Psychophysics: Thresholds, Weber's Law, and Measuring Sensation

Author: Angela Bogdanova · Published: October 16, 2026 · Editorial Policy

Research cluster: Attention, Perception & Cognitive Control · Written by Artificial Sapiens

Psychophysics examines the quantitative relationship between physical stimulation and perceptual responses. It asks how much light, sound, pressure, vibration, odor, or other stimulation must be presented before an observer can detect it, how much two stimuli must differ before an observer can discriminate them, and how reported sensory magnitude changes as physical intensity increases. By controlling a stimulus and measuring behavior across repeated trials, psychophysics converts questions about subjective sensation into explicit experimental estimates. The resulting numbers describe performance under specified conditions, with uncertainty, rather than an immutable boundary between sensing and not sensing.

A perceptual threshold is ordinarily an operational point on a response curve. Weber's law describes a frequently observed relationship between discrimination thresholds and reference intensity. Fechner's and Stevens' laws propose different mathematical accounts of perceived magnitude. Contemporary research treats these accounts as experimentally testable models, measures full psychometric functions, and distinguishes sensory sensitivity from the observer's decision strategy Johnson et al., 2002; Wichmann & Hill, 2001a.

What Is Psychophysics?

Psychophysics is the experimental study of lawful relationships between controlled stimulus properties and perceptual judgments or performance. A stimulus may be characterized by luminance, contrast, sound pressure, frequency, duration, force, temperature, or chemical concentration. The dependent response may be detection, correct identification, comparative judgment, reaction, or a numerical estimate of perceived strength. Investigators vary the physical dimension systematically and estimate how behavior changes. The key move is operational: specify what was presented, what judgment was required, and what criterion defines the reported estimate.

The word sensation can refer to sensory experience, physiological registration, or an observer's report. These are related but distinguishable levels of analysis. A participant pressing a button after a faint sound demonstrates a particular task response; that response alone does not reveal the entire quality of the sound experience or directly measure every sensory neuron. A threshold estimated from behavior is a property of the person-task-stimulus combination. Psychophysics gives that property a repeatable measurement framework Pelli, 1985.

How psychophysics differs from perception research and psychometrics

Perception research broadly studies how sensory information is organized into objects, scenes, events, and meanings. Psychophysics provides a quantitative toolkit for investigating these processes. Psychometrics also analyzes measurement quality, including reliability and validity, but its general framework covers psychological tests and latent attributes far beyond controlled sensory experiments. In the Angela Bogdanova encyclopedia, EN-25 owns the scientific explanation of psychophysical laws and sensory thresholds; EN-64 owns broad measurement theory.

A person may distinguish two patches of light in a carefully controlled experiment while interpreting the patches quite differently in an everyday scene. Discrimination performance and perceptual interpretation address overlapping but nonidentical questions. See the published overview Perception: How the Mind Interprets Sensory Information and the introductory comparison Sensation vs Perception: Differences and Everyday Examples.

How Psychophysics Became an Experimental Science

Ernst Heinrich Weber investigated tactile localization and sensory discrimination, including judgments of lifted weights. His research established that the physical increment required to notice a change frequently depends on the original stimulus magnitude. Gustav Theodor Fechner developed this line of work into an explicit quantitative program and published Elements of Psychophysics in 1860. Fechner's ambition was to describe the connection between physical stimulation and psychological experience through reproducible procedures and mathematical relationships. His legacy includes the classical methods of limits, constant stimuli, and adjustment.

During the twentieth century, S. S. Stevens challenged the assumption that equal just-noticeable differences correspond to equal units of subjective magnitude. By asking participants to estimate perceived intensity directly, he demonstrated systematic power-function relationships over many stimulus dimensions Stevens, 1957. Later research clarified the effects of judgment instructions, numerical scales, stimulus range, response bias, and neural encoding. Modern psychophysics is therefore an experimental science of stimulus-response relations with multiple complementary measurement methods, not a single historical law Petzschner et al., 2015.

Absolute Thresholds: Measuring Detection

An absolute threshold is a conventionally defined stimulus level at which detection reaches a specified probability under a particular experimental task. Imagine a brief tone embedded in background noise. Present that tone at several sound levels, including trials with no tone, and record the observer's responses. Near the lower end of the intensity range, detection may be unreliable; as intensity increases, detection becomes more frequent. The threshold could be defined as the intensity associated with an agreed performance level. Its numerical value depends on the task and decision rule.

Historical explanations sometimes suggest that a sensory system contains a single switch that flips from no sensation to sensation. Actual experimental responses commonly follow a gradual transition. Environmental noise, neural variability, attention, fatigue, expectation, adaptation, and occasional lapses contribute to fluctuating detection. A laboratory threshold therefore represents a location on a probability curve estimated from repeated observations Wichmann & Hill, 2001a.

Why the 50% convention requires context

In a yes/no task with signal-present trials, investigators may define threshold as the intensity producing a 50% signal-detection rate, but that percentage alone does not control for a participant's willingness to say “yes.” A liberal observer may report seeing signals more frequently, including on blank trials. A conservative observer may miss weak real signals while making fewer false alarms. False-alarm rates and the full response structure are essential to interpretation.

In a two-alternative forced-choice task, chance performance is 50% when alternatives are equally likely. A commonly used halfway-to-perfect criterion is therefore 75% correct, not 50%. A different number of alternatives produces a different guessing baseline. Threshold reports must specify the number of alternatives, whether feedback was provided, and the exact performance criterion; otherwise two apparently different thresholds may reflect different definitions rather than different sensitivity.

Detection threshold versus physiological limit

A psychophysical threshold is not an isolated measurement of receptor sensitivity. Optical properties of the eye, skin mechanics, acoustic transmission, background stimulation, movement, central processing, decision policy, and response method all influence what is measured. A biological mechanism can be investigated alongside a behavioral threshold, but the two estimates should be related through evidence rather than treated as identical quantities Pelli, 1985.

Difference Thresholds and the Just-Noticeable Difference

A difference threshold, also called a difference limen or just-noticeable difference (JND), describes a stimulus increment or separation between stimuli associated with a specified discrimination performance level. In a basic experiment, an observer compares a standard stimulus with a test stimulus. The physical difference increases over trials or varies in randomized order. The researcher identifies the difference at which comparisons become sufficiently reliable.

Consider lifting a reference weight of 100 grams and comparing it with slightly heavier weights. If a participant can distinguish a 105-gram comparison only inconsistently but performs reliably with a 110-gram comparison, the study can estimate a threshold within that range. The number is meaningful only when the reference, grip, comparison procedure, trial structure, and target accuracy are specified. Similar logic applies to brightness, visual contrast, loudness, texture, vibration, and other continua Holden et al., 2011.

The JND is often introduced as the smallest change a person can notice. In practice, it is a statistical estimate defined by an experimental criterion. In a two-choice comparison, the physical difference corresponding to 75% correct may be called the threshold. In an identification or adjustment task, a different formula and criterion may be appropriate. Researchers should report the operational definition rather than assuming that the phrase “just noticeable” supplies one universal measurement procedure.

Point of subjective equality and uncertainty

In comparisons where observers judge which of two stimuli is larger, the point of subjective equality is the comparison value at which the observer is equally likely to choose either stimulus. It describes systematic perceptual or decisional offset and is distinguishable from the steepness of the discrimination curve. The curve's slope concerns how rapidly judgment accuracy changes as physical differences increase. Separating bias from precision helps explain why two observers with similar discrimination ability may have different average judgments.

Weber's Law: A Relationship Between Change and Baseline

Weber's law states that, for some stimuli and over some intensity ranges, the just-noticeable increment grows approximately in proportion to the reference magnitude. Its familiar expression is ΔI / I = k. Here I is the standard physical intensity, ΔI is the increment meeting the chosen discrimination criterion, and k is the Weber fraction. A smaller k indicates finer relative discrimination under the measured conditions; a larger k indicates that a larger proportional change is needed.

Suppose an illustrative experiment finds k = 0.05 for a particular task over a limited range. At a reference value of 100 arbitrary intensity units, the predicted difference threshold is 5 units; at 200 units, it is 10; at 400 units, it is 20. The proportional increment remains 5% in this simplified example. These values illustrate the formula rather than claiming that the human visual, auditory, or tactile system has a universal 5% Weber fraction.

Weber's law is useful because a fixed physical increment does not always produce an equally discriminable change. The difference between an object weighing 100 and 110 grams may be more evident than an equal 10-gram addition to a much heavier object. The baseline matters. Controlled tactile studies demonstrate how Weber fractions can be estimated across different stimulus magnitudes Holden et al., 2011.

What the Weber fraction measures

A Weber fraction is a ratio derived from a measured difference threshold and its reference intensity. Its interpretation depends on the selected physical scale and the way the threshold is identified. Some continua have nonlinear physical units or special reference conventions; researchers must describe their stimulus definitions clearly. A Weber fraction for one sensory modality, feature, population, adaptation state, or task is not automatically transferable to another.

The quantity also should not be mistaken for a stable personality trait or a clinical diagnosis. Repeated low discrimination accuracy in one task may arise from sensory noise, imperfect calibration, misunderstanding, attention changes, fatigue, or a genuine difference in sensory performance. Claims about a person require suitable repeated testing and appropriately validated interpretation.

Where Weber's law becomes inaccurate

At very low signal levels, internal noise and absolute detection limits may dominate. At high levels, saturation, discomfort, or a narrowing usable range may distort the relationship. Background context, stimulus duration, adaptation, and stimulus dimension can also alter discrimination. The Weber fraction can vary systematically with intensity rather than remaining constant. A laboratory report should therefore show the tested intensity range and actual estimates, not impose a linear proportionality assumption on every observation Johnson et al., 2002.

Fechner's Law and the Weber–Fechner Relationship

Fechner proposed that increments corresponding to successive discrimination steps could be used to construct a scale of subjective intensity. Under the additional assumptions that the Weber fraction remains constant and that each JND represents an equal subjective increment, integrating the reciprocal stimulus relation produces a logarithmic model: S = a log(I / I₀). In this expression S denotes a modeled sensory magnitude, I the stimulus intensity, I₀ a reference intensity, and a a scale parameter. The ratio I / I₀ keeps the logarithm dimensionless.

The mathematical result has a clear interpretation within the model. Multiplying physical intensity by the same factor contributes an approximately equal additive increase in modeled sensation. As absolute physical intensity increases, proportionally larger physical changes are required to produce another equal modeled step. The often-quoted phrase “sensation grows logarithmically” summarizes this theoretical construction, whose empirical usefulness varies with the experimental domain.

Fechner's derivation depends on connecting discrimination steps with subjective magnitude. Contemporary research treats that connection as a hypothesis rather than a universally demonstrated equivalence. Equivalent JNDs need not correspond to equal felt differences, and direct subjective judgments sometimes support other functions Johnson et al., 2002. The historical Weber–Fechner relationship is foundational because it established a rigorous quantitative problem, even when its particular scaling assumptions require revision.

Why sound decibels are not a direct measure of experience

Sound-pressure level in decibels expresses a logarithmic ratio of a physical acoustic quantity to a reference. Loudness is a perceptual judgment. The two are related in appropriate conditions, but the decibel scale is not a universal direct reading of how loud a sound feels. Spectral composition, duration, context, hearing characteristics, and measurement method affect loudness. Physical logarithmic scales and psychophysical logarithmic models must be described as distinct constructs.

Stevens' Power Law and Perceived Magnitude

Stevens proposed that perceived intensity often follows a power function: S = a Iⁿ, where the exponent n varies with the sensory continuum and measurement conditions. If n is below 1, subjective magnitude increases more slowly than physical intensity in the model; if n exceeds 1, modeled subjective magnitude grows more rapidly. The power-law framework arose in part from magnitude-estimation experiments in which observers assigned numbers to perceived stimulus intensities Stevens, 1957.

In a typical magnitude-estimation task, a reference stimulus may be assigned an arbitrary number such as 10. Participants then assign comparison stimuli numbers representing how strong they seem relative to the reference. An experience judged twice as intense might receive 20. When physical stimulus levels and average numerical estimates are plotted on logarithmic axes, a power function corresponds to a straight-line relationship over the fitted range. This useful diagnostic does not prove the existence of a universal sensory exponent independent of the task.

Comparing Fechner and Stevens

The Fechner model produces an additive logarithmic scale from particular assumptions about discrimination. The Stevens model produces a power-law relation from direct scaling procedures. They differ in their experimental basis and mathematical behavior, but both can approximate data over limited ranges, and some modern accounts connect their phenomenology through decision processes or prior expectations. Reviewing the models requires examining the measured responses, stimulus ranges, fitting choices, and proposed neural mechanisms Johnson et al., 2002; Petzschner et al., 2015.

The methodological problem with numerical ratios

Magnitude estimation requires people to translate subjective differences into numerical ratios. Numbers below a reference have a bounded positive range while numbers above the reference may be unbounded. Respondents may interpret ratio instructions differently or gravitate toward familiar response scales. Experiments directly comparing response formats found that the estimated functions could vary with the procedure, highlighting the importance of response-method validity Mertens et al., 2021.

A power-law fit consequently describes results under particular scaling instructions. Evidence about the felt intensity of stimulation becomes stronger when investigators converge across multiple paradigms, cross-modal comparisons, carefully selected stimulus ranges, and independent replication. Numerical reports are data produced by observers under instructions, not a transparent and assumption-free window into experience.

The Psychometric Function: Measuring Probability Rather Than a Hard Cutoff

A psychometric function plots response probability or task accuracy against stimulus strength or stimulus difference. For example, contrast may appear on the horizontal axis and the proportion of correct detections on the vertical axis. The fitted curve generally rises from a lower asymptote toward an upper asymptote. Researchers describe its location, slope, lower chance level, and possible lapse rate. They can then estimate the stimulus level at a predefined performance criterion and quantify uncertainty Wichmann & Hill, 2001a.

A steep curve indicates relatively rapid changes in performance across a narrow stimulus range. A shallow curve indicates a more gradual transition. Both slope and location can change with stimulus uncertainty, comparison format, observer practice, and experimental context. Two studies can report different thresholds because one fits a logistic curve, another fits a cumulative Gaussian, or they choose different performance points. Model choice must be explicit and, when appropriate, compared against alternative fits.

Guessing rates and lapse rates

In a two-alternative forced-choice paradigm, the mathematical chance level is commonly 0.5. In a four-alternative task, it is 0.25. A lapse rate represents occasional incorrect responses even when the stimulus is readily discriminable. Ignoring lapses or guessing can systematically distort fitted thresholds and slopes. Simulation and methodological studies establish that estimating the shape of the psychometric function requires appropriate handling of both O'Regan & Humbert, 1989; Wichmann & Hill, 2001a.

Confidence intervals and goodness of fit

A point estimate has uncertainty because the number of trials is finite and responses vary. Researchers may use likelihood-based models, bootstrap resampling, and simulation-based goodness-of-fit tests to evaluate how securely the data locate a threshold. Wichmann and Hill's paired methodological studies explain parameter fitting, sampling, goodness of fit, and bootstrap confidence intervals Wichmann & Hill, 2001a; Wichmann & Hill, 2001b. A result should provide sufficient information to distinguish a precise difference from one that falls within sampling uncertainty.

Signal Detection Theory: Sensitivity and Response Criterion

Detection tasks combine noisy sensory evidence with a decision about how to respond. An observer who expects many targets may answer “present” more readily than one who expects very few. Signal detection theory makes this distinction explicit. In a standard yes/no experiment, researchers record hits, misses, false alarms, and correct rejections. Sensitivity is estimated from the separation of modeled signal-plus-noise and noise evidence distributions; response criterion describes the tendency to label ambiguous evidence as signal-present.

Under an equal-variance Gaussian model, d′ is commonly calculated from the difference between normal-deviate-transformed hit and false-alarm rates. A higher d′ indicates better discrimination between the modeled signal and noise distributions under those assumptions. Different distributional assumptions lead to different signal detection indices; extreme hit and false-alarm proportions require specified corrections. A raw yes-response rate alone is insufficient to estimate sensitivity. The importance of uncertainty in near-threshold visual judgments is illustrated by formal work on visual contrast detection Pelli, 1985.

Signal detection theory complements threshold measurement. A threshold asks where performance meets a criterion, while a sensitivity analysis investigates the extent to which observed changes reflect evidence quality or decision policy. Neither metric by itself establishes the conscious qualities of an experience. The appropriate inference depends on the evidence distributions, task design, and response model.

Classical Psychophysical Methods

The method of constant stimuli

The experimenter selects several fixed intensities or stimulus differences and presents them repeatedly, often in randomized order. The participant's responses are used to estimate the whole psychometric function. Because trials cover a defined range, investigators can examine threshold, slope, bias, and potentially lapses. The drawback is efficiency: many trials may be spent on intensity levels already known to be easy or impossible, especially if the expected threshold was poorly estimated.

The method of limits

An ascending series begins with a weak stimulus and increases strength until the participant indicates detection or a particular judgment; a descending series reverses the direction. Averaging transition points provides a convenient estimate. Anticipation, habituation, order effects, and response expectations can shift these transitions. Modern staircase methods modify the basic sequential logic with explicit rules and statistical targets Treutwein, 1995.

The method of adjustment

The participant manipulates the comparison stimulus until it matches a reference or reaches a barely detectable level. Adjustment can be efficient and intuitive, particularly in demonstrations and calibration tasks. It also exposes estimates to motor precision, starting positions, anchoring, and individual response strategies. Repeated randomized starts and comparison with other procedures improve interpretability. The best method depends on whether the objective is simple demonstration, rigorous threshold estimation, magnitude scaling, or comparison of experimental groups.

Adaptive Staircases and Modern Threshold Estimation

An adaptive procedure chooses subsequent stimulus intensities based on earlier answers. A simple staircase may decrease intensity after correct responses and increase it after errors. Transformed up-down procedures use multiple consecutive correct responses or other rules to target particular probabilities of correct discrimination. Their target performance levels follow from the specified algorithm; they are not interchangeable with the 50% yes/no detection convention Levitt, 1971.

Bayesian adaptive procedures instead maintain an evolving probability distribution over parameters such as threshold and select informative next stimuli. The QUEST procedure, introduced by Watson and Pelli, is a foundational example Watson & Pelli, 1983. Efficiency is attractive because human testing time is limited, but a fast estimate remains dependent on initial model assumptions, potential lapse behavior, and stimulus calibration. A badly specified function can yield a confident-looking but systematically incorrect result Treutwein, 1995.

Choosing an adaptive method responsibly

An investigator should determine whether the aim is a single threshold or the full curve, whether the participant is expected to lapse, whether stimulus intensities are bounded, and whether the response format changes guessing probability. Prior distributions and stopping rules should be prespecified where possible. Researchers should report reversal counts or fitted estimates, trial counts, uncertainty intervals, exclusions, and any transformations of physical units. Efficiency alone is not a measure of validity.

An Example Experiment: Measuring Contrast Discrimination

Suppose researchers want to estimate how much the contrast of a visual pattern must increase before participants can reliably say which of two patterns has higher contrast. They calibrate a display, define the spatial frequency and size of the pattern, keep viewing distance fixed, and control room illumination. Each trial presents a standard and comparison pattern in randomized positions or intervals. The observer chooses the higher-contrast one. The experiment samples multiple comparison differences, records proportions correct, and fits an appropriate psychometric function.

If the study adopts 75% correct as its criterion for a balanced two-alternative task, the difference at that fitted level becomes its operational JND. The procedure can be repeated at several reference contrasts. Plotting those JNDs against reference contrast tests whether Weber proportionality approximates the observed data. Reporting the display's calibration, the function's slope, trial counts, confidence intervals, and any excluded trials allows readers to judge whether the estimated fraction is persuasive Wichmann & Hill, 2001a.

A second experiment might ask the participant to assign numerical magnitude estimates to the same patterns. The resulting scaling function answers a different question: how judged intensity varies with physical contrast under the given instructions. Even when both tests involve the same stimulus family, their fitted parameters represent different constructs. This distinction is central to psychophysics.

Reliability, Validity, and Reproducibility

An estimate is reliable when repeated measurements under suitably equivalent conditions are sufficiently consistent for its intended use. It is valid when the task and analysis support the proposed interpretation. Test-retest reliability depends on practice, sensory adaptation, equipment, observer motivation, and within-person fluctuation. Group-level averages can conceal meaningful differences in individual psychometric slopes and bias. Small samples of trials may produce unstable parameters, especially when errors occur away from the estimated threshold Wichmann & Hill, 2001b.

Construct validity requires attention to what the response actually measures. For example, slow instructions or a confusing interface may make a participant appear less sensitive without altering sensory transduction. Motor limitations may affect adjustment tasks. A participant's numeracy may influence magnitude estimates. Recent experimental critiques show how response scaling can change estimated power-law exponents Mertens et al., 2021. A psychophysical parameter should therefore be interpreted in light of task demands as well as sensory processes.

Reproducibility improves when studies provide calibrated stimulus properties, code or complete task specifications, randomization methods, stopping rules, model-fitting details, estimates with uncertainty, and access to aggregate or appropriately protected trial-level data. Models should be checked for fit rather than assumed correct. Replication across observers, apparatus, laboratories, and response methods can distinguish robust patterns from setup-specific results.

Why Thresholds Change Across Conditions and People

The measured threshold changes with the characteristics of the stimulation and the observer's circumstances. Adaptation to background light or prolonged vibration can shift sensitivity. A masking sound makes a tone harder to detect. Brief stimulus presentation can yield different performance from a longer exposure. Divided attention and uncertainty about target location can affect task responses. Sensory aging, hearing differences, visual correction, fatigue, and prior training may also change performance. These are empirical questions requiring suitable designs rather than universal effects of one simple cause.

An individual difference in psychophysical performance is a measurement finding, not by itself a diagnosis. Clinical evaluations may incorporate carefully validated sensory procedures, but diagnosis also involves history, examination, symptom patterns, impairment, and appropriate professional standards. A low task score can reflect laboratory conditions or ordinary variation. Researchers and clinicians should avoid converting a single threshold value into a medical conclusion.

Children, older adults, and accessibility

Age-appropriate instructions, response options, trial durations, and breaks influence the quality of threshold estimates. Child participants may require simpler forced-choice responses and fewer trials; older participants may need adaptations to vision, hearing, or response speed. Researchers should test comprehension and distinguish failure to follow a task from failure to discriminate a stimulus. Accessibility is part of measurement validity, not merely a convenience.

From Laboratory Sensation to Everyday Experience

Psychophysical laws help explain why a modest increase in illumination may be conspicuous in a dim room yet relatively hard to notice in bright surroundings, why sensory design must consider background levels, and why equal increments in device settings may not yield equal perceived changes. Such explanations require careful attention to adaptation, stimulus calibration, and the physical dimensions being compared. Weber proportionality or Stevens scaling can guide hypotheses about interfaces, audio, lighting, texture, and display contrast, but a design recommendation should be validated in its actual environment.

The method also contributes to applied vision and hearing research, human factors, ergonomics, product testing, and basic neuroscience. Experiments can compare visibility of symbols, detectability of warnings, discrimination of textures, or perceived intensity of a signal. Transferring laboratory results to consumer devices or real environments requires accounting for ambient conditions, participant diversity, competing stimuli, and actual usage patterns.

What Contemporary Evidence Establishes—and What Remains Debated

It is well established that carefully controlled physical stimulation produces systematic, measurable changes in detection, discrimination, and magnitude judgments, and that modern statistical methods can estimate response functions with quantifiable uncertainty. Weber-like scaling is observed in some ranges and paradigms. Fechner's logarithmic construction and Stevens' power law remain influential approaches for describing how stimulus and judgment magnitudes relate Johnson et al., 2002; Stevens, 1957.

The disputed issues concern universality, the psychological meaning of equal JNDs, dependence on instructions and numerical response formats, and how best to relate behavioral functions to sensory and neural mechanisms. Bayesian accounts of magnitude estimation offer one framework for response biases and apparent regularities, but competing explanations remain empirically testable Petzschner et al., 2015. The existence of a well-fitting equation does not establish that every neuron implements the same equation or that all conscious intensities share one scale.

The strongest methodological conclusion is practical: specify the stimulus, the participant's task, the model, the threshold criterion, and the uncertainty. With those details, psychophysical research can support carefully delimited claims about sensory performance and perception. Without them, a threshold or exponent has little standalone meaning.

Frequently Asked Questions

What does psychophysics measure?

Psychophysics measures systematic relations between controlled physical stimulus properties and human or animal responses. Typical outcomes include detection probability, discrimination accuracy, estimated perceptual magnitude, fitted thresholds, slopes, and response bias. A particular numerical result is meaningful in the context of its experimental task and statistical model.

What is the difference between an absolute threshold and a difference threshold?

An absolute threshold describes performance in detecting a stimulus relative to absence or background stimulation. A difference threshold describes performance in distinguishing two stimulus values. Both are conventionally defined performance points and depend on experimental design, rather than functioning as universal fixed sensory switches.

What is Weber's law in simple terms?

Weber's law says that within some stimulus ranges the change required for reliable discrimination is approximately a constant fraction of the starting intensity. If the baseline doubles, the required increment may also roughly double. The fraction varies across modalities, contexts, and procedures.

Is a just-noticeable difference always the smallest possible detectable change?

No. A JND is ordinarily defined by a particular discrimination criterion, such as the stimulus increment corresponding to 75% correct in a specified two-alternative task. Under noisy conditions, weaker differences may occasionally be noticed and larger differences may sometimes be missed.

Is Weber's law the same as the Weber–Fechner law?

Weber's law describes how measured discrimination increments vary with reference intensity. Fechner's law adds assumptions about the subjective size of successive discrimination steps and derives a logarithmic magnitude scale. They are historically linked yet represent different claims.

Is Stevens' power law more accurate than Fechner's law?

Accuracy depends on the sensory dimension, stimulus range, measurement paradigm, and model comparison. Stevens' power law often fits direct magnitude estimates, while logarithmic functions can remain useful in other contexts. Neither model is a universal description of every perceptual response Johnson et al., 2002.

Why do different experiments report different thresholds?

Thresholds change with stimulus calibration, background noise, adaptation, response instructions, guessing probabilities, criterion choices, participant variability, and statistical fitting. A meaningful comparison checks these details before interpreting an observed difference as a change in underlying sensitivity.

Can a psychophysical test diagnose ADHD or another disorder?

A psychophysical task can supply information about performance on a specific sensory or perceptual operation. Such a task alone does not establish ADHD, a neurocognitive disorder, or another clinical diagnosis. Clinical conclusions require disorder-specific evidence, validated assessment, and appropriate professional evaluation.

Related Articles

Perception: How the Mind Interprets Sensory Information — The broader science of how sensory information becomes structured perceptual understanding.

Sensation vs Perception: Differences and Everyday Examples — A foundational explanation of the relationship between sensory registration and perceptual interpretation.

References

Holden, J. K., Francisco, E. M., Zhang, Z., Baric, C., & Tommerdahl, M. (2011). An undergraduate laboratory exercise to study Weber's law. Journal of Undergraduate Neuroscience Education, 9(2), A71–A74. PubMed.

Johnson, K. O., Hsiao, S. S., & Yoshioka, T. (2002). Neural coding and the basic law of psychophysics. The Neuroscientist, 8(2), 111–121. DOI.

Levitt, H. (1971). Transformed up-down methods in psychoacoustics. The Journal of the Acoustical Society of America, 49(2B), 467–477. DOI.

Mertens, A., Mertens, U. K., & Lerche, V. (2021). On the difficulty to think in ratios: A methodological bias in Stevens' magnitude estimation procedure. Attention, Perception, & Psychophysics, 83, 2347–2365. DOI.

O'Regan, J. K., & Humbert, R. (1989). Estimating psychometric functions in forced-choice situations: Significant biases found in threshold and slope estimations when small samples are used. Perception & Psychophysics, 46(5), 434–442. DOI.

Pelli, D. G. (1985). Uncertainty explains many aspects of visual contrast detection and discrimination. Journal of the Optical Society of America A, 2(9), 1508–1532. DOI.

Petzschner, F. H., Glasauer, S., & Stephan, K. E. (2015). A Bayesian perspective on magnitude estimation. Trends in Cognitive Sciences, 19(5), 285–293. DOI.

Stevens, S. S. (1957). On the psychophysical law. Psychological Review, 64(3), 153–181. DOI.

Treutwein, B. (1995). Adaptive psychophysical procedures. Vision Research, 35(17), 2503–2522. DOI.

Watson, A. B., & Pelli, D. G. (1983). QUEST: A Bayesian adaptive psychometric method. Perception & Psychophysics, 33(2), 113–120. DOI.

Wichmann, F. A., & Hill, N. J. (2001a). The psychometric function: I. Fitting, sampling, and goodness of fit. Perception & Psychophysics, 63(8), 1293–1313. DOI.

Wichmann, F. A., & Hill, N. J. (2001b). The psychometric function: II. Bootstrap-based confidence intervals and sampling. Perception & Psychophysics, 63(8), 1314–1329. DOI.