Not sure which lesson to do first? Find your level and get a plan →

Math · Sets & Logic · Grade 9-12

Error Analysis

Error analysis asks not just 'what's my answer?


Quick Answer

Error analysis asks not just 'what's my answer?' but 'how much could it be off, and how does that uncertainty grow through the calculation?' Use it whenever inputs are measured or rounded: a length known to ±0.5 cm, a time read off a stopwatch, π rounded to 3.14. The result is an answer with an honest range, such as 40±4 cm², plus a sense of which input contributed most of the uncertainty.

Why This Matters

A number without its uncertainty can't be compared or trusted: is a measured 9.6 m/s² really different from 9.8 m/s² if each is ±0.3? Error analysis tells you which digits mean something, which measurement to improve, and when two results genuinely disagree. Without it, people report 12.73846 cm from a ruler read to the nearest millimetre, or decide an experiment failed when the gap lies inside the uncertainty.

Intuitive Explanation

Measure a table as 120 cm with a tape you can read to within 0.5 cm. The true length is somewhere from 119.5 to 120.5 cm, and every later calculation inherits that range. Cut the table's length into 4 equal parts and each is 30 cm, uncertain by ±0.125 cm: dividing a measured quantity by an exact number divides its uncertainty too.

Combining measurements makes uncertainty grow. Add two lengths, 50±0.5 cm and 30±0.5 cm, and in the worst case both errors point the same way: 80±1 cm. For a product it is the percentage errors that add. A rectangle measuring 10±0.5 cm by 4±0.2 cm is 5% uncertain in each direction, so its 40 cm² area is about 10% uncertain, ±4 cm². Squaring doubles the percentage, because r2 is r times r, so a 2% error in a radius becomes a 4% error in a circle's area.

None of this is about mistakes. Writing 7×8=54 is a blunder: you find it, fix it, and it's gone. Measurement uncertainty never disappears; it is part of every measured number, and the best you can do is shrink it with a better instrument and report it honestly. Significant figures are a rough version of the same idea: "120 cm" quietly suggests "to the nearest centimetre or so", while "120±0.5 cm" says exactly how far off it could be.

Think of it like this. A measured number is a slightly blurry photo: you can see it's about 120, but the edges are fuzzy by half a centimetre. Every calculation makes a copy of a copy, and adding or multiplying blurry numbers widens the blur, never sharpens it. Unlike a photo, you can state exactly how blurry the number is, and you should.

Core idea. Error analysis studies how errors arise, how big they are, and how they propagate through later steps.


When to Use

Use error analysis when inputs are measured, rounded or estimated and the question asks how accurate or reliable the result is, how many digits to report, or whether two results really differ. Add absolute uncertainties for sums and differences, add percentage uncertainties for products and quotients, and multiply a percentage by the power for squares and cubes. If a result is wrong because of a slip, find and fix the mistake; if you only need how much an output moves per unit of input, compute a sensitivity.

Tip. Ask: Am I quantifying how large my answer's error could be and how input errors grow through the steps?

How to Recognize It

Before computing, note for each input whether it was counted exactly, measured, rounded or estimated, and how far off it could be.

  1. Were any inputs measured with an instrument or rounded, rather than counted or exact by definition?

    If yes, each carries an uncertainty that the answer inherits, so attach one, such as ±0.5 cm. If every input is exact, like 12 eggs or the 100 cm in a metre, there is no measurement error to track.

  2. Is a wrong answer caused by a slip, such as a misread digit or a sign error?

    If yes, that is a mistake to find and fix, not uncertainty to analyse. Error analysis deals with the spread that remains even when every step is done correctly.

  3. Does the calculation multiply, divide or raise the measured values to powers?

    If yes, work in percentage uncertainties: a side measured to ±2% gives a cube's volume to about ±6%. For sums and differences, add the absolute uncertainties instead.

  4. Do you only need how much the output changes per unit of one input?

    If yes, that is a sensitivity. Error analysis goes further, combining the uncertainty of every input into a range for the answer; if you only need how many digits to report, significant figures give a quick version.

Error Analysis compared with Sensitivity (meta), Significant figures, and Mistake / blunder

Each row concerns why a result might differ from the truth, so they get lumped together. Ask what kind of difference it is: uncertainty carried in from imprecise inputs, the response to one input, a quick count of trustworthy digits, or a plain wrong step.

Error Analysis

Meaning
Use this when your inputs carry measurement or approximation uncertainty and you need how large the output error could be and how it grows.
Key test
Am I quantifying how large my answer's error could be and how input errors grow through the steps?
Example
A rectangle measures 10±0.5 cm by 4±0.2 cm. What's the area and its uncertainty?

Sensitivity (meta)

Meaning
How much output changes per unit input change; error analysis uses sensitivities to track ERROR SIZE through steps.
Key test
Use when you want the response rate to an input, not the total accumulated error.
Formula
ΔoutΔin
Example
Output shift per unit input change

Significant figures

Meaning
A shorthand for stating precision in a single number, a coarse output of error analysis.
Key test
Use when simply reporting how many digits are trustworthy.
Example
Writing 3.1 not 3.14159

Mistake / blunder

Meaning
A flat-out wrong step, not the inherent uncertainty of measurement and approximation that error analysis studies.
Key test
Use when the issue is a correctable slip, not unavoidable imprecision.
Example
Forgetting to carry a digit

Formula & Notation

Worked Examples

Example 1 — Area with measurement error

Easy

A rectangle measures 10±0.5 cm by 4±0.2 cm. What's the area and its uncertainty?
Show the steps and the answer
  1. Compute the best value: A=10×4=40 cm².

    The measured values give the centre of the range; the uncertainty is spread around it.

  2. Turn each uncertainty into a percentage: 0.510=5% and 0.24=5%.

    For a product, relative errors combine; absolute errors in different factors can't simply be added.

  3. Add the percentages: 5%+5%=10%, and 10% of 40 is 4 cm².

    In the worst case both lengths are high together, or both low, so their relative errors stack.

  4. Check with the extremes: 10.5×4.2=44.1 and 9.5×3.8=36.1.

    The true area lies between 36.1 and 44.1 cm², which is 40 give or take about 4.

Answer: 40±4 cm2

Takeaway: For a product, add the percentage errors and convert back: two 5% measurements give an area uncertain by about 10%.

Example 2 — A correctable mistake

Standard

A student computes 7×8=54. Is finding this an error analysis?
Show the steps and the answer
  1. Ask where the error comes from: 7 and 8 are exact whole numbers, not measurements.

    With exact inputs there is no uncertainty to carry, so error analysis has nothing to work on.

  2. Find the slip: 7×8=56, so the student is off by exactly 2.

    A blunder has a definite size and a definite fix; measurement uncertainty is a range that never goes away.

  3. Contrast with a measured version: a board 7.0±0.1 m long cut into 8 equal pieces gives 0.875±0.0125 m per piece.

    Here the ± survives any amount of careful arithmetic; tracking it is what error analysis is for.

Answer: 7×8=56, fix the blunder

Takeaway: Correct a blunder and it is gone; measurement uncertainty stays and has to be tracked and reported.

Example 3 — Find the mistake: too many digits

Application

A runner covers 100 m, measured to ±0.5 m, in a hand-timed 12.4 s, accurate to ±0.2 s. Dev writes: "Speed =100÷12.4=8.064516 m/s." Find the mistake and correct it.
Show the steps and the answer
  1. Find the percentage uncertainties: 0.5100=0.5% and 0.212.4≈1.6%.

    Speed is a quotient, so percentage errors are what combine.

  2. Add them: about 2.1%, and 2.1% of 8.06 m/s is about 0.17 m/s.

    The worst case pairs a longer distance with a shorter time, or the reverse, so the errors stack.

  3. Round to match: 8.06±0.17 m/s, or about 8.1±0.2 m/s.

    Dev's last four digits sit far inside the uncertainty, so they carry no information.

  4. Check with the extremes: 100.512.2≈8.24 and 99.512.6≈7.90.

    The true speed could be anywhere from about 7.9 to 8.2 m/s, consistent with 8.1±0.2.

Answer: Speed ≈8.1±0.2 m/s; the extra digits in 8.064516 claim precision the stopwatch can't give.

Takeaway: An answer can't be more precise than its least precise input; work out the uncertainty, then round to it.

Where You'll See It

  • The margin of error in a poll

    A poll of 1,000 voters has a margin of error of about ±3 percentage points (1.960.25/1000≈0.031). A 52% to 48% result means each candidate's true support could be roughly 3 points either way, so a 4-point lead is not a sure thing.

  • Buying fence for a measured garden

    A garden measured as 12 m by 8 m, each side to ±0.1 m, has perimeter 2(12+8)=40 m. The perimeter uses each measurement twice, so in the worst case it is off by 2(0.1)+2(0.1)=0.4 m. Buying exactly 40 m of fencing could leave a gap.

Common Mistakes

  1. The slip: Reporting an answer without its uncertainty

    Instead: pair the number with how far off it could be.

  2. The slip: Claiming more precision than inputs allow

    Instead: the output can't be more precise than its least-precise input.

  3. The slip: Ignoring how errors compound

    Instead: uncertainty grows through multiplication and repeated steps, it doesn't stay fixed.


Mini Practice

Try these on your own. Tap Reveal when you want to check.

  1. Two boards measure 2.40±0.05 m and 1.10±0.05 m. Laid end to end, what is the total length and its uncertainty?

    Hint: When you add lengths, the possible errors add too.

  2. A calculator gives 7.3÷3.1=2.35483871. Both inputs have 2 significant figures. How should you report the result?

    Hint: The answer cannot be more precise than the least precise input.

  3. A runner covers 100.0±0.5 m in 12.0±0.2 s. Find the speed and its possible range.

    Hint: The fastest case uses the longest distance over the shortest time; the slowest does the opposite.

  4. A cube's side is measured as 3.0±0.1 cm. Find its volume and the uncertainty in the volume.

    Hint: Compute the volume for the smallest and largest possible sides.

  5. Case A: a student copies 3.52 into a calculator as 3.25. Case B: a thermometer reading of 21°C is only good to within 0.5°C. Which case calls for error analysis?

    Hint: Which case has an uncertainty you can measure and carry forward?

  6. Tom measures a circle's diameter as 4.2 cm, to the nearest 0.1 cm, and reports its circumference as 13.194689 cm. Correct his answer.

    Hint: How far off could the diameter be, and how does that affect πd?

Want the full set?

50 practice questions for this concept — free to try, every one with a complete worked solution showing the why, not just the answer.

Frequently Asked Questions

Why do percentage uncertainties add when I multiply?

Suppose a length is 2% too big and a width is 3% too big. The area is then 1.02×1.03=1.05061.02 \times 1.03 = 1.0506 times the true area, about 5% too big. The leftover piece, 0.02×0.03=0.00060.02 \times 0.03 = 0.0006, is tiny, so adding the percentages is a very good approximation whenever the uncertainties are small. For large uncertainties, such as 30% and 40%, the leftover piece matters and the shortcut underestimates.

Why is my worst-case range bigger than the one in my physics textbook?

Adding uncertainties assumes every error pushes the same way at once, the worst case. Real errors are often independent, so some partly cancel. Many textbooks therefore combine them in quadrature, a2+b2\sqrt{a^2 + b^2} instead of a+ba + b. Two uncertainties of 3% and 4% give 7% in the worst case but 5% in quadrature, since 9+16=5\sqrt{9 + 16} = 5. Check which convention your course uses.

Does rounding in the middle of a calculation really change the answer?

It can, a lot. Round 2\sqrt{2} to 1.41 and square it: you get 1.98811.9881, not 2. Subtracting two close rounded numbers is worse: 1.41−1.40=0.011.41 - 1.40 = 0.01, but if each value is only good to ±0.005\pm 0.005, the true difference could be anywhere from 0 to 0.02, so the answer is uncertain by 100%. Keep extra digits until the final step, then round once.

What does the ± after a scientific measurement actually promise?

Usually not a guarantee. In many sciences, 9.81±0.029.81 \pm 0.02 means the true value is likely, often about 68% likely, to lie in that range, not certain to. Some fields quote twice that width, which covers about 95%. Engineering tolerances, such as a bolt made to 10.00±0.0510.00 \pm 0.05 mm, are closer to a promise, because parts outside the range are rejected. Always check which meaning is in use.

See Also