iSixSigma

Accuracy

Definition of Accuracy:

« Back to Glossary Index

All processes should have a target value you wish to achieve. How close you come to hitting that target will be the measure of your accuracy. Let’s explore this concept in a little more detail. 

Overview: What is accuracy? 

When speaking in terms of statistics, you will often hear the concepts of accuracy and precision used together. 

  • Accuracy can be defined as, on average, how far your measurements or results are from your target. In other words, accuracy is the extent to which the average of the measurements deviate from the true value. 
  • Precision relates to how consistent you are. Assuming the center bullseye is your target, the graphics below show what accuracy and precision would look like.

Most Six Sigma professionals will agree that achieving precision should be the first goal of any improvement activity, and then you can worry about the accuracy. 

For example, the average of these values will equal 5: 0, 0, 0, 0, 10, 10, 10, 10. So will these values: 4, 4, 4, 4, 6, 6, 6, 6. The second set of values have much less variation and greater precision. This will allow for great predictability. The goal should be to reduce the variation of the first set of values.

Dr. Ken Feldman is fond of saying, “Accuracy in the absence of precision is meaningless.” In statistics, the mean (or average) is often used as a measure of accuracy, and the range (or standard deviation) a measure of precision.

An industry example of accuracy 

The weights of raw material bags seemed to be lower than expected. While the weights from bag to bag showed little variation, the average was lower than the target weight provided by the vendor. 

After doing an MSA study, the company Black Belt discovered the scale had experienced some drift and was now consistently weighing three pounds lighter. After adjusting and recalibrating the scale, the bag weights were now accurate and precise.  

Frequently Asked Questions (FAQ) about accuracy

1. What is accuracy? 

The degree to which the average of measurements differs from the target or true value of what you are measuring. 

2. Which is more important, accuracy or precision? 

Precision is more important. Accuracy is the degree to which the measurements center around your target or true value. If the precision of those measurements is low, the accuracy doesn’t mean much since the variation will hinder your ability to predict. 

3. Is a lack of accuracy easy to fix? 

Generally yes. If your outcomes are not centered on your target, accuracy can sometimes be achieved by simply shifting the process mean. For example, if your weighing process is consistently off by 0.5 pounds, you can adjust the scale, which should result in more accurate weights.

« Back to Dictionary Index