Propagation of uncertainty lets you report a calculated physics result with an honest range of values. This guide gives you a reusable workflow, reference formulas, and worked examples for density, pendulum measurements, resistance, and other laboratory quantities.
Overview
A measured quantity is never perfectly exact. A ruler has a smallest scale division, a balance has limited resolution, and repeated readings vary. If you use one or more measurements in a calculation, the uncertainty in the result must also be calculated.
Write a measurement as x ± δx, where x is the best estimate and δx is its absolute uncertainty. For example, a length might be recorded as 12.4 ± 0.1 cm. The relative uncertainty is δx / x, and the percentage uncertainty is:
Percentage uncertainty = (δx / x) × 100%
There are two common approaches. The first uses simple rules that are appropriate for many introductory laboratory exercises. The second uses a general derivative formula for independent random uncertainties. Always check the convention specified by your laboratory manual or instructor, because courses may use slightly different rounding or instrument-resolution rules.
Reference table for common calculations
| Result | Useful uncertainty rule |
|---|---|
| q = a ± b | For a conservative maximum estimate, add absolute uncertainties: δq = δa + δb. |
| q = ab or q = a/b | Add relative uncertainties: δq/q = δa/a + δb/b. |
| q = an | Multiply relative uncertainty by the magnitude of the power: δq/q = |n|δa/a. |
| q = f(a,b,...) | For independent random uncertainties, use δq = √[(∂q/∂a · δa)2 + (∂q/∂b · δb)2 + ...]. |
The simple addition rules are often called worst-case or maximum uncertainty propagation. For independent random uncertainties, the root-sum-square method is usually less conservative. Do not combine the methods casually: identify which convention your course expects before completing the calculation.
Checklist by scenario
1. Addition and subtraction
When quantities are added or subtracted, combine their absolute uncertainties. Suppose a distance is calculated from two readings:
d = (8.2 ± 0.1) cm + (3.6 ± 0.1) cm
The result is d = 11.8 cm, and the maximum absolute uncertainty is 0.1 cm + 0.1 cm = 0.2 cm. Report:
d = (11.8 ± 0.2) cm
Using a root-sum-square estimate for independent uncertainties would give √[(0.1)2 + (0.1)2] = 0.14 cm, which might be rounded to 0.1 cm or 0.14 cm depending on the course convention.
2. Multiplication and division: density example
Density is calculated from ρ = m/V. Assume a measured mass and volume are:
m = (50.0 ± 0.5) g
V = (20.0 ± 0.5) cm3
The best estimate is ρ = 50.0/20.0 = 2.50 g/cm3. For the maximum-uncertainty method, calculate relative uncertainties:
δm/m = 0.5/50.0 = 0.010
δV/V = 0.5/20.0 = 0.025
Add them to obtain the relative density uncertainty:
δρ/ρ = 0.010 + 0.025 = 0.035
Thus, δρ = 2.50 × 0.035 = 0.0875 g/cm3. With appropriate rounding, report ρ = (2.50 ± 0.09) g/cm3. The percentage uncertainty is about 3.5%.
3. Powers: pendulum measurements
For a simple pendulum, the period relation can be rearranged to estimate gravitational acceleration:
g = 4π2L/T2
Because length is to the first power and period is squared, the relative uncertainty is:
δg/g = δL/L + 2δT/T
Suppose L = (0.800 ± 0.002) m and T = (1.79 ± 0.01) s. The relative contributions are:
δL/L = 0.002/0.800 = 0.0025
2δT/T = 2(0.01/1.79) ≈ 0.0112
The period contributes more strongly because it is squared. The estimated value is approximately g = 9.86 m/s2, with relative uncertainty about 0.0137. This gives an absolute uncertainty of approximately 0.14 m/s2, so a suitable result is g = (9.86 ± 0.14) m/s2 under the maximum-uncertainty convention.
4. Resistance from voltage and current
Ohm’s law gives R = V/I. If V = (5.00 ± 0.05) V and I = (0.250 ± 0.005) A, then:
R = 5.00/0.250 = 20.0 Ω
The relative uncertainties are 0.05/5.00 = 0.010 and 0.005/0.250 = 0.020. Therefore, the maximum relative uncertainty is 0.030, and the absolute uncertainty is 20.0 × 0.030 = 0.6 Ω. Report R = (20.0 ± 0.6) Ω. This result also shows why carrying units through every step matters: volts divided by amperes produces ohms.
5. General formula for independent measurements
For a result q = f(x, y, ...), independent uncertainties can be propagated using partial derivatives:
(δq)2 = (∂q/∂x · δx)2 + (∂q/∂y · δy)2 + ...
For density, ρ = m/V, this becomes:
(δρ)2 = (δm/V)2 + (mδV/V2)2
This method assumes the input uncertainties are independent and represent random uncertainty. If measurements are correlated, or if a dominant systematic error is present, the simple formula may not describe the full uncertainty.
What to double-check
- Define every input. Record the measured value, absolute uncertainty, unit, and source of the uncertainty before substituting into an equation.
- Distinguish precision from accuracy. Repeated readings can reveal random variation, but they do not automatically remove a calibration error or other systematic effect.
- Use consistent units. Convert centimetres to metres, milliamperes to amperes, or other units before calculating a derived quantity.
- Choose the correct rule. Addition and subtraction use absolute uncertainties; products, quotients, and powers use relative uncertainties.
- Check sensitivity. A variable raised to a power contributes that power’s magnitude to the relative uncertainty. In the pendulum example, period uncertainty matters twice as much because of T2.
- Keep extra digits during calculation. Round only after the uncertainty has been calculated, then round the central value to the same decimal place.
- Check dimensional consistency. The units of the final value should match the physical quantity you intended to calculate.
Common mistakes
One frequent mistake is adding percentage uncertainties when the formula requires absolute uncertainties. For example, lengths in a subtraction must be combined in centimetres or metres, not by adding their percentages. The reverse mistake occurs with multiplication and division: adding absolute uncertainties directly produces the wrong scale.
Another error is reporting too many digits, such as 2.50000 ± 0.08753 g/cm3 when the measurements do not support that precision. A practical convention is to keep one significant figure in the uncertainty, or two when the first digit is small, and round the measured result to the same decimal place. Follow your laboratory’s stated rule if it differs.
Do not treat uncertainty as the same thing as the difference between an experimental result and a reference value. That difference may indicate systematic error, a model limitation, or an experimental mistake. Uncertainty propagation describes how input uncertainty affects a calculated result; it does not guarantee that every source of error has been included.
Finally, do not hide the calculation. A lab report should show the equation, substituted values, uncertainty rule, and final rounded result so another reader can evaluate the method.
When to revisit
Return to this checklist whenever a new laboratory experiment uses a derived quantity, when an instrument or measurement method changes, or when your course manual introduces a different uncertainty convention. Recalculate the result if you replace a measured input, revise an instrument resolution, average repeated trials differently, or discover that units were inconsistent.
Before submitting a lab report, use this final action list:
- Write each input as value plus or minus uncertainty.
- Convert all inputs to compatible units.
- Identify whether the calculation uses sums, differences, products, quotients, or powers.
- Apply the required propagation rule and show the intermediate relative or absolute uncertainties.
- Consider whether random and systematic effects are being treated separately.
- Round the uncertainty and central value consistently.
- Report the final result with units and, when useful, the percentage uncertainty.
Keeping this workflow with your physics lecture notes or lab template makes uncertainty calculations faster without turning them into a black box. For related calculation checks, see the guide to physics units and significant figures.