A measurement result is influenced by more than the instrument used to obtain it. The instrument may have calibration uncertainty, limited resolution, drift or nonlinear response. Environmental conditions such as temperature, humidity, pressure and vibration may influence the instrument or the measurand itself. The observer may introduce uncertainty through alignment, reading, timing or interpretation. The measurement method may rely on approximations or assumptions that do not perfectly represent the physical process. The sample may be heterogeneous, contaminated or unstable. Reference standards also have uncertainty, which can be transferred through a calibration chain. Mathematical models, numerical approximations and data-processing methods introduce additional contributions. These influences should be identified systematically. A useful classification distinguishes uncertainty evaluated using statistical analysis of repeated observations from uncertainty evaluated using other available information. These are called Type A and Type B evaluations. This classification describes the method of evaluation, not whether a source is random or systematic. A systematic influence can be evaluated statistically in some circumstances, and a random influence can be evaluated using non-statistical information. Uncertainty sources may be independent or correlated. Assuming independence without justification can lead to an incorrect combined uncertainty. A complete evaluation should consider both direct and indirect influences, their probability distributions, sensitivity coefficients and possible correlations.
Sources and Classification of Measurement Uncertainty
Measurement uncertainty originates from multiple influences associated with instruments, measurement procedures, environmental conditions, observers, samples, reference standards and mathematical models. Identifying and classifying these sources is the foundation of a complete uncertainty evaluation.
For a measurement model y = f(x₁, x₂, ..., xₙ), uncertainty contributions are evaluated from the uncertainties of the input quantities and their influence on the output quantity.
Explanation
Variables
| Symbol | Meaning | Si Unit |
|---|---|---|
y | Output quantity or measurement result | Unit of the output quantity |
xᵢ | An input quantity in the measurement model | Unit of the corresponding input quantity |
u(xᵢ) | Standard uncertainty of input quantity xᵢ | Same unit as xᵢ |
cᵢ | Sensitivity coefficient describing how the output changes with an input quantity | Unit of y divided by unit of xᵢ |
Example or application
A worked example has not yet been added.
Definition
A source of measurement uncertainty is an influence, input quantity or limitation that contributes to the uncertainty associated with a measurement result.
Purpose
To identify, organize and evaluate all significant influences that contribute to measurement uncertainty, preventing important sources from being overlooked.
Importance
A structured classification helps laboratories build uncertainty budgets, improve measurement procedures, identify dominant uncertainty contributions and make reliable comparisons between measurement results.
Classification By Evaluation Method
Type A
Description
Evaluation of uncertainty by statistical analysis of a series of observations obtained under defined measurement conditions.
Typical Information
Repeated measurements
Sample means
Experimental variance
Standard deviation
Type B
Description
Evaluation of uncertainty using information other than statistical analysis of the current series of observations.
Typical Information
Calibration certificates
Manufacturer specifications
Reference data
Previous measurements
Resolution limits
Scientific knowledge
Important Distinction
Type A and Type B classify evaluation methods. They do not directly classify uncertainty as random or systematic.
Classification By Physical Source
| Source | Mechanisms | Example |
|---|---|---|
Instrumental |
| A balance has a calibration uncertainty associated with its displayed mass. |
Environmental |
| Thermal expansion changes the measured length of a metal component. |
Observer-related |
| Different observers may identify the endpoint of a scale differently. |
Method-related |
| A simplified mathematical model may neglect air resistance. |
Sample-related |
| A small sample may not represent the composition of an entire material batch. |
Reference-related |
| A reference weight has a certified value accompanied by uncertainty. |
Data-acquisition and computational |
| A digital sensor converts a continuous signal into discrete numerical values. |
Classification By Influence
Direct Influence
An influence that affects the measured quantity or instrument indication directly.
Indirect Influence
An influence that affects the result through another quantity or through the measurement model.
Independent Contributions
Contributions that can reasonably be treated as statistically independent under the adopted model.
Correlated Contributions
Contributions that are statistically related and therefore require covariance terms during combination.
Systematic And Random Influences
Systematic Influence
An influence that contributes to a consistent or predictably changing deviation.
Random Influence
An influence that produces variation in an unpredictable manner.
Relationship To Type AAnd Type B
Either kind of influence may be evaluated using Type A or Type B methods, depending on the information available.
Identification Procedure
Define the measurand and the complete measurement process.
Draw a process diagram or list every measurement stage.
Identify instrument, environmental, observer, sample, reference and computational influences.
Examine calibration certificates and instrument specifications.
Study repeated observations and previous measurement records.
Determine which input quantities are correlated.
Estimate the magnitude and relevance of each contribution.
Document the assumptions used in the evaluation.
Worked Example
Problem
A laboratory measures the length of a metal rod. Identify possible uncertainty sources.
Sources
Calibration uncertainty of the measuring instrument.
Limited instrument resolution.
Temperature-dependent expansion of the rod.
Alignment and reading variation.
Surface irregularities at the measurement points.
Uncertainty in the reference standard used for calibration.
Interpretation
The final uncertainty must account for the significant contributions through an appropriate measurement model. The sources should not simply be added as raw numerical values.
Common Mistakes
Considering only instrument resolution.
Treating every Type B contribution as systematic error.
Assuming that all uncertainty sources are independent.
Ignoring uncertainty in reference standards.
Including the same physical influence twice in an uncertainty budget.
Ignoring sample variability and measurement-method limitations.
Applications
Uncertainty-budget development
Laboratory quality assurance
Instrument calibration
Experimental design
Measurement-method validation
Scientific data analysis
Reference
JCGM 100:2008, Evaluation of Measurement Data — Guide to the Expression of Uncertainty in Measurement (GUM).
JCGM 200, International Vocabulary of Metrology (VIM).
NIST Technical Note 1297, Guidelines for Evaluating and Expressing the Uncertainty of NIST Measurement Results.
Related Topic
SCI-REF-003-073
SCI-REF-003-083
SCI-REF-003-086
SCI-REF-003-088
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