In science, precision refers to how closely repeated measurements or results agree with each other. It has nothing to do with whether those measurements are correct. A set of measurements can be highly precise (tightly clustered together) yet completely wrong if they’re all off by the same amount. This distinction between precision and accuracy is one of the most important concepts in scientific measurement.
Precision vs. Accuracy
The classic way to understand precision is the bullseye analogy. Imagine throwing darts at a target. Accuracy describes how close your darts land to the center. Precision describes how close your darts land to each other. You can have one without the other.
If your darts cluster tightly in the upper left corner, far from the bullseye, your throws are precise but not accurate. If they scatter all over the board but average out near the center, they’re accurate but not precise. The ideal is both: a tight cluster right on the bullseye. In measurement terms, you want results that are repeatable and correct.
These two qualities are usually independent of each other. A bathroom scale that consistently reads 3 pounds too heavy is precise (it gives the same reading each time) but inaccurate (the reading is wrong). A scale that sometimes reads 1 pound over and sometimes 2 pounds under is imprecise but might average out to accurate. This independence is why scientists evaluate precision and accuracy separately.
What Causes Poor Precision
Precision is degraded by random errors, the unpredictable fluctuations that cause measurements to scatter. These might come from tiny vibrations in equipment, slight variations in how a person reads an instrument, temperature changes in a lab, or electrical noise in a sensor. Random errors are not repeatable, which is exactly why they spread results apart rather than shifting them all in one direction.
Systematic errors, by contrast, affect accuracy. A thermometer calibrated incorrectly will consistently read too high or too low. Every measurement shifts the same way. That kind of error doesn’t reduce precision at all because the results still cluster tightly. They’re just clustered around the wrong value. Understanding which type of error you’re dealing with determines whether you need a better technique (for precision) or a corrected calibration (for accuracy).
Repeatability and Reproducibility
Precision actually breaks down into two subcategories. Repeatability is the simpler one: when the same person uses the same equipment under the same conditions and gets consistent results close together in time. If you weigh a sample five times on the same balance in the same lab within an hour and get nearly identical values, your measurement has good repeatability.
Reproducibility is harder to achieve. It asks whether consistent results hold up across different operators, different equipment, or different laboratories. A measurement method with high reproducibility means any competent lab following the same protocol should get similar numbers. The international standard ISO 3534-1 defines precision as “the closeness of agreement between independent test results obtained under stipulated conditions,” and those conditions can range from tightly controlled (repeatability) to broadly varied (reproducibility).
This matters enormously for science as a whole. Limits to measurement precision are one of the key constraints on whether scientific results can be replicated by other research groups. If a finding depends on measurements that only one lab can reproduce, the precision of the method is suspect.
How Precision Shows Up in Numbers
When you see a measurement reported as 36.7 cm rather than 37 cm, those extra digits communicate precision. The three significant figures in 36.7 tell you the measuring tool was precise enough to distinguish tenths of a centimeter. The last digit in any reported measurement is understood to carry some uncertainty, so writing 36.7 means the true value is somewhere close to that, but the ones and tens digits are solid.
A more precise instrument yields more significant figures. A ruler marked in millimeters lets you report 36.72 cm (four significant figures), while a ruler marked only in centimeters limits you to 37 cm (two significant figures). When combining measurements in calculations, the result can only be as precise as the least precise input. If you multiply a four-significant-figure measurement by a two-significant-figure one, your answer gets two significant figures.
Scientists also quantify precision statistically. The most common measure is standard deviation, which captures how spread out a set of measurements is from their average. A small standard deviation means the values are tightly grouped, indicating high precision. When reporting the precision of an average value rather than individual measurements, researchers use standard error, calculated by dividing the standard deviation by the square root of the number of measurements. Taking more measurements shrinks the standard error, which is why repeating experiments improves confidence in the result.
Why Precision Matters in Practice
In everyday science, precision determines whether you can detect real differences between things. If your measurement tool has poor precision, the natural scatter in your readings might be larger than the actual difference you’re trying to measure. A scale precise only to the nearest gram can’t help you study a chemical reaction that produces half-gram changes.
In medical testing, precision determines whether a blood test can reliably distinguish between normal and slightly elevated levels of a substance. In manufacturing, it determines whether parts will fit together consistently. In climate science, it determines whether a temperature trend of a fraction of a degree per decade is detectable above the noise in the data.
Precision also sets the ceiling on what conclusions you can draw. NIST, the U.S. agency responsible for measurement standards, recommends that scientists express the uncertainty of their results using standard deviation rather than vague terms like “high precision” or “low precision.” Quantifying precision forces researchers to be honest about the limits of what their data can actually tell them, and it lets other scientists judge whether a finding is meaningful or could simply be an artifact of imprecise measurement.