Percent Error Calculator
Percent error tells you how far a measurement is from the accepted value, as a percent of that value: |measured − accepted| ÷ |accepted| × 100. If you measured 23.4 g and the accepted value is 25.0 g, the error is 1.6 g, and 1.6 ÷ 25 × 100 = 6.4% (see the steps).
Find the percent error
Percent error = |measured − accepted| ÷ |accepted| × 100. The answer updates as you type.
Accepted is the true or textbook value; measured is what you got. Add a unit if you like (23.4 g), or paste the whole question.
Separate measurements with commas, spaces or new lines, or paste a column from a spreadsheet (one value per line, so 1,250 stays one number).
For two measurements when neither one is the true value. Percent difference = |first − second| ÷ average × 100.
A percent error hides whether the measurement was too high or too low, so you get both answers.
Check your input
On the number line
Each trial
| Trial | Value | Error | % error | Direction |
|---|
How to Use This Calculator
-
Pick what you have
One measurement, several trials, two measurements to compare, or a percent error to work back from.
-
Type the values
Decimals, negatives and numbers like 6.02e23 work. Add a unit if you like, or paste the question from your worksheet.
-
Read the answer
You get the percent error with the absolute and signed error, a number line, and every step for your report.
What is percent error?
Percent error measures how close an experimental result is to the value it should have been. The accepted value (also called the theoretical, true or literature value) comes from a textbook, a data table or a calculation. The measured value (also called the experimental or observed value) is what you got. Percent error states the gap between them as a percent of the accepted value, so a 1 g miss on a 10 g sample counts for more than a 1 g miss on a 1,000 g sample.
The percent error formula
The formula comes down to three steps:
- Subtract: measured − accepted. This is the error, in the same unit as your values.
- Drop the sign and divide by the accepted value. This is the relative error, a plain number.
- Multiply by 100 to turn it into a percent.
The bars | | mean absolute value: you ignore the minus sign, so percent error is never negative. The accepted value goes on the bottom. Dividing by your own measurement instead is the most common slip, and it gives a slightly different answer.
Example: a chemistry mass
You expect 25.0 g of product and weigh 23.4 g. The error is 23.4 − 25.0 = −1.6 g, so the measurement is 1.6 g too low. Then 1.6 ÷ 25.0 = 0.064, and 0.064 × 100 = 6.4%.
Show the chemistry example in the calculator
Example: gravity in physics
A pendulum experiment gives g = 9.62 m/s², against the accepted 9.81 m/s². The error is 0.19 m/s², and 0.19 ÷ 9.81 × 100 ≈ 1.94%. This one does not come out even, so the calculator shows ≈ 1.937% and keeps four significant figures.
Show the gravity example in the calculator
Example: a worksheet question
"Experimental value = 1200 ml, accepted value = 800 ml." The error is 1200 − 800 = 400 ml, and 400 ÷ 800 × 100 = 50%. Percent error can go above 100%: a measurement of 2000 ml would be 150% off. Paste a line like this into the calculator and it fills both fields.
Show the worksheet example in the calculator
Signed percent error: too high or too low
The standard formula throws away the direction of the error. Some lab reports ask you to keep it:
A negative result means you measured too low, and a positive one means too high. In the chemistry example the signed error is −6.4%. The sign is useful when you discuss sources of error: if every trial comes out low, something is pushing your results one way.
Percent error with several trials
When you repeat a measurement, there are two common ways to report the error, and they give different numbers. Take five boiling points of water against the accepted 100 °C: 99.5, 100.2, 99.8, 100.5 and 99.1 °C.
- Error of the mean: the mean is 99.82 °C, so the percent error is 0.18%. This is the usual figure for a lab report, because averaging cancels out some of the random scatter.
- Average of the trial errors: the five trials are off by 0.5%, 0.2%, 0.2%, 0.5%, 0.9%, which average to 0.46%. This is larger, because errors in both directions all count as positive.
The calculator gives both, the best and worst trial, and the standard deviation. A small error of the mean with a large spread means you were accurate on average but not precise.
Show the five boiling points in the calculator
Percent error vs percent difference
Percent difference (sometimes called percent discrepancy) is for two measurements when neither one is the right answer, such as two lab groups measuring the same thing. With no accepted value to divide by, you divide by their average:
For 4.6 and 5.0 the gap is 0.4 and the average is 4.8, so the percent difference is 0.4 ÷ 4.8 × 100 ≈ 8.33%. It gives the same answer whichever value you call a. Percent error does not: the accepted value always goes on the bottom.
Show 4.6 and 5.0 in the calculator
Error, difference, relative error and percent yield
These terms turn up in the same labs and are easy to confuse:
| Measure | Formula | 25.0 g accepted, 23.4 g measured |
|---|---|---|
| Absolute error | |measured − accepted| | 1.6 g |
| Relative error | absolute error ÷ |accepted| | 0.064 |
| Percent error | relative error × 100 | 6.4% |
| Percent difference | |a − b| ÷ average × 100 | 6.61% |
| Percent yield | actual ÷ theoretical × 100 | 93.6% |
Percent yield and percent error use the same two numbers. When the yield is below 100%, they add up to 100%: a 93.6% yield is a 6.4% error.
Absolute error and relative error
The absolute error is the plain gap, in the same unit as the measurement. It tells you how far off you were, but not whether that matters. Being 1 g off when weighing a 10 g sample is a 10% error. Being 1 g off on a 1,000 g bag of flour is a 0.1% error. The absolute error is the same; the relative error is a hundred times smaller.
Relative error is the absolute error divided by the accepted value, a plain number with no unit. Percent error is the same thing times 100. Scientists often quote relative error in engineering notation, such as 2 × 10⁻³, while school reports almost always ask for the percent.
Percent yield and percent error together
Chemistry labs often ask for both. Take a reaction with a theoretical yield of 25.0 g. If you collect 23.4 g, the percent yield is 23.4 ÷ 25.0 × 100 = 93.6%, and the percent error is 6.4%. If the product is still damp and weighs 26.1 g, the yield comes out at 104.4%, which is impossible for a dry product, and the percent error is 4.4%. A yield over 100% almost always means something extra was weighed: water, solvent or an impurity.
Temperatures: the scale changes the answer
Percent error divides by the accepted value, so it depends on where the scale puts zero. Measure water boiling at 99.5 °C instead of 100 °C and the error is 0.5%. Do the same in kelvin, 372.65 K against 373.15 K, and it is only 0.134%, even though the thermometer was off by the same half degree. For an accepted value of 0 °C, percent error cannot be worked out at all. When a lab asks for percent error on a temperature, check which scale it expects; when in doubt, kelvin is the one that starts at a true zero.
Rounding the answer
Round the percent error at the end, never part way through. Rounding the relative error 0.019368 to 0.02 before multiplying turns a 1.94% error into 2%. A percent error usually needs no more than two or three significant figures, because the error itself is rarely known better than that. The calculator keeps every digit until the last step and marks a rounded figure with ≈.
Working backwards from a percent error
Some questions give the percent error and ask for a value: "Your experiment had a 12% error. Your measured value is 10.1 g. What is the theoretical value?" Write the error as a decimal, 0.12. The percent error does not say whether the measurement was too high or too low, so there are two answers:
- Too high: measured = accepted × 1.12, so accepted = 10.1 ÷ 1.12 ≈ 9.02 g.
- Too low: measured = accepted × 0.88, so accepted = 10.1 ÷ 0.88 ≈ 11.48 g.
If the question tells you which way the result went, keep that answer. If you know the accepted value instead, multiply: accepted × (1 + error) and accepted × (1 − error) give the two possible measurements.
Show this question in the calculator
What is a good percent error?
No single number counts as good everywhere, and no official standard sets one. What you can expect depends on the equipment, the method and the size of the quantity. The table is a common rule of thumb for school labs, not a rule from any one textbook:
| Percent error | Usually means |
|---|---|
| under 1% | Very close; typical of calibrated instruments and careful technique |
| 1% to 5% | Good for most school experiments |
| 5% to 10% | Acceptable for rough methods; worth discussing the causes |
| over 10% | Check the method, the readings and the arithmetic for a mistake |
Your lab manual or teacher sets the real target. A 3% error measuring g with a hand-timed pendulum is a fine result; the same 3% from a digital balance would mean something is wrong.
Where the error comes from
Errors come in two kinds, and the trials view helps you tell them apart.
- Systematic errors push every result the same way: a balance that reads 0.2 g heavy, a thermometer that is never zeroed, heat lost to the air in every run. Repeating the experiment does not cancel them. Look for trials that all fall on one side of the accepted value.
- Random errors scatter results both ways: reaction time on a stopwatch, reading a scale between marks, small draughts. Averaging more trials shrinks their effect on the mean.
This is the difference between accuracy and precision. Accuracy is how close you are to the accepted value, which percent error measures. Precision is how close your trials are to each other, which the standard deviation measures. Results can be precise but not accurate (tightly grouped around the wrong value) or accurate on average but not precise.
Writing it up in a lab report
State the accepted value and where it came from, show the formula with your numbers in it, and round the answer sensibly, usually to two or three significant figures. Then say whether the result was too high or too low and name a likely cause that fits that direction. "6.4% low, probably product lost while filtering" says more than "6.4% error due to human error".
Frequently Asked Questions
What is the formula for percent error?
Percent error = |measured − accepted| ÷ |accepted| × 100. Subtract the accepted value from your measured value, drop the minus sign, divide by the accepted value and multiply by 100. For 23.4 g measured against 25.0 g accepted, that is 1.6 ÷ 25 × 100 = 6.4%.
What is a good percent error?
There is no fixed standard; it depends on the experiment and the equipment. A common rule of thumb in school labs is that under 5% is good and under 10% is acceptable, and careful work with calibrated instruments can land under 1%. Your teacher or lab manual sets the real target.
Is 20% percent error bad?
For most school lab measurements, yes, it is high. An error that large usually points to a mistake in the method, a misread instrument or a systematic error such as an uncalibrated scale. For rough experiments, like timing a fall with a hand-held stopwatch, 20% can still happen.
Can percent error be negative?
Not in the usual formula, because it takes the absolute value of the error. Some courses keep the sign on purpose: (measured − accepted) ÷ accepted × 100 is negative when the measurement is too low and positive when it is too high. The calculator shows both.
What is the difference between percent error and percent difference?
Percent error compares a measurement with an accepted value and divides by that accepted value. Percent difference compares two measurements when neither one is the true value, and divides by their average. For 4.6 and 5.0 the percent difference is about 8.33%.
How do you find the accepted value from percent error?
Divide the measured value by (1 + error) if the measurement was too high, or by (1 − error) if it was too low, with the error as a decimal. With a 12% error and a measured 10.1 g, the accepted value is about 9.02 g or 11.48 g.
What if the accepted value is zero?
Then percent error is not defined, because you would divide by zero. Report the absolute error instead, the plain difference in units, such as 0.3 °C.