A measurement system can be affected by several error-producing mechanisms at the same time. Classifying errors helps explain their behavior and determines how they should be investigated. The three commonly taught categories are systematic errors, random errors and gross errors. These are practical categories; a complete uncertainty evaluation should also consider the underlying physical and statistical effects rather than relying only on labels.
Systematic error is a component of measurement error that remains constant or changes in a predictable manner under specified measurement conditions. It may be caused by an incorrect calibration, zero offset, instrument drift, an unsuitable measurement model or a consistent environmental influence. A systematic effect can shift many observations in approximately the same direction. Taking more readings generally does not remove this effect, although repeated measurements may reveal it when compared with a suitable reference.
Random error is the component of measurement error that varies unpredictably among repeated observations. It may result from electronic noise, small environmental fluctuations, mechanical vibration, limited observational ability or other changing influences. Random effects produce a distribution of measured values rather than a single consistent shift. Statistical analysis of repeated observations can estimate the associated variation. Averaging independent observations may reduce the standard uncertainty of the mean, provided the measurement conditions and statistical assumptions are appropriate.
Gross errors, sometimes called blunders, are substantial mistakes caused by incorrect actions or processing. Examples include reading the wrong scale, selecting the wrong unit, recording 15.2 instead of 12.5, using an incorrect conversion factor or connecting an instrument improperly. Gross errors are not simply unusually large random errors. They often arise from failures in the measurement procedure and require investigation, correction of the underlying process and transparent treatment of affected data.
Systematic and random effects are not always completely separate in a practical instrument. For example, an instrument may have a fixed calibration offset and also display random electronic noise. A temperature sensor may experience a slowly changing drift while individual readings fluctuate around the drifting value.
Systematic effects are commonly investigated using calibration, reference standards, comparison measurements, environmental control and validated models. Random effects are investigated using repeated observations, statistical summaries, time-series inspection and appropriate experimental design. Gross errors are investigated using procedural checks, independent verification, data validation and review of original records.
An important distinction is that a correction can compensate for an estimated systematic effect, but the correction itself may have uncertainty. A corrected result must therefore retain the uncertainty associated with the correction and the remaining measurement process.
Systematic, Random and Gross Errors
Measurement errors can arise through different mechanisms and produce different patterns in experimental results. Systematic errors introduce consistent or predictable deviations, random errors cause observations to vary unpredictably, and gross errors generally result from mistakes in measurement, recording or calculation. Recognizing these categories helps investigators select appropriate methods for detecting, evaluating and reducing measurement problems.
Measurement model: xᵢ = μ + b + εᵢ
Here, μ represents the reference or underlying quantity value, b represents a systematic bias under the stated conditions, and εᵢ represents a random contribution for observation i.
Mean of n observations: x̄ = (1/n) Σxᵢ
Sample standard deviation: s = √[Σ(xᵢ − x̄)²/(n − 1)]
Standard error of the mean: sₓ̄ = s/√n, when the observations are appropriately independent and the statistical model is suitable.
Corrected estimate: x꜀ = xₘ + C, where C is an applied correction.
Explanation
Variables
| Symbol | Meaning | Unit |
|---|---|---|
xᵢ | The ith measured observation | Unit of the measurand |
μ | Underlying or reference quantity value in the simplified measurement model | Unit of the measurand |
b | Systematic bias represented in the model | Unit of the measurand |
εᵢ | Random contribution in the ith observation | Unit of the measurand |
x̄ | Arithmetic mean of the observations | Unit of the measurand |
s | Sample standard deviation | Unit of the measurand |
n | Number of observations | Dimensionless count |
C | Correction applied to compensate for an estimated systematic effect | Unit of the measurand |
Example or application
A worked example has not yet been added.
Definition
Systematic Error
A component of measurement error that, in repeated measurements, remains constant or varies in a predictable way under specified conditions.
Random Error
A component of measurement error that, in repeated measurements, varies in an unpredictable manner.
Gross Error
A major measurement or data-processing mistake caused by an incorrect action, reading, recording, calculation or procedure. It is commonly treated as a blunder rather than as an ordinary statistical error component.
Purpose
The purpose of classifying measurement errors is to identify the mechanisms responsible for deviations, determine whether repeated observations or calibration are appropriate, and choose effective corrective and uncertainty-evaluation methods.
Importance
The distinction is important because different error types require different responses. Calibration can address recognized systematic effects, statistical analysis can characterize random variation, and procedural verification can prevent gross mistakes. Treating all errors as if they were random can lead to incorrect conclusions.
Systematic Errors
Main Characteristics
Can shift repeated observations in a consistent direction.
May remain approximately constant or vary predictably with time, temperature, load, frequency or another influence.
Can affect both the mean and the accuracy of a measurement result.
Are not generally eliminated by simply increasing the number of repeated observations.
Common Causes
| Cause | Explanation |
|---|---|
Zero error | An instrument indicates a nonzero value when the true input should be zero. |
Calibration bias | The instrument response differs consistently from the reference relationship. |
Environmental influence | Temperature, humidity, pressure or magnetic fields produce a consistent influence under the measurement conditions. |
Method error | The measurement procedure or mathematical model does not correctly represent the measurand. |
Instrument drift | The instrument response changes over time because of ageing, component changes or other influences. |
Reduction Methods
Calibrate instruments using appropriate reference standards.
Determine and apply justified corrections.
Control or monitor environmental influences.
Use an independently validated measurement method when possible.
Check instrument zero and operating condition before measurements.
Evaluate drift and repeat calibration at suitable intervals.
Random Errors
Main Characteristics
Produce variation among repeated observations.
May be positive or negative relative to an underlying value.
Can often be described statistically when enough suitable observations are available.
May be reduced in the estimate of a mean by averaging independent observations.
Common Causes
| Cause | Explanation |
|---|---|
Electronic noise | Unpredictable fluctuations in electronic circuits or sensor output. |
Environmental fluctuations | Small changing variations in temperature, airflow, vibration or illumination. |
Mechanical instability | Small changes in position, friction, contact force or mechanical motion. |
Observation variability | Small variations in reading an analogue scale or identifying an event. |
Reduction Methods
Take repeated observations under controlled conditions.
Use suitable averaging or statistical estimation.
Improve shielding, stabilization and environmental control.
Use an instrument with suitable noise performance.
Avoid unnecessary changes in the measurement procedure.
Gross Errors
Examples
Recording a length in centimetres as though it were in millimetres.
Reading the wrong scale or selecting the wrong measurement range.
Transcribing a decimal point incorrectly.
Using the wrong sample identification.
Applying an incorrect formula during data processing.
Forgetting to account for an instrument's known zero correction.
Detection And Prevention
Check original records against processed data.
Use independent verification of important calculations.
Apply plausible-range and consistency checks.
Use clear instrument labels and standardized procedures.
Repeat a measurement when a documented procedural problem is discovered and repetition is scientifically justified.
Preserve original observations and document any correction or exclusion.
Data Integrity
An observation should not be discarded merely because it differs from the others. An unusual value must be investigated using a justified procedure. If a gross mistake is confirmed, the reason for correction or exclusion should be documented. If no cause is established, the observation should be assessed using an appropriate statistical and scientific method.
Comparison
| Feature | Systematic | Random | Gross |
|---|---|---|---|
Primary behavior | Consistent or predictable deviation | Unpredictable variation | Mistake or blunder |
Effect of repeated measurements | Usually remains in the result unless corrected or controlled | Can be characterized statistically and may be reduced in the mean | May recur if the faulty procedure continues |
Typical investigation | Calibration and reference comparison | Repeated measurements and statistical analysis | Record review and procedural verification |
Example | A scale consistently reads 0.2 g too high | Small fluctuations in repeated voltage readings | A value is copied into the record with the wrong decimal point |
Worked Examples
| Title | Problem | Solution | Interpretation |
|---|---|---|---|
Identifying a systematic offset | A reference block has a length of 50.00 mm. An instrument repeatedly indicates approximately 50.30 mm under the same conditions. | Estimated bias = 50.30 mm − 50.00 mm = +0.30 mm. An estimated correction is C = −0.30 mm, provided the comparison is valid and the offset is applicable under the measurement conditions. | The consistent positive deviation suggests a systematic effect. Further calibration and uncertainty evaluation are needed. |
Evaluating random variation | Five suitable repeated readings of a voltage are 5.01 V, 4.99 V, 5.02 V, 5.00 V and 4.98 V. Calculate their mean. | x̄ = (5.01 + 4.99 + 5.02 + 5.00 + 4.98) / 5 = 5.00 V. | The readings vary around the mean. Their spread can be evaluated using the sample standard deviation. |
Recognizing a gross error | A laboratory record contains a reading of 125.0 cm, while the original instrument record and all procedural evidence show that the reading was 12.50 cm. | The discrepancy is a decimal-place transcription mistake. The original record should be retained, the transcription corrected with an audit trail, and the corrected value used in subsequent analysis. | This is a data-recording blunder, not ordinary random variation. |
Common Misconceptions
Systematic error is not necessarily constant forever; it can change predictably with operating conditions.
Random errors do not always follow a normal distribution.
Averaging repeated readings does not automatically remove systematic bias.
An outlier is not automatically a gross error.
A gross error should not be hidden by simply deleting an inconvenient observation.
Calibration can reveal and help correct systematic effects, but calibration itself has uncertainty.
Applications
Calibration laboratories
Repeated experimental measurements
Industrial production and inspection
Electronic sensor testing
Scientific data validation
Quality assurance systems
Environmental measurement networks
Medical laboratory instrumentation
Related Topics
Introduction to Measurement Errors
Sources and Mechanisms of Measurement Errors
Accuracy, Precision, Trueness and Repeatability
Error Detection, Correction and Reduction Techniques
Statistical Analysis of Repeated Measurements
Reference
| Title | Organization | Url |
|---|---|---|
International Vocabulary of Metrology (VIM), JCGM 200 | Joint Committee for Guides in Metrology | https://www.bipm.org/en/committees/jc/jcgm/publications |
Guide to the Expression of Uncertainty in Measurement | JCGM | https://www.bipm.org/en/committees/jc/jcgm/publications |
NIST/SEMATECH e-Handbook of Statistical Methods | National Institute of Standards and Technology | https://www.itl.nist.gov/div898/handbook/ |
Related Topic
SCI-REF-003-073
SCI-REF-003-075
SCI-REF-003-076
SCI-REF-003-077
SCI-REF-003-080
Sources and verification
Sources are shown as provided in the project data. Not every source has yet been independently verified. Please report errors or missing citations.
- Source verification pending; check an authoritative textbook, university, standards body, or peer-reviewed source.
Suggest a correction or source · More in Units & Measurement