Limitation of physical measurements — AQA A-Level Physics
Teaching note checked against the official specification.
Limitation of physical measurements covers the inherent uncertainties in experimental physics, including random and systematic errors, precision, accuracy, resolution, and the use of significant figures. Understanding these limitations is essential for evaluating experimental results and calculating uncertainty intervals in AQA A-Level Physics.
Test yourself on Limitation of physical measurements with AQA A-Level practice questions.
A week of Premium from your first lesson · Then the free plan continues · No card, no charge
Limitation of physical measurements explained
Limitation of physical measurements covers 7 specification points, including Random and systematic errors.
Read the full explanation
and Precision, repeatability, reproducibility,…. Open a point below for its explanation and revision guidance.
Random and systematic errors.
Every measurement is affected by error, and AQA expects you to distinguish two fundamentally different kinds. Random errors scatter repeated readings unpredictably about a mean, usually because of fluctuating conditions or the limits of reading a scale; they are reduced by taking many readings and averaging, since positive and negative deviations tend to cancel. Systematic errors shift all readings in one direction by a consistent amount or proportion, for example a zero error on a micrometer, a wrongly calibrated thermometer or a parallax habit; averaging does not remove them, so they must be identified and corrected, perhaps by re-zeroing the instrument or applying a calibration correction. A useful test is to plot repeated readings: random scatter looks like a spread, whereas a systematic shift moves the whole set away from the true value.
Precision, repeatability, reproducibility, resolution and accuracy.
These five terms describe different qualities of measurement and are easily confused. Precision describes how closely repeated readings agree with each other, so a precise set is tightly clustered. Repeatability is the closeness of repeated readings taken by the same person, using the same method and equipment, over a short time; reproducibility is closeness when the measurement is repeated by a different person, or with different equipment or conditions. Resolution is the smallest change an instrument can detect, such as 0.01 mm on a digital micrometer or 1 mm on a metre rule. Accuracy is how close a reading is to the true value. A set can be precise but inaccurate if a systematic error shifts every reading, and an instrument with fine resolution can still give inaccurate results if it is miscalibrated.
Uncertainty: Absolute, fractional and percentage uncertainties represent uncertainty in the final answer for a quantity.
Every measurement carries doubt, and this statement requires you to express that doubt in three linked ways. Suppose a metre rule reads a length as 0.842 m with a reading uncertainty of ±0.001 m. The absolute uncertainty is the ±0.001 m itself, carrying the same unit as the quantity. The fractional uncertainty is the absolute uncertainty divided by the measured value: 0.001 ÷ 0.842 ≈ 0.00119. The percentage uncertainty is that fraction multiplied by 100, giving about 0.119%. All three describe the same doubt, so converting between them changes only the form, never the size. The final answer for a quantity should therefore be quoted with its uncertainty, and the form chosen must match what the question asks. A small absolute uncertainty on a large value gives a tiny percentage uncertainty, which is why percentage form is useful for comparing precision across different quantities.
Combination of absolute and percentage uncertainties.
When a quantity is calculated from several measurements, their uncertainties must be combined. For addition and subtraction, add the absolute uncertainties. If a rod is made of two lengths, 0.842 m ± 0.001 m and 0.315 m ± 0.001 m, the total is 1.157 m ± 0.002 m. For multiplication and division, add the percentage uncertainties instead. If a rectangle has length 0.842 m ± 0.119% and width 0.315 m ± 0.317%, the area has a percentage uncertainty of about 0.436%. Raising a quantity to a power multiplies its percentage uncertainty by that power. Choose the rule that matches the arithmetic, convert between forms as needed, and state the final absolute uncertainty to one significant figure.
Represent uncertainty in a data point on a graph using error bars.
Error bars show the uncertainty attached to each plotted value. For a point (x, y), the absolute uncertainty in the dependent variable gives a vertical bar from y − Δy to y + Δy; uncertainty in the independent variable gives a horizontal bar from x − Δx to x + Δx. Draw bars with a ruler, centred on the plotted point, with caps if you wish. For example, a current of 0.50 A with uncertainty ±0.02 A is plotted as a vertical bar from 0.48 A to 0.52 A. If only one variable has a stated uncertainty, draw only that bar. Error bars let you judge whether a line of best fit is consistent with the data and whether a point is anomalous.
Determine the uncertainties in the gradient and intercept of a straight-line graph.
The gradient and intercept of a straight-line graph are found from a line of best fit, but that line is not unique. Draw the steepest and shallowest lines that remain consistent with the error bars, then calculate the gradient of each. Half the difference between the steepest and shallowest gradients gives the uncertainty in the gradient. Repeat for the intercept: read the intercept of each extreme line and halve the difference. For example, if gradients are 2.4 and 2.0, the best gradient is 2.2 ± 0.2 in the stated units. This method links graphical work to the uncertainty in a derived quantity.
Individual points on the graph may or may not have associated error bars.
When plotting experimental data, each point represents a measured value. An error bar is drawn through the point to show the absolute uncertainty. Individual points on a graph may or may not have associated error bars. Crucially, if a point lacks an error bar, it is typically because the absolute uncertainty is smaller than the size of the plotted point itself, not because uncertainty is zero. For example, a highly precise digital measurement might yield an uncertainty too small to draw. When evaluating a line of best fit, you must check if it passes through the drawn error bars. The size of the error bar directly reflects the precision of the measurement technique used for that specific data point.
Your focus
- Define random and systematic errors and identify the cause of each in a given measurement context.
- Explain how repeating readings and averaging affects random error and why it does not remove systematic error.
- Select and justify an appropriate correction or improvement for a stated random or systematic error.
Show all 21 objectives
- Define precision, repeatability, reproducibility, resolution and accuracy and use each term correctly in context.
- Compare sets of measurements in terms of precision and accuracy and explain how each is affected by random and systematic effects.
- Justify the choice of instrument for a measurement by referring to its resolution and the required accuracy.
- Convert a stated absolute uncertainty into fractional and percentage forms correctly.
- Explain that absolute, fractional and percentage uncertainties describe the same doubt in different ways.
- Quote a final answer for a quantity with an uncertainty in the form requested.
- Select the correct rule for combining uncertainties in addition, subtraction, multiplication, division and powers.
- Convert between absolute and percentage uncertainties during a multi-step calculation.
- Present a calculated quantity with its absolute uncertainty rounded to the correct decimal place.
- Draw a correctly centred error bar of the right length for a stated absolute uncertainty.
- Convert a percentage uncertainty into an absolute uncertainty before plotting.
- Interpret error bars to judge whether a line of best fit is consistent with the data.
- Draw steepest and shallowest lines consistent with error bars.
- Calculate gradient and intercept uncertainties by halving the difference between extreme values.
- Report a gradient or intercept with its uncertainty, units and appropriate significant figures.
- Identify error bars on a graph and state what they represent.
- Explain why some points on a graph may lack error bars due to the scale of the plot.
- Use error bars to evaluate the acceptability of a line of best fit.
Which command word is this?
Give the precise meaning of a term, often with a formula or example. For example, 'Define resolution' expects 'the smallest change in a quantity that an instrument can detect'.
Limitation of physical measurements exam tips
Topic Overview
This topic explores the fundamental limitations inherent in all physical measurements, including the concepts of resolution, precision, accuracy, and the distinction between random and systematic errors. It equips students with the skills to quantify uncertainty, propagate errors through calculations, and critically evaluate experimental data, which is essential for practical work and exam questions.
Understanding measurement limitations is crucial for interpreting experimental results in all areas of physics, from mechanics to quantum physics. It forms the basis for assessing the reliability of conclusions and is a key component of the AQA A-Level Physics practical endorsement and written examinations.
Key Concepts
- →Random errors cause unpredictable variation in measurements and can be reduced by taking multiple readings and calculating a mean, while systematic errors cause a consistent shift and cannot be reduced by averaging.
- →Precision refers to the closeness of repeated measurements to each other, while accuracy refers to how close a measurement is to the true value. A measurement can be precise but not accurate.
- →Resolution is the smallest change in a quantity that an instrument can detect, and it contributes to the uncertainty of a measurement.
- →Absolute uncertainty is the margin of error in a measurement, while percentage uncertainty expresses this as a fraction of the measured value.
- →When combining measurements, absolute uncertainties are added for addition/subtraction, and percentage uncertainties are added for multiplication/division or powers.
Marking Points
- Defines random errors as unpredictable variations in repeated measurements about a mean value, caused by factors such as fluctuating temperature, vibration or inconsistent reading of a scale.
- Explains that random errors are reduced by repeating measurements and calculating a mean, because deviations in opposite directions partially cancel.
- Defines systematic errors as consistent shifts of all readings from the true value, caused by apparatus faults such as a zero error, miscalibration or a flawed technique like parallax.
- Explains that systematic errors are not reduced by averaging and require identification and correction, for example re-zeroing a balance or applying a calibration correction.
- Distinguishes the two by their effect on a set of repeats: random errors produce scatter, whereas systematic errors displace the whole set in one direction.
- Applies the distinction to a practical context, such as timing oscillations with a stopwatch (random reaction-time scatter) versus a stopwatch running slow (systematic timing shift).
- Defines precision as the closeness of agreement among repeated measurements of the same quantity, indicated by a small spread of values.
- Defines repeatability as agreement between repeats made by the same person, with the same method and equipment, over a short interval.
- Defines reproducibility as agreement between measurements made by a different person, or with different equipment, method or conditions.
- Defines resolution as the smallest change in the measured quantity that produces a detectable change in the instrument reading.
- Defines accuracy as the closeness of a measured value to the true value, and explains that precision and accuracy are independent qualities.
- Uses a context to separate the terms, for example a miscalibrated digital balance giving closely agreeing but inaccurate masses, or a metre rule whose 1 mm resolution limits the precision of a length.
- Absolute uncertainty is expressed in the same unit as the measured quantity, for example ±0.001 m for a length of 0.842 m.
- Fractional uncertainty equals absolute uncertainty divided by the measured value, so 0.001 ÷ 0.842 ≈ 0.00119.
- Percentage uncertainty equals fractional uncertainty multiplied by 100, giving about 0.119% in this example.
- The three forms represent the same doubt, so converting between them must not change the underlying size of the uncertainty.
- A final answer for a quantity should state its uncertainty, and the form used should match the form requested.
- Percentage uncertainty allows fair comparison of precision between quantities measured in different units.
- For addition and subtraction, combine the absolute uncertainties by adding them, keeping the unit of the quantity.
- For multiplication and division, combine the percentage uncertainties by adding them.
- For a quantity raised to a power, multiply its percentage uncertainty by that power.
- Convert between absolute and percentage forms when the calculation mixes operations.
- State the final combined uncertainty to one or two significant figures, matching the decimal places of the calculated value.
- Identify which variable carries the stated uncertainty and whether it is the independent or dependent variable.
- Calculate the upper and lower limits by adding and subtracting the absolute uncertainty from the plotted value.
- Draw the bar centred on the point, of correct total length 2Δy or 2Δx, using a sharp pencil and ruler.
- Add caps or clear end markers so the extent of the uncertainty is unambiguous.
- If both variables have uncertainties, draw both vertical and horizontal bars to form a cross.
- Keep the scale consistent so the bar length genuinely represents the uncertainty.
- Draw a line of best fit through the data, taking account of error bars.
- Draw a steepest and a shallowest acceptable line that still pass through the error bars.
- Calculate the gradient of the best, steepest and shallowest lines using well-separated points on each line.
- Find the uncertainty as half the difference between the steepest and shallowest values.
- Read the intercept of each extreme line and halve the difference to find the intercept uncertainty.
- State the final result with correct units and a sensible number of significant figures.
- State that an error bar represents the absolute uncertainty associated with a plotted measurement.
- Recognise that error bars may be present on some points and absent on others within the same graph.
- Explain that a missing error bar indicates the absolute uncertainty is smaller than the plotted point, not that uncertainty is zero.
- Use error bars to judge whether a straight line of best fit is consistent with the data.
- Describe how the size of an error bar reflects the precision of the measuring instrument or technique.
Examiner Tips
- 💡When asked to classify an error, state whether it shifts all readings consistently (systematic) or scatters them unpredictably (random), then justify with the cause.
- 💡Link each error type to its remedy: repeats and means for random errors, calibration checks and zeroing for systematic errors.
- 💡Use a practical example from your own required practicals, such as a zero error on a micrometer or reaction time when using a stopwatch, to make the distinction concrete.
- 💡Define each term in one sentence before applying it, so the distinction is explicit in your answer.
- 💡When comparing two data sets, comment separately on their spread (precision) and their closeness to the true value (accuracy).
- 💡Quote the resolution of an instrument with its unit, for example 0.01 mm for a digital micrometer, and link it to the uncertainty in a single reading.
- 💡Write the conversion chain explicitly: absolute, then divide by the value, then multiply by 100, so the examiner can follow your reasoning.
- 💡Check the size of your percentage: a precise measurement usually gives a small percentage uncertainty, so a value above 100% signals an error.
- 💡When a question asks for the uncertainty in the final answer, decide first whether it wants an absolute, fractional or percentage form before calculating.
- 💡Identify the arithmetic first: addition and subtraction use absolute uncertainties, while multiplication, division and powers use percentage uncertainties.
- 💡Round the final absolute uncertainty to one significant figure and ensure the calculated value is given to the same number of decimal places.
- 💡Read the question carefully to see whether uncertainty is given for x, y or both before drawing.
- 💡Use a ruler and label the bar or state its value so the examiner can see the intended uncertainty.
- 💡Check that the bar is symmetrical about the plotted point; asymmetry suggests an arithmetic slip.
- 💡Show the coordinates you used for each gradient calculation so your method is clear.
- 💡Keep units consistent and quote the gradient and intercept with their uncertainties and units.
- 💡If error bars are small, say so and explain that the uncertainty is correspondingly small.
- 💡When asked to comment on a graph, explicitly state whether error bars are present and what that means for judging the fit.
- 💡If a point lacks an error bar, explicitly state that the uncertainty is smaller than the plotted point to secure marks.
- 💡Always specify the type of error (random or systematic) and link it to a specific improvement, such as 'repeat and average' for random errors or 'calibrate' for systematic errors.
- 💡When calculating uncertainties, show your working clearly, including whether you are adding absolute or percentage uncertainties, and round the final uncertainty to one significant figure.
- 💡In practical questions, always consider the resolution of the instruments used and how it affects the overall uncertainty in the final result.
Common Mistakes
- Believing that taking more readings removes systematic error: averaging only reduces random scatter, so a zero error or miscalibration must be found and corrected instead.
- Calling any large error random: a consistent offset, such as all readings being 0.5 cm too high, is systematic and needs a correction rather than more repeats.
- Treating a mistake or misreading as an error type: a one-off blunder is neither random nor systematic in the intended sense and should be discarded and the measurement repeated.
- Using precision and accuracy interchangeably: precision concerns agreement between repeats, whereas accuracy concerns closeness to the true value, so a precise set can be inaccurate.
- Confusing resolution with accuracy: a fine-resolution instrument can still be inaccurate if it is miscalibrated, because resolution only describes the smallest detectable change.
- Treating repeatability and reproducibility as identical: repeatability involves the same person, method and equipment, while reproducibility involves a change of person, equipment, method or conditions.
- Dividing the measured value by the absolute uncertainty instead of dividing the absolute uncertainty by the measured value; correct this by keeping the uncertainty on top of the fraction.
- Forgetting to multiply the fractional uncertainty by 100 when a percentage is requested; correct this by checking whether the answer is sensibly below 100% for a precise measurement.
- Attaching a unit to a fractional or percentage uncertainty; correct this by leaving fractional and percentage forms unitless while the absolute form keeps the unit.
- Adding percentage uncertainties when quantities are added or subtracted; correct this by adding absolute uncertainties for those operations.
- Adding absolute uncertainties when quantities are multiplied or divided; correct this by converting to percentage uncertainties and adding those.
- Giving the absolute uncertainty to too many significant figures; correct this by rounding the uncertainty to one (or two) significant figures and matching the value's decimal places.
- Drawing the bar from the point to one limit only, so the length represents Δy rather than 2Δy; correct this by centring the bar on the point.
- Using the percentage uncertainty as if it were an absolute uncertainty; correct this by converting to an absolute value before plotting.
- Drawing bars on every point when only one variable has a stated uncertainty; correct this by adding bars only where uncertainty is known.
- Using the difference between the steepest and shallowest values directly as the uncertainty instead of halving it; correct this by dividing by two.
- Choosing extreme lines that miss the error bars entirely; correct this by ensuring each extreme line still passes through the bars.
- Reading gradient points too close together, which magnifies reading error; correct this by using points near opposite ends of the line.
- Assuming every point must have an error bar; correction: error bars are omitted if the uncertainty is smaller than the plotted point.
- Treating a point without an error bar as exact; correction: absence of an error bar does not imply zero uncertainty, only that it is too small to display.
- Drawing error bars of equal length on all points regardless of the measurement method; correction: error bar size should reflect the actual uncertainty of each measurement.
- Students often think that taking more readings reduces systematic errors. Correction: Systematic errors are consistent and cannot be reduced by averaging; they require calibration or correction of the instrument.
- Students sometimes believe that a more precise instrument always gives a more accurate result. Correction: Precision and accuracy are independent; a precise instrument may still have a systematic error, making it inaccurate.
- Students may confuse the resolution of an instrument with its uncertainty. Correction: Resolution is the smallest detectable change, while uncertainty is the range within which the true value lies, often taken as half the resolution for analogue instruments.
Revision Plan
- 1Day 1-2: Learn definitions of key terms: resolution, precision, accuracy, random error, systematic error. Create flashcards for each.
- 2Day 3-4: Practice calculating absolute and percentage uncertainties for single measurements and simple combinations.
- 3Day 5-6: Work through past paper questions on uncertainty propagation, focusing on combining uncertainties in calculations.
- 4Day 7-8: Review practical experiments you have done and identify sources of random and systematic errors, and suggest improvements.
- 5Day 9-10: Complete a full exam-style question on measurement limitations and mark it using the mark scheme to identify gaps.
Exam Question Types
- 📋Definition and explanation questions: e.g., 'Distinguish between random and systematic errors and give an example of each.' Advice: Use precise terminology and link to experimental context.
- 📋Calculation questions: e.g., 'Calculate the percentage uncertainty in the density of a material given measurements of mass and volume.' Advice: Show all steps, including uncertainty in each quantity, and combine correctly.
- 📋Practical-based questions: e.g., 'Describe how you would reduce the uncertainty in measuring the period of a pendulum.' Advice: Suggest specific techniques like timing multiple oscillations and using a fiducial marker.
- 📋Data analysis questions: e.g., 'Analyse the graph and determine the uncertainty in the gradient.' Advice: Use error bars and lines of best/worst fit to estimate uncertainty.
Command Word Expectations (AQA)
Give the precise meaning of a term, often with a formula or example. For example, 'Define resolution' expects 'the smallest change in a quantity that an instrument can detect'.
Provide a detailed account of why something happens, including a cause-and-effect link. For example, 'Explain why repeating measurements reduces uncertainty' expects 'random errors cancel out when averaged, reducing the spread of results'.
Use given data and formulas to find a numerical answer, showing all working and units. For uncertainty calculations, expect to combine absolute or percentage uncertainties correctly.
How Students Lose Marks (Examiner Pitfalls)
Step-by-Step Worked Solutions
Question: A student measures the length of a pendulum as 1.245 m using a metre ruler with a resolution of 1 mm. The time for 10 oscillations is measured as 18.4 s using a stopwatch with a resolution of 0.01 s. Calculate the percentage uncertainty in the period T and hence the uncertainty in the calculated value of g using the formula g = 4π²L/T². Take π = 3.1416.
- 1.Step 1: Identify given facts: L = 1.245 m, resolution of ruler = 0.001 m, time for 10 oscillations = 18.4 s, resolution of stopwatch = 0.01 s. Period T = 18.4 / 10 = 1.84 s.
- 2.Step 2: Calculate absolute uncertainty in L: ±0.001 m (half resolution is sometimes used, but AQA often uses resolution as uncertainty). Percentage uncertainty in L = (0.001 / 1.245) × 100% = 0.0803%.
- 3.Step 3: Calculate absolute uncertainty in T: The time measurement has uncertainty ±0.01 s for the total time, so for one period, uncertainty is ±0.001 s. Percentage uncertainty in T = (0.001 / 1.84) × 100% = 0.0543%. Since T is squared, percentage uncertainty in T² = 2 × 0.0543% = 0.1086%.
- 4.Step 4: Calculate g: g = 4π²L/T² = 4 × (3.1416)² × 1.245 / (1.84)² = 4 × 9.8696 × 1.245 / 3.3856 = 49.14 / 3.3856 = 14.52 m/s² (this is unrealistic but used for calculation).
- 5.Step 5: Total percentage uncertainty in g = percentage uncertainty in L + percentage uncertainty in T² = 0.0803% + 0.1086% = 0.1889%. Absolute uncertainty in g = (0.1889 / 100) × 14.52 = 0.0274 m/s². Round to 0.03 m/s². So g = 14.52 ± 0.03 m/s².
Question: A student measures the diameter of a wire using a micrometer with a resolution of 0.01 mm. The readings are 0.52 mm, 0.53 mm, 0.51 mm, 0.52 mm, 0.54 mm. Calculate the mean diameter, the random uncertainty, and the percentage uncertainty. Also state one systematic error that could affect the measurement and how it could be reduced.
- 1.Step 1: Calculate mean diameter: (0.52 + 0.53 + 0.51 + 0.52 + 0.54) / 5 = 2.62 / 5 = 0.524 mm.
- 2.Step 2: Calculate random uncertainty: The range is 0.54 - 0.51 = 0.03 mm. The uncertainty is half the range = ±0.015 mm. So mean = 0.524 ± 0.015 mm.
- 3.Step 3: Calculate percentage uncertainty: (0.015 / 0.524) × 100% = 2.86%.
- 4.Step 4: Identify systematic error: A zero error on the micrometer, where the reading is not zero when the jaws are closed. This shifts all readings by a constant amount. It can be reduced by checking for zero error and subtracting it from all readings.