A manager signs the final approval on a project. A specialist without a senior title has shaped every option that reached the manager’s desk. Who had power? The signature records formal authority; the specialist’s control over information may have narrowed the choice before the decision. Measuring organizational power means specifying whose behavior or outcome could be changed, through which resources or dependencies, at which stage of a decision. One hierarchy chart or self-rating cannot settle all those questions.
Organizational power is a capacity to affect outcomes or other actors under particular conditions. It can arise from formal rights, resource control, specialized knowledge, network position and alternatives available to others. It differs from status, which concerns esteem or respect, and from influence, which concerns observed change in a particular interaction. A person can hold power without exercising it in every episode, or be admired without controlling a budget. This Methods & Measures entry focuses on choosing evidence for a defined unit, decision and time window.
Choose the level and outcome before the indicator
Power can be attributed to a person, role, team, department or outside stakeholder. Those units have different resources and targets. A department may be influential in budget negotiations while a junior analyst controls an indispensable dataset. A supervisor may decide schedules but not product strategy. A study should name the unit and the domain of possible control. Otherwise a single score invites the false claim that someone is generally “powerful” across all decisions.
The outcome matters as much as the title. Do we ask who can allocate funds, set the agenda, veto a proposal, assign work, withhold information or shape performance criteria? Each requires different records. A formal delegation schedule can establish authorized decision rights. Meeting notes may show agenda control. Resource records can show allocation, though an allocation also depends on need, policy and history. An interview can reveal perceived influence, but perceptions are themselves shaped by fear and reputation.
Finkelstein’s study of top management teams proposed multiple dimensions of executive power and a measurement approach rather than one proxy. Its particular population does not define every workplace, yet the lesson travels: structural, ownership, expert and prestige indicators should be described separately before being combined. An index that mixes them must explain why its components belong together and whether one dimension can compensate for another.
There is a temporal distinction between capacity and exercise. A leader may have a veto but never use it because everyone anticipates the veto and adapts. Counting vetoes would understate capacity. Conversely, someone may win one dispute through a persuasive argument despite lacking stable authority. That episode demonstrates influence under those conditions, not necessarily enduring organizational power. Repeated decisions and records of anticipated constraints can help distinguish the two.
Map dependence instead of reading the chart literally
A formal hierarchy is a starting map, not the full network of dependence. A team that solves an urgent uncertainty or controls a non-substitutable workflow may gain power beyond its rank. Hickson and colleagues’ strategic-contingencies framework identifies uncertainty coping, centrality in workflow and substitutability as reasons for subunit power. The framework is a theory to operationalize, not proof that every specialist automatically dominates a decision.
Operational measures might ask which units must approve a handoff, whose absence delays a task, what expertise can be replaced and how quickly, and whose resources have viable alternatives. These can be measured through process maps, logs, interviews and observed contingencies. A high count of incoming requests may signal dependence, but it may also signal routine administrative workload. The investigator needs to connect the indicator to control over a valued outcome.
Salancik and Pfeffer examined power in university decisions and related resource and decision evidence in a defined setting. Their case reminds us that the basis and use of power can diverge. A powerful department may shape the rules by which allocations are justified, not merely receive a larger final amount. If the measure starts only at the final distribution, it may miss the earlier choice of criteria and the exclusion of alternatives.
Network position provides another angle. Brass’s organizational study connected structural position with individual influence. A person who bridges otherwise separate groups can affect what information circulates, but centrality is not a magic causal score. A highly central coordinator may be overburdened and unable to choose outcomes. A peripheral person may hold a unique legal or technical veto. Report which ties are mapped, who nominated them and whether the network reflects advice, communication, approval or friendship.
Network surveys can miss people who are feared or avoided, and electronic logs capture channels that leave digital traces. Informal conversations may be decisive. Combining network data with decision histories can test whether a structural position translated into an outcome. It also helps avoid reading every tie as an expression of voluntary cooperation. A subordinate may communicate frequently with a manager because reporting is mandatory.
Separate felt power from control
The Personal Sense of Power scale developed by Anderson, John and Keltner measures a person’s perceived ability to influence others in a specified relationship or context. That perception is meaningful: it can shape behavior and willingness to speak. It is not a complete inventory of formal authority or actual resource control. A confident respondent may overstate options; someone with a real veto may feel constrained by norms or consequences.
Self-reports can be paired with peer reports, formal records and behavioral observation. Agreement across methods may strengthen an interpretation, but disagreement can be informative. If colleagues say a person is influential while the person reports little power, perhaps the individual has hidden constraints or perhaps peers mistake visibility for control. Do not average contradictory measures into one number before asking what each observes.
Asking directly about power can itself be sensitive. Staff may fear that naming an informal gatekeeper will expose them. Anonymous surveys reduce some pressure but cannot guarantee candor in a tiny team. Interviews should protect confidentiality and avoid leading questions that assume the hierarchy’s official account is true. Retaliation risk is part of measurement validity: a method that makes honest answers costly can systematically distort the record.
Cross-cultural and organizational norms affect how people describe authority. A respondent might report “we decide together” while one actor still sets the acceptable options. Another may call a manager powerful because they carry responsibility, though the actual budget is controlled elsewhere. Behavioral examples and concrete decision episodes can clarify what respondents mean without forcing a single cultural vocabulary onto them.
Study decisions as sequences
A decision trace begins before the meeting at which a vote is taken. Who defined the problem? Who had the data? Which options were never considered? Who could delay or accelerate a recommendation? Who approved implementation, and who could reverse it? These stages can locate power more precisely than a final yes-or-no result. The project in the opening example may have been settled by the specialist’s framing long before the manager signed.
Collect contemporaneous records where possible. Retrospective interviews are useful but can be revised after an outcome becomes known. Winners may describe the choice as inevitable; others may minimize their own influence. Timestamped drafts, agenda versions and policy documents can show how alternatives changed. Yet documents reflect formal channels and may omit an influential private conversation. Triangulation should preserve these gaps rather than present a seamless story it cannot verify.
A counterfactual question sharpens power inference: what would likely have happened if the actor had opposed the choice or withdrawn the resource? The answer cannot be observed directly for the same decision. Comparative episodes, policy thresholds or changes in staffing can offer partial evidence. Claims should remain conditional. Showing that an analyst’s report preceded a decision does not prove the report caused it, even if the analyst later receives credit.
Power can shift during a crisis. An expert may gain temporary authority when uncertainty rises; a routine manager may recover control once procedures stabilize. Repeated measurements across ordinary and exceptional episodes help reveal scope. A single organizational snapshot can confuse an emergency role with a stable hierarchy. A longitudinal study should record reorganizations, turnover and changing external constraints that alter what actors can control.
Test whether the score measures power
Construct validity asks whether the indicators behave as the proposed capacity would. A measure of budget power might predict ability to redirect funds within the rules, not simply job grade. A network brokerage measure might predict information access, not automatically favorable decisions. A self-report might predict willingness to challenge a superior more strongly than formal veto rights. These are distinct predictions; making them in advance prevents selective interpretation after the data arrive.
Discriminant validity is equally important. Status, expertise, popularity and performance can correlate with power but should not be assumed identical. A respected clinician may have high professional status yet little authority over staffing. An unpopular administrator may hold binding approval rights. Measuring both conditions separately lets the study test where they converge. A single “leadership” rating can conceal the difference.
Measurement invariance matters if comparing departments, organizations or periods. A title like “director” can carry different rights in different structures. A five-point survey item about influence can be understood differently by managers and frontline staff. Before interpreting group differences, inspect actual decision rights and item meanings. Where the same scale cannot be made comparable, report within-setting patterns rather than a falsely precise universal ranking.
Be cautious with indicators that may be consequences of power as well as its sources. Pay, office size and meeting invitations can follow from prior influence or policy. Treating them as causes without temporal evidence reverses the argument. Finkelstein’s dimensional approach is useful for organizing proxies, but a validated measure in one executive sample does not certify every proxy in a different workforce.
Another source of distortion is procedural power: the ability to decide which evidence counts and when a proposal is “ready” for approval. Salancik and Pfeffer’s university case illustrates why the criteria governing resource allocation deserve attention alongside the allocation itself. A department that persuades the organization to use a metric favorable to its work may influence repeated decisions without appearing as the winner of any single negotiation. To study this, preserve successive versions of criteria and ask who proposed, challenged and authorized them. A criterion can also be adopted for technical reasons, so authors should not attribute every favorable rule to a hidden contest.
Power is relational. The same actor can be powerful relative to one colleague and dependent on another unit that controls legal approval, data access or skilled labor. A measure that lists the actor’s assets without specifying others’ alternatives misses that relation. Hickson and colleagues’ attention to substitutability is especially useful here: scarce expertise matters differently when a replacement is available tomorrow than when a whole project would stall for months. Report the relevant alternative and time horizon. What counts as power in a one-day emergency may not count as power over a multiyear strategy.
Access to decision records is itself uneven. Senior staff may authorize an archive of formal meetings while informal bargaining remains invisible. Workers with little formal authority may be reluctant to describe how pressure was applied. A credible study can record these limitations in a data-source matrix and avoid interpreting silence as agreement. When possible, compare accounts from multiple roles and show where they conflict. Disagreement is not merely noise to average away; it can reveal that actors see different stages of the decision process or face different costs for speaking openly.
What the evidence does not show
The cited studies do not provide one universal organizational power score. They examine executives, subunits, decision processes, subjective power and network position at different levels. None shows that a job title alone captures the capacity to shape every outcome. Neither does a central network position prove control, nor does one decision victory establish enduring power.
Power measurement does not by itself determine whether its use was legitimate, fair or effective. An actor may control a scarce resource for good reasons, or use a formal right abusively. Those judgments require normative and outcome evidence beyond the structural indicator. A study should not equate more power with better leadership or lower power with lower competence.
The evidence also does not license confident causal stories from a single meeting transcript or retrospective account. Agenda setting, anticipated compliance and absent alternatives are hard to observe. A responsible report states the decision domain, sources of data and rival explanations, then limits the conclusion to what those records support.
Show who could change what
A strong report can tell readers who could approve, delay, reshape and implement a choice, with evidence at each stage. It can show formal authority beside dependence, network position, perceived power and actual decision influence. The specialist and the manager in the opening example may each have power over different parts of the process. The useful question is not which one “really” had it in general, but which capacity mattered for which outcome.
Organizational power becomes measurable when the study names the alternatives available to each actor and the conditions under which those alternatives mattered. That account may be less tidy than a hierarchy chart. It is more faithful to how decisions are made.
Research references
- Finkelstein, S. (1992). Power in top management teams: Dimensions, measurement, and validation. Academy of Management Journal, 35(3), 505–538. View source
- Hickson, D. J., Hinings, C. R., Lee, C. A., Schneck, R. E., & Pennings, J. M. (1971). A strategic contingencies’ theory of intraorganizational power. Administrative Science Quarterly, 16(2), 216–229. View source
- Salancik, G. R., & Pfeffer, J. (1974). The bases and use of power in organizational decision making: The case of a university. Administrative Science Quarterly, 19(4), 453–473. View source
- Brass, D. J. (1984). Being in the right place: A structural analysis of individual influence in an organization. Administrative Science Quarterly, 29(4), 518–539. View source
- Anderson, C., John, O. P., & Keltner, D. (2012). The personal sense of power. Journal of Personality, 80(2), 313–344. View source