Reliable physics results depend not only on the measured value, but also on a clear estimate of how uncertain that value is. This guide shows how to choose an appropriate uncertainty for common instruments, combine uncertainties in calculations, round results sensibly, and write conclusions that communicate what your experiment actually established.
Overview
Every physical measurement is an estimate. If a ruler reading is recorded as 12.4 cm, the object is not known to have exactly that length; the reading represents a range of plausible values determined by the ruler, the observer, and the measuring method. Measurement uncertainty expresses that range in a compact, useful form.
Uncertainty is not the same as a mistake. A mistake might be reading the wrong scale, recording a value incorrectly, or using an unsuitable instrument. Those errors should be corrected where possible. Uncertainty describes the limitation that remains even after the procedure has been carried out carefully.
A standard result has the form:
measured value ± absolute uncertainty
For example, 24.6 ± 0.1 cm states both the best estimate and the approximate range associated with the measurement. The percentage uncertainty makes the size of that limitation easier to compare:
percentage uncertainty = (absolute uncertainty ÷ measured value) × 100%
Use the uncertainty convention required by your course or laboratory instructions. The rules below are common classroom methods, but some specifications use instrument accuracy, calibration data, standard deviation, or a defined confidence interval instead.
How to estimate uncertainty
1. Identify the measuring instrument and its scale
Start with the smallest marked division, often called the resolution. For an analog scale, a common school-laboratory estimate is half the smallest division when the reading is taken carefully. Thus, a ruler marked in millimeters may be recorded as having an uncertainty of about ±0.5 mm for a single reading. Some courses instead use ±1 smallest division, so check the stated convention.
For a digital instrument, the displayed resolution is often used as a starting point. A display reading of 2.37 V might therefore have a resolution-based uncertainty of about ±0.01 V. However, a manufacturer’s stated accuracy, calibration uncertainty, or range setting may be larger than the last displayed digit. If that information is available, use the more appropriate specification rather than relying on resolution alone.
2. Decide whether repeat measurements are useful
Repeat measurements help reveal random variation, such as small timing differences or fluctuating sensor readings. Measure the same quantity several times without changing the intended conditions, then calculate the mean:
mean = sum of measurements ÷ number of measurements
For a simple school experiment, one practical estimate of random uncertainty is half the range:
half-range uncertainty = (largest value − smallest value) ÷ 2
For example, stopwatch readings of 4.8 s, 5.0 s, and 4.9 s have a mean of 4.9 s and a half-range uncertainty of 0.1 s. You should still consider the stopwatch resolution and reaction time. Do not report an uncertainty smaller than the limitations of the method without a clear reason.
3. Convert to percentage uncertainty
Once you have an absolute uncertainty, divide it by the measured value and multiply by 100%. If a mass is measured as 250.0 ± 0.5 g, its percentage uncertainty is:
(0.5 ÷ 250.0) × 100% = 0.2%
Percentage uncertainty is especially helpful when comparing measurements or deciding whether an instrument is suitable. Measuring a 5 mm thickness with a ruler may produce a much larger percentage uncertainty than measuring a 500 mm length with the same ruler.
4. Propagate uncertainty through a calculation
When measured quantities are combined, their uncertainties affect the calculated result. Common introductory rules are:
- Addition or subtraction: add the absolute uncertainties. If x = 12.0 ± 0.2 and y = 5.0 ± 0.1, then x + y has an uncertainty of about ±0.3.
- Multiplication or division: add the percentage uncertainties. If a = 4.0 ± 2% and b = 3.0 ± 1%, then a/b has an uncertainty of about ±3%.
- Powers: multiply the percentage uncertainty by the power. If a quantity is squared, its percentage uncertainty is approximately doubled.
These are deliberately conservative rules often used in introductory physics. More advanced courses may use root-sum-square methods or statistical uncertainty propagation, particularly when uncertainties are independent and random.
Inputs and assumptions
Your uncertainty estimate is only as good as the information behind it. Record the following before calculating a final result:
| Input | What to record | Why it matters |
|---|---|---|
| Instrument | Type, range, and scale | Different instruments have different resolutions and accuracy limits. |
| Resolution | Smallest division or displayed increment | Sets a practical limit on a single reading. |
| Measured values | Every repeat, not only the mean | Shows spread and makes random variation visible. |
| Method | Where the scale was read and how timing was performed | Parallax, reaction time, and alignment can affect results. |
| Calibration or specification | Any stated accuracy or correction | Instrument accuracy may exceed the limitation suggested by resolution. |
| Model | Equation used to calculate the result | Determines whether absolute or percentage uncertainties are combined. |
Common instrument examples include a ruler or meter stick for length, a balance for mass, a stopwatch or light gate for time, a multimeter for voltage and current, and a temperature or motion sensor for continuously changing data. A sensor’s software display may show many decimal places without making the physical measurement equally precise. Treat displayed digits as information about the interface, not automatic proof of accuracy.
Keep units consistent before calculating. Converting units first reduces avoidable mistakes; the guide to physics units and conversions can help with prefixes and dimensional checks. Also distinguish uncertainty from a signed correction: an instrument that consistently reads high may require a calibration correction, while uncertainty describes what remains unknown.
Worked examples
Example 1: measuring a length with a ruler
A block is measured as 12.4 cm using a ruler with 1 mm divisions. Using a half-division estimate gives an uncertainty of ±0.05 cm. The percentage uncertainty is:
(0.05 ÷ 12.4) × 100% ≈ 0.4%
A suitable result statement is length = (12.40 ± 0.05) cm, provided the measuring method supports recording to that decimal place. If your class uses one smallest division instead, report the uncertainty using that convention and keep the value rounded consistently.
Example 2: calculating density
A sample has mass 50.0 ± 0.5 g and volume 20.0 ± 0.5 cm3. Density is mass divided by volume:
ρ = 50.0 ÷ 20.0 = 2.50 g/cm3
The mass percentage uncertainty is 1.0%, and the volume percentage uncertainty is 2.5%. Using the introductory division rule, the density percentage uncertainty is approximately 3.5%. The absolute density uncertainty is:
2.50 × 0.035 = 0.0875 g/cm3
Rounded sensibly, the result is density = (2.50 ± 0.09) g/cm3. The uncertainty is reported to one or two significant figures, and the measured value is rounded to the same decimal place.
Example 3: timing a repeated motion
Suppose five timing readings are 2.1 s, 2.0 s, 2.2 s, 2.1 s, and 2.0 s. The mean is 2.08 s. The range is 0.2 s, so the half-range estimate is 0.1 s. A practical statement might be time = (2.1 ± 0.1) s. If the measurement was made with a handheld stopwatch, reaction time may be important; repeating the trial or timing several cycles and dividing by the number of cycles can often reduce the relative effect of starting and stopping delays.
For graph-based experiments, uncertainty can also be shown with error bars. Explain whether each bar represents instrument uncertainty, half-range, or another defined estimate. When analyzing a graph, use the same data treatment throughout; the guide to reading physics graphs covers slopes, areas, and motion data.
When to recalculate
Recalculate the uncertainty whenever an input or assumption changes. This includes changing the instrument, selecting a different measurement range, adding or removing repeat trials, correcting a unit conversion, or replacing a single reading with a mean. Recheck the result if your calculation changes from addition to multiplication, or if you use a new equation with powers or several measured quantities.
Before submitting a lab report, use this short checklist:
- Write down the instrument resolution and any available accuracy specification.
- Record all raw readings and calculate the mean when repeats are appropriate.
- Choose an uncertainty method and state it briefly.
- Calculate absolute and percentage uncertainty with consistent units.
- Propagate uncertainty using the rule that matches the equation.
- Round the uncertainty first, then round the measured value to the same decimal place.
- Report the final result with units and explain the largest limitation.
Finally, connect the uncertainty to the experiment’s purpose. If the uncertainty is large enough to overlap two competing predictions, the data may not distinguish between them. That is a useful conclusion, not a failed experiment: it tells you whether a finer-scale instrument, more repeats, a longer measurement interval, or a better-controlled procedure is needed next time. For broader lab planning, see the physics lab report guide and the collection of physics experiment videos and lab-prep guidance.