Dacher Keltner’s approach-inhibition theory

Dacher Keltner and colleagues’ research on power and social cognition established the foundational mechanism. Their approach-inhibition theory of power predicts that elevated power produces a shift toward approach-oriented processing, the attentional orientation focused on personal goals and reward pursuit, and away from inhibition-oriented processing, the attentional orientation that monitors social context, attends to other people’s signals, and modulates behaviour in response to them. The high-power person, by the theory’s prediction, becomes more focused on their own goals and less sensitive to the social environment in which those goals are pursued.

The research confirmed the prediction. High-power individuals showed reduced attention to the people around them, lower accuracy in reading emotional states from facial expressions, and lower spontaneous perspective-taking than their lower-power counterparts in the same interaction. The reduction was not a matter of motivation or intention. The high-power participants were not choosing to pay less attention. The attentional reorientation was operating below the level of deliberate choice, as a function of the power state itself. The finding has been replicated across experimental and naturalistic settings and is among the more robustly established findings in the social psychology of power.

Adam Galinsky’s perspective-taking findings

Adam Galinsky and colleagues’ research on power and perspective-taking documented the most consequential specific effect for decision quality. Elevated power significantly reduced the accuracy of perspective-taking: the capacity to accurately represent another person’s mental state, their beliefs, concerns, and the information available to them. The high-power person took fewer perspectives, took them less accurately, and was more confident in the accuracy of their inaccurate perspectives than the low-power person.

The confidence component of the finding is the most practically important. The leader who has accumulated power not only becomes less accurate at representing their team members’ perspectives; they become more confident that their inaccurate representations are correct. The combination produces the specific epistemic condition in which the leader who most needs accurate social information is simultaneously the least likely to receive it, the least likely to seek it, and the most confident that they already have it. The Galinsky finding is the research account of why power corrupts epistemic quality even when it leaves moral integrity intact.

The information filtering mechanism

The Keltner and Galinsky findings describe what power does to the leader’s processing of social information. A complementary body of research describes what power does to the social information that reaches the leader in the first place. Research on information transmission in power-differentiated organisations consistently documents that information is systematically filtered on its way to the most powerful people. Bad news is softened. Disagreements are suppressed. The concerns of low-status team members are omitted, reframed, or delayed. The filtering is not primarily the result of deliberate deception; it is the result of the rational social calculation that every team member makes when deciding what to share upward with someone who has power over their position, opportunities, and recognition.

The founder who believes they are receiving accurate, complete information because they have encouraged open communication and signalled that they want to hear the truth is contending with a filtering mechanism that operates below the level of deliberate decision-making in every member of the team. The team member who softens bad news, delays sharing a concern, or omits a dissenting view before a meeting with the founder is not consciously calculating. They are responding to the social threat that power differentials install at the level of automatic social cognition. The founder’s stated preference for honesty is a far weaker regulatory force than the team member’s lived experience of what happens when honest assessments are shared with powerful people.

You might also like
The sunk cost problem: the psychological mechanism that keeps entrepreneurs committed to failing positions
The sunk cost is not an accounting concept. It is a psychological mechanism — and it operates most powerfully at exactly the moment when the evidence for changing course is clearest.
Why the people who speak first and most frequently in meetings are rarely the people with the best ideas
There is a persistent and costly assumption embedded in most meeting cultures: that the person who speaks most is the person who thinks most. The research on group participation and collective intelligence consistently contradicts this assumption. Speaking frequency in meetings correlates more reliably with extraversion, status, and the anchoring effect of early contribution than with the quality of ideas being contributed. The information most valuable to the decision is often held by the participant who speaks least, not most. The meeting format that privileges confident early talkers is not selecting for quality of thinking; it is selecting for comfort with social risk-taking and willingness to fill silence, which are genuine and useful attributes but not the same thing as the quality of the contribution being made.
Why entrepreneurs are poor at recovery and the psychological barriers specific to high-drive individuals
The barriers to recovery for high-drive entrepreneurs are not logistical — they are psychological, and they are more pronounced in this population precisely because the traits that make entrepreneurship possible are the same ones that make genuine recovery feel threatening.
Why the same leadership style that works in one country produces resistance in another
Leadership is among the most culturally loaded forms of human behaviour. Every leader develops a style within a cultural context that shapes what authority looks like, how trust is earned, when deference is appropriate, and what decisions a person in a position of responsibility should make alone versus with input from others. When that style is transported into a different cultural context — a cross-cultural hire, an international team, a founder who relocates — it carries those cultural assumptions with it. The research on cross-cultural leadership is consistent on what follows: the same behaviours that produce trust, commitment, and high performance in one cultural context reliably produce confusion, resistance, and disengagement in another.

The cookie study finding

The most accessible single demonstration of the power-disinhibition mechanism is Keltner’s research finding that high-power participants in a group task took more cookies from a shared bowl, ate them more messily, and were unaware of how their behaviour appeared to others. The finding is not about cookies. It is about the specific attentional reorientation that power produces: the high-power person has reduced their monitoring of the social signals that would otherwise moderate their behaviour, and they have lost awareness of the gap between their own perception of how they are behaving and how they appear to the people around them.

For the founder, the cookie study finding predicts the specific perceptual gap that power accumulation produces: the founder’s self-assessment of how they come across, how open they appear, how well they listen, and how their communications land is progressively diverging from the experience of the team members on the receiving end of those communications. The divergence is not visible from the inside, because the attentional mechanism that would close it has been reduced by the power state. The founder who is confident that their team feels heard and genuinely informed is operating on a self-assessment that the power research predicts is less accurate than it feels.

Amy Edmondson’s correction mechanism

Amy Edmondson’s psychological safety research predicts the specific structural correction for the power-information-filtering problem. The leader who deliberately creates the psychological safety conditions that make genuine upward information transmission possible, who explicitly invites challenge, responds to bad news with curiosity rather than anxiety, and shares their own uncertainty, is counteracting the filtering mechanism that their power differential otherwise installs. The psychological safety environment does not eliminate the power differential. It provides the social evidence that the rational filtering calculation is less necessary than the default power dynamic suggests.

The correction mechanism requires deliberateness because the power state that has reduced the leader’s attentional orientation toward social signals is the same state that makes the correction feel less urgent. The Keltner research predicts that the leader who most needs to actively seek genuine information is the leader who, left to their natural attentional orientation, will seek it least. The implication is that the structural habit of information-seeking, the regular, disciplined, protocol-supported practice of genuinely asking and genuinely listening, is more reliable than the motivational commitment to do so, because motivation is subject to the attentional reorientation that power produces and structural habit is not.

The self-compassion prerequisite

Kristin Neff’s self-compassion research predicts the foundational internal condition for the power correction. The leader who treats their own limitations, errors, and uncertainties with genuine self-compassion rather than defensive self-protection is the one who can genuinely receive and engage with the challenging information that their power differential would otherwise suppress. The defensive response to difficult feedback, the response that rationalises, dismisses, or reframes it in ways that preserve the existing self-assessment, is the internal analogue of the external filtering mechanism: it removes the accurate social information at the point of receipt rather than before it arrives.

The self-compassion mechanism predicts a specific and practically applicable relationship: the leader who is genuinely kind to themselves about their own limitations is the leader who least needs the information environment to be filtered before it reaches them, because they have the internal resource to receive and use unfiltered information without experiencing it as a threat to their self-concept. The power correction that Edmondson’s psychological safety research describes at the structural level is supported by the internal condition that Neff’s self-compassion research describes at the psychological level.

If the experiences described in this article are significantly affecting your wellbeing, speaking with a psychologist is the right next step. UK: Samaritans (116 123, free, 24/7). Mind (0300 123 3393). BACP: bacp.co.uk/search/Therapists. Crisis Text Line — text HOME to 741741 (US, UK, Canada, Ireland). International: internationaltherapistdirectory.com.

This article is for educational and informational purposes only. Sources: Keltner, D., Gruenfeld, D.H. and Anderson, C. (2003). Power, approach, and inhibition. Psychological Review, 110(2), 265-284. Galinsky, A.D., Magee, J.C., Inesi, M.E. and Gruenfeld, D.H. (2006). Power and perspectives not taken. Psychological Science, 17(12), 1068-1074. Edmondson, A.C. (1999). Psychological safety and learning behavior in work teams. Administrative Science Quarterly, 44(2), 350-383. Neff, K.D. (2003). Self-compassion: An alternative conceptualization of a healthy attitude toward oneself. Self and Identity, 2(2), 85-101. Diamond, J. (2016). Power: A User’s Guide. Belly Song Press. Schein, E.H. (2013). Humble Inquiry. Berrett-Koehler.