Locked lesson.

Upgrade

About this lesson

Two of the most important attributes of a measurement system are accuracy and precision. These concepts are explained in this lesson and a simple example is used to illustrate the difference. Also included is a brief discussion on the sources of accuracy and precision errors.

Exercise files

Download this lesson’s related exercise files.

Precision and Accuracy.docx
61 KB Precision and Accuracy - Solution.docx
61.2 KB

Quick reference

Precision and Accuracy 

Precision and accuracy are two components of the measurement system error. These components are often linked because their similar root causes for the error, although the error manifests itself differently.

When to use 

Whenever a measurement system is being used, it should be checked for precision and accuracy. Depending upon the characteristics of the system, this check should be done periodically.

Instructions

Precision and accuracy are two of the components of measurement system error. They have similar causes, but the impact they have on the measurement is different. 

 

Accuracy errors cause a shift in the measured result. On average, the results are higher or lower than the true value. The amount of this shift is called bias. This shift is often addressed through calibration, which attempts to adjust the measurement system so as to remove bias. 

 

Precision error is the common cause error of the system and reflects the normal variation in the results.  Precision error is usually categorized as repeatability error and reproducibility error.

  • Repeatability is often referred to as equipment variation. It addresses the question, “Can the same person, using the same measurement system, in the same way to measure the same item get the same answer?”
  • Reproducibility is often referred to as operator variation. It addresses the question, “Can a different person, using the same measurement system, in the same way to measure the same item get the same answer?”

Both of these precision error categories are evaluated using a Gage Repeatability and Reproducibility (R&R) Study. The impact of accuracy errors is to shift the average value of the measurement by the amount of the bias. The impact of the precision errors is to increase the variation from measurement to measurement. Although the impacts are different, they have similar causes.

  • Operator impact – in this case the error is due to changing operators. One operator’s technique may create a bias in the measurement that does not exist with a different technique (accuracy). Also, an operator’s training or skills may lead to different levels of variation and uncertainty in the use of the equipment (precision).
  • Equipment impact – some equipment may have a built-in bias due to system fabrication, assembly or calibration that is not found in other equipment (accuracy).  Also, the complexity of measurement equipment can lead to difficulties in setup or operation that create significant levels of uncertainty and variation in the measurement. This can be further compounded by variation in the system components due to wear and tear creating variation (precision).
  • Other environmental factors – these can impact both operators and equipment.  Environmental factors such as temperature may affect the operation of components and create a bias in equipment (accuracy).  Or poor lighting may make it difficult for an operator to correctly read a dial gauge (precision).

Hints & tips

  • It is important to understand the nature of the measurement error – accuracy or precision.  Chasing the variation of a poor precision like it is an accuracy problem is likely to lead to even greater errors due to over-control. A Gage R&R study will show you the level of precision errors. First attain precision, then with calibration attain accuracy.
  • Improving the impacts on measurement systems will often improve both accuracy and precision.
Login to download
  • 00:04 Hi, I'm Ray Sheen.
  • 00:06 Two of the categories of measurement error are precision and accuracy.
  • 00:10 Let's look at what each of those mean and their impact on the data measurement.
  • 00:15 I'll start by talking about accuracy.
  • 00:19 Accuracy is defined as whether the average of the measurements of an item reflect
  • 00:23 the true value of that item.
  • 00:25 To the extent that the average is different than the true value,
  • 00:28 we say that the system has a bias.
  • 00:30 The magnitude of the bias is considered to be the accuracy error
  • 00:34 inherent in the system.
  • 00:36 Asking if a system is accurate is the same as asking
  • 00:39 when I use these measurement system on average, do I get the right answer?
  • 00:43 If the answer is yes, then the system is considered accurate. But
  • 00:47 we're talking about on average, that means that on any specific instance, we may have
  • 00:52 a bias error but the average of those cancel out, we say the system is accurate.
  • 00:57 Accuracy errors are corrected by calibrating the measurement system.
  • 01:01 Through the calibration process, the bias error is reduced to near zero.
  • 01:06 So, let's switch gears and now look at Precision.
  • 01:09 Precision is the extent to which independent measurements made
  • 01:13 on the same entity result in the same data values.
  • 01:17 So a system that is precise, which means low precision error,
  • 01:21 gets virtually the same value with every measurement.
  • 01:24 Systems that are imprecise,
  • 01:26 which means high precision error, each measurement is different.
  • 01:30 Precision represents a common cause variation in the system.
  • 01:34 It's inherent in the measurement system design and the measurement process.
  • 01:38 Precision error is determined by conducting a Gage Repeatability and
  • 01:42 Reproducibility study, which is normally called a Gage R&R study.
  • 01:46 The results of the Gage R&R study will determine whether the measurement
  • 01:49 system is acceptable for the application.
  • 01:52 If the variation is too high, the system needs to be improved or
  • 01:56 a different system needs to be used.
  • 01:59 Since a Gage R&R study quantifies the precision error,
  • 02:02 let's look deeper at that topic.
  • 02:04 Precision error is normally allocated to one of two categories,
  • 02:08 repeatability error and reproducibility error.
  • 02:12 Repeatability error is sometimes called equipment variation.
  • 02:16 It answers the question, can the same person using the same measurement
  • 02:20 equipment, in the same way, to measure the same item get the same answer?
  • 02:25 This assessment minimizes the individual as a major component of the variation and
  • 02:30 focuses on everything else in the program.
  • 02:33 In the diagram shown, low repeatability has a much wider distribution and
  • 02:38 results, as compared to the distribution with high repeatability.
  • 02:42 Reproducibility error is sometimes called operator variation.
  • 02:46 It answers the question, can a different person, using the same measurement system,
  • 02:50 in the same way to measure the same item, get the same answer.
  • 02:54 This assessment minimizes the measurement equipment as a major component of
  • 02:58 variation and focuses on the effective changing operators.
  • 03:03 In the diagram shown, high reproducibility results
  • 03:06 in the operator distribution being almost identical.
  • 03:09 While low reproducibility has three distinct distributions of the three
  • 03:13 operators.
  • 03:14 The Gage R&R study will quantify the level of variation
  • 03:18 which makes up the precision error in the measurement system.
  • 03:22 These diagrams illustrate the difference between accuracy and precision.
  • 03:26 The two terms are often used interchangeably but
  • 03:29 they are very different with different effects.
  • 03:31 The target in the upper right is for a system that is accurate and precise.
  • 03:36 It has low accuracy error, everything is near the center of the target and
  • 03:40 low precision error, the grouping is very tightly clustered.
  • 03:44 The lower left corner is the opposite condition, it is inaccurate and imprecise.
  • 03:50 The high accuracy error is shown by the clustering of points down into the left.
  • 03:54 In addition,
  • 03:56 there is high precision error which is shown by a wide dispersion of data points.
  • 04:01 Upper left is precise but inaccurate.
  • 04:04 It is a type cluster of data points but off center.
  • 04:07 The lower right is accurate but imprecise.
  • 04:11 On average, the data points are on target but the spread is so
  • 04:15 large that any given point may be way off center.
  • 04:19 So let's now look at some of the sources of accuracy and precision error.
  • 04:23 Once we understand these sources, we can address them and
  • 04:26 reduce the measurement system error.
  • 04:28 Accuracy error is due to bias.
  • 04:31 One type of bias is operator bias.
  • 04:33 This is the case when each operator has a different magnitude of bias.
  • 04:38 This type of error can be reduced through training and through the measurements
  • 04:41 system designs that reduces the opportunity for operator interactions.
  • 04:46 Another bias is the instruments bias, this is normally addressed through
  • 04:50 the calibration of the equipment in the measurement systems.
  • 04:53 We include a third category of bias that is just a catch-all for any other effect.
  • 04:57 Normally, this will be environmental factors such as location or time of day.
  • 05:02 This is usually a very minor effect on accuracy of the measurement system.
  • 05:06 There are also several sources of precision error.
  • 05:09 The equipment can create variation, due to variation in the physical properties
  • 05:13 of the elements of the equipment.
  • 05:15 A typical one is the effective wear and
  • 05:17 tear on the component parts of the equipment.
  • 05:19 Another source of variation is setup variation.
  • 05:22 This is inherent in the design of the measurement system.
  • 05:26 Some measurement systems require complex set ups, which is prone to variations,
  • 05:30 of course, there's the operator variation,
  • 05:33 beyond the operator bias discussed above, some operators will just have higher
  • 05:37 levels of variation in their measurements than others do.
  • 05:41 This could be due to training, skills or technique.
  • 05:44 Finally, there are environmental factors that can create variation.
  • 05:47 As you can see,
  • 05:48 some of the sources of variation affect both accuracy and precision.
  • 05:52 This is one factor why the two errors effects are often confused.
  • 05:56 A particular error source could contribute to both.
  • 05:59 For instance, an operator technique when using a measurement system may lead to
  • 06:03 both a high variation in their measurement and to a bias in the average value.
  • 06:08 Accuracy and
  • 06:09 precision are two of the major categories of measurement system error.
  • 06:14 Understanding the source and magnitude of these errors will help us to determine
  • 06:18 if the measurement system is appropriate for its application.

Lesson notes are only available for subscribers.

PMI, PMP, CAPM and PMBOK are registered marks of the Project Management Institute, Inc.

Gift this course