A-Level Physics Uncertainty, Errors and Significant Figures: A Practical Calculation Guide
A-Leveluncertaintyexperimental errorssignificant figuresdata analysisgraphs

A-Level Physics Uncertainty, Errors and Significant Figures: A Practical Calculation Guide

SStudy Physics Editorial Team
2026-11-14
8 min read

A practical A-level physics guide to uncertainty, errors, significant figures, repeated measurements, graph gradients and data analysis.

Uncertainty questions become much easier when you use the same workflow every time. This A-level physics guide explains how to calculate absolute and percentage uncertainty, combine uncertainties, use significant figures, assess graphs and describe errors in the precise language expected in data-analysis questions.

Overview

Experimental results are measurements, not perfectly exact values. A measured length of 12.4 cm does not mean the true length is exactly 12.400000 cm; it means the measurement is best reported as approximately 12.4 cm, with a range of reasonable values around it.

In A-level physics, keep three ideas separate:

  • Uncertainty describes the range within which a measured value is expected to lie.
  • Random error causes readings to vary unpredictably between repeats. Repeats and a mean can reduce its effect.
  • Systematic error shifts readings consistently in one direction. Repeating the measurement does not remove it.

A useful starting statement is:

Measured value = best estimate ± absolute uncertainty

For example, 12.4 ± 0.1 cm gives a best estimate of 12.4 cm and an absolute uncertainty of 0.1 cm. The corresponding percentage uncertainty is:

Percentage uncertainty = (absolute uncertainty ÷ measured value) × 100

For this example:

(0.1 ÷ 12.4) × 100 = 0.806...% ≈ 0.8%

Always check the terminology and mathematical conventions required by your own specification. AQA, Edexcel and OCR questions use closely related principles, but the wording of practical and data-analysis questions can differ.

What to track

1. Instrument resolution and absolute uncertainty

The resolution is the smallest scale division or display increment that can be read. For an analogue instrument, a common school-laboratory convention is to take the uncertainty as about half the smallest division when a reading is taken between marks. For a digital instrument, the displayed increment is often used as the uncertainty. However, follow the method or data provided in the question if it specifies a different convention.

For a ruler marked in millimetres, a single length reading might be written as 86.0 ± 0.5 mm. If you measure a distance by recording two ruler positions and subtracting them, the absolute uncertainties may need to be added. Thus two readings each with ±0.5 mm can produce a distance uncertainty of ±1.0 mm.

2. Repeated measurements

Record repeats rather than quietly discarding values that do not agree. For readings of 2.41 s, 2.38 s, 2.44 s and 2.40 s, the mean is:

(2.41 + 2.38 + 2.44 + 2.40) ÷ 4 = 2.4075 s

Report the mean as 2.41 s if the measurements are recorded to two decimal places. The spread can help you judge random variation. A simple range is the largest reading minus the smallest reading: 2.44 − 2.38 = 0.06 s. Do not automatically label the range as the uncertainty unless the question or your practical method tells you to do so. Explain what you have used.

3. Propagation through calculations

For quantities added or subtracted, add the absolute uncertainties:

If Q = A + B, then ΔQ = ΔA + ΔB.

If a quantity is multiplied or divided, add the percentage uncertainties:

If Q = AB or Q = A ÷ B, then the percentage uncertainty in Q is approximately the percentage uncertainty in A plus the percentage uncertainty in B.

For a power relationship, multiply the percentage uncertainty by the power:

If Q = An, then the percentage uncertainty in Q is approximately n times the percentage uncertainty in A.

Worked example: A speed is calculated from a distance of 1.20 ± 0.01 m and a time of 0.80 ± 0.02 s. The speed is 1.20 ÷ 0.80 = 1.50 m s−1. The distance has a percentage uncertainty of 0.01 ÷ 1.20 × 100 = 0.83%. The time has a percentage uncertainty of 0.02 ÷ 0.80 × 100 = 2.5%. Therefore, the speed has an uncertainty of approximately 3.3%. The absolute uncertainty in speed is 1.50 × 0.033 = 0.0495 m s−1, so a sensible report is approximately 1.50 ± 0.05 m s−1.

4. Significant figures and decimal places

Do not give a calculated result more precision than the measurements justify. A practical convention is to give an uncertainty to one significant figure, or sometimes two when the first digit is small, and to round the measured value to the same decimal place.

For example, 4.836 ± 0.173 N could be reported as 4.84 ± 0.17 N if two significant figures are retained in the uncertainty. The value and uncertainty finish at the same decimal place. Avoid reporting 4.836 ± 0.2 N, because the decimal places do not match.

5. Percentage difference

When comparing an experimental value with an accepted or reference value, use:

Percentage difference = |experimental value − reference value| ÷ reference value × 100

This describes how far the result is from the reference value. It is not automatically the same as percentage uncertainty. A result may differ from the reference value by more than its uncertainty, suggesting a systematic issue, an inappropriate model or an error in the procedure.

6. Graph gradients and intercepts

For a straight-line graph, calculate the gradient using two widely separated points on the best-fit line, not necessarily two raw data points:

gradient = change in y ÷ change in x

Suppose a graph gives a gradient of 2.40 units and the steepest and shallowest acceptable lines give gradients of 2.55 and 2.25. A reasonable gradient uncertainty is half the difference:

(2.55 − 2.25) ÷ 2 = 0.15

Report the gradient as 2.40 ± 0.15, or as a percentage uncertainty of 0.15 ÷ 2.40 × 100 = 6.25%. State how the uncertainty was obtained. An intercept that is significantly different from zero can indicate a systematic error or an incomplete model, although you should consider the uncertainty in the intercept before making a firm conclusion.

Results-table template

QuantityTrial 1Trial 2Trial 3MeanUncertainty
Independent variable / unitvaluevaluevaluevalue± value
Dependent variable / unitvaluevaluevaluevalue± value

Put units in the column headings, keep consistent decimal places, and record a zero where appropriate rather than leaving a cell ambiguous. Add a separate column for percentage uncertainty when the calculated quantity is central to the conclusion.

Error language examiners can credit

IssueBetter explanation
Random variationRepeat the measurement, calculate a mean and use more readings to reduce the effect of scatter.
ParallaxRead the scale at eye level or use a pointer and fixed viewing position.
Zero errorCheck the instrument before use and subtract the zero offset where appropriate.
Heat lossInsulate the apparatus, reduce the time between readings or include a correction based on a justified method.
Small gradient uncertaintyUse a larger range of independent-variable values and plot points across the available graph area.

“Improve accuracy” is usually too vague. Name the source of error, explain its effect and give a practical change that addresses it.

Cadence and checkpoints

Use this article as a repeatable data-analysis checkpoint rather than reading it only before an examination.

  • After each required practical: check units, resolution, repeats, means, uncertainty calculations and the quality of your evaluation.
  • After each topic-question set: record whether mistakes came from selecting a formula, converting units, propagating uncertainty or rounding.
  • Monthly during revision: complete one unfamiliar graph or uncertainty question without notes, then compare your method with the mark scheme.
  • Before mock or final exams: practise a complete data-analysis question under time pressure, including a results table, graph, gradient and evaluation.
  • After an assessment: update your error log with the exact step that failed and a short rule for preventing the same mistake.

Keep a small tracker with five columns: question or practical, error type, correct method, units or precision issue, and next review date. This turns recurring mistakes into specific revision targets.

How to interpret changes

If your percentage uncertainties are consistently large, first identify the dominant measurement. Improving a quantity with a 5% uncertainty will have less effect if another quantity contributes 15%. Increase the measured value where possible, use an instrument with finer resolution, or repeat readings to reduce random scatter.

If repeated readings become closer together, random variation may be under better control, but this does not prove that systematic error has disappeared. If a graph remains well aligned but its intercept is displaced, investigate zero error, background readings or an omitted constant. If the gradient changes substantially when one point is removed, the result may be sensitive to an outlier or the data range may be too small.

When comparing a result with a reference value, ask whether the difference is consistent with the combined uncertainty. Avoid claiming that two values are “the same” merely because they look close, and avoid claiming a method is invalid solely because one result differs slightly. Use cautious conclusions such as: “The values are consistent within the stated uncertainty” or “The difference is larger than the estimated uncertainty, so a systematic error may be present.”

When to revisit

Revisit this guide whenever you start a new required practical, meet a new calculation involving powers or gradients, or notice a repeated error in physics past papers or worked solutions. It is also worth checking your exam board’s current specification, formula sheet and practical guidance before final revision, because terminology and assessment conventions can be updated.

Before submitting any data-analysis answer, use this final checklist:

  1. Have all quantities got units, including graph axes?
  2. Are measured values recorded to consistent decimal places?
  3. Have you distinguished absolute uncertainty from percentage uncertainty?
  4. Did you add absolute uncertainties for sums and differences?
  5. Did you add percentage uncertainties for products and quotients?
  6. Did you multiply percentage uncertainty by the power for a power relationship?
  7. Did you calculate a graph gradient from suitable points on the best-fit line?
  8. Have you explained the source and effect of any error rather than naming it vaguely?
  9. Do the value and uncertainty have matching decimal places?
  10. Does the conclusion reflect the size of the uncertainty?

Save your completed tracker with your A-level physics notes and return to it at the next practical or revision checkpoint. The aim is not to memorise isolated rules, but to build a dependable routine that makes uncertainty, errors and significant figures part of every calculation.

Related Topics

#A-Level#uncertainty#experimental errors#significant figures#data analysis#graphs
S

Study Physics Editorial Team

Physics Education Editors

Senior editor and content strategist. Writing about technology, design, and the future of digital media. Follow along for deep dives into the industry's moving parts.