Leaders’ Use of Moral Justifications Increases Policy Support

Abstract

Leaders must choose how to justify their organization’s actions to stakeholders. We differentiate moral frames, or justifications based on moral values, from pragmatic frames, or justifications based on practical costs and benefits. In Experiments 1a and 1b, we found that moral policy frames elicited more support than pragmatic frames across a variety of scenarios. This effect was mediated by the perception that leaders who offer moral justifications possess relatively greater moral character. In Experiment 2, we found that perceptions of a leader’s private motives had a stronger influence on policy support than did the leader’s public stance. Experiment 3 demonstrated that, irrespective of how a policy was framed, people were most supportive of a policy championed by a leader high in moral character. In Experiment 4, we documented an additional benefit of moral policy frames: They allow leaders to mitigate the moral outrage generated by reneging on a policy.

Keywords

morality judgment values social perception policy making open materials open data preregistered

William Clay Ford, Jr., the former CEO of Ford Motor Company, wrote that “creating a strong business and building a better world are not conflicting goals—they are both essential ingredients for long-term success” (Ford, 2010). Indeed, leaders seen as ethical are also perceived as effective and elicit commitment from their followers (Brown, Treviño, & Harrison, 2005; Mayer, Aquino, Greenbaum, & Kuenzi, 2012; Mayer, Kuenzi, Greenbaum, Bardes, & Salvador, 2009).

Yet by focusing on what leaders do at the expense of considering why they do it, prior research has neglected the impact of how leaders frame or justify their decisions. Many policies have both economic and ethical implications; leaders like William Clay Ford, Jr., must decide how to justify them to stakeholders. Stakeholders’ support of policies affects their willingness to sacrifice for them, take action for them, and join organizations supporting them (Kahneman & Knetsch, 1992; Kahneman, Ritov, Jacowitz, & Grant, 1993; McGraw, Schwartz, & Tetlock, 2012). Therefore, obtaining the support of stakeholders both inside and outside the organization is critical for the success of any policy. However, little research has compared how outside support for an organization’s actions is affected by its leaders’ justifications for those actions. We attempt to do so here.

We capitalize on the distinction between moral frames and pragmatic frames offered by Kreps and Monin (2011). Whereas moral frames justify policies on the basis of moral values, pragmatic frames justify policies on the basis of economic and organizational benefits. For an illustration of the distinction between these two means of justifying a policy, consider a CEO’s plan to provide employees with free, healthy meals. On the one hand, the CEO could justify the policy on the basis of a moral obligation to care for employees’ health (a moral frame). On the other hand, the CEO could explain that the availability of meals will motivate employees to work longer hours (a pragmatic frame).

We propose that moral frames will enhance support for leaders and their policies relative to pragmatic frames. The moral-character account hypothesizes that leaders’ justifications signal something about their moral character. Scholars have suggested that leaders with high moral character may be perceived as more effective and persuasive than those with low moral character (e.g., Den Hartog, House, Hanges, Ruiz-Quintanilla, & Dorfman, 1999; Lowe, Kroeck, & Sivasubramaniam, 1996). Because actors’ intentions are fundamental to making sense of their actions (Knobe, 2003), moral justifications may signal moral character (cf. Sripada, 2012) and rally more support than pragmatic justifications.

We contrast this account with a more straightforward moralization account. Moralization involves the attachment of moral significance to an issue (Rozin, 1999). When an issue acquires moral weight, it grows in perceived importance (Rozin, Markwith, & Stoess, 1997), and people become resistant to considering trade-offs that render the issue economically fungible (Tetlock, 2003; Tetlock, Kristel, Elson, Green, & Lerner, 2000). Although this account suggests that moral frames may generate more support than pragmatic frames because they induce people to moralize policies, it also suggests that perceptions of leaders’ moral character should be relatively unimportant drivers of policy support.

Experiment 1

In Experiments 1a and 1b, we compared the impact of moral frames and pragmatic frames on policy support. In Experiment 1b, we replicated and extended Experiment 1a by including a measure of moralization along with a control measure to investigate whether any effect of policy frame on policy support was driven by particular frames (i.e., pragmatic frames) making policies appear unethical or immoral as opposed to amoral.

Pilot testing

Before conducting Experiment 1, we ran a pilot study to validate a measure of policy support. On the basis of an intuition about a likely effect size, we aimed to enroll 100 participants via Amazon Mechanical Turk and wound up with 107. They were randomly assigned to read either moral or pragmatic justifications for six hypothetical policies. Participants then evaluated the policies by responding to four items that we adapted from measures of policy support used in prior research (McGraw et al., 2012; McGraw & Tetlock, 2005; Staw & Ross, 1980). For each scenario, participants indicated their attitudes toward the leader (very negative to very positive), perception of the leader’s plan (immoral to moral), rating of the leader’s performance (very poor to outstanding), and the extent to which they agree with the leader’s plan (not at all to very strongly). All items were scored on 7-point scales (from 1 to 7) on which higher numbers indicated more support. A factor analysis found that all items loaded onto a single factor (α = .96), so we averaged all four items into a single measure. Policies framed in moral terms (M = 5.77, SD = 0.64) generated more support than those framed in pragmatic terms (M = 5.10, SD = 0.92), t(105) = 4.44, p < .001, d = 0.87.

Experiment 1a

Participants

Aiming to exceed Simmons, Nelson, and Simonsohn’s (2013) recommendation for confirming the existence of an effect using 2.5 times the sample size of a prior study, we enrolled 385 participants via Amazon Mechanical Turk. They participated in a survey on managerial decision making in exchange for $1.25. Among those participants, 374 successfully passed two reading comprehension checks. The remaining 11 participants were prevented from continuing with the study, so we did not collect any dependent measures from them. Participants’ mean age was 35 years (SD = 12.37), and 50% of them were female.

Procedure

To examine the effect of policy frames across both the private and the public sector, we used a 3 (policy frame: moral, pragmatic, or ambiguous) × 2 (policy sector: private or public) mixed design with policy frame manipulated between subjects and policy sector manipulated within subjects. We accounted for policy sector in our analyses to see whether any effect of policy frame depended on whether the organization was expected to accomplish communal goals (public sector) or to follow market-pricing norms (private sector; see McGraw et al., 2012). We preregistered the study design, including all measures and planned analyses, at Open Science Framework, https://osf.io/wptk5/.

Manipulations

Each participant read six policy proposals. Three came from the public sector (a politician’s plan to fund a retirement planning agency, a state governor’s plan to repave state highways, and a president’s plan to outlaw child labor in a developing country), and three came from the private sector (a CEO’s plan to invest in greener technology, a marketing executive’s plan to target sales of a new tablet PC toward public schools, and a CEO’s plan to provide healthy meals to employees). The presentation order of the scenarios was randomized for each participant.

In the framing manipulation, participants were randomly assigned to read about six scenarios that were all framed in either moral terms, pragmatic terms, or ambiguous terms. For example, participants who read about a politician’s policy to fund a retirement planning agency heard about the importance of ensuring that retirees “live with comfort and dignity” (moral frame), “do not drain public funds” (pragmatic frame), or “have sufficient funds” (ambiguous frame). Likewise, participants who read about a CEO’s plan to provide healthy meals for employees saw the plan justified in terms of either the organization’s moral imperative to look out for its employees’ well-being (“increased access to meals should improve our employees’ well-being”) or the organization’s self-interest of keeping employees productive (“increased access to meals should improve our employees’ productivity”). In the ambiguous frame, they saw a justification with ambiguous motives (“increased access to meals should improve the status-quo”). See the Supplemental Material available online for the full versions of each scenario.

Leader moral character

Individuals who are kind, compassionate, and caring are generally perceived to possess high moral character (Aquino & Reed, 2002). Therefore, we asked participants to indicate the extent to which the traits kind, compassionate, and caring described the leader in each scenario on a scale ranging from 1 (not at all) to 7 (a great deal). The items were averaged to form an index (α = .98).

Policy support

After reading each scenario, participants responded to the same measure of policy support that had been validated in the pilot study (α = .93).

Results

As described in our preregistered data-analysis plan, we used 3 (policy frame) × 2 (policy sector) mixed analyses of variance (ANOVAs) to evaluate the effects of policy support and each leader’s moral character. Finally, we conducted a mediation analysis.

Policy support

The results revealed a main effect of policy frame, F(2, 371) = 4.92, p = .008, η_p² = .03, replicating the results of the pilot study. Participants were more supportive of policies framed in moral terms (M = 5.82, SD = 0.74) than of policies framed in pragmatic terms (M = 5.54, SD = 0.72), t(246) = 3.08, p = .002, d = 0.39, 95% confidence interval (CI) = [0.14, 0.64]. Policy support in the ambiguous frame fell in the middle (M = 5.70, SD = 0.70) and did not significantly differ from either the moral frame condition or the pragmatic frame condition (both ps > .071). We also found an unexpected main effect of policy sector; participants were more supportive of policies in the private sector (M = 5.72, SD = 0.77) than of those in the public sector (M = 5.65, SD = 0.84), F(1, 371) = 4.09, p = .044, η_p² = .01, d = 0.21, 95% CI = [0.01, 0.41]. The interaction between policy frame and policy sector did not attain statistical significance, F(2, 371) = 2.86, p = .059.

Leader moral character

As expected, the results show a main effect of policy frame on leader moral character, F(2, 371) = 22.74, p < .001, η_p² = .11. Planned comparisons revealed that participants perceived leaders who framed issues in moral terms (M = 5.69, SD = 0.75) to possess greater moral character than leaders who framed issues in pragmatic terms (M = 4.96, SD = 1.05), t(246) = 6.30, p < .001, d = 0.80, 95% CI = [0.54, 1.06]. Furthermore, participants perceived leaders who framed issues in ambiguous terms (M = 5.35, SD = 0.73) to be less moral than leaders who framed issues in moral terms, t(246) = 3.57, p < .001, d = 0.46, 95% CI = [0.21, 0.71], but more moral than leaders who framed issues in pragmatic terms, t(250) = 3.49, p < .001, d = 0.44, 95% CI = [0.19, 0.69]. We also found a main effect of policy sector; participants perceived leaders in the private sector to possess less moral character (M = 5.23, SD = 0.99) than leaders in the public sector (M = 5.42, SD = 0.96), F(1, 371) = 23.66, p < .001, η_p² = .06, d = 0.50, 95% CI = [0.29, 0.71]. We did not find evidence of an interaction between policy frame and policy sector, F(2, 371) = 1.25, p > .250.

Mediation analysis

We assessed whether, across all policies, leader moral character mediated participants’ greater support of policies framed in moral terms than of policies framed in pragmatic terms. When regressing policy support on leader moral character and policy frame, we found that leader moral character continued to predict policy support, β = 0.84, t(247) = 20.43, p < .001, but the effect of policy frame reversed in direction (from β = 0.19 to β = −0.12), t(247) = 2.93, p = .004. A bootstrap procedure with 10,000 replications revealed an indirect effect of leader moral character, z = 6.02, p < .001, indirect effect = 0.47, 95% CI = [0.33, 0.60]. Taken together, these results suggest that the power of moral frames is attributable to perceptions of leaders’ moral character. This effect held across policies in both the public and private sectors. We also note that when we controlled for the influence of leaders’ moral character, issues framed in pragmatic terms actually generated more support than issues framed in moral terms.

Experiment 1b

Participants

Given the inclusion of an additional control variable to this study and its potential to reduce statistical power, we planned to recruit 575 individuals via Amazon Mechanical Turk to complete a survey in exchange for $2.00. A total of 606 attempted to complete the study, and we collected dependent measures from 578 who successfully passed two reading comprehension checks. Their mean age was 35.26 years (SD = 11.30), and 47% of them were female.

Procedure

Like Experiment 1a, this experiment had a 3 (policy frame: moral, pragmatic, or ambiguous) × 2 (policy sector: private or public) mixed design with policy frame manipulated between subjects and policy sector manipulated within subjects. Unlike Experiment 1a, all measures were presented in a randomized order in Experiment 1b. We preregistered the study design, including all measures and planned analyses, at Open Science Framework, https://osf.io/wptk5/.

Manipulations

The framing manipulations were identical to those in Experiment 1a.

Leader moral character

We assessed perceptions of each leader’s moral character using the same measure as in Experiment 1a (α = .98).

Issue moralization

To understand whether participants’ moralization of issues might explain the impact of policy frame on policy support, we presented participants with four items assessing the extent to which they viewed the issue addressed by a particular policy as being morally significant. Specifically, we informed participants that a policy they had just read about was relevant to a particular issue. We then asked them to indicate their agreement with four statements about the issue on a scale ranging from 1 (strongly disagree) to 7 (strongly agree): “There are moral implications associated with this issue,” “My stance on this issue is guided by my moral convictions,” “It would be foolish to attach moral significance to this issue,” and “This issue is no different from any other routine economic issue that leaders face on a daily basis.” Because the measure was intended to capture the extent to which a participant had moralized an issue, we reverse-scored the last two items. The four items were reliable (α = .83) and were thus averaged to form a single index.

Policy support

We measured policy support using the same items as in Experiment 1a (α = .94).

Policy ethicality

As a control variable, we asked participants to indicate their agreement with the statement “the policy is unethical” on a scale ranging from 1 (strongly disagree) to 7 (strongly agree). We reverse-scored this measure.

Results

In keeping with our preregistered data-analysis plan, we analyzed policy support, leader moral character, and issue moralization using 3 (policy frame) × 2 (policy sector) mixed ANOVAs. We also conducted a series of follow-up analyses of covariance that treated policy ethicality as a covariate. Finally, we conducted a mediation analysis using leader moral character and issue moralization as potential mediators. See Table 1 for the conditional means and standard deviations of all measures.

Table 1.

Results From Experiment 1b: Means and Standard Deviations for Each Policy-Frame Condition

Variable	Moral frame	Ambiguous frame	Pragmatic frame
Policy support	5.73_a (0.71)	5.67_a (0.70)	5.46_b (0.76)
Leader moral character	5.45_a (0.76)	5.27_b (0.79)	4.93_c (0.98)
Issue moralization	4.47_a (0.81)	4.42_a (0.87)	4.32_a (0.90)
Policy ethicality	6.13_a b (0.92)	6.24_a (0.90)	6.03_b (0.86)

Note: The table presents conditional means, with standard deviations in parentheses. Within a row, values with different subscripts differ to a statistically significant extent (p < .05).

Policy support

As in Experiment 1a, there was a main effect of policy frame, F(2, 575) = 7.51, p < .001, η_p² = .03. Participants were more supportive of policies framed in moral terms than of policies framed in pragmatic terms, t(383) = 3.62, p < .001, d = 0.37, 95% CI = [0.24, 0.50]. As in Experiment 1a, the level of policy support in the ambiguous-frame condition fell between those in the other policy-frame conditions but did not significantly differ from that in the moral-frame condition, t(385) = 0.80, p > .250. However, policy support in the ambiguous-frame condition did differ from that in the pragmatic-frame condition to a statistically significant degree, t(382) = 2.87, p = .004, d = 0.29, 95% CI = [0.16, 0.42]. As in Experiment 1a, participants were more supportive of policies in the private sector (M = 5.67, SD = 0.78) than policies in the public sector (M = 5.58, SD = 0.86), F(1, 575) = 9.34, p = .002, η_p² = .02, d = 0.25, 95% CI = [0.09, 0.41]. The interaction between policy frame and policy sector did not attain statistical significance, F(2, 575) = 1.28, p > .250.

When we controlled for the extent to which a policy was perceived as unethical, the effect of policy frame held, F(2, 574) = 6.33, p = .002, η_p² = .02. Thus, the effect of policy frame on policy support was not eliminated when we accounted for variation in a policy’s perceived ethicality, nor was the effect of policy sector, F(1, 574) = 6.51, p = .011, η_p² = .001, d = 0.21, 95% CI = [0.05, 0.37]. The Policy Frame × Policy Sector interaction remained nonsignificant when we controlled for policy ethicality, F(2, 574) = 1.17, p > .250.

Leader moral character

Replicating the pattern of results in Experiment 1a, we found a main effect of policy frame on leader moral character, F(2, 575) = 18.55, p < .001, η_p² = .06. Leaders who framed issues in moral terms were perceived to possess greater moral character than leaders who framed issues in pragmatic terms, t(383) = 5.82, p < .001, d = 0.60, 95% CI = [0.40, 0.80]. Furthermore, participants perceived leaders who framed issues in ambiguous terms to be less moral than leaders who framed issues in moral terms, t(246) = 3.57, p < .001, d = 0.46, 95% CI = [0.21, 0.71], but more moral than leaders who framed issues in pragmatic terms, t(385) = 2.33, p = .021, d = 0.24, 95% CI = [0.04, 0.44]. We also replicated the main effect of policy sector in Experiment 1a, in which participants perceived leaders in the private sector to possess less moral character (M = 5.14, SD = 1.02) than leaders in the public sector (M = 5.29, SD = 0.91), F(1, 575) = 18.88, p < .001, η_p² = .03, d = 0.65, 95% CI = [0.48, 0.82]. We did not find evidence of an interaction between policy frame and policy sector, F(2, 575) = 0.27, p > .250.

The effect of policy frame remained statistically significant when we included policy ethicality as a covariate, F(2, 574) = 17.55, p < .001, η_p² = .06, but the effect of policy sector did not, F(1, 574) = 2.22, p = .14, η_p² = .004, d = 0.12, 95% CI = [–0.04, 0.28]. When we controlled for the influence of policy ethicality, the Policy Frame × Policy Sector interaction remained nonsignificant, F(2, 574) = 0.15, p > .250.

Issue moralization

Although we did not identify a main effect of policy frame on issue moralization, F(2, 575) = 1.48, p = .228, we did find a Policy Frame × Policy Sector interaction, F(2, 575) = 6.75, p = .001, η_p² = .02. For policies in the public sector, we found a main effect of policy frame on issue moralization, F(2, 575) = 3.10, p = .046. Participants ascribed greater moral significance to public-sector policies framed in moral terms (M = 4.67, SD = 0.81) than to public-sector policies framed in pragmatic terms (M = 4.50, SD = 0.86), t(383) = 2.08, p = .038, d = 0.21, 95% CI = [0.01, 0.41]. Although participants did not moralize issues framed in ambiguous terms (M = 4.50, SD = 0.86) any more than they did issues framed in pragmatic terms, t(382) = 0.12, p > .250, they did moralize issues framed in ambiguous terms to a lesser extent than they did issues framed in moral terms, t(385) = 2.26, p = .025, d = 0.23, 95% CI = [0.03, 0.43]. In contrast, we did not identify evidence of a framing effect on private-sector issue moralization, F(2, 575) = 1.98, p = .139. We also found a main effect of policy sector such that participants moralized issues in the public sector (M = 4.55, SD = 0.84) to a greater extent than they did issues in the private sector (M = 4.25, SD = 1.04), F(1, 575) = 82.18, p < .001, η_p² = .13, d = 0.76, 95% CI = [0.59, 0.93]. The results imply that moral policy frames may cause people to moralize public-sector issues more than private-sector issues.

When we controlled for the extent to which participants perceived a policy as unethical, the Policy Frame × Policy Sector interaction held, F(2, 574) = 7.15, p < .001, η_p² = .02, but the main effect of policy sector did not, F(1, 574) = 0.05, p > .250, η_p² < .001, d = 0.02, 95% CI = [–0.14, 0.18]. We again failed to find evidence of a main effect of policy frame on issue moralization, F(2, 574) = 1.49, p = .227.

Mediation analysis

We wanted to understand whether differences in leaders’ perceived moral character accounted for the relatively greater support for policies framed in moral terms rather than pragmatic terms above and beyond any potential impact of issue moralization. Thus, we conducted a mediation analysis treating leader moral character and issue moralization as potential mediators. As illustrated by Figure 1, when we regressed policy support on leader moral character, issue moralization, and policy frame, the main effect of policy frame was reduced to nonsignificance (from β = 0.18 to β = −0.04), t(381) = 1.25, p = .212. Furthermore, both leader moral character, t(381) = 21.51, p < .001, and issue moralization, t(381) = 4.55, p < .001, predicted policy support. As in Experiment 1a, a bootstrap procedure with 10,000 replications revealed an indirect effect of leader moral character, z = 6.08, p < .001, indirect effect = 0.32, 95% CI = [0.21, 0.43]. However, the indirect effect of issue moralization did not attain statistical significance, z = 1.49, p = .136, indirect effect = 0.02, 95% CI = [−0.004, 0.04]. In a follow-up analysis that included policy ethicality as a covariate, the indirect effect of leader moral character held, z = 5.86, p < .001, indirect effect = 0.29, 95% CI = [0.18, 0.40], whereas the indirect effect of issue moralization continued to be nonsignificant, z = 1.53, p = .125, indirect effect = 0.02, 95% CI = [−0.004, 0.04]. Overall, Experiment 1b replicated the findings of Experiment 1a, which suggests that the relatively greater support elicited by moral frames relative to pragmatic frames could be attributed to their ability to signal leaders’ moral character. In addition, this set of analyses suggests that these effects were not merely a byproduct of participants moralizing issues framed in moral terms. We found limited evidence that participants’ moralization of policies could explain participants’ greater support for policies framed in moral as opposed to pragmatic terms; this does not lend support to the moralization account. Instead, we found that perceptions of leaders’ moral character uniquely mediated the link between policy frames and policy support in a manner that was robust to a policy’s perceived ethicality.

Fig. 1.

Results from Experiment 1b: mediation model showing the effect of policy frame on policy support, as mediated by leader moral character and issue moralization. Along the middle path, the value below the arrow shows the total effect, and the value above the arrow shows the direct effect after we controlled for the mediators. Asterisks indicate significant path coefficients (**p < .001). Policy frame was coded as 1 for moral frames and 0 for pragmatic frames.

Experiment 2

In Experiment 2, we explored whether it is more important that a leader take a public moral stance or believe it privately. If the relative effectiveness of moral policy frames is driven by perceptions of leader moral character, public justifications for a policy should be relatively inconsequential when private motives are transparent.

Participants

With the goal of obtaining enough participants to detect a private-frame main effect of d = 0.39 (the effect size obtained in Experiment 1a) with at least 80% power, we recruited 262 participants via Amazon Mechanical Turk to complete a survey on managerial decision making in exchange for $0.50. Ten participants failed a reading comprehension test and were not allowed to complete the dependent measures, which left a final sample of 252. Participants’ mean age was 29.76 years (SD = 8.49), and 33% of them were female.

Procedure

Participants read an adapted version of the “healthy meals for employees” policy used in Experiments 1a and 1b. We manipulated a leader’s private and public policy frames in a 2 (private frame: moral or pragmatic) × 2 (public frame: moral or pragmatic) between-subjects design. We preregistered the study design, including all measures and planned analyses, at Open Science Framework, https://osf.io/wptk5/.

For the private-frame manipulation, participants read a paragraph about a CEO’s goal to improve either employee well-being (moral frame) or productivity (pragmatic frame). The paragraph explained the motives behind the policy (the first set of words in brackets was used in the moral-frame condition, and the second set was used in the pragmatic-frame condition):

Looking to improve [employee well-being/productivity], the CEO of a large company assigned a task force to identify solutions that could [encourage employees to adapt a healthy lifestyle while improving their quality of life/motivate employees to work longer hours in the office while taking fewer sick days]. Ultimately, the task force recommended a plan where full-time chefs would be hired to provide free, healthy meals to employees throughout the day. Convinced that this plan would [make for a healthier workforce/improve the company’s bottom line], the CEO decided to adopt the plan immediately.

Participants then learned about the CEO’s public justification for the decision. This served as the public-frame manipulation. The following paragraph explained the public justification (the first set of words in brackets was used in the moral-frame condition, and the second set was used in the pragmatic-frame condition):

In a company-wide announcement, the CEO proclaimed that “increased access to meals should prevent our employees from engaging in unhealthy dietary habits. By ensuring that our employees eat healthier food, we can [help them lead healthier lives/ensure they are more productive]. Easy access to healthy meals should improve our [employees’ well-being/company’s bottom line].”

Leader moral character

After reading about the policy, participants indicated the extent to which nine different traits described the CEO. The traits were selected from Aquino and Reed (2002), who found them to be associated with moral identity. In addition to the three traits used in Experiments 1a and 1b (i.e., kind, compassionate, and caring), participants indicated the extent to which the traits honest, helpful, hardworking, friendly, generous, and fair described the CEO on a scale ranging from 1 (not at all) to 7 (a great deal). The nine items were reliable (α = .95) and were averaged into a single measure.

Policy support

We used the same measure of policy support as in Experiments 1a and 1b (α = .91).

Results

We analyzed leader moral character and policy support using 2 (private frame) × 2 (public frame) ANOVAs before conducting a mediation analysis.

Policy support

As expected, the private-frame manipulation affected policy support. Participants were more supportive of the policy when the leader’s private motivation was moral (M = 6.00, SD = 1.03) rather than pragmatic (M = 5.33, SD = 1.18), F(1, 248) = 23.06, p < .001, η_p² = .09, d = 0.72, 95% CI = [0.47, 0.97]. As for our manipulation of the leader’s public stance, neither its main effect (moral public frame: M = 5.77; pragmatic public frame: M = 5.55), F(1, 248) = 2.07, p = .151, nor its interaction with private motives, F(1, 248) = 2.89, p = .091, attained statistical significance.

Leader moral character

We also identified a main effect of private frame such that participants perceived the CEO whose private motives were moral (M = 5.64, SD = 1.00) to possess greater moral character than the CEO whose motives were pragmatic (M = 5.07, SD = 1.13), F(1, 248) = 17.45, p < .001, η_p² = .07, d = 0.53, 95% CI = [0.28, 0.78]. As for the leader’s public stance, neither its main effect (moral public frame: M = 5.44; pragmatic public frame: M = 5.26), F(1, 248) = 1.57, p = .212, nor its interaction with private frame attained significance, F(1, 248) = 0.38, p > .250.

Mediation analysis

As illustrated by Figure 2, after we regressed policy support on leader moral character and private-frame condition, leader moral character continued to predict policy support, t(249) = 17.22, p < .001. Although the private-frame condition continued to predict policy support, t(249) = 2.40, p = .017, the effect reduced in magnitude (from β = 0.29 to β = 0.10). A bootstrapping procedure with 10,000 replications revealed an indirect effect of leader moral character, z = 4.09, p < .001, indirect effect = 0.43, 95% CI = [0.24, 0.64]. Overall, the analysis suggested that leader moral character partially mediated the relatively greater support generated by the private moral frame. Consistent with the moral-character account, we found that the benefits of moral frames were contingent on the perception that a leader was privately motivated by moral values.

Fig. 2.

Results from Experiment 2: mediation model showing the effect of private frame on policy support, as mediated by leader moral character. Along the bottom path, the value below the arrow shows the total effect, and the value above the arrow shows the direct effect after we controlled for the mediator. Asterisks indicate significant path coefficients (*p < .05, **p < .001). Private policy frame was coded as 1 for moral frames and 0 for pragmatic frames.

Experiment 3

In Experiment 3, we aimed to directly test whether moral policy frames promoted policy support via their ability to signal a leader’s moral character. We hypothesized that people would lend more support to leaders of upstanding moral character, regardless of whether the leader framed policies in moral or pragmatic terms.

Participants

We sought to obtain a sample size similar to that obtained in Experiment 2 (a minimum of 250 participants). We wound up getting 254 participants via Amazon Mechanical Turk. Of those, 251 passed a comprehension check and completed all dependent measures in exchange for $0.50. Their mean age was 32.46 years (SD = 11.10), and 36% of them were female.

Procedure

We adapted the materials from Experiment 2 for a 2 (moral character: high or low) × 2 (policy frame: moral or pragmatic) between-subjects design. We preregistered the study design, including all measures and planned analyses, at Open Science Framework, https://osf.io/wptk5/. Participants followed the same procedure as in Experiment 2 with two exceptions: First, we did not assess their perceptions of a leader’s moral character because it was directly manipulated. Second, we adapted the first paragraph of the scenario to include a manipulation of the leader’s moral character. To manipulate the leader’s moral character, we used a high-moral-character condition in which we described the leader using the three adjectives from the moral-character measure used in Experiments 1a and 1b. In the low-moral-character condition, we used antonyms of these adjectives. These adjectives were inserted into the following paragraph (the first set of words in brackets was used in the high-moral-character condition, and the second set was used in the low-moral-character condition):

The CEO of a large company has been described by employees as [kind, compassionate, and caring/inconsiderate, cruel, and neglectful]. Recently, the CEO devised a plan for the company to hire full-time chefs to provide free, healthy meals to employees throughout the day.

The second paragraph that participants read was the same as the second paragraph in Experiment 2. This paragraph contained the same manipulation of policy frame as the public-frame manipulation in Experiment 2. After the participants read the scenario, they were presented with the same measure of policy support used in Experiments 1a, 1b, and 2.

Results

In keeping with our preregistered data-analysis plan, we analyzed policy support using a 2 (leader moral character) × 2 (policy frame) ANOVA. As hypothesized, we found a main effect of leader moral character such that participants were more supportive of the leader’s policy when the leader was described as high in moral character (M = 6.18, SD = 1.13) as opposed to low in moral character (M = 5.24, SD = 1.08), F(1, 247) = 45.82, p < .001, η_p² = .16, d = 0.86, 95% CI = [0.60, 1.12]. Neither the main effect of policy frame (moral frame, M = 5.73; pragmatic frame, M = 5.67), F(1, 247) = 0.16, p > .250, nor its interaction with leader moral character attained significance, F(1, 247) = 0.47, p > .250. Irrespective of how the policy was framed, participants’ support for it was largely contingent on whether they perceived the leader as being high or low in moral character.

Experiment 4

Reneging on a policy commitment tends to generate disapproval (Staw & Ross, 1980) that manifests itself in the form of moral outrage when it renders moral values economically fungible (Tetlock et al., 2000). However, if perceptions of a leader’s moral character dictate policy support, then framing an issue in moral terms should signal that the leader possesses ethical motives—even if the policy is abandoned. In contrast, if policy support is driven by issue moralization, then a leader’s decision to renege on a policy framed in moral terms should elevate moral outrage. Therefore, the moral-character account would predict that leaders should generate less moral outrage after reneging on a policy framed in moral terms, whereas the moralization account would predict the opposite. In Experiment 4, we aimed to pit these accounts against each other.

Participants

With the goal of obtaining enough participants to have at least 80% power to detect a main effect of a policy’s initial framing given an effect size (d) of 0.3, we recruited 371 participants via Amazon Mechanical Turk to complete a survey on managerial decision making in exchange for $0.75. Seventeen participants failed a comprehension check and were prevented from responding to the study’s dependent measure, which resulted in a final sample of 354 participants from whom we collected data. Their mean age was 30.88 years (SD = 10.80), and 37% of them were female.

Procedure

We randomly assigned participants to a 2 (policy frame: moral or pragmatic) × 2 (postabandonment frame: moral or ambiguous)¹ between-subjects design. Participants read about a nascent clothing company known for developing young, talented designers. After the company began to achieve financial stability, its CEO decided to increase designers’ salaries.

Policy-frame manipulation

We manipulated policy frame by altering the manner in which the CEO justified the planned salary increase for employees. In the moral-frame condition, the CEO made an appeal to fairness by stating that “it is only fair for us to make sure that their talent and dedication is recognized the way it deserves to be.” In the pragmatic-frame condition, the CEO instead emphasized the benefits to the organization: “We need to pay them better to keep them productive and likely to stay with our company for the long term.”

After this explanation, all participants read that, in light of unanticipated economic difficulties, the CEO was unable to raise salaries. See the Supplemental Material for the full versions of each scenario.

Moral outrage

Our key dependent measure was participants’ moral outrage in response to the CEO’s decision to renege on the prior commitment. We used an index (developed by Tetlock et al., 2000) composed of six bipolar items rated on 7-point scales (from 1 to 7): bad-good, wise-foolish, negative-positive, immoral-moral, unfair-fair, disgusted–not at all disgusted. With the exception of wise-foolish, all items were reverse-scored so that higher scores indicated more outrage (α = .91).

Results

We conducted a 2 × 2 ANOVA and identified a main effect of policy frame such that the CEO’s decision to abandon the planned salary increase was met with less moral outrage when the policy had originally been framed in moral terms (M = 3.32, SD = 1.22) as opposed to pragmatic terms (M = 3.67, SD = 1.33), F(1, 350) = 6.66, p = .010, η_p² = .02, d = 0.28, 95% CI = [0.07, 0.49]. Neither the main effect of postabandonment frame (moral postabandonment frame: M = 3.50; pragmatic postabandonment frame: M = 3.50), F(1, 350) = 0.002, p > .250, nor its interaction with policy frame, F(1, 350) = 0.74, p > .250, attained statistical significance. Results were consistent with the moral-character account and counter to the moralization account; we found that use of a moral policy frame somewhat insulated a leader from moral outrage—even after the leader reneged on a commitment.

General Discussion

Some economists have argued that a corporation’s prime directive should be to maximize profits and that the pursuit of any other goal, including contributing to the broader welfare, is wrong (Friedman, 1962). However, not everyone shares this prime economic directive. Our results suggest that, at least for our research participants, the perceived morality of organizational leaders was a powerful determinant of their support for policies championed by those leaders. The current results are consistent with prior findings that actions are often judged on the basis of actors’ moral character (Nadler & McDonnell, 2012). However, the current findings also suggest that despite people’s attempts to be vigilant against hypocrisy (Valdesolo & DeSteno, 2008; Wagner, Lutz, & Weitz, 2009), they will overlook inconsistent framing and broken commitments that appear to be motivated by moral concerns.

Footnotes

Acknowledgements

We thank Phil Tetlock for his inspiration early in the life of this project.

Declaration of Conflicting Interests

The authors declared that they had no conflicts of interest with respect to their authorship or the publication of this article.

Funding

This work was funded by the Dean’s Office at the Haas School of Business, University of California, Berkeley.

Supplemental Material

Additional supporting information can be found at

Open Practices

All data and materials have been made publicly available via Open Science Framework and can be accessed at https://osf.io/c2gaf and https://osf.io/2htuq, respectively. The plans for the experiments were preregistered at Open Science Framework (Experiments 1a and 2: https://osf.io/qur68; Experiment 1b: https://osf.io/ixfh4; Experiment 3: https://osf.io/wptk5). The complete Open Practices Disclosure for this article can be found at http://pss.sagepub.com/content/by/supplemental-data. This article has received badges for Open Data, Open Materials, and Preregistration. More information about the Open Practices badges can be found at https://osf.io/tvyxz/wiki/view/ and .

Notes

References

Aquino

Reed

II . (2002). The self-importance of moral identity. Journal of Personality and Social Psychology, 83, 1423–1440.

Brown

M. E.

Treviño

L. K.

Harrison

D. A.

(2005). Ethical leadership: A social learning perspective for construct development and testing. Organizational Behavior and Human Decision Processes, 97, 117–134.

Den Hartog

D. N.

House

R. J.

Hanges

P. J.

Ruiz-Quintanilla

S. A.

Dorfman

P. W.

(1999). Culture specific and cross-culturally generalizable implicit leadership theories: Are attributes of charismatic/transformational leadership universally endorsed? Leadership Quarterly, 10, 219–256.

Ford

W. C.

Jr. (2010). Letter from William Clay Ford, Jr. (Sustainability Report 2009/10). Retrieved from http://web.archive.org/web/20150327030615/http://ophelia.sdsu.edu:8080/ford/09-05-2011/microsites/sustainability-report-2009-10/overview-letter-ford.html

Friedman

(1962). Capitalism and freedom. Chicago, IL: University of Chicago Press.

Kahneman

Knetsch

J. L.

(1992). Valuing public goods: The purchase of moral satisfaction. Journal of Environmental Economics and Management, 22, 57–70.

Kahneman

Ritov

Jacowitz

K. E.

Grant

(1993). Stated willingness to pay for public goods: A psychological perspective. Psychological Science, 4, 310–315.

Knobe

(2003). Intentional action in folk psychology: An experimental investigation. Philosophical Psychology, 16, 309–324.

Kreps

T. A.

Monin

(2011). “Doing well by doing good”? Ambivalent moral framing in organizations. Research in Organizational Behavior, 31, 99–123.

10.

Lowe

K. B.

Kroeck

K. G.

Sivasubramaniam

(1996). Effectiveness correlates of transformational and transactional leadership: A meta-analytic review of the MLQ literature. Leadership Quarterly, 7, 385–425.

11.

Mayer

D. M.

Aquino

Greenbaum

R. L.

Kuenzi

(2012). Who displays ethical leadership, and why does it matter? An examination of antecedents and consequences of ethical leadership. Academy of Management Journal, 55, 151–171.

12.

Mayer

D. M.

Kuenzi

Greenbaum

Bardes

Salvador

(2009). How low does ethical leadership flow? Test of a trickle-down model. Organizational Behavior and Human Decision Processes, 108, 1–13.

13.

McGraw

A. P.

Schwartz

J. A.

Tetlock

P. E.

(2012). From the commercial to the communal: Reframing taboo trade-offs in religious and pharmaceutical marketing. Journal of Consumer Research, 39, 157–173.

14.

McGraw

A. P.

Tetlock

P. E.

(2005). Taboo trade-offs, relational framing, and the acceptability of exchanges. Journal of Consumer Psychology, 15, 2–15.

15.

Nadler

McDonnell

M.-H.

(2012). Moral character, motive, and the psychology of blame. Cornell Law Review, 97, 255–304.

16.

Rozin

(1999). The process of moralization. Psychological Science, 10, 218–221.

17.

Rozin

Markwith

Stoess

(1997). Moralization and becoming a vegetarian: The transformation of preferences into values and the recruitment of disgust. Psychological Science, 8, 67–73.

18.

Simmons

J. P.

Nelson

L. D.

Simonsohn

(2013, January). Life after p-hacking. Paper presented at the annual meeting of the Society for Personality and Social Psychology, New Orleans, LA.

19.

Sripada

C. S.

(2012). Mental state attributions and the side-effect effect. Journal of Experimental Social Psychology, 48, 232–238.

20.

Staw

B. M.

Ross

(1980). Commitment in an experimenting society: A study of the attribution of leadership from administrative scenarios. Journal of Applied Psychology, 65, 249–260.

21.

Tetlock

P. E.

(2003). Thinking the unthinkable: Sacred values and taboo cognitions. Trends in Cognitive Sciences, 7, 320–324.

22.

Tetlock

P. E.

Kristel

O. V.

Elson

S. B.

Green

M. C.

Lerner

J. S.

(2000). The psychology of the unthinkable: Taboo trade-offs, forbidden base rates, and heretical counterfactuals. Journal of Personality and Social Psychology, 78, 853–870.

23.

Valdesolo

DeSteno

(2008). The duality of virtue: Deconstructing the moral hypocrite. Journal of Experimental Social Psychology, 44, 1334–1338.

24.

Wagner

Lutz

R. J.

Weitz

B. A.

(2009). Corporate hypocrisy: Overcoming the threat of inconsistent corporate social responsibility perceptions. Journal of Marketing, 73(6), 77–91.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.15 MB

0.01 MB