Science Scientific Method and Critique

What Counts as “Science”

Sociology Optional Notes {PDF}

Complete Notes Package (Paper I & II) | 1000+ Pages | Instant Download

Discounted Price | Buy Now @ ₹600 →

• Theodorson and Theodorson define science as the building of a body of knowledge through observation, experimentation, generalization, and verification — it is both a body of certified, changing knowledge and the method used to acquire it.
  • This double meaning matters: when someone asks “is sociology a science?”, they could be asking either whether sociology has produced a reliable, cumulative body of knowledge, or whether it follows the method that natural science follows. The two questions get different answers.
• A discipline is usually called scientific when it shows five features together:
  • Empiricism — claims rest on observable, verifiable, quantifiable facts, not on authority, tradition, or intuition.
  • Objectivity and value-neutrality — the researcher’s personal preferences should not distort the finding.
  • Reliability, validity, predictability and generalization — a finding should hold up if the study is repeated, should measure what it claims to measure, and should let one anticipate future cases.
  • Self-correction — science stays open to revising itself in light of new evidence, and takes spatial and temporal variation seriously rather than freezing one context into a universal law.
  • Theoretical orientation and cumulativeness — findings build on each other into an organized body of theory, rather than remaining a pile of disconnected facts.
• Anthony Giddens puts the sociological case directly: sociology is a scientific endeavour to the extent that it involves systematic methods of empirical investigation, the analysis of data, and the assessment of theories in the light of evidence and logical argument — while still being, in his own description, a “soft” science next to the “hard” sciences of physics or chemistry, because its subject matter (self-aware human beings) resists the same degree of control and prediction.
• Immanuel Kant’s older distinction is worth keeping in mind as the philosophical background to this whole debate: he separated knowledge from philosophy (statements of value) from knowledge from science (statements of fact). Much of the “is sociology a science” argument is really an argument about whether social life can ever be reduced entirely to statements of fact.

The Four Postulates That Decide the Answer

• Whether a discipline is treated as science-like does not just depend on its methods; it depends on prior assumptions a researcher makes before any method is even chosen. Social theorists have identified four such sets of postulates:
  • Ontological — what is the nature of social reality? Is it an objective structure “out there” (realism), or is it constructed through the meanings people attach to it (nominalism/constructionism)?
  • Epistemological — how is knowledge of that reality obtained? Through causal explanation and hard data (empiricism), or through grasping subjective meaning (interpretivism)?
  • Human nature — are people shaped by external forces (determinism), or do they exercise genuine choice (voluntarism)?
  • Methodological — should inquiry aim at general laws applicable everywhere (nomothetic), or at the deep, particular understanding of one case (ideographic)?
• These four postulates line up into two competing packages:

• This is why the “is sociology a science” question never gets settled by pointing to a single method — the disagreement starts one level below method, in what each side already assumes about the nature of social reality itself.

Comte’s Optimism and Its Critics

• Auguste Comte placed sociology at the top of his hierarchy of the sciences — mathematics, astronomy, physics, chemistry, biology, and finally sociology — on the reasoning that each science depends on the ones below it and studies more complex phenomena. His claim was that with enough time and effort, sociology would mature into a fully positive science: empirical, value-neutral, testable, and capable of universal theories, exactly as physics or chemistry had already become.
• Against this, Karl Popper, and thinkers like Goode and Hatt, reframed the question itself: science, for them, is not a fixed body of subject matter that a discipline either reaches or doesn’t — it is a method of approaching and studying a phenomenon. On this reading sociology does not need to “wait its turn” in Comte’s hierarchy; it is already a science in the sense that matters, because it already uses a recognisable scientific method.
  • Goode and Hatt break this method into three components: empirical observation, experimentation to test a hypothesis, and verification — a process also called inductivism, because it builds toward a general law from repeated particular observations. Its risk is that a researcher who has already formed a hypothesis can end up looking only for confirming evidence and quietly setting aside what contradicts it.
  • The ultimate purpose of this process is to arrive at a law; the intermediate purpose is classification by common attribute, which is what makes future prediction possible at all.

The Steps of Scientific Research

• Horton and Hunt list the sequence a piece of scientific research is expected to follow:
  1. Identification of a research problem, chosen on scientific rather than personal grounds.
  2. Review of the existing literature, so the study builds on what is already known rather than repeating it.
  3. Construction of a hypothesis — a tentative generalization whose validity has not yet been tested, stating which aspect of reality will be examined.
  4. Design of the research plan.
  5. Collection of data using scientific tools and techniques.
  6. Checking the reliability, validity, and authenticity of that data.
  7. Generating theory — drawing a conclusion from the data.
  8. Replicating the study, to see if the same result holds up independently.
• Horton and Hunt also specify what makes the conduct of this research itself count as scientific: verifiable evidence, accuracy, precision, systematization, objectivity, careful recording, control of conditions, and trained investigators.
• A research design is the blueprint that operationalizes all of this. Black and Champion identify what a sound design must do: gather enough evidence to support a hypothesis while genuinely testing it against alternatives, allow the study to be repeated rather than describing a one-off situation with no wider relevance, work out correlations among variables in a way that produces interrelated propositions, signal the need for a pilot study before a full project is committed to, and economize on time and resources by choosing appropriate data-collection techniques.
  • Mannheim similarly identifies five purposes a research design must serve, converging on the same point from a different angle: a design exists to make a study defensible, repeatable, and efficient, not simply to give a researcher a starting point.
• Serendipity, a term brought into methodology by Robert Merton, names the situation where a researcher, while investigating one hypothesis, stumbles onto an unanticipated, anomalous, and strategic finding that becomes the seed of a new theory — Fleming noticing that a mould had killed his bacterial culture is the classic outside-sociology example. Merton’s point is that rigid, hypothesis-only research designs can miss the very discoveries that most advance a field, because a design built to test one thing is not built to notice something else.
• Explanatory and exploratory research designs serve different moments in the research process:
  • An exploratory design is used when a phenomenon is not yet well understood — its purpose is to generate hypotheses, clarify concepts, and identify variables worth studying further, typically through flexible, open-ended methods like case studies, focus group discussions, or unstructured interviews. A researcher entering a newly emerging subculture with no prior sociological literature on it would use this design first.
  • An explanatory design comes after enough is already known to state a specific hypothesis about a cause-effect relationship, and then tests that hypothesis using more structured methods — surveys, controlled comparison, statistical analysis. Durkheim’s study of suicide rates against religious affiliation is explanatory in this sense: he already had a hypothesis (social integration affects suicide) and designed a comparison to test it.
  • The two are not rival philosophies so much as successive stages: exploration typically comes first and narrows the field; explanation follows and tests a specific claim within it.

The Comparative and Historical Method: Sociology’s Substitute for the Laboratory

• Sociology cannot easily put a whole society into a laboratory, so it relies on the comparative method, which Durkheim treated as a form of indirect experiment. He distinguished three ways of applying it:
  • Within one society — comparing married and unmarried people in the same society.
  • Across different societies at the same point in time — comparing a Catholic country like Spain with a Protestant one like Germany.
  • Over a period of time within the same society — comparing the same society at different stages of development.
  • By holding some variables constant and letting others vary across these comparisons, a researcher can approximate the logic of a controlled experiment without ever running one directly.
• The historical method extends this same logic across time: by tracing how an institution, custom, or social pattern has actually changed, a researcher can identify which features are contingent and which are more enduring — something a single cross-sectional snapshot of one society at one moment cannot reveal on its own. It remains relevant precisely where laboratory experimentation and even the comparative method run into limits, such as the study of long-run institutional change (the evolution of the family form, or of land-tenure systems, for instance), where only a historical record supplies the necessary variation.

Critique of the Scientific Method Itself

• This critique is distinct from the critique of positivism as a philosophy of social reality (that argument is taken up in a separate article) — it concerns the scientific method as a procedure, regardless of who is applying it.
• Karl Popper reframes what science can even claim: it does not arrive at absolute truth, only at the truth that is currently subjected to observation and not yet falsified — Pluto was classified a planet and later was not, without either classification ever having been “science getting it wrong” in a naive sense. His proposed corrective to the inductive method is falsification: a theory earns scientific status not by accumulating confirming instances (which can always be found for almost any theory) but by surviving genuine attempts to disprove it. He also argues that because scientific methods are themselves built on existing scientific theories, they are just as vulnerable to error as any other method — there is no view from nowhere that a “scientific” method occupies.
• Thomas Kuhn, in The Structure of Scientific Revolutions, argues that a discipline’s development is not the smooth, cumulative process the scientific-method model implies — it moves in sudden jumps he calls paradigm shifts, through three discernible phases:
  1. Pre-scientific/pre-paradigmatic phase — competing, incomplete theories with no consensus.
  2. Normal science — a single paradigm (a dominant theory or family of theories) becomes established, and researchers work to increase its precision rather than question its foundations.
  3. Paradigm shift — anomalies accumulate that the existing paradigm cannot explain, the paradigm is abandoned, and a new one takes its place, offering a more complete account.
  • Kuhn also points out a specific flaw in the scientific method’s logic: because a hypothesis is constructed before the research begins, the result is, in a sense, already known in advance — which makes the process far more predictable and far less creative or innovative than it presents itself as being.
  • Contemporary sociologists have turned Kuhn’s own framework back onto the “is sociology a science” debate: since functionalism, conflict theory, symbolic interactionism, and phenomenology compete as paradigms much as rival theories do within natural science’s pre-paradigmatic phase, the presence of multiple paradigms need not disqualify sociology from being scientific — it may simply mean sociology has not yet, or may never, settle into a single dominant paradigm the way physics has.
• Paul Feyerabend goes further than Popper or Kuhn: he treats the scientific method as only one way of approaching truth among many, arguing in Against Method that binding inquiry to a single, fixed methodology restrains what can be discovered rather than protecting it. His position is sometimes summarized as epistemological anarchism — “anything goes” — meaning that the history of genuine scientific progress shows repeated departures from supposedly fixed methodological rules, not loyal adherence to them.
• J. S. Mill takes a more moderate version of the same worry, proposing methodological pluralism: rather than glorifying any one method as the method, different questions call for different methods, and a discipline should stay open to using several.
• A cluster of thinkers press a cultural and moral critique of scientism itself, distinct from Popper’s or Kuhn’s technical objections:
  • Jacques Barzun argues that an uncritical faith in science’s conclusions turns science into a kind of divinity — a new form of divination, believed in with the same fanaticism once reserved for religious dogma.
  • Theodor Adorno offers a periodization of science’s relationship to nature: in the 17th and 18th centuries science sought to study nature, in the 19th it exploited nature, the 20th revealed its destructive power, and by the 21st century science has been colonized by the state — a trajectory that should temper any impulse to simply celebrate science as an unqualified good. Elsewhere he argues that science is inherently presumptive: it assumes its own frame is the right one before it has even tested that assumption. He and E. F. Schumacher both press a related point — an exclusive focus on quantitative measurement leaves science structurally unable to recognize important qualitative aspects of experience.
  • Carl Jung argues that subjective phenomena — happiness, beauty, pleasure — cannot be measured by the methods of natural science at all, and concludes that methodology should be discipline-specific rather than imported wholesale from physics or chemistry.
  • Jürgen Habermas likewise treats science’s claim to be a neutral, interest-free mode of knowing with suspicion, arguing that it, too, serves particular human interests (technical control) even while presenting itself as disinterested.
• Beyond these philosophical objections, sociology faces distinctly practical limits in applying a natural-science method:
  • Experimentation carries ethical limits that natural science mostly avoids — a sociologist cannot randomly assign people to poverty or war to observe the effects. (This limit is not fatal to scientific status by itself: mature sciences like astronomy also cannot experiment on their subject matter directly, yet remain sciences.)
  • Quantification struggles with the large stretches of social life that are inescapably qualitative — meaning, emotion, motive.
  • Generalization is harder because human behaviour, unlike the behaviour of physical objects, does not reliably repeat itself in fixed patterns across contexts.
  • Cause and effect are easy to misread. Durkheim’s own data found a correlation between suicide rates and the season of the year — which could tempt the conclusion that temperature itself affects suicide, when in reality the more plausible mechanism is that people are more socially active in spring and summer than in winter, and that social integration (not weather) is doing the causal work.
  • Every research process risks distortion by the researcher’s own value bias, and by a conflicting-paradigms problem in which a researcher who already holds a paradigm tends to notice confirming evidence and overlook what contradicts it.
  • Underlying all of this is the deeper objection that social life cannot be “laboratorized” — placed under controlled, repeatable conditions — without losing exactly the qualities (meaning, context, historical specificity) that made it worth studying in the first place.

Is Sociology a Science? Where the Question Actually Stands

• One influential line — running through Comte, Spencer, Durkheim, Malinowski, Talcott Parsons, and the Chicago School — answers yes, on the grounds that sociology shares the natural sciences’ unity of method: it seeks laws that determine human behaviour, in the spirit of Comte’s original “positive philosophy,” and social sciences can legitimately borrow both the methodology and the underlying insight of the physical sciences as a model.
• A second line resists this, not always by denying that sociology can be systematic, but by denying that “science” should mean only what the natural-science model means. Max Weber’s own position is genuinely a bridge rather than a flat rejection: he argued that sociology is neither purely ideographic nor purely nomothetic, and that it must still strive to be value-free in its conclusions even while its subject matter (meaningful human action) requires an interpretive method the natural sciences do not need. It is a common error to read Weber as simply “anti-positivist” — his actual position tries to keep science’s discipline of evidence and value-freedom while adding a method suited to meaningful action, which is a subtler claim than either side of the debate is often given credit for.
• The most honest summary is probably Giddens’s own: sociology can be scientific in the procedural sense — systematic, evidence-based, self-correcting, logically argued — without being scientific in the sense of producing the same kind of universal, predictive laws that physics produces. Being a “soft” science is not a failed attempt at being a “hard” one; it reflects a genuinely different kind of subject matter, in which the units of study can interpret themselves, argue back, and change their behaviour once they learn what has been found about them — something no physical particle does.
• This last point is itself sometimes called the double hermeneutic (a concept developed most fully by Giddens): sociological knowledge feeds back into the society it studies and changes it, which means sociology’s object of study is a moving target in a way physics’s is not — a structural difference in kind, not merely a temporary shortfall in method.

A Live Illustration: the Replication Crisis

• The debate over whether the natural-science method transfers cleanly to social inquiry is not only a historical argument between Comte’s critics and defenders — it has a very current version. Since the early 2010s, psychology and other social sciences have faced a well-documented replication crisis: a significant share of published findings, when independently repeated, fail to reproduce the original result.
• Recent scholarship (2024-25) reads this as more than a technical problem of statistical power or publication bias — some researchers now argue it reflects a deeper epistemological rupture, questioning whether replicability itself, borrowed directly from the natural-science model, is even the right standard to hold social research to when the object being studied (a meaning-making human subject) is not independent of the researcher observing it in the way a chemical compound is.
• This gives Popper’s, Kuhn’s, and Durkheim’s older methodological worries a concrete, contemporary anchor: it is not merely that sociology struggles to replicate natural-science rigor, but that recent natural-science-adjacent disciplines are themselves discovering the same problem, which complicates any argument that treats the natural sciences as a stable benchmark sociology has simply failed to reach.

In Sum

• Sociology’s claim to be a science was never a settled fact waiting to be confirmed or denied by one clever test — it depends on prior ontological and epistemological commitments, on which definition of “science” (a body of laws, or a method of disciplined inquiry) is being used, and on how much weight is given to the practical and philosophical limits that separate a self-interpreting human subject from a chemical compound. The most defensible position treats sociology as scientific in its discipline and procedure while remaining distinct in the kind of knowledge it can ultimately produce.