We measured locus of control using Rotter’s (1966) internal–external (I–E) locus-of-control scale. The I–E scale has a Cronbach’s ˛ of .73 and consists of nine factors with eigenvalues greater than 1.0. The two factors with the highest eigenvalues were labelled control over politics and misfortunes, and together they explained 23% of the total variance. The control-over- politics dimension considers the perceived control a person can have over politics. The misfortunes dimension considers the perceived causes of peoples’ misfortunes. Because the I–E scale has been shown to oversimplify the multidimensionality of the locus-of-control construct, these two dimensions were also taken separately into analyses (Duffy et al., 1977).

5. Pre-tests

We conducted two pre-tests to test the manipulated variables crisis response and crisis type.

5.1. Pre-test 1

Twelve respondents participated in an order task to check the internal validity of the crisis responses. The respondents were instructed to read each of the three crisis response scenarios and then to order them based on a list containing all five-crisis response strategies. Some neutral sentences were mistakenly interpreted as strategies. We removed these lines from the crisis responses used in the main experiment.

5.2. Pre-test 2

According to the SCCT typology (Coombs, 2007; Coombs & Holladay, 2002), we can manipulate crisis types through the assessment of crisis responsibility. Therefore, the second pre-test assessed (with a seven-point scale) the manipulation of corporate responsibility for each crisis type. Twenty-one respondents participated. The questionnaire used a within- subjects design, and all respondents saw the three crisis types. The results show that in the victim crisis (M = 4.31, SD = 1.49), respondents perceived the company as less responsible than in the accidental crisis (M = 5.31, SD = .98; t(40) = 2.56, p = .015) and the preventable crisis (M = 5.79, SD = 1.01; t(40) = 3.75, p = .001). However, there were no significant differences between the accidental crisis (M = 5.31, SD = .98) and the preventable crisis (M = 5.79, SD = 1.01; t(40) = 1.55, p = .13). We adapted all scenarios for the main study on the basis of these results.

6. Results

6.1. Manipulation checks

We performed a one-way analysis of variance (ANOVA) to check for the manipulation of crisis type, measured by the amount of crisis responsibility. The mean difference in crisis responsibility among the three crisis types is significant (F(2, 313) = 82.93, p < .001). The Scheffé follow-up procedure shows that the victim crisis (M = 4.78, SD = 2.11) results in signifi- cantly lower attributions of responsibility to the organization than the accidental crisis (M = 6.81, SD = 1.87, p < .001) and the

260 A.-S. Claeys et al. / Public Relations Review 36 (2010) 256–262

Fig. 1. Interaction between crisis response cluster and crisis cluster.

preventable crisis (M = 8.07, SD = 1.62, p < .001). Furthermore, the accidental crisis results in significantly lower attributions of crisis responsibility than the preventable crisis (MAccidental crisis = 6.81, SD = 1.87; MPreventable crisis = 8.07, SD = 1.62, p < .001).

6.2. Impact of crisis type and response strategy on reputation

To address the main effects of crisis types (H1) and response strategies (H2) on organizational reputation and to test the interaction effect of crisis type and crisis response strategies predicted in H3, we use a univariate two-way ANOVA (general linear model). Two main effects occur. Both crisis type (F(8, 307) = 94.72, p < .001) and crisis response strategy (F(8, 307) = 4.68, p = .01) have a significant main effect on organizational reputation (see Fig. 1). To detect the significant differences between the three conditions of crisis type and crisis response pairwise, we also conduct separate one-way ANOVA’s with Scheffé follow-up tests. This enables us to assess H1 and H2.

The results of a separate one-way ANOVA (F(2, 313) = 95.33, p < .001) show that respondents perceive the organization’s reputation less favourably in the case of a preventable crisis (M = 2.72, SD = 1.42) than in the cases of both the victim crisis (M = 5.95, SD = 1.97, p < .001) and the accidental crisis (M = 5.48, SD = 2.04, p < .001). No significant difference in reputation occurs between the victim crisis (M = 5.95, SD = 1.97) and the accidental crisis (M = 5.48, SD = 2.04, p = .18). These results partially support H1. Preventable crises, in which the organization is responsible for a crisis, have the most negative impact on organizational reputation.

The findings also confirm the main effect of crisis response strategy on organizational reputation (F(2, 313) = 4.52, p = .012). The Scheffé procedure reveals that rebuild strategies (M = 5.26, SD = 2.36) significantly lead to a more positive reputation than diminish strategies (M = 4.35, SD = 2.24, p = .018). The difference between the rebuild strategies and the deny strategies is only marginally significant (p < .1; MRebuild strategies = 5.26, SD = 2.36; MDeny strategies = 4.57, SD = 2.28, p = .087). Therefore, H2 is partially supported. Respondents assess the reputation of organizations using rebuild strategies (i.e., when firms accept responsibility) more positively than the reputation of organizations using diminish strategies (i.e., when organizational responsibility is minimized).

The interaction effect of crisis type and crisis response strategies on reputation is not significant (F(8, 307) = 1.28, p = .28). Therefore, H3 is not supported. Crisis responses that are matched to crisis type, in accordance with the SCCT recommendation, do not differ in their impact on organizational reputation compared with mismatched responses.

To investigate the impact of the perceived severity of the crisis on organizational reputation, we calculate a correla- tion coefficient. There is a significant, negative correlation between the severity of the crisis and organizational reputation (r(314) = −.12, p = .034). These results show that the more severe respondents perceive a crisis to be, the worse the reputation of the organization confronted with the crisis will be, in support of H4.

6.3. Moderating influence of locus of control

We tested H5a to H5c with a univariate two-way ANOVA (general linear model). A significant interaction effect occurred between crisis response strategy and one of the factors in the locus-of-control scale—namely, misfortunes (F(5, 310) = 3.82, p = .023). The plot in Fig. 2 shows that when a company uses a deny response strategy, its reputation is perceived more positively by respondents with an external locus of control (M = 5.30, SD = 2.15) than by respondents with an internal locus of control (M = 3.90, SD = 2.20; t(105) = 3.33, p = .001). Therefore, H5a is supported. However, when an organization uses diminish strategies, its reputation is perceived as equally positive by both externally focused respondents (M = 4.27, SD = 1.91)

A.-S. Claeys et al. / Public Relations Review 36 (2010) 256–262 261

Fig. 2. Interaction between crisis response cluster and locus of control factor.

and internally focused respondents (M = 4.44, SD = 2.53; t(99) = .38, p = .70), in support of H5b. We do not find support for H5c—rebuild strategies do not lead to a less negative reputation for respondents with an internal locus of control compared with respondents with an external locus of control. The other dimensions of locus of control do not offer significant results.

7. Conclusions and discussion

The SCCT advises communication managers to select crisis response strategies that match the specific crisis type (Coombs & Holladay, 2002). However, matching crisis types and crisis responses does not lead to more positive perception of firm reputation than mismatches. This contradicts Coombs and Holladay’s (1996) findings. An explanation for this might be the difference in methodology. Coombs and Holladay (1996) use a within-subjects design, whereas we use a between-subjects design. Another explanation may be that we combine two response strategies into one response, whereas Coombs and Holladay (1996) only use one strategy.

However, we find two main effects for crisis type and crisis response in relation to organizational reputation. First, corporate reputation is least favourable when organizations are confronted with a preventable crisis. In this type of crisis, companies are considered responsible for the crisis (Coombs, 2007; Coombs & Holladay, 2002). We find no significant difference in reputational perceptions between the accidental crisis and the victim crisis. Second, there is a main effect of crisis response strategy on organizational reputation. Specifically, the reputations of organizations using rebuild crisis response strategies are more positive than the reputations of organizations using diminish strategies. The results also provide evidence for the relationship between the perceived severity of the crisis and organizational reputation, confirming Coombs and Holladay’s (2002) assumption. Specifically, the more severe people perceive a crisis to be, the more negative are there perceptions of the organization’s reputation.

Locus of control (more specifically, the misfortunes factor) as a personality trait has a moderating impact on the effect of response strategy on reputation. For the deny strategy, respondents with an external locus of control perceive the corporate reputation more positively than respondents with an internal locus of control. Because diminish strategies attempt to take on some of the responsibility, while minimizing part of it, such strategies lead to similar results for both internally focused respondents and externally focused respondents. However, contrary to our expectations, respondents with an external locus of control do not prefer the use of rebuild strategies less than those with an internal locus of control. Further research is needed to explore this unexpected result.

8. Limitations and further research

The limitations of the current study provide some possibilities for further research. First, we measured reputation after only one exposure to a fictitious company. Therefore, it can be argued that the reputation measure resembles more of an attitude rather than a reputation developed over time. Second, our use of the I–E locus-of-control scale might be questioned. Despite its popularity, there has been evidence that it may oversimplify the actual dimensionality of the construct (Duffy et al., 1977). Third, the current study does not incorporate a condition in which the company does not react to the crisis. This would be worthwhile to measure in further research to detect the impact of a company’s reaction to a crisis versus a situation in which a company does not react in any way. Fourth, we based the selection of respondents on a convenience sample. Further research should use a more systematic procedure to select the respondents. In addition, the number of respondents could be increased. Fifth, further research should investigate whether there are different personality traits that influence the impact of crisis response strategies. Finally, the current study compared the impact of matching response strategies

262 A.-S. Claeys et al. / Public Relations Review 36 (2010) 256–262

and crisis types across the different clusters of each. It would be worthwhile to examine the impact of each of the different strategies within each cluster.

References

Arpan, L. M. (2002). When in Rome? The effects of spokesperson ethnicity on audience evaluation of crisis communication. Journal of Business Communication, 39(3), 314–340.

Benoit, W. L. (1995). Accounts, excuses, and apologies: A theory of image restoration strategies. Albany, NY: State University of New York Press. Benoit, W. L. (1997). Image repair discourse and crisis communication. Public Relations Review, 23(2), 177–186. Collins, B. E. (1974). Four components of the Rotter internal–external scale: Belief in a difficult world, a just world, a predictable world, and a politically

responsive world. Journal of Personality and Social Psychology, 29(3), 381–391. Coombs, W. T. (1998). An analytic framework for crisis situations: Better responses from a better understanding of the situation. Journal of Public Relations

Research, 10(3), 177–191. Coombs, W. T. (2004). West Pharmaceutical’s explosion: Structuring crisis discourse knowledge. Public Relations Review, 30(4), 467–473. Coombs, W. T. (2007). Protecting organization reputations during a crisis: The development and application of situational crisis communication theory.

Corporate Reputation Review, 10(3), 163–176. Coombs, W. T., & Holladay, S. J. (1996). Communication and attributions in a crisis: An experimental study in crisis communication. Journal of Public Relations

Research, 8(4), 279–295. Coombs, W. T., & Holladay, S. J. (2002). Helping crisis managers protect reputational assets: Initial tests of the situational crisis communication theory.

Management Communication Quarterly, 16(2), 165–186. Coombs, W. T., & Holladay, S. J. (2008). Comparing apology to equivalent crisis response strategies: Clarifying apology’s role and value in crisis communi-

cation. Public Relations Review, 34(3), 252–257. Duffy, P. J., Downey, R. G., & Shiflett, S. (1977). Locus of control: Dimensionality and predictability using Likert scales. Journal of Applied Psychology, 62(2),

214–219. Folkes, V. S. (1988). Recent attribution research in consumer behavior: A review and new directions. Journal of Consumer Research, 14(4), 548–565. Fombrun, C. J., Gardberg, N. A., & Sever, J. M. (2000). The reputation quotient sm: A multi-stakeholder measure of corporate reputation. Journal of Brand

Management, 7(4), 241–255. Griffin, M., Babin, B. J., & Darden, W. R. (1992). Consumer assessments of responsibility for product-related injuries: The impact of regulations, warnings,

and promotional policies. In J. F. Sherry, & B. Brian Sternthal@@Jr. (Eds.), Advances in consumer research, vol. 19 (pp. 870–878). Provo, UT: Association for Consumer Research.

Huang, Y. (2006). Crisis situations, communication strategies, and media coverage: A multicase study revisiting the communicative response model. Communication Research, 33(3), 180–205.

Huang, Y., Lin, Y., & Su, S. (2005). Crisis communicative strategies in Taiwan: Category, continuum, and cultural implication. Public Relations Review, 31(2), 229–238.

Lefcourt, H. M. (1966). Internal versus external control of reinforcement: A review. Psychological Bulletin, 65(4), 206–220. Rotter, J. B. (1966). Generalized expectancies for internal versus external control of reinforcement. Psychological Monographs, 80(1), 1–27. Spillan, J. E. (2003). An exploratory model for evaluating crisis events and managers’ concerns in non-profit organizations. Journal of Contingencies and Crisis

Management, 11(4), 160–169. Weiner, D. (2006). Crisis communications: Managing corporate reputation in the court of public opinion. Ivey Business Journal, 70(4), 1–6.