Significant Figures and Measurement Uncertainty

Introduction

A calculator will display more digits than an experiment can justify. Rounding a result prevents the display from implying unrealistic detail, but significant-figure rules alone do not evaluate the full uncertainty. A trustworthy analytical report states the measurand, value, unit and an uncertainty or context appropriate to the decision. Digits are a communication tool, not proof of accuracy.

Core explanation

Measured quantities have limited resolution and uncertainty. A mass recorded as 0.5234 g suggests more detail than a balance reading of 0.5 g, but its final digit still has some uncertainty. Exact counted numbers and defined conversion factors do not limit significant figures in the same way as measurements. In multiplication or division, a common classroom rule rounds the result to the least number of significant figures among measured inputs. In addition or subtraction, the limiting decimal place is the coarse one. Carry extra digits internally and round once at the end to reduce accumulated rounding error.

These rules are convenient approximations. They do not capture systematic errors, correlated uncertainties or the actual tolerance of glassware and calibration. For example, writing 10.00 mL for a pipette does not mean its delivered volume is perfectly known to 0.01 mL; use its calibration and tolerance to estimate uncertainty. Conversely, a result may legitimately require an extra guard digit in intermediate work. An evaluated uncertainty is more informative than relying only on digit counting.

Standard uncertainty is an uncertainty component expressed like a standard deviation. Some components come from repeated observations; others come from calibration certificates, stated tolerances or physical models. Combined standard uncertainty brings relevant contributions together through an appropriate propagation model. Expanded uncertainty U = k u c multiplies the combined standard uncertainty by a coverage factor k. A statement such as 12.4 ± 0.3 mg L⁻¹ must specify whether 0.3 is a standard or expanded uncertainty and, for the latter, the chosen k or approximate coverage interpretation.

Usually round an uncertainty to one or two significant figures, then round the reported value to the same decimal place. For example, 12.437 mg L⁻¹ with expanded uncertainty 0.286 mg L⁻¹ might be reported as (12.44 ± 0.29) mg L⁻¹ when two digits in the uncertainty are useful. The exact convention depends on the reporting standard and decision context. A result near a regulatory threshold needs transparent uncertainty treatment, not a strategically chosen rounding direction.

Always preserve units and sample basis. A number in mg L⁻¹ for a final extract is not necessarily the original sample concentration if dilution or extraction occurred.

Step-by-step reasoning

1. Keep unrounded values while applying stoichiometry, calibration and dilution factors. 2. Identify uncertainty sources from sampling, preparation, calibration and measurement. 3. Combine appropriate components and decide whether standard or expanded uncertainty is needed. 4. Round the uncertainty sensibly and align the value's decimal place. 5. State analyte, sample basis, unit and uncertainty interpretation in the final report.

Visual explanation

Draw a result box with four labels: measurand, value, unit and uncertainty. Under it show a raw calculator display, 12.437286, turning into (12.44 ± 0.29) mg L⁻¹. Arrows from balance, flask, calibration and replicate scatter enter the uncertainty box, demonstrating that more than one instrument reading controls confidence.

Real-world analogy

Reporting the travel time of a bus as 37.482916 minutes because a computer calculated it would be misleading if traffic changes it by several minutes. A rounded time and an honest range better communicate what a rider can expect. Chemical measurements similarly need a precision of expression matched to the quality of evidence.

Real-world example

A water laboratory obtains 4.876 mg L⁻¹ nitrate from its calibration equation. Replicates, standards and sample preparation together support an expanded uncertainty of 0.21 mg L⁻¹. It reports (4.88 ± 0.21) mg L⁻¹ with the stated coverage factor, rather than listing all calculator digits or simply writing 4.9 without explaining the uncertainty relevant to the decision.

Why?

Why round the uncertainty before aligning the value? The uncertainty sets the meaningful scale at which differences in the value can be interpreted. If uncertainty is around 0.3 mg L⁻¹, hundredths or thousandths beyond the reporting convention can imply false certainty. Rounding too coarsely can also hide information, so the report should follow its method's stated practice.

Common misconception

“Four significant figures means four-figure accuracy” is false. Systematic bias can affect all four digits. Another misconception is that ± always means a 95% interval; it may be a standard deviation, standard uncertainty, expanded uncertainty or another quantity. A report should identify which one.

Worked example

An analyte mass is calculated from 0.1256 g precipitate multiplied by a gravimetric factor of 0.6043, giving 0.1256 × 0.6043 = 0.07590008 g before final rounding. The calculated mass to four significant figures is 0.07590 g. If evaluated standard uncertainty is 0.00018 g, a more informative report is (0.07590 ± 0.00018) g with “standard uncertainty” stated. The arithmetic digits and uncertainty statement serve different purposes.

Quick check

1. Is “7.3456 mg L⁻¹” automatically a high-quality result because it has five significant figures? Answer: No. The digits only show how it was written. Sampling, calibration, bias and uncertainty must support them; otherwise extra decimal places create false precision.

Exam focus

Apply classroom rounding rules correctly but explain their limits. Keep guard digits during calculations. Distinguish standard uncertainty from expanded uncertainty, and give k when using the latter. State units and measurand; a number without those is incomplete even if rounded perfectly.

Advanced insight

Uncertainty is not a bound on every possible mistake. A mislabeled sample or unrecognised interferent can place a result far outside a calculated interval because the measurement model omitted the problem. Quality assurance therefore combines numerical uncertainty propagation with procedural controls, independent checks and clear definition of the measurand.

Summary

Significant figures prevent reporting unjustified calculator digits, while measurement uncertainty evaluates how well a stated result is supported. The two ideas complement one another but are not interchangeable. A useful report gives a defined analyte and sample basis, value, unit, rounded uncertainty and its interpretation.

Practice questions

1. A result is 3.4567 g with expanded uncertainty 0.12 g. Give a consistent report. Answer: Report approximately (3.46 ± 0.12) g, stating that 0.12 g is expanded uncertainty and giving the coverage factor if known.

2. Why should intermediate calculations retain more digits than the final result? Answer: Early rounding can accumulate errors through subsequent multiplications or subtractions. Guard digits preserve numerical accuracy until the uncertainty and reporting precision are assessed.

3. What is missing from “chloride = 20 ± 2”? Answer: Units, sample basis and the meaning of ±2 are absent. A complete report specifies, for example, mg L⁻¹ in a defined water sample and whether 2 is standard or expanded uncertainty.