✦ScienexySCIENCE KNOWLEDGE LIBRARY
reference · SCI-REF-002-091

Precision, Accuracy and Significant Figures

Precision, accuracy and significant figures are essential concepts for interpreting and reporting numerical measurements. Accuracy describes closeness to a reference value, precision describes the agreement among repeated measurements, and significant figures communicate numerical resolution without implying unsupported certainty.

Measurement error relative to a reference value:

E = xmeasured − xreference

Relative error, when the reference value is suitable and nonzero:

Relative error = |E|/|xreference|

Sample standard deviation as a measure of dispersion:

s = √[Σ(xᵢ − x̄)²/(n − 1)]

Relative precision is not a universal standalone quantity; precision is commonly characterized using standard deviation, repeatability or another defined statistical measure.

Explanation

Scientific measurements must be interpreted not only by their numerical values but also by how reliably those values represent the quantities being measured. Precision and accuracy describe different aspects of measurement quality.

Accuracy refers to the closeness of agreement between a measured value and a reference value. In metrology, measurement accuracy is a qualitative concept rather than a numerical quantity. It is commonly assessed using measurement error, bias and uncertainty relative to an appropriate reference.

Precision describes the closeness of agreement among repeated measurements under specified conditions. A highly precise measurement produces results that are closely grouped, even if their average differs from the reference value. Precision is related to repeatability and reproducibility.

Repeatability refers to precision under nearly unchanged measurement conditions, such as the same instrument, operator, procedure and location over a short period. Reproducibility refers to precision under changed conditions, such as different laboratories, operators, instruments or measurement systems.

A measurement can be precise but inaccurate when a systematic bias affects all observations. It can also be relatively accurate on average but imprecise when repeated results fluctuate widely. The ideal measurement is both sufficiently accurate and precise for the intended purpose.

Significant figures are the meaningful digits used to represent a measured or calculated quantity. They communicate the scale of numerical resolution and help prevent results from implying greater precision than the measurements justify. However, significant figures are not a substitute for a complete uncertainty statement.

For example, writing a length as 2.5 cm communicates a different numerical resolution from writing it as 2.500 cm. The additional zeros in 2.500 are significant and indicate that the value is reported to a finer decimal place. Whether that precision is justified depends on the measurement method and its uncertainty.

In multiplication and division, a common introductory rule is to report the result with the same number of significant figures as the input having the fewest significant figures. In addition and subtraction, the result is commonly rounded to the least precise decimal place among the inputs. These rules are useful for basic reporting, but rigorous scientific uncertainty propagation is preferable in precision work.

Exact counted quantities and exact defined conversion factors do not limit significant figures. For example, the number 12 in a definition of a dozen is exact, and the SI conversion 1 m = 100 cm is exact. Measured quantities, in contrast, have uncertainty.

Rounding should generally be performed only at the final reporting stage. Keeping guard digits during intermediate calculations reduces the risk of accumulated rounding effects.

Variables

SymbolMeaningSi Unit

xmeasured

Measured value

Same as the measured quantity

xreference

Reference value used for comparison

Same as the measured quantity

E

Measurement error relative to the reference value

Same as the measured quantity

x̄

Mean of repeated measurements

Same as the measured quantity

s

Sample standard deviation

Same as the measured quantity

Example or application

A worked example has not yet been added.

Definition

Accuracy is the qualitative closeness of agreement between a measured value and a reference value. Precision is the closeness of agreement among repeated measurements. Significant figures are the digits used to represent the meaningful numerical resolution of a value.

Purpose

To evaluate measurement quality, communicate numerical resolution and report results without suggesting a level of certainty unsupported by the measurement process.

Importance

These concepts are essential for interpreting laboratory data, comparing experimental results, evaluating instruments and preventing misleading numerical claims. They help distinguish consistent measurements from measurements that are close to the accepted reference.

Accuracy And Precision Comparison

ConditionDescriptionInterpretation

High accuracy and high precision

Measurements are closely grouped and their results agree well with the reference value.

Desirable measurement performance.

High precision and low accuracy

Measurements are closely grouped but consistently displaced from the reference.

A systematic bias may be present.

Low precision and potentially high average accuracy

Measurements are widely dispersed, but their average may be near the reference value.

Random variation is substantial.

Low accuracy and low precision

Measurements are widely dispersed and do not agree well with the reference.

Both systematic and random effects may require investigation.

Significant Figure Rules

  • All nonzero digits are significant. For example, 347 has three significant figures.

  • Zeros between nonzero digits are significant. For example, 1002 has four significant figures.

  • Leading zeros are not significant. For example, 0.0045 has two significant figures.

  • Trailing zeros after a decimal point are significant. For example, 2.500 has four significant figures.

  • Trailing zeros in a whole number without a decimal point may be ambiguous. Scientific notation can remove the ambiguity.

  • In scientific notation, the digits in the coefficient determine the significant figures. For example, 3.20 × 10⁵ has three significant figures.

  • Exact counted quantities and exact defined conversion factors do not limit the significant figures of a calculation.

  • Rounding should normally be delayed until the final result.

Calculation Rules

Addition And Subtraction

Round the final result to the least precise decimal place among the measured inputs. For example, 12.11 cm + 0.3 cm = 12.4 cm.

Multiplication And Division

For introductory reporting, round the result to the fewest significant figures among the measured inputs. For example, 2.5 × 3.42 = 8.6 when reported to two significant figures.

Exact Numbers

Exact quantities do not restrict the significant figures of a result. For example, multiplying a measured value by an exact conversion factor does not introduce additional measurement uncertainty.

Uncertainty Based Reporting

For rigorous measurements, report the uncertainty explicitly and round the measured value to a compatible decimal place. Significant-figure counting alone cannot replace uncertainty analysis.

Repeatability And Reproducibility

Repeatability

Agreement among repeated measurements made under specified, nearly unchanged conditions.

Reproducibility

Agreement among measurements made under changed conditions, such as different operators, instruments, laboratories or procedures.

Importance

A method may show good repeatability in one laboratory but poorer reproducibility across laboratories. Both properties are important when evaluating scientific methods.

Bias And Random Variation

Bias is an estimate of a systematic displacement from a reference value. Random variation describes the spread of repeated observations. Correcting an identified bias can improve agreement with the reference, while reducing random variation can improve precision. Both should be considered when evaluating measurement performance.

Worked Example

Problem

A reference mass is 100.00 g. Two measurement methods produce the following results: Method A gives 100.40 g, 100.41 g and 100.39 g; Method B gives 99.80 g, 100.20 g and 100.00 g. Compare their precision and average accuracy.

Solution

  • Method A mean = (100.40 + 100.41 + 100.39)/3 = 100.40 g.

  • Method B mean = (99.80 + 100.20 + 100.00)/3 = 100.00 g.

  • Method A measurements are tightly grouped, so Method A is more precise.

  • Method B mean agrees with the reference value of 100.00 g, while Method A mean is 0.40 g higher.

  • Therefore, Method B has a more accurate average in this example, while Method A has greater precision.

Result

Method A: higher precision but a biased average. Method B: lower precision but an average equal to the reference value. This illustrates why precision and accuracy must not be treated as identical.

Rounding Example

Problem

Report 0.00045678 m to three significant figures.

Solution

  • Write the number in scientific notation: 4.5678 × 10⁻⁴ m.

  • Keep the first three significant digits: 4.56.

  • The next digit is 7, so the third digit is rounded upward.

Result

4.57 × 10⁻⁴ m

Common Misconceptions

  • A measurement with many decimal places is not necessarily accurate.

  • High precision does not guarantee closeness to the reference value.

  • A small measurement error in one experiment does not establish high precision.

  • Significant figures are not a complete representation of measurement uncertainty.

  • Trailing zeros should not be removed if they communicate meaningful measurement resolution.

  • Rounding every intermediate calculation can create avoidable numerical errors.

Limitations And Considerations

  • Accuracy cannot usually be established without an appropriate reference value.

  • Precision depends on the conditions under which repeated measurements are performed.

  • Significant-figure rules are simplified reporting conventions and may be inadequate for advanced uncertainty analysis.

  • A reference value may itself have uncertainty, which must be considered when evaluating agreement.

  • Different instruments may have different resolutions and systematic limitations.

Applications

  • Reporting laboratory measurements.

  • Comparing measurement instruments.

  • Evaluating repeatability and reproducibility.

  • Scientific calculations and numerical reporting.

  • Quality assurance in industrial processes.

  • Assessment of experimental methods and reference data.

Related Topics

  • SCI-REF-002-087

  • SCI-REF-002-089

  • SCI-REF-002-090

  • SCI-REF-002-092

  • SCI-REF-002-095

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.

  • JCGM 200 — International Vocabulary of Metrology (VIM).
  • Taylor, J. R. — An Introduction to Error Analysis: The Study of Uncertainties in Physical Measurements.
  • NIST — Guide for the Use of the International System of Units (SI).

Suggest a correction or source · More in Physical Constants