How to Calculate Percentage Uncertainty (With Worked Examples)

Introduction

Two students hand in lab reports on the same day.

The first measured a wooden block with a ruler and got 24.30 ± 0.05 cm. The second timed a trolley with a stopwatch and got 1.20 ± 0.05 s.

Same uncertainty. Both ± 0.05. So both measurements are equally good, right?

Not even close. The first student’s doubt is about a fifth of a percent of what they measured. The second student’s doubt is more than four percent. One of those experiments is solid. The other is shaky, and the ± sign alone doesn’t show it.

That’s the whole reason percentage uncertainty exists. Absolute uncertainty tells you how big your doubt is. Percentage uncertainty tells you whether that doubt actually matters.

It’s also one of the easiest calculations you’ll ever do in a lab a single division and a multiplication. The tricky part isn’t the arithmetic. It’s knowing when to use percentages and when to leave them alone, which is where most of the lost marks come from.

What Percentage Uncertainty Actually Tells You

Percentage uncertainty expresses your doubt as a fraction of what you measured, rather than in units.

Go back to those two students. Both had ± 0.05, but:

• 24.30 ± 0.05 cm → the doubt is 0.21% of the measurement
• 1.20 ± 0.05 s → the doubt is 4.2% of the measurement

The stopwatch measurement is roughly twenty times worse in relative terms, even though the ± number looks identical. Once you convert to percentages, you can compare measurements of completely different things a length against a time, a mass against a voltage and immediately see which one is dragging your experiment down.

This is genuinely useful, not just an exam requirement. When your final answer comes out with a huge uncertainty and you need to know why, percentages are how you find the culprit.

The three forms you’ll meet

Students often think absolute, relative and percentage uncertainty are three different things to learn. They’re not. They’re the same doubt written three ways:

Relative and percentage differ only by a factor of 100. Some syllabi say “relative uncertainty” when they mean the percentage version, so read the question carefully rather than assuming.

The Percentage Uncertainty Formula

Here it is:

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

Or in symbols, where Δx is the absolute uncertainty and x is your reading:

% uncertainty = (Δx / x) × 100

That’s it. The only genuine difficulty is that you need Δx first, and where that comes from depends on how you measured.

If you took one reading from a scale, the absolute uncertainty is usually half the smallest division. A millimetre ruler gives ± 0.5 mm; a thermometer marked in whole degrees gives ± 0.5 °C.

If you took several readings, use half the range that is, (largest − smallest) ÷ 2 and quote the mean as your value. Some courses want the standard error s/√n instead, which gives a smaller figure. Check which one yours expects.

If the value came off a calibration certificate, divide the stated uncertainty by the coverage factor k before you do anything else.

If a component is already labelled with a percentage like a resistor marked ± 5% then you’ve been handed the answer and you can work backwards to absolute if you need it.

One thing worth noticing: the measured value sits in the denominator. So for a fixed instrument, measuring something bigger always gives you a smaller percentage uncertainty. That single fact quietly explains a lot of experimental design, and we’ll come back to it.

Worked Examples

A book measured with a ruler

You measure a book as 24.30 cm using a metre rule with 1 mm divisions, so the absolute uncertainty is ± 0.05 cm.

(0.05 ÷ 24.30) × 100 = 0.21%

Under a fifth of a percent. That’s a good measurement.

A mass on a digital balance

The balance reads 12.47 g and its resolution is 0.01 g, giving ± 0.005 g.

(0.005 ÷ 12.47) × 100 = 0.040%

Digital balances usually produce tiny percentage uncertainties, which is exactly why mass is so rarely the weak link in a school experiment.

A time measured by hand

You time a trolley over a track and get 1.20 s. Human reaction time makes ± 0.05 s a fair estimate.

(0.05 ÷ 1.20) × 100 = 4.2%

Twenty times worse than the ruler. If this trolley experiment gives a disappointing final answer, you now know where to look first.

The same instrument, two very different results

This is the example worth remembering. Suppose you’re working with something that gives ± 1 mm every time.

Measure a doorframe at 2000 mm:

(1 ÷ 2000) × 100 = 0.05%

Measure a screw at 5 mm:

(1 ÷ 5) × 100 = 20%

Identical instrument, identical absolute uncertainty, and one result is four hundred times worse than the other. The instrument didn’t change the size of what you measured did.

This is why physics labs tell you to time twenty pendulum swings rather than one, and why you stack twenty coins before measuring their thickness. You’re not making the instrument better. You’re making the measured value bigger, and letting the denominator do the work.

Working backwards from a percentage

A resistor is marked 470 Ω ± 5%. What’s the absolute uncertainty?

(5 ÷ 100) × 470 = 23.5 Ω

So you’d write 470 ± 24 Ω. Plenty of students see “± 5%” and write “± 5 Ω”, which is a different claim entirely and a wrong one by a factor of nearly five.

What Happens When You Combine Measurements

Most lab answers aren’t measured directly. You measure two or three things, then calculate. And the rule you apply depends entirely on what the calculation does.

Multiplying or dividing – add the percentages

If y = a × b   or   y = a ÷ b
then  % uncertainty in y  =  % in a  +  % in b

Example – density. You have m = 250.0 ± 0.5 g and V = 100.0 ± 1.0 cm³.

Mass:    (0.5 ÷ 250.0) × 100 = 0.2%
Volume:  (1.0 ÷ 100.0) × 100 = 1.0%
Density = 250.0 ÷ 100.0 = 2.500 g/cm³
% uncertainty = 0.2 + 1.0 = 1.2%
Absolute = 0.012 × 2.500 = 0.03 g/cm³

Result: 2.50 ± 0.03 g/cm³

The volume measurement contributed five times more doubt than the mass did. If you wanted to improve this experiment, a better balance would be almost pointless.

(At university level you’d combine these in quadrature instead √(0.2² + 1.0²) = 1.02% which gives a slightly smaller, less pessimistic figure. Simple addition is the standard school method and is deliberately cautious.)

Powers multiply the percentage by the exponent

If y = aⁿ   then   % in y = n × (% in a)

Example – volume of a sphere. Radius r = 2.50 ± 0.05 cm, and V = (4/3)πr³.

Radius: (0.05 ÷ 2.50) × 100 = 2%
Since r is cubed:  3 × 2% = 6%
V = (4/3)π(2.50)³ = 65.4 cm³
Absolute = 0.06 × 65.4 = 3.9 cm³

Result: V = 65 ± 4 cm³

A 2% wobble in the radius became a 6% wobble in the volume. Whenever a variable is raised to a power, it deserves your most careful measuring the formula amplifies its doubt.

Adding or subtracting – do NOT add the percentages

This is the one that catches people out, so read it twice.

If y = a + b   or   y = a − b
then work in ABSOLUTE uncertainties, and convert to a percentage at the end.

Example. Two lengths: 12.5 ± 0.1 cm and 8.3 ± 0.1 cm. What’s their total length, with percentage uncertainty?

The percentages are 0.8% and 1.2%. If you add those you get 2.0%, which on a total of 20.8 cm means ± 0.42 cm.

That’s wrong.

Do it properly:

Total = 12.5 + 8.3 = 20.8 cm
Absolute uncertainty = 0.1 + 0.1 = 0.2 cm
% uncertainty = (0.2 ÷ 20.8) × 100 = 0.96%

Just under 1%, not 2%. Adding percentages in an addition problem more than doubled the answer.

There’s a neat reason behind this. When you add measurements, the value grows along with the doubt, so the percentage tends to stay put or shrink. When you subtract, the value shrinks while the doubt still grows which is why subtracting two similar numbers is such a fragile way to measure anything. Subtract 10.02 ± 0.01 from 10.05 ± 0.01 and you get 0.03 ± 0.02, a percentage uncertainty of nearly 70%.

Where Students Lose Marks

Forgetting to multiply by 100. You calculate 0.0021 and write “0.0021%”. The correct answer is 0.21%. Easy slip, and it appears in exams every year.

Adding percentages when you should be adding absolutes. Covered above, but it’s the single most common structural error in this topic.

Confusing percentage uncertainty with percentage error. These sound similar and mean completely different things:

Percentage uncertainty = (Δx ÷ x) × 100
    → how precise your measurement was
Percentage error = (|accepted − experimental| ÷ accepted) × 100
    → how far your result was from the known value

Uncertainty is about the reliability of your method. Error is about whether your answer came out right. You can have a tiny uncertainty and a huge error that combination usually means a systematic problem, like an uncalibrated instrument.

Quoting too many digits. 0.2058% should be 0.21%. One or two significant figures is standard; anything more implies a precision you don’t have.

Dropping the % sign. Writing “0.21” when you mean “0.21%” changes the answer by a factor of a hundred.

Is your percentage uncertainty any good?

There’s no official scale, but as rough working guidance for school and undergraduate labs:

• Under 1% – a well-designed measurement
• 1–5% – normal and perfectly acceptable
• 5–10% – worth commenting on in your evaluation
• Above 10% – something in the method needs rethinking, not just repeating

If you land in that last band, resist the urge to simply take more readings. Look at the denominator instead. Can you measure something bigger? Time more swings, stack more coins, use a longer track. That’s almost always the faster fix.

Frequently Asked Questions

What is the formula for percentage uncertainty?

Divide the absolute uncertainty by the measured value, then multiply by 100. So a measurement of 24.30 cm with an absolute uncertainty of 0.05 cm has a percentage uncertainty of (0.05 ÷ 24.30) × 100 = 0.21%.

Do you add percentage uncertainties when adding measurements?

No, and this is the most common mistake in the topic. For addition and subtraction you combine the absolute uncertainties first, then convert the total to a percentage at the end. You only add percentages directly when you’re multiplying or dividing.

What’s the difference between percentage uncertainty and percentage error?

Percentage uncertainty describes how precise your measurement was, based on your instrument and method. Percentage error describes how far your result sits from an accepted or true value. A measurement can have very low uncertainty and still have a large error — which usually points to a systematic problem such as a zero-offset or an uncalibrated instrument.

Can percentage uncertainty be greater than 100%?

Yes, though it’s a sign something has gone wrong with the experimental design. It happens when the absolute uncertainty is larger than the measured value itself, which is common when you’re subtracting two similar numbers or measuring something close to your instrument’s resolution limit.

How do I reduce percentage uncertainty?

You have two levers. Reduce the absolute uncertainty by using a finer instrument or taking repeated readings, or increase the measured value so the same doubt becomes a smaller fraction of it. The second lever is usually cheaper and more effective timing twenty oscillations instead of one cuts your percentage uncertainty by a factor of twenty without buying anything.

How many significant figures should percentage uncertainty have?

One or two. Round 0.2058% to 0.21%, and 4.166% to 4.2%. Quoting more digits suggests a precision your data doesn’t support.

Is relative uncertainty the same as percentage uncertainty?

Almost. Relative uncertainty is the plain ratio Δx/x, while percentage uncertainty is that same ratio multiplied by 100. Be careful though some textbooks and syllabi use “relative uncertainty” loosely to mean the percentage version, so check what your course intends.

Conclusion

Percentage uncertainty is a one-line calculation that answers a question absolute uncertainty can’t: is this doubt actually a problem?

Divide, multiply by a hundred, and you have it. Where it gets interesting is in what the number then tells you. A 0.05% measurement and a 20% measurement can come from the same instrument on the same afternoon — the difference is entirely in what you pointed it at. Once you can see that, experimental design stops feeling arbitrary. Timing twenty swings instead of one isn’t a ritual your teacher invented. It’s the cheapest way there is to shrink a denominator.

The two things most worth carrying out of this: multiply and divide means add the percentages, add and subtract means don’t, and always check whether the question wants uncertainty or error. Those two distinctions account for the overwhelming majority of marks lost on this topic.

Next time you finish a lab, convert every measurement you took into a percentage before you calculate anything. The largest one is your experiment’s weak point, and you’ll have found it in about thirty seconds.