What is uncertainty in measurement?

What Is Uncertainty in Measurement?

By Prof. Ada Sinclair·October 19, 2026·Related course

If you have ever used a ruler, stopwatch, thermometer, or digital sensor, you have already met one of the most important ideas in experimental physics: **uncertainty in measurement**. In everyday speech, “uncertainty” sounds like a mistake or a sign that we do not know what we are doing. In physics,

What Is Uncertainty in Measurement?

If you have ever used a ruler, stopwatch, thermometer, or digital sensor, you have already met one of the most important ideas in experimental physics: uncertainty in measurement. In everyday speech, “uncertainty” sounds like a mistake or a sign that we do not know what we are doing. In physics, it means something more precise and much more useful: every measurement has a limit to how exact it can be.

That does not mean measurements are useless. In fact, the whole practice of physics depends on measuring carefully and reporting honestly how confident we are in the result. Understanding uncertainty helps us compare data, judge whether two values really differ, and avoid pretending our measurements are more exact than they actually are.

Why Measurements Are Never Perfect

Suppose you measure the length of a pencil with a ruler marked in millimeters. You might report something like 14.2 cm14.2 \text{ cm}. But is it really exactly 14.2 cm14.2 \text{ cm}? Probably not. The pencil edge may fall somewhere between two marks, your eye may not line up perfectly, and the ruler itself may have tiny manufacturing imperfections.

This is the key idea: a measured value is an estimate, not an exact truth.

There are many reasons measurements have uncertainty:

  • The instrument has limited precision.
  • The person reading the instrument may make a small judgment error.
  • The object being measured may change slightly during the measurement.
  • Random environmental effects, such as vibration or temperature, may influence the result.

A good measurement report therefore includes both:

  1. the measured value
  2. the uncertainty

For example:

L=14.2±0.1 cmL = 14.2 \pm 0.1 \text{ cm}

This means the best estimate of the length is 14.2 cm14.2 \text{ cm}, and the true value is expected to lie within about 0.1 cm0.1 \text{ cm} of that estimate.

Absolute and Relative Uncertainty

Uncertainty can be described in two common ways: absolute uncertainty and relative uncertainty.

Absolute uncertainty

This is the uncertainty in the same units as the measurement itself.

Example:

m=52.0±0.5 gm = 52.0 \pm 0.5 \text{ g}

Here the absolute uncertainty is 0.5 g0.5 \text{ g}.

Relative uncertainty

This compares the uncertainty to the size of the measurement:

relative uncertainty=absolute uncertaintymeasured value\text{relative uncertainty} = \frac{\text{absolute uncertainty}}{\text{measured value}}

For the mass above:

0.552.0≈0.0096\frac{0.5}{52.0} \approx 0.0096

So the relative uncertainty is about 0.96%0.96\%.

This is useful because 0.5 g0.5 \text{ g} means something very different when measuring a 52 g52 \text{ g} object than when measuring a 5000 g5000 \text{ g} object. A small absolute uncertainty can still be large or small in relative terms, depending on the situation.

Random Error and Systematic Error

Not all uncertainty comes from the same source. In physics, we often distinguish between random error and systematic error.

Random error

Random error causes measurements to scatter unpredictably above and below the true value. If you repeat the same measurement many times, the results might look like this:

  • 10.1 s10.1 \text{ s}
  • 10.0 s10.0 \text{ s}
  • 10.2 s10.2 \text{ s}
  • 10.1 s10.1 \text{ s}

These small variations are random. They may come from reaction time, noisy sensors, or tiny changes in the environment.

The good news is that random errors can often be reduced by repeating measurements and averaging.

Systematic error

Systematic error shifts all measurements in the same direction. For example, if a stopwatch starts 0.2 s0.2 \text{ s} late every time, then every measurement is too large by about 0.2 s0.2 \text{ s}. If a ruler is stretched or incorrectly zeroed, every length measurement may be off.

Systematic errors are more dangerous than random errors because averaging does not remove them. You can repeat a bad measurement many times and still get the wrong answer.

A simple analogy: random error is like taking a few wobbly steps around the correct path; systematic error is like walking confidently in the wrong direction.

How to Estimate Uncertainty

There are several practical ways to estimate uncertainty, depending on the situation.

1. From the instrument resolution

If a ruler is marked in millimeters, a reasonable estimate might be half the smallest division. So a length might be reported as

L=14.2±0.05 cmL = 14.2 \pm 0.05 \text{ cm}

if you can estimate to the nearest half millimeter.

2. From repeated measurements

If you take several measurements, you can find the average and see how much the values vary. For example, if you measure the time for a pendulum five times, you might get:

  • 1.52 s1.52 \text{ s}
  • 1.49 s1.49 \text{ s}
  • 1.51 s1.51 \text{ s}
  • 1.50 s1.50 \text{ s}
  • 1.48 s1.48 \text{ s}

The spread gives you an idea of the uncertainty. In more advanced work, statisticians use the standard deviation, but at an introductory level it is enough to say: more scatter means more uncertainty.

3. From the measurement method

Sometimes the uncertainty is not just about the instrument. For example, timing a falling object by hand includes human reaction time. Even if the stopwatch displays milliseconds, your body may not react that precisely. In that case, the measurement uncertainty is larger than the display resolution alone suggests.

Reporting Measurements Properly

A good physics report should never present a number without context. Instead, it should tell the reader how sure you are.

Here are a few rules of thumb:

  • Round the uncertainty to one significant figure, or sometimes two if needed.
  • Round the measured value to the same decimal place as the uncertainty.
  • Include units every time.

Examples:

x=3.4±0.2 mx = 3.4 \pm 0.2 \text{ m} t=12.03±0.05 st = 12.03 \pm 0.05 \text{ s}

A measurement like 12.0317 s12.0317 \text{ s} without uncertainty can be misleading. It suggests a level of precision the experiment may not really have.

A Simple Example: Measuring the Width of a Book

Imagine you measure the width of a book with a ruler and get 21.3 cm21.3 \text{ cm}. The ruler has millimeter marks, so a realistic uncertainty might be 0.1 cm0.1 \text{ cm}.

You would report:

w=21.3±0.1 cmw = 21.3 \pm 0.1 \text{ cm}

This means the width is likely between 21.2 cm21.2 \text{ cm} and 21.4 cm21.4 \text{ cm}.

Now suppose you measure the same book five times and get:

  • 21.2 cm21.2 \text{ cm}
  • 21.3 cm21.3 \text{ cm}
  • 21.3 cm21.3 \text{ cm}
  • 21.4 cm21.4 \text{ cm}
  • 21.3 cm21.3 \text{ cm}

The average is about 21.3 cm21.3 \text{ cm}, and the spread is small. That tells you the measurement is fairly reliable. If the results had ranged from 20.8 cm20.8 \text{ cm} to 21.7 cm21.7 \text{ cm}, the uncertainty would be much larger.

Why Uncertainty Matters in Science

Uncertainty is not a weakness of science. It is one of its strengths.

Physics is not about claiming perfect certainty; it is about making careful, testable statements. When uncertainty is reported honestly, we can answer questions like:

  • Are two measurements actually different?
  • Does a result agree with a theory?
  • How much confidence should we place in a conclusion?

For example, suppose a theory predicts a mass of 5.0 kg5.0 \text{ kg} and your experiment gives

m=5.1±0.2 kgm = 5.1 \pm 0.2 \text{ kg}

That result is consistent with the theory because 5.0 kg5.0 \text{ kg} lies within the uncertainty range. But if you measured

m=6.3±0.2 kgm = 6.3 \pm 0.2 \text{ kg}

then the theory would be in serious trouble.

So uncertainty helps us decide whether differences are meaningful or just noise.

Common Misconceptions

“Uncertainty means the experiment failed.”

Not at all. Every real measurement has uncertainty. A result with uncertainty is more scientific, not less.

“If I repeat the experiment enough times, uncertainty disappears.”

Repeating measurements can reduce random error, but it does not eliminate systematic error. If your method is biased, more trials will not fix the bias.

“A smaller uncertainty always means a better answer.”

Usually yes, but only if the measurement is also accurate. A very precise result can still be wrong if there is a systematic error.

“Uncertainty is the same as being unsure.”

In physics, uncertainty has a technical meaning. It is a quantified estimate of how much a measured value may vary from the true value.

“The more digits I write, the more accurate my result is.”

Extra digits do not create accuracy. They only create the illusion of precision. Always match the number of digits to the

This article was generated by an AI physics persona for educational purposes. While we strive for accuracy, always verify important information with qualified instructors or academic sources.

← Back to Blog
Physics textbook with handwritten annotations

Want to learn the physics behind the articles?

Our blog articles are written by AI physics personas — the same guides available in the courses. Pick a course, choose your guide, and start a real conversation about physics.