Abstract
Four experiments and a correlational study explored the relationship between power and perspective-taking.
In Experiment 1, participants primed with high power compared to low power were more likely to draw an “Ε” on their forehead in a self-oriented direction, indicating less of an inclination to spontaneously adopt another’s visual perspective.
In Experiments 2a and 2b, high-power participants were less likely to take into account that others did not possess their privileged knowledge, suggesting that power leads individuals to anchor too heavily on their own vantage point, insufficiently adjusting to others’ perspectives.
In Experiment 3, compared to a control condition, high-power participants were less accurate in determining others’ emotion expressions, suggesting a power-induced impediment to experiencing empathy. An additional study also found a negative relationship between individual difference measures of power and perspective-taking. Across these studies, power was associated with a reduced tendency to comprehend how others see, think, and feel.
The powerful are often accused of being predominantly concerned with their own desires and well-being, of being insensitive to the social implications of their behavior, of being poor perspective-takers (Fiske, 1993; Keltner, Gruenfeld, & Anderson, 2003; Kipnis, 1972). Indeed, perspective-taking – the capability to step outside of one’s own experience and imagine the emotions, perceptions, and motivations of another – seems the antithesis of the self-interested behavior often displayed by the powerful: it has been linked to moral reasoning (Kohlberg, 1976), altruistic behavior (Batson, 1991), and social competence (Davis, 1983).
In this paper, we demonstrate empirically that power is associated with greater difficulty in taking others’ perspectives. We find that individuals primed with power anchor too heavily on their own vantage points and demonstrate less accuracy when assessing the emotions and thoughts of others.
In Experiment 1, participants primed with high power compared to low power were more likely to draw an “Ε” on their forehead in a self-oriented direction, indicating less of an inclination to spontaneously adopt another’s visual perspective.
In Experiments 2a and 2b, high-power participants were less likely to take into account that others did not possess their privileged knowledge, suggesting that power leads individuals to anchor too heavily on their own vantage point, insufficiently adjusting to others’ perspectives.
In Experiment 3, compared to a control condition, high-power participants were less accurate in determining others’ emotion expressions, suggesting a power-induced impediment to experiencing empathy. An additional study also found a negative relationship between individual difference measures of power and perspective-taking. Across these studies, power was associated with a reduced tendency to comprehend how others see, think, and feel.
The powerful are often accused of being predominantly concerned with their own desires and well-being, of being insensitive to the social implications of their behavior, of being poor perspective-takers (Fiske, 1993; Keltner, Gruenfeld, & Anderson, 2003; Kipnis, 1972). Indeed, perspective-taking – the capability to step outside of one’s own experience and imagine the emotions, perceptions, and motivations of another – seems the antithesis of the self-interested behavior often displayed by the powerful: it has been linked to moral reasoning (Kohlberg, 1976), altruistic behavior (Batson, 1991), and social competence (Davis, 1983).
In this paper, we demonstrate empirically that power is associated with greater difficulty in taking others’ perspectives. We find that individuals primed with power anchor too heavily on their own vantage points and demonstrate less accuracy when assessing the emotions and thoughts of others.
Definitions and Theories of Power
Power is often defined as the capacity to influence others, with power emerging from control over valuable resources and the ability to administer rewards and punishments (French & Raven, 1959; Keltner et al., 2003). The power-approach theory (Keltner et al., 2003) suggests that power increases goal-directed activity. As a result, the powerful act more (Galinsky, Gruenfeld, & Magee, 2003) and with greater variability (Guinote, Brauer, & Judd, 2001). Although power is considered a structural variable, a property of social relationships, its psychological properties can be activated by exposing people to cues related to power or by recalling past experiences with power; activating power through these manipulations leads to the same subsequent effects as those obtained using structural and role-based manipulations of power (Anderson & Galinsky, in press; Chen, Lee-Chai, & Bargh, 2001; Galinsky et al., 2003).
Power and Perspective-Taking: Opposing Effects
There are a number of reasons why power may diminish perspective-taking. First, those in power – by definition - have control over valuable resources and are therefore less dependent on others. Thus, to accomplish their goals the powerful do not need to rely on an accurate, comprehensive understanding of others. Second, power is typically associated with increased demands on attention, making it difficult to take the perspective of everyone under the power-holder’s charge (Fiske, 1993).
In addition, the effects of perspective-taking stand in sharp contrast to those of power, with power and perspective-taking affecting a number of variables in opposite ways. For example, whereas the perspective-taker seems the consummate adapter, including the traits of others in his or her own self-representation (Galinsky, Ku, Wang, 2005), the high-power individual’s self-concept remains more rigid; individuals with more power in their marriages resist the identities imposed on them by their spouses (Cast, 2003), and when relationship partners become more emotionally similar, it is the lower power partner that does the majority of the adaptation (Anderson, Keltner, & John, 2003). Perspective-taking is associated with increased similarity (Davis, Conklin, Smith, & Luce, 1996) whereas Lee and Tiedens (2001) have argued that having power creates psychological distance from others.
The differential effects of power and perspective-taking extend to social perception. The powerful, on the one hand, are less accurate than the powerless in estimating the interests and positions of others (Keltner & Robinson, 1997) and are more likely to make self-serving attributions (Kipnis, 1972). Perspective-takers, on the other hand, accurately perceive the interests of others (Eisenberg, Murphy, & Shepard, 1997) and are more other-serving in their attributions (Regan & Totten, 1975). Those with more power form less complex interpersonal impressions (Woike, 1994), basing their impressions of others on expectancies (Copeland, 1994) and stereotypes (Fiske, 1993, but see Overbeck & Park, 2001). In contrast, perspective-takers have been shown to battle successfully the seemingly inevitable forces of stereotyping (Galinsky & Moskowitz, 2000) and leads to more integratively complex contemplations (Tetlock, Skitka, & Boettger, 1989). The opposing forces of perspective-taking and power have important social implications: Perspective-taking has been associated with altruism and helping behavior (Batson, 1991), whereas power has been linked with such malfeasant social behaviors as sexual harassment (Bargh, Raymond, Pryor, & Strack, 1995).
Based on these conflicting effects of power and perspective-taking, we hypothesized that power would decrease perspective-taking. We conducted an initial study to test for this proposed inverse relationship. Thirty-two participants (13 male and 19 female undergraduate students) reported their personal sense of power, a measure that assesses an individuals’ power in general, across their social relationships (8 items; alpha = .82; e.g., “I think I have a great deal of power in my relationships with others”; Anderson, John, & Keltner, 2005). The sense of power scale is correlated with people’s standing in power hierarchies, whether they occupy powerful roles (Anderson, John, & Keltner, 2005), and predicts the same behaviors as structural manipulations of power (Anderson & Berdahl, 2002; Anderson & Galinsky, in press). Participants also reported their tendency to engage in perspective-taking, a scale that measures the “tendency to adopt the point of view of others in everyday life” (7 items; alpha = .64; e.g., “I sometimes try to understand my friends better by imagining how things look from their perspective”; Davis, 1983). We regressed the perspective-taking scale on the power scale (controlling for participant sex) and found a significant negative relationship between power and perspective-taking, B = -.35, SE = 0.17, p-rep = .88.
To establish that power directly leads to a lack of perspective-taking, we conducted four experiments using a diverse set of perspective-taking variables. We explored whether high-power participants would be less inclined to spontaneously adopt another’s visual perspective (Experiment 1), less likely to take into account that others did not possess their privileged knowledge (Experiments 2a and 2b), and less accurate in determining the emotion expressions of others (Experiment 3) than individuals in low power or control conditions. Because structural manipulations of power (e.g., providing direction or evaluating others) often increase cognitive load (Fiske, 1993), which might reduce the ability to engage in perspective-taking, we used a priming procedure – having individuals recall an experience with power (Galinsky et al., 2003) – in which power was manipulated while cognitive load was held constant.
In addition, the effects of perspective-taking stand in sharp contrast to those of power, with power and perspective-taking affecting a number of variables in opposite ways. For example, whereas the perspective-taker seems the consummate adapter, including the traits of others in his or her own self-representation (Galinsky, Ku, Wang, 2005), the high-power individual’s self-concept remains more rigid; individuals with more power in their marriages resist the identities imposed on them by their spouses (Cast, 2003), and when relationship partners become more emotionally similar, it is the lower power partner that does the majority of the adaptation (Anderson, Keltner, & John, 2003). Perspective-taking is associated with increased similarity (Davis, Conklin, Smith, & Luce, 1996) whereas Lee and Tiedens (2001) have argued that having power creates psychological distance from others.
The differential effects of power and perspective-taking extend to social perception. The powerful, on the one hand, are less accurate than the powerless in estimating the interests and positions of others (Keltner & Robinson, 1997) and are more likely to make self-serving attributions (Kipnis, 1972). Perspective-takers, on the other hand, accurately perceive the interests of others (Eisenberg, Murphy, & Shepard, 1997) and are more other-serving in their attributions (Regan & Totten, 1975). Those with more power form less complex interpersonal impressions (Woike, 1994), basing their impressions of others on expectancies (Copeland, 1994) and stereotypes (Fiske, 1993, but see Overbeck & Park, 2001). In contrast, perspective-takers have been shown to battle successfully the seemingly inevitable forces of stereotyping (Galinsky & Moskowitz, 2000) and leads to more integratively complex contemplations (Tetlock, Skitka, & Boettger, 1989). The opposing forces of perspective-taking and power have important social implications: Perspective-taking has been associated with altruism and helping behavior (Batson, 1991), whereas power has been linked with such malfeasant social behaviors as sexual harassment (Bargh, Raymond, Pryor, & Strack, 1995).
Based on these conflicting effects of power and perspective-taking, we hypothesized that power would decrease perspective-taking. We conducted an initial study to test for this proposed inverse relationship. Thirty-two participants (13 male and 19 female undergraduate students) reported their personal sense of power, a measure that assesses an individuals’ power in general, across their social relationships (8 items; alpha = .82; e.g., “I think I have a great deal of power in my relationships with others”; Anderson, John, & Keltner, 2005). The sense of power scale is correlated with people’s standing in power hierarchies, whether they occupy powerful roles (Anderson, John, & Keltner, 2005), and predicts the same behaviors as structural manipulations of power (Anderson & Berdahl, 2002; Anderson & Galinsky, in press). Participants also reported their tendency to engage in perspective-taking, a scale that measures the “tendency to adopt the point of view of others in everyday life” (7 items; alpha = .64; e.g., “I sometimes try to understand my friends better by imagining how things look from their perspective”; Davis, 1983). We regressed the perspective-taking scale on the power scale (controlling for participant sex) and found a significant negative relationship between power and perspective-taking, B = -.35, SE = 0.17, p-rep = .88.
To establish that power directly leads to a lack of perspective-taking, we conducted four experiments using a diverse set of perspective-taking variables. We explored whether high-power participants would be less inclined to spontaneously adopt another’s visual perspective (Experiment 1), less likely to take into account that others did not possess their privileged knowledge (Experiments 2a and 2b), and less accurate in determining the emotion expressions of others (Experiment 3) than individuals in low power or control conditions. Because structural manipulations of power (e.g., providing direction or evaluating others) often increase cognitive load (Fiske, 1993), which might reduce the ability to engage in perspective-taking, we used a priming procedure – having individuals recall an experience with power (Galinsky et al., 2003) – in which power was manipulated while cognitive load was held constant.
Experiment 1: Drawing an E
Experiment 1 was designed to examine the effect of power on the tendency to spontaneously adopt another person’s visual perspective, an important dimension in understanding others’ beliefs and intentions (Baron-Cohen, 1995). We used a procedure created by Hass (1984) in which participants are asked to draw an “E” on their foreheads. One way to complete the task is to draw an “E” as though the self is reading it, which leads to a backward and illegible “Ε” from the perspective of another person. The other way to approach the task is to draw the “Ε” as though another person is reading it, which leads to production of an “Ε” that is backward to the self (see below). We predicted that participants in the high power condition would be more likely to draw the “Ε” in the “self-oriented” direction, indicating a lesser tendency to spontaneously adopt another’s perspective, than would participants in the low power condition.
Method
Participants
Participants were 57 undergraduates (41 women and 16 men) with an average age of 20.02 years who participated for payment of $10 and entry into a $300 lottery.
Design and Procedure
The experiment involved two conditions, a high-power and a low-power condition. Participants came in groups of 2 or 3 to the lab where they were given a high- or low-power experiential prime, which has been shown to reliably manipulate a sense of power (Galinsky et al., 2003). Participants assigned to the high-power condition were instructed to recall and write about a personal incident in which they had power over another individual or individuals. Power was defined for them as a situation in which they controlled the ability of another person or persons to get something they wanted or were in a position to evaluate those individuals. Participants assigned to the low-power condition were instructed to write about a personal incident in which someone else had power over them. Participants were unaware of the power prime manipulation that the other participants in the session received.
After completing this power-priming task, participants were brought to separate rooms, and the experimenter explained that the next tasks would be described in a packet on the desks in their rooms. The packets instructed the participants to begin working on a resource allocation task, which was designed to reinforce the power prime. Those participants who received the high-power prime were asked to allocate lottery tickets to themselves and another participant, whereas those participants who completed the low-power prime were asked to estimate how many lottery tickets they would receive. Participants next encountered the following instructions:
Please follow the instructions below and perform the tasks requested in the order they are presented. They are coordination tasks.
Task 1. With your dominant hand, as quickly as you can, snap your fingers 5 times.
Task 2. With your dominant hand, as quickly as you can, draw a capital E on your forehead with the marker provided. Don’t worry, the marker is non-toxic, and we will make sure it is removed before you leave today.
At the end of the experiment, each participant was thoroughly debriefed and fully probed for suspicion. Not a single participant expressed any suspicion that the power manipulation and the dependent measure were related.
After completing this power-priming task, participants were brought to separate rooms, and the experimenter explained that the next tasks would be described in a packet on the desks in their rooms. The packets instructed the participants to begin working on a resource allocation task, which was designed to reinforce the power prime. Those participants who received the high-power prime were asked to allocate lottery tickets to themselves and another participant, whereas those participants who completed the low-power prime were asked to estimate how many lottery tickets they would receive. Participants next encountered the following instructions:
Please follow the instructions below and perform the tasks requested in the order they are presented. They are coordination tasks.
Task 1. With your dominant hand, as quickly as you can, snap your fingers 5 times.
Task 2. With your dominant hand, as quickly as you can, draw a capital E on your forehead with the marker provided. Don’t worry, the marker is non-toxic, and we will make sure it is removed before you leave today.
At the end of the experiment, each participant was thoroughly debriefed and fully probed for suspicion. Not a single participant expressed any suspicion that the power manipulation and the dependent measure were related.
Results
As Hass (1984) points out, handedness could affect the direction of the “E”; thus, we controlled for handedness (49 right-handed and 8 left-handed participants) and participant sex in all analyses. Using logistic regression, we regressed the direction of the “E” (0 = self-oriented, 1 = other-oriented) on power condition, handedness, and participant sex. The only significant effect was for power, B = -1.51, SE = 0.76, p-rep = .88. High-power participants were almost three times as likely as low-power participants to draw a self-oriented “E” (33% [n = 8 of 24] vs. 12% [n = 4 of 33]).
To determine whether the amount of power people expressed possessing in their essays predicted perspective-taking, we had one coder, who was blind both to condition and hypotheses, rate all essays using a 7-point scale measuring how much power the participant reported having. This coder had achieved high reliability (α = .94) with another coder on a large sample of power essays from another set of studies. Participants described themselves as having more power in the high-power essays (M = 5.75, SD = 0.53) than in the low-power essays (M = 2.21, SD = 0.55), t(55) = 24.44, p-rep = .99, d = 6.57. Using logistic regression, we regressed the direction of the “E” on the amount of power that people expressed possessing in their essays, handedness, and participant sex. The only significant effect was for the amount of essay power, B = -0.41, SE = 0.21, p-rep = .88. The more power participants described possessing in their essays, the more likely they were to draw a self-focused “E”.
Consistent with our theorizing, priming power led individuals to give less consideration to others’ perspectives when given an opportunity to spontaneously adopt another’s visual point of view. In the next set of experiments we sought to determine whether power affects the tendency to take another person’s perspective when doing so is required for effective communication.
To determine whether the amount of power people expressed possessing in their essays predicted perspective-taking, we had one coder, who was blind both to condition and hypotheses, rate all essays using a 7-point scale measuring how much power the participant reported having. This coder had achieved high reliability (α = .94) with another coder on a large sample of power essays from another set of studies. Participants described themselves as having more power in the high-power essays (M = 5.75, SD = 0.53) than in the low-power essays (M = 2.21, SD = 0.55), t(55) = 24.44, p-rep = .99, d = 6.57. Using logistic regression, we regressed the direction of the “E” on the amount of power that people expressed possessing in their essays, handedness, and participant sex. The only significant effect was for the amount of essay power, B = -0.41, SE = 0.21, p-rep = .88. The more power participants described possessing in their essays, the more likely they were to draw a self-focused “E”.
Consistent with our theorizing, priming power led individuals to give less consideration to others’ perspectives when given an opportunity to spontaneously adopt another’s visual point of view. In the next set of experiments we sought to determine whether power affects the tendency to take another person’s perspective when doing so is required for effective communication.
Experiments 2a and 2b: Considering Communication Intentions
Most messages can be interpreted in multiple ways, and effective communication requires taking the knowledge and perspectives of one’s audience into account. The same semantic content (e.g., “nice suit!”) can be received as a compliment or a thinly veiled insult, depending on knowledge of the speaker’s tastes and previous interactions. When message receivers have privileged knowledge about a speaker’s intentions, they often have difficulty recognizing and adjusting for the fact that other listeners do not share this privileged perspective (Keysar, 1994). They are cursed by their knowledge, inaccurately predicting that others see the world as they do. Epley, Keysar, Van Boven, and Gilovich (2004) have suggested that individuals initially anchor on their own vantage point and then adjust for another’s perspective. We contend that power leads this adjustment to be particularly insufficient.
Experiment 2a
Following Keysar (1994), we gave participants a message and asked them to interpret how a friend of the speaker might perceive the message. The message on its face seemed sincere, but privileged background knowledge about the speaker’s intentions suggested a sarcastic interpretation. We predicted that high-power participants would be more likely than low-power participants to assume that the friend understood the sarcasm, even though a sarcastic interpretation depended on privileged background knowledge that the friend did not possess.
Method
Participants
Participants were 42 undergraduate students who participated for payment of $10.
Procedure
Participants were greeted in the laboratory by an experimenter who explained that they would complete several questionnaires related to decision-making. The experimental manipulations and our dependent variables were embedded in the packet of materials that participants received.
Power manipulation.
The power manipulation was the same experiential prime used in Experiment 1.
Message interpretation.
After completing the power prime and a filler task, participants were given a scenario in which they and a colleague had gone to a fancy restaurant recommended by the colleague’s friend but had a particularly poor dining experience. The next day the colleague had sent an email to the friend stating only that: “About the restaurant, it was marvelous, just marvelous.” Participants were asked to respond to the question, “How do you think the colleague’s friend will interpret the comment?” on a scale anchored at “very sarcastic” (1) and “very sincere” (6). There was no information in the email itself to suggest anything other than sincerity. However, if participants anchored on their privileged knowledge of the speaker’s intention then they would think that the friend would interpret the message as sarcastic in nature.
As in Experiment 1, no participants expressed any suspicion that the power manipulation and the dependent measure were related.
Results
High-power participants (M = 3.74, SD = 1.54) thought the message would be perceived as more sarcastic by the naïve recipient than did low-power participants (M = 4.84, SD = 1.30), t(40) = 2.47, p-rep = .93, d = .77. These findings support our prediction that power leads individuals to anchor too heavily on their own vantage point, insufficiently adjusting to other’s perspectives.
Experiment 2b
We ran a follow-up experiment using a different scenario to rule out the possibility that power simply produces more sarcastic attributions. Fifty-one participants read that they and a colleague had gone to a restaurant where the colleague’s friend always had poor dining experiences. They, however, really enjoyed the meal. The colleague sent the friend the same “marvelous, marvelous” email, and participants predicted how the friend would interpret the comment using the same sarcastic-sincere scale from Experiment 2a. In this experiment, however, the privileged information implied sincerity, but the naïve listener would have inferred sarcasm. Replicating Experiment 2a, high-power participants were significantly more anchored on their privileged knowledge. Specifically, high power participants (M = 3.31, SD = 1.46) thought that the naïve recipient of the message would interpret it as more sincere and less sarcastic than did low-power participants (M = 2.36, SD = 1.35), t(49) = 2.4, p-rep = .93, d = .68.
Experiment 3: Interpreting Emotion Expressions
The previous experiment suggests that power can lead to less accurate social judgment. Power might also inhibit the ability to pay attention to and comprehend others’ emotional states. Accurate perception of another’s emotional state is part and parcel of the capacity to experience empathy. Indeed, empathy has been defined as “the ability to perceive accurately how another person is feeling” (Levenson & Ruef, 1992). Since the ability to discern others’ emotions bears resemblance to performance on theory-of-mind tasks related to perspective-taking, this skill has also been called affective perspective-taking (Denham, 1986). We predicted that power would be associated with worse ability to accurately detect others’ emotion expressions. Consistent with this prediction, previous studies have determined that men, who typically have more relative power in society compared to women, are worse judges of others’ emotion expressions than women (Hall, Gaul, & Kent, 1999; McClure, 2000).
To test our hypothesis, we presented to participants pictures from the Diagnostic Analysis of Nonverbal Accuracy (DANVA2; Nowicki & Carton, 1993) materials of young men and women expressing emotions that differed in both their quality and intensity. Because the first two experiments did not use an experimental control condition, it is not entirely clear whether activating high power decreases perspective-taking. To resolve this ambiguity, in the final experiment we compared high power to a control condition.
To test our hypothesis, we presented to participants pictures from the Diagnostic Analysis of Nonverbal Accuracy (DANVA2; Nowicki & Carton, 1993) materials of young men and women expressing emotions that differed in both their quality and intensity. Because the first two experiments did not use an experimental control condition, it is not entirely clear whether activating high power decreases perspective-taking. To resolve this ambiguity, in the final experiment we compared high power to a control condition.
Method
Participants
Seventy undergraduate students (16 men and 54 women, average age of 20.17 years) were recruited from an online participant pool at a private West coast university and paid $12.
Procedure
When participants arrived at the lab they were seated at semi-private cubicles and were told they would be completing a study on experiential cognition.
Power manipulation.
There were two experimental conditions: high-power and control. The high-power condition completed the same experiential prime that was used in the previous two experiments. Participants randomly assigned to the control condition responded to the following: “Please recall your day yesterday. In the space below, describe your day including thoughts, feelings, events, etc.”
Emotion expression interpretation.
Next, participants were told they would complete an ostensibly unrelated task that involved “looking at the experience of others.” Specifically, they were told that they would be observing a series of faces and their task would be to determine the emotion being expressed by the target. The Adult Faces and Adult Paralanguage Scale associated with the DANVA2 is comprised of 24 images of faces that express happiness, fear, anger or sadness. After viewing each stimulus, participants selected which of the four emotions was being expressed.
Results
Accuracy was measured by recording the number of errors participants made when judging the target’s expressed emotion. We submitted the total number of errors to a 2 (high-power vs. control) X 2 (participant sex) analysis of variance. Replicating prior work, a main effect for participant sex emerged, F(1, 66) = 4.59, p-rep = .90, h2 = 07, with men (M = 4.56, SD = 3.01) making more errors than women (M = 3.61, SD = 2.12). More important to the present research, a main effect of power also emerged, F(1, 66) = 10.81, p-rep = .98, h2 = .14. Participants primed with high power (M = 4.54, SD = 2.80) made more errors when judging the emotion expressions of others compared to control participants (M = 3.11, SD = 1.57). The interaction was not significant F(1, 66) = 2.32, p-rep = .79, h2 = .03. These findings support the prediction that power is associated with less accuracy in emotion detection and suggest an additional consequence of diminished perspective-taking: greater difficulty in experiencing empathy.
Discussion
Across four experiments we found that power was associated with a reduced tendency to comprehend how others see the world, how others think about the world, and how others feel about the world. Priming power led participants to be less likely to spontaneously adopt another’s visual perspective, less likely to take into account that another person did not possess their privileged knowledge, and less accurate in detecting the emotional states of others. This inverse relationship between power and perspective-taking emerged across multiple forms of perspective-taking; regardless of whether participants were explicitly told to be accurate (Experiment 3) or whether the consideration of other’s perspectives was spontaneous (Experiment 1), power was associated with less perspective-taking. In addition, evidence from the third experiment in which those who were primed with power deviated from a control condition, demonstrates that it is high power that is an impediment to perspective-taking. Given that the use of our experiential prime led participants to be unaware of the connection between our manipulation of power and the perspective-taking dependent variables, we can say this relationship is at least partially a non-conscious one. Instead of a conscious decision to ignore others’ perspectives, we believe that power leads to a psychological state that makes perspective-taking less likely. The four experiments, combined with our initial study demonstrating a negative relationship between a measure of power and a measure of perspective-taking, suggest that high-power individuals are less focused on the meaningful psychological experiences of those around them.
Although the current studies collectively point in the direction of our hypothesized inverse relationship between power and perspective-taking, we acknowledge that each of our designs were limited to a dichotomized independent variable, and our findings always depended on a relative difference between two conditions. The use of multi-level manipulations of power in the future would provide for greater precision in interpreting power’s effects.
Although the current studies collectively point in the direction of our hypothesized inverse relationship between power and perspective-taking, we acknowledge that each of our designs were limited to a dichotomized independent variable, and our findings always depended on a relative difference between two conditions. The use of multi-level manipulations of power in the future would provide for greater precision in interpreting power’s effects.
Implications and Moderators
Diminished perspective-taking may connect to the finding that power is associated with action and increased goal pursuit (Galinsky et al., 2003). By not attending to the concerns of others, the powerful can plunge headfirst into action and pursue goals without restraint (Galinsky et al., 2003; Magee, Galinsky, & Gruenfeld, 2005). Although lack of perspective-taking may lead the powerful towards malfeasance, it can be an adaptive response to attentional demands, one that allows for efficient navigation of social and organizational worlds.
Finding that power reduces perspective-taking and empathy is consistent with recent views that power can lead to objectification (Gruenfeld, Inesi, Magee, & Galinsky, 2005; Keltner et al., 2003), the tendency to view others only in terms of qualities that serve the perceiver’s personal goals and interests, while failing to consider those features of others that define their humanity. With less attention to the human aspects of others, it should be easier for a power-holder to use others as tools in the service of his or her goals and interests (Gruenfeld et al., 2005).
The current results also indirectly support an integrated relationship between power, perspective-taking, and stereotyping: perspective-taking decreases stereotyping (Galinsky & Moskowitz, 2000), power increases stereotyping (Fiske, 1993), and power decreases perspective-taking. Fiske (1993) has argued that stereotyping is one way that the powerful maintain their privileged position. Failing to take others’ perspectives, self-interested objectification of others, and stereotyping of others may all be part of the cognitive toolbox that power-holders use to stay in control. However, lack of perspective-taking may also sew the seeds of power’s demise. When disregard for the concerns, emotions, and individuality of others – their humanity – persists, the powerful can start to inspire enmity, bitterness, and incipient rebellion. The inverse relationship between power and perspective-taking may allow the powerful to accomplish short-term goals but lead to the long-term loss of power.
Although our studies demonstrate that power leads to less perspective-taking, we suspect this relationship is not invariant. One important moderator might be the extent to which accurate knowledge of another’s perspective would increase the likelihood of power-holders achieving their goals. In addition, the degree of responsibility that high-power individuals feel for their subordinates is likely to influence perspective-taking tendencies. Indeed, when the powerful feel a sense of responsibility, their behavior resembles the ideal perspective-taker, showing increased generosity (Chen et al., 2001) and more individuated impressions (Overbeck & Park, 2001). Similarly, manipulations of accountability (the pressure to justify one’s decisions or view to others [Tetlock et al., 1989]) may direct the attention of power-holders to other’s perspectives, checking and balancing the cognitive effects of power. Finally, culture may be a critical variable in determining when power leads to perspective-taking and when it leads to egocentric self focus (Zhong, Magee, Maddux, & Galinsky, in press).
The observation that responsibility, accountability, and culture may influence how power is experienced and enacted offers insight into when and how power may be delivered from the forces of corruptibility: action-oriented power combined with perspective-taking may be a particularly constructive force. By leading the powerful to take the perspectives of others, their actions ultimately might benefit the lives of the less powerful, while also possibly helping them to remain in power.