Quick answer
Use percent error when you know the accepted (true) value: |measured − accepted| ÷ accepted × 100. Use percent difference when you compare two measurements and neither is “true”: |a − b| ÷ ((a + b) ÷ 2) × 100. A g of 9.62 m/s² against 9.81 m/s² is a 1.94% error.
Key takeaways
- Percent error → there is an accepted value. Divide by it.
- Percent difference → two equal measurements. Divide by their average.
- Both are usually reported as positive numbers.
- Lower is better: it means more accurate (error) or more consistent (difference).
On this page
Lab reports ask for one or the other, and they are easy to mix up. Here are both formulas, when to use each, and worked examples.
Percent Error CalculatorMeasured vs. accepted value.Open →The two formulas
percent error = |measured − accepted| ÷ |accepted| × 100
percent difference = |a − b| ÷ ((a + b) ÷ 2) × 100
Worked example: Percent error: gravity lab
- Measured g = 9.62 m/s²; accepted g = 9.81 m/s²
- |9.62 − 9.81| = 0.19
- 0.19 ÷ 9.81 × 100 = 1.94%
Worked example: Percent difference: two lab partners
- Partner A: 9.62 m/s²; Partner B: 9.70 m/s²
- Difference = 0.08; average = 9.66
- 0.08 ÷ 9.66 × 100 = 0.83%
Which one should I use?
| Question | Use |
|---|---|
| How close am I to the textbook value? | Percent error |
| How close are my two trials? | Percent difference |
| How much did a value change over time? | Percent change |
Signed percent error
Some teachers want the sign: (measured − accepted) ÷ accepted × 100. A negative answer means your measurement was too low. In the gravity lab that is −1.94%.
Accuracy first
Put the accepted value in the denominator for percent error. Dividing by your measured value gives a slightly different — and wrong — answer.
What NOT to do
- Don’t use percent difference when there is an accepted value.
- Don’t divide by the measured value in percent error.
- Don’t report a percent error without saying which accepted value you used.
Related: percent yield formula, percentage change vs difference, all science calculators.
Accuracy vs. precision
Percent error measures accuracy: how close you are to the true value. Percent difference between repeated trials measures precision: how close your results are to each other. A set of trials can be precise but not accurate — for example, three readings of 9.50, 9.52 and 9.51 m/s² agree closely (low percent difference) but are all about 3% below 9.81 m/s².
Where you’ll use this
Most school lab reports ask for percent error in the conclusion: you measured the density of aluminum, the speed of sound, or the value of g, and you compare it with the accepted value from a reference table. A low percent error supports your method; a high one is a prompt to discuss possible sources of error.
Percent difference is more useful when there is no reference value — for example, when two lab groups measure the same object, or when you repeat a trial. A small percent difference shows your results are consistent, even if you don’t yet know whether they are accurate. Many reports include both.
Practice problems (with answers)
Try each one, then check it with the Percent Error Calculator. More tools: all science and math calculators.
- Measured 2.65, accepted 2.7. What is the percent error? Answer: |2.65 − 2.7| ÷ 2.7 × 100 = 1.85% (low).
- Measured 101.2, accepted 100. What is the percent error? Answer: |101.2 − 100| ÷ 100 × 100 = 1.2% (high).
- Measured 330, accepted 343. What is the percent error? Answer: |330 − 343| ÷ 343 × 100 = 3.79% (low).
- Measured 0.52, accepted 0.5. What is the percent error? Answer: |0.52 − 0.5| ÷ 0.5 × 100 = 4% (high).
Frequently asked questions
What is a good percent error?
It depends on the experiment. In many school labs, under 5% is good and under 1% is excellent.
Can percent error be negative?
Not with the absolute-value formula. The signed version can be negative if your value was too low.
Is percent difference the same as percent error?
No. Percent error compares with an accepted value; percent difference compares two measurements using their average.
How do I lower my percent error?
Repeat trials and average them, use more precise instruments, read scales at eye level, and control the variables that affect your measurement.
Should I round percent error?
Report it to the same number of significant figures as your least precise measurement, usually two or three.
Sources
Spotted an error? Tell us and we will fix it and note the change here.