Differentiating the Effects of Social and Personal Power

[First published on Nathen’s Miraculous Escape, June 11, 2009.]

Thanks to my advisor and collaborator, Sean M. Laurent, and my second reader, Sara D. Hodges


The current research attempted to differentiate the effects of social power (i.e., having control of others’ outcomes) from personal power (i.e., control of one’s own outcomes) on variables related to perspective-taking. Using methodology adapted from Galinsky, Magee, Inesi, and Gruenfeld (2006), 224 participants were primed with high social power, high personal power, or low power, and then completed two perspective-taking measures. While these measures did not significantly vary with power condition, evidence was found that supports the differentiation of personal power from social power. Low power, as experienced by participants (rather than as manipulated) had a strong negative correlation with personal power, and a weak negative correlation with social power. Personal power and social power were not significantly correlated.

Differentiating the Effects of Social and Personal Power

To some extent, all human beings want power—even if it is just enough power to ensure one’s safety, allow oneself to prosper, and obtain or keep freedom for oneself and one’s groups. This desire for power, to get just enough or to get even more than one needs for one’s own comfort and safety, may explain the frequent power struggles that populate our history books. And given the strong desire humans have to attain power, it is no wonder that psychologists are fascinated with the construct; understanding power is central to understanding human behavior.

Social power

Power can be defined in many ways, but in psychological research it is usually defined in relational terms, and is called “social power.” Individuals are said to have social power when they can control the outcomes of another person, influencing that person’s states, providing that person with rewards, inflicting punishments on that person, or governing the flow of resources to that person (e.g., Fiske & Berdahl, 2007; Galinsky, Magee, Inesi, & Gruenfeld, 2006; Keltner, Gruenfeld, & Anderson, 2003; Smith & Trope, 2006). Common to all definitions of social power is that there exist relationships between people, where some people control and are important sources for valued resources for other people. These valued resources are not limited to any particular physical or social object; a valued resource might range from a kind word to a grade in a class, a promotion, an offer of protection, or a glass of water for a thirsty child.

The approach-inhibition theory of social power ties high and low power states to psychological approach and inhibition mechanisms (Keltner et al., 2003). According to this theory, people who are high in social power will tend to have an approach orientation, will typically experience more positive affect, and will be trait-congruent—that is, they will act in alignment with their own traits, be less careful, more automatic in their social cognitions, and attend primarily to elements of the environment which might be useful or pleasurable, including a tendency to see others as a means to their own ends. People who are low in power, on the other hand, will tend to have an inhibition orientation, where they tend to experience more negative affect, be more cautious and observant of potential risks in their environment, and tend to see themselves as means to others’ ends.

Research has provided some support for approach-inhibition theory: Compared to people low in power, people high in power have been found to be more trait-congruent, more sensitive to potential rewards, and to experience more positive affect (Anderson & Berdahl, 2002). They are more prone to take risks and be optimistic about risk outcomes (Anderson & Galinsky, 2006). Higher power has also been associated with increased abstract information processing (Smith & Trope, 2006), increased reliance on the availability heuristic (Weick & Guinote 2008), and increased stereotyping (Fisk, 1993), while low power has been shown to increase metastereotyping (Lammers, Gordjin, & Otten, 2008) and impair executive functions (Smith, Jostmann, Galinsky, & van Dijke, 2008). Supporting an approach orientation, high power is associated with a tendency to act rather than not act (Galinsky, Gruenfeld, & Magee, 2003), to be more expressive (Snodgrass, Hecht, & Ploutz-Snyder, 1998), and to be less influenced by a variety of situational variables such as pressure to conform (Galinsky, Magee, Gruenfeld, Whitson, & Liljenquist, 2008). Furthermore, high power has been linked to devaluing others’ work for self-gain (Kipnis, 1972), to more extreme, less accurate judgments of opponents (Keltner & Robinson, 1997), and in some cases, to sexual harassment (Bargh, Raymond, Pryor, & Strack, 1995).

While many of the behaviors associated with power are self-serving and might serve to entrench power holders in their positions, at least some of these effects can be shifted by values and accountability; for example, individuals with a communal relationship orientation were found to be more generous when primed for high power (Chen, Lee-Chai, & Bargh, 2001), and familial responsibilities reduced several kinds of risky reward seeking in high need-for-power individuals (Winter & Barenbaum, 1985). It may also be true that some effects may not be the direct result of having power—instead they may be the result of the responsibilities that come along with power: For example, high power individuals have been shown to be better at individuating others unless burdened with organizational responsibility (Overbeck & Park, 2001).

Personal power

Van Dijke and Poppe (2006) asserted that striving for power is better explained by distinguishing between social power (i.e., having control of others’ outcomes) and personal power (i.e., having control of one’s own outcomes). While social power is related to authority, status, and dominance (Hall, Coats, & LeBeau, 2005), personal power is also related to autonomy, independence, agency, and competence (Overbeck & Park, 2001; Van Dijke & Poppe, 2006). Van Dijke and Poppe presented evidence that, given the choice, individuals act to increase their personal power instead of their social power, and will at times even voluntarily reduce their own social power to obtain personal power. Van Dijke and Poppe argued further that when individuals act to increase their power over others, they are motivated not by a wish to control others, but by a wish to not be controlled by others, and that increasing social power is a means to that end.

While most researchers have been careful to define power in social terms, it is evident that personal and social power have at least sometimes been confounded in research on the psychological effects of having power. That is, because personal and social power may tend to vary together, if they are not carefully separated, the effects of one type of power may be masked by the effects of the other. This may be especially true in experimental paradigms where social power is manipulated by giving some participants control over others or where the idea of control over others is primed. In these cases, personal power might be inadvertently manipulated at the same time. For example, one commonly used power-priming manipulation is to have participants write about a time in which they had power over someone else, or when someone else had power over them. While manipulating social power, this manipulation will tend to prime personal power as well, because people low in social power will tend to have less control of their own outcomes than do those high in social power, and those high in social power will tend to have more power over their own outcomes. This leaves open the possibility that any observed effects of this power manipulation could be due to the participants having control of others’ outcomes, having control over their own outcomes, or some combination of the two.

If personal power is the most likely reason that people seek power over others, then it is important to distinguish the effects of personal power from the effects of social power. It may be that having social power is associated with more negative or antisocial outcomes than is having personal power. If this is true, then encouraging people to seek personal power but not social power may be indicated.

Power and perspective-taking

Perspective-taking, conceptualized as the act of imagining or ability to imagine the experiences of others (Galinsky et al., 2006), is an important skill, associated with social competence and self-esteem (Davis, 1983). It is considered a necessary component of moral development and moral reasoning (Kohlberg, 1976; Walker 1980), and of empathy (Eisenberg, Murphy, & Shepard, 1997). Furthermore, perspective-taking has been linked to reductions in stereotyping (Galinsky & Moskowitz, 2000), increases in self-other overlap (Davis, Conklin, Smith, & Luce, 1996), and helping behaviors (Batson, 1991).

Perspective-taking, then, is clearly a beneficial behavior, and a necessary part of healthy social functioning. Social power, however, seems to have a negative effect on the ability to take others’ perspectives; Galinsky et al. (2006) found that after participants were primed with social power, they were less likely to take an outside visual perspective (i.e., less likely to draw an “E” on their foreheads in the direction that would be easily readable from another person’s perspective), less able to recognize other people’s facial emotional expressions, and were more likely to believe that others knew what they knew, even when they were presented with contrary evidence.

The present research

The purpose of the current research was to take a preliminary step toward differentiating the effects of having social and personal power on perspective-taking. Toward this end, Galinsky and colleagues’ (2006) methodology was used and extended somewhat, in order to test hypotheses about personal as well as social power. In the original paradigm, participants were first primed for either high social power or low power by writing a power-related essay. In the current paradigm, another condition was added where participants were primed for personal power. Following the priming tasks, participants performed one of two perspective-taking tests (these were also the same dependent variables used by Galinsky et al., 2006). One of these was the Diagnostic Analysis of Nonverbal Accuracy for adult faces (DANVA; Nowicki & Duke, 2001). In this test, participants view photographs on a computer screen, and are asked to identify the emotions shown on facial expressions of the photographed target. This task is thought to be a measure of interpersonal sensitivity related to perspective-taking, because the ability to recognize what emotions others are feeling is an important part of being able to take their perspectives. The other task asked participants to draw an E on their own foreheads; this measure assesses whether participants tend to spontaneously adopt the visual perspective of a person facing them, rather than their own internal perspective.

As a first outcome, we expected to replicate the work of Galinsky and colleagues, showing that the high social power group would perform worse on the DANVA and draw Es from their own perspective more often than the low power group. In terms of what personal power would predict, our hypotheses were not as clear. It was possible that personal and social power would activate approach equivalently and thus have the same effect on perspective-taking. It was also possible, however, that personal power would affect perspective-taking more negatively than social power, because people in the personal power condition might be less cued in to relationships with others. We reasoned, however, that the personal power group would perform somewhere between the low and social power groups on measures of perspective-taking, because similarly to social power, envisioning personal power should create associations with rewards and freedom, but different from social power, it should not contain the same dominance or hierarchy power cues.



Two hundred and twenty-four undergraduates (172 females) from the University of Oregon’s Psychology and Linguistics human subjects pool participated in this experiment. The mean age was 19.6 years and the age range was between 17 and 49 years. In addition to partially fulfilling course requirements by participating, participants received entry into a lottery drawing for $100.


Participants each drew an E on their own foreheads as a measure of perspective-taking (Galinsky et al., 2006; Hass, 1984). An E that read in the correct direction for an outside observer facing the participant was treated as an indication that perspective-taking had occurred. This was presented to participants as a coordination task, rather than a test of perspective-taking. For a full description of the methodology, see Galinsky et al. (2006).

Participants also completed the Diagnostic Analysis of Nonverbal Accuracy (DANVA2-AF), a task where participants guess the emotions shown in 24 photographs of young adults’ faces: happiness, sadness, anger, or fear (Nowicki & Duke, 2001). Each image appeared on a computer screen for 2 seconds, followed by 5 seconds for the participant to mark their answer on a paper test form.

Next, participants completed four questionnaires that had been slightly modified from their original forms such that participants were instructed to respond based on how they were currently feeling, rather than assessing themselves globally. The first of these was a measure of Unmitigated Agency (UA; Spence, Helmreich, & Holahan, 1979), focusing on the self to the exclusion of others. On this scale, participants rated themselves on a five-point scale anchored by two contradictory characteristics (e.g., “Not at all arrogant” and “Very arrogant”).The second was the WHO-5 Well-being Index (WHO-5; Bech, 1993), where participants responded to questions such as “I feel cheerful and in good spirits” on a 5-point scale anchored by “Does not describe the way I feel at all” and “Describes how I’m feeling well.” The next questionnaire was Hegelson’s Unmitigated Communion Scale (UC; Hegelson & Fritz, 1998), which measures a focus on others’ needs to the detriment of the self. Participants responded on a five-point scale anchored with “Strongly agree” and “Strongly disagree” (e.g., “I always place the needs of others above my own.”)The last measure assessed conservative attitudes (Short Right-Wing Authoritarianism Scale; RWA; Altemeyer, 1998; Zakrisson, 2005). In this measure, participants rated their agreement with statements such as, “Our country needs a powerful leader, in order to destroy the radical and immoral currents prevailing in society today,” on a six-point scale anchored with “Strongly agree” and “Strongly disagree.” One additional, single-item measure of self-esteem was used: “How much do you like your name, in total?” This was measured on a 9-point scale, and was taken from Gebauer, Riketta, Broemer, and Maio (2008), who found that higher scores on this single-item measure indicated higher self-esteem.

As a manipulation check, participants rated themselves on several power-related adjectives or short statements (7-point scale where 1 = “Not at all” and 7 = “Extremely”). The following adjectives were used: High Social Power—dominant, in-charge, leader, organizer; High Personal Power—self-directed, independent, controls own choices, free; Low Power—accountable (reverse-scored), submissive, restricted, powerless. As an additional manipulation check, participants completed a post-experimental questionnaire about how powerful they actually felt during the manipulation. Questions included, “As you wrote about your life, to what extent did you feel as if you had power over someone else?” and, “As you wrote about your life, to what extent did you feel as if another person had power over you?” Each of these questions was rated on a 7-point scale anchored by either “Not at all,” and “Very much,” or “I agree completely,” and “I disagree completely.” See Appendix A for the complete list. Demographic questions were also included.


After giving written consent to participate, participants were run in groups of 1-4. In Study 1, participants were then seated at computers in separate rooms (in Study 2, participants were first informed about a $100 drawing and resource allocation task—see below). Participants were randomly assigned to one of three conditions: high social power (HSP), high personal power (HPP), or low power (LP). HSP and LP participants completed an experiential priming task commonly used to manipulate a sense of high or low social power (Anderson & Berdahl, 2002; Galinsky et al., 2003; Galinsky et al., 2006; Smith & Trope, 2006) in which they wrote about a time when they had power over someone (HSP) or when someone had power over them (LP). HPP participants wrote about a time when they could do whatever they wanted, without interference. See Appendix B for the complete set of instructions.

Study 2 added a second task to amplify the prime (Galinsky et al., 2006). HSP participants were instructed that they had been randomly assigned to a “resource allocation task” in which they were to divide seven chances to win in a drawing for $100 between themselves and one other randomly selected participant. LP participants were instructed that they had been assigned to the resource allocation task, but were not to do the allocating. Instead, they would guess how many chances out of 7 another participant would give them. HPP participants were not assigned to the resource allocation task, but asked which of several prizes they would prefer, should they win the drawing. See Appendix C for the full text of these tasks.

Following this, participants completed the “E” task and the DANVA, counterbalanced. Next, participants responded to the questionnaires described above: UA, WHO-5, UC, RWA, adjectives, and self-esteem measure. This was followed by the post-experimental manipulation check questions and demographics questions, as well as written suspicion probes. Participants were then probed for whether they had guessed the connection between the manipulation and measures debriefed, thanked, and excused.


The central research question of this study was how personal power would affect perspective-taking compared to low power and social power. To address this, participants were primed with either low power (LP), high personal power (HPP), or high social power (HSP). Based on past research (Galinsky et al., 2006), we expected high social power to decrease perspective-taking relative to low power. Additionally, we hypothesized that high personal power would decrease perspective-taking, but not to the same extent as high social power.

Contrary to our hypotheses, the primary dependent variables (the direction participants drew their Es and scores on the DANVA), did not vary significantly across condition. Because previous research (i.e., Galinsky et al., 2006) has shown that the same methodology used in the current study reliably affected perspective-taking, we undertook to discover why this finding was not replicated in our sample. As a first step, we assessed the reliability of the DANVA. This analysis revealed low internal reliability of the measure (Cronbach’s α = .345). To improve the reliability of the measure, we used the criterion of removing any item that had an item-total correlation lower than .1 (see Patterson & Stockbridge, 1998, for a similar method). Eleven of the 24 DANVA items were thus removed, based on this criterion, increasing reliability to .49. The item-total correlation of one further item dropped below .1 after adjusting the scale, so this final item was removed, leading to a final scale that contained 12 items. The reliability of the adjusted DANVA was still low compared to the alpha of .77 reported for college students by Nowicki and Carton (1993), but improved somewhat (Cronbach’s α = .504). Rerunning the previous analysis using the adjusted DANVA, however, returned the same result: no significant differences across condition in DANVA scores.

Next, we assessed whether there were any significant differences across conditions on any of the questionnaire measures that were included (UA, UC, WHO-5, RWA, Self-esteem). Although the majority of these questionnaires did not differ, the global self-esteem measure that asked participants how much they liked their names differed significantly across conditions, F(2, 217) = 5.79, = .004. Post-hoc tests revealed that participants in the low power condition (M = 7.80,SD = 1.52) liked their names significantly more than participants in the high personal power (M= 7.34, SD = 1.80) and high social power (M = 6.75, SD = 2.21) conditions, = .005, and that the high personal and high social power conditions differed marginally, = .058. That is, as measured by name-liking, LP participants had significantly higher self-esteem than both HPP and HSP participants, and HPP participants had marginally higher self-esteem than HSP participants. That self esteem varied by condition is most likely a contrast effect. Intuitively, participants primed with low power should have lower self esteem than those primed with high power, unless being made to think of themselves as low in power caused them to react and reassert themselves, increasing name-liking. This idea is supported by other research, where well-being and self-esteem were positively correlated, and name-liking predicted well-being better than explicitly measured self-esteem (Gebauer et al., 2008), while in this study, name-liking was positively correlated with well-being only in the HPP condition.

Manipulation Check

Because our primary hypotheses were not supported, and improving the reliability of the DANVA did not affect this finding, we turned to another idea: Perhaps the reason for the null effect is that power was not reliably manipulated. To test this, we checked the effectiveness of our manipulation, which had been assessed in two ways. On one questionnaire, participants rated themselves on a series of low, personal, and social power-related adjectives. On another questionnaire, participants answered either two or three questions about how much low, personal, and social power they had felt during the power-priming task.

Correlations between the four power-related adjectives and the questions meant to assess each type of manipulated power were examined. In each case, the correlations supported averaging measures together, forming three final composite manipulation check scales. For each, the reliability was adequate (low power α = .63, personal power α = .61, social power α = .64). Next, three separate one-way analyses of variance (ANOVAs) were run for each of the composite measures.

In the first, the composite measure for low power differed significantly across condition, (2, 178) = 20.32, < .001. Post-hoc tests showed the low power condition (M = 3.57, SD = 1.14) scored significantly higher (ps < .001) than the high personal power (M = 2.47, SD = 1.07) and high social power (M = 2.58, SD = .93) conditions, which did not differ from one another. That is, participants in the low power condition felt significantly less powerful than participants in the other two conditions, who did not differ in their experience of low power.

In the next analysis, the composite measure for personal power also differed significantly across condition, (2, 178) = 8.69, < .001. In this case, the low power condition (M = 4.63, SD = 1.05) scored significantly lower than the personal power (M = 5.36, SD = 1.04) and social power (M = 5.23, SD = 1.01) conditions (ps < .005), and again, the personal and social power conditions did not differ. That is, participants in the low power condition felt less personal power than participants in the other two conditions, which did not differ.

Last, on the composite measure of social power, the three conditions again significantly differed,(2, 178) = 12.415, < .001. Post-hoc tests revealed that on this measure, participants in the social power condition (M = 4.14, SD = 1.00) felt significantly more social power than participants in the personal power (M = 3.28, SD = 1.05) and low power (M = 3.44, SD = .97) conditions (ps < .001), which did not differ in the amount of social power they felt.

This pattern of results suggests that the experimental manipulation had the intended effects: Participants in the low power condition felt less powerful than participants in the other two conditions, and also felt less personal power than participants in the other two conditions. At the same time, on both of these measures, participants in the personal and social power conditions did not differ, as expected. On the last measure, which assessed feelings of social power, people in the social power condition felt like they had more power over others than did participants in the low power and personal power conditions, which did not differ.

Exploratory Correlational Analyses

After finding that the results of our primary tests did not support past research or our hypotheses, we decided to explore the data further. For example, although the manipulation check measures suggested that people in the different conditions actually did experience power in different ways, it is possible that the subjective experiences of power as measured by the manipulation checks themselves (i.e., rather than the manipulation) might shed some light on the differential effects of low, personal, or social power on perspective-taking and other outcome measures. Examining patterns of correlations between the manipulation check composite items (i.e., experienced power) and other variables allowed us to explore these questions.

Initially, correlations of the experienced power items (i.e., the composite manipulation check items) with each other were computed. As can be seen in Table 1 (along with the correlations of all the major variables), low power was negatively correlated with both social power (r = -.14, p= .055) and personal power (r = -.55, < .001). Although personal and social power were not significantly correlated across condition, within conditions, an interesting pattern emerged. In the social power condition, there was a significant and positive association between personal and social power (r = .39), but in the personal power condition, they were negatively correlated (=

-.21). These correlations significantly differ using Fisher’s r to z transformation, z = 3.34, p < .001. This provides further evidence that the experience of personal and social power can be differentiated, and that the manipulation served to do so effectively (see Table 2 for correlations of measures within conditions). This suggests that participants who had been reminded of a time when they were in control of themselves evaluated social power differently than those reminded of a time when they had power over someone else; in the social power condition, personal and social power were associated, but in the personal power condition, they were negatively associated. Perhaps, from a standpoint of social power, more social power means more freedom, but from a standpoint of personal power, more social power just means more responsibility.


The current study set out to determine how personal power, a type of power that in the literature has been potentially confounded with social power, would differentially affect perspective-taking compared to social power and low power. We hypothesized that, as in past research (Galinsky et al., 2006), participants who were primed with social power would exhibit less perspective-taking than people primed with low power. In addition, we hypothesized that being primed with personal power, or the feeling of not being under anyone else’s control and able to take charge of one’s own destiny, would have effects on perspective-taking somewhere between the two. Our results, however, did not support these hypotheses. The primary dependent variables—the direction of an E drawn by participants on their own foreheads and a facial emotional expression recognition task—did not significantly differ across condition. Further exploration of the data, however, did allow a few conclusions to be drawn.

Using composite measures of power that were created from what were originally manipulation check items, it appears that there are some clear distinctions between social and personal power, particularly in how each type of power relates to the experience of not having power at all. In support of this, experiences of low power and high personal power were strongly negatively correlated, while low power and high social power were only marginally correlated. This was true even though the definition of low power provided was in terms of social power: being under the control of someone else. This indicates that low power and personal power might be part of a single dimension, while low power and social power are less related. This supports the view that personal power is a broader category than social power. Furthermore, it suggests that similarly to how people may seek social power as a way of increasing their personal power (e.g., Van Dijke & Poppe, 2006), people who have little power also are less concerned about others having power over them than lacking power over themselves. In further support of this, the same social and personal power measures were positively correlated in the social power condition, and negatively correlated in the personal power condition, indicating that participants primed with social power failed to make a clear distinction between social and personal power, while participants primed with personal power not only made the distinction, but conceived of the types of power as negatively related. If this finding can be strengthened by future research, it could have implications for power theorists: Not only do people seem to prefer personal to social power (Van Dijke & Poppe, 2006), but a reminder of personal power is all it takes to make the distinction salient. 

Although it is difficult to interpret null results, the lack of significant findings in this case implies several possibilities. It is possible that, despite our best efforts, our experimental procedure was flawed in some way. For example, some element in our instructions may have been restricting the range of outcomes for the DANVA scores and E direction: In the previous research that we were trying to replicate by using this design element, Galinsky et al. (2006) found that 33% of participants in the high social power condition drew a self oriented E compared to 12% for the low power participants, while we found 29% of high social power participants drew a self-oriented E and 31% of low power participants, a difference in magnitude that was not only much smaller, but in the opposite direction. Galinsky and colleagues also found that high social power participants made an average of 4.54 errors on the DANVA and low power participants made 3.11, while we found that high social power participants made an average of 4.20 errors, and low power participants made an average of 4.16 errors. These differences suggest that while the social power manipulation might have had the intended effect, participants in the low power condition did not respond in the same way in the current study as in Galinsky et al. (2006).

Another possibility is that the relationship between power and perspective-taking, at least as perspective-taking is measured by the variables used here and in Galinsky et al., may not be a stable one. It may be subject to change, for example, based on minor situational variables such as the wording of auxiliary instructions. Last, our inability to replicate past research on the topic might imply a difference between the populations of the human subjects pools at the University of Oregon who take part in studies for partial class credit, and the paid participants found at business and public service graduate schools used by Galinsky and colleagues. Perhaps power was particularly salient to their population or particularly unimportant to ours.

Personal and social power

One limitation of this research is that personal and social power were still confounded in both the low power condition and the high social power condition. Ideally, some way might be found to further disentangle these types of power, although this might be difficult. Theoretically, low and high social power are not distinct from personal power, but social power is one of the routes by which people seek personal power. The question of whether it is possible to separate these effects may depend on how we conceptualize personal power.

Personal power, or the ability to control one’s own outcomes, has been related to independence, agency, competence, feelings of expertise, autonomy, and personal causation (Overbeck & Park, 2001; Van Dijke & Poppe, 2006). Personal causation, “the initiation by an individual of behavior intended to produce a change in his environment” (DeCharms, 1968, p. 6), is related to personal power, but only tangentially so; an individual with personal power would likely exhibit this initiating behavior, but the behavior is not synonymous with personal power. Agency defined as self-efficacy (see, for example, Bandura, 1989, and Van Dijke & Poppe, 2006) is very closely related to personal power, but another psychological definition of agency—as a focus on self and separation (e.g., Abele, Uchronshi, Suitner, & Wojciszke, 2008; Ghaed & Gallo, 2006; Hegelson & Fritz, 1999)—confuses the issue in much the same way that operationalizing social power as dominance orientation has been a problem for social power theorists (e.g., Hall et al., 2005). Similarly, autonomy, which is commonly understood to mean self-governance, is a near synonym of personal power. Self-determination theorists, however, distinguish autonomy, defined as the extent to which an individual personally endorses the activities they engage in, from independence, defined as the extent to which people are reliant on others’ resources, assistance, or permission (e.g., Chirkov, Ryan, Kim, & Kaplan, 2003; Ryan & Deci, 2000). This leaves open the possibility that individuals whose behaviors are tightly controlled by others, but who enjoy the tasks they are given, or who feel that the control is justified and wise, could be labeled personally powerful, which stands in contrast to our working definition of personal power.

It may be then that personal power needs to be defined as a combination of other factors: In order to control one’s own outcomes, there must be the freedom to do so (i.e., independence as defined above), but also the ability to do so, or competence. If this is the case, then only one of these preconditions to personal power is confounded with social power—the extent to which we have independence is constrained by the social power others hold over us, while competence is a dimension of personal power unrelated to social power.

With this understanding of personal power, it may be possible in future research to create experimental conditions of both high and low personal power that do not vary as a function of social power. One possibility is using a manipulation of task difficulty, as in studies on the effects of control deprivation (Ric, 1997; Ric & Scharnitzky, 2003) and learned helplessness (Hiroto & Seligman, 1975; Simkin, Lederer, & Seligman, 1983). In these paradigms, participants in a “no-control” group are given some form of unsolvable task, putting them in a position of non-social, low personal power; their inability to control their own outcomes stems from the difficulty of the task, not from the control of someone else. A symmetrical non-social, high personal power condition could be created using practice to achieve competence at a task, or relative mastery of a task as a quasi-independent variable.

In its two-faceted sense, personal power is a broad concept, including social power as one of its limiting or enabling components. It should also be possible and instructive, however, to create experimental conditions in which social power varies independently from personal power. Participants in both low and high social power conditions could be instructed to perform the same task, perhaps relaying orders from superiors to inferiors in a hierarchy, and the number of superiors and inferiors could be manipulated: Low social power participants could be placed low in the hierarchy, with many superiors and only one underling, while high social power participants could be placed in the opposite position, second from the top, with many underlings.

There is also a need for control groups in power research. While high and low power are conceptualized as having symmetrical effects, with high power increasing positive affect and automatic social cognition, for example, and low power increasing negative affect and controlled social cognition (e.g., Keltner et al., 2003), most power research has not used control groups. In research that has used a control group, results have sometimes been asymmetrical. For example, in some experiments the low power condition was not different from the control, and in others it was the high power condition that was not different from the control (Smith & Trope, 2006). Without controls, it is impossible to tell whether it is high or low social power that is responsible for effects that are found, or whether the effects are symmetrical, as predicted by approach-inhibition theory. Control groups could also be helpful in determining what outcomes are a direct result of power and which are mediated by stress or mood.

Future research

How does personal power affect perspective-taking compared to social power? To the extent that these types of power are confounded, their effects will be equivalent, but it is also possible that their effects will remain equivalent even if they can be disentangled. Intuitively, personal power should produce an approach orientation. It may be, however, that personal power negatively affects perspective-taking more than social power: Personal power, on a broader continuum, has more extreme positions of power available; imagine being at the top of your hierarchy and a master of the skills required of you, versus at the bottom of your hierarchy and, additionally, inept. Social power, on the other hand, involving relationships by definition, may be associated with increased perspective-taking compared to personal power; at least someperspective-taking is necessary when dealing with subordinates. There is some evidence for this line of reasoning: Tjosvold and Deemer (1980) found that independence decreased perspective-taking compared to dependence, which in turn decreased perspective-taking compared to interdependence. To the extent that personal power is akin to independence, low power to dependence, and social power to interdependence, then, personal power should decrease perspective-taking more than low power, which in turn should decrease perspective-taking more than social power.

On the other hand, it may be that personal power affects perspective-taking less than social power does. The subjective sense of power has been shown to mediate at least some of the effects of social power (Anderson & Berdahl, 2002), and the cues of status and dominance that social power provides may be important to social power’s effect on perspective-taking.

Broadening our view beyond perspective-taking, this general line of research—distinguishing between the psychological outcomes of social and personal power—could be useful in expanding approach-inhibition power theory to include personal power. Many questions remain. Does high and low personal power come along with the same biological markers as social power, such as increased testosterone or glucocorticoids (Sapolsky, 2004)? How would a non-social, personal power condition compare to social power on measures of relative action, abstract thinking, stereotyping, executive functioning, and risk taking? In which areas are the effects of power the results of status and dominance cues and in which are they the results of freedom and efficacy?

The “empowerment” of oppression theorists such as feminists and other post-modern philosophers, anarchists and other libertarian philosophers, post-modern sociologists, and some alternative-education proponents, is often construed as more than just a human right. It is a benign form of power, (theoretically) coming with all of the benefits and few of the drawbacks and ethical quandaries inherent in hierarchical power. And empowerment, the ability of individuals and groups to be independent, self-governing, and to take part in decision-making processes, has much in common with personal power. Power corrupts, but empowerment does not. By distinguishing the psychological effects of having personal power and social power, we can begin to examine this assertion as an empirical question.


Abele, A. E., Uchronski, M., Suitner, C., & Wojciszke, B. (2008). Towards an operationalization of the fundamental dimensions of agency and communion: Trait ratings in five countries considering valence and frequency of word occurrence. European Journal of Social Psychology, 38, 1202-1217.

Altemeyer, B. (1998). The other “authoritarian personality.” In M. P. Zanna (Ed.), Advances in Experimental Social Psychology, Volume 30 (pp. 47-92). St. Louis, MO: Academic Press.

Anderson, C., & Berdahl, J. L. (2002). The experience of power: Examining the effects of power on approach and inhibition tendencies. Journal of Personality and Social Psychology, 83,1362-1377.

Anderson, C., & Galinsky, A. D. (2006). Power, optimism and risk-taking. European Journal of Social Psychology, 36, 511-536.

Bandura, A. (1989). Human agency in social cognitive theory. American Psychologist, 44, 1175-1184.

Bargh, J. A., Raymond, P., Pryor, J. B., & Strack, F. (1995). Attractiveness of the underling: An automatic power -> sex association and its consequences for sexual harassment and aggression.Journal of Personality and Social Psychology, 68, 768-781.

Batson, C. D. (1991). The altruism question: Toward a social-psychological answer. Hillsdale, NJ: Erlbaum.

Bech, P. (1993). Rating scales for psychopathology, health status, and quality of life : a compendium on documentation in accordance with the DSM-III-R and WHO systems. New York : Springer-Verlag.

Chen, S., Lee-Chai, A. Y., & Bargh, J. A. (2001). Relationship orientation as a moderator of the effects of social power. Journal of Personality and Social Psychology, 80, 173-187.

Chirkov, V., Ryan, R. M., Kim, Y., & Kaplan, U. (2003). Differentiating autonomy and independence: A self-determination theory perspective on internalization of cultural orientations and well-being. Journal of Personality and Social Psychology, 84, 97-110.

Davis, M. H. (1983). Measuring individual differences in empathy: Evidence for a multidimensional approach. Journals of Personality and Social Psychology, 44, 113-126.

Davis, M. H., Conklin, L., Smith, A., & Luce, C. (1996). Effect of perspective taking on the cognitive representation of persons: merging of self and other. Journal of Personality and Social Psychology, 44, 113-126.

DeCharms, R. (1968). Personal causation: The internal affective determinants of behavior.New York: Academic Press.

Eisenberg, N., Murphy, B. C., & Shepard, S. (1997). The development of empathic accuracy. In W. J Ickes (Ed.), Empathic accuracy (pp. 73-116). New York: Guilford Press.

Fiske, S. T. (1993). Controlling other people: The impact of power on stereotyping. American Psychologist, 48, 621-628.

Fiske, S. T., & Berdahl, J. (2007). Social power. In A. W. Krulanski & E. T. Higgins (Eds.),Social psychology: Handbook of basic principles (2nd ed.) (pp. 678-692). New York, NY: Guileford Press.

Galinsky, A. D., Gruenfeld, D. H., & Magee, J. C. (2003). From power to action. Journal of Personality and Social Psychology, 85, 453-466.

Galinsky, A. D., Magee, J. C., Gruenfeld, D. H., Whitson, J. A., & Liljenquist, K. A. (2008). Power reduces the press of the situation: Implications for creativity, conformity, and dissonance.Journal of Personality and Social Psychology, 95, 1450-1466.

Galinsky, A. D., Magee, J. C., Inesi, M. E., & Gruenfeld, D. H. (2006). Power and perspectives not taken. Psychological Science, 17, 1068-1074.

Galinsky, A. D., & Moskowitz, G. B. (2000). Perspective-taking: Decreasing stereotype expression, stereotype accessibility, and in-group favoritism. Journal of Personality and Social Psychology, 78, 708-724.

Gebauer, J. E., Riketta, M., Broemer, P., & Maio, G. R. (2008). “How much do you like your name?”: An implicit measure of global self-esteem. Journal of Experimental Social Psychology,44, 1346-1354.

Ghaed, S. G., & Gallo, L. C. (2006). Distinctions among agency, communion, and unmitigated agency and communion according to the interpersonal circumplex, five-factor model, and social-emotional correlates. Journal of Personality Assessment, 86, 77-88.

Hall, J. A., Coats, E. J., & LeBeau, L. S. (2005). Nonverbal behavior and the vertical dimension of social relations: A meta-analysis. Psychological Bulletin, 131, 898-924.

Hass, R. G. (1984). Perspective taking and self-awareness: Drawing an E on your forehead.Journal of Personality and Social Psychology, 46, 788-798.

Hegelson, V. S., & Fritz, H. L. (1999). Unmitigated agency and unmitigated communion: Distinctions from agency and communion. Journal of Research in Psychology, 33, 131-158.

Hegelson, V. S., & Fritz, H. L. (1998). A theory of unmitigated communion. Personality and Social Psychology Review, 2, 173-183.

Hiroto, D. S., & Seligman, M. E. P. (1975). Generality of learned helplessness in man. Journal of Personality and Social Psychology, 31, 211-327.

Keltner, D., Gruenfeld, D. H., & Anderson, C. (2003). Power, Approach, and Inhibition.Psychological Review, 110, 265-284.

Keltner, D., & Robinson, R. J. (1997). Defending the status quo: Power and bias in social conflict. Personality and Social Psychology Bulletin, 23, 1066-1077.

Kipnis, D. (1972). Does power corrupt? Journal of Personality and Social Psychology, 24, 33-41.

Kohlberg, L. (1976). Moral stages and moralizations: The cognitive-developmental approach. In T. Lickona (Ed.), Moral development and behavior (pp. 31-53). New York: Holt, Rinehart & Winston.

Lammers, J., Gordjin, E. H., Otten, S. (2008). Looking through the eyes of the powerful.Journal of Experimental Social Psychology, 44, 1229-1238.

Nowicki, S., & Carton, J. (1993). The measurement of emotional intensity from facial expression.Journal of social Psychology, 133, 749-750.

Nowicki, S., & Duke, M. P. (2001). Nonverbal receptivity: The diagnostic analysis of nonverbal accuracy (DANVA). In J. A. Hall & F. T. Bernieri (Eds.) Interpersonal Sensitivity: Theory and Measurement (pp. 183-198). Philadelphia, PA: Lawrence Erlbaum Associates.

Overbeck, J. R., & Park, B. (2001). When power does not corrupt: Superior individuation processes among powerful perceivers. Journal of Personality and Social Psychology, 81, 549-565.

Patterson, M. L., & Stockbridge, E. (1998). Effects of cognitive demand on judgment strategy on person perception accuracy. Journal of Nonverbal Behavior, 22, 253-263.

Ric, F. (1997). Effects of control deprivation on subsequent use of stereotypes. Journal of Social Psychology, 137, 333-342.

Ric, F., & Scharnitzky, P. (2003). Effects of control deprivation on effort expenditure and accuracy performance. European Journal of Social Psychology, 33, 103-118.

Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American Psychologist, 55, 68-78.

Sapolsky, R. M. (2004). Why zebras don’t get ulcers: The acclaimed guide to stress, stress-related diseases, and coping (3rd ed.)New York: Henry Holt and Company.

Simkin, D. K., Lederer, J. P., & Seligman, M. E. (1983). Learned helplessness in groups.Behavior Research and Therapy, 21, 613-622.

Smith, P. K., Jostmann, N. B., Galinsky, A. D., & van Dijk, W. W. (2008). Lacking power impairs executive functions. Psychological Science, 19, 441-447.

Smith, P. K., & Trope, Y. (2006). You focus on the forest when you’re in charge of the trees: Power priming and abstract information processing. Journal of Personality and Social Psychology, 90, 578-596.

Snodgrass, S. E., Hecht, M. A., & Ploutz-Snyder, R. (1998). Interpersonal sensitivity: Expressivity or perceptivity? Journal of Personality and Social Psychology, 74, 238-249.

Spence, J. T., Helmreich, R. L., & Holahan, C. K. (1979). Negative and positive components of psychological masculinity and femininity and their relationships to self-reports of neurotic and acting out behaviors. Journal or Personality and Social Psychology, 37, 1673-1682.

Tjosvold, D., & Deemer, D. K. (1980). Effects of controversy within a cooperative or competitive context on organizational decision making. Journal of Applied Psychology, 65, 590-595.

Van Dijke, M., & Poppe, M. (2006). Striving for personal power as a basis for social power dynamics. European Journal of Social Psychology, 36, 537-556.

Walker, L. J. (1980). Cognitive and perspective-taking prerequisites for moral development.Child Development, 51, 131-139.

Winter, D. G., & Barenbaum, N. B. (1985). Responsibility and the power motive in women and men. Journal or Personality, 53, 335-355.

Weick, M., & Guinote, A. (2008). When subjective experiences matter: Power increases reliance on the ease of retrieval. Journal of Personality and Social Psychology, 94, 956-970.

Zakrisson, I (2005). Construction of a short version of the right-wing authoritarianism (RWA) scale. Personality and Individual Differences, 39, 863-872.

Appendix A

Post-Experimental Questionnaire

All questions were answered on a 7-point scale. Questions 1 through 3 were anchored by 1 = “Not at all” and 7 = “Very much.” Question 4 was anchored by 1 = “Not at all powerful” and 7 = “Very powerful.” Questions 5 through 8 were anchored by 1 = “I agree completely” and 7 = “I disagree completely,” and reverse coded.

For the following 8 questions, please select the answer that best describes your experience.

1. As you wrote about your life, to what extent did you feel as if you had power over someone else?

2. As you wrote about your life, to what extent did you feel as if another person had power over you?

3. As you wrote about your life, to what extent did you feel as if you had power over yourself?

4. As you wrote about your life, how powerful did you feel?

5. As you wrote about a situation in your life, you felt as if you had control over someone else’s destiny.

6. As you wrote about a situation in your life, you felt as if someone else had control over your destiny.

7. As you wrote about a situation in your life, you felt as if you had control over your own destiny.

8. As you wrote about a situation in your life, you felt very powerless.

Appendix B

High Social Power Writing Task Instructions

On the next screen, you will be asked to spend about 5 minutes thinking about and writing about a time in your life when you felt like you had power over someone else. That is, you will write about a time when you had control over someone else’s actions or something that they wanted. Any time or event that fits this description will do. When it happened or the context in which it happened is not important – for example, it could concern school, family, work, or your personal life.

This task is important. Therefore, before continuing and writing about this time in your life, spend a moment RIGHT NOW to think of the time or event in your life that best fits this description. When you have this time or event firmly in mind, you may click “continue” and begin writing

Low Power Writing Task Instructions

On the next screen, you will be asked to spend about 5 minutes thinking about and writing about a time in your life when you felt like someone else had control over your actions or something that you wanted. Any time or event that fits this description will do. When it happened or the context in which it happened is not important – for example, it could concern school, family, work, or your personal life.

This task is important. Therefore, before continuing and writing about this time in your life, spend a moment RIGHT NOW to think of the time or event in your life that best fits this description. When you have this time or event firmly in mind, you may click “continue” and begin writing.

High Personal Power Writing Task Instructions

On the next screen, you will be asked to spend about 5 minutes thinking about and writing about a time in your life when you felt like you had complete control over your own life and were free to do whatever you wanted, without interference. Any time or event that fits this description will do. When it happened, or the context in which it happened, is not important – for example, it could concern school, family, work, or your personal life.

This task is important. Therefore, before continuing and writing about this time in your life, spend a moment RIGHT NOW to think of the time or event in your life that best fits this description. When you have this time or event firmly in mind, click “continue” to progress to the next screen.


Appendix C

High Social Power Resource Allocation Task Instructions


You have been assigned to the resource allocation task and are in charge of deciding how many tickets for the drawing for $100 that you and the other person involved in this task each will receive.

Decide on a division of 7 tickets between yourself and the other person:

7                      7 for yourself and 0 for the other person

6                      6 for yourself and 1 for the other person

5                      5 for yourself and 2 for the other person

4                      4 for yourself and 3 for the other person

3                      3 for yourself and 4 for the other person

2                      2 for yourself and 5 for the other person

1                      1 for yourself and 6 for the other person

0                      0 for yourself and 7 for the other person

Low Power Resource Allocation Task Instructions


You have been assigned to the resource allocation task, but you do not have control over the allocation of the tickets for the drawing. Instead, predict how many tickets the person making the allocation decision will take and how many will be given to you.

Remember not to pretend you are determining the allocation. You should try to predict what the person making the decision is going to do.

0          0 for you and 7 for them

1          1 for you and 6 for them

2          2 for you and 5 for them

3          3 for you and 4 for them

4          4 for you and 3 for them

5          5 for you and 2 for them

6          6 for you and 1 for them

7          7 for you and 0 for them

High Personal Power Resource Allocation Task Instructions


As part of this study, with your consent, you will be entered into a drawing at the end of the study. You will have seven full chances to win.

There are four prizes to choose from. Please choose which of these prizes you would prefer, should you win. If you choose number 4, a text box will open for you to specify which business you would like a gift certificate from.

1          $100 cash

Appendix C continued

2                      $110 gift certificate at Market of Choice on Franklin Blvd.

3                      Bose Triport In-Ear Headphones, worth $99.95

4                      $105 gift certificate to local or online business of your choice <specify>

5          Opt out of drawing

Table 1

Correlations and reliability (Cronbach’s α) of major variables

E           DANVA UC            UA         WHO-5      RWA     Name     LP comp   HPP comp  HSP comp


DANVA     -.039         .504

UC             -.028         .224*      .651

UA              -.075         -.195*     -.232*    .693

WHO         -5 .047      -.004       -.031      -.132*      .698

RWA          -.068         .058        -.021      .017         .065       .820

Name         -.122         .000         -.115      -.008        .121       .132

LP comp     .123          .159         .157*     -.001        -.225*    .011       -.011         .632

HPP comp  .021          .013         -.111      -.018        .201*     -.074      .046          -.554*    .608

HSP comp  -.006         .084       -.024      .109       .207*     .164*     .081          -.143      .089           .635

Note. Reliability coefficients are on the diagonal. E = E direction (coded 1 = perspective-taking, 0 = no perspective-taking); DANVA = adjusted DANVA scores; UC = Unmitigated Communion Scale; UA = Unmitigated Agency Scale; WHO-5 = the WHO-5 Well-Being Index; RWA = Right-Wing Authoritarianism Scale; Name = self-esteem measure; LP comp = composite low power measure; HPP comp = personal power composite measure; HSP comp = composite social power measure.

* Correlation is significant, p < .05

Table 2

Correlations of composite power variables with other dependent variables

Condition  ____

Variables LP HPP HSP Overall

LP comp HPP comp                   -.421a*    -.561a*    -.502a*    -.554a*

HSP comp                                    -.233a -.043a -.100-.143a

HPP comp  HSP comp              .089abc -.213a .388b*     .089c

E direction LP comp                 .212a .234a -.056a .123a

HPP comp                                   .021a -.112a .   160a .021a

HSP comp                                   -.098a -.171a .267b*     -.006ab

DANVA LP comp                      .020a .038a .397b*     .159ab*

HPP comp                                  .205a -.200b .115ab .013ab

HSP comp                                 -.139a .232.140ab .084a

UA          LP comp                     .040a .016.096-.001a

HPP comp                                 -.126a .042a -.061-.018a

HSP comp                                .146a -.008a .083.109a

WHO-5   LP comp                   -.142a -.278a*    -.369a*    -.225a*

HPP comp                                 .239a .204a .191.201a*

HSP comp                                .104a .219a .371a*     .207a*

UC          LP comp                    .303a*     .211a .143.157a*

HPP comp                                -.011a -.351a*    -.097-.111a

HSP comp                                 .039a      -.002a     -.135a     -.024a

RWA      LP comp                    -.043a .115a -.103.011a

HPP comp                               -.039a -.071a -.082-.074a

HSP comp                                  .135a .207a .223.164a*

Name-liking    LP comp         -.036a -.262a*    -.001-.011a

HPP comp                                 .005a .163a .126a         .046a

HSP comp                                  .026a .284a*     .128.081a

Note. Within rows, correlations with different subscripts differ significantly, < .05, using Fisher’s r to z transformations. LP comp = composite low power measure; HPP comp = personal power composite measure; HSP comp = composite social power measure. For condition, LP = low power condition; HPP = personal power condition; HSP = social power condition. DANVA = adjusted DANVA scores; UA = Unmitigated Agency Scale; WHO-5 = the WHO-5 Well-Being Index; UC = Unmitigated Communion Scale; RWA = Right-Wing Authoritarianism Scale; Name-liking = self-esteem measure.

* Correlation is significant, p < .05