Precision Versus Accuracy

accuracy-vs-precision

Precision and accuracy are two very interesting concepts which are very often misunderstood.

To understand the concept of quality management better it is very important for you to understand these two quality management terms, which are often thought to be synonyms for each other.

There is another pair of terms which is also as important as precision and accuracy: quality and grade. We will discuss these concepts in my next blog post.

These topics are also very important from a PMP exam point of view; therefore you must know the definition of these concepts and understand the difference between them.

Anyway, let’s get started.

Precision

The definition of precision says that precision is when the values of repeated measurements are clustered and have little scatter.

If you get all measurements very precise, it does not mean that the measurements are close to the target value. It means that the measurements are only close to one another, and they may or may not be near the target value.

Precision is about how the measured values are close to one another. If there is less scatter, measurements are said to have a high precision.

For example, let’s say that you got an order to supply 10,000 nos of rods to your client with a length of 10 meters. You started producing rods, and during quality inspection you randomly take the measurements of five rods.

The length of each rod is as follows:

  • First Rod:        10.490 meters
  • Second Rod:   10.495 meters
  • Third Rod:      10.500 meters
  • Fourth Rod:    10.505 meters
  • Fifth Rod:       10.510 meters

If you analyze these measurements, you will notice that although these measurements are not close to the target (10 meters), they are very close to one another. There is very little difference in the measurements among them, so you can say that the scatter is very small.

So, in this case you will say that the measurements are very precise.

Precision is a measure of variation among the measured values.

Accuracy

Accuracy is defined as how the measured values are close to the target value.

If someone says that the measurements are accurate, then you should understand that those measurements are very close to the target value. Scatter doesn’t have any significant role here. The scatter of accurate measurements may or may not be close or dense.

For example, let’s take the same example again. This time you find the measured length of each rod as follows:

  • First Rod:         9.900 meters
  • Second Rod:    9.800 meters
  • Third Rod:       10.000 meters
  • Fourth Rod:     10.100 meters
  • Fifth Rod:         10.200 meters

You can clearly see that all measurements are very close to the target length of the rod, which is 10 meters. You can also see that although the values are closer to the target value, they are not closer to each other and scatter is high.

So you can say that these measurements are accurate but not precise.

You may now wonder which one is desirable during a quality control inspection.

The answer is “accuracy”. This is because in accuracy, all data are close to the true value which is a sign that the deliverables are of the right quality. However, if the measurements are precise as well, it is the best because in this case the measurements are close to the target value and very close to each other.

The Difference between Accuracy and Precision

The following are a few differences between accuracy and precision:

  • Accurate data need to very close to the target value, but precise data need not be close to the target value. Precise data only need to be close together.
  • For quality purposes, accuracy is desired while precision is not. Precision is required when it is coupled with accuracy.
  • Accurate data can be precise but precise data may or may not be accurate. Precision is not dependent on accuracy.
  • For accuracy, one measurement is enough; however, for precision many measurements are required.

Summary

Precision and accuracy are two different terms in quality management which are very often misunderstood as synonyms of each other; however they are not. It is not necessary for precise measurements to be accurate, or accurate measurements to be precise.

Precise measurements can be accurate or not, and vice versa. During the quality inspection of the deliverables, it is the responsibility of the project management team to decide the level of accuracy and precision for their project deliverables.

Here is where this blog post on precision versus accuracy completes; I hope I made the points clear to you. If you have something to share, you can do so through the comments section.

Image credit => NOAA’s National Ocean Service.

Comments

  1. Sam says

    Good examples and explanation, but can you please elaborate the below:

    in your first example where results are 10.490, 10.495, 10.500…. can we say it is both accurate and precise as result is close to 10.

    Similarly in the second example, 9.8, 9.9 and so on… can we still say it is precise and accurate??? as again values are less scattered (in the given sample) and close to 10?

    Also, i will appreciate if you can clarify another question. (based on Oliver Lehmann’s questions)… Qs.9 from 175 questions http://www.oliverlehmann.com/contents/free-downloads/175_PMP_Sample_Questions.pdf

    the options are “process should be adjusted” or improved… the correct answer goes with adjusted…

    i want to know if there is any rule of when to adjust or improve? this is not clear at all and i can’t find any good reference :(

    • Fahad Usmani says

      In my first example, all values are close to 10.5 meters means they are precise (note that, required length is 10.00 m). Though they are not close to the target length.

      In second example, values are moving around 10.0 meters, meaning they are close to the target length but not very precise (they are not very close to each other).

    • Fahad Usmani says

      If the process is precise but not accurate you should adjust it but if the process is neither accurate nor precise you should improve it.

Leave a Reply

Your email address will not be published. Required fields are marked *