Accuracy is the degree to which an expression or measurement conforms to a true value. It also has to do with correctness or absence of error. The term is more tightly defined in certain fields, e.g. scientists distinguish it from precision.
According to Merriam-Webster, the term covers three meanings:
Correctness – as in lacking errors or mistakes
Exactness – as in agreeing with a standard or conforming to truth
Precision – as in the extent to which a measure conforms to a true value or standard
The following sentences give examples of use of the term.
There are several ways to improve location accuracy in Google Maps.
Sara scored high on factual accuracy and average on style in her history exam.
In measurement, high accuracy corresponds to low error.
The researcher tested the accuracy of the World Wide Web in answering some general knowledge questions.
To ensure accuracy, the scientists carried out the experiments three times.
Accuracy and reliability
Certain fields that rely on measurement – such as science and engineering – distinguish between accuracy and reliability. In these instances:
Accuracy is about how close the measurement is to the true value.
Reliability is about getting the same value each time you take the measurement.
Imagine that taking a measurement is like shooting at a target. The bullseye represents the true value that you are trying to measure. The following image illustrates the difference between the two terms:
Scientists also differentiate between accuracy and precision. Precision is to do with the degree to which a measurement is correct.
A clock that measures time in seconds might be accurate in telling the time. However, a stopwatch that can measure tenths of a second has greater precision.
Accuracy of medical tests
When talking about the accuracy of medical tests, doctors and researchers use the terms sensitivity and specificity. These measures tell us how well a test can identify whether an individual does or does not have a disease.
A highly sensitive test returns few false negatives, and so misses fewer true positive cases. A highly specific test returns few false positives.
To illustrate this, let us consider a test for a virus. In clinical trials, the test showed 99% sensitivity and 96% specificity. This means that, for every 100 people known to have the virus, we can expect the test to correctly identify 99 of them and miss just one.
Also, for every 100 without the virus, the test would likely correctly identify 96 as virus free and also incorrectly return four positive results.
In medicine, it is important to differentiate between these two measures of accuracy. For example, a test with low specificity would not make a useful screening tool. It would screen too many people without the disease as positive, causing them to undergo unnecessary diagnostic procedures.