Experimental Psychology

Abstract

Experimental psychology is a branch of psychology that studies mind and behavior through controlled experiment, establishing cause by manipulation and measurement rather than by introspection or anecdote. It was founded when nineteenth-century physiologists showed that mental events could be measured: Weber and Fechner turned sensation into a quantitative law, Wundt opened the first dedicated laboratory in Leipzig in 1879, and Donders timed thought itself by subtraction. From that beginning the field built a method rather than a subject matter, and its history is largely the refinement of that method, through Fisher's logic of randomized experiment and the null-hypothesis test, to the modern reckoning with replication that has made the reliability of the experiment itself an object of study. Its enduring problem is the one it started with: how to draw a valid inference about an unobservable mind from observable behavior.

Keywords: psychophysics, mental chronometry, experimental method, null-hypothesis testing, replication

Experimental psychology is less a topic than a way of asking questions. Where earlier philosophy of mind reasoned from armchair reflection, experimental psychology insists that a claim about perception, memory, or thought earn its place by surviving a controlled test, in which the investigator varies one thing, holds others constant, measures an outcome, and reasons from the result to its cause. The commitment is methodological, which is why the same experimental tradition underwrites the study of sensation, learning, attention, and decision alike. Its founders were physiologists and physicists who found, against the expectation of their age, that the mind was not beyond measurement, and the discipline has spent a century and a half making that measurement more rigorous and more honest.

What Experimental Psychology Is

Experimental psychology is the study of mental processes and behavior by the experimental method: the investigator manipulates an independent variable, measures a dependent variable, and controls the rest, so that a difference in outcome can be attributed to the manipulation rather than to chance or confound. Its defining commitment is not to a set of topics but to a mode of inference. Sensation, perception, attention, memory, learning, emotion, and thought are all studied experimentally, and what unifies the work is the demand that every claim be grounded in a measurement made under controlled conditions rather than in report or intuition (Boring, 1950).

That demand was radical when the field began, because the prevailing view held the mind to be private and immeasurable. The founders' achievement was to show otherwise. By the end of the nineteenth century the reaction time, the just-noticeable difference, and the forgetting curve were all quantities an experiment could yield, and the mind had become, in principle, an object of measurement like any other in natural science (Ebbinghaus, 1913). The discipline's later history is the working-out of the consequences: better designs, sharper statistics, and, most recently, a hard look at whether its published measurements can be trusted.

The Founding of the Laboratory

Experimental psychology began not in a philosopher's study but in a physiologist's, with the discovery that sensation obeys a law. Ernst Heinrich Weber, studying the discrimination of weights and lengths, found that the smallest detectable change in a stimulus, the just-noticeable difference, is not a fixed amount but a constant fraction of the stimulus already present: to tell 100 grams from a heavier weight one needs an increment of about two grams, but to tell 200 grams apart one needs four (Weber, 1996). Gustav Theodor Fechner, seeking to relate the physical and mental worlds, integrated Weber's fraction into a law of his own, holding that sensation grows as the logarithm of stimulus intensity, so that equal ratios of intensity produce equal steps of sensation (Fechner, 1860). Fechner named the enterprise psychophysics and, in doing so, gave the new science its first quantitative program. The interactive below finds a just-noticeable difference and shows Weber's fraction holding across intensities.

Weber's law holds the fraction k constant, so the just-noticeable difference grows in proportion to the intensity already present. Because each step is a fixed fraction of the current intensity, equal ratios of intensity span equal numbers of steps, which is Fechner's logarithmic law of sensation.

Wilhelm Wundt turned this program into an institution. In 1874 he published the Grundzüge der physiologischen Psychologie, arguing that psychology could be an experimental science of immediate experience on the model of physiology (Wundt, 1874), and in 1879 he founded at Leipzig the first laboratory devoted to psychological experiment, training a generation of investigators who carried the method across Europe and to America. One of them, Edward Bradford Titchener, brought the experimental method to Cornell and codified it in a laboratory manual that drilled students in introspection under controlled conditions, the systematic report of the elements of conscious experience (Titchener, 1901). Titchener's structuralism, the attempt to analyze consciousness into its sensory elements, did not survive as a theory, but his insistence on standardized procedure helped fix the laboratory as the field's home. Hermann Ebbinghaus, working outside any laboratory tradition, showed how far the method could reach: using himself as subject and nonsense syllables as material, he plotted the rate of forgetting as a curve, proving that even memory, seemingly the most private of faculties, yielded to experiment (Ebbinghaus, 1913).

Mental Chronometry

If sensation could be measured, could thought be timed? Franciscus Cornelis Donders answered yes, inventing the method that still underlies much of experimental cognitive research. Donders reasoned that a mental operation takes time, and that the time of an unobservable operation can be recovered by subtraction: measure the reaction time for a simple task, then for a task identical except that it requires one additional mental step, and the difference estimates the duration of that step (Donders, 1969). A simple reaction, pressing a key the instant any signal appears, measures detection and response alone. A choice reaction, pressing one key for one signal and another for a second, adds stimulus identification and response selection. Subtract the first from the second and what remains is an estimate of the time the mind spends choosing.

Donders's subtraction method rested on the assumption that mental stages are discrete and additive, that inserting a stage leaves the others unchanged, and later work has shown that assumption to be an idealization rather than a law. Yet the core idea, that differences in reaction time across carefully matched tasks can isolate the duration of a hidden mental process, remains one of the most productive experimental logics in psychology, and the chronometric study of the mind runs directly from Donders to the modern measurement of reaction time (Boring, 1950). The interactive below runs the three classic task types and recovers a stage duration by subtraction.

Donders assumed mental stages are discrete and additive, so inserting one stage leaves the others unchanged and the reaction-time difference measures the inserted stage alone. Raising the number of choice alternatives lengthens response selection, an early hint of Hick's law relating choice time to the logarithm of the number of options.

Measuring the Mind: Psychophysical Scaling

Fechner's logarithmic law dominated psychophysics for most of a century, but it rested on an indirect method: counting up just-noticeable differences and assuming each felt equally large. Stanley Smith Stevens challenged both the method and the law. Asking observers to assign numbers directly to the apparent magnitude of a stimulus, a procedure he called magnitude estimation, he found that sensation grows not as the logarithm of intensity but as a power of it, so that the ratio of two sensations equals the ratio of the intensities raised to an exponent characteristic of the modality (Stevens, 1957). The exponent varies widely: for brightness it is well below one, so that sensation compresses as light intensifies; for line length it is near one, a nearly faithful scale; and for electric shock it is far above one, so that sensation expands steeply, a difference with obvious survival logic. Modern reviews have argued that the venerable Weber-Fechner law conflates Weber's well-supported discrimination law with Fechner's shakier scaling assumption, and should be retired as a single named law (Algom, 2021).

The dispute matters beyond psychophysics because it is a template for the field's method. Two laws fit the same rough data; deciding between them required a new experimental procedure, magnitude estimation, that put the rival predictions to a direct test rather than settling the question by assumption. The scaling of sensation also seeded a broader theory of how observers turn graded evidence into discrete judgments, developed in signal detection theory, which separated an observer's sensitivity from the criterion they set for responding and so cleaned up a confound that had muddied psychophysical measurement for decades.

The Logic of the Experiment

A measurement is only as good as the design that produced it, and the twentieth century's great methodological contribution came from outside psychology. Ronald Aylmer Fisher, working in agricultural research, formalized the randomized experiment: assign subjects to conditions by chance, and the many uncontrolled differences between them are converted from systematic biases into quantifiable random error, so that a statistical test can gauge whether an observed difference exceeds what chance alone would produce (Fisher, 1935). Randomization, replication, and the null-hypothesis significance test became the standard grammar of the psychological experiment, and remain so. Fisher's logic gave the field a disciplined way to answer its recurring question, whether a manipulation made a real difference, by asking how surprising the result would be if it had not.

That grammar also exposed a fault line within the discipline. Lee Cronbach, in a famous address, observed that psychology had split into two research cultures that barely spoke: an experimental tradition that manipulated variables and averaged over individuals, and a correlational tradition that measured the differences between individuals and never manipulated anything (Cronbach, 1957). The experimental wing treated individual variation as noise to be controlled away; the correlational wing treated it as the very signal of interest. Cronbach argued that a mature science of psychology would have to unite the two, studying how treatments interact with the traits of the people who receive them, and his diagnosis still frames debates about what a psychological experiment can and cannot tell us about any particular person.

Worked Example

Weber's law gives experimental psychology one of its cleanest quantitative predictions, and working through it shows how a single measured constant generalizes across an entire range of intensities. Suppose a subject in a weight-discrimination task reliably detects the difference between a 100-gram standard and a comparison only once the comparison reaches 102 grams. The just-noticeable difference (JND) is then 2 grams at a baseline of 100 grams, and Weber's fraction, the ratio of the JND to the standard, is

k = ΔI / I = 2 / 100 = 0.02.

Weber's law asserts that this fraction, not the absolute increment, is constant. So the same subject's JND at any other intensity is predicted by ΔI = k × I. At a 500-gram standard the predicted JND is 0.02 × 500 = 10 grams; at 1,000 grams it is 0.02 × 1,000 = 20 grams. The absolute increment needed grows with the load, but the relative increment stays fixed at two percent, which is the content of the law and the reason a pianist's touch or a grocer's hand keeps constant relative precision across very different weights.

Fechner built his scale of sensation on exactly this fact. If each JND is one just-perceptible step of sensation, then counting the steps between two intensities measures the sensation difference between them. Because each step is a constant fraction of the current intensity, the number of steps from a baseline I₀ to an intensity I is n = ln(I / I₀) / ln(1 + k). For our subject, the number of just-noticeable steps between 100 grams and 1,000 grams is

n = ln(1,000 / 100) / ln(1.02) = ln(10) / ln(1.02) = 2.3026 / 0.0198 ≈ 116 steps.

A tenfold increase in physical weight corresponds to only about 116 equal steps of felt heaviness, and this compression, sensation growing as the logarithm of intensity, is Fechner's law derived from Weber's. The table below shows the predicted JND climbing in proportion to intensity while the Weber fraction holds constant, the signature of the law.

Table 1. Weber's law applied to weight discrimination from a single measured Weber fraction of 0.02. The just-noticeable difference scales with intensity while the fraction stays constant. Computed locally.

Discussion

Experimental psychology's coherence comes from its method, and its recurring difficulties come from the gap the method must always bridge, between an observable behavior and the unobservable mental process inferred from it. Every technique the field prizes is a way of narrowing that gap: Weber's fraction turns a private sensation into a measurable threshold, Donders's subtraction turns a hidden mental stage into a difference of milliseconds, Fisher's randomization turns uncontrolled variation into calculable error. None closes the gap entirely. A reaction-time difference is consistent with more than one arrangement of mental stages; a significant p-value is consistent with more than one underlying truth. The field advances by building designs whose inferences are progressively harder to explain away.

That is also why experimental psychology connects outward to every content area of the discipline rather than sitting beside them. The experimental study of learning, of attention, and of perception are not neighbors of experimental psychology but instances of it, applications of the same controlled method to different faculties. Cronbach's warning still marks the boundary of what the method delivers: an experiment that averages over people establishes what a manipulation does in general, and needs a correlational complement to say what it does for anyone in particular. The most consequential recent development, though, has turned the method's lens on itself.

Current Directions

The defining movement in experimental psychology this century has been a crisis over replication and the reforms it produced. In 2011 a demonstration showed that ordinary, undeclared flexibility in how data are collected and analyzed, choosing when to stop testing, which conditions to compare, which covariates to include, can push the rate of false-positive findings from a nominal five percent to over sixty, without any single step feeling like misconduct (Simmons et al., 2011). The scale of the problem became concrete when the Open Science Collaboration repeated one hundred published psychology studies and found that fewer than half yielded significant effects the second time, and that the replicated effects were on average about half the original size (Open Science Collaboration, 2015). A parallel project replicating social-science experiments from Nature and Science reached a similar verdict, reproducing about two-thirds of the effects at markedly reduced magnitude (Camerer et al., 2018).

The response has been methodological rather than merely cautionary, and it targets exactly the flexibility Simmons identified. Preregistration asks researchers to specify their hypotheses and analysis plan before seeing the data, converting the analytic choices that inflate false positives from hidden researcher degrees of freedom into a public commitment (Nosek et al., 2018). Registered reports go further, submitting the introduction and method for peer review before data collection, so that a study is accepted on the quality of its design rather than the appeal of its results, which removes the incentive to chase significance (Chambers & Tzavella, 2022). A broad manifesto gathered these correctives, transparency, preregistration, replication, and reformed incentives, into a program for cumulative, reproducible science (Munafò et al., 2017), and a decade on, reviews now treat replicability, robustness, and reproducibility as distinct, separately measurable properties of a literature rather than a single virtue (Nosek et al., 2022). The interactive below switches these researcher degrees of freedom on and off and tracks the false-positive rate as it climbs and falls.

Each toggle is one researcher degree of freedom, and the solo figures are the false-positive rates Simmons and colleagues reported for each in isolation. The gauge compounds the active choices as roughly independent analytic paths; the real study found that combining them all pushed the rate above 60%. Preregistration removes the flexibility rather than the analysis.

Common Misconceptions

Experimental psychology is a topic, like memory or perception.

It is a method, not a subject. Memory, perception, attention, and learning are all studied experimentally; what makes the work experimental psychology is the controlled manipulation-and-measurement design it shares, not the faculty it examines (Boring, 1950).

A statistically significant result means the effect is real and important.

Significance gauges only how surprising the data would be if there were no effect, and undisclosed analytic flexibility can manufacture significance from noise. A small p-value is neither a measure of effect size nor a guarantee of replication (Simmons et al., 2011).

Sensation is a direct copy of the physical stimulus.

It is a lawful but nonlinear transform of it. Weber's law makes discrimination relative to the baseline, and Stevens's power law compresses or expands sensation by a modality-specific exponent, so equal physical steps are almost never equal perceptual steps (Stevens, 1957).

Glossary

Dependent variable.

The outcome an experiment measures, expected to change as a function of the manipulation; in psychology often a response time, accuracy, or judgment.

Experimental method.

The controlled manipulation of an independent variable while holding others constant, so that a change in the dependent variable can be attributed to the manipulation.

Independent variable.

The factor an experimenter deliberately varies across conditions in order to observe its effect on the dependent variable.

Introspection.

The systematic, trained observation and report of one's own conscious experience; the core method of Wundt's and Titchener's laboratories, later largely displaced by behavioral measurement.

Just-noticeable difference.

The smallest change in a stimulus that an observer can reliably detect; the basic unit of psychophysical measurement, denoted the JND.

Mental chronometry.

The use of reaction-time measurement to infer the timing and organization of unobservable mental processes, originating in Donders's subtraction method.

Null-hypothesis significance test.

A statistical procedure that gauges how improbable the observed data would be if the manipulation had no effect, formalized for experimental science by Fisher.

Preregistration.

The public specification of a study's hypotheses and analysis plan before data are seen, which converts hidden analytic choices into a fixed commitment and curbs false positives.

Psychophysics.

The quantitative study of the relation between physical stimuli and the sensations they produce, founded by Fechner and the origin of experimental psychology.

Registered report.

A publication format in which a study's rationale and method are peer-reviewed and accepted before data collection, so acceptance rests on design rather than on the results obtained.

Researcher degrees of freedom.

The undisclosed analytic choices, such as optional stopping or selective reporting, that can inflate the false-positive rate of an experiment far above its nominal level.

Structuralism.

Titchener's program of analyzing conscious experience into its elementary sensations, images, and feelings through trained introspection; historically important but abandoned as a theory.

Subtraction method.

Donders's technique of estimating the duration of a mental stage by subtracting the reaction time of a task lacking that stage from one that includes it.

Weber's law.

The principle that the just-noticeable difference is a constant fraction of the stimulus intensity, so that discrimination is relative rather than absolute.

Key Researchers

Edwin G. Boring (1886-1968). Harvard psychologist whose A History of Experimental Psychology became the field's standard historiography and shaped how generations understood its origins. Wikipedia - Wikidata

Christopher D. Chambers (living). Cognitive neuroscientist at Cardiff University and a principal architect of the registered-reports format, which reforms publication to reward rigorous design over striking results. Wikidata - Faculty Page - ORCID

Hermann Ebbinghaus (1850-1909). German psychologist who, experimenting on his own memory with nonsense syllables, produced the forgetting curve and proved that higher mental processes could be studied experimentally. Wikipedia - Wikidata

Gustav Theodor Fechner (1801-1887). Leipzig physicist and philosopher who founded psychophysics, formulating a logarithmic law of sensation and giving experimental psychology its first quantitative program. Wikipedia - Wikidata

Brian A. Nosek (living). Psychologist at the University of Virginia and co-founder of the Center for Open Science, who led the large-scale reproducibility project and championed preregistration. Wikidata - Faculty Page - ORCID

Stanley Smith Stevens (1906-1973). Harvard psychophysicist who introduced magnitude estimation and the power law of sensation, and whose taxonomy of measurement scales reshaped how psychologists quantify data. Wikipedia - Wikidata

Edward Bradford Titchener (1867-1927). Cornell psychologist who brought Wundt's experimental method to America and codified structuralist introspection in an influential laboratory manual. Wikipedia - Wikidata

Simine Vazire (living). Personality and meta-scientist at the University of Melbourne whose work on research credibility and self-correction helped define the reproducibility reform agenda. Wikidata - ORCID

Eric-Jan Wagenmakers (living). Mathematical psychologist at the University of Amsterdam known for Bayesian alternatives to significance testing and for advancing preregistered, adversarial methods. Faculty Page - ORCID

Ernst Heinrich Weber (1795-1878). Leipzig physiologist whose discovery that the just-noticeable difference is a constant fraction of stimulus intensity gave psychophysics its founding law. Wikipedia - Wikidata

Wilhelm Wundt (1832-1920). Founder of the first psychological laboratory at Leipzig in 1879 and author of the programmatic text that established psychology as an independent experimental science. Wikipedia - Wikidata

Frequently Asked Questions

What is experimental psychology?
It is the branch of psychology that studies mind and behavior through controlled experiment, manipulating a variable and measuring its effect so that outcomes can be attributed to cause rather than chance or confound (Boring, 1950).

Who founded experimental psychology?
Its founding is usually credited to Wilhelm Wundt, who opened the first dedicated psychological laboratory at Leipzig in 1879, building on the psychophysics of Weber and Fechner that had already made sensation measurable (Wundt, 1874).

What is Weber's law?
Weber's law states that the smallest detectable change in a stimulus is a constant fraction of the stimulus intensity, so that discriminating heavier weights or brighter lights requires proportionally larger increments (Weber, 1996).

What is the difference between Fechner's and Stevens's laws of sensation?
Fechner held that sensation grows as the logarithm of intensity, a single curve for all senses, while Stevens found it grows as a power of intensity with an exponent that differs by modality, compressing for brightness and expanding for shock (Stevens, 1957).

What is mental chronometry?
It is the measurement of reaction times to infer the duration and arrangement of hidden mental processes, originating in Donders's subtraction method, which estimates a mental stage by comparing tasks that do and do not require it (Donders, 1969).

Why does experimental psychology rely on statistics?
Because randomization converts uncontrolled differences between subjects into quantifiable random error, and a significance test then gauges whether an observed effect exceeds what chance would produce, a logic formalized by Fisher (Fisher, 1935).

What is the replication crisis?
It is the finding that a large share of published psychological effects do not reappear when studies are repeated, driven partly by undisclosed analytic flexibility, which prompted reforms in how experiments are planned and reported (Open Science Collaboration, 2015).

How does preregistration improve experiments?
By committing hypotheses and analysis plans publicly before the data are seen, preregistration removes the hidden researcher degrees of freedom that can inflate false-positive rates, and registered reports extend this to peer review of the design itself (Nosek et al., 2018).

References