Accuracy vs Precision: What's the Difference in Chemistry?
Learn the difference between accuracy and precision in chemistry measurements, with the classic target model, systematic vs random error, and percent error calculations.
Quick answer
Accuracy is how close a measured value is to the true (accepted) value, while precision is how close repeated measurements are to each other. A measurement can be precise without being accurate, accurate without being precise, both, or neither.
Unit: percent (%)
Example: A student measures the density of water as 0.98 g/mL when the accepted value is 1.00 g/mL. The percent error is |0.98 − 1.00| ÷ 1.00 × 100% = 2%, so the measurement is accurate to within 2% of the true value.
What you will learn
- Define accuracy and precision using the target model
- Distinguish between systematic and random error
- Calculate percent error for a measurement
- Explain how repeated trials improve precision
- Identify common sources of error in laboratory measurements
Why the difference matters
Every chemistry measurement is a little bit wrong. No balance, thermometer, or graduated cylinder gives a perfectly exact value. Analytical chemists therefore ask two separate questions about any set of measurements: how close are they to the truth, and how close are they to each other? The first question is about accuracy; the second is about precision.
Confusing the two leads to bad conclusions. A laboratory can produce beautifully consistent numbers that are all wrong, or a single lucky reading that happens to hit the true value while the method itself is sloppy. Keeping the concepts separate is the first step toward trustworthy data.
What accuracy means
Accuracy describes how close a measured value comes to the true value — the value accepted as correct, such as a known density, a certified reference standard, or a theoretically calculated result. A measurement is accurate when it lands near the truth.
Accuracy is usually reported with percent error, which expresses the size of the mistake relative to the true value:
A percent error of 0% means the measurement matched the true value exactly. In school laboratories, a percent error under about 5% is generally considered good accuracy, though professional analytical work demands far better.
What precision means
Precision describes how close repeated measurements of the same quantity are to each other, regardless of whether they are near the true value. If you weigh the same coin five times and get 5.01 g, 5.02 g, 5.01 g, 5.02 g, and 5.01 g, your technique is precise — the readings cluster tightly.
Precision reflects the reproducibility of a measurement. It is limited by random variation: tiny fluctuations in how you read a meniscus, slight drafts on a balance, or small differences in reaction timing. Taking more trials and averaging them improves precision because random ups and downs tend to cancel out.
You cannot judge precision from a single measurement. Precision is a property of a set of measurements, so it only appears when you repeat the experiment.
The target model
The classic way to picture the two ideas is an archery target, where the bullseye is the true value and each arrow is a measurement:
- Accurate and precise: arrows cluster tightly in the bullseye. This is the goal of good analytical work.
- Precise but not accurate: arrows cluster tightly, but off-center. Something is systematically skewing every shot — the classic sign of a calibration problem.
- Accurate but not precise: arrows scatter around the bullseye. The average may land near the truth, but individual measurements are unreliable.
- Neither accurate nor precise: arrows scatter widely and off-center. Both the method and its execution need work.
Notice the most dangerous case is the second one: precise-but-wrong data looks convincing because the numbers agree with each other.
Precise but inaccurate results are the hardest errors to catch, because agreement between trials creates false confidence. Only a comparison against a known true value — a standard or reference — reveals the problem.
Systematic vs random error
Errors fall into two families, and each one attacks a different half of the accuracy–precision pair.
Systematic error shifts every measurement in the same direction by a roughly constant amount. A balance that reads 0.05 g too high, a thermometer miscalibrated by the manufacturer, or always reading a burette from above the meniscus (parallax) are systematic errors. They damage accuracy while leaving precision untouched — the readings still cluster, just around the wrong value. Because the shift is consistent, systematic error can be found and corrected by calibration against a known standard.
Random error scatters measurements unpredictably above and below the true value. Estimating between scale marks, slight variations in drop size from a pipette, or air currents around a sensitive balance are random errors. They damage precision. Random error cannot be calibrated away, but repeating measurements and averaging reduces its influence.
Improving your measurements
Good laboratory technique targets both problems at once:
- Calibrate instruments against known standards to remove systematic error and protect accuracy.
- Repeat trials (at least three) and average them to shrink random error and improve precision.
- Read scales at eye level to avoid parallax error with burettes, graduated cylinders, and thermometers.
- Estimate one digit beyond the smallest scale mark, and record it — that estimated digit is a legitimate significant figure.
- Control conditions such as temperature and sample handling so that trials are truly comparable.
Before trusting any experimental result, ask both questions: is it close to the accepted value (accuracy) and would I get the same number if I did it again (precision)? Data you can answer yes to twice is data you can build on.
Variables and units
| Symbol | Name | Unit | Meaning |
|---|---|---|---|
| measured value | same as the quantity | The value obtained from the instrument or experiment. | |
| true (accepted) value | same as the quantity | The accepted or theoretically correct value used as the reference. |
Common mistakes
Precision is only repeatability. A miscalibrated balance gives tightly clustered readings that are all wrong by the same amount — precise, but inaccurate.
An average can land near the true value even when individual measurements scatter widely. A good average does not guarantee that any single reading was accurate.
Better equipment reduces random error, but it can never remove it entirely. Random error is managed by repeating trials and averaging, while systematic error is corrected by calibration.
A single measurement can hit the true value purely by luck while the method has poor precision. One lucky bullseye does not prove the archer is skilled — repeat the measurement to find out.
Real-world applications
Quality control laboratories verify that product measurements stay within accuracy limits before batches ship
Pharmaceutical manufacturing depends on precise, accurate dosing measurements for patient safety
Environmental monitoring agencies track pollutant concentrations against legal accuracy requirements
Forensic laboratories must document measurement precision when results are presented as evidence
Frequently asked questions
What is the difference between accuracy and precision in one sentence?
Accuracy is how close a measurement is to the true value, while precision is how close repeated measurements are to each other — you can think of accuracy as hitting the bullseye and precision as hitting the same spot every time.
Which is more important, accuracy or precision?
Both matter, and which one limits you depends on the task. Precise-but-inaccurate data misleads quietly, so accuracy is usually checked first with a known standard; then precision is tightened with repeated trials and better technique. Reliable science needs both.
How do you improve precision?
Repeat the measurement several times and average the results, use instruments with finer scale divisions, estimate consistently between marks, and keep experimental conditions the same across trials. Averaging works because random errors scatter in both directions and partially cancel.
Can a measurement be accurate but not precise?
Yes. If repeated measurements scatter widely but their average lands near the true value, the average is accurate while the individual measurements are imprecise. This is common with noisy instruments — the mean is trustworthy, but no single reading is.
What does percent error actually tell you?
Percent error tells you how large your mistake is relative to the size of the true value, which makes it comparable across different measurements. A 0.1 g error matters little when weighing 100 g (0.1%) but is huge when weighing 0.2 g (50%).