Analyzing Power: Empirical Challenges and Methodologies

How do you study something you cannot see directly? Power leaves footprints, but it rarely shows its face. A decision gets made, a project gets funded, a complaint gets buried, and somewhere behind these outcomes sits the question that political scientists have struggled with for decades: who actually held the power, and how would we ever prove it? The empirical study of power is the attempt to answer this question through evidence rather than assumption. It sounds straightforward. In practice, it is one of the most contested and difficult tasks in all of political theory.

Why power resists measurement 🔗

Power is central to political analysis, yet it is notoriously hard to quantify. A weight can be measured on a scale and a temperature with a thermometer, but power has no such instrument. It is relational, contextual, and often invisible. We usually only notice it when it produces a visible outcome, and even then we cannot be sure who exercised it or why.

The British historian Maurice Cowling captured a deeper version of this problem. He argued that politics is driven by a small circle of actors operating within institutions like Parliament, and that their real motives and calculations are largely closed off to outside observers. If the genuine workings of power happen behind closed doors, among people who have every reason to conceal their intentions, then the researcher is left studying shadows on the wall. This is the access problem: the people who hold power are rarely willing to be studied, and the most important decisions are often the least documented.

This challenge becomes sharper in democracies. In authoritarian systems, power may be concentrated and visible in a single ruler or party. In a democracy, power is dispersed across legislatures, bureaucracies, courts, parties, interest groups, and the electorate. Influence flows through committees, informal networks, and negotiations that leave little public trace. The very features that make democracy attractive also make its power structures fiendishly difficult to map.

The problem of defining what to measure 🔗

Before researchers can measure power, they must agree on what power is, and here consensus collapses. One influential definition, advanced by Robert Dahl, treats power as the ability of one actor to make another do something they would not otherwise do. This definition has a useful quality: it points toward something observable, namely a change in behaviour following an attempt to influence. But critics such as Steven Lukes argued that this definition failed to capture other dimensions of power, such as the capacity to shape the very norms and values that people hold. If power can operate by changing what people want, then measuring only observable conflict misses its most profound effects.

Dahl and the New Haven study 🔗

The most famous attempt to study power empirically is Robert Dahl’s Who Governs?, published in 1961. Dahl chose a single city, New Haven, Connecticut, and asked a deceptively simple question: who really runs this place? Rather than assuming that a hidden elite controlled everything, he set out to trace power through actual decisions.

His method was the heart of his contribution. Dahl examined three distinct areas of local policy: political nominations, urban redevelopment, and public education. Through detailed interviews, he tried to identify the specific individuals who shaped the outcome in each area. The findings were striking. Different people held power in each policy area, and only the mayor, whose position depended on voters, exercised influence across all three. From this Dahl concluded that power in the city was not monopolised by one ruling group but distributed among competing interests, a position known as pluralism.

Dahl was reacting against an earlier study by Floyd Hunter, who had used a reputational method, asking residents to name the most powerful people in their city. Hunter found that a small group of businessmen formed a unified power elite. Dahl rejected this approach. He believed that asking people who they thought was powerful told you about reputations, not about who actually prevailed when real decisions were contested. By studying concrete decisions instead, he claimed to find pluralism where Hunter had found domination.

The logic of the small-scale study 🔗

Notice what Dahl did. He took one city, studied it in depth, and used his findings to make a broad argument about the nature of American democracy. This is a recurring strategy in the empirical study of power. Because power is so hard to observe at the national or societal level, researchers narrow their focus to a setting small enough to study carefully. A single community, a single institution, or a single set of decisions becomes a window into a much larger system.

The appeal is obvious. A researcher cannot interview every powerful person in a nation, but they might reasonably map the key actors in one town. The depth of understanding gained in a small setting can be remarkable, revealing the texture of how power actually works rather than offering abstract generalisations.

The challenges that small-scale studies cannot escape 🔗

Yet the small-scale approach carries serious risks, and these go to the heart of why studying power empirically is so difficult.

The first is the problem of generalisation. New Haven is not the United States. Findings from one community, however careful, may not hold elsewhere. A city dominated by a single industry, a homogeneous population, or a particular political culture may produce power structures that look nothing like those of a different community. When researchers stretch local conclusions to cover an entire nation, they risk claiming far more than their evidence can support. In a country as varied as India, with its enormous differences across states, castes, languages, and levels of development, the danger of over-generalising from a single case is especially acute.

The second is methodological bias. The way researchers design a study often determines what they find. This was the core of the famous critique that followed Dahl’s work. By choosing to study only visible decisions where there was observable conflict, the pluralist method was almost guaranteed to find pluralism. It could not detect power that worked by keeping issues off the agenda entirely.

The faces of power that decisions cannot show 🔗

This critique deserves a closer look, because it transformed how scholars think about studying power. Peter Bachrach and Morton Baratz argued that the decision-making approach captured only one face of power. There was a second face: the power to control the agenda, to ensure that certain issues never reach the point of decision at all. They called this non-decision-making, the exercise of power to limit debate to issues that are safe for those already in control. They introduced the idea of the “mobilisation of bias,” the set of dominant values and procedures that systematically benefit some groups while excluding others.

If power can operate by suppression, then Dahl’s method had a hole in it. A researcher studying only the decisions that get made will never see the demands that were quietly strangled before they became demands. The poor family that never petitions the council because it knows the request is hopeless leaves no trace in any record of decisions.

Steven Lukes then added a third dimension. Power’s most insidious form, he argued, is the ability to prevent conflict from arising in the first place by shaping people’s perceptions, preferences, and sense of what is possible. If people accept their disadvantage as natural or inevitable, there is no observable conflict at all, yet power is clearly at work. This third face is the hardest of all to study empirically. How do you produce evidence of a grievance that the people affected do not even recognise as a grievance? Lukes effectively pushed the study of power toward questions that observable data alone may never resolve.

Studying power in the Indian context 🔗

These debates are not confined to American cities or British universities. They bear directly on how we study power closer to home. The system of Panchayati Raj offers a rich and accessible setting for empirical research, much as New Haven did for Dahl.

The 73rd Constitutional Amendment of 1992 created a three-tier structure of local self-government and reserved seats for Scheduled Castes, Scheduled Tribes, and women, in the hope of redistributing power away from entrenched rural elites. This makes panchayats a natural laboratory. A researcher can study a village council the way Dahl studied a city, tracing who shapes decisions about welfare schemes, housing allocations, or local works.

And the findings echo the old methodological warnings. A surface-level study counting reserved seats might conclude that power has been successfully redistributed to marginalised groups. But deeper field research repeatedly uncovers a different reality. Studies in states like Karnataka show that local development schemes remain prone to elite capture, where dominant castes and landed interests continue to control outcomes despite the formal change in who holds office. A formally elected woman or Dalit representative may hold the seat while real decisions are made elsewhere, a phenomenon that purely formal analysis would miss entirely.

This is the second face of power in action. Research on welfare programmes finds that households connected to local political executives are more likely to obtain benefits than unconnected households, revealing informal channels of influence that never appear in official records. A study that looked only at the formal rules of distribution would conclude the system was fair. Only fieldwork that follows the actual flow of resources exposes the hidden mechanism.

What good empirical work demands 🔗

None of this means that studying power empirically is futile. It means the researcher must be honest about the scope and limits of what they can claim. Several principles follow from the debates above.

Match the method to the dimension. Interviews and decision-tracing can reveal who prevails in open conflicts, but uncovering agenda control or shaped preferences requires different tools, including long-term observation, attention to what is absent, and willingness to study the issues that never made it to a vote.

Be cautious with generalisation. A single well-studied case is valuable, but its conclusions should travel only as far as the evidence allows. Scholars increasingly recommend triangulation, combining multiple methods and larger-scale data collection to reduce the distortions any single approach introduces.

Acknowledge bias in design. Because the choice of what to study so strongly shapes what is found, researchers must justify their selection of cases, decisions, and actors, and remain alert to the possibility that their method is quietly determining their conclusions.

The contextual nature of political power 🔗

Underlying all these challenges is a single stubborn fact. Power is contextual. The same resource that confers power in one setting confers none in another. Caste standing may dominate a village council yet count for little in a corporate boardroom. Control of information may be decisive in one decision and irrelevant in the next. Comparing power across different issue-areas or different systems has long frustrated researchers precisely because there is no common currency in which to measure it.

This is why the empirical study of power will always require judgement as much as data. The numbers and interviews matter, but so does the researcher’s interpretation of what they mean. The honest scholar treats their findings as evidence within a context, not as universal laws. The dishonest one dresses a local observation in the language of certainty and calls it a theory of how power works everywhere.

Dahl’s New Haven study endures not because its conclusions were correct in every detail, indeed they were fiercely contested, but because it took seriously the demand for evidence. It forced the field to argue about method, not just about theory. That argument, carried forward by Bachrach, Baratz, Lukes, and the researchers studying panchayats today, is the real legacy of trying to study power empirically. The difficulty is not a reason to give up. It is a reason to be careful.

What do you think? If the most powerful uses of power are the ones that leave no visible trace, can the empirical method ever fully capture how power works, or will some of it always remain beyond the reach of evidence? And when you look at a panchayat, a student union, or a workplace near you, how would you go about finding out who truly holds power there?

References

1. https://www.britannica.com/biography/Robert-A-Dahl
2. https://www.britannica.com/topic/Who-Governs-Democracy-and-Power-in-an-American-City
3. https://www.routledge.com/An-Analysis-of-Robert-A-Dahls-Who-Governs-Democracy-and-Power-in-an/Nilsson-Xidias/p/book/9781912128006
4. https://s-usih.org/2018/03/two-faces-of-power-two-faces-of-fun/
5. https://www.powercube.net/analyse-power/forms-of-power/academic-debate-on-forms/
6. https://www.epw.in/journal/2021/20/book-reviews/decentralisation-development-and-%E2%80%98elite-capture%E2%80%99.html
7. https://link.springer.com/article/10.1057/s41287-021-00447-4
8. https://journals.sagepub.com/doi/10.1177/1465116508095151
9. https://www.cambridge.org/core/journals/american-political-science-review/article/abs/measuring-the-concentration-of-power-in-political-systems/458C15AC2068603821ADB94F3E161A2F